8839.pdf

8839_pdf.txt

PAGE 1

Technology and the American Economic Transition: Choices for the Future May 1988 NTIS order #PB88-214127

PAGE 2

Recommended Citation: U.S. Congress, Office of Technology Assessment, Technology and the American Economic Transition: Chokes for the Future OTA-TET-283 (Washington, DC: U.S. Government Printing Office, May 1988). Library of Congress Catalog Card Number 87-619892 For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, DC 20402-9325 (order form can be found in the back of this report)

PAGE 3

Foreword In many studies during the past decade and a half, the Office of Technology Assessment has analyzed hundreds of changes in American life that have been brought about by new technologies. In this study, for the first time, OTA steps back from the detailed analyses of individual industries and sectors to take a broad look at the combined impact of new technologies on American society. From this perspective it is possible to describe the opportunities and risks for the nation as a whole that are not apparent in studies targeted on topics that are more narrowly defined. The report highlights strategic choices available to Americans as we negotiate a period of major transformation. The choices we make will have profound consequences for the quality of work and the amenities available to Americans and for Americas role of leadership in the free world. Eight committees of Congress asked the Office of Technology Assessment to assess the new environment in which Americas economy now operates as a result of new technologies, a global market, and related changes in consumer attitudes and behavior. OTA was asked to identify areas where existing policy might block attractive avenues of growth and where new policies could facilitate growth. The requesting committees represent a wide spectrum of congressional interest. They include the Senate Committee on Commerce, Science and Transportation; the Committee on the Budget; the House Committee on Energy and Commerce; the Committee on Post Office and Civil Service; the Committee on the Judiciary; the Committee on Public Works and Transportation; the Committee on Education and Labor; and the Committee on Science Space and Technology. The analysis begins and ends by focusing on people in their role as consumers and as employees. It uses conventional economic accounting procedures to document economic growth, but also employs more qualitative standards for measuring progress in eight basic categories of demand or amenity: food, housing, transportation, health, clothing & personal care, education, personal communication & business, and recreation & leisure. It also uses standard methods for measuring gains in compensation paid to workers, but introduces other ways of evaluating job quality such as opportunities for learning and career advancement, the extent to which work and family responsibilities can be combined, and whether a person can take pride in his or her work. The first chapter serves as an introduction to the themes developed in the document and a summary and guide to the entire report. Readers will find this a useful way to identify sections on consumption, business structure, trade, employment, education, or other topics that may be of special interest to them. We feel that the major accomplishment of this document, however, lies in the way it helps develop a perspective on these issues showing how the networks of production and consumption are interconnected, how international and domestic economies are connected, and how changes in one area spread through the nations entire economic and social fabric. The study was undertaken with the help of many individuals and institutions around the United States. We owe a particular debt to individuals in the U.S. Department of Commerce and the U.S. Bureau of Labor Statistics who not only provided data and reviewed OTAs work but gave us crucial insights and guidance about their complex resources. Responsibility for the contents of this document, of course, rests with OTA. Director ,.. Ill

PAGE 4

Technology and the American Economic Transition Advisory Panel David Saxon, Chairman Chairman of the Corporation, Massachusetts Institute of Technology Cambridge, MA Claude Ballard John J. McNamara Partner McCann Erickson Goldman Sachs New York, NY New York, NY Kathleen OReilly, Esq. William Baumol Executive Director Department of Economics Citizens Utility Board Princeton University Madison, WI Princeton, NJ Harvey Brooks John F. Kennedy School of Government Harvard University Cambridge, MA Richard Crowder Senior Vice-President Pillsbury Co. Minneapolis, MN Thomas G. Denomme Vice-President Chrysler Corp. Highland Park, MI Charles F. Sabel Department of Science, Technology and Society Massachusetts Institute of Technology Cambridge, MA George M. Scalise Senior Vice President Advanced Micro Devices Sunnyvale, CA Al Sobey Director of Energy and Advanced Product Economies General Motors Corp. Detroit, MI Judith Gregory Barbara Starfield Department of Communications School of Hygiene and Public Health University of California at San Diego John Hopkins University La Jolla, CA Baltimore, MD Henry A. Lichstein Arthur G. Wirth Vice President Department of Education Citibank Washington University New York, NY St. Louis, MO Mary Jo Manning, Esq. Howard Young Partner Livonia, MI Wilkes, Artis, Hedrick, & Lane Washington, DC Ray Marshall L.B.J. School of Public Affairs University of Texas Austin, TX NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the advisory panel members. The panel does not, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for the report and the accuracy of its contents. iv

PAGE 5

OTA Project Staff: Technology and the American Economic Transition Lionel S. Johns, Assistant Director, Energy, Materials, and International Security Division Henry C. Kelly, Project Director Andrew W. Wyckoff Daniel J. Chenok Joy Dunkerley Review & Counsel Peter Blair Audrey Buyrn Walter Parham Peter Sharfman Contributors Food: Mark Belle, Charles Abdalla, John Connor, Kenneth Cook, Jack Giacin, Roger Griffin, Steve Hiemstra, Eric Hyman, Chester Mackson, Rosanna Morrison, Stanley Johnson, Bruce Harte, Robert Evenson, Carl Pray, Jonathan Putnam, Theron Dowries, Ronald Cotterrill. Health: Zeda Rosenberg, Nicholas Ashford, Tony Bale, Howard Berliner, Robert Curtis, Eli Ginzberg, Peter Goldschmidt, Dale Hattis, Irene Jillson-Boostrom, Louanne Kennedy, Andrew Martin, Mariam Ostow, Curtis Priest, Herbert Semmel, Alvin Strelnick Housing: David Dowall, Robert Gold, Vincent Brannigan, Edward Starostovic, Steven Winter, Alton Bradford, Don Carlson, David Claridge, Michael Clevenger, Albert Dietz, Eric Dluhosch, John Eberhard, James Gross, Harry Mileaf, John Millhone, Piero Patri,* Richard Reisman,* Richard Tucker,* Charles Thornton,* Wendel Wendel,* Raymond Whitten* Transportation: Elizabeth Deakin, William Garrison, Robert Cervero. Education: Francis Fisher, Alfred Bork, Lewis Branscomb, John Seely Brown, Andrew Catkins, Margaret CoX, Richard Cyret, Christopher Dede, Dorothy Deringer,* David Docterman, J. Dexter Fletcher, William Gattis, John Goodlad, John Kemeny, Daniel Levin, J.C.R. Licklider, Milbrey McLaughlin, Arthur Melmed)* Andrew Molnar, Billy Reagan, Theodore Sizer, Thomas Snyder, Patricia Sturdivant* Communications: Robert Bolter, Donald Dunn, Jamie Grodsky, Jim McConnaughey Recreation and Leisure: Steven Ackerman, Anindya Bose, Arlene Fleming, Margaret Fischer, Haines Gaffner, Robert Howard, Steve Sieck, Robert November, Peter Verhoven Textiles and Apparel: Ruth Ruttenberg, Gordan Berkstresser, N.C. Williamson. Internationaf Trade issues: Steven Haggard, Michael Aho, Steven Cohen, peter Kenen, Raymond Vernon, William Wallace, John Zysman* Manufacturing Issues: Steven Cohen, Barbara Baran, Bruce Guile, Carol Parsons, Robert Scott, Jean Ross. Analytics: Frank Bean, William Booth, Mary Feng, Jan Graf, David Gillman, Gail McGrew, Lori Renner Time Use: John Robinson, Jeffrey Holland Editorial & Inspirational: Robin Clarke, Ann Kelly, Robert Niblock, Emilia Govan Administrative Staff: Phyllis Brumfield, Lillian Chapman, Linda Long OTA Publkhing Staff: Kathie Boss, James Agee, John Bergling, James Brown, Debra Datcher, Cheryl Davis, Dorinda Edmondson, Chip Moore *Prepared papers for workshops held in connection with this project.

PAGE 6

Technology and the American Economic Transition List of Reviewers Karen J. Baehler Roosevelt Center Washington, DC Richard S. Belous The Conference Board Washington, DC Howard Berliner Conservation for Human Resources Columbia University New York, NY Anne P. Corter Brandeis University Boston, MA John Connor Department of Agricultural Economics Purdue University West Lafayette, IN Elizabeth Deakin Institute of Urban and Regional Development University of California at Berkeley Berkeley, CA Fay Duchin Director Institute for Economic Analysis New York University New York, NY Kenneth Friedman Special Assistant to the Deputy Assistant Secretary for Conservation Conservation and Renewable Energy U.S. Department of Energy Washington, DC William Garrison Institute for Transportation Studies University of California at Berkeley Berkeley, CA Bruce Guile National Academy Washington, DC Carl Hodges Director of Engineering Environmental Research Laboratory Tucson International Airport Tucson, AZ Kaname Ikeda Science Counselor Embassy of Japan Washington, DC David R. Johns Chairman and Chief Executive Officer David R. Johns Group Phoenix, AZ Ronald Kutscher Associate Commissioner Office of Economic Growth and Employment Projections Bureau of Labor Statistics U.S. Department of Labor Washington, DC Linda LeGrande U.S. Congressional Research Service Washington, DC Jeff S. Luke Director Bureau of Governmental Research and Service University of Oregon Eugene, OR Larry Mishel Economic Policy Institute Washington, DC Ronald Moroney Director Innovative Technology and Special Projects Division U.S. Department of Housing and Urban Development Washington, DC Mark Planting Interindustry Economics Division Bureau of Economic Analysis U.S. Department of Commerce Washington, DC Art Rosenfeld Lawrence Berkeley Laboratories Berkeley, CA Marc Ross Energy and Materials Program Department of Physics University of Michigan Ann Arbor, MI Laura D. Tyson Department of Economics University of California at Berkeley Berkeley, CA Amy Walton Jet Propulsion Laboratory Pasadena, CA George Wino American Textile Manufacturers Institute Washington, DC Paula Young Bureau of Rconomic Analysis U.S. Department of Commerce Washington, DC

PAGE 7

Contents Page Chapter I: Overview . . . . . . . . . . . . . . 3 Part I: The Networks of Consumption Chapter Chapter PartI 1 Chapter Chapter Chapter 2: Defining the Consumption Recipe . . . . . . . . . 61 3: Eight Amenity Groups . . . . . . . . . . . 89 The Networks of Production 4: Defining the Production Recipe . . . . . . . . .. ....147 5:Three Perspectives on Structural Change ..............................165 6: The Networks That Produce Amenity . . . . .................203 Part III: The International Connection Chapter7: The Framework of U.S. Trade . . . . . . . . ......285 Chapter8: How Trade Enters U.S. Production Recipes . . ...................303 Chapter 9: The Composition of Trade . . . . . . . . . . . 321 Part IV: People in the Production Recipe Chapter 10: How America Uses People at Work Chapter 11: Matching People to Jobs . . Chapter 12: Work and the Amenity Networks . . . . . . . . . 341 . . . . . . . . 363 . . . . . . . . 395 PartV: Policy and the Future Chapter 13: Alternative Paths for the U.S. Economy . . ......................425 Chapter 14: Public Regulation and Incentives . . . . . . . . .. ...445 Appendix on Data and Methods . . . . . . . . ...............467 Index . . . . . . . . . . . . . . ..............485 vii

PAGE 8

List of Related OTA Reports Technology and the American Economic Transition Supporting Reports l Special Report Technology, Trade, and the U.S. Residential Construction industry OTA-TET-315, September 1986, NTIS order #PB 87-117 594/AS The U.S. Textile and Apparel industry: A Revolution in Progress OTA-TET-332, April 1987, NTIS order #PB 87-196 762/XAB l Technical Memorandum A Review of U.S. Competitiveness in Agricultural Trade OTA-TM-TET-29, October 1986, GPO stock W52-OO3-O1O54-2, NTIS order #PB 87-140 760/AS l Panel Proceedings Technology and the Future of the U.S. Construction Industry AIA Press, 1986, NTIS order #PB 86-209 442/AS Other OTA Reports l New Electric Power Technologies OTA-E-246, .hdy 1985, GPO stock W52-003-01005-4, NTIS order #PB 86-121 746/AS International Competition in Services OTA-ITE-328, July 1987, GPO stock W52-OO3-O1O67-4, NTIS order #PB 87-212 403/AS l Trade in Services: Exports and Foreign Revenues OTA-ITE-316, September 1986, GPO Stock W52-OO3-O1O49-6, NTIS order #PB 87-118 204/AS l Technology and Structural Unemployment OTA-lTE-250, February 1986, GPO stock W52-OO3-O1OI7-8, NTIS order #PB 86-206 174/AS l Technology, Public Policy, and the Changing Structure of American Agriculture OTA-F-285, March 1986, GPO stock W52-OO3-O1O18-6, NTIS order #PB 86-184 637/AS l Preventing illness and injury in the Workplace OTA-H-256, April 1985, NTIS order #PB 86-115 334/AS l Life Sustaining Technologies and the Elderly OTA-BA-306, .h]y 1987, GPO stock W52-OO3-O1O74-7 NTIS order #PB 87-222 527/AS l The Electronic Supervisor: New Technology, New Tensions OTA-CIT-333, September 1987, GPO stock W52-OO3-O1O82-8 NTIS order # PB 88-156 351/AS l Automation of America Offices OTA-CIT-287, December 1985, NTIS order #PB 86-185 055/AS l Computerized Manufacturing Automation OTA-CIT-235, April 1984, GPO stock W52-003-00949-8, NTIS order #PB 84-196 500/AS l Technology, Innovation, and Regional Economic Development OTA-ST1-238, .hdy 1984, GPO stock W52-003-00959-5, NTIS order #PB 85-150 894/AS Note: Reports are available through the U.S. Government Printing Office, Superintendent of Documents, Washington, DC 20402-9325, (202) 783-3238; and the National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22162-0001, (703) 487-4650,

PAGE 9

Chapter 1 Overview

PAGE 10

CONTENTS Introduction . . . . . . . . . . . . Guide to the Analysis. . . . . . . . . . . . . . Analysis by Networks . . . . . . . . . . . . Policy Consequences. . . . . . 4........ . 9 . . . The Structure of the Report ........ . . . . . . . . . Summary . . . 0 . 0 . . . . . 0 . . The New Rules.. . . . . . . . . . . . . . . New Technology. . . . . . . . . . . . . . New Challenges From Abroad . . . . . . . . . . . New Resource Constraints . . . . . . . . . . . New Values and Tastes . . . . . . . . . . . . Changes in Public and Private Regulations and lncentives . . . . . Choices Created by the New Rules . . . . . . . . . . Economy-Wide Structural Change . . . . . . . . . . Stmctural Change in the Eight Major Amenity Networks of the American Economy Food . . . . . . . . . . . . . . . Health . . . . . . . . . . . . . . . Housing . . . . . . . . . . . . . . . Transportation . . . . . . . . . . . . . Clothing and Personal Care . . . . . . . . . . . Education . . . . . . . . . . . . . . Personal Business and Communication . . . . . . . . Recreation and Leisure . . . . . . . . . . . . Pulling the Pieces Together . . . . . . . . . . . 3 5 5 12 14 15 15 15 18 19 19 20 21 21 40 40 41 43 45 46 47 49 50 51

PAGE 11

Chapter 1 Overview INTRODUCTION During the next two decades, new technologies, rapid increases in foreign trade, and the tastes and values of a new generation of Americans are likely to reshape virtually every product, every service, and every job in the United States. These forces will shake the foundations of the most secure American businesses. Few features of the change seem inevitable. The Nations future has probably never been less constrained by the cost of natural resources or the limits of human strength, dexterity, or memory. Much less depends on physical limits to what can be done and much more on what Americans choose to do, acting privately as consumers, investors, and employees and publicly as voters. The choices will affect the options available to consumers, the rate and nature of growth in different business sectors, the geography of growth, and the role played by large and small enterprises. They will affect Americas position in the world economy and the number and quality of jobs the American economy produces. This document is designed to describe the nature and consequences of some of these choices. Given the importance of choice, it would be foolish to make confident predictions about the future of Americas economy. It is possible, however, to outline a set of exciting possibilities. They include opportunities to: extend life and reduce sickness and disability; create more choices in recreation and entertainment; expand access to information about products and services; have products and services more precisely tailored to personal needs and tastes; and make learning more productive and accessible. There are also opportunities for making work more rewardingin all senses of that term. Technology can replace many of the most tedious, dangerous, and dehumanizing tasks while creating jobs that require more intellectual and social skills. Machines are likely to plant seeds, weave cloth, fabricate metal parts, handle routine paperwork, enter data, and perform a vast number of other repetitive tasks more efficiently and more productively than people. By default, the majority of jobs created in the economy could be those requiring human, and not machinelike skills: designing; tailoring products and services to unique customer needs; teaching; caring; entertaining; promoting; and persuading. Ironically, one result of sophisticated technology may be a work force whose primary task is dealing with people as customers or as colleagues. With each opportunity for progress, of course, there are opportunities for serious missteps. Even change that clearly leads to overall economic growth can have very uneven effects. Change can lead to wrenching dislocation and pain for workers with obsolete skills, for management unable to recognize opportunity, and for communities where traditional businesses have failed. Change can create an America in graceless decline its living standards falling behind those of other world powers. This could mean an America less able to ensure the operation of free international markets and less able to ensure the security of the free world. Change can weaken the bargaining position of some groups while strengthening that of others. Change can result in a growing gap between those fortunate enough to have the talents, education, and connections needed to seize emerging opportunities and those forced into narrowly defined, heavily monitored, temporary positions, This latter group could be forced to bear most of the costs of uncertainty. A central issue is at the core of the choices: will change increase or decrease the power of individuals? Specifically: l l Will it become easier to purchase products tailored to specific interests, or will choices be constrained as national production systems substitute standard products for specialized products reflecting individual interests or local tastes? Will businesses change and grow under the assumption that workers will be well educated and intellectually flexible, or will they plan with the assumption that workers will be ignorant, untrainable, and unreliable? Another way of asking this question is whether people will be able to find a variety of attractive opportunities for work, or whether only a credentialed elite will enjoy such opportunities. 3

PAGE 12

4 From another perspective, the issue is whether the flexibility and dynamism essential to progress in the emerging economy will come at the expense of individuals, or whether individuals themselves will become more flexible because of continuous opportunity for learning and growth. In the past it seemed necessary to make a Faustian compact with technology: efficiency could be improved only by sacrificing individuality. Efficiency demanded mass production of uniform products, and a reasonable income often required acceptance of a constraining and narrowly defined occupation. It is at least possible that emerging technology will make it possible to avoid such a choice. Can America operate a dynamic and growing economy and navigate a transition to a new, more flexible, economic structure without falling into any of many potential traps? Specifically, can the U.S. Government create a set of marketplace rules that bring private and public goals into greater harmony without stifling the innovation and entrepreneurship needed to reach these goals? Can both workers and investors have incentives to undertake major changes in production systems? Can a major transformation be managed so that unavoidable trauma will not all be borne by a single group? The following pages argue that the answer to all these questions is yesbut only given an unflinching reexamination of some of the most cherished notions about the way businesses are managed internally, and about the way networks of enterprises work together. In the end, the choices governing the direction taken by the economy will be made by individual Americans acting as consumers with diverse interests, employees interested in rewarding work, and investors in search of profits. The rules under which these choices will be made are the issue. The collection of rules, regulations, and incentives adopted over the past several decades for perfectly good reasons may send misleading signals today. Programs designed to create growth in new directions require an ability to think in clear, practical ways about the way skills and investment are connected to the provision of good health, recreation, and other amenities measured in human terms. The baffling complexity makes it easy to be misled about where real opportunities for progress lie. Indeed, it can make it difficult to believe in the possibility of system-wide progress. The analysis that follows provides a practical set of tools for evaluating the performance of an economy that operates increasingly as a set of complex networks, which add value in many different ways and in many different locations before a product or service finds its way to a household. It also develops a set of concrete hypotheses about the way economic networks could operate in the future. The document does not attempt to forecast the future. Forecasting implies that choice plays a minor role. Instead, the analysis attempts to provide the clearest possible description of the available choices and their implications. The changes discussed in this volume appear to make prescriptive government planning less desirable. This does not necessarily mean that the responsibilities of government are reduced. Government maintains a central role because it creates many of the rules under which private choices are made. It may also have a growing responsibility for ensuring that Americans have adequate access to education throughout their lives, for ensuring a continuous flow of invention and innovation, and for protecting individuals from the dangers and risks of rapid economic change. Eight congressional committees asked the Office of Technology Assessment to take a broad look at the opportunities and risks created by the new environment in which Americas economy now operates, in order to identify areas where existing policy might block attractive avenues of growth and where new policy could improve incentives. Real economic growth requires both a belief that progress is possible and a vision of progress that is broadly shared. Government cannot create such vision, but it can provide a place where such visions emerge. The genius that has driven U.S. prosperity throughout its history has been an ability to combine collective vision with diversity and individualismto unite grand ideals with hard pragmatism. As the United States enters the 21st century, this genius will be put to its severest test.

PAGE 13

5 GUIDE TO THE ANALYSIS Analysis by Networks Given the broad goals of this analysis, it is necessary to take a very basic look at the way the economy operates as a whole. The vocabulary and the accounting techniques used to describe the economy, however, can limit our ability to imagine fundamental change. They embody many implicit assumptions about values and the way an economy operates. There is no obvious way, for example, to know whether a shift to an economy heavily based on services is either a fate to be embraced or avoided. In many manufacturing industries most employees never touch production equipment; more than half the cost of producing a computer, for example, typically results from software development. Should computer manufacturing now be considered a service? Many service workers are now involved in facilitating the design and delivery of manufactured products tailored to specific needs, or in facilitating the formation of production networks needed for timely delivery of new products. Should growth in these enterprises be read as the decline of manufacturing? In an effort to avoid the traps of these and other abstractions, this analysis concentrates on some basic concepts: l However sophisticated the economy may become in the future, its final products must always be the production of what will be called amenity throughout this volume. Amenity is intended to mean anything that contributes to the comfort, convenience, or happiness of an individual or household. While amenity is measured differently by every person, the amenity categories remain essentially unchanged in basic areas like health, food, housing, entertainment, and security. Progress in the economy must ultimately be measured by the extent to which the quality of these amenities has been improved, and by the extent to which these improvements are shared by the least fortunate groups in America. There are no easy ways to monitor gains in amenity. Growth in average income per person l provides only a limited view of changes taking place. Averages, for example, can mask increasing differences between wealthy and poor households. Growth in spending on burglar alarms is obviously not a good measure of security. Increased medical spending is not a good measure of national health. However sophisticated the economy becomes, the networks providing amenity directly or indirectly generate all the employment created by the economy. Barring the development of a pill that increases native intelligence or removes the need for sleep, the basic resources of time, talent, and enthusiasm available in the work force will not be altered by any economic transformation. Change can, however, affect the mix of skills demanded throughout the economy, and the links connecting income to skill. It can alter the quality of education offered by the work force, and the capacity of workers to learn and adapt to shifting circumstances. It can alter the quality of jobs, the texture of a working day, and the extent to which a person can take pride and pleasure in the work. The quality of working life is itself an important amenity. Most of the following analysis is devoted to an examination of the way amenity and jobs are connected in todays economy and the way these connections may change during the next two decades. Understanding these connections, of course, requires a clear understanding of the way a modern economy operates. The goal is to enter these intricacies without losing track of the fact that the analysis must begin and end with people. Paradoxically, even in an economy increasingly based on information flows the simplest questions seem more difficult to address. The information available can be overwhelming rather than helpful. The analytical strategy proposed here attempts to close the gap between speculation based on anecdotes and quantitative analysis based on national economic statistics. It uses statistical analysis wherever possible, but combines these results with the insights of experts in areas where statistics are not available or

PAGE 14

6 are difficult to interpret. Above all, the method is designed to be flexible enough to describe opportunities for basic structural change. A Parable Before describing the methods of this study in more detail, it is useful to look at an example of a contemporary production network. What could be more basic than a frozen pizza? A man cooking a frozen pizza in a microwave oven cares about what the pizza costs, how it tastes, how its preparation fits into his increasingly harried lifestyle, and maybe a bit about whether it is good for his diet or health. In effect, he cares about the net productivity of the network of activities (including the time he invested in cooking and learning about health) that brought the pizza to his palate. He probably couldnt care less whether it was the product of a manufacturing economy, a service economy, an information economy, or an international economy. But consider a likely chain of events that culminated in the pizza. Knowledge about health effects of food came from a TV talk show and information about a sale on pizza from a newspaper ad. Wheat for the pizza crust was grown in Kansas using sophisticated seeds and pesticides. The pizza was assembled automatically and wrapped in materials that are themselves the product of considerable research. The pizza was probably purchased at a grocery store where a clerk passed it over a laser scanner, which entered data into a computer and communication system designed to adjust inventories, restock shelves, and reorder products. This system in turn made it possible to operate an efficiently dispatched transportation system, placing a premium on timely and safe delivery rather than on low bulk hauling charges. The checkout data were probably also used to analyze consumer response to the previous days advertisement and to ensure that the store was closely following trends in neighborhood tastes. This pizza parable is important because food remains a major part of the U.S. economy. The parable is even more important, however, because it contains many of the themes now reshaping the American economy, It is difficult to argue that the United States has moved beyond an agricultural economy when one person in seven still works directly or indirectly to bring food to American tables. Demand for food continues, but the jobs involved in supplying it are very different from those of our grandparents. In 1984, only 4 of every 100 jobs in food production were on farms. The number of lawyers, bankers, scientists, and accountants needed to supply food in 1984 was about equal to the number of farmers. More than half the jobs supplied by the food network were for sales workers (20 percent), precision craftsmen (12 percent), managers (11 percent), and data entry clerks (10 percent). Technology plays a crucial role at each step of the network of business activity just described. Foreign products enter at many points: the pizza may contain tomatoes from Mexico, may be prepared by food handling equipment from West Germany, and may be sold using checkout equipment from Japan. While the rules governing the network are primarily those of private markets, the role of government is pervasive. Government regulations control the safety and labeling of products. Public funds supply the highways crucial to food delivery. Many farmers spend hours finding ways to benefit from government farm programs. Basic Tools for Representing Networks The analysis presented in this study begins by dividing the entire output of the U.S. economy into 10 components, identified by the amenity they provide: l l l l l l l l l l Food, Housing, Health, Transportation, Clothing and Personal Care, Education, Personal Business and Communication, Recreation and Leisure, Defense, and government activities not elsewhere classified. Only the first eight amenity groups are analyzed in detail in the material that follows. Changes in demand and in production efficiencies, have reduced the fraction of Americas income spent on each amenity. During the past three decades, the fraction of spending used to purchase Food has declined while the fraction spent on Health, Education, Per-

PAGE 15

7 sonal Business and Communication, and Recreation and Leisure has increased (see figure l-l). The amenity ultimately enjoyed by each individual or household is achieved through a complex series of steps, each of which adds value of some kind. The value may be contributed by investments of household time, or maybe added by a business that contributes to a purchased product or service. The final value of a purchased frozen pizza, for example, includes value contributed by farmers, packers, truckers, wholesalers, grocery stores, and the collection of legal, financial, insurance, and other firms that facilitate transactions needed by the complex network of activities. The methods this analysis uses to keep track of the flow of value in economy are summarized in figure 1-2, which illustrates four basic steps in the network connecting people as consumers to people as employees and investors: 1. Consumption Recipes. These recipes describe the way households achieve each amenity by combining time invested by household members, purchased goods, purchased services, and goods and services purchased by the government. (The mix of private and public spending in 1985 is summarized in figure 1-3.) A recipe Figure 1-1.-How America Spent Its Money (percent of all personal and government spending in current dollars) Food Housing Transportation Health Clothlng & Pers. Care Education Pers. Busi, & Comm, Recreation & Leisure Government (nec) Defense r Figure 1-2.-Structure of the Analysis 1 Food Housing Transportation Health Clothing & Pers. Care Education Pers, Busi, & Comm, Recreation & Leisure Government (n. e c.) Defense 1 1 ) 5 10 15 20 25 Percent of current dollar spending Household ~ Government How To Read This Figure: Of ail household and government spending in 1985, 14.5 percent was spent by households to purchase food and 0.5 percent by government to purchase food or support food production. SOURCE: Office of Technology Assessment (see table 2-2 of ch. 2).

PAGE 16

8 a mathematical analysis of the way spending correlates with price, income, household structure, and other factors); and second, by examining the possibility that entirely new directions may be taken because of unexpected new products, changes in consumer education, changes in tastes, or changes in formal regulations. 2. production Recipes. These recipes represent the mix of ingredients needed by producers to supply household and government buyers with products and services. Each business produces output by combining material and services purchased from other businesses with value that is added by the business itself. This value takes the form of labor performed by the business Figure 1-4.-Where Is Value Added? (valu e employees and the returns paid to investors for use of capital. Tracing the value that ultimately appears in the price of a consumer product requires an understanding of the way value is added by a complicated network of businesses that supply each other. If the purchasing recipe of each business is known, it is possible to estimate the wages and profits generated in each sector in the process of satisfying the eight major classes of consumption. The result of such an analysis for 1984 is shown in figure 1-4. ] l Information in this and other tables displayed in this chapter must be treated with some caution. They are based on a large number of assumptions discussed in greater detail in later chapters. -added by production sector in 1984) Food HOUSING TransportatIo n Heal t h Clothing & Pers Care Education Pers Busi & Comm Recreation & Leisure Government (n e c.) De fen se Exports n Construction & Other EZZl Social Services = Transactions D Transport & Trade = Manufacturin g Natural Resources* 11111111 1 0 100 200 300 400 500 600 700 800 BI I lions of $198 0 How To Read This Figure: In 1984, the U.S. gross national product (GNP) was $2,890 billion (1980 dollars). Of this total, approximately $430 billion went to produce food for U.S. household and government purchasers. Of this $430 billion, $64 billion ended up as value-added (primarily employee compensation and profits) earned in Natural Resource businesses (a category that includes farming, fisheries, energy, and mining businesses). If the bars for each amenity group were placed end to end, the sum would equal the total U.S. GNP. l Farming, Mining, and Energy. Education, Health, and other Government. SOURCE: Office of Technology Assessment (see table 4-Sa of ch. 4).

PAGE 17

9 International Connections. Imported products have insinuated themselves into the domestic production recipes in complex ways (see figure 1-5), while nearly 10 percent of the U.S. gross national product (GNP) is sold abroad as exports. Trade of this magnitude obviously changes some of the basic rules under which domestic production networks operate. Different parts of the economy vary greatly in the way they are affected by trade. One dollar in five spent for Transportation ends up abroad, either directly (as in purchases of foreign cars or fuel) or indirectly (as in expenditures for steel used to produce automobile parts). About 12 percent of the price of products exported by U.S. firms results from imported products. Patterns of involvement can change rapidly. Housing, for example, was comparatively isolated from world trade in 1984 but there may soon be rapid growth in imported housing components. Figure 1-6 provides a rough estimate of the way trade affects domestic employment. Estimates of the displacement effects of imports are highly artificial, since they depend on assumptions about how U.S. firms would produce substitutes for imports. Gains and losses exceed 25 percent in some sectors. Figure l-5.-lmports Used Directly or Indirectly To Produce Amenity in 1984 Tranportatio n Clothing & Pers Care Federal Defense Exports Recreation 8 Leisure Food Housing Government (n e c ) Health Pers Busi & Comm Education 1 I o 5 10 15 20 25 Percent of domestic spending on amenity How To Read This Figure: 21.2 percent of the money spent by U.S. households and government for transportation in 1984 was spent abroad to purchase petroleum, automobiles, automobile components, and other products and services. SOURCE: Off Ice of Technology Assessment (see table 7-11 of ch 7). 4 4 People in Production Recipes. The value that is added by each U.S. business depends on the contributions of employees with different kinds of skills. Figure 1-7 completes the network described at the beginning of this discussion by showing how effort by people with different occupations translates into amenity. Occupation is used as a proxy for skill since no other good substitute exists. The American economy in 1900 allocated occupations to amenity very differently than the economy described by figure 1-7. A much larger fraction of all jobs would have been devoted to the production of Food and Clothing, for example, and a larger share of these jobs would have been directly involved with farming or other natural resources. What might happen to such structures in the future? The method of analysis just described provides a systematic framework for examining hypotheses about changes in consumption recipes, production recipes, trade, and the use of skills in production. Public policy sets the rules under which choices are made in each of these four areas. These policies are discussed in greater detail in chapter 14.

PAGE 18

10 Figure 1-7.-J Jobs Required To Provide Amenity in 1984 Food HousIng Transportation Heal t h Clothing & Pers, Care Education Pers, BusI, & Comm, Recreation & Leisure Government (nec,) Defense Exports 1 I 1 I I Sales & Other Customer Contact Managers & Support o 5 10 15 2 0 2 5 M I I I ions of jobs How To Read This Figure: In 1984, the U.S. economy produced approximately 107 million jobs. Of this total, approximately 17.5 million jobs resulted from the production of food for U.S. households and government. Of these, approximately 1.6 milI ion jobs were created for managers and management support occupations. All reported jobs in the U.S. economy are recorded somewhere in this figure, SOURCE: Office of Technology Assessment (sea table 10-6 of ch. 10). Chapter 13 of this volume traces plausible hypotheses about changing choices in consumption recipes, production recipes, trade, and staffing by occupation. Figure 1-8 shows one of the many different scenarios examined. It represents an extreme case in that it shows what could happen if a major transformation occurred in each of the four analytical areas: consumption recipes that might improve amenity (for example, it assumes that investments in health promotion and disease prevention succeed); production recipes built around new paradigms (for example, it assumes a shift to a tightly integrated network connecting fiber production to retail apparel outlets); reduced reliance on trade; and a work force built around comparatively well-educated workers. It reflects comparatively rapid productivity growth with the real U.S. GNP rising at 3 percent per year. Under the hypotheses leading to the economy described in figure 1-8, productivity would greatly reduce employment in most mechanical jobs while employment in managerial, technical, and sophisticated clerical tasks would increase. Fewer people would be needed to provide basic materials or material processing. Many more would be needed to manage complex transactions and tailor products to individual needs. Since it is difficult to alter the productivity of tasks where direct human contact is essential, the share of all jobs required to provide education, entertainment, and basic government services (like fire, police, and other services) would increase in proportion to the rest of the economy. Chapter 13 describes a variety of other scenarios. It may be possible, for example, to achieve high levels of amenity and a decline in the number of hours worked even though the dollar value of GNP grows comparatively slowly. There is no absolute way to measure the desirability of any of these scenarios. Basic Characteristics of the Networks Economic networks have been growing in size and complexity for some time. Each generation of new technology appears to increase the interdependence of businesses and individuals as well as the number of people and businesses tied together. Specifically:

PAGE 19

11 Figure 1-8. -A Scenario of Jobs Required To Produce Amenity Foo d Housing Transportatio n Healt h Clothing & Pers Car e Educat[o n Pers Busi & Com m RecreatIon & Leisure Government ( n e c ) Defens e Export s in 2005 (one of several considered) Machine & Transportation Operators, Laborers, and Farmers Managers & Support I I 1 I I I o 5 10 1 5 20 25 Millions of jobs How To Read This Figure: Several scenarios for the future structure of the U.S. economy are considered throughout this document. Under the assumption that significant, but plausible, changes occur in consumption recipes, production recipes, trade, and staffing patterns, this figure shows where jobs would be created in the U.S. economy in the year 2005. The format is identi cal to figure 1-7. SOURCE: Office of Technology Assessment (see table 13-12 of ch. 13). Virtually no part of the economy now operates without heavy reliance on elaborate production networks. Production, marketing, sales, and transactional activities combine to deliver everything from frozen pizza to health care. l Technology can increase the efficiency of a network acting as a whole as well as the performance of each enterprise in the network. Data provided by laser readers in grocery checkouts and improved communication systems permit efficient operation of tightly integrated networks of wholesalers, truckers, food processing firms, advertisers, and other businesses. This can be more important than efficiency improvements of any single business in the network. l Taken together, new technologies can greatly increase the efficiency with which energy and materials are used. The efficiency gains, however, can typically only be achieved with increased spending for design and management. l Intricate interconnections mean that problems and prosperity propagate quickly and affect l l l many parts of the economy indirectly. Imports that affect a manufacturing firm also affect its web of suppliers. Exports create an elaborate trail of wealth. The new networks can have a surprising capacity for regeneration. New connections grow quickly when others are cut. Network components can shift rapidly to produce new products and services or to adopt new production strategies. Many of the networks have become international. Products, money, and ideas move about the world with new ease and speed. When production networks consist of many independent components, it is easier for both foreign and domestic businesses to enter. The complex networks have greatly expanded the role of transactional businessesbanks, lawyers, accountants, and communication specialists. Transactional costs can be measured by the increase in purchases of legal or financial services needed to make and enforce contracts, or

PAGE 20

12 in purchases of other specialized business services. They can also be measured by the growth of management occupations and other transactional overhead costs in corporate bureaucracies, l While disaggregate production networks could in principle spread economic activity widely in the United States, in fact they appear to have resulted in the concentration of economic activity in coastal cities. Businesses are choosing locations where they can find workers with adequate skills, and where extensive networks of personal contacts can be maintained. In the past much of American business needed easy access to natural resources. l New networks are not only changing the relations between businesses, they are also changing the relationship between the market economy and the unpaid work done by family members. Capital investment in things like microwave ovens and video tape recorders ties households to food and entertainment networks in new ways. Child care, care for the elderly, and other household tasks are entering the market economy. l People most likely to prosper in these networks are proteanable to change, adapt to unfamiliar work, and learn new trades as a continuous part of working experience. The talents needed are not clever hands or a strong back but rather the ability to understand instructions and poorly written manuals, ask questions, assimilate unfamiliar information, and work with unfamiliar teams. In short, the new networks require the skills provided by a solid basic education. There is a paradox in all of this. Countries, establishments, communities, and individuals are finding themselves ever more tightly connected, and yet the networks allow more independence and choice. In particular, technology may tie production systems in different countries more closely together while nations may become less and not more dependent on imported supplies of energy, food, and manufactured products. The green revolution used technology to make many nations that were formerly food importers self-sufficient in food production. Material substitution and efficiency can reduce needs to import energy and minerals. In such situations the movement of materials may have decreased while the strength of linkages moving information, technology, ideas, and capital equipment has increased. The shift to any new economic structure leads to effects resulting from the transition itself. These can often be difficult to distinguish from more lasting patterns of change. In a period of transformation one can expect rapid abandonment of older equipment (depreciation has grown steadily as a percentage of GNP), changes in business structure (rates of mergers, acquisitions, divestitures, and other transactions are extraordinarily high), and changes in demand for labor (there are large differences between the average job in the work force and the new jobs added during the past few years). Policy Consequences The depth of the changes underway in business networks, and the speed of change, requires a fresh look at policies designed to stimulate growth and mitigate the pain of adjustment. It is useful to divide options into two basic classes: 1. policies that affect the performance of the economy as a whole, changing rules in ways likely to facilitate choices made by consumers and pro ducers; and 2. policies that facilitate the performance of specific networks like Education, Housing, or Health. It is essential that the two policy strategies be coordinated. Policies designed to improve the performance of specific networks affect and are affected by programs designed to improve business performance through basic fiscal policy. Analysis presented later suggests that production networks in diverse areas are becoming much more similar in the way they are managed, in the way they are linked to the rest of the economy, in the range of skills expected of employees, in the way they depend on the products of research and development, in the way they are regulated, and in the way they react to signals from the financial markets. Each network, however, retains unique features that must receive separate attention. Programs designed to facilitate economic growth and expand the range of choices available to individuals can be built around the major elements of the network structure described in figure 1-2: l The performance of consumption networks can be improved by providing individuals with more information, by ensuring that their education

PAGE 21

13 is adequate to make informed choices in an increasingly complex society, by reviewing the incentives created by tax and regulatory policy, and by improving the way the government itself behaves as a consumer. l Production recipes can be improved by changing the financial rules that may reward shortterm profit taking over long-term investment, and by reshaping regulations that may place unnecessary constraints on the emergence of new production networks. l The growth of international production networks makes it increasingly difficult to manage the economy in the absence of international coordination and cooperation. Regulation in areas as diverse as banking, communications, and antitrust needs to be undertaken with even greater attention to the way domestic networks are tied to international ones. Policy can influence the way employers use People in production recipes in a variety of ways. Policy can encourage employers to achieve flexibility through teams of skilled employees rather than by relying on disposable workers. A variety of incentives can be used to encourage compensation schemes that reward investment in training and that allow firms to adjust to hard times by reducing bonuses instead of laying off the most vulnerable employees. Policy can also encourage workers to be more flexible by reducing the pain of adjustment. Public investment and incentives can ensure employees access to training needed throughout a career. Retirement and health benefits can be made more transportable. While the strategies are diverse, a number of themes emerge. Complex networks make centralized management of economic activity less feasible and less desirable. Networks in which productivity demands freedom to adapt and change may be impeded by regulation. This does not mean that planning is unwise, but rather that a new kind of planning is required. A firm confident in future demand for a specific product can plan to reduce costs through mass production. A firm certain that it cannot predict future demand must plan to have the flexibility needed to accommodate change. The changing networks make it important to understand how the effects of public policy ripple through the economy. The flexibility of the emerging networks, for example, makes it comparatively easy for firms to evade conventional regulation. It can only become harder, and not easier, to define and regulate banks, emergency care facilities, communication common carriers, or even electric utilities. Restrictions applied to groups of products seem increasingly easy to subvert. The problems associated with many kinds of traditional business regulation have led to an increased emphasis on programs designed to protect individuals rather than to regulate institutions. It is likely that this trend will continue. Steps can be taken to improve the information available to consumers as they make complicated decisions in areas ranging from health care to home purchasing. New rights to education, insurance, and pensions can be given to individual employees. New steps can be taken to ensure that markets take adequate account of safety, environmental, and other costs. These are difficult enough to regulate when production systems remain comparatively static. They are more so when systems are in flux. Earlier economic transformations were associated with a major public investment in infrastructure: canals, railroads, electric lines, and highways. The transformation taking place today seems to require an entirely different kind of public involvement. An educated population is the most critical infrastructure of the emerging economy. It is critical for both the economic growth of the Nation as a whole, and the success of individuals acting as either consumers or employees. Education has, of course, always been a central interest of American policy. Two things are new. First, the emerging economy places an unprecedented demand on the intellectual skills and knowledge of American workers. Old standards of competence are no longer adequate. Second, technology is making it possible to look for significant changes in the productivity and quality of teaching and learning for the first time. A system allowing any person, anywhere, with any background, and any assortment of gaps in education, access to training on any subject is within the state of the art of existing technology.

PAGE 22

14 The Structure of the Report It is obviously difficult to take a many dimensional network and describe it in a linear prose essay. Figure 1-9 illustrates how the remainder of this document is organized. Each of the four major parts covers a different element of the American economic system. Each introduces a set of analytical tools for using national statistical series to look at economywide patterns of change. Each of the four parts, however, also contains a discussion of issues unique to each amenity network. These analyses are based on an enormous variety of sources ranging from national statistical accounts to specialized journals. The analysis of specific networks includes an effort to define progress in areas like Health and Education, a description of the peculiar features of each network, an analysis of trends in the way the network operates, and speculation about how it could operate in the future. These examinations include a review of changes in patterns of consumer and government purchases, changes in the way producers combine to deliver products and services (and the net productivity of these combinations), and changes in the way people with different skills are linked together directly and indirectly in these networks. The final two chapters pull the pieces together. Chapter 13 develops a set of hypotheses about the future structure of the economy built from a series of specific hypotheses concerning consumption, production, trade, and labor. Chapter 14 reviews options for revising the regulations and incentives that shape the direction of U.S. economic growth. Figure l-9.-Organization of the Volum e Chapter 1: 1ntroduction and Summary Consumptio n Net works Setting U p the Analysi s n Chapter 2 Productio n Net works Trad e Net works Chapter 7 Chapter 8 Chapter 9 Labor in Productio n Net works Chapter 10 Chapter 11 r Chapter 12 I Chapter 13: Defining Alternative s Chapter 14: Defining Choice s SOURCE: Office of Technology Assessment.

PAGE 23

75 SUMMARY The following section provides a brief summary of the topics covered throughout the remainder of this volume. The reader is due two warnings before proceeding: 1. No attempt is made in this summary to provide references for facts or detailed arguments in support of conclusions. For this, the reader is referred to the sections of the report cited. 2. The discussions result in an assessment of strategic choices and provide a list of specific policies that could be used to implement these choices. The discussions of choice are divided THE NEW The rules under which the economy operates are being reshaped by four major forces: l l l l new technologiesprimarily those built around microelectronics; the loss of U.S. preeminence in international markets; the possibility that the price of energy and other resources may increase sharply by the turn of the century; and changes in consumer and labor markets and a new attitude toward public regulation of economic activity, resultingat least in part-from new values and tastes. Taken together, these forces appear to open more opportunities than they foreclose. New Technology Technology has forced the U.S. economy through profound changes in the past and shows every indication that it will do so again. The introduction of steam power, railroads, and mass production equipment at the beginning of the 19th century serves as one example; the development of electric power, inexpensive steel, automobiles and telephones at the beginning of the 20th century provides another. These inventions did much more than improve on the way things had been done in the past. They changed basic conceptions about the limits of human ingenuity, removing seemingly insurmountable barriers. Moreover, they had effects going far beinto two parts. The first part outlines a set of public policy goals and objectives that grow directly from the analysis of this report. The second part, which suggests options for achieving these objectives, enjoys no such connection and has not been analyzed in any detail. The options are offered to demonstrate that public choice can have a significant effect in moving the economy toward a desired objective. They are not intended to form a comprehensive fist. Neither the costs nor the benefits of specific options are estimated in detail. RULES yond the markets for which specific inventions were originally developed. Each cluster of technologies led to rapid growth in wealth, standards of living, and employment. The texture of everyday life was transformed. Weaving machines, for example, took work that had been performed in most households for generations and moved it to towns and factories. Inexpensive cloth improved comfort and sanitation and revolutionized fashion. New techniques used in the production of textiles and apparel turned villages into cities, changed the terms of international trade, and helped make England a world power. The automobile quickly became more than just an improved horse. Affordable cars reshaped everything from the design of cities and suburbs to styles of courtship. They generated noise, pollution, accidents, poetry, and an unimagined range of personal mobility. A curiosity at the beginning of the century, auto production dominated U.S. industry by the 1950s. Do the technologies now entering the economy have the potential to so transform society that their impact can be considered revolutionary? Can any of them do something that could simply not be done before, or reduce costs to a point where they can change basic paradigms of production throughout the economy? These are not easy questions. An overwhelming body of evidence suggests, however, that new technologies for collecting, storing, manipulat-

PAGE 24

16 ing, and communicating information do have the potential to revolutionize the structure and performance of the national economy. They would have this potential even if the technology were to freeze at present levelswhich it plainly will not. This is not to say that other emerging technologies will not also have a significant effect on the economy. Bioengineering can reduce the incidence of disease, extend life expectancy, reduce the cost of producing food, and improve the quality of American life in a variety of other ways. New materials can improve the performance of machines ranging from automobile engines to space stations. But while surprises are to be expected, it is likely that most of these technologies will do something familiar in a better wayat least during the next two decades. Information technologies have the potential to change the performance of the economic system itself. If a revolution of some sort is underway, measuring its impact with any precision has proven to be exasperating. As in the past, technologies with the power to reshape the basic structure of production have effects where they are least expected and where official government statistics are poorly designed to chronicle the change (statistical series are usually best suited to chronicle the previous technical revolution). Some evidence of the astonishing impact of the new technology, however, can be seen in the fact that over 40 percent of all new investments in plant and equipment are now in a category called information technology -computers, copying machines, and the like. This is double its share in 1978 (see figure 1-10). Much of this equipment is purchased by businesses specializing in transactions (law, banking, insurance, etc.), education, retail and wholesale trade, and in parts of manufacturing, health, and transportation traditionally associated with overhead or margins. These are businesses where output cannot be measured in concrete terms. Convincing evidence of productivity growth has yet to be measured in many of these businesses. It is difficult to determine whether this is a transitional effect associated with learning during a period of massive change, a defect of measurement, or a real limitation of the technology. The transitional problems associated with learning how best to use information technologies are Figure I-lO.-Investment in Information Equipment as a Percent of all Investment in Producer Durables 40 30 20 10 0 1950 1955 1960 1965 1970 1975 1980 1985 How To Read This Figure: Measured in constant 1982 dollars, 40.4 percent of all producers durable equipment (PDE) was spent to purchase computers, communication equipment, or related information equipment in 1984. About half of all investment in the U.S. is spent for PDE, the rest is used to purchase houses, and other construction products. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 5.7 (see figure 4.1 of ch. 4). large. As was the case with earlier innovations, however, the more sophisticated information technology becomes, the easier it is to operate and the more widespread is its adoption. It would have been difficult for James Watt to believe that heat engines (in automobiles) could be operated by people unable to distinguish a wrench from a peony. Information equipment has made an important transition during the past few years. Computing power once confined to specially equipped, air-conditioned rooms has already insinuated itself into everything from automobile carburetors to greeting cards and teddy bears. Like the carburetor, the computer has become invisible for most users. One of the problems of measurement is that information technology affects the structure of the economy in unexpected ways. Three areas appear to be of particular importance: 1. Information technologies can increase the productivity of operations where real productivity changes once seemed so remote that they may never have been considered seriously. Paper shuffling occupations provide a prime example. Clerical or quasi-clerical data entry, processing, communication, or manipulation is the principal

PAGE 25

17 occupation of at least 16 percent of the work force (27 percent if managers are included). The potential productivity gains in this areathe movement and organization of information are at least as great as those produced when the first Industrial Revolution began to revolutionize productivity in moving and transforming material objects. Education provides another unexpected example. At least 8 percent of the work force is engaged in teaching. The new information technologies have the potential to increase the quality and the quantity of learning that occurs per unit of a students time and per unit of teaching time. Quality means learning in greater depth. It means an ability to tailor teaching techniques to individual abilities and more time for teachers to talk to students. Making full use of information equipment, however, requires a willingness to undertake a basic reexamination of management strategies and job descriptions. This is always a painful and uncertain process. Successful use of new information equipment seldom involves the direct replacement of a task by automation. Instead, entire operations (like data entry, filing, and routine management tasks) may be replaced with a combination of information equipment and jobs with new responsibilities (typically quasi-professional jobs involving clerical, interpersonal, and analytical skills). The potential for using technology in education depends on a willingness to reshape what is taught, where it is taught, how it is taught, the point in a persons career when it is taught, and the range of talents needed by the teaching staff. 2. Information technologies have the potential to link production systems together in ways that improve the performance of entire networks. These technologies can make it easier to serve large numbers of highly specialized markets. They make it possible to tie together complex networks of producers around the Nation or around the world by forging tighter links between retail, wholesale, transportation, and manufacturing operations. They also make it possible to concentrate production in areas where labor skills, wages, business conditions or living conditions are judged to be favorable. Fundamental changes in management practices are needed throughout business networks if the full potential of the new technology is to be harnessed. Many such changes are already underway. Unprecedented scrutiny of management practices, inventory control, sales strategies, and other activities has already led to the growth of tightly integrated systems connecting manufacturers and business services. Tight networks now connect grocery stores to warehouses to food producers. Networks will shortly tie clothing retailers to apparel manufacturers to textile producers to the fiber industry. These changes impose a range of new demands on production technology and supporting services. For example, the new systems place a premium on rapid response and efficient batch production rather than on mass production. They place a premium on fast, reliable transportation of comparatively high-value products, rather than on low-cost transportation of bulk materials. As a result, they demand considerable support from service enterprises. 3. Information systems can change the ways business performance and financial information are gauged, and can increase the rate at which markets respond to new business conditions. Information technology that can make information both cheaper and easier to use can obviously have a deep, though subtle, effect on the efficiency of a free economy. It can affect both how producers organize their activities, and how consumers decide to spend their money. Time-sensitive prices of such things as electricity can be continuously communicated to individual residences, for example. Businesses can also be made much more sensitive to market trends, and can tailor production planning more closely to market demand. The international stock market crash of October 1987 provided a stunning reminder of the way that information processing equipment (facilitated by regulatory change) allows enormous amounts of money to shift hands quickly. The equations imbedded in the computer software conducting such trading do not just anticipate the performance of the national economic system but, to a major extent, have become part of the system. Regardless of whether this practice should be celebrated or condemned, there can be little doubt that information is influenc-

PAGE 26

18 ing the way national capital resources are allocated. New Challenges From Abroad The networks that combine to deliver amenities to Americans have taken on an increasingly international dimension. Exports and imports nearly doubled their share of GNP during the 1970s (see figure 1-1 1). The factors that make it possible for producers to divide production into separate steps in comparatively small establishments located across the country also make it possible to tie foreign producers into these networks. Once confidence is developed in foreign suppliers, it is easier for U.S. firms to expand operations abroad; once foreign producers establish a reputation for quality, they can build on this reputation. The process is cumulative and, barring catastrophic events, irreversible. While the establishment of familiar links with foreign producers irreversibly opens the potential for foreign production, these links do not necessarily imply a continuing expansion of foreign production. Long transportation and communication lines add costs, both directly through increased shipping and inventory costs, and indirectly through inflexibility. Domestic ingenuity can be substituted for foreign supplies of energy and materials. It is entirely possible that the advantages of quick access to domestic markets and production links will offset any advantages of foreign production. This topic will appear repeatedly in this analysis. The growth in trade over recent decades has resulted largely from the economic recovery of Japan and Western Europe following World War 11. This recovery was, of course, a central goal of U.S. foreign policy for four decadesbased on the sound belief that prosperity is a strong bulwark against Soviet destabilization. A second major factor has been the explosive growth of Korea, Taiwan, Hong Kong, Singapore, and other producers on the Pacific Rim who have developed rapidly, in large part because of their ability to offer competent workers with a sound basic education at low wages. In 1970, 40 percent of South Koreans of appropriate age were in the equivalent of high school. In 1982, 82 percent were enrolled. It also appears that the quality of the schooling received is high. Figure l-11.-Exports and Imports (percent of GNP in constant 1982$) Percent of GNP 15 10 5 0 5 -1 0 -1 5 -2 0 1950 1955 1960 1965 1970 1975 1980 1985 Net Exports = Exports ~ Import s How To Read This Figure: In the third quarter of 1987, exports totaled about 11 0 / 0 of the U.S. GNP while imports totaled about 15%. The trade deficit was about 4% of the GNP. third quarter. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 1.2 (see discussion In ch. 8). The gap separating productivity levels in Japan, Australia, the United States, Canada, and major European economies has been cut in half during the past 100 years. This has meant a convergence in living standards, incomes, and labor costs. There are some obvious reasons why followers gain on the leaders. Imitation is much easier once the basic paths have been revealed. Ideas flow rapidlyincreasingly so, if some recent studies are to be believedacross international borders. Expanding economies are more likely to take risks with state of-the-art production systems than established firms with large investments in existing equipment. While loss of overwhelming U.S. dominance in world trade was expected, the rapid erosion of U.S. capabilities in advanced technologies was not. The theory of the product cycle seemed to explain U.S. leadership for many years. The theory was based on an assumption that U.S. producers operated in an economy with some of the worlds most affluent consumers and best paid workers. Technologies developed for such an economy were unique, emphasizing high labor productivity and sophistication. As other nations approached U.S. living standards, and as the technologies involved became comparatively routine, they would naturally pass out of U.S. hands while U.S. producers moved to ever more sophisticated production.

PAGE 27

19 This cycle has not only been slowed by the convergence of living standards, it may have been reversed. Foreign firms, recognizing that dynamic exploitation of technology is the key to international competitiveness, may have taken the lead. Having broken the product cycle, foreign producers are suddenly in a position to finance the next generation of technology with income earned from sales of the current generation. U.S. producers are left trying to leapfrog competitors without such a revenue source. In the case of consumer electronics, U.S. producers appear to have all but abandoned the effort. The rules of trade have changed in fundamental ways for U.S. producers. Increased linkages in the domestic economy also mean that trade problems in one area (e.g., manufacturing) quickly translate into prob Iems in many parts of the economy (e.g., business service firms that owe a good portion of their livelihood to domestic manufacturing). Loss of U.S. dominance does not mean that U.S. living standards will necessarily fall; indeed, living standards in the United Kingdom rose steadily for a century after it lost its dominant position in the world economy. It simply means that undisputed U.S. economic leadership may be lost. Certainly, it is now possible that the United States will find its living standards in decline with respect to its competitors, and discover its role as an economic and military leader of the free world called into question during the next two decades. New Resource Constraints The availability of natural resources once dictated the pattern of the U.S. economy. Today, as industries learn to make more with less and to substitute light, high-value products for heavy, cheap ones, far fewer firms are constrained by lack of resources. This is already altering the geography of U.S. industry in major ways. But two natural resource issues will continue to demand attention: the availability of petroleum, and the limits of the environments ability to absorb waste. Both can be surmounted given adequate planning. Without such planning, the comparatively heavy energy dependence of the United States could become a major liability. U.S. lifestyles would suffer if the price of petroleum increased suddenly (60 percent of U.S. petroleum consumption is for personal vehicles); furthermore, the comparative energy inefficiency of U.S. products (cars and appliances, for example) and production systems could make U.S. products less attractive in international markets and foreign products more attractive in U.S. markets. Resource policy leading to inefficient energy use can also lead to inefficient use of capital. Lack of attention to energy efficiency resulting from poor utility regulation, for example, can lead to massive diversion of capital to energy projects. With current technologies, rising energy use translates into growth in solid waste and increases in air and water pollution. Fortunately, technology that improves the efficiency with which energy and materials are applied to products and services usually has the serendipitous effect of simultaneously reducing emissions. While energy use per dollar of GNP in the United States is still nearly twice that of Europe and Japan, it fell 25 percent during the 1973-86 period. U.S. consumption of steel, cement, paper, ammonia, chlorine, and even aluminum per dollar of GNP and per capita has either stopped growing or begun to decline. Electricity was the major exception to the trend, increasing nearly as fast as GNP since 1970. New Values and Tastes Another basic change in the rules involves changes in the tastes and values that motivate decisions in consumer markets, public spending, and labor markets. Changes in these areas have had dramatic and sometimes complex influences. Strenuous objections to nuclear power in the United States, the growth of the environmental protection industry, new interest in exercise and nutrition, and the growth of female participation in the work force have all resulted primarily from changes in values having littIe to do with economic forces. Demographic and Social Factors The size and structure of U.S. households and the role of women in the economy have both changed dramatically over the past decade. There has clearly been a change in the behavior that Americans find acceptable, and a consequent increase in the range of choices available for individuals. These changes come on top of an underlying pattern of demographic change of no small consequence. Taken together,

PAGE 28

20 they have been responsible for qualitative changes in patterns of consumer expenditure and in labor markets. Those born during the baby boom recently left the Nations educational system. They are now at an age to make major consumption decisionsdecisions that often reflect changes in values from the generation they have replaced. This generation has created booms in a number of markets (such as increased spending on health, fitness, and travel), and in some areas this has led to congestion. Many young families are discovering, for example, that population pressures in areas where new jobs are being created have driven prices for housing to levels that are prohibitive for new home buyers. This process is paralleled by a growing population of elderly people and a radical transformation in the size and structure of households. Divorces, late marriages, and growing acceptance of previously unacceptable living arrangements, such as singleparent households, have led to a rapid growth in comparatively small households. The average size of an American household fell from 3.3 people in 1960 to 2.7 in 1986. Women are now much less likely to leave the work force even when they have young children. Many are forced to work since they are the sole source of support for their families. Moreover, the expansion of female employment opportunities has meant that women are much less willing to subsidize the economy with poorly paid work. Markets for nurses, teachers, and other traditionally female occupations have changed greatly as a result. The increase in female participation has been offset by a sharp trend toward early retirement, resulting in part from more generous retirement programs and in part from a troubling trend toward the disposal of older workers for whom retraining is not judged to be profitable. Young married women are now as likely to be in the work force as older men (see figure 1-12). The number of hours worked in the economy by the average adult (age 16 to 65) seems to actually have declined during the past decade. Since the baby boom has recently increased the percentage of all Americans who are adult, however, the number of hours worked per member of the total U.S. population has increased. Figure l-12.-Work Force Participation for Older Men and Young Married Women Percent oo~ 80 Men aged 66-64 f~ .. 60 --Married women aged 26-34 --L -~. .t ~40 -----------------------20 i o~ 1955 1960 1965 1970 1975 1980 1985 How To Read This Figure: In 1966, approximately 67 percent of married women aged 25 to 34 participated in the work force (e.g., were working for pay or actively looking for work). Approximately the same percentage of all men aged 55 to 64 participated in the work force. SOURCE: U.S. Department of Labor, Bureau of Labor Statistics, Handbook of hborSt Wst/cs, Bulletin 2217 (Washington, DC: U.S. Government Printing Office, Juna 1965); data since 1963 provided by the Bureau of Labor Statistics (see discussion in ch. 11). Changes in Public and Private Regulations and Incentives The phrase regulations and incentives is used throughout this document to mean the answer to the question What explains the behavior of the individuals involved in economic activity? These rules fall into two basic types: the profit motive of free markets, as modified by formal incentives and regulations promulgated by the government; and informal patterns of management that have evolved historically, often without reference to economic logic. For better or for worse, both the formal and informal regulations governing the U.S. economy are undergoing changes that increase the freedom of American businesses. This is not a return to a halcyon past that was free of regulation. Most business activity was much more heavily regulated in the past. Regulatory reform has eliminated or greatly reduced controls that once tightly constrained competition in enterprises throughout the economy. Many emerging technologies have the effect of reducing the size of economically competitive facilities. In some circumstances, privately owned phone

PAGE 29

21 switches can compete with utility telephone switches, and cogeneration plants can compete with utility generation of electricity. Other technologies have introduced competition in areas once dominated by a single technology (such as alternatives to broadcast television and business communication systems). The result has been to undermine the logic of using principles of natural monopoly to regulate broad areas of the economy. Pressures of international competition are also destroying many traditional management mechanisms in businesses ranging from textiles to insurance. While the logic of monopoly control may have been altered in many ways by the emergence of new technologies and demands for new network flexibility, the changes may increase the need for regulation to protect the health, safety, and privacy of individuals. Perhaps the most critical question has to do with the motivation of individuals. Hard work, dedication, and inspiration cannot be entirely explained by looking at how a rational person would maximize income under prescribed rules. What keeps an inventor working through the weekend? The prospect of fortune plainly helps, but interviews with innovators continually suggest that much more is involved. A sense of pride, the admiration of colleagues and employers and, at least in traditional societies, a sense of contribution to a collective or national accomplishment all play a critical role. Interviews with entrepreneurs who left large firms to start their own repeatedly suggest that a desire for independence contributed as much as a desire to earn more money. Public or private rules that frustrate such intangible individual incentives are unlikely to promote real national growth. CHOICES CREATED BY THE NEW RULES The rules shaping opportunities for economic growth have changed in ways that undermine many basic paradigms for public and private management. Deep structural changes are underway in virtually every U.S. business sector. Some are painfully obvious: firms threatened by extinction in the face of foreign competition, or a health care industry revolutionized by new technology that has overwhelmed traditional means for managing health financing. In other parts of the economy, change is more subtle: an insurance industry in which products and production processes have been transformed by competition and new information technology, or a transportation system reshaped by new demands for speed, reliability, and quality in the delivery of comparatively small, lightweight products. Still other parts of the economy (education and home building, for example) remain comparatively stable, but there is potential for major change. The choices created by the new rules are viewed from two perspectives: 1) structural change underway at the level of the entire economy and the policy choices presented by these changes, and 2) structural change underway in specific amenity networks and the choices presented by these changes. Changes affecting the entire economy are measured by shifts in the relative role of different types of businesses (e.g., farming, manufacturing, and service enterprises), in the nature and volume of trade, and in the kinds of jobs created and lost. Choices exposed by an examination of economic performance at this level suggest a review of the incentives and disincentives created for businesses by laws affecting finance, trade, and labor policy. In many cases these signals seem incompatible with the course of action most likely to capture the opportunities of the new rules. Changes in the performance of specific amenity networks can only be seen by examining the unique features of Health, Education, Transportation, and other sectors. Each require programs specialized to their unique problems and opportunities. Choices made about such programs work best when coordinated with programs designed to stimulate innovation and growth throughout the economy, and when made with a clear view of problems and opportunities shared by different sectors. Economy-Wide Structural Change The following discussion examines structural change in the U.S. economy by tracing changes at each of the four stages that connect people as consumers to people as employees and investors described in figure 1-2:

PAGE 30

22 1. changes in consumer and government spending (consumption recipes), 2. changes in the way purchased goods and services are produced (production recipes), 3. changes in the way trade modifies these production recipes, and 4. changes in the use of people in production recipes. 2 Each of these discussions lead to a distinct set of choices for public policy. Consumption Recipes (Ch. 2) The term consumption recipe is defined to mean the combination of private and public investment used to achieve an amenity like good health (see table l-l). These recipes are based on a complex combination of logic and whimsy. They are shaped by changes in demographics, incomes, income distribution, prices, and by the emergence of new products and services. They can be strongly influenced by public regulation, tax policy, and information. Describing Recent Trends.American households are looking forand findinggreater quality and diversity in products and services. This results in part from the fact that U.S. households are growing more diverse, and in part from the fact that technology and trade bring a greater variety of products to market at prices not significantly higher than those of mass produced goods. Technology also makes it possible to reach small markets efficiently. While the rate of change in household types is likely to slow during the next two decades, the new diversity of American households will continue to shape demand for products and services tailored to a range of interests as the baby boom generation moves through middle age. Only one-fifth of house2 The compensation and unearned income generated in step 4, of course, provide a basic constraint on purchasing patterns described in step 1. The analysis is designed to be internally consistent, but no attempt is made to produce a closed economic model. Table I-l. -Elements of the Consumption Recipe Household purchases from labor and capital income + Household purchases from transfer and other income + Unpaid time of household members +Government purchases of goods and services =Total final demand SOURCE: Office of Technology Assessment, 1988. holds now have a working father, a housewife, and children, down from about one-third in 1972. Households headed by the elderly are increasing rapidly. Changing incomes and income distribution also shape purchasing patterns. Higher income families, for example, tend to spend a higher fraction of their incomes on Recreation, Education, and Clothing and less on Food and Health. Demographic effects (particularly the increase in households headed by single women) and inequality in the distribution of property-type income has led to greater inequality in the distribution of after-tax income among families. While income distribution is obviously important for reasons of social policy, it has a surprisingly small effect on levels of relative spending in major consumption categories. Direct consumer purchases of services are increasing. Much of the growth has been in the form of demand for financial services (including credit), hired substitutes for work formerly performed by unpaid household labor (like child care and care for the elderly), and increased expenditures on health care. A declining share of U.S. income has gone to purchases of Food, but the share spent on food-related services (primarily in restaurants) has increased over the past 30 years. Purchases of energy and materials are declining as comparatively light, high-value products (such as personal computers) represent an increasing share of purchases. Purchases of heavy durables are leveling off. And people are buying more and more products tailored to their specific needs, from magazines to insurance policies. There has always been a demand for tailored services of this kind, and new technology is now making it possible to serve specialized markets by lowering the differences in cost that have traditionally made batch production more expensive than mass production. Purchasing decisions also depend heavily on the availability of a persons time. Americans reported spending 108 minutes more per week in paid work, commuting, and household-related work in 1985 than in 1975; they therefore had less free time for eating, personal care, and other household activities (see figure 1-13). There was less time for housekeeping as well, Men did approximately 1 hour more housework per week in 1985 than in 1975, but women did about 5 hours less per week. Time avail-

PAGE 31

23 Figure l-13.-American Time Budgets Work, Commuting 8 Education Housework, child care, shopping, family travel Eating & personal care. Free time -.---- Work, Commuting & Education Housework, child care, shopping, family travel Eating & personal care Free time //////////////////&/// Men o 10 20 30 40 50 Hours per week per adult m 1975 = 1985 How To Read This Figure: In 1975, American women 18 years of age or older had 40 hours per week for free time activities (organizations, social life, recreation, television and other electronic media, leisure travel, etc.). By 1985, this had fallen to 36.6 hours. l Not including 7.5 hours/night. SOURCE: Office of Technology Assessment (see table 2-3 of ch. 2). able for leisure also seems to have declined in the past decade, in part due to the 40 percent increase in work-related hours. Women were affected the most, losing an average of 3 hours free time per week. Both sexes spent much less time spent socializing and participating in social organizations. There was even a slight decline in time spent watching tele vision. The mix of private and public spending used to achieve amenity has remained remarkably stable at levels that were shown in figure 1-3. There are some exceptions. For example, two-thirds of medical spending came directly from patients in 1950. In 1983, 73 percent of all spending came from insurance companies or government programs. Strategic Choices. An economy that increasingly takes the form of complex and interlocked networks can have the effect of putting more power in the hands of consumers. While the complexity of choice increases, the time available for making informed decisions is declining for many households. The joy of frivolous and capricious purchases and the principle of caveat emptor are obviously important driving forces in a free economy. Enthusiasm for them diminishes when it appears that ignorant consumers may place their health or savings at risk out of ignorance, or when the social costs of poorly informed consumers are high. The high social costs of poor consumer choices can be illustrated by the following, admittedly extreme example: poorly informed consumers might buy air conditioners that are much less efficient than ones that would have been purchased with the benefit of an informed analysis of the returns to an investment in higher efficiency. These poorly informed decisions would create a demand for electricity that is higher than would have been the case if all consumers were better informed. This, in turn, could force the construction of a new generating station, for which real returns to capital invested are much lower than the returns that consumers would have earned on added investments in air-conditioners. Leaving all other issues aside, the net result of such a series of events is an inefficient use of national capital resources. New technologies can help to some extent by making it easier for consumers to sort out information. New information systems could even permit purchasers to play a more active role in designing products and services tailored to their personal needs, and even to receive on-line guidance about price trends in products and services of interest to them. Policy can help improve consumption recipes in several ways: l Regulations designed to protect consumers, which may have made perfectly good sense when first enacted, may have outlived their usefulness in areas ranging from banking to housing to electric utilities. Many of the specific issues are discussed in the individual amenity network discussions. Three principles, however, seem to have general value in reviewing how regulations might be rebuilt to facilitate the emergence of a more productive, restructured economy: 1. Arguments about regulations governing the pricing of regulated monopolies, or those governing business entry or exit, need to be kept distinct from arguments about the need to protect consumers against fraud, unsafe practices, or environmental damage. Changing technologies may undermine the logic of regulated monopolies or elaborate programs of cross subsidies, but increase the need for government inspection and safety controls.

PAGE 32

24 l l l 2. Subsidies for low-income groups are commonly used to justify regulation of transportation, housing, communications, and other major parts of the economy. They may, however, be inefficient ways of allocating public funds when a large portion of the subsidy does not end up helping the poor. 3. Standards for public regulation of hazards in the environment vary widely. High standards are set in one area (e.g., mobile source emissions), while other problems (e.g., indoor air) are largely unregulated. Contaminants in some kinds of foods are held to very low levels while others are more loosely regulated. In neither case does the response seem to correlate with levels of risk. It would be useful to have a uniform way to measure environmental exposure and associated risks from all sourcesfood, water, and air. There has been no comprehensive assessment of environmental regulation parallel to the omnibus review of the penal code that occurred several years ago. Labels can provide consumers with information not otherwise available for making intelligent choices. Energy efficiency labels on air-conditioners encourage consumers to consider energy costs when making an investment decision, much as miles per gallon labels help consumers make choices about cars. Warning labels on cigarettes and nutrition labels on food products also help in this process. There are many areas where labeling to give quality ratings could be improved, ranging from food productswhere information about cholesterol, alcohol, sodium, saturated fat, and pesticide residues appears only voluntarilyto houses, where precise information about energy efficiency and other quality features is extremely hard to obtain. Personal income taxes strongly affect consumer incentives, by doing such things as making investments in houses (of any size) more easily deductible than investments in education. These priorities could be reviewed. The Federal Government, considered a consumer because of the peculiarities of standard accounting practices, can improve the way it makes critical purchasing decisions and can support research designed to improve its own productivity. Government buildings and vehicles are often less efficient than those purchased by the private sector. The U.S. Government, for example, pays more than $1 billion per year to subsidize heating for public housingstructures often having energy bills far higher than necessary. Major government enterprises, such as public education, are not backed by consistent research support or funds for capital investment. Production Recipes (Chs. 4 and 5) The term production recipe refers to the combination of products and services a firm purchases from other firms, It includes items from both foreign and domestic suppliers, and both capital equipment and supplies. 3 Producing for the diverse and rapidly changing markets just described requires alert and integrated production networks. These production networks connect many kinds of manufacturing enterprises, service firms, transportation operations, wholesalers, and retailers. The performance of these networks cannot be measured effectively by looking only at the performance of component businessesthe customer sees the performance of the system acting as a whole. A trucking firm specializing in on-time delivery of comparatively small batches may seem to have lower productivity than another firm with lower charges per ton-mile. Yet the flexibility offered by the more responsive trucking service could reduce costs throughout a complex production and retailing network. Describing Recent Trends.Change in the structure of production networks is measured in three ways: 1. Changes in the recipes used by individual businesses. Each business creates a product by purchasing inputs from other businesses (materials, capital equipment, business services, etc.) and adds value of its own through the skill of its employees and the capital supplied by its owners (see table 1-2). 2. Changes in the size of the establishments that comprise the production network, and in their ability to produce more than one product. Fords sin techn~ca] terms, the recipes considered here include both fixed investment and intermediate inputs.

PAGE 33

25 Table 1.2.Components of the Production Recipe Products purchases from domestic producers + Products purchased from abroad +Value added Capital costs Labor costs Indirect taxes =Total value of industry output SOURCE Office of Technology Assessment, 1986 3. enormous Rouge River plant, which took raw materials in one end and pushed finished cars out the other, seems unlikely to be the model for future production. Instead, production networks are being organized in ways that defy easy classification. This applies to both the size of establishments and the size of firms that may own multiple establishments. Changes in the geography of production. Changes in production networks can also be measured by shifts in the physical location of businesses. Changes in recipes, changes in the scale of manufacturing and service enterprises, and a variety of other factors have redefined the logic of firm location. Changing lnputs. New technologies, and the complex networks in part made possible by these technologies, make increasingly efficient use of natural resources and energy inputs. The decline in the share of GNP contributed by natural resource businesses (farming, mining, and energy) resulted mostly from a decline in domestic demand for their products (see figure 1-14). The declining share of natural resource businesses is more than offset by growth in transactional businesses (law, finance, insurance, business services, etc.). Between 1972 and 1984, these gains were almost equally divided between increases in consumer and government demand and increased use of transactional activities as inputs by other businesses (see figure 1-1 5). The rising use of transactional inputs results in part from increased demands for information, and in part from the fact that the management of these networks has become more complicatedthe emerging networks require both more complex external contracting and more complex flows of products and information within firms. The three fastest growing direct inputs into the Nations production processes are wholesale and retail Figure l-14.-Shares of GNP (in constant 1982$) ~F Percent of GNP I Manufacturing I 20How To Read This Figure: Measured in constant 1982 dollars, the fraction of the gross national product contributed by farms, mines, energy companies, and other natural resource businesses fell from 12.8 percent in 1950 to 8.7 percent in 1985. The share contributed by manufacturing appears to have remained close to 22/0 during the entire period but there is reason to believe that some of the procedures leading to this estimate are fiawed (see chapter 5). SOURCE. U.S. Department of Cammerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 62 (see figures 5-2b, 5-3b, and 5-4 b). Figure l-15.-Explaining Changes in Sector Shares: 1972-1984 Losing Share of GNP Gaining Share of GNP r~ .. ~~ 4 2 0 2 4 6 Percentage point change ($1980) ~ Domestic Demand How To Read This Figure: Between 1972 and 1984 the percent of the U.S. GNP produced by Natural Resource businesses fell 2.4 percentage points. Changes in househoid and government demand reduced the share by 1.1 percentage points, changes in trade reduced the share by 0.4 percentage points, and changes in the production recipe of businesses reduced the share by 0.9 percentage points. SOURCE Office of Technology Assessment (see table 5.1 of ch, 5).

PAGE 34

26 trade services, business services like advertising and consulting, and communications. While the contribution of manufacturing appears constant in figure 1-14, its share of overall employment has been falling rapidly. Calculations showing that manufacturing commands a fixed share of the GNP, however, depend heavily on techniques used to adjust for inflation. Half of the growth in manufacturings contribution to GNP estimated for the 1980-86 period resulted from the difference between the inflation adjustment for computers and other information equipment and the average adjustment made for all U.S. businesses. Techniques discussed in chapter 5 suggest that the part of manufacturing paying the highest wages (such as the automobile and steel industries) may have fallen as a percent of GNP during the past few years, resulting in an overall decline in the share of manufacturing. Deficiencies in available data leave the issue clouded. International trade had a relatively minor effect on the changing GNP shares of most sectors in the U.S. economy before 1980, but has since had an effect nearly as large as shifts in domestic demand. The negative effect of trade was felt most heavily in manufacturing. Transportation, wholesale and retail trade, and transactional services all benefited from trade because many of these enterprises gained business from exports. There is no good index for measuring the growth of networks. One way is to see how much business a firm generates for its suppliers. An economy-wide index of linkage based on this concept increased 5 percent between 1972 and 1980. While the entire economy became more tightly linked during this period, conflicting forces are at work. The links connecting manufacturing firms to the rest of the economy, especially to the service sector, seem to be growing: the linkage index for manufacturing industries paying the highest wages increased 15 percent between 1972 and 1980. On the other hand, the transactional businesses gaining share in the economy are not very highly linked to other parts of the economy, since a large fraction of their sales result from their own value-added (rather than from the cost of inputs purchased from other firms). On average, manufacturers keep only about one-half of the value of their sales in the form of profits or wages paid to employeespaying the rest to their suppliers. In contrast, service firms keep nearly two-thirds. Service businesses are linked to the rest of the economy primarily through purchases of capital goods, such as buildings and computers. In 1984, demand for manufactured goods generated approximately 6.3 million service sector jobsabout 1 out of every 1 lwhile demand for service products indirectly created about 4.3 million manufacturing sector jobsabout 1 out of every 5. Changes in Scale and Scope.Given the striking transformations underway in production networks, it is not surprising that many traditional management strategies are under intense scrutiny. The prevailing wisdom about the optimum size of production facilities, relations to domestic and foreign suppliers, and investment in research and new production technology is being challenged across a broad front. As expected, responses from established management vary from enthusiasm to intransigence. In some cases it seems to swing wildly from one pole to the other. There is much debate, and no convincing data, about the kinds of business structure that will prove most successful in the emerging economy. Indeed the very vocabulary available for describing the different formal and informal business networks and management strategies being tried proves to be inadequate. It is obvious that an enormous amount of experimentation is underway. Many large firms are adopting a management style that allows subsidiaries to act with greater independence. It is common to find many establishments owned by a large firm that provides services such as communication networks, billing, advertising, and personnel management. Clusters of small firms can operate under the umbrella of a larger firm (e.g., suppliers to computer industries). Under the pressure of intense domestic and international competition, the marketplace should soon provide some clarifying signals. In the meanwhile, there is great confusion over whether smaller enterprises are more flexible, more innovative, more likely to invest in long-or short-range research and development, or more likely to be attractive places to work than larger, more bureaucratic institutions. Economies of scale in individual establishments appear not to be as much of an advantage as they once were. Indeed, there appear to be diseconomies if es-

PAGE 35

2 7 tablishments become so large than management loses touch with employees. Rigidly managed larger firms have suffered from an inability to move quickly to exploit new market niches and new production technologies. In their attempts to impose uniformity, such firms may be unable to learn from their employees. On the other hand, the larger firms can sustain major development projects requiring large, long-term investments. They can also be better employers, offering employees more benefits and more job stability. One measure of the rate of restructuring can be found in the surge of mergers, takeovers, and divestitures that have taken place in the past decade, made possible by reduced regulatory restrictions (see figure 1-16). Hostile takeovers represented only a small fraction of the total. The net result of mergers, and of patterns of new business formation and bankruptcy, has been a change in the size of the average U.S. establishment and in the range of products that it produces. Management styles seem to be converging across the U.S. economy. Large farms, for example, are being managed as sophisticated, heavily capitalized enterprises. The owner is often, in effect, a professional Figure l-16.-Mergers and Takeovers, 1972-86 (constant 1982$) ~111 OnS of dollars ( 1982$] ~ ~lc I 1972 1974 1976 1978 1980 1982 1984 198 6 = All merger activity = Hostile takeovers How To Read This Figure: In 1986, the value of mergers and takeovers totaled more than $150 billion (1982 dollars), up from $40 billion (1982 dollars) in 1972. The value of takeovers considered hostile totaled about $30 billion in 1986. Data for hostile takeovers not available prior to 1982. SOURCE: Julius Allen, Corporate Takeovers: A Survey of Recent Developments, U.S. Congressional Research Service report No. 87-726-E, Washington, DC, Aug. 6, 1987; and Carolyn Kay Brancato, Merger and Acquisition Activity: The Level of Hostile Mergers, U.S. Congressional Research Service report No 87-507 E, Washington, DC, June 16, 1987 (see discussion in ch. 5). manager. The construction industry is increasingly resembling a manufacturing enterprise, with more components assembled in factories. Fragmented low-technology enterprises, such as those found in the textile and apparel industries, are becoming consolidated and are coming to depend heavily on advanced production technology. Large integrated facilities, such as those in automobile production, are moving toward smaller establishments and shorter production runs. Large banking and insurance businesses are investing heavily in computer and communications equipment. Hospitals and groups of physicians are moving toward professional management, with physicians as paid employees. Schools may move toward a more differentiated teaching staff and increased capital investment. In 1984, 73 percent of all manufacturing employees worked in establishments employing more than 100 people, a 2.3-percent decline from 1982, while only 40 percent of employees in finance, insurance, and real estate worked in establishments with more than 100 employees (up 3.2 percent). Nevertheless, enterprises with fewer than 100 employees added 47 percent of all jobs gained by the economy between 1976 and 1984, even though they supplied only 37 percent of all jobs in 1976. The largest gain in employment share between 1978 and 1982, however, was in small establishments of large enterprises. While small firms have generated a growing percentage of all jobs, they appear to generate a decreasing share of output. Businesses with 20 to 500 employees increased their sales annually by 4.3 to 9 percent between 1976 and 1982, while firms with more than 10,000 employees grew at 16 percent. Part of this growth can be attributed to recent acquisitions and mergers. The strategies behind the mergers are frequently difficult to ascertain since most recent mergers have been classified as conglomerates, a category that includes both mergers made to achieve system-wide productivity gains and mergers that represent portfolios of convenience for speculators. While only 50 percent of mergers were classified as conglomerate in 1950, 90 percent were classified this way in 1979. Mergers tend to make the parent firm more diverse. Between 1963 and 1982, the number of different products made by a single manufacturing firm increased 15 percent. At the same time, establish-

PAGE 36

28 ments that are a part of larger firms have become more specialized. The number of different products made by single manufacturing establishments declined by a factor of 3 over the same period. The Geography of Production.-A map showing such natural features as rivers and harbors, or even minerals, no longer says a great deal about where economic activity will occur in the United States. New forms of communication, declines in the scale of individual establishments (in both manufacturing and services), a decline in the resource demands of manufacturing, movement of jobs from the shop floor to offices, improved flexibility and responsiveness in transportation, and a variety of other factors remove many of the constraints that formerly limited choices about business location. The theoretical possibility of locating a manufacturing or back office service firm virtually anywhere has, however, apparently been more than offset by other factors. The attraction of locations offering a well-educated labor pool (even where this means higher taxes), and access to major research centers and financial institutions, increases as other constraints on location diminish. A survey of high technology manufacturing firms taken in 1982 indicated that access to raw materials was 12th in a list of 12 reasons for selecting a plant location. The availability of skilled labor topped the list, ahead of concern about labor costs. Moreover, rapidly expanding businesses, such as banking, insurance, law, and real estate, still depend on networks of personal contacts and communications. This often translated into a decision to locate new facilities in the immediate vicinity of urban centers: New York, San Francisco, and more recently Los Angeles, Dallas, and Houston. The excitement and variety of these areas are plainly also a lure. The result has been an increasing concentration of employment, particularly high-wage employment, around major metropolitan areas of both coasts (see figure 1-17). Indeed, there has been rapid movement of economic activity down the east coast, from Boston to Miami; continuing expansion in Texas, Arizona, and southern California; and continuing losses of population and economic activity in the rust belt regions of the northern midwest. While inequalities in regional incomes declined during the 1970s, the decline appears to have stopped. Figure l-17.-Regional Wages: Percent Above or Below National Average Percent 16 I I n 14 12 10 8 6 4 2 0 2 1970 1975 1980 1986 = East Coast ~ West Coast = Pennsylvania, Ohio, Michigan, Illinois, Indiana How To Read This Figure: In 1988, wages in States bordering the Atlantic coast were 8.1 Io above the national average and wages in States bordering the Pacific coast were nearly 11.7/0 above average. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Information System, tables SA1, SA2, and SA3, unpublished, August 1987 (see discussion in ch. 5). In fact, during the 1980s New England and the Pacific coast regions appear to have regained their significant lead in average incomes. Incomes in Ohio, Illinois, Indiana, Michigan, and Wisconsin, which were above the national average in 1970, are now below average. Although they remain below the national average, incomes in States on the southern Atlantic coast showed a steady gain from 1970 to 1985. The southern Appalachian and Gulf Coast States, however, have incomes that remain 20 to 25 percent below the national averagea position unchanged since the 1960s. On the whole, the economies of the U.S. coastal States appear to be prospering, while the center of the Nation falters. It is difficult to account for the continuing attractiveness of coastal cities and the failure of U.S. companies to move to low-wage regions. Employers appear to be willing to pay a premium for access to people with good training and job skills, and use their new mobility to pay lower wages to relatively welltrained people abroad. An employee in Taiwan with a solid high school training in mathematics, for example, costs far less than a person with a mediocre education in the United States. With the exception of New York, most of the growth has occurred in centers associated with a major city but not in the traditional core of the city. Pop

PAGE 37

29 ulation shifts to the suburbs have allowed many people to live in relatively inexpensive areas while keeping commuting times reasonable. However, this has also made public transportation virtually impossible. And since the headquarters of most major firms are usually located in downtown areas, movement to suburbs may have blocked contacts critical for career advancement. Suburban business growth can also create problems for mid-level minority workers, many of whom cannot afford to live in suburban areas and are unable to commute thereon public transit. Population has followed economic activity, and Americans remain extremely mobile. Only half of the population lived in the same house in 1985 as in 1980, and one-third had moved from a different U.S. county or a different country. A survey of movers in the early 1980s indicated that the 50 percent moved to take a new job or military assignment. Nearly one out of every six people classified as displaced between 1981 and 1985 moved. Surprisingly, the highest percentage of displaced workers who moved were not the youngest workers but those aged 45 to 54. Strategic Choices. lt is apparent that a deep structural transformation is taking place in production networks. Trends are difficult to read because of complex patterns of experimentation. The challenge of policy here is to create a system of rules that encourages experimentation, transformation, and growth without sacrificing the traditional responsibilities of government to provide protection against monopoly abuse and protection for individuals and communities forced to bear the burden of adjustment. The challenge is to find practical ways to enjoy the advantages of competition and creative destruction in fostering change, while minimizing the risk that such a process will destroy the ability to engage in long-term development work. There is no magic way to do this. The most promising places to look for solutions are first in the financial rules that guide investment decisions, and second in research and development policy. Financial Rules.The rate at which constructive changes are made in production networks, as well as the kinds of networks created, depend heavily on the financial rules under which investors operate. Typically, failure of the U.S. financial system does not occur in its support of invention or innovation. The United States continues to lead the world in many research fields, and has a strong venture capital community. Problems seem to develop when hundreds of millions of dollars must be found for comparatively risky projects that require many years to reach maturity; in effect, the system seems to fail once the invention in the garage starts working, Pub Iic policy in areas like tax law, as well as the techniques used to evaluate managers, can influence the planning horizon of investors and businesses. Some financial instruments are obviously keyed to very short-term, speculative trading that is only weakly tied to the long-term interests of a firm. On the other hand, the pressure of aggressive financial markets is necessary to ensure effective management of national capital assets. Takeover threats and the potential for stockholder rebellion can provide a defense against inattentive management. The following policy options may improve the mix of patient capital and effective competition for capital resources: l l l l Tax reform: reduce or abolish income tax paid on capital gains from assets that are held for more than 3 years; limit the deduction for housing interest to a fixed amount; reform or abolish the corporate income tax (it contributes only 7 percent of government revenue; its complexity distorts private investment decisions; the transactional costs associated with tax compliance are high; the tax acts as a non-progressive sales tax for U.S. consumers; and it increases the cost of U.S. exports). Reform of the system regulating financial institutions (limitations on merchant banks, for example), which could allow for greater diversity in capital assets available to corporations and could encourage more long-term plans for new products and production networks. Reduction of transactional costs, by developing better techniques for combining regulation, negotiation, and freedom to sue for damages. Review of priorities for the $200 to $300 billion in federally supported loans and loan guarantees now largely directed to housing and agriculture.

PAGE 38

30 Research and Innovation. A competitive and prosperous economy depends both on a system of incentives that encourages investors to make money from new ideas, and on a continuous supply of new ideas; neither is adequate in itself. Research and innovation play a growing role in the health of every sector in the economy. Even areas where research has not traditionally played a major role (e.g., home construction and apparel) are likely to find their productivity and competitive position strongly influenced by the ability to conduct and absorb the products of research. It is essential that the United States improve on the way it combines corporate and public resources in the pursuit of innovation. Like education, this is an area where government expenditure should be considered as investment and not consumption. There are many areas of pure research that cannot be sponsored by corporate funds because of the high risks involved. Also, areas of applied engineering can play a crucial role in national economic development, but their value is likely to be spread so broadly that no firm can expect to regain its investment in full. The challenge is to find ways of spending public funds in a manner that neither competes with existing private innovation nor disappears in pursuit of white elephants. Two basic strategies deserve attention: 1. Cooperative Research Facilities. Measured in terms of the development of inventions that have provided a key to world economic growth, Bell Laboratories surely was one of the best research facilities in the history of the world. It has, of course, been largely reorganized following the breakup of the AT&T system. A variety of mechanisms have been proposed for building laboratories of this quality around clusters of pure and applied research issues: a. a mechanism for mingling public and private funds in ways that allow public disclosure of discoveries affecting the common interest, while permitting private firms to maintain rights to specific proprietary ideas; b. a reputation for research excellence that would attract the best people in the field (it may be necessary for the facility to be associated with a degree program at a major university); c. d e. f. research freedom adequate to permit exploration of novel ideas, coupled with a set of general themes tied to practical markets; a way to disburse funding through grants and venture allocations to small research groups; a way to combine academic research with research keyed to practical needs that does not compromise the intellectual freedom of universities; and a way to encourage research capable of improving productivity in business services (since service productivity is of growing importance to overall economic growth). The Bell management provided guidance on the balance between pure and applied research. Research ventures pooling resources from many firms could be managed by a private board with an interest in seeing that funds are well used. A number of sectors of the economy do not conduct significant amounts of research and might benefit from a coherent research program, such as new technologies for learning and teaching, residential construction, and textiles and apparel. Government itself is a major information enterprise, and could justify research designed to improve the net efficiency of data processing and communication within its own bureaucracy. This, coupled with a careful procurement policy, could work to stimulate efficiency in private firms and to encourage new information technology. 2. National Goals. Research in some areas could be stimulated by the clear articulation of a national visiona goal with the power of the space program that could be accepted as important by all Americans. Candidates include a Learning Research Institute, the renewal of space exploration, or even the development of a radically new form of personal transportation. Such projects are, by definition, impossible to justify on the basis of a purely rational calculus of cost and benefit. But they can couple public imagination and support of private genius in ways not otherwise possible. Properly constructed, the project could provide a home for a variety of research efforts and a market for many of the products of research. Defense-related projects can focus research efforts, but they may not be an adequate substitute; security classification, specialization in areas without clear commer-

PAGE 39

31 cial applications, and other factors limit the extent to which defense research facilitates opportunities for civilian economic growth. International Connections (Chs. 7 and 8) Describing Recent Trends.Production networks have always operated across international borders. In the past two decades, however, the networks connecting the United States to the rest of the world have changed in both size and scope. Nearly 7 million American workers now owe their jobs to exports, a percentage that has nearly doubled in a decade. Because of the large trade imbalance, however, U.S. employment might have been 10 percent higher in 1984 without imports.~ Figures 1-5 and 16 showed that trade has entered virtually every production network and has had a significant impact on virtually every U.S. business. The few areas that are not already involved heavily in trade (home construction, for example) may soon find themselves parts of international production networks. If nothing else, trade creates direct competition between American styles for managing networks and those of other countries. Until very recently, it could comfortably be assumed that American firms would be at the lead in development of new products and labor-saving technology. America led the world in both supplies and demand for innovation. American commitment to research and innovation was strong, and America had the most affluent consumers and most expensive labor. Forces driving both supplies and demand, however, have changed rapidly during the past few years. The ambiguity about whether new technology would encourage geographically dispersed production networks in the United States (as seems theoretically possible), or would lead to further concentration in favored regions translates directly to ambiguity over the effect of technology on trade. In some cases, technology has encouraged the spread of economic activity: 4These estimates, and many that follow in this summary, are based on an input/output analysis approach discussed in chapter 7. They must be treated with great care. Obviously, it is difficult to estimate how the economy would operate without imports since the price of domestically produced items is different. Assumptions needed to make the estimates, and the limitations of the methods, are discussed at length in chs. 4 and 7. l l l l The complexity of production networks, and the decline of large production establishments, has allowed more points of entry for foreign products. The declining significance of resources, and the growing importance of a firms ability to manage innovation, has made it possible for alert and aggressive foreign firms to compete successfully in U.S. markets. Communication and efficient transportation of lightweight, high-value products makes it possible to tie diverse networks together. Advanced telecommunication equipment has created worldwide financial networks. The previous discussion showed how technology can affect the scale of production. It proves difficult, however, to demonstrate the effect of scale on international competition. Anecdotes can be provided showing how small, entrepreneurial U.S. firms have fallen victim to large, patient, bureaucratic foreign organizations (e.g., producers of semiconductors). Examples can also be found where small, aggressive foreign firms entering U.S. market niches have beaten bureaucratic firms in the United States (e.g., manufacturers of textile machinery). There are also reasons to believe that technology may undermine trade and exaggerate, rather than shrink, the difference between living standards in rich l l l l and poor nations: Technology capable of making efficient use of energy and material resources can limit needs for world trade in these products. Technology capable of making small batch production nearly as efficient as large mass production can reduce the advantage of large plants producing for world markets. The growth of niche markets, and of domestic production networks requiring rapid movement of products between establishments, can increase the advantage of proximity. Networks connecting retailers and producers in ways that minimize inventories depend on an ability to respond quickly to changes in marketsthis increases the value of producing in locations physically close to the market served. Changing Comparative Advantage. -Under ideal conditions, the economics of trade should encourage each nation to specialize in the production

PAGE 40

32 of goods and services that its firms make at comparatively low cost. If the ratio of exports to imports of a given product decline, trade may have revealed a decline in the nations comparative advantage in that product or service. While real international markets fall far short of the ideal markets that in theory would reveal comparative advantage, since national policies frequently shape trade patterns, it is useful to see what the record shows about changes in Americas comparative position in different areas. Trade forces attention to the links connecting different parts of production networks. Figure 1-18 shows, for example, that a significant fraction of all jobs hypothetically lost from U.S. imports of manufactured products in 1984 were lost in service enterprises that played a role in production networks. About 1.5 million jobs in transportation and trade and over 600,000 jobs in transactional activities were lost indirectly because imports displaced domestic manufacturing activity. All told, 4 of every 15 jobs hypothetically displaced by imported manufactured goods are in the service sector, Figure l-18.-The Jobs Lost Because of Imports of Manufactured Products in 1984 (illustrating trade linkages) 11 1 i 0 500 1000 1500 2000 2500 3000 Thousands of jobs How To Read This Figure: Imports of manufactured products in 1984 resulted in the loss of 1.5 million jobs in Low Wage Manufacturing industries, 2.3 million jobs in Medium Wage industries and 1.6 million jobs in High Wage Manufacturing businesses. The loss of business for manufacturing, however, also creates losses for businesses that supply manufacturers with resources and services. Imports of manufactured products in 1984, for example, resulted in the loss of 300,000 jobs in Natural Resource businesses and 660,000 jobs in Transactional businesses. SOURCE: Office of Technology Assessment (see table 7-10 of ch. 7). Considering both direct and indirect effects, trade appears to have reduced business for nearly all U.S. high technology production between 1972 and 1984. Chemical products was one of the only industries that gained in relative rank between 1972 and 1984, while businesses in electronic components, office and computer equipment, engines, aircraft, and other technology-sensitive industries lost rank. In general, 18 of 20 industries designated as high technology lost rank over this period. Businesses with heavy natural resource components (i.e., lumber, leather, tobacco, and paper) gained rank, as did service and support businesses. Measured from the perspective of occupations gaining or losing advantage from trade, manufacturing employees lost in comparative terms between 1972 and 1984 while lawyers, communication equipment operators, and farmers gained rank. Of all hand working, machine setting, and precision production jobs in 1984,23 to 32 percent were displaced by imports. These occupations, however, also gained the most from exports, with 12 to 17 percent of their 1984 jobs resulting directly or indirectly from exports. The period showed a sharp break between the way engineers and scientists were affected by trade. Scientists benefited from changes in trade, while engineers were among those most heavily disadvantaged. Had all imports that could be manufactured in the United States in 1984 been so manufactured, there would theoretically have been 21 percent more jobs for engineers. On the other hand, 14 percent of U.S. engineers owed their jobs to exports. Natural scientists, computer experts, and mathematicians were less heavily affected, with a 12 percent loss from imports and an 8 percent gain from exports. The occupations least affected by trade were in services (i.e., teachers, health services, clericals, and sales personnel). Strategic Choices. It is theoretically possible for trade to improve living standards throughout the world, but this promise remains clouded by the stubborn realities of nationalism, protectionism, and distrust. The central objective of trade policy must be to make trade a positive sum game for all participants, including the United States. Even in the best of circumstances, the theoretical advantages of free trade can be demonstrated only for the average American over the long-run. There

PAGE 41

33 will be winners and losers as trade expands production networks across international borders. A second objective, therefore, must be to find ways to share the pain and the advantages of trade as widely as possible. At a minimum, no group should bear the bulk of the burden of adjustment. It is worth beginning by being clear about some of the problems associated with trade even under favorable circumstances: l l l l l Expanding trade increases the difficulty of managing the domestic economy without international cooperation. Expanding trade takes jobs from the individuals with the lowest skills in the United States, and increases uncertainty throughout the economy. Trade has a comparatively large effect on those with poor educations. In the economy as a whole, one job in five is held by a college graduate. Only one job in seven lost to trade, however, is lost by people likely to have college degrees. Statistics underestimate the disruption caused by expanding trade, since the jobs created by exports are not necessarily in the same companies, or located in the same place, as the jobs lost due to imports. Foreign producers able to expand their productivity rapidly may have a significant short-term advantage in crucial areas before wages in their home countries increase to reflect productivity growth. New or infant industries may be hurt by foreign targeting, although with rapid economic change, successful businesses prosper by being perpetual adolescents. A productive and innovative domestic economy is essential for national security-particularly in areas where the United States relies on technical advantages rather than on advantages growing out of higher levels of spending or manpower. The most important tool for improving the U.S. trading position is clearly the restoration of leadership in areas critical to Americas export performance, as well as productivity gains in manufacturing. Indeed, manufacturing exports are likely to be the only way that the United States will avoid massive trade deficits in the foreseeable future. But improvements in domestic productivity alone are unlikely to be adequate in a global environment where explicit manipulation of trade by foreign governments is the norm. To have any positive effect, management of trade must be patient and skillful. Under present circumstances, unilateral steps taken even by the largest economy in the world are unlikely to be effective. Only multinational action is likely to move constructively toward this objective. U.S. leadership is of vital importance. This leadership can grow only from a coherent strategy built around a clear set of objectives and from skillful and patient negotiations in a variety of international forums. Finding a way to put the U.S. trade and spending accounts in order by shrinking the deficits in government budgets and trade accounts is an essential first step. The following options deserve consideration: l l l l Banking and fiscal policies among nations belonging to the Organization for Economic Cooperation and Development (OECD) could be coordinated in fact, not just in rhetoric, to ensure that such policies are not working at cross purposes. Policies coordinating international development assistance and trade policy in the OECD could ensure the economic growth in the third world necessary for a long-term solution for their debt problems. Domestic policy decisions could be scrutinized more closely for their effects on the U.S. trading position. Prior to the post-war era, trade played such a small role in U.S. business transactions that public choices in critical areas, ranging from telecommunications to banking to agriculture, were made with little attention to the impact of these choices on the competitive position of U.S. firmsor, in the case of telecommunications, to the impact of opening U.S. markets to foreign competition without obtaining reciprocal opportunities for U.S. suppliers to enter foreign markets. Cooperative research programs in areas of clear mutual interest could be expanded. There are many areas where the combined world benefit of having a problem solved rapidly is far greater than the short-term gain that any single country could enjoy by reaching the goal first. Breakthroughs in medicine or safety for a variety of

PAGE 42

34 industrial operations would be mutually beneficial. l Tariffs, rather than non-tariff barriers, could be used when temporary protection is warranted, Tariffs maintain price competition and the value of the higher price paid by consumers remains in the United States, whereas quotas allow benefits of price increases to flow primarily to foreign suppliersthereby enabling them to further improve their long-term competitive position and expand their markets. l A variety of programs in export promotion could help businesses unfamiliar with the complex world of international markets find outlets for their products. Many of the Nations competitors have far more sophisticated systems than the United States in this area. Information alone can help. The Department of Agriculture, for example, has introduced a successful data network that assists comparatively small farming operations in bidding on foreign solicitations. l Increased language training, translation, and education about the culture of U.S. trading partners could help U.S. exporters understand export markets and more easily penetrate the labyrinth of foreign marketing problems. People in the Production Recipe (Chs. 10 and 11) The cycle suggested in figure 1-2 connects people in their role as workers and investors to people in their role as consumers. Some of these people know each other. Some are connected only by passing products to each other, and some are connected through paper flows and electronics. Figure 1-7 provided a crude description of how these networks operated in 1984. Capital invested in new technology, and the way management chooses to use this technology, has the potential to reshape these relationships in very basic ways. Specifically, information technology has the ability to mechanize tedious and repetitive intellectual tasks in much the same way that farming and manufacturing equipment mechanized tedious and repetitive physical tasks. Many of the jobs being created in the emerging economy do not appear in conventional categories. While the labor part of many tasks proves comparatively easy to automate, the part of a job that requires imagination, a capacity for learning, or an ability to work with other people is not easy to automate and therefore represents a growing fraction of all jobs. The new networks can make the links connecting a persons daily work and a jobs contribution to human welfare more transparent. This can mean that an apparel worker operating as a part of a responsive team tailors batches of products for specific clients instead of doing repetitive mass production. It can mean an insurance para-professional) discussing policies with a customer on the phone and entering data into a terminal rather than keypunching anonymous data. It can mean a teacher having more time to spend with individual students. Productive use of new technologies will require deep changes in management strategies and job descriptions in areas as diverse as insurance, education, health, and apparel production. These changes in turn can undermine long-standing assumptions about the role of people with different skills, the power commanded by different groups, the role of managers, the factors that motivate work, and the links connecting pay with skill. Some connections can be explained using the economics of human capital. The networks are obviously also connected by pride, loyalty, prejudice, compassion, whimsy, love, and many other factors not measured by bookkeepers. For better or for worse, the influence of all these factors is likely to be reshaped as production networks are transformed. The impact is enormously important but impossible to measure. Systems based on one kind of prejudice or irrationality may well be replaced by systems based on others. The task of management is changed when jobs become less precisely defined or when the job definitions themselves are changing. This can mean that the bargaining power workers have achieved on the basis of unique skills will be diminished, and that employers will have difficulty developing satisfactory techniques for measuring individual contributions. Education is likely to play a central role in determining who is advantaged and disadvantaged by structural change. Education seems important not just for making a person productive as a member of a team with a fixed purpose, but also because it helps people learn and change to keep pace with dynamic production networks.

PAGE 43

35 Since little is known about the merits of different approaches to using much modern production and information equipment, there is little certainty about the kinds of jobs that should be created. Radically different strategies are possible: l l It appears possible to create efficient production networks basing their performance and flexibility on the intellectual and social skills of employees. It is also possible to envision a system where a well-educated elite enjoys a majority of the benefits, while the costs of flexibility are paid by workers forced to take temporary jobs with narrowly defined tasks monitored by electronic surveillance. While it is possible that the second of these alternatives would prove most effective in a highly competitive economy, a wealth of anecdotes suggests that it would not. Choices made in the next few years can play a critical role in shaping the outcome. Policy designed to capture the first of the alternatives must accomplish two things, neither of which is adequate in itself: first, it must ensure an adequate supply of well-trained individuals (meaning education available throughout a persons career); and second, it must ensure a productive and innovative economy capable of providing jobs that can take advantage of a skilled work force. Describing Recent Trends.while changes in the structure of the economy are changing supplies of new jobs, the entry of the baby boom into the work force, the rapid growth of women in the work force, a move toward early retirement for older men, and a new wave of immigrants have changed the supply of skills and experience in the work force. There has also been a change in the racial composition of the work force. An absolute majority of all people joining the work force between 1985 and 2000 will be minorities, many of whom will enter with comparatively poor educations. Changes in the role played by transactional, manufacturing, natural resource, and other business sectors translate directly into changes in their share of U.S. employment. The effect is magnified by differences in the rate of growth of labor productivity in each sector. In part because of the comparatively rapid growth in productivity, the proportion of manufacturing jobs in the economy has declinedfrom 27 percent of all full-time equivalent employment in 1950 to 18 percent in 1986. Over the same period, jobs in business services and related activities grew from 6 to 15 percent. Changing demand and changing regulations were largely responsible for the sharp increase in health care, education, and other government employment during the 1950s and 1960s, and for their relatively unchanged share of national employment since 1975. Changes in production recipes, and even changes in the overall number of people a firm requires to produce its products, explain only a small part of the radical change in occupations emerging in the economy. Many of the new jobs fall into unfamiliar categories. This is particularly true in areas where information equipment has transformed processes and management strategies. Awkward terms like para-professional, para-librarian, and super clerk are used to bridge the gap between old paradigms about job categories and ones that remain to be developed. Professions like teacher are likely to be transformed into many identifiable specialties in the next few years. Since figure 1-7 contains all the jobs in America, it is possible to follow the contributions of different jobs to their ultimate contribution in the form of amenity. Figure 1-19 shows, for example, that jobs for natural scientists, engineers, and other technical professionals are heavily concentrated in Defense and Exports. LawFigure l-19.-Where Managers and Professionals Contribute Value* Percent of all workers [n occuoatlon 1 1 0 5 1 0 15 20 25 30 = Managers = Tech Professionals m Other Professionals How To Read This Figure: Nearly 15/0 of all managers and 9.6/0 of all technical professionals in the U.S. are emp~oyed directly or indirectly in the process of producing Food for domestic markets. Not including education and health professionals. SOURCE: Office of Technology Assessment (see table 10-6 of ch. 10)

PAGE 44

36 yers, social scientists, and other professionals are concentrated in Health, Education, and Recreation. Looking at the national statistics, managers and professionals were responsible for nearly 40 percent of net jobs added between 1972 and 1986 (see figure 1-20). Undoubtedly, many are doing tasks that would have been handled by clerical personnel in an era where entering and thinking about data could be clearly separated. Jobs in sales, administrative support, and services (other than household or protective services) also increased sharply, while there was a net decline in the number of persons holding jobs as laborers, fabricators, and operatives. Changes in the scale and scope of firms and establishments that form parts of production networks, and changes in the scale and scope of work teams within individual establishments, obviously reshape the nature of work and the kinds of jobs required. Occupations once part of rigid hierarchies, like mass production manufacturing, may be redefined if productive use of production equipment requires more independent intervention on the part of operators. Attempts to introduce mass production management into clerical work have not enjoyed much success. On the other hand, occupations with a tradition of independence (e.g., teachers, home builders, Figure l-20.-New Jobs and Existing Jobs Techmclans & Sales Service Occupations Precision Production, Craft, and Repair Operators, Fabricators, and Laborers Farming & Forestry by Occupation I I -10 0 10 20 313 40 5(3 %,:~.all 1972 ~ ~o~:all 198 6 m % of 1980-1986 job growth How To Read This Figure: Managers and Professionals held 19.6 percent of all U.S. jobs in 1972 and 24.5/0 of all jobs in 1986. About 45% of all job growth between 1960 and 1986, however, resulted from the addition of jobs in these occupations. SOURCE: Office of Technology Assessment (see table 10-10 of ch. 10), and physicians) may find themselves in more bureaucratic settings. The changes in occupations are extremely diverse. It appears, however, that on average new jobs are being created most rapidly in areas that require significant amounts of education (see figure 1-21). Growing unemployment rates among young workers with 4 or fewer years of high school was responsible for a significant part of all growth in U.S. unemployment rates in the last decade. High correlations between education and income persist, although the connection between skills learned in formaI education and skills actually required on the job may be becoming less clear. There may be a growing gap separating people able to grow and prosper with economic change from people who, for lack of good education or other reasons, remain trapped in positions with little future. Janitors, employees in fast-food emporiums, farm labor, maids, and a variety of other occupations are likely to represent a large part of the American work force for the foreseeable future. Changes in Pay and Job Quality.While a review of changes in the kinds of jobs produced provides one measure of how the new production networks have changed the nature of work, it is also important to see how they have changed the quality of Figure l-21.-The Growing Demand for Education (percent of jobs for indicated level of education) 45 I I = All jobs in 1986 m Jobs added 1983-86 How To Read This Figure: About 22 percent of all jobs in 1986 were held by people with four or more years of college education. Nearly 27/0 of all new jobs created between 1983 and 1986 were for positions held by people with 4 or more years of college education. SOURCE: Office of Technology Assessment (see table 11-4 of ch. 11).

PAGE 45

37 the jobs created. Job quality can be measured in a variety of ways: (a) the wages and benefits paid; (b) the equality in allocation of wages and benefits (it is entirely possible for the average wage to grow while the real wage of many employees falls); (c) the extent to which working hours can be adjusted as needed to combine work with raising children, an education, or partial retirement, while maintaining the security of the job; and (d) the extent to which a job provides non-cash rewards in the form of selffulfillment, pride, and fun. Each of these ways of measuring quality require a different approach. (a) Changes in the wages and benefits paid in American jobs are filled with paradoxes. Real hourly wages for non-supervisory production workers have not increased for 15 years while real GNP per capita has grown almost constantly. Employee compensation increased at almost exactly the same rate as productivity until the late 1970s, but comparatively little of productivity growth since the 1970s has been passed to employees in the form of increased compensation (see figure 1-22). Some of these paradoxes can be explained by the fact that increased international and domestic competition put pressure on U.S. wages. The effect was reinforced by the rapid growth in the work force resulting from the entry of the baby boom. The fall in the fraction of GNP paid as wages and salaries, however, resulted from several other factors. Economic transformation meant that a larger fraction of GNP was needed by businesses replacing equipment. This meant high rates of depreciation and a decline in the fraction of GNP paid as personal income. Non-wage benefits represent a growing fraction of compensation; more personal income was paid in the form of retirement benefits and welfare assistance. interest, dividends, rents, and other unearned income became a larger fraction of all income. The many forces at work have made changes in compensation difficult to explain. The industries where compensation increased faster than the average between 1970 and 1985 make a rather curious collection: legal services, the armed services, telephone and telegraph, metal mining, government enterprises (like the Tennessee Valley Authority), motion pictures, health services, and investment companies. Among other things, new patterns of capital investment have redefined the way wages are paid. Figure 1-22.-The New Gap Between Growth in Productivity and Compensation (real output & compensation per hour) 1977=1000 oo~ 1 / 900 I I I 1 ( I I 1 1 1 I 1 1 I 1 1 1 1 ? 1 1960 1965 1970 1975 1980 1985 Real Output/Hour Real Cc~mlp/Hour How To Read This Figure: Between 1977 and 1987 U.S. productivity increased 8.4 percent but compensation remained nearly unchanged. Preliminary. SOURCE: Economic Repo~ of the President 1988 (Washington, DC. US. Government Printing Office, 1988), p. 300 (see discussion in ch. 11). Some service industries have invested more money per employee than manufacturing firms since the recession of the early 1980s. Competitive markets, however, mean that these differences are not necessarily linked to wages. High rates of capital investment per worker may actually reduce the skill levels required of workers. Connections between pay and skill become more ambiguous when the most critical skill offered by an employee may be a capacity to learn new tasks quickly, to perform well in unique situations, and to work well with people. These are general skills, rather than specific skills that allow workers to benefit from a limited supply. (b) The factors just discussed, coupled with the shifting pattern of occupations, have reshaped the way income is distributed. The wages of men have become somewhat less equally distributed, as have the wages of women. On average, however, the gap separating male and female wages is closing, thereby removing a major source of wage inequality in the work force. The net effect of these changes is that overall wage inequality has not changed significantly. Inequality in compensation has probably grown more rapidly than inequality in wages because the benefits offered for different kinds of jobs differ greatly.

PAGE 46

38 Even if inequality in wages and salaries is not increasing, demographic and other factors are increasing inequality of household income. Growing numbers of households headed by single women, the growing importance of unearned income (which is very unequally distributed), and the fact that highincome people tend to marry each other created a situation where income per household became strikingly less equal during the 1980s. These and other effects have resulted in a situation where real after-tax income has fallen for more than 90 percent of American households. Some of this has resulted from the decline in family size. Household incomes per adult-equivalent family member have increased for most household types, 5 Inequality in household income per family member has, however, also increased during the past decade. (c) Apart from pay, a workers ability to controf working hours, and ultimately the ability to keep a job, is obviously a critical measure of job quality. Two perspectives on this issue are important. Flexibility is important to employees as well as to employers, although for different reasons. Working parents, students, people in semi-retirement, and people interested in job mobility often find flexibility in working hours an asset. Employers need flexibility to respond to new competitive pressures, business cycles, and to introduce new production schemes. The record on balancing the employer and employee interest in flexibility is mixed. The search for greater flexibility has caused many employers to minimize commitments to permanent staffs. The number of people working in temporary or part-time jobs has increased rapidly. This is an advantage for some and a liability for others. The key question is the extent to which changes in working hours are under a workers control. Mothers wishing to work at home, independent owner-operators, self-employed construction managers, and farmers are often willing to trade uncertainty for the sense of independence they can achieve by working for themselves or by themselves. For many, however, the uncertainty is highly unwelcome. The number of women who worked part time involuntarily, for example, increased 300 percent between 1967 and 1984. SAn adu]t-cquiva]cnt counts children as a fraction of an acfult. See ch. 11 for a further discussion. Surprisingly few people in the United States hold stable, 40 hour a week jobs. Throughout the economy, temporary, part-time, and self-employed workers increased 25 percent between 1975 and 1985 and now represent 27 percent of the work force. High volatility in employment has always been endured by nonunion employees with comparatively few skills. Temporary employment, however, has become common even for individuals with basic skills. Most of the burden of adjustment seems to be born by people with comparatively poor educations (see figure 1-23). College graduates weather recessions with comparatively low rates of unemployment, while unemployment for high school drop-outs went over 20 percent in the early 1980s. Annual turnover rates for blue-collar occupations may be as high as 20 percent. While there appears to be some growth in the number of firms attempting to stabilize employment by encouraging wage flexibility through profit-sharing and other means, firms have consistently achieved flexibility in the United States through new hires and lay-offs. This is particularly tempting when firms employ older, comparatively expensive workers with obsolescent skills, while new production systems place a premium on credentials and basic skills that these older workers may lack. It is difficult for a firm Figure 1-23.-Who Bears the Burden of Change? (unemployment and education) Percent unemployed 25 20 15 10 5 0 1970 1972 1974 1976 1978 1980 1982 1984 :ol~;ore years of IZZ 1-3 years of college m ~c;::~s of high = less than 4 years of high school How To Read This Figure: In 1985, 16/0 of people with less than 4 years of high school were unemployed but only 2.6/0 of people with four or more years of college were unemployed. SOURCE: U.S. Department of Education, Center for Education Statistics, The Corrdition of Education (Washington, DC: U.S. Government Printing Office, 1986), table 2-3 (see discussion in ch. 11).

PAGE 47

39 to justify retraining when workers with adequate training are looking for work, and when the costs of training can be lost if an employee leaves a job to work elsewhere. The shifts in occupation by industry are painfully visible in statistics on experienced workers who lost their jobs between 1981 and 1986 because of plant closings or lay-offs. Of the 5 million people unemployed due to these factors, half had worked in manufacturing, where nearly 1 job in 8 was eliminated. The service industries were virtually untouched. In spite of the severe recession of the early 1980s, only 2 percent of people working in finance, insurance, and real estate, and only 4 percent of the workers in business services, were displaced. These businesses added jobs throughout the 1980s. Workers displaced by economic restructuring faced a difficult period of adjustment. Of those displaced between 1981 and 1986, less than one-third were reemployed with salaries equal to or greater than their previous jobs. Nearly half failed to find full-time employment. Displacement hit older workers with particular force; less than half of displaced workers over the age of 55 had any kind of employment in 1986. (d) However important, there is plainly no good index for non-cash rewards of working life. As earlier discussions indicate, there appears to be an enormous range of possibilities in the character of the jobs created by emerging production networks. It is clearly possible that mechanical tasks can be largely eliminated, freeing workers for tasks they find personally more rewarding. It is also possible that they will find jobs more narrowly defined and monitored. The rapid shift away from jobs in mining, agriculture, and heavy manufacturing has helped create a steady decline in rates of occupational injuries and deaths. Accident rates for manufacturing, for example, are five times higher than those in transactional activities. New jobs, however, present risks that may be difficult to detect. Stress, sometimes induced by job insecurity, may be the most common health problem of high-pressure jobs demanding responsiveness in unusual situations. Rapid changes in production techniques may mean that by the time the health problems of a process are established, the process will be long obsolete. Strategic Choices. The challenge of policy designed to influence the way people are used in production recipes is to find a way to encourage supplies of people empowered by their education and experience to assume more attractive positions in the work force, and to encourage employers to build productive systems that create attractive jobs. A specific challenge is avoiding a potential mismatch between the skills offered by young workers who may have been poorly served by the Nations educational system, and an economy that increasingly links jobs to education and intellectual skills. A second goal is to ensure that the costs and benefits of needed business flexibility are equitably shared. Alternatives must be found to flexibility built around lay-offs of people least able to bear the burden. The following specific programs deserve consideration: l l l l l l Encouraging profit-sharing, in which the consequences of poor economic performance of a firm would be shared by employees in an equitable way rather than being, in effect, passed on to those individuals with the least potential for reemployment. Establishment of a Learning Research Institute, designed to provide local school systems, universities, and corporate training and retraining programs with an expanded set of tools for increasing the productivity of teaching and learning. The institute would conduct basic and applied research on the use of advanced learning technology, and on the new strategies of pedagogy required to make use of this technology. Encouraging, or requiring, universally retainable pensions and health insurance. Subsidies for education beyond high school, based on a flexible system of loans and vouchers that could be used to supplement corporate training funds. Welfare programs designed to provide incentives for education, training, and reemployment. A program designed to ensure that the emergence of flexible production networks often built around comparatively small firms or establishments does not undermine provisions ensuring employee safety, freedom from unnecessary stress, and privacy.

PAGE 48

40 Structural Change in the Eight Major Amenity Networks of the American Economy (Chs. 3, 6, 9, and 12) The previous discussion focused on a nationwide view of structural change. Many of the most important features of this change, however, cannot be understood without a careful look at the way production networks are being defined to deliver specific amenities. Virtually every production network is either in the midst of a basic change or may be entering one. They share many of the themes already identified: expanded consumer choice and fragmentation of markets, complex production networks, growing involvement in international production, and a profound redefinition of the nature of work. The quality of amenity consumers receive, and the kinds of jobs created to produce this amenity, may change in basic ways. The following section can provide only the most cursory sketch of the complex changes underway in each amenity group. In some ways the structure of the amenity networks are becoming more similar in the way they are managed, in the quality of the jobs they create, in the way they conduct research, and in the way they are involved in international trade. The unique features of each network, however, merit unique approaches to public policy. Strategies for improving the performance of each network will clearly work best if they are undertaken with a clear grasp of the changes taking place throughout the economy. Similarly, programs designed to improve prospects for growth in the economy as a whole are unlikely to succeed unless they are tested in the way they affect the performance of individual sectors. While the objectives of policy may be read from the following discussion of trends and the potential for change in each major network, no attempt is made to define alternative courses of action. Food (15 percent of personal and government spending) The Food amenity can be measured by the quality, variety, and convenience of food products purchased by people in different income groups, and by the extent to which eating habits contribute to good health. The two measures can, of course, give contradictory results. Better information about the connection between diet and health, changes in taste, and a variety of other factors have reshaped American consumption recipes for Food over the past few years. For example: l l l an increased number of working women and more hurried lifestyles have led to growth in restaurant eating, even for breakfast; greater concern for health has led to a decline in consumption of red meats and other sources of cholesterol, and to an increase in consumption of fish, poultry, low-fat milk, fruits, and vegetables; and there has been an emergence of more varied tastes both in restaurant and stores. While most Americans enjoy an increasingly varied and healthy diet, poor nutrition continues to be a problem for the elderly, the homeless, poor pregnant women, and other groups. The network of businesses that bring food to American forks looks increasingly like the networks that bring other retail products to marketsimilar in the sophistication of their management, in the intensity with which they employ research, in the way manufacturing and business services are integrated into production networks, in the way they are linked to international markets, and in the education and training of the people they employ. Food production remains a major part of the economy, but the productivity of farming has grown so quickly that farming now provides only 3 percent of all jobs. The number of farmers may continue to decline since productivity growth is likely to remain high in most farming activities. There is some concern, however, that inadequate attention has been paid to the problem of increasing efficiency in products other than wheat, corn, and other bulk commodities. The pizza parable discussed earlier provides some clues about places where productivity growth can be expected. New information technology, improved packaging, and a variety of other technologies can allow a greater variety of fresh products to be available at comparatively low cost. There are some clouds on the horizon. Increases in the scale of firms, from farms to grocery chains, could decrease competition. Foreign firms could continue to capture do mestic markets in processed foods and other highvalue food products.

PAGE 49

41 The following trends can be identified: Consolidation of farm ownership and ownership of food processing and retail operations is likely to continue. By the year 2000, the 14 percent of farms with annual sales of more than $250,000 are likely to be taking 80 to 90 percent of cash receipts. Farms are operated increasingly like businesses in other sectors. The share of managers and professionals in farming has doubled since 1976. Net imports of high-value processed foods since 1979 have grown so fast that their value nearly equals the U.S. trade advantage in bulk commodities. Foreign producers are also gaining markets in some food processing technology. Grocery stores are getting bigger and offering a greater variety of products. Some stores offer 20,000 products and cover 200,000 square feet. These superstores account for 28 percent of all grocery sales. Convenience stores, typically franchise chains, have grown rapidly to fill market needs not met by these super-stores. They are also major financial centers. Supermarkets cash one-third of all non-government checks. Restaurants and other food service firms now account for over 5 percent of all U.S. employment. Fast-food, typically in franchises or subsidiaries of large firms, has rapidly gained share. The fast-food format and a variety of technologies like microwaves and deep-fat fryers increased productivity until the mid 1970s, but productivity has not grown significantly since. Productivity gains in food manufacturing will continue, but research levels may be inadequate to the challenge. Packaging technology seems particularly critical both to improve quality and to reduce costspackaging costs more than the contents for a growing number of products ranging from beer to breakfast cereal. Perhaps most critically, information technologies have the potential to improve the performance and reshape relationships throughout the Food system. Data entered automatically through the scanners that price more than half of all groceries purchased today can improve the productivity of check-out clerks. More importantly, the data entered drives an information system capable of reducing paperwork and improving the efficiency of activities throughout the system. It allows rapid price changes, measurement of consumer response to advertising, closer inventory control, precise dispatch orders for trucking, improved coupon management ($2 billion in coupons were redeemed in 1983) and reduced error rates in billing, ordering, and pricing. Jobs throughout the Food system are already in flux. Automation is replacing much labor in food preparation and is demanding new sets of skills. Consolidation of wholesale and retail chains, coupled with automated equipment, has eliminated many jobs and has resulted in the substitution of part-time for full-time work. The elimination of skilled jobs in grocery stores (meat-cutting in particular) has reduced employee leverage in bargaining for wages. In spite of the tradition of independence in American farming, the government is heavily involved in the enterprise. Major Federal programs support prices, regulate international trade, and subsidize research. Farm production commands a significant amount of basic research, a large fraction of which is supported by public funds. The research program has an exceptionally fine record in moving ideas from the laboratory to productive practice. Government involvement also shapes the geography of Food production. Subsidies for water supplies and price supports have allowed production of commodities like wheat and corn to be profitable in parts of the country that might not otherwise support such production. Health (11 percent of personal and government spending) Consumption recipes for good health involve increasingly complex decisions about investments in health promotion (e.g., appropriate exercise), disease prevention (e.g., changing smoking and drinking habits, controlling hypertension), public and private investments in environmental controls and safety devices, and investments in clinical treatment of medical problems. In 1983, approximately $350 billion was paid to health enterprises; $17 billion was spent for self-care medication, fitness equipment, and nutrition aids; and $50 billion for environmental controls. It seems possible to improve national health status without significant increases in spending given

PAGE 50

42 a way that provides a better match between spending and amenity measured in terms of improved health and longevity. Calculations presented in chapter 3 show that average life expectancy in the United States could well reach 80 years without extraordinary breakthroughs in medical technology. This would almost require greater incentives for individuals and households to take responsibility for their own health. Smoking habits, seat belt use, modest attention to diet, prenatal care, the strength of a persons social contacts, and other aspects of lifestyle have a major effect on life expectancy. Improved health would also require strong incentives for research, development, and innovation. It is also possible to imagine a system choking on regulations and red tape, a system bureaucratized to the point where patient care was compromised in the interest of avoiding litigation or following rigid guidelines. Pressure to reduce health care costs when an aging population needs more health care services could lead to a real decline in health status and increasing inequality among different segments of the population. It proves difficult to draw a clear line between consumption and production recipes for Health. It is clear that both are changing. A better understanding of the connections between diet, smoking, and other aspects of lifestyle and good health seems to have altered many American habits. The high cost and sophistication of medical technology have made management, research, productivity, paperwork, litigation, finance, and even advertising an important part of a business once dominated by independent private practitioners. These new tasks, coupled with new alternatives for women once willing to work at low wages as nurses, are reshaping the nature of work in Health professions, New technology, better consumer information, and improved management of health resources have improved the health status of Americans during the past few decades. Life expectancy in the United States has increased 6.5 years since 1950 (the penalty for living longer is that each person spends more time in some state of ill health). These averages belie great inequalities. Life expectancy in Hawaii is higher than in Sweden or Japan, while life expectancy in Louisiana is lower than in Singapore. Infant mortality rates in Washington, DC are close to those of the poorest developing countries. The differences are due partly to inadequate access to health care, and partly to the fact that differences in lifestyles contribute to differences in health. The comparatively poor U.S. health statistics occur in spite of the high quality of U.S. medicine because of defects in other parts of the American recipe for Health. The health status of the poor is often limited by a variety of social and environmental conditions beyond the reach of traditional health-related spending. A large number of households are not covered by any health insurance program, public or private. This group may grow as the work force turns to greater use of temporary and part-time workers. These people tend to be much sicker by the time they enter the health care system. American mortality statistics are, therefore, not necessarily a good measure of the quality of the U.S. medical facilities. They may instead be a measure of the availability of insurance coverage, inequality in income, racial discrimination, and other social phenomena beyond the reach of medical enterprises. Technology has obviously made a significant contribution to the quality of American health care. Surgical techniques continue to improve. Biotechnologies promise both new diagnostics and, hopefully, new treatments for a variety of conditions. Engineering advances permit prosthetics, artificial organs, and greater assistance for the visually and hearing handicapped. Expensive new technologies like Magnetic Resonance Imagery and Computer Assisted Tomography have centralized some activities into tertiary hospital centers. Other technologies have permitted treatments once requiring extensive hospitalization to be conducted in smaller clinics or on an outpatient basis, and have permitted home treatment to replace some hospital care. Implantable pumps, for example, allow controlled dosage of a variety of drugs outside of hospital settings. Information equipment may affect health care costs, health care quality, and the structure of Health businesses as profoundly as any of the technologies just listed. More sophisticated management systems, growing complexity of procedures, and an avalanche of paperwork connected with cost-containment and (increasingly) legal issues have made information management a major part of the health care busi-

PAGE 51

43 ness. Computer-based record management has improved efficiency and reduced costs in everything from patient billing to inventory control. It is now possible to store high-resolution photographs of specimens and other records electronically. Information equipment permits small home care firms to manage complex paperwork. New technology can permit rapid communication between specialists in different parts of a city, or indeed different parts of the country. It can permit telemetry to monitor patients being treated at home. The challenge of better integration of health treatment can be seen in the baffling statistics on use of different therapeutic techniques. The probability that a women will have a hysterectomy, for example, is 20 percent in some markets and 70 percent in others. At least one (possibly both) of these regional practices must be wrong. Information equipment should be able to permit more precise sharing and analysis of procedures. In some cases, information equipment can even provide a second opinion for a physician otherwise forced to rely on his or her own memory about techniques. The effects of these and other changes have already transformed the structure of production in health enterprises. Once dominated by general practitioners who would occasionally refer a patient to a hospital, the business of medicine now integrates individual practitioners, group practices, and tightly managed hospital corporations, as well as a growing array of satellite institutions: testing laboratories, out-patient clinics, hospices, nursing homes, homecare specialists, and even doc in a box emergency care facilities in shopping centers. The number of people receiving medical care from a health maintenance organization (HMO), as opposed to a privately practicing physician, nearly doubled from 1981 to 1985. Using modern management techniques, hospitals and other health organizations are beginning to behave like other service organizations. Enormous investments in environmental protection have created wholly new business opportunities. Cleaning hazardous waste sites alone promises to generate a significant amount of future business. The changes just discussed are also transforming the nature of health as a profession. Physicians are frequently hired on fixed salaries much as any other technical professional. A shortage of nurses is developing as women realize that other occupations result in greater returns for equivalent levels of experience and education, and typically offer less erratic hours. The cost of new technology, coupled with rapid growth of the elderly population and expanded pub lic insurance programs, have also led to explosive growth in health care costs and have spawned a number of regulatory remedies to control costs. While it may well be that Americans would freely choose to spend an increasing share of their incomes on Health, the question of whether spending levels are or are not appropriate cannot be resolved by a simple appeal to market forces. The task of designing a financing system that provides incentives to maximize the quality of health outcomes while minimizing costs has proved to be vexing for both public and private insurance programs. By the early 1980s, there was general agreement that the incentive and regulatory systems in the health industry needed revision. Both private and public insurers moved away from fee for service reimbursement. The challenge has been to find a substitute that does not collapse under ever more invasive regulation. Government and private insurance regulations have attempted to force greater attention to the costs and benefits of different treatments. As a result there has been some reduction in patient use of hospitals and physicians. Hospitalization rates for prepaid plans, where hospitals have a strong incentive to minimize costs, are 10 to 40 percent lower than those for people with insurance that pay fees for services. Physician visits per person have fallen since 1976. Housing (20 percent of personal and government spending) Owning a detached home remains a firm part of nearly every Americans vision of a good life. Decent housing is largely a matter of taste, but it surely includes security, a pleasant landscape, low maintenance, and low operating costs. It also means a pleasant home with access to jobs (frequently jobs for several members of the household), shopping, schools, and day care facilities. Changes in taste and demographics also redefine definitions of housing quality. The rapid growth of the elderly population, for example, creates new

PAGE 52

44 housing needs. Fully 40 percent of Americans aged 65 to 74 have some kind of disability that limits their activity, and 63 percent of those over 75 have some disability. Shrinking leisure time has made the amenities of a comfortable home substitute for other forms of recreational spending. Homes have also become centers for home health care, and even places of part-time work. Housing is, of course, an investment in addition to an amenity. Home ownership differentiates those with assets from those that lack them in contemporary America. The net worth of households owning homes is nine times higher than the net worth of renters and home equity represents two-thirds of the net worth of home owners. The high mobility of Americans reinforces the tendency to consider a home a temporary investment, with purchasers paying as much attention to the hypothetical desires of the next buyer as to their own. On average, the quality of American housing is improving. The average home has more rooms per person, and is more likely to have air-conditioning, a full garage, and a well-equipped kitchen, than ever before. On the other hand, home ownership has moved beyond the means of many young Americans, particularly those moving to areas where employment is expanding rapidly. The fraction of median household income needed to purchase a new home is more than 50 percent above what it was in 1968. Housing costs are also an increasing burden for poor families, particularly homes with a single wage earner. The cost of buying the average new house would require 75 percent of the average income of single women. One result of these trends has been an increase in renting. The extreme symptom is growing homelessness. Many households have found affordable housing only at considerable distances from their jobsa major problem for households with two earners. In effect, housing costs are traded against the cost of time spent commuting. Only half of the people in the San Francisco area now work and live in the same town. Since men commute 35 percent more miles and spend 15 percent more time commuting than women, it appears that women may be taking jobs closer to home, possibly sacrificing income in order to combine lives as a homemaker and paid employee. The high cost of housing results from several factors: l l l l Land costs have soared, particularly in rapidly growing urban areas. Growth constraints have exaggerated the problem in many areas. Changes in regulations have forced purchasers of home mortgages into competition with wider financial markets. Financing expenses remain a dominant housing cost. Productivity has fallen in construction. Rising energy prices have increased the price of home operations. The costs associated with land prices are obviously not touched by changes in the production recipe just described. They require changes in the efficiency of transportation, changes in tax laws that may encourage speculation, and policies that may affect the location of jobs, retail facilities, recreational facilities, and other destinations. The recipe by which houses are produced has not changed significantly in some time, In spite of forecasts about factory-built housing dating to the early 1930s, little has actually happened. Radical swings in demand for new housing, and a history of fragmentation in home construction businesses, has kept research and capital investment in this industry far below that of most manufacturing firms. There are reasons to believe, however, that significant change is not only possible but necessary. The U.S. housing industry could follow the trends of the industry in Sweden and Japan, moving construction away from highly fragmented small firms with few investments in capital equipment or trained employees, and toward production networks that more closely resemble other manufacturing operations. The result would be a profound change in the nature of the institutions supplying Housing in the United States, as well as a change in the quality of the product: l The house itself could be designed on the basis of more systematic research. At present, most innovations come from component suppliers, with little research directed at the integrated performance of the structure itself. There is compelling evidence that simple, inexpensive changes in home design make new homes not only more

PAGE 53

45 l l l l comfortable but less expensive to operate. Factory construction could lead to both greater reliability and energy efficiency. Houses could be brand name products with multi-year guarantees of reliability. Productivity of home manufacturing could benefit from greater capital investment in production facilities and the use of advanced assembly techniques. The home building industry invests little in improving production technology. Homes could be assembled by firms specializing in site preparation and rapid site assembly of factory components. These firms could either be independent or franchises of larger firms (as they are in Sweden). Homes could be marketed from showrooms, where prospective clients can design floor plans and explore alternative exterior and interior finishings on computer screens. The Japanese claim to be able to convert designs into finished houses in 3 weeks. Changes of the sort just described could radically change the nature of construction work. The construction industry has always achieved a high degree of flexibility by using temporary work teams. It has an extraordinary ability to assemble many disciplines quickly. Changes in production recipe could make construction work more closely resemble jobs in other manufacturing industries. (Of course, it is also likely that manufacturing enterprises will change in ways that may make them more closely resemble construction in the way work teams are assembled on short notice.) If the construction industry began to resemble the organizational structure of manufacturing, a large number of design, research, and customer support jobs would be created. Sales operations would combine design, finance, and marketing. Production work would involve programming equipment to produce unique products, maintaining sophisticated equipment, and learning to operate new tools. The changes in construction networks could, however, also be used to replace skilled construction jobs with low-paid, low-skilled employees in production facilities. Many factory construction facilities in the United States use temporary labor. The U.S. home building industry has largely been shielded from foreign competition but the situation has changed rapidly. Foreign appliances, fittings, and other components are entering U.S. markets rapidly. There is a real risk that without some significant change in the productivity of U.S. construction techniques, U.S. firms could become little more than assemblers of foreign products using foreign-made tools. Without some change in housing technology or policy, a growing number of Americans may find themselves faced with painful choices between cramped quarters and long commutes. It is possible that productivity would continue to decline, and U.S. home owners may be burdened with continuing increases in home prices, and with inefficient structures in which operating costs could rise sharply if energy prices increase. Since residences consume nearly 30 percent of all electricity generated by U.S. utilities, and are largely responsible for the peak demands that motivate new plant construction, there is a critical need to communicate accurate costs to residential customers. Residences also consume approximately 20 percent of U.S. oil and gas. It should be possible to cut energy consumption of new homes in half with improved windows, lighting, appliances, better heating and airconditioning controls, and advanced materials. More over, communications technologies are being tested that can help households adjust electricity use to minimize cost both for themselves and the utilities. This will require pricing electricity not as a commodity but as a product whose value depends on the season and the time of day. Transportation (11 percent of personal and government spending) The amenity of Transportation is measurable by the extent to which people can go where they want, when they want. This has become almost synonymous with travel by automobile; 86 percent of all work trips in cars were made alone in 1983, and the driver was alone in a car for 60 percent of all vehicle miles traveled. In spite of massive public investments, travel on public transit continues to decline. The comparatively low densities of most U.S. living and working areas makes transit a poor alternative for all but the most densely populated urban centers, such as New York, Chicago, Boston, or Philadelphia. Mass transit is often used only as a last resort. One-third of mass transit passengers live in

PAGE 54

46 households that do not own cars. The central challenge of personal transport in the United States remains one of preserving the amenity inherent in a personal vehicle while providing needed mobility for those unable to drive. Complex networks of production have altered demand for business transportation services (typically demanding more reliable and timely deliveries), and have changed the geography of production and the location of jobs. Demographic changes and changes in housing costs have shifted the location of housing. The result is new patterns of demand for both personal and business transportation. In 1980, 60 percent of the 31 million commuters in the Nations 25 largest urban areas lived in suburban areas and traveled to jobs outside the central business district. There is now one car for every licensed driver. Public transport is increasingly dominated by air travel, which has appeared as a substitute for long bus and train trips. There has been little net improvement in the real efficiency of auto travel. The real costs of driving a mile have not changed significantly in 15 years. Imports of inexpensive and durable cars increased competition for domestic markets and placed new emphasis on quality and fuel economy. These improvements have largely offset the higher cost of fuel. On the other hand, travel times to work and shopping remain comparatively unchanged because of a combination of congestion and decentralization. Little can or will change without both private efforts to improve personal vehicles and a public effort to provide improved highways. The recipe for Transportation plainly involves a coordination of public and private investment. Transportation provides an enormous variety of jobs, ranging from high-skill/high-wage positions for airline pilots to minimum wage jobs in gasoline stations. Since personal transportation is by private automobile, Transportation provides as many jobs in manufacturing as it does for paid vehicle operators. It should be possible to make the Nations Transportation system faster, more flexible, less costly, and safer by increasing the variety and capabilities of personal vehicles. High-performance, two-passenger vehicles could be given privileged commuter lanes and parking. Low-speed vehicles with minimal licensing requirements could provide mobility for the elderly or young teenagers in communities. Improvements are also possible in the coordination of traffic signals, and better information on congestion and road conditions can be transmitted to individual vehicles again, a combination of private and public equipment. Improved communication systems may permit more people to work at home, but it appears unlikely that this will significantly reduce transport needs. On the contrary, communications technology may increase travel, by making more people aware of opportunities for business, shopping, and recreation. Cellular telephones can even make cars double as offices. The near doubling of fuel economy of automobiles played a major role in reducing U.S. and world petroleum imports. It is possible to achieve another doubling or even tripling in fuel economy without significant sacrifices in vehicle performance. A coordinated movement to a fuel other than gasoline would also ease pressure on the system. This, too, is likely to require public as well as private investment if such a transition is to occur before a crisis strikes. Without some imagination to break the current stagnation in the performance of the personal transportation system, it is possible to imagine a system that not only fails to improve but that offers declining levels of amenity. The system could become increasingly congested and poorly matched to the diverse needs of a complex society. It could also fail to serve the needs of production systems dependent on fast, reliable, flexible transport systems. Without adequate investment, the existing Transportation infrastructure could deteriorate, and congestion along existing highways could increase. A system heavily dependent on oil could be very vulnerable to the availability and price of foreign oil. Clothing and Personal Care (7 percent of personal and government spending) Americans are clearly impressed by the symbolism of clothing, and clothing purchases rise sharply with income. Americans consume twice as many square yards of cloth per person as the average French or German consumer.

PAGE 55

47 The technology of production in textiles and apparel is going through revolutionary changes. Computer-driven shuttleless looms and cutting machines now dominate the market. Apparel assembly, which the industrial revolution seems to have bypassed, may soon be automated using robotic sewing equipment now entering the market. Perhaps more importantly, the network connecting producers together appears to be on the brink of a major change. The change is forced both by new technology and the pressure of imported products that have devastated markets for domestic producers in a growing number of products. The industry is moving quickly to integrate operations from fiber production to retail outlets. The goal is to make a radical reduction in the 65 weeks now required to move fiber to a retail store in the form of finished apparel. This is done through adoption of standardization, responsive transportation systems, and production equipment that allows high productivity in comparatively small batch production. These responsive systems allow retail outlets to carry more products in more sizes without increasing inventories. The cost of producing products tailored to a persons unique measurements may soon be little more than the cost of mass-produced garments. If this occurs, the industry will have come full circle from tailoring to commodities and back to tailoring in a hundred years. Ironically, the productivity of individual steps in this network (e.g., trucking or weaving) may decrease while the productivity of the system taken as a whole may increase. Given the intense nature of foreign competition, it seems apparent that the network delivering apparel to U.S. consumers will either have to undergo major improvements in productivity or become little more than a vendor for foreign products. On the other hand, it is entirely possible that a flexible and responsive domestic system can provide customers with fashions and service impossible to provide from foreign suppliers. Whether reshaped by trade or by new technology and management systems, employment in the collection of enterprises just described will change. The trend is already evident. Between 1977 and the end of 1987, employment in the apparel industry fell 16 percent and textile employment fell 18 percent. Heavy investment in modern production equipment has eliminated many jobs in fiber production and textiles. Automation will also reduce jobs in apparel assembly. The jobs eliminated will largely be those held by people with comparatively low skills. Many of them are minorities with few resources for other jobs. The jobs created will require skill in managing orders, maintaining and programming sophisticated equipment, and other functions likely to require a significant amount of specialized training. Some effort will be needed to provide people now employed by the industry with the new skills needed. Job loss in textile and apparel production is partly offset by rapid growth in apparel retailing. Retail apparel and accessory stores provide jobs for nearly 1 million people, and over 2 million people are employed by department stores. In spite of the importance of this sector of the economy, the U.S. Government spends few resources on research and development work for textile or apparel technology. Education (7 percent of personal and government spending) Education is both an end and a means to an end. It provides tools critical for finding rewarding employment in a complex economy. Education also proves to be strongly correlated with health, and with an ability to recover from such personal disasters as illness, divorce, or a job loss. It opens doors to cultural opportunities that are otherwise inaccessible. Education is also one of the Nations largest enterprises, Between $300 and $500 billion are spent on Education each year (the range results from an inability to count corporate investments in Education). The total investment in learning exceeds the $240 billion annual private investment in buildings and structures. About one American in three is enrolled in some kind of educational program during the year. The Nations educational system may be on the brink of a major change. This results both from a fundamental change in the demand for educational services and from the fact that new technology makes it possible to consider real improvements in the productivity of both teaching and learning. Taken together, these forces could change what is taught, when it is taught, where it is taught, and the nature

PAGE 56

48 of teaching as a profession. There are some stark alternatives: l l The system could change in a way that makes learning more productive and fun while allowing teachers more time to spend with individuals as coaches or tutors. It could put more power in the hands of the learner, tailor instruction to each persons level of understanding and learning speed and technique, and make it easier for an individual to learn when instruction is most needed. The system could create rigid centralization of course design, mechanical-and impersonal instruction, national regulation, and a contraction of choice for both students and instructors. Adult training programs could widen the gap between those with good educational skills and those lacking them. The difference between the two depends heavily on public choice. Demand for Education has changed both in qualitative and quantitative terms. Demand for formal educational services (K-12 and college education) de clined with the graduation of the baby boom generation, while demand for adult training has increased. Not only do the emerging jobs require a higher level of educational skills, but jobs are changing in ways that require a continuous renewal of skills. The kinds of basic skills required of entry-level employees are reshaped by these new demands. Skills in working with people in groups, in self-education, in coping with ambiguity, and in coping with too much or too little information become requirements as important as any specialized training. The real costs of employee training are almost impossible to measure. Formal training can be tracked, but the costs of informal training and of problems created by inadequate training (e.g., Three Mile Island) are more ambiguous. They are obviously very large. It appears, however, that people most likely to benefit from formal and informal on-the-job training are those who come to the position with a good basic education. If anything, retraining appears to be widening, not reducing, the gap separating highly skilled from unskilled employees. The basic characteristics of Americas educational system have not changed significantly in a century, and each student continues to have unique learning styles and interests. The existing educational system appears to be highly personal and differentiated. However, the isolation of individual instructors may have the opposite effect. Studies indicate that a typical student in primary or secondary school (grades K-12) receives less than 1 minute per day of individual attention. Teachers throughout the country teach nearly the same course from similar texts, and spend large fractions of their time repeating standard lectures or performing routine administrative functions. Apart from buildings, desks, and blackboards, virtually no capital equipment is used in this, the Nations largest information enterprise. Television has not proved to be a major asset. Over the next two decades, capital equipment could improve the productivity of Education at all levels given adequate research and capital investment. New information technologies allow active response to students not possible with televisiona critical distinction. Existing computer-based instruction systems are often disappointing, primarily because of the limitations of existing software. But there is reason for considerable optimism. Artificial intelligence techniques could help diagnose defects in what a student understands, while less exotic software could help with needed drill and practice. Simulations that use images, sounds, and text can lend realism, and can remove the barriers of abstraction that so often impose themselves between formal education and practical mastery of a subject. Indeed, since the working life of many graduates will increasingly be spent viewing reality through computer screenswhether they are operating nuclear powerplants or analyzing a commercial insurance policythe distinction between simulation and reality can be extremely small. Communication systems could forge links between teachers working in similar areas, permitting specialization not possible in comparatively small schools. They could also allow students to work more easily with each other, and could help tie homework to school work. Significant changes are likely in the nature of teaching as a profession even if technology is not widely introduced. There are new demands for instructional quality and efficiency. At the same time, women who might once have entered teaching be-

PAGE 57

49 cause they had no other professional opportunities are now being drawn into a variety of other occupations. Introduction of more capital equipment and expansion of teaching beyond traditional school settings will lead to much greater differentiation in an occupation now called teachers. Tasks involving software development, equipment maintenance, holding tutorials, working with students around the country in a highly specialized area, identifying learning pathology, and a range of other areas will lead to jobs for a variety of specialists, and doubtlessly a variety of pay scales. Firms ranging from insurance underwriters to producers of metal parts have found that the potential efficiency gains from new information technology cannot be captured without a profound change in management strategies. Education will be no different. Management reforms are, however, notoriously difficult in Education because of the sectors fragmented nature. Hundreds of State and local governments manage K-12 instruction. This has a healthy effect by allowing teachers considerable power in resisting grand schemes unlikely to further instruction. In some cases, however, the extreme fragmentation can frustrate progress. Corporations and the Armed Forces have moved most rapidly to exploit the advantages of new instructional technology. This is due in part to a different style of management, and in part to the fact that since these organizations pay salaries of both students and teachers, they are as interested in the productivity of a students time as the productivity of the teachers time. As in many other sectors examined here, making full use of the potential of emerging technology requires a major investment in research. At present there is no national center for focusing research on education equivalent to the National Institutes of Health or the Agricultural Research Center. By far the bulk of research on educational technology supported by the Federal Government is undertaken by the Department of Defense. While a private information company typically spends several percent of gross revenue on research, virtually nothing is allocated for research directed at the real problems of teaching and learning. If the fraction of gross expenditures invested in research were the same for Education as for the average privately owned business in the United States, about $9 billion a year would be spent for Education research. This is 60 times more than the present allocation. Personal Business and Communication (6 percent of personal and government spending) to 90 Astonishingly little is known about the volume or the function of information entering or leaving U.S. homes. By far the greatest volume of information is broadcast to homes through television, radio, and direct mail. Large amounts of information enter through selective purchases or rentals of video tapes, magazines, books, newspapers, and cable television. Two~way point-to-point communication, involving telephone conversations and correspondence, is responsible for a comparatively small fraction of the volume of information, though this activity traditionally receives much more attention. Most of this information is used for entertainment. The economics of information are curious, since much of it is subsidized by advertisers. Advertising pays for nearly all broadcast costs, and 50 to 75 percent of the cost of newspapers and magazines. The costs paid by individual consumers come entirely in what they pay for receivers, and in the time spent watching advertising messages. Apart from entertainment, information is used largely for personal business-primarily for retail purchasing, but also for financial transactions and bill paying. Keeping in touch with friends and even gossiping represent other critical markets. Direct marketing has enjoyed an expanded business, using television and direct mail to communicate from businesses to homes, as well as 800number telephone systems and credit cards to communicate from homes to businesses. In the future, some of this electronic shopping will certainly become part of computer-based systems. One key barrier to computer-based systems is the lack of serendipity: when people enter files looking for something in particular, spontaneous response to products is not possible. Cable television has introduced a novel way to bypass this problem, by holding what amounts to a continuous sale where proprietors take telephone orders for products shown. How should amenity be measured in this case? By analogy to Transportation, improvements in in-

PAGE 58

50 formation resources depend on a persons ability to get the information he wants when he wants it (leaving a little room for serendipity). It seems that this system is moving rapidly toward one with greater consumer choice, more control over the information that enters the home, and more direct payments for services actually used. VCRs, cable television, and the explosion of specialized magazines have all allowed special tastes to flourish where broadcast television and general interest magazines once dominated markets. Similarly, the once undifferentiated telephone has been replaced by a variety of specialized products, ranging from $6 disposable phones to ports of entry into sophisticated data management services. Advanced communications to homes can also mean more control over electricity bills, better home health care services, and greater opportunities for education, while cellular communications can extend such freedoms to private vehicles. Many more services are likely to be available as the struggle over regulation of telephone services is resolved. The deregulation of banking, coupled with sharp increases in the incomes of many households, has generated a boom in individualized financial services. A growing fraction of all households use tax accountants and financial advisors to provide specialized services. Few now buy whole life policies that combine savings with insurance. While deregulation has spawned an explosion of competition, it has not had the effect of creating much real diversity in most markets. AT&T, for example, retains 90 percent of the long-distance market. Local telephone companies maintain regulated monopoly control over local telephone services, and are gradually being allowed to compete with other firms for lucrative advanced communications services. There is a possibility that these firms could expand on their regulated business to become regional monopolists in a variety of unregulated areas. The businesses in this group provide jobs dominated by clerical and information management occupations. The complex pattern of change possible in these occupations has already been discussed. In some cases, modern communication equipment has fragmented jobs into narrow functions that can be carefully monitored by electronic means. This has been true for telephone operators or clerks working on routine insurance accounts. In others, clerical tasks have been upgraded and combined with quasi-managerial tasks. This is particularly the case when products cannot be handled using mass production techniques (e.g., commercial insurance). Between 1983 and 1986, employment growth in clerical and other administrative support personnel slowed in finance, insurance, and real estate businesses, while jobs for people classified as managers, professionals, and sales workers increased rapidly. There is much uncertainty about the future of employment in these occupations, and much room for choice. Recreation and Leisure (9 percent of personal and government spending) Of all the sectors examined in this analysis, the amenity of Recreation and Leisure is most difficult to measure objectively or even define precisely. How can the recreational value of dining be separated from biological necessity? How much of a purchase of a home or a car can be considered recreational? Obviously, both time and money are critical when considering this amenity. Free time available to Americans has been shrinking for at least a decade and, as a result, spending for Recreation and Leisure has typically become more intense (more spending per unit of time but less time). Moreover, Americans earn far fewer vacation days than their counterparts in Europe. U.S. manufacturing employees averaged 29 days of paid vacation and holidays in 1982, while the West Germans took 41 days and the French 35. Outdoor recreation and travel command a firm share of U.S. leisure time in spite of advances in home electronics. In fact, home electronics may have expanded interest in travel and in a greater variety of away-from-home activities. Time constraints remain a bigger problem than money for many families. Vacations tend to be short and intense, and many households now look for short trips close to home in lieu of the traditional two weeks at the beach vacation. There is little sign that national tastes in recreation are becoming more homogeneous. Business is booming for everything from theme parks to National Barbecue cooking contests to Mississippi Catfish Festivals. Hotels and motels cater to specific client groups, such as older couples or families with children.

PAGE 59

51 In the future, technical capabilities of home electronics should continue to expand quickly. investments in home capital equipment now include video equipment (where high-definition TV can greatly improve picture and sound quality) and sophisticated audio equipment (digital disks and tapes offer remarkable improvements). New information technologies also promise to expand opportunities for travel. Airlines were among the first to use sophisticated national reservation systems. Travel agencies (an industry that has been growing rapidly in recent years) offer growing variety of information services. Reservation services can be provided for hotels and motels, automobile rentals, sports and theater events, and a variety of other activities. Some agencies are introducing expert systems to help match unique interests to travel reservations. In the near future, reservation systems will permit a video tour of prospective destinations, hotels, and even individual rooms. Prototypes are already in place. It should be possible to enrich opportunities for using leisure time through the use of advanced technologies, which can bring a greater variety of entertainment, education, and shopping services to individual households and improve access to recreation and entertainment facilities away from home. On the other hand, recreation and entertainment opportunities could be curtailed for all but the affluent, given a decline in advertiser-supported TV and newspapers or a decline in public support for parks, town centers (as opposed to privately managed shopping malls), and cultural facilities and activities. Many outdoor recreation activities involve an integration of publicly supported infrastructures, such as parks, beaches, civic centers, and sports stadiums, and private investment, such as hotels, travel agencies, and theme parks. Many businesses depend on public investment in these infrastructures, and on public efforts to maintain the quality of the environment in parks and other recreation areas. There may also be a role for public information services such as the city of Baltimore, Marylands INFOTOUCH. Not counting jobs in transportation services or jobs in restaurants, the Recreation and Leisure industries are responsible for 8 percent of American jobs. Many, like those for hotel employees or employees in theme parks, are comparatively low paid and are in occupations where productivity increases are difficult to envision. In fact, hotels appear to be becoming more, and not less, labor-intensive. On the other hand, the production and distribution of consumer electronics and the software used on these systems can create many rewarding jobs. American production of many consumer electronics products has all but disappeared in the face of foreign production, Foreign producers have used their command of manufacturing the current generation of consumer electronics products to fund the development of a continuing series of innovations. U.S. firms will need to move aggressively to recapture these promising new markets. Pulling the Pieces Together (Ch. 13) The kinds of change discussed so far will have far-reaching effects on the future of the U.S. economy. Unfortunately, none of these effects can be confidently forecast because virtually none depends on immutable natural forces. On the contrary, the Nations economic future depends as never before on choices made by households, producers, workers, and the government. These choices will, of course, also be shaped by the regulations and incentives the government oversees, and by the skill with which the government manages international affairs. Forecasts where so much depends on choice are absurd. What can be done, however, is to sketch out self-consistent descriptions of the different kinds of economic structure that could emerge in the United States during the next two decades. The implications of public and private choices become clearer when displayed in this way. Four scenarios are used in this report to explore the implications of alternative futures for the U.S. economy. These postulate: 1 2 3 4 continuation of present trends, a recovery of the manufacturing sector, a stagnating economy, and a transformed economy making optimum use of new technology. The results range from cases where GNP grows slowly (1.5 percent per year) and unemployment reaches 23 percent (measured in conventional terms) to one where GNP grows rapidly (3 percent per year) with possible labor shortages. While there seems to

PAGE 60

52 be no question that the second case is preferred, this quick judgment should perhaps be suspended while some of the fine print is examined. A case could be made that GNP growth does not necessarily provide a good guide to progress. Indeed, the scenarios reveal that the choices made by consumers can greatly affect the level of amenity they achieve, particularly in areas such as Health, mobility, and Housing, even if their incomes do not rise. Improved amenity achieved with a working week of 35 hours (which would drastically reduce unemployment) may look better than 3 percent GNP growth that fails to raise living standards equitably. With careful management, however, it should be possible to have both rapid growth and improved amenity. Changes in personal and government consumption recipes, and in income distribution, have a surprisingly small impact on the economy in terms of the share of value-added or jobs in manufacturing or other sectorsalthough individual enterprises would obviously be affected much more strongly. The different scenarios also reveal interesting information about the future of natural resources and energy consumption. Changes in household and government spending recipes can have a significant impact on national use of energy. The combination of greater attention to energy efficiency by consumers, and production recipes that demand less energy and material inputs, can lead to a significant decline in national use of resources. Optimal use of new technology could result in a 40 to 60 percent decline in use of natural resources, even when there is rapid (3 percent) economic growth. Since air and water pollution, and the generation of hazardous waste, scale roughly with demand for energy and material resources, these differences obviously translate into large differences in future environmental quality. The scenarios thus demonstrate that it is possible to have vigorous economic growth while reducing demands on the natural environment. Indeed, most innovative production schemes result in a decline in environmental releases simply because most new technologies are intrinsically more efficient. The impact of trade depends both on whether trade volume continues to grow as a fraction of GNP (in constant dollars) and on the nature of U.S. exports and imports. High levels of trade can lead to a net increase in employment since in many cases the number of jobs generated per dollar of exports exceeds the number of jobs lost per dollar of imports. Employment in manufacturing and agriculture is particularly vulnerable to different patterns of trade. Changes in production recipes, and assumptions about productivity in each industry, have the largest effect on the industries likely to produce jobs in the future. Changes in the type of occupations required in individual industries, however, have an even more profound effect on the number of jobs created in each occupation (see figure 1-24). All the scenarios share some characteristics. Employment in manufacturing during the next two decFigure 1-24.-Why Occupation Shares May Change Managers Technical Professionals Educational Professionals Other Professionals Tech nicians Sales Workers Other Customer Contact Information Data Entry Food and Beverage Other Service Precision Craft Machine Operators Transportation Handlers & Laborers Farming & Forestry

PAGE 61

53 ades will not decline as rapidly as it has in the past 20 years. Manufacturing employment seems likely to fall from about 20 percent of all jobs to between 16 and 18 percent in 2005. Measured as a (constant dollar) percentage of GNP, manufacturing may fall by only 1 to 2 percentage points, with declines occurring primarily in manufacturing sectors that now pay comparatively high wages. Also breaking a trend, jobs in transactional activities (such as banking, insurance, law, and business services) seem unlikely to grow substantially as a fraction of all jobs. This results in part from the fact that productivity may grow sharply in paper pushing activities, thereby reducing the work force required. The characteristics of employment possible in the future differ most strikingly when measured by the type of job created. An economy that manages to build growth primarily around a revitalization of traditional manufacturing produces many more craft workers, precision production operators, and other hands-on manufacturing occupations. An economy moving sharply to new production networks generates many fewer jobs of this type, twice as many jobs for technical professionals (compared with an economy following 1984 staffing patterns), nearly 25 percent more managers and management support personnel, and 30 to 40 percent more sales workers. There would also be a 30 percent decline in jobs involving routine data entry and manipulation, and fewer jobs for the most low-paid occupations (food preparation, low-paid service workers, laborers, and farm labor). The transformation scenarios have the effect of reducing employment in categories that traditionally pay low wages (the fraction of all workers now in occupations paying less than two-thirds of median wages could fall by 40 percent) while increasing employment in traditionally well-paid jobs (the fraction of all workers in occupations now paying more than 33 percent above the median wage could grow 50 to 60 percent). Of course, nearly all these occupations would be redefined if such a transformation took place. Many managers are likely to have more routine jobs than todays managers. Sales workers in the future could have more demanding jobs if they were more tightly integrated into inventory control and production networks, and given greater control over the design and tailoring of products sold. The statistics say little about the quality of the working environment, an employees control over working life, or satisfaction with what is accomplished. It does appear, however, that a restructured economy could generate large numbers of interesting and rewarding jobs, ranging from interesting face-to-face sales positions, to technicians installing and repairing a continuing series of new hardware and software, to managers wrestling to build new teams for specialized projects. These jobs will be in occupations that currently require a high level of education. The educational attainment level experiencing the fastest increase in its share of jobs, both currently and under the Transformation and Trend scenarios, is the highestfour or more years of college (see figure 1-25). On the other hand, it is also possible to produce jobs that allow individual employees few initiatives and monitor performance with Or well ean precision using advanced communication networks. The combination of technology and large numbers of entrylevel people with poor basic education could force (or tempt) employers to produce large numbers of jobs requiring minimal training for people considered interchangeable and disposable. Figure 1-25.-Educational Requirements of Future Scenarios 10 15 20 25 30 35 40 45 Percent of all jobs 1984 ~ Manufacturing = Trend m Transformation ------------Scenar ids ---------How To Read This Figure: College graduates held about 21 0 / 0 of all jobs in 1984 but would hold nearly 28/0 of all jobs under the assumptions of the Transformation scenarios. SOURCE Office of Technology Assessment (see figure 13-2 of ch. 13)

PAGE 62

54 A Concluding Note jor types of businesses, the relative position of the The calculations just presented, coupled with the United States in the world economy, and the nature and rewards of jobs available. The range of possimore qualitative descriptions that preceded them, describe starkly contrasting futures for the U.S. econbilities is shown not as forecasts but as possibilities possibilities that illustrate the power of and the need omy. They paint very different pictures of the future for careful choice in the construction of public policy. of critical amenities, the potential for growth in ma-

PAGE 63

Part I The Networks of Consumption While many of the most basic aspects of economic structure are in flux, human needs and desires provide a fixed point of reference. Whatever happens in the future, the output of the U.S. economy must ultimately be measured by the extent to which it allows Americans to achieve amenity: to be in good health, to have a varied and healthy diet, to be well clothed, to live in attractive housing, to receive a useful and interesting education, to expand options for personal communication, to travel, and to enjoy leisure time. Together with national defense and other functions necessarily provided by collective rather than individual purchasing, these amenities represent the real output of an economic system. The net productivity of the U.S. economy must be measured by the efficiency with which human time and talent are applied to their achievement. The next two chapters define methods for measuring growth in amenity, and describe the consumption recipes used by Americans to achieve these amenities. These consumption recipes combine household purchases of goods and services, investment of unpaid time by household members, and public spending. The chapters provide a basis for describing possible changes in purchasing which, in turn, affect the structure of the producing sectors of U.S. economy, and ultimately the jobs that the economy creates. It is not possible to develop a rigid definition of quality in any of the major classes of amenity. Concepts of quality are often readily understoodfreedom of choice, good health, or happinessbut may be impossible to quantify. Some purchases are based on need, and some on choice. Some are spontaneous while others are induced by advertising. Many are shaped (or misshaped) by regulation, instinct, ignorance, or hasty decisions. Even if it were possible to develop a perfect way to measure amenity for an individual, it would be impossible to develop a perfect formula describing amenity for society as a whole. ] Once incomes are adequate to provide for Kenneth Arrow argues that even if a function could be written to describe the way each individual would rank the value of different patterns of expenditure, it would be impossible to combine these functions to develop a method for ranking expenditure patterns for the group taken as a whole. K. Arrow, Social Choice and lndividua] Values (New Haven, CT. Yale University Press, 1963). basic necessities, the extent to which spending buys satisfaction depends heavily on expectations. The quality of amenities available to a person considered wealthy at the turn of the century is below the expectations of even the poorest family today. On the other hand, the well-educated baby boom generation may have expectations that will be difficult for the economy to meet. 2 The problem of defining and measuring amenity can, of course, be avoided if growth in national income can be used as a proxy for economic progress. Given this assumption, productivity can be measured by the efficiency with which labor and other inputs are converted into measurable quantities like CAT scanners, and not by the facility with which resources achieve longer lives or greater freedom in personal transport. Yet while such an approach solves many analytical problems, it is insufficient for the purposes of the analysis that follows for two critical l reasons: Measures of amenity that are independent of spending levels are needed to consider public policy choices affecting the way consumers combine goods and services. There are obviously many cases where income measures of economic welfare are not adequate. Defects in public regulation or inappropriate public spending can be responsible for inefficiencies in the way consumers convert money to amenity. The quality of life in a community can decline while incomes increase if environmental quality deteriorates or social unrest undermines security. There may actually be a negative correlation between spending for health care and life expectancy, or between spending for burglar alarms and security. Policies designed to facilitate private choice can only be assessed given a clear understanding of the performance of networks that connect spending with amenity measured in human terms. 2A straightfo~ard calculation shows that if the baby boom generation is to be paid as much as the previous generation on the basis of age and education, the U.S. gross national product (GNP) would need to grow at 2.5 percent per year. This rate of growth is needed if the baby boom generation is to fulfill its expectations for their investment in education. W.H. Esselman and OS. Yu, Economic Growth to Meet Income Expectations, Journal olfoky Artafvsis and Management, vol. 2, No, 1, fall 1982, pp. 111-118. 55

PAGE 64

56 l Analysis of trends in spending patterns may not provide a good basis for anticipating spending patterns over the next two decades. Technology can create: new products and services, radical declines in the prices of existing products, an increased ability to tailor products to individual needs, new sources of information about products, new retailing methods, changes in time available for making purchases, changed tastes, and new government regulations affecting both price and quality. There is no obvious way to estimate consumer response. It is also difficult to anticipate changes in public expendituresnearly one-quarter of all personal and government spending combined. Defining the choices Americans have about the future is necessarily an inexact process. It is essential to begin with hypotheses about how public choices could affect the structure of future consumption recipes. The next two chapters combine a formal analysis of trends in U.S. consumption over the last two decades with an assessment of choices about the direction of future spending, which is necessarily much less rigorous, While generalizations are difficult, the discussion traces a number of themes affecting the recipe for consumption. All stem from a growth of choice made possible by increased national income, new technologies (particularly those that improve communication and those that reduce the price difference between mass-produced items and items tailored to narrow markets), and an increase in international and domestic competition. The themes include: the growing complexity of consumption recipes and the increased demand for sophisticated consumer decisions; the fragmentation of some markets into a series of niches and the integration of others into undifferentiated commodities; advances in communication, marketing, and retailing technologies, which can improve the match between specialized products and specialized tastes; the growing desire for quality in both products and services; the rising importance of purchased services, partly associated with an increased interest in specialized products; l l l a shift of activity formerly undertaken in the home to the marketplace, such as child care, coupled with substitution of household time for services formerly provided in the market, such as VCRs for movies and home health care for hospital care; the growing importance of timeas opposed to moneyas a constraint on the achievement of amenity for many groups; and persistent signs that large groups of the population fall far below standard norms in a wide range of amenities, including health, housing, transportation, and even basic nutrition. Chapter 2 begins by defining the eight categories of amenity that, taken together, constitute most of the output of the U.S. economy. It describes recent trends in private and public spending and personal time investments used to achieve each amenity, and the major forces behind these trends: income, income distribution, demographics, prices, new technologies, new patterns of regulation, and changes in taste. The chapter examines spending at the level of the economy as a whole, focusing on the major factors that influence consumer purchasing and applying standard methods to estimate how consumption may change. Among other things, it demonstrates how demand could change as a result of changed demographics, income and income distribution, and product prices, given consumer spending that follows trends established during the past two decades. Chapter 3 leaves the comparatively safe world of macroeconomic statistics and plunges into the admittedly qualitative issueswhether amenities such as health care, housing, and recreation are improving for Americans, and whether they are likely to improve in the future. In each of the eight amenity categories, the discussion in chapter 3 proposes a way to measure change in the basic amenity (e.g., life expectancy, morbidity rates, and infant mortality in the case of health care), and then describes the way the quality of the amenity has changed for different groups during the past few decades. It pays special attention to groups most disadvantaged by the existing distribution of resources. Finally, building on the base established in chapter 2, chapter 3 develops Alternative scenarios for each amenity. These alternatives are explicitly de-

PAGE 65

5 7 signed to improve the quality of the amenity delivulation. It should be noted that the Alternative ered within fixed financial constraints. Economic facscenarios are speculative, and necessarily reflect tors are combined with factors that could not be some of the values of the people constructing them. predicted by extrapolation alone, such as shifts in Nonetheless, they may not be any less accurate than government policy and radically new technologies. an assumption that trends in consumer behavior can Significant improvements in amenity appear to be be confidently extrapolated to describe the structure possible in virtually every case, given the developof demand over the next two decades. ment of key technologies or changes in public reg

PAGE 66

Chapter 2 Defining the Consumption Recipe

PAGE 67

CONTENTS Defining Categories of Amenity. . . . . . . . . . . . Trends in the Consumption Recipe . . . . . . . . . . Consumer and Government Spending . . . . . . . . . Time . . . . . . . . . . . . . . . . Qualitative Factors . . . . . . . . . . . . . Forces Affecting the Consumption Recipe . . . . . . . . . Households . . . . . . . . . . . . . . Income . . .. .. .. .. .. .. .. .. $ . . . . . . . . Prices . . . . . . . . . . . . . . . . Demographics, Income, and Price: The Combined Effect . . . . . . Alternatives for the Future . . . . . . . . . . . . Constructing Scenarios . . . . . . . . . . . . Summarizing the Results . . . . . . . . . . . . 61 63 63 66 70 71 71 73 80 80 80 80 82

PAGE 68

Chapter 2 Defining the Consumption Recipe Americans achieve amenity through the consumption recipe combining household spending for goods and services, investments of time by house hold members (for which no compensation is received), and government spending.1 These recipes I]n most Caxs, the consumption recipes considered here can be considered equivalent to the utility functions used in most analyses of demand. Moreover, a number of authors have shown how time can are in constant flux, influenced by changes in house hold income, in the price of goods and services, and in demographics, as well as changes in technology that can result in new products, can affect the ways producers communicate with customers, and can even influence tastes and values. Changes in values are reflected directly through new patterns of consumer purchasing, and indirectly through new patterns of public spending and regulations designed be included in the utility equation. See, for example, G.R. Ghez and to guide private spending. This chapter reviews how G.S. Becker, The Allocation of Time and Goods Over the Life Cycle, these forces have shaped and may continue to shape working paper, National Bureau of Economic Research, New York, NY 1975. U.S. consumption recipes. DEFINING CATEGORIES OF AMENITY For the purposes of the analysis that follows, all final household and government purchases are assigned to one of ten amenity categories (see box 2-A). The selection was necessarily somewhat arbitrary. Any classification scheme helps to illustrate some features of economic structure and obscure others. The categories were chosen because they cluster networks of spending where the underlying purpose is comparatively easy to describe. Chapter 3 will undertake the task of describing these purposes in ways that permit a working definition of progress in each of the first eight areas. The category of recreation and leisure proves to be the most difficult to define, since there is plainly an element of recreation involved with spending in all other categories. Is a walk in the park an investment in health or an investment in recreation? What about a car that is fun to drive, a home with a pleasant yard, a meal eaten in a good restaurant? The present study takes the restrictive definition of recreation used in the National Income and Product Accounts.* The spending shown in box 2-A includes only personal and government spending that goes directly for the purchase of amenity. The total does not include investment and savings. In 1985, consumption categories accounted for about 85 percent of the U.S. Department of Commerce, Bureau of Economic Analysis, 4National fncome and Product Accounts, Survey of Current Business, July 1987. table 2.4. U.S. gross national product (GNP). mainder represents investment. The national statistical accounts, Most of the rewhich serve as the base for of most of the data in this study, have several peculiar features that are necessarily reflected in the analysis presented here. 3 One peculiarity is the distinction made between consumption and investment. In general, anything with value after a year is considered to be a form of savings and not consumption. As applied, however, this results in a situation where spending on a gold-tiled bathroom is considered to be an investment, while money spent on education is considered consumption. Similarly, all government spending is considered to be consumption. Government purchases of research and development, roads, dams, and education are not considered a form of national investment. National accounts are also schizophrenic in the way they treat the nonwash economy. Nearly 8 percent of GNP reported by the U.S. Department of Commerce results from the imputed value of services that involve no real market transaction. This value includes the imputed income that homeowners receive by renting houses from themselves, in itself equal to nearly 6 percent of GNP; the imputed 3A formal technique to remedy some of the problems described is proposed by Robert Eisner in The Total Incomes System of Accounts, Survey o/Current Business, U.S. Department of Commerce, Bureau of Economic Analysis, January 1985, pp. 24-48. 61

PAGE 69

62 Box 2-A.The 10 Basic Amenity Groups (and as a percent of all personal and government spending on goods and services in 1985) 1 Percent of personal and Definition government spending 1. The FOOD category includes food and alcohol consumed in restaurants and purchased from stores. Government contributions include agricultural research and a variety of support programs for farmers. 15.1 % 2. The HOUSING category includes everything needed for the operation of a purchased or rented home, including rent and mortgage payments, and purchases of fuel, electricity, furniture, china, draperies, housecleaning, and other goods and services used to maintain a home. Government spending goes for both housing redevelopment programs and spending for infrastructure like water and sewer systems. 19.7% 2 3. Personal spending for TRANSPORTATION includes all spending for mobility, including purchases of automobiles and other personal vehicles, vehicle maintenance, gasoline, and oil; and purchases of public transportation services (air, rail, bus, and taxi). Public spending includes highway construction attributable to personal travel, and maintenance and operation of air and rail facilities. 11 .4% 4. Personal consumption in the HEALTH category includes purchases of drugs, physicians fees, hospital costs (including payments made by Medicare and Medicaid), and spending for health insurance. Public spending includes hospital construction and operation, and community health services. 11 .4% 5. Expenditures on CLOTHING AND PERSONAL CARE go for products and services ranging from apparel and footwear to toiletries and beauty salons. Clothing is by far the largest item in this category, accounting for more than 80 percent of the total. There are no direct government purchases of clothing and personal care as defined here. 6.7% 6. Personal EDUCATION spending includes payments to private schools and colleges. Government spending includes the operation of public school systems and libraries, subsidies for colleges, and worker training programs. 3 7.2% 7. PERSONAL BUSINESS AND COMMUNICATION includes personal communication by tele phone and writing, as well as personal financial, legal, and insurance activities. 6.4% 8, The category of RECREATION AND LEISURE is particularly difficult to define. It is clear that eating out, living in a comfortable home, and transportation are to some extent recreational activities. The more restrictive definition used here includes foreign travel, hotel accommodations, social and religious activities, and purchases of books, magazines, toys, home electronics, movies, and admissions. Government contributions include the operation of parks and recreation areas. 8.9% 9. The category DEFENSE AND SPACE includes only Federal Government purchases of goods and services for military and space. 7.7% 10. GOVERNMENT NOT ELSEWHERE CLASSIFIED (NEC) includes government purchases of goods and services other than defense. Activities include the operating costs of government not directly attributable to a specific amenity group (operation of the Congress and State legislatures); transportation spending not attributable to personal travel; the court system; police, fire, and correctional institutions; and the work of authors of this study. 5.6% Iln standard accounts, the gross national product is conventionally divided into the following components: Personal Consumption Expenditures, Government Purchases of Goods and Services, Net Exports, Gross Private Domestic Investment (consisting mostly of personal purchases of housing and business purchases of products with an expected life of more than 1 year), and an adjustment for changes in inventories. Government purchases of goods and services do rrof include transfer payments such as social security. Spending resulting from transfer payments is counted under Personal Consumption Expenditures. The spending shown in this table includes only personal consumption and government purchases. See the appendix for a detailed table showing how standard categories are mapped into the amenity categories shown here. The spending on housing shown here does not include any payments that result in an increase in the value of the U.S. housing stock from new construction or major renovation. Construction spending of this kind is considered savings and not consumption in conventional accounts. If purchases of new housing are included in the accounts, housing would be responsible for approximately 24 percent of the total. 3 The accounts treat both private and public spending for education as consumption and not as a form of savings. Corporate training costs are not included. SOURCES: Consumer purchases in these amenity categories are derived from the Personal Consumption Expenditure (PCE) categories used in the National Income and Product Accounts (NIPA), U.S. Department of Commerce, Bureau of Economic Analysis, Survey of Curren( Business, July 1987, table 2.4, A precise map connecting these PCE categories to amenities is shown in the appendix. Categories of government purchases of goods and services are derived from tables 3.15 and 3.16 of the National Income and Product Accounts. The map connecting spending categories in these tables with the amenity groups is also shown in the appendix.

PAGE 70

63 rent paid by non-profit organizations that own their national accounts do not, however, impute the value own buildings; the imputed value of liquidity in bank of housework, education provided at home, or time accounts and insurance funds; and the imputed invested by standing in line at fast food restaurants. value of food produced and eaten on farms. 4 The 4 See U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Survey o/Current Business, July 1987, table 8,9. TRENDS IN THE CONSUMPTION RECIPE Before proceeding to examine the forces driving the change, it is worth taking a brief look at recent changes in the way Americans construct recipes to achieve amenity. These trends are described in three ways: by examining changes in overall patterns of spending for the amenities, by exploring trends in the way time was spent, and by reviewing qualitative features of spending patterns that are not adequately reflected in statistics on how Americans spend their time and money. Consumer and Government Spending Share of Total Expenditures The history of household and government spending during the past three decades is traced in figures 2-la, 2-lb, and 2-lc. Looking at trends in expenditures on each of the different amenity groups, perhaps the most striking feature of the statistics is the Figure 2-la. -Spending by Amenity Type (percent of all consumer and government spending) Percent of current dollars 25 I 1 20 --------------------------------151 lo 5 Foo d Housin g Transpor t Healt h SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, tables 2.4,3.15, 3.16. Figure 2-lb.-Spending by Amenity Type (percent of all consumer and government spending) Percent of current dollars ~121 I lo 8 6---/------, 1 I 2 I 1 I I 1 1 I ( ( I r I ( 1955 1960 1965 1970 1975 1980 1985 Clothlng 8 Pers Care --Educatio n ~ Pers Busi /3 Comm. Recreation 8 Leisure SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, tables 2.4,3.15, 3.16. Figure 2-le.-Spending by Amenity Type (percent of all consumer and government spending) Percent of current dollars 14 I I 8 ,----.. --.6 4, 1 1 2-I I 1 1 1 I I 1 1 I I 1 I 1 1 I I I 1 1 1 I 1 1 1955 1960 1965 1970 1975 1980 1985 Government n e c. ---Defense SOURCE: U.S. Depafiment of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, tables 2.4,3.15, 3.16.

PAGE 71

64 continued high level of personal consumption expenditure (PCE) on the most basic of amenities. Housing, Transportation, Food, Health, and Clothing and Personal Care accounted for roughly twothirds of U.S. consumer spending in 1985, changing l l l l l little since 1950. The figures show: a rapid and consistent decline in the percent of income spent for Food, which has been almost precisely offset by a rapid and continuous growth of Health spending (Food lost and Health care gained about 8 percent of total spending between 1955 and 1985); a sharp rise of spending on education followed by declines after 1975 as the baby boom generation passed through the system; continuous growth of spending in Personal Business (primarily financial services) and Communication, which appears to have accelerated since 1981; a steady decrease in the share of spending going for Clothing and Personal Care; and a decline in Defense spending as a proportion of all spending, which reversed in 1979. Spending for Housing and Transportation, two of the largest categories, has remained surprisingly constant since World War IIdespite large changes in the prices of energy, automobiles, and housing; in the size and characteristics of households; and in real per capita spending, which has doubled in the last 40 years. Whenever Housing costs go above 21 percent of spending, or Transportation costs go above 12 percent, a subtle alarm seems to sound. This results from a combination of factors, including historically fixed formulas for the percentage of gross income that borrowers can expect for mortgage or car payments. People have apparently used discretionary income to increase the quality and variety of purchases in amenity areas like Food (through more varied eating), Transportation (through higher quality cars and more cars per family), and Clothing and Personal Care, rather than spending new income on Recreation. As a fraction of all spending, Recreation and Leisure seems to have gone through a one-time jump between 1961 and 1972, with slow subsequent increases. Dollars and Value.-The previous discussion showed how Americans spent their income in any given year in current dollars values measured in the currency of the year in which the data was collected. The statistics say little about either the quality or the quantity of the products and services purchased, since many relative prices changed rapidly during the period shown. Legal services that cost $1,000 in 1955 might still be considered worth the price in 1987, but a computer equivalent to one that cost $1,000 in 1955 would be worth very little in 1987. The effects of inflation can, in principle, be removed from product prices by converting spending to constant dollars. Converting current into constant dollars involves the vexing problem of developing a consistent set of prices for goods and services of constant quality. This process becomes more difficult during periods of rapid technological change. It can also become more difficult as a greater fraction of economic activity involves activities like legal services, where quality is inherently more difficult to measure (see discussion in ch. 5). Most of the qualitative features of figures 2-la to 2-lc remain when spending is converted into constant dollars (see figure 2-2), with one important exception: in constant dollars, spending on Clothing and Personal Care has risen sharply since the mid 1970s. Constant dollar spending by government for amenities cannot easily be estimated, since data is not collected on constant quality government services in much details Another way to view the changes in average prices that have occurred during the past three decades, compared to changes in average incomes, is to see how long it has taken an average full-time worker in America to earn enough to buy items of comparable quality during the past few decades. In this respect, most manufactured goods have become a bargain while the real price of most services has changed little (see table 2-l). A television set that could be bought for 1 days work in 1972 needed over 4 days work in 1950 and only 4 hours work in 1986. Clothing, new cars, and communication services showed strong declines. On the other hand, most forms of medical services now require more work to purchase than was the case in 1950. 5Fo~ years in which data is avai]able, constant dollar estimates of gov ernment spending can be computed by converting government spending in each category to spending in commodity categories for which output deflators are available. This could not be done for most of the period shown in figure 2-2.

PAGE 72

65 Figure 2-2a. -Constant Dollar Shares of Consumer Spending on Health, Housing, Clothing & Personal Care, and Education Percent of all consumer expenditures 30 25 20 j 15 ----lo ---------------------------------5 1950 1955 1960 1965 1970 1975 1980 1985 --Healt h Housin g Clothing & Pers Care EducatIon SOURCE: U.S. Department of Commerce, Bureau of Economic Anatysis, National Income and Product Accounts, historical diskettes, table 2.5. Figure 2-2 b.-Constant Dollar Shares of Consumer Spending for Food, Personal & Business Communication, Transportation, and Recreation & Leisure SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 2.5. Patterns of price reduction have also changed over time. Between 1950 and 1970, there was a significant decline in the amount of work needed by the average worker to make all purchases. Of the items listed in table 2-1, only mass transit required more labor to buy in 1970 than in 1950. Since 1970, however, many items require more work to purchase. Spending on Services. -Attempts to separate services from goods have never been completely satisfactory. The category services combines housecleaning, brain surgery, banking, and car repair. Most amenities are satisfied by a complex combination of purchased goods and services. Later chapters will show the extent to which even manufactured products embody non-manufacturing service activities. However defined, services represent an increasing fraction of consumer expenditures. The price of manufactured goods has declined much faster than that of services (table 2-l), while demand for services appears to have increased rapidly with rising incomes. The curious pattern of decline and growth in demand for services between 1930 and 1986 is charted in figure 2-3. Rising from a low of less than one-third of all constant dollar spending at the end of World War II, services now command more than 40 percent-even given a comparatively narrow definition. Most recent growth in demand has not resulted from purchases of personal services like housekeeping, but rather from purchases of medical care, education, and professional services like banking, law, and insurance. Savings and Investment.-In recent years, personal and government consumption has grown as net savings have declined. Figure 2-4 illustrates the decline in national savings rates during the past decade or so, measured as a percentage of GNP. A slight increase in rates of business savings, taken as retained earnings and in other ways, has been more than offset by a sharp decline in net government savingsthe budget deficitand by the drop in personal savings. Gross savings fell from over 18 percent of GNP in 1979 to about 13.5 percent in 1986. As will be outlined in chapter 8, this savings shortfall has been offset by large foreign investments in the United States. The Private/Public Mix The mix of public and private spending used to purchase amenity differs greatly from one part of the economy to another (see table 2-2). Taken together, public spending categories together have held a surprisingly constant share of total spending for three decades. The combination of spending directly related to the first eight amenity groups, Defense, and Government spending not elsewhere classified has remained near 23 percent of all spending since 1955, with a brief excursion to 26 percent during the VietNam War. Defense purchases fell steadily as a fraction of all current dollar spending toward the end

PAGE 73

66 Table 2-1 .Change in Time Needed by the Average a American To Purchase Goods and Services (time worked to buy item in 1972=1.0) Item purchased 1950 1960 1970 1980 1986 Radio and television receivers . . . Womens and childrens clothing . . Shoes. . . . . . . . . Telephone and telegraph . . . . Durable goods (average) . . . . Recreation (BEA definition) . . . . New autos . . . . . . . Admissions to spectator events . . . Nondurable goods (average) . . . Food (grocery) b . . . . . . Average personal consumption. . . Transportation c . . . . . . Homeowning d . . . . . . . Transit systems . . . . . . Home renting.. . . . . . . Drug preparations and sundries . . . Food (restaurants) e . . . . . Gasoline and oil. . . . . . . Health insurance . . . . . . Airline . . . . . . . . Personal business . . . . . Medical care (all private) . . . . Privately controlled hospitals . . . Household electricity . . . . . Physicians . . . . . . . Household gas. . . . . . . 4.09 2.02 1.47 2.21 2.42 1.37 2.18 1.19 1.80 1.80 1.75 1.39 1.72 0.85 1.72 2.40 1.25 1.98 1.28 1.82 1.28 1.15 1.11 2.22 1.13 1.73 2.20 1.36 1,24 1.63 1.66 1.21 1.72 0.99 1.34 1.35 1.35 1.20 1.40 0.94 1.40 1.80 1.02 1.54 0.91 1.34 1.09 1.05 1.06 1.53 1.05 1.46 1.13 1.07 1.07 1.03 1.09 1.06 1.10 1.05 1.06 1.06 1.05 0.99 1.05 1.02 1.05 1.11 1.01 1.11 0.90 1.03 1.02 1.02 1.04 1.02 1.03 1.01 0.63 0.71 0.85 0.66 0.88 0.86 0.90 0.87 1.04 1,06 1.00 1.02 0.94 0.81 0.94 0.89 1.01 1.89 0.72 1.23 1.09 1.08 1.13 1.19 1.12 1.66 0.39 0.55 0.68 0.72 0.75 0.80 0.81 0.87 0.90 0.94 0.96 0.99 1.01 1.02 1.03 1.04 1.07 1.08 1.08 1.17 1.20 1.21 1.23 1.25 1.28 1.82 How To Read This Table: The fraction of television and radio receivers that could have been purchased for 1 day of workin 1972 would have required 4.09 days of workto purchase in1950and 0.39 days of work to purchase inl986.NOTE: Thecategorles are arepresentatlve sample andnota complete set. They are ranked by growth In the ratio between 1972and 1988. For details, see table 2.4 of the National Income and Product Accounts. ao4Average defined by average wage and salary earnings per fulhtimeequivdent emPloYee. bF~purchmed for off.premise consumption. cNot including vehicle purchases. dMofigagepayment$ Iesspayment for equity. eFood purchse$d for on-premise consumption. fB~kse~~ce charges, Insurance (non-health), brokers. SOURCE: Based on U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Survey of Currerrt Business, historical diskettes, tablea 6.8b and 7.10, July 1987. of the Viet Nam War (figure 2-lc) while civilian spending increased. The trend was reversed in 1979 but Defense spending remains below the share it held during the 1955-68 period. In most cases, patterns of purchasing have been comparatively stable. By long tradition, for example, government pays for roughly 85 percent of all Education; privately supported education has not made serious inroads. There have, however, been some changes in the mix of public and private spending. In Transportation, though most highways and some other transportation infrastructure are purchased publicly, many private developers have recently been required to build roads, sewers, parks, water supplies, sidewalks, and other infrastructure formerly provided at public expense. Table 2-2 indicates a sharp decline in government spending on highway construction and other infrastructure. It also shows the effect of declines in support for Housing and parks (in the Recreation and Leisure category) that occurred in 1980. Time Incomes can increase due to productivity, but time available to spend that income remains fixed unless the economy moves to shorter work weeks and longer vacations. If anything, however, Americans are working harder and longer than they have in the recent past. Increased female participation in the work force has meant that many chores once purchased through unpaid housewife time (child care,

PAGE 74

6 7 Figure 2-3.-Share of Consumer Spending on Services Percent of all consumer spending (constant 1982$) 1 50 45 40 35 30 25 7r? i930 1940 1950 1960 1970 1980 SOURCE: US. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 2.5, Figure 2-4.-Savings and Investment by Type Percent of GNP 20~ 5 ~ 1, .-. 1 1 1 1 I 1 I -5 1 7 1 r 1 1 r 1 I 1 1 1 1 1 1 r 1955 1960 1965 1970 1975 1980 1985 Personal Saving s Business Saving s .Government Savings = Foreign Investment How To Read This Table: In 1986, personal savings (investments in savings accounts, stocks, etc.) fell to 3V0 of GNP. The U.S. economy received more investment money from foreigners (nearfy 4/0 of GNP in 1966). Because of the government deficit, government savings were a negative 3/0 of GNP. Net business savings (retained earnings, depreciation, etc.) rose to 16/0 of GNP. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts) historical diskettes, table 5,1, care for the elderly, and even cooking) are now bought in the market. Technology has made it possible to move many activities formerly available only through the marketplace into the home. The private automobile continues to replace purchased transit. The video cassette recorder and other home electronic equipment are substituting for entertainment purchased away from home and seem to have had an impact on many of the social clubs and organizations that once occupied a significant amount of American time. Cost containment in health care, combined with the emergence of elaborate home care equipment, has forced many kinds of health delivery out of the hospital and into the home. Interest in self-administered health promotion, such as diet and exercise programs, has grown. Direct marketingusing 800numbers, credit cards, home shopping networks, and the United Parcel Service-has risen dramatically. In effect, technology has capitalized household time to make it more productive. But while this has increased the productivity and decreased the burden of housework in principle, real savings are difficult to finds Personal investment in self-training appears to have increased as well, in areas related to both employment and recreation. Home information is one of the most curious examples. In the past, information acquired from radio and television broadcasting was purchased primarily through unpaid investment of a consumers timewaiting through commercials. The consumer bought only the receiver. Purchases of information in the form of cable television and video cassettes (including rentals), however, are beginning to approach the total value of equipment sales. Trading Time and Money A number of attempts have been made to understand how individuals trade personal time against time spent in the marketplace. 7 Will an individual choose to work fewer hours and enjoy more leisure as a result of higher wages, and is it institutionally possible to do so? Alternatively, will a person with a higher income feel that the effective cost of leisure time is also high and prefer to work longer? The choice requires an understanding of the way individuals substitute leisure time for goods, but theory does not pretend to predict an answer and available 6J. Vanek, Time Spent in Housework, Scientific American, vol. 231, 1974, pp. 116-120. TSee J.D. Owen, Working Lives (Lexington, MA: Lexington Books, 1979); G.R. Ghez and G.S. Becker, op. cit., footnote 1. Early work on trade-offs between leisure time and work time can be found in Lionel Robbins, On the Elasticity of Income in Terms of Effort, Econornica, No. 10, pp. 123-29, 1930. Also see discussion in ch. 11 tying income to the desire for more or less work.

PAGE 75

68 Table 2-2.Government Spending as a Percent of Final Consumption in Each Amenity Group (current dollars, in percent) Amenity 1955 1965 1975 1985 Food . . . . . . . . . 2.6 % 0.7% 1.4% 3.8% Housing . . . . . . . . 1.5 2.5 2.9 2.0 Transportation. . . . . . . . 11.3 13.1 11.8 9.1 Health . . . . . . . . . 6.0 19.7 22.1 17.5 Clothing and Personal Care . . . . 0.0 0.0 0.0 Education . . . . . . . . 82.5 84.0 85.2 82.7 Personal Business and Communication . . 0.0 0.0 0.0 0.0 Recreation and Leisure . . . . . 2.2 2.9 4.2 3.3 Defense . . . . . . . . 100.0 100.0 100.0 100.0 Government n.e.c. . . . . . . 100.0 100.0 100.0 100.0 Total (GNP). . . . . . . . 21.4 23.7 24.6 23.5 How To Reed ThlsTable: !rt1955,2.6 percent of aff final consumption of Food resulted from Federal, State, or local government purchases of goods and services related to this amenity. NOTE: The large lncreaaa in government spendingforthe Health amanity batwean 1955andl~resultad from aradefinition ofthls category betwean 1956and 1959. SOURCE: Based on U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Survey of Current Bush?ess, tables 2.4, 3.15, and 3.16, July 1987. behavioral evidence is ambiguous. An obvious shortcoming of conventional approaches to the issue is that they assume workers do not derive any satisfaction from their jobs 8 (see ch. 11 for a discussion of these issues). Trenda in Time Use There have been many changes in how time is divided between work, including housework, and leisure. Comparisons between time use surveys conducted in the 1930s and the 1960s, for example, show considerable increases in work time and decreases in free time. One of the major areas of growth in work time was housework, in particular the time spent shopping and traveling on household errands. It appears that the impact of the deluge of timeand labor-saving home appliances during this period was to lengthen rather than shorten the amount of time spent on housework.9 Between 1965 and 1975, however, Americans experienced considerable gains in free time and declines in the amount of time spent both working for pay and working in the home. More women were Tibor Scitovsky, The Joyless Economy (Oxford: Oxford University Press, 1976). The impact of new home technologies on time use is complicated. C.L. Long, Labor Force Under Changing Income and Employment (Princeton, NJ: Princeton University Press, 1958), argued that improved household technologies made women more productive in the home and increased female leisure, thereby freeing them for more market work. However J. Vanek, in Time Spent in Housework, op. cit., footnote 6, refuted the notion of increased household productivity. working for pay, but on average they were working considerably fewer hours than before. 10 Time spent on housework also declined considerably, by 20 percent overall. The combined decline in labor market time of about 40 minutes per day and in housework of about 14 minutes per day resulted in a 10percent increase in leisure time. What were Americans doing with their increased free time during this period? Primarily, they were watching television. On average, TV viewing time in 1975 totaled about 14.7 hours per week, the equivalent of all time spent on housework and 1.5 times as much as time spent eating. TV viewing constituted 50 percent of all leisure, and more than 60 percent of all time spent on passive leisure. More recent time use data, collected in 1985, suggest that over the past decade the use of time has changed once again (see table 2-3). 1 1 A sharp increase in work time, largely the result of women working longer hours, is mirrored by a decline in free time.1 2 IOJOhn P RObin~n, changes in Americans Use Of Time: 1!%5-1975: A Progress Report, Communications Research Center, Cleveland State University, August 1977. I IJOhn p. Robinson, Trends in Americans Use of Time: some PreIiminafy 1975-85 Comparisons, Survey Research Center, University of Maryland, December 1986. 12Data provided by the U.S. Bureau of Labor Statistics indicate that the number of hours worked per adult grew at a slower rate than that captured in the survey discussed here, due largely to earlier retirement among older men, which has tended to offset the entry of women into the U.S. work force (see discussion in ch. 11). This time use survey however, does not capture the offsetting effects of earlier retirement, primarily because the number of men surveyed who were over age 55 did not constitute a statistically significant sample.

PAGE 76

69 Table 2-3.WeekIy Time Budgets: Men and Women, 1975 and 1985 (in hours) 1975 1985 Change in hours 1975 v. 1985 Men Women Average Men Women Average Men Women Average Contracted time . . 32.9 14.5 23.7 33.2 21.3 27.3 0.3 6.8 3.5 work . . . . 30.0 13.4 21.7 29.6 19.3 24.5 .4 5.9 2.7 travel to work . . 2.9 1.1 2.0 3.6 2.0 2.8 0.7 0.9 0.8 Committed time . . 15.6 33.1 24.3 16.6 28.6 22.6 1.0 .5 1.7 housework. . . . 8.1 21.6 14.8 9.4 17.7 13.5 1.3 .9 1.3 child care. . . . 1.0 3.6 2.3 0.8 3.0 1.9 .2 .6 .4 shopping . . . 2.7 4.1 3.4 2.8 4.1 3.5 0.1 0.0 0.1 family travel . . 3.8 3.8 3.8 3.6 3.8 3.7 .2 0.0 .1 Personal time . . . 75.3 78.6 76.9 75.9 79.6 77.7 1.0 0.8 eat at home. . . 6.6 6.6 6.6 6.3 5.8 6.1 .3 .8 .5 eat out . . . . 3.0 1.9 2.4 2.0 .7 0.1 .3 personal care . . 65.7 70.1 67.9 67.3 71.8 69.5 1.6 1.7 1.6 Free time. . . . 44.2 41.6 42.9 42.1 38.3 40.2 .1 .3 .7 education . . . 3.0 1.8 2.4 2.6 1.7 organization . . 1.4 2.1 .4 .1 .3 2.4 1.9 1.4 1.7 1.5 0.0 .7 .4 social . . . . 6.9 7.0 6.9 5.5 4.7 5.1 1.4 2.3 1.8 recreation . . . 4.7 4.5 4.6 5.0 4.1 4.5 0.3 .4 .1 electronic med . . 18.4 15.9 17.1 17.2 15.6 16.4 .2 .3 .7 other media . . . 6.6 7.5 7.1 6.6 7.8 7.2 0.0 0.3 0.1 Ieisure travel . . 3.2 2.5 2.9 3.8 2.7 3.3 0.6 0.2 0.4 SOURCE: John P. Robinson, Trends in Americans Use of Time: Some Preliminary 1975-85 Comparisons, Survey Research Canter, University of Maryland, December 19s6, p. 34. All age groups of women have increased their hours of paid work by about 6 hours a week, while there has been relatively little change in the numbers of hours of paid work for men. The increase among women was offset to some extent by a decline in hours of housework, but there was also an average loss of more than 3 hours per week in womens leisure time. In fact, women in most age groups have 5 to 6 hours less leisure time per week than men. The exception is the 35 to 54 age group, where men and women appear to have the same amount of leisure time. Ten years ago, women had more leisure time than men in this age group. There have also been changes in how leisure time is used. The two largest uses of leisure time are social activities and electronic media. Time spent in sociaI activities declined by 26 percent between 1975 and 1985. Television viewing, expressed in the table as part of electronic media, also declined by varying amounts among the different age groups. At the same time, a sharp increase in time spent in telephone conversations was recorded. For example, time spent on the telephone doubled for men aged 25 to 44, and increased by 10 percent for men over 65. Although there has traditionally been a gender gap in the use of the telephone, by 1981 this difference disappeared in the youngest age cohort. Allowing for statistical vagaries, Americans seem to spend more days and more hours on the job per day than is the norm for industrialized nations. Between 1953 and 1983, average weekly hours worked by U.S. production workers remained virtually constant, moving from 40.7 to 40.1 with only minor fluctuations. During the same period, however, hours in other industrial countries decreased, often markedly. Canadian workers averaged 40.1 hours in the mid 1950s against 37.1 in 1982, during which time the Italians declined from 44.7 to 37.7 hours and Belgian averages fell from 41.6 to 31.7. Between 1965 and 1983, the French average declined from 45.8 to 39.3. The West Germans, Dutch, and British workers now work more hours than their U.S. counterparts (40.3 to 40.6 hours in 1983). The Japanese worked 41.1 hours.1 3 Long commutes also constrict American leisure. The United States shares with Australia the second lowest percentage of people commuting to work 15 minutes or less (36 percent), exceeding only the Netherlands (21 percent) .14 Vacation time and paid leave offer other indications of the character of Americans as workers. The United States has, with the enactment of Martin Luther King Day, nine national holidays, with addi-

PAGE 77

. 70 tional holidays particular to some States such as Patriots Day in New England. This is lower than all but 2 of 14 OECD (Organization for Economic Cooperation and Development) nations surveyed. ]5 Unlike most of OECD, the United States does not impose a legal minimum of compensated leave. Time and Consumer Decisions The amount of free time available also affects the quality of purchasing decisions made by Americans. lG Conflicting forces are again at work. On the one hand, future consumers are likely to be better educated than those of earlier generations, and will have a wide range of new technologies available for gaining access to information. On the other hand, consumers are likely to have less time to analyze information and make major decisions. Increased incomes, dual earner households, and the growing complexity of choice resulting from diversified product offerings means that decisions are becoming more frequent and more difficult while less time is available to make them. 17 In fact, many retailers find that major purchasing decisions are being made by teenagers whose parents may have little time to shop. s While imperfect consumer decisions that result from minimizing the time spent collecting and analyzing information may be a rational response by individuals, society as a whole may pay a much greater price.19 This raises a number of real challenges for those wishing to forecast the future of consumer spending, Qualitative Factors The data just cited provide a crude guide to changes in the ways Americans spend their income .-. sIbid., p. 89. IWne of the Centra] axumptions in much economic analysis is that consumers are perfectly informed and rational in the way they make decisions, that this information is free, and that any defects in decisionmaking will continue into the future. There is now a considerable literature discussing defects in this assumption. The defects are not particularly important if the degree or the significance of misinformation do not change significantly over time. The discussion in ch. 3, however, shows many places where significant changes may occur in the way consumers obtain and use information. 17S. B. Linder, The Harried Leisure C/ass (New York, NY: Columbia University Press, 1970). IEC. RIJS@], The New Homemakers, American Demo~raphks, OCtober 1985, p. 23. l~ibor Scitovsky, op. cit., footnote 8. and timechanges whose influence on the structure of the economy as a whole will be illustrated in later chapters. As spending has moved further from needs and necessities to decisions reflecting a range of tastes and choice, it has generally been distributed in familiar ways. Americans spent about one dollar in five on Housing in 1950 and the same fraction in 1986, in spite of the fact that real per capita income more than doubled during the period. Many factors in the structure of household and government demand that affect the quality of amenity achieved, as well as the structure of producing enterprises, are not well reflected in these statistics. They include such things as changes in the quality of the houses purchased, the type of food Americans eat, and the kinds of cars they drive. These issues are discussed at much greater length in chapter 3, One general observation, however, deserves attention. Many markets formerly dominated by a comparatively small number of relatively homogeneous products are becoming boutique markets, combining a wide range of specialties. There are, of course, markets for low-priced commoditiesbut the low cost, vanilla-flavored product is now itself a kind of niche. Chapter 3 will demonstrate that specialization is replacing commodity-like products in a remarkable variety of marketsand will suggest a number of areas where specialization is likely to grow rapidly. Financial packages in areas such as insurance and investment can be designed and analyzed rapidly by trained people working with a computer terminal. The mass media are becoming differentiated, as specialized magazines replace such publications as the Saturday Evening Post. An industry directory indicates that there were 11,000 periodical titles in 1986, an increase of 60 percent from 1985. 20 Broadcast television has had its market fragmented by cable TV and video cassette recorders. Eating habits have diversified, as both restaurants and larger groceries cater to a wider variety of tastes. This product differentiation is driven by a variety of factors, including the growing diversity of Amerz~he directory inc]udes titles like: Walking, Running, wildfowl @Ving, Gambling Times, Vegetarian Times, North Texas Golfer, Croquet Today, and The Quarterly Magazine of The Rocky Mountain Elk Foundation. See S. Fatsis, A New Leaf: Magazine Industry Flourishing, Asse ciated Press, Oct. 11, 1987.

PAGE 78

71 ican households, the movement of comparatively well-educated baby boomers into their peak earning years, and technology that allows both production of specialized products without an enormous price premium and highly targeted advertising. 21 However, it proves difficult to disentangle cause and effect. It is likely, for example, that the demand for tailored products was always latent but difficult to express because of the large cost premiums involved. Changes in taste also play a role, but are far harder to document. It does little good to produce for a specialized market if it is impossible to market the product within that niche. New technology, however, has also changed the way that producers can reach specialized markets. Printing and mailing catalogs to spe Zlsee a]~ Aimee L, Stern, The Baby Boomers Are Richer and Older, Business Month, October 1987, pp. 24-28. cialty lists has become an enormously sophisticated enterprise. The Montgomery Ward catalog has given way to a plethora of specialized pamphlets. Active consumer participation in the design of products, including information products, is another likely development. Homes can be designed, with the assistance of a skilled salesman, on a computer in showrooms. Even clothing may soon be tailored at an affordable price using robotic sewing equipment. As later chapters will show, the shift to specialized goods and services is having a profound impact on the structure of the business networks that produce amenities. In particular, technologies that appear to have economies of scale in a situation where demand is predictable may perform poorly in a rapidly changing market. A flexible production system may use equipment with comparatively inexpensive set-up times and small truck deliveries instead of bulk freight, even though these systems appear to be less productive than those they replace. FORCES AFFECTING THE CONSUMPTION RECIPE The changes in spending patterns just described have been driven by a variety of forces that will continue to play a powerful role in shaping future recipes of consumption. Demographic factors, such as changes in the number and type of American households, the aging of the baby boom generation, and the growing number of elderly people, will affect spending patterns, as will changes in household income, income distribution, and prices. The following discussion examines these forces in some detail, in order to develop a portrait of how Americans may spend their money in the future assuming that such an estimate can be derived from data on demographics, income, and prices. This allows for the creation of Trend scenarios, using standard extrapolative techniques to project spending patterns. These Trend cases will also serve as the basis for the Alternative scenarios, which will be described at some length in chapter 3. The Alternatives make assumptions about the role of new products, new regulations, new values, and other factors that may not be captured in standard statistical serieshowever clever the extrapolation technique. Households National averages can provide a grossly misleading view of national living standards. Every American household has its own needs and resources, and its own recipe for happiness. Depending on its nature, economic growth can benefit some groups but not others. One way of looking below the veil imposed by averages is to examine spending patterns by household type and household income. 22 Household analysis is also needed to understand how well the economy is meeting the needs of a diverse population, and how it is serving those most likely to be disadvantaged. Trends in the Formation of Households Characteristics of households have changed rapidly over the past two decades, following changes Zzsee, for example, Roberta Barnes and Robert Gillingham, Demo graphic Effects in Demand Analysis: Estimation of the Quadratic Expenditure System Using Microdata, The Review o/ Econornks and Stafjsfks, No. 591, 1984; and Robert Gillingham, Measuring the Cost of Shelter for Homeowners: Theoretical and Empirical Considerations, Revjew of Economjcs and Statjstjcs, vol. LXV, No. 2, 1983.

PAGE 79

72 in both the age structure and lifestyle of the population. The average American household shrank from 3.33 people in 1960 to 2.65 in 1983 and has continued to decline, largely as a result of the striking increase in the number of people living alone. The number of all single households has nearly doubled since 1960 and amounted to nearly one-quarter of all such American households in 1983; the share of young singles grew over 60 percent between 1972 and 1983 (see table 2-4). One in every eleven Americans now lives alone. Since the average size of a household is shrinking, the total number of households has grown more rapidly than the U.S. population in recent decades. At the same time, significant changes in lifestyle have taken place. The post-war era has seen more women join the work force, incomes rise, a growth in divorce rates, and an increase in the number of young Americans who have decided to live away from home. These factors are giving a new look to the American household. Only about one-fifth of American households now include a working father, Table 2-4.The Composition of Consumer Households: 1972, 1983, and Estimates for 2005 (In percent) Percent of total Household (HH) type 1972 a 1983 2005 Singles . . . . . 11.5% Age 15-34 . . . . 4.2 Age 35-64 . . . . 7.4 Elderly (65+) . . . 19.2 Single . . . . 8.3 Couples . . . . 10.9 Couples . . . . 58.0 No children . . . 16.9 Child <6 years old . . 9.2 Child 6-17 years old . 20.8 Child >17 years old . 11.1 Single parent . . . 5.1 Other. . . . . . 6.2 Unrelated adults . . 1.5 Other. . . . . 4.7 Total (percent) . . 100.0 Total (millions of HH) 69.3 15.2% 6.9 8.3 20.3 8.9 11.4 47.4 14.0 7.4 16.0 10.0 6.6 10.5 4.0 6.5 100.0 88.8 14.7% 4.9 9.8 20.7 9.1 11.6 48.4 14.5 5.8 16.3 11.8 5.9 10.3 3.3 7.0 100.0 118.6 alg72 shu~ co~spond to families, as defined by the U.S. BUreSU of the census Consumer Population Survey. For the 11 household types, the differences In 1983 share between families and consumer households were ail less than 1 percent. NOTE: Totals may not equal 100 due to rounding. SOURCES: 1972 figures taken from U.S. Bureau of the Census, Consumer Population Survey; 1983 and 2005 figures taken from Household Formation Program, working paper prepared for the Office of Technology Assessment, Washington, OC, May 1986. a housewife, and children, 23 down from nearly onethird in 1972. As table 2-4 indicates, all categories of married couples under age 65 have declined in share since 1972when the oldest members of the baby boom were 26, an age after which the traditional expectation would bean increase in the share of married couples. Indeed, the baby boom generation, born just after World War 11, has played a major role in reshaping American demographics, and its weight will continue to be felt as its ranks pass through different age cohorts. In 1960, 41 percent of the population was under the age of 21. By 1982, this figure had fallen to 34 percent. Other age groups have become more populous as the median age of the population has risen. The demographic model used for this analysis 24 indicates that during the next 20 years the number of U.S. consumer households 25 will increase by w. Rusge]], op. cit., footnote 18. 24 Household Formation Program, working paper prepared for the Office of Technology Assessment, Washington, DC, May 1986. See box 2-B and the appendix for more detail. For other projections of household composition, see U.S. Bureau of the Census, Current Population Reports, Series P-25, No. 986, Projections in the Number of Households and Families: 1986-2000 (Washington, DC: U.S. Government Printing Office, 1986); John R. Pitkin and George S. Masnik, Households and Housing Composition in the United States, 1985 to 2000: Pro jections by a Cohort Method, Joint Center for Housing Studies of MIT and Harvard University, Research Report Rf8&l, Cambridge, MA, 1986; and Patricia H. Hendershott, Household Formation and Home Ownership: The Impact of Demographics and Taxes, National Bureau of Economic Research, Working Paper No. 2375, Cambridge, MA, September 1987. zSThe term consumer household in this analysis is used in phiCe of the term consumer unit used by the U.S. Bureau of Labor Statistics. A consumer household is a single person or group of persons in a sample household related by blood, marriage, or adoption or other legal arrangement, or who share responsibility for at least two out of three major types of expenses-food, housing, and other expenses. Consumer households are divided into 11 household types, for which trends and scenarios on composition and spending are developed. See Household Formation Program, working paper prepared for the Office of Technology Assessment, Washington, DC, May 1986. The term consumer household is used in this discussion to differentiate it from the generic term household, used in other parts of ch, 2. Consumer households (or consuming units) should not be confused with other precise definitions of households or families. As defined by the U.S. Bureau of the Census, Current Population Survey (CPS), Technical Documentation, March 1984, a household is as follows: A household consists of all the persons who occupy a house, an apartment, or other group of rooms, or a room, which constitutes a housing unit. A group of rooms or a single room is regarded as a housing unit when it is occupied as separate living quarters; that is, when the occupants do not live and eat with any other person in the structure, and when there is direct access from the outside or through a common hall (this does not include group quarters like rooming houses, military barracks, or institutions).

PAGE 80

73 about one-thirdan average annual increase of 1.3 percent (again see table 2-4). This is a higher rate of increase than that of total population because of an expected continuing decline in household size. The distribution of major consumer household categories should remain fairly stable, with the largest change on the order of 1 percent (married couples under age 65). The distribution of types within categories, however, will shift somewhat, due to the aging of the baby boomers. Single consumer households age 35 to 64 and married couples with children over 17the two groups in which baby boomers will be counted most frequently in 2005will increase more than any other type; at the same time, young singles and couples with children under 6 will lose share. 2G Spending by Type of Consumer Household Data describing expenditure by type of consumer household are sparse. The primary source is the U.S. Bureau of Labor Statistics Consumer Expenditure Survey (CES), which uses only a small sample and suffers from a number of other drawbacksthere was no survey taken between 1972/73 and 1982/83. If the CES data are used in an aggregated form, however, they give reasonable results, and the 1982/83 survey is used as the basic household database in this study .27 Table 2-5 shows how different consumer households spent their money in 1972/73 and 1982/83. Housing costs represent a significant fraction of the spending of couples with young children, singles, and single parents; housing captured the highest fraction of spending among single elderly households. Not surprisingly, smaller households spend more per capita on housing than their larger counterparts. The share of spending devoted to housing has increased The consumer household is also distinct from the CPS family, which is defined as a group of two persons or more (one of whom is the house holder) residing together and related by birth, marriage, or adoption. There may be several families in a single household. ZG]t should be noted that the number 6f elderly households will begin to increase rapidly after 2011, 6 years beyond the scope of the scenarios used in this analysis. ZTExpenditure categori~ for consumer households are taken from the Consumer Expenditure Survey, and are not an exact match with the amenity categories outlined earlier in this chapter. See Consumer Expenditure Demand Projection Program, working paper prepared for the Office of Technology Assessment, April 1986. For more detail, see the appendix. since the 1972/73 CES survey for all but two consumer household types. Per-person expenditures on food are high in single consumer households, which do not benefit from economies of scale in food preparation. Within the food category, young singles and childless couples both spend far more on food away from home than their larger and/or older counterparts, although all households spend a smaller fraction of their incomes eating out than cooking at home. Still, in households headed by someone 34 years of age or less, almost half of all expenditures on food are away from home, compared with 27 percent in the over 65 group. Similarly, per capita spending for away-from-home entertainment is higher for younger singles and childless couples than for the elderly. On the other hand, spending on air fares, hotels and motels, home electronic and other devices, and alcohol increased for every household type. Out-of-pocket medical expenses increased sharply for the elderly during the 1970s and early 1980s, and remain much higher than for other cohorts; health care also became a greater burden for single parents. Out-of-pocket spending for education increased for most consumer household types; the increases were particularly dramatic for young singles and single parents, while parents with children of college age continued to spend more on education than any other group. All household types increased the percentage of spending on personal business and communication, while the converse was true for spending for clothing and personal care. The proportionately greater spending that the single elderly devote to housing and medical care reduces the money they can spend elsewhere. Not surprisingly, the single elderly, a group with comparatively little mobility, devote fewer resources to transportation and recreation and leisure than other groups, while the share going to personal business largely time spent on the telephoneis greater than that of any other. Income Household income affects not only the level but also the type of expenditures made by Americans. As incomes increase, families typically spend a smaller proportion of their income on foodthough

PAGE 81

74 . . . . . . . . . . . . . . . .

PAGE 82

75 in absolute terms, the highest income group spends twice as much per family member for groceries as the lowest, and eight times more for meals at restaurants. 28 Wealthier families also devote more of their resources to clothing, education, and recreation, while lower income groups spend comparatively more on health care. Such differences reflect the fact that as incomes fall, basic expenditures must be maintained while such luxuries as new clothes and recreation are cut back. Tracing spending patterns by income and income cohort is needed to understand how changes in national income may affect spending, and also to understand the fate of low income families who may be hurt by changes in income distribution. It is difficult, however, to disentangle the effects of social class and fashion on spending from the effects of income. Given the lack of data, a snapshot of spending in one year (cross-sectional data) must often be used to anticipate future changes over timean admittedly perilous assumption for extrapolation. It is not obvious, for example, that if future economic growth gave middle class families purchasing power equivalent to that of wealthy households today, they would spend the income in the same way as todays rich. Trends in Income and Income Distribution The long upward trend in median family income per family member stopped in the late 1960s (see figure 2-5). While the income available for spending has remained roughly constant, there has been a striking change in the distribution of income among families since the mid 1970s. The average incomes of all but the wealthiest families fell between 1977 and 1984, while those of families in the very highest income groups increased sharply (see table 2-6). Indeed, only the wealthiest 10 percent of American families enjoyed any growth in average income between 1977 and 1984. In effect, the wealthiest 10 zsThis study divides the 11 consumer household types introduced in the previous discussion into 7 income cohorts. The cohorts are selected by first ranking all consumer households by income per consumer household member (e.g., a family of four with a household income of $20,000 would have an income per consumer household member of $5,000); this ranked set is then divided into 7 groups consisting of equal numbers of consumer households. Expenditure categories are taken from the Consumer Expenditure Survey, developed by the Bureau of Labor Statistics at the U.S. Department of Labor. See Consumer Expenditure Demand Projection Program, ibid. 12 1 0.8 0.6 Figure 2-5.-Changes in Family Income (in 1984$) Index where 1972 = 10 1 Median Family Income --Median Family Income ($27,599 in 1972) Per Family Member ($7,818 in 1972) How To Read This Table: Median family income (half of all families earn more and half earn less than the median income) and median family income per family member (median income divided by the number of people in each family) both approximately doubled between 1950 to 1972, when growth stopped abruptly. In spite of growth in the 1960s, median income returned to 1972 levels only in 1986. SOURCE: U.S. Bureau of the Census, Current Population Reports, Series P60, No. 150, Money Income of Households, Families, and Persons in the UnRed States: 7$k75 (Washington, DC: U.S. Government Printing Office, 1986). percent of U.S. families enjoyed all of the increase in economic output of this period. The U.S. Congressional Budget Office (CBO) forecast shown in table 2-6 anticipates real income growth for nearly all family groups between 1984 and 1988, though the rate of growth will be greatest within the wealthiest decile. 29 The sharp differences in rates of income growth mean that the wealthiest families are capturing a growing share of all income available for personal spending. The wealthiest 20 percent of all families now command 50 percent of U.S. after-tax family income. Assuming that this trend continues, half the work of anticipating how Americans will spend their money during the next few decades involves anticipating the spending habits of this wealthiest quintile. The share of all national income earned by the wealthiest 1 percent of all households increased 4 percentage points between 1977 and 1984 (see figure 2-6); the share of the wealthiest groups is expected to increase through 1988, even after accounting for the effects of 1986 tax reform. 30 NU.S. congressional Budget Office, The Changing Distribution Of Ftieraf Taxes: ]975-1990 (Washington, DC: U.S. Government Printing Office, October 1987). 30(J.5. Congressional Budget Office, ibid., P TO.

PAGE 83

76 tober 1987). 1 : 4 ; ; 9 10 ToP 5% Top 1% Figure 2-6.-Distribution of Family Income After Taxes t J I J J 1 Income decile I 1 Upper income cohorts O 5 10 15 20 25 80 85 40 Percent of all family income m 1977 = 1988 How To Read This Table: In 1977, the average family in the 1 % of families with the highest incomes received about 7/0 of all family income, while in 1988 these families are expected to receive 11 % of all family income. On the other hand, there was no significant change in the share of income received by the families in the eighth income decile, which remained close to 10/0 of all family income. Families are placed in deciles as follows: all families are ranked by their total income; the bottom 10/0 are in the first decile, the next 10/0 are in the second decile, etc. (the incomes reported assume that corporate taxes are allocated to capital and not labor income). SOURCE: U.S. Congress, Congressional Budget Off Ice, The Changing Dlstribu. tion of Federal Taxes: 1975-1990, Washington, DC, October 1987. Diverging paths of income growth represents a reversal of an earlier trend. Between 1950 and 1967, for example, the wealthiest 20 percent of American families actually lost their share of total U.S. personal income, dropping from 42.7 to 40.4 percent, while the shares of all other income groupsparticularly those of the lowest 40 percent-increased. Since that time, the trend has been toward increased share among the upper income groups and a decline elsewhere; the pace of this movement doubled during the 1979-84 period. 31 Changes in the distribution of family incomes depend on changes in several different areas (discussed below): 32 l the distribution of individual wage and salary earnings, sl~~ard S mIouS, Linda H. ffirande, and Brian W. Cashell, Middle Class Erosion and Growing Income Inequality: Fact or Fiction? U.S. Congressional Research Service, Library of Congress Report No. 85-203 E, Nov. 28, 1985. Trends in income distribution between 1979 and 1984 must be treated with caution since they represent different points in the business cycle. See also the discussion on page 376. 32AII of these factors are discussed at greater length in ch. II.

PAGE 84

77 l the number of wage earners in each family (a family may have a high income by combining a number of low wages), l the distribution of capital earnings, and l the distribution of transfer payments and other earnings. While a considerable amount of work has been and is being done to disentangle these factors, much remains unknown. 33 Given available data, it appears that growing inequality of family incomes is driven primarily by shifting demographics and inequality in capital income, rather than by inequality in wage and salary earnings of workers (see figure 2-7). 34 The Gini coefficient that measures wage inequality indicates little change since 1967, while inequality in .. ssNumerous studies argue that the distribution of American income is becoming less equal, propelled by the arguments made in Barry Bluestone and Bennett Harrison, The Deirrdustriaiization of America (New York, NY: Basic Books, 1982). Recent discussions include, among others, Lester C. Thurow, A Surge in Inequality, Scientific American, vol. 256, No. 5, May 1987, pp. 30-37; Gary Burtless, Inequality in America: Where Do We Stand? The Brookings Review, summer 1987, pp. 9-16; Kathryn L. Bradbury, The Shrinking Middle Class, New England Economic Review, Federal Reserve Bank of Boston, September/October 1986; and J. Rose, The American Profile Paster (New York, NY: Pantheon Books, 1986). Others argue that such conclusions should not be drawn so quickly. Neal H. Rosenthal, in The Shrinking Middle Class: Myth or Reality? Month/y Labor Review, March 1985, pp. 3-10, indicated that trends in weekly wage and earnings (as distinct from income) measured by occupation point to movement away from both lower and middle wage groups and into the upper third. Other studies finding that income distribution is not growing less equal include Marvin H. Kosters and Murray Ross, Deficits, Taxes, and Economic Adjustments, in American Enterprise Institute, Contemporary Economic Problems, Philip Cagan (cd.), Washington, DC, 1987; Sar. A. Levitan and Peter E. Carlson, Middle Class Shrinkage? Across the Board, October 1984, pp. 55-59; and Robert J. samuelson, Middle-Class Media Myth, Nadonal.fournal, Dec. 31, 1983, Pp. 2673-2678. Attempts have been made to reconcile these conflicting views. Robert Z. Lawrence, in Sectoral Shifts and the Size of the Middle Class, The Brookirrgs Review, fall 1984, pp. 3-11, argued that growing unequal distribution of income between 1969 and 1983 was real, but was due to the demographic effects of young baby boomers entering the work force and not to the wage effects of an employment shift to lower paying industries. Patrick J. Mchfahon and John H. Tschetter, in The Declining Middle Class: A Further Analysis, Month/y Labor Review, September 1986, pp. 22-27, found that while the proportion of all jobs in highwage occupations (again, as distinct from income) increased during the 1973-82 period, the earnings distribution within those occupations grew somewhat poorer. For more on this subject, see U.S. Congressional Research Service, Middle Class Decline? Bibliography-in-Brief, 1983-1986, No. 87-68 L, Washington, DC, January 1987. JiMcKin]ey L. B]ackburn and David E. Bloom, The Effects of Technological Change on Earnings and Income Inequality in the United States, National Bureau of Economic Research, working paper No. 2337, Cambridge, MA, August 1987. Figure 2-7.-income Distribution from Three Different Perspectives 048 Earnings of Individual workers 0,46~ 0,44 j I 0,42 Total family income (from ali sources) 0 < 4 1 0,38 [ 036Totai family Income (all sources) per adult famiiy member 0!34 1 I 0,32 1 1 T I I 1 I 1 1 I 1 I 1 I 1 1 1967 1972 1977 1982 SOURCE: McKinley L. Blackburn and David E. Bloom, The Effects of Technological Change on Earnings and Income Inequality in the United States, Nationai Bureau of Economic Research, working paper No. 2337, Cambridge, MA, August 1987. family income (which includes capital income) has increased sharply since 1980. 35 These findings are based on data from the U.S. Bureau of the Census that brings with it a limitation in analysis of income distribution: the actual income of the wealthiest families is not reported. Wealthy families report only that their incomes are above some threshold level, or top code-a level that has increased at irregular intervals over the last decade. Since much income growth has occurred in the wealthiest families, it is obvious that the Census data introduces some distortion. Using data which avoids the threshold difficulty, the CBO analysis discussed above suggests that between 1977 and 1984, the Gini coefficients for aftertax income increased from 0.42 to 0.47; before taxes, the coefficients were 0.45 in 1977 and 0.48 in 1984. 3 6 sSThe Gini coefficient measures inequality. A coefficient Of O meZNIS perfect equality, a coefficient of 1 would mean that all earnings or income is received by one family. JGU.S. C ongre s s iona l] Budget Office, op. cit., footnote 29. The pre-tax Gini coefficients assume that corporate income taxis allocated to capital earnings.

PAGE 85

78 The inclusion of the income of wealthier families leads to a coefficient considerably higher than those shown in figure 2-7, Wage and Salary Earnings of Individuals. Technical change, shifts in industry structure, changing trade patterns, changing management strategies, an increase in part-time work, and a variety of other factors all play a role in determining individual wage and salary earnings; the issue is examined more thoroughly in chapter 11. In brief, there appears to be no net change in the earnings of individuals (see the top line in figure 2-7). This results in part because the convergence of male and female wages has removed a major source of inequality .37 The CBO data, which include the distribution of wages and salaries among high-income groups, indicates that high-income families are increasing their total share of all wage income as well as their share of total income. The wealthiest 10 percent of families increased their share of all labor income from 29 percent in 1977 to 32 percent in 1984.38 Unfortunately, it is impossible to use the data to separate differences in labor earnings per family from differences in earnings by each family member. Family Composition. Changes in the composition of families have had a major effect on the distribution of family income. Families with only one earner, particularly those headed by women with children, have significantly lower incomes than those with multiple earners.39 This alone contributes to family income inequality, Indeed, the shrinking size of households has produced a decline in the number of earners per household in recent years. The average household had 1.16 earners in 1979, but only 0.94 earners in 1984. 40 Inequality is further increased by a strong correlation between the wages earned by individuals in the same family. A man earning the minimum wage is most likely to be married to a person earning the minimum wage, while virtually no men earning more than $75,000 in 1984 were married to working women making less than $10,000 a year. 41 Taken together, demographic factors had the effect of increasing inequality in family income since 1980, even while there was no greater inequality in the earnings of individuals (the middle line in figure 2-7). When family income is expressed as income per family member, inequality has grown even more rapidly (the bottom line in figure 2-7). Capital Income .Income from capital sources is distributed much less equally than income from labor. Not only is capital (or unearned) income becoming a larger share of all family income, but the way this income is allocated among families is becoming less equal (see figure 2-8). In 1984, the wealthiest 10 percent of all families had 63 percent of all income from capital, and the wealthiest 1 percent of families had 37 percent.42 Similarly, ranked by net worth, the bottom 26 percent of American households owned only 10 percent of total net worth in 1984, while the top 12 percent of households owned 38 percent of total net worth. 43 4ZU.S, congressjona] Budget Office, op. cit., footnote 29. 43u.s. f)epartrnent of Commerce, Bureau of the Census, Household Weafth and Asset Ownership: 1984, Current Population Reports, Household Economic Studies, series P-70, No. 7. Figure 2-8.-Change in the Share of Family Income: 1977 to 1988 (change in percentage share of all personal income) How To Read This Table: The share of all income earned by families in the 10th income decile increased over 4 percentage points between 1977 and 1966. The share of families in the 6th income decile fell by about 1/2 of a percentage point. The incomes reported assume that corporate taxes are allocated to capital and not labor income. SOURCE: U.S. Congress, Congressional Budget Off Ice, The Changing Dlstrfbutlon of Federal Taxes: 1975-1990, Washington, DC, October 1987.

PAGE 86

79 The Influence of Income Distribution on National Spending The impact that changing income distribution could have on expenditures is shown in table 2-7. Two alternatives are explored: one, called the middle case, in which it is assumed that in 2005 all consumer households are in the middle income cohort; and another, the extremes case, in which consumer households are divided equally between the two highest and two lowest income cohorts, with none in the middle. 44 440f the 7 income cohorts introduced earlier in this chapter, sdected by ranking all consumer households by income per consumer household member and dividing the ranked set into 7 heptiles consisting of equal numbers of consumer households, the fourth heptile is the middle. Expenditure categories are taken from the Consumer Expenditure Survey, developed by the Bureau of Labor Statistics at the U.S. Department of Labor. See Consumer Expenditure Demand Projection Program, op. cit., footnote 27. The results suggest that if equality in income distribution is increased, more would be spent on food, housing (mostly home maintenance), transportation (mostly auto-related), and recreation (goods). Less would be spent on clothing and personal care (with the exception of non-apparel services like health clubs and beauty parlors) and personal business (other than phone use). Of the amenities listed here, only spending on personal business would change significantly. If there were less equalitythe extreme case proportionally more would be spent on clothing and personal care and recreation (mostly services), and less on transportation (especially new cars, though air fares would also grow). There would be little difference among the other amenity categories, though increased spending on household maintenance services and appliances would be offset by Table 2.7.The Potential Impact of Changing Income Distribution on Personal Consumer Expenditures in the Year 2005 (assuming 3 percent annual economic growth, no change indexed to 100.0) Change in spending on Change in spending on selected amenities or items selected amenities or items Amenity or item purchased Middle b Extremes c Amenity or item purchased Middle b Extremes c Food . . . . . . 104.0 98.7 Clothing and Personal Care . 94.2 106.9 Food and beverages at home. 105.4 98.7 Food and beverages away from home . . . 102.3 98.6 Tobacco . . . . 98.5 99.1 Housing . . . . . 101.9 100.9 Owner occupied . . . 104.2 93.2 Renters. . . . . 87.3 106.3 Maintenance services . 112.7 154.8 Maintenance commodities Tenants insurance. . . House furnishings . . House appliances. . . Water and sewer . . Transportation. . . . New vehicles. . . . Used vehicles . . . Vehicle maintenance . . Other private transportation Air fare . . . . . Other public transportation 111.8 105.0 107.9 97.0 101.3 104.1 103.8 113.1 110.6 113.0 91.3 84.7 86.6 99.8 98.4 132.8 112.6 90.0 73.8 90.7 89.3 108.5 95.8 Personal care commodities 92.9 109.8 Personal care services . 113.5 107.2 Mens and boys clothing . 86.8 105.8 Womens and girls clothing . . . . 98.4 105.1 Other (including jewelry). . 82.9 106.1 Footwear . . . . 94.2 114.1 Apparel services. . . 93.7 108.9 Personal Business and Communication . . Telephone . . . . Personal business . . Recreation and Leisure . . Entertainment services . Entertainment commodities . . . TV and sound . . . Lodging . . . . 83.8 103.7 77.5 104.0 91.9 117.3 110.7 98.0 100.4 100.3 100.4 103.2 120.3 88.9 105.1 96.6 How To Read This Table: Assuming 3 percent annual economic growth through 2005, in an economy where all households earned an income in the middle cohort (#4 of 7, where 7 cohort ranges are divided into equal numbers of households), the spending index for food would be 104.Oin other words, Americans would purchase 4 percent more food than would be the case with no change in income distribution. If all households were divided between the two lowest and two highest cohorts (extremes), the spending index on food would be 98.7, or 1.3 percent less than the no change scenario. ~he totals for PCE on amenity groups include certain items which have not been listed separately because the 2005 scenarios assume that expenditures on them would remain constant even with changing income distribution. See the appendix for details. bAss ume s all households In the middle inCOI?W cohort (#a of 7). cAssumes households evenly distributed between two low and two high hICOITIe cohorts (#l, 2, & 7, of 7). SOURCE: U.S. Congress, Office of Technology Assessment, Consumer Expenditure Demand Projection Program, April 1966, based on data provided by the U.S. Department of Labr, Bureau of Labor Statistics; the U.S. Department of Commerce, Bureaus of Census and of Economic Analysis; and the U.S. Department of Health and Human Services, Social Security Administration.

PAGE 87

80 declines in maintenance commodities and new homes. Comparing the middle case with the extremes, an economy consisting entirely of middle income consumer households would spend more on home purchasing, food (particularly food at restaurants), vehicles, and entertainment commodities. Households would also enjoy more of their entertainment at home. Price s The response of consumers to changes in price has received considerable theoretical attention, but applications of the theory are often frustrated by the scarcity of data. Complex interactive relationships (cross elasticities) make the problem difficultconsumption of beef may decline even when beef prices remain constant if chicken prices fall. Looking to the future, price effects pose an even greater dilemma, due to both the great difficulty of forecasting relative prices and the need to consider such issues as: will spending for travel decline if the price of communications falls? In practical terms, these issues can only be handled by examining particular amenity networks in detaila task undertaken in chapter 3. For present purposes, however, it is useful to have a feeling for the influence that a continuation of recent price trends might have on spending patterns. A complete set of recent price elasticities, constructed in a way that accounts for demographic effects, provides the basis of this analysis (again see the appendix).45 45P, Devine, FOreCasting Personal Consumption Expenditures From Cross-Section and Time-Series Data, Ph.D. Dissertation, University of Maryland, 1983. Demographics, Income, and Price: The Combined Effect The possible effects of changes in household structure, income, and prices on household expenditure are shown in table 2-8. The table shows relative changes, not absolute ones. Overall, Americans would spend more on two-thirds of the items identified. The figures assume annual 3 percent growth in GNP over the next 20 years, and that future household spending patterns can be estimated from current ones. Because incomes would increase rapidly in a high-growth economy, most of the changes presented in the table are caused by income effects rather than demographic ones (changes in population and household structure). If economic growth were slower, demographic effects would become more important. The biggest change in share of spending comes in the food amenity, which loses 5,26 percentage points as a share of national spending. Most of this results from declines in grocery eating; eating out holds a virtually constant share. Under the assumptions of these calculations, spending on recreation would capture a growing fraction of all consumer spending because of both price and income effects. Changes in household types would result in some increase in housing expenditures, but the overall effect would be more than compensated for by changes in income. Income increases would also lead to sharp growth in demand for restaurant eating, air travel, clothing, personal business, and entertainment services, while Americans would spend proportionally less on apartment rentals, smoking, and time spent on the telephone. ALTERNATIVES FOR THE FUTURE Constructing Scenarios households, income, and prices can simply be extrapolated into the future-the Trend scenarios (see The remainder of this chapter describes alternate appendix). Under these assumptions, spending tive hypotheses about the ways consumers might for 2005 can account for an aging population by asspend their money (and instruct their governments suming that the aging baby boomers in a given into spend money in their name) in 2005. One such come class will spend money in the same way as hypothesis is that things will continue much as they similar groups spend money today. There is, howare now, and that the statistics already compiled on ever, a qualification: the older baby boomers will

PAGE 88

81 Table 2-8.-The Effects of Demographic Change, Income Growth, and Price Change on U.S. Personal Consumption in 2005 (changes in percent of all spending) Percent change from 1983 due to various factors Amenity or item purchased All Demographic Income Price Interactive Food . . . . . . . . . Food and beverages at home . . . Food and beverages away from home . Tobacco. . . . . . . . Housing . . . . . . . . Owner occupied. . . . . . Renters . . . . . . . Maintenance services . . . . Maintenance commodities . . . Tenants insurance ..,. . . . . House furnishings . . . . . House appliances . . . . . Water and sewer . . . . . Transportation . . . . . . New vehicles . . . . . . Used vehicles . . . . . . Vehicle maintenance. . . . . Other private transportation. . . . Airfare . . . . . . . Other public transportation . . . Clothing and Personal Care . . . . Personal care commodities . . . Personal care services . . . . Mens and boys clothing . . . . Womens and girls clothing . . . Other (including jewelry) . . . . Footwear . . . . . . . Apparel services . . . . . Personal Business and Communication . Telephone . . . . . . . Personal business . . . . . Recreation and Leisure . . . . . Entertainment services . . . . Entertainment commodities . . . TV and sound . . . . . . Lodging . . . . . . . .26 0.07 ,41 0.04 .15 0.03 .70 0.00 .48 0.40 0.28 0.19 ,91 0.32 0.01 0.01 .49 0.02 ,02 0.04 0.95 0.03 0.71 0.01 .01 0.01 .51 0.06 .33 0.03 .24 0.01 .13 0.04 .02 0.00 0.37 0.00 .17 0.00 1.84 0.10 0.06 0.00 0.11 0.01 0.39 0.02 0.94 0.06 0.24 0.01 0.03 0.01 0.06 0.00 0.95 0.07 0.17 0.02 0.78 0.09 3.47 0.10 1.26 0.06 0.89 0.02 0.83 0.00 0.49 0.02 .3 7 .0 5 1.31 .6 3 .9 1 0.00 .81 0.10 .30 0.02 0.06 0.17 .15 0.38 .00 0.05 .06 .01 0.37 0.02 1.63 0.06 .02 0.38 0.87 0.25 0.02 0.06 1.26 .35 1.61 2.02 1.43 0.20 0.11 0.28 .8 7 .4 5 .3 0 .1 1 1.72 0.47 0.19 .0 7 .2 4 .0 9 0.84 0.47 0.16 .9 6 .3 3 .2 9 .1 4 .0 1 .0 1 .1 8 0.01 .0 0 0.11 .0 2 .0 5 .0 1 .0 1 .0 0 .1 6 0.61 .7 7 1.26 .2 1 0.64 0.68 0.14 .10 0.06 .1 9 0.04 0.11 0.00 0.03 .01 0.03 0.01 0.02 0.06 .0 3 0.02 .0 2 0.01 0.04 0.00 .0 0 .0 0 0.10 0.00 0.01 0.01 0.05 0.01 0.01 0.00 .2 3 .0 8 .1 5 0.09 .0 2 0.03 0.03 0.06 How To Read This Table: Assuming 3 percent annual economic growth through 2005,2005 household distribution, and prices and incomes adjusted to this growth, the percentage of American spending on food eaten at home (as a share of the items Iisted here-roughly three-quarters of all personal spending) would decline by 4.41 percentage points. Changing incomes would account for a drop of 4.05 points and changing prices would account for a 0.45 point drop, while demographic changes would exert a slight positive trend of O.04 percentage points; the effect of interaction between these factors would be a rise of O.06. NOTES This table estimates how U.S. consumer spending on selected items and amenities could change, and attempts to isolate what factors may contribute to that change. The All column assumes 30/0 annual economic growth through 2005, 2005 household structure as developed earlier in this chapter, and a set of possible price changes for these items in 2Q05 as outlined in the appendix. Incomes are then raised by 35.50/., the level at which Americans will have enough purchasing power to satisfy the estimate of personal spending in 2005 (also developed in the appendix). For individual components of change l For demographic changes in 2005: see table 2-4; price and income held at 1983 levels. l For income changes In 2005: incomes raised by 35.5 percent; demographics and prices held at 1983 levels. l For price changes in 2G05: see the appendix; demographics and income held at 1963 levels. The effects of these three components that cannot be traced individually but are rather the result of a combination of factors are captured in the Interactive column SOURCE U S Congress, Office of Technology Assessment, Consumer Expenditure Demand Projection Program, April 1986, based on data provided by the U.S. Department of Labor, Bureau of Labor Statistics; the U.S. Department of Commerce, Bureaus of Census and of Economic Analysis; and the US Department of Health and Human Services, Social Security Administration

PAGE 89

82 have had life-histories significantly different from older Americans today. They will be better educated and healthier, and will have a far higher share of women retired from work than is currently the case. It is extremely difficult to predict whether this effect will result in new spending patterns. 4G Similarly, Americas ethnic composition will change. Between now and 2000, roughly one-half of the increase in the U.S. population will be from minority ethnic groups.47 These developments pose problems for analyzing future spending patterns, because it is nearly impossible to allow for the effects of changing ethnic composition. For these and many other reasons already discussed, trends can be misleading during periods of fundamental change. The new environment in which the economy operates allows for a number of major shifts in the way households elect to achieve amenity through private and public choice. Given the wide array of possible choices, and the enormous range of uncertainties about the cost and capability of emerging technology, there is no completely satisfactory method for outlining the possibilities, nor any mathematical technique for producing them. Potential changes are represented though an illustrative set of Alternative scenarios, which will be developed further in chapter 3. These scenarios were designed with the help of individuals familiar with the operation of the eight major amenity categories (Food, Health, Education, etc.), who were charged with describing a way that new technology could lead to a significant increase in productivity during the next 20 years given appropriate changes in regulation, information flows, and other factors. Along with the Trend scenarios, the Alternatives share common assumptions about such factors as population growth, rates of consumer household formation, and rates of income and GNP growth (see box 2-B and the appendix for details). dGWilliam &er and Eric H. Shaw, How Older Americans Spend Their Money, American Demographics, September 1987, p. 41. 47 U.S. Bureau of the Census, Statistic/Abstract of the United States: 1987 (107th cd.), Washington, DC, 1986. Changing ethnic composition, of course, has occurred for some time; the U.S. Bureau of the Census has recently reported that the U.S. Hispanic population has increased by 30 percent since 1980, as opposed to a 6-percent increase in the non-Hispanic population. See U.S. Bureau of the Census, The Hispanic Population in the United States: March 1986 and 1987, Current Population Reports, Series P-20, No. 416, Washington, DC, August 1987. Given these scenarios, two patterns of national income growth are examined: a pessimistic assumption, represented by 1.5 percent annual growth in GNP (as in the late 1970s and early 1980s); and an optimistic assumption, represented by 3 percent per year GNP growth. This range is plainly not meant to be a forecast, but rather is designed to bracket a wide yet reasonable range of possibilities. In principle, it is possible to develop a closed, internally consistent model that can accommodate these connections. It proves extremely difficult to use such models in practice when exploring the possibility of significant changes in technology or in private and public management during a period of 20 years. While the analysis presented throughout this work is consistent in that the income generated by output is equal to the money spent on final and intermediate products, it is not based on a closed model of the economy.48 Summarizing the Results The scenarios are summarized in table 2-9; chapter 3 and the appendix provide further detail. The proportion of expenditure on Food goes down no matter what growth rate or assumption about technology is usedas does, to a lesser extent, that spent on Transportation. On the other hand, more is spent on Recreation and Leisure in all cases, and both Clothing and Personal Care (mostly clothing) and Personal Business and Communication gain in all but the 1.5 percent Trend scenario (in which they hold the same share as 1983). Considering first the 3 percent scenarios, Food purchases are expected to decline as a fraction of all spending, though more slowly than in the past largely because of an increase in the proportion of more expensive food eaten away from home. The Trend and Alternative scenarios differ considerably in estimates of future Health spending. The Trend assessment shows spending for Health reaching 14 percent of the total in 2005 while the Alternative scenario suggests 11 percent, due to greater reliance on preventative (as opposed to curative) techniques. d~onslstent Sets of forecasts were developed by beginning with rough estimates of structural change in production and productivity, using the resulting GNP growth to estimate purchasing, and then using this purchasing to estimate production and GNP. The estimates are combined explicitly in ch. 13.

PAGE 90

83 Box 2-B.Basic Strategies in Developing Scenarios 1. Demographics Population was forecast using a version of the projection model used by the U.S. Social Security Administration ] modified for use on a personal computer. z The projection includes an allowance for illegal immigration somewhat higher than that used by the U.S. Bureau of Census middle series. For more detail, see the appendix. Estimates of the number of people in each age group can be converted to estimates of household types given assumptions about future marriage and divorce rates. Annual marriage rates have remained within ranges of 9.9 to 10.9 per thousand since 1968, and of 10.2 to 10.6 since 1978. Annual divorce rates, after rising consistently between 1960 and 1979, have since remained fairly constant at around 5.0 per thousand. The hypothesis used here is that these recent steady rates will continue into the foreseeable future. Assuming that the likelihood that a person of a given age and sex becomes a member of any one type of household is the same in the year 2005 as it was in 1984, the number and type of future households can then be calculated (see table 2-4). 4 2. GNP Growth Two rates of gross national product (GNP) growth are considered, one based on the optimistic assumption that rates of productivity growth characterizing the years 1965 to 1975 can be recovered, and one based on the assumption that productivity would grow at the much slower 1975 to 1985 rates. Together with assumptions about size of the work force and other factors (discussed in more detail in ch. 13), these result in GNP growth rates of approximately 1.5 and 3 percent per year. 3. Allocating GNP to Consumption All the cases considered in this analysis share the assumption that the sum of personal consumption and government spending will remain at a constant share of 85 percent of GNP. This ratio has not varied by more than 3 percentage points for nearly 30 years. Personal consumption expenses have fluctuated in recent years from a low of 61.5 percent of GNP in 1973 to a high of 65.5 percent in 1983, and have since remained around 65 percent. The analysis also assumes that defense will maintain its present 7 percent share of GNP. 5 5. Prices Relative prices are assumed to change at roughly the rate of the past two decades, with exceptions based on changes in trade and production technology discussed in the appendix. The set is made consistent in the sense that a shift to the new price set would not change total spending. 6. Allocating Consumption by Product Type With the exceptions discussed in the text, money available for consumption is assumed to be allocated given estimates of price, income, and household size using the methods described to produce table 2-9. The calculations are based on an analysis of spending by income cohort and household type. G They are not based on an assumption that spending rises or falls exponentially with income, but allow for more complex relationships. For example, spending on used cars first rises and then falls as incomes increase. 7 IThe assumptions on life expectancy and fertility used here have been selected from the ranges developed by the Social S@curity Administration (SSA); see (J S Department of Health and Human Services, S~ial Security Administration, Social Security Area Population Projections, 1984, Actuarial Study No 92, Washington, DC, May 1984
PAGE 91

84 Table 2-9.Spending on the Amenities: 1983 and the Scenarios (in percent of personal and total spending in 1983 dollars) year 2005 1983 Trend 3 % Trend 1.5/0 ALT 3 % ALT 1.50/0 PCE Total PCE Total PCE Total PCE Total PCE Total Food . . . . 200/0 16% 16% 13%0 19% 15% 16% 12% 16% 13% Housing . . . 26 20 25 19 27 21 23 18 23 18 Transportation . . 13 12 11 10 11 11 10 10 11 11 Health . . . 12 11 15 14 14 13 12 11 13 13 Clothing and Personal Care . . . 9 7 11 8 9 7 12 9 11 8 Education . . . 2 7 1 5 1 6 1 6 2 7 Personal Business and Communication . 8 6 9 7 8 6 11 8 10 7 Recreation and Leisure. 10 8 13 10 11 9 15 12 14 11 Defense . . . 0 7 0 7 0 7 0 7 0 7 Other Government Expenditure . . 0 5 0 7 0 6 0 6 0 4 Total (percent) . 100 100 100 100 100 100 ($1983 billion) ....2,229 100 100 100 100 2,905 4,270 5,565 3,093 4,031 4,270 5,565 3,093 4,031 NOTE: Total Includes all government purchases of goods and services. ALT Alternatives (see text for deflnltlons). Totals may not add to 100 due to rounding, SOURCE: 1983 statistics from U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Survey of Current Business, table 2.4; Trend projections drawn from U.S. Congress, Office of Technology Assessment, Consumer Expenditure Demand Projection Program, April 1986, based on data provided in the Consumer Expenditure Survey, U.S. Department of Labor, Bureau of Labor Statistics (see the appendix for details): Alternatives from the Office of Technology Assessment. The share of Housing also differs between the two 3 percent scenarios, being 1 percent lower in the Alternative than in the Trend. This is due to the assumed sharp fall in the cost of household energy utilities, and to an assumption that real shelter costs can be kept at 1970 levels using technical improvements in the production and operation of homes. The shares of income spent for Personal Business and Communication, Recreation and Leisure, and Clothing and Personal Care are all higher in the Alternative scenario. In the case of Personal Business and Communication, the difference is created by greater use of both the telephone and services outside the house, Similarly, the Alternative assumes that technology allows for an expansion of recreational telephone use, and of the information, educational, and entertainment resources available in the home through television and computers. The Alternative case also envisions a 15 percent drop in the real price of clothing, which boosts spending considerably. Overall, as might be expected, the 3 percent Alternative case differs most from 1983 in terms of distribution of personal spending, since high economic growth is combined with liberal assumptions about the impact of technology on purchasing patterns. In the low growth (1.5 percent) scenarios, the Trend patterns are quite similar to those of 1983. Food expenditures are lower in the Alternative case than in the Trend, despite generous assumptions about family diet. The share of Housing falls even more sharply in the Alternative, largely due to the reasons given for the 3 percent cases. On the other hand, expenditures on Recreation and Leisure are distinctly higher in the Alternative, in which new technologies allow for an expansion of entertainment products and services at lower pricesmaking a wide range of entertainment resources available to all. Government spending as a whole (counting all public expenditures on the amenities) increases somewhat more rapidly in the Alternative case than the Trend, primarily because of assumed increases in spending on education. The statistics just discussed do not directly address some of the most critical questions. What, for example, will happen to the real quality of housing, to the flexibility and convenience of transportation, or to the state of American health in these scenarios? To what extent will future changes in consumer demand bring real improvements in amenity to the lives of Americans? Such questions cannot be answered with precision. To the extent that it is possi-

PAGE 92

85 ble to talk in terms of quality as well as quantity, progress be clearly linked to an improvement in the discussion of changes in amenity will be apamenity provides a focus for the abstract calculus preached on a sector-by-sector basis in chapter 3. of productivity and technology that is the subject of If nothing else, a patient insistence that economic much of this report.

PAGE 93

Chapter 3 Eight Amenity Groups

PAGE 94

CONTENTS Page Food . . . . . . . . . . . . . . . . 89 Quality and Variety. . . . . . . . . . . . . . 89 Nutritional Consequences . . . . . . . . . . . . 91 Choices and Consequences . . . . . . . . . . . 94 Health . . . . . . . . . . . . . . . . 94 Societys Recipe for Health . . . . . . . . . . . 94 The Record . . . . . . . . . . . . . . 96 Choices and Consequences . . . . . . . . . ..........103 Housing . . . . . . . . . . . ....................105 The Housing Recipe . . . . . . . . . ..............105 Trends in U.S. Housing Quality and Affordabihty. ..........................107 Choices and Consequences . . . . . . . . ............113 Transportation . . . . . . . . . ....................114 Trends in Personal Transport . . . . . .....................116 Looking to the Future. . . . . . . . . .................119 Choices and Consequences . . . . . . . . .............121 Clothing and Personal Care. . . . . . . . . . .........125 Education . . . . . . . . . .......................126 Defining Demand for Education . . . . . ....................126 Trends in lnputs and Outputs . . . .............................127 Choices and-Consequences . . . . . .........................130 Personal Business and Communication . . . . . . . . ...132 Demand for Communications and Information .............................133 I)emand for Business Services . . . . . . . . . . ...135 Choices and Consequences . . . . . .................,.......136 Recreation and Leisure . . . . . . . ....................137 Recent Trends.. . . . . . . . . . ...................137 Choices and Consequences . . . . . . . . . . . 140

PAGE 95

Chapter 3 Eight Amenity Groups This chapter examines characteristics of eight major household amenity groups and explores options for the next two decades. It does not cover national defense or government activities not directly associated with a household amenity. The discussions form a diverse set but are all designed around the following plan. They begin with a description of the way amenity might be measureda description of the units that might be used to gauge progress even if measurement is difficult because of the absence of good data. Second, they show how themes raised in the previous chapter apply in each amenity group. The themes include trade-offs between time, consumer spending, and government spending; growth in specialized markets; and changes in regulations that affect consumer purchasing. They end with a brief description of hypothetical patterns of future spending, These descriptions are used later in this volume to explore the potential impact of future consumption patterns on economic structure and on jobs. ] IA reader interested in tracing the performance of a specific amenity network from consumption recipe to job generation can read appropriate sections from chs. 3,6,9 and 12. FOOD There are two criteria against which to measure changes in the price, quality, information, and other factors that influence a nations diet: 1. do the changes improve the quality, variety, and convenience of the food purchased and 2. has the new diet improved health? Growth in choice available to most American households does not necessarily translate into a diet more likely to promote health. Indeed, higher incomes may lead to diets that are less healthy if consumption of fats and alcohol increases while consumption of bulk fiber declines.z An unhealthy diet may be purchased as a result of an informed decision to ignore the advice of physicians, ignorance of established links between diet and health (or lack of information about product contents), or the absence of credible information linking diet and health. The relative role of these factors, and therefore the influence of programs designed to facilitate consumer choice, remains ambiguous. These issues are also relevant in the section on spending for health that follows. Quality and Variety While people still go hungry in the United States, the quality and variety of the food eaten by most ZL, A Cohen, Diet and cancer, Scie~t;k American, VOI. 257, No. 5, November 1987, pp. 42-48. Americans have improved considerably. Many foods have become cheaper, production and packaging technologies have improved, the health implications of different diets are more widely understood, increased trade has brought a wider variety of foods to the U.S. consumer, and more meals are eaten outside the home, partly because more women are in the work force. Net Food Purchases The total quantity of food purchased per person increased by only 5.5 percent between 1965 and 1985 (largely the result of demographic factors), but food quality and diversity appears to have increased substantially. A variety of large and small grocery stores have emerged. The typical new supermarket may sell 15,000 to 20,000 separate products, and large stores may sell many more types. Red meat consumption per capita peaked in the mid-1970s (red meat prices increased substantially during the late 1970s as lucrative grain exports increased feed prices), but per capita consumption of fish and poultry increased 34 and 69 percent respectively between 1965 and 1985. 3 During the same two decades, per capita purchases of fresh fruits increased 12 percent, with declines in staples such as oranges SK.L. Bunch, Food consumption, Prices and Expenditures U.S. Department of Agriculture, Economic Research Service, Statistical Bulletin 749, Washington, DC, January 1987, 8 9

PAGE 96

90 and grapefruit offset by large increases in demand for more exotic fruits; per capita consumption of fresh vegetables grew 30 percent while that of frozen vegetables grew 50 percent. Increased variety in American cooking is measured in part by the near doubling in per capita consumption of spices over the period. There was also a 30-percent increase in per capita consumption of sugars and other sweeteners. A survey by the U.S. Department of Agriculture (USDA) in 1978 showed that the average respondent ate 25 to 28 different types of food over a 3-day period, although the range was from 4 to 63 unique food items. 4 Total consumption varied greatly with age and sex; males older than 23 appear to eat more than the average number of foods, and females aged 15 to 22 consume less than the averages Eating Away From Home A system for tracking food expenditures developed by USDA suggests that by 1983, some 40.9 percent of all money spent on food was eaten away from home.G Only about 20 percent of the quantity of food and beverages consumed are eaten away from home; 3,7 out of 21 meals are eaten away from home.7 By the turn of the century, the average American may spend half of every food dollar away from home. The choice of eating place depends heavily on household type and income (see table 3-l). The addition of a family member over the age of 65 typically increases household spending in cafeterias by 9 percent. The addition of a child of age 12 or younger reduces overall spending in all food service estab 4J Krebs.smith, et al., Variety in Foods, What /S America Eating? (New York, NY: National Academy Press, 1986). Ibid. filJ s f)epartment of Agriculture, Economic Research Service, U.S. Food Expenditures, 1954-78: New Measuresat Point of Sale and by Type of Purchaser, AER-431, Washington, DC. The USDA statistic is considerably higher than that reported by the U.S. Bureau of Labor Statistics. [n that survey, eating away from home represented 28 percent of all personal consumption expenditures for food by 1983, an increase from 24 percent 10 years earlier. The USDA series includes expenditures for foods in places where food sales are not the main enterprise (such as hotels and motels, airlines, recreation places, military installations, and retail stores) and also includes the value of government food subsidies (such as the school lunch program). It is difficult to determine whether differences in definitions can account for the magnitude of the difference. TJohn M. Con nor, et al., The Food Manufacturing Industries (LeXIWkm, MA: D.C. Heath & Co., 1985), pp. 24-25. lishments by 7 percent, but increases spending in fast food restaurants by 26 percent and in take-out restaurants by 16 percents Single-person households, whether young or old, spend a large share of their food budget away from hometypically 45 percent.g Families with working wives spend 28 percent more on food away from home than comparable families with housewives, although the increase drops to 7 percent after controlling for income;l most of the rise goes to specialty restaurants. In contrast, elderly households spend more in cafeterias and less in specialty restaurants. It is apparent that the food purchased in fast food restaurants has become more diverse. In addition to the traditional hamburgers, french fries, soft drinks, sweets, and pizza, fast food now embraces seafood, chicken, sandwiches, breakfast items, baked potatoes, and salad. The number of cuisines available as fast food has also increased, and now includes Mexican, Chinese, Indian, and other types. However, since fast food outlets have often expanded their markets at the expense of family-run restaurants, often serving regional menus, it is not obvious that real diversity has increased.11 Growth in other segments of the food service sector, particularly in more expensive restaurants, has introduced many Americans to varied cuisines. The upgrading of food purchasing patternsresulting in higher prices per amount of food consumed has been expressed most recently in the growth of the up-market hamburger restaurants, which not only offer a more diverse and expensive menu but also sell alcohol. Traditional restaurants still claimed 51 percent of all sales in separate eating places in 1982, but fast food was a close second with 44 percent. 12 Of franchised restaurants, 42 percent serve hamburgers, 17.5 percent pizza, 12.4 percent chicken, fJUs Department of Agriculture, Economic Research Service, Food Consumption, Prices and Expenditures 1963-83, Statistical Bulletin No. 713, Washington, DC, 1984. 9A. Arbe], Higher Energy Cost and the Demand for Restaurant ServicesA Time-Series Analysis, lrrternational Journal of Hospitality Management, vol. 2, No. 2, 1983, pp. 83-87. 10.f, Kin~y, Food Away From Home Expenditures by source of HOUS hold Income, paper presented at the annual meeting of Agricultural Economics Association, Clemson, South Carolina, 1981. I IJohn M. Connor, review comment, Department of Agricultural E CO nomics, Purdue University, 1987. ITUs f)epaflment of commerce, Bureau of Economic Analysis, Franchising in the Economy, 1982-1 984, report, Washington, DC, 1984.

PAGE 97

91 Table 3-1 .Dining Out, 1985 Proportion Average number Restaurant type of occasions of meals per week A. Diners Preferences Quick-service (primarily fast-food) . . Hamburger . . . . . . . Pizza . . . . . . . . Chicken . . . . . . . . Ice cream . . . . . . . Sandwich . . . . . . . Donut . . . . . . . . Mexican . . . . . . . Fish/seafood . . . . . . Oriental . . . . . . . . All other . . . . . . . Midscale (table service, Iimited alcohol). Varied menu . . . . . . . Family style . . . . . . . Cafeteria . . . . . . . Family steak . . . . . . Oriental . . . . . . . . Fish/seafood . . . . . . Italian . . . . . . . . Mexican . . . . . . . All other . . . . . . . Upscale (full-service) . . . . . Varied menu . . . . . . . Hotel . . . . . . . . Fish/seafood . . . . . . Mexican . . . . . . . Oriental . . . . . . . . Italian . . . . . . . . Steak . . . . . . . . All other . . . . . . . 64.l% 24.7 10.8 5.3 4.4 3.5 2.1 1.9 0.6 9.4 27.1% 9.7 4.0 3.1 2.2 0.9 0.6 0.6 0.6 5.4 8.8% 4.2 1.2 0.8 0.8 0.7 0.4 0.4 0.3 B. Who Dines Out All individuals . . . . . . Men . . . . . . . . Women . . . . . . . . Household size: One member . . . . . . Two members . . . . . . Three members . . . . . . Four members. . . . . . . Five or more members . . . . Household income: Under $l0,000 . . . . . . $10,000-14,999 . . . . . . $15,000-19,999 . . . . . . $20,000-24,999 . . . . . . $25,000-34,999 . . . . . . $35,000-49,999 . . . . . . $50,000 and over . . . . . . 3.7 4.2 3.4 4.3 3.1 3.9 4.1 3.7 2.8 3.2 3.6 3.7 3.8 4.5 4.6 Percent of food dollar spent away C. Dining Dollars from home 1955 . . . . . . . . 25.00/6 1967 . . . . . . . . 29.2 1972 . . . . . . . . 33.6 1977 . . . . . . . . 37.4 1982 . . . . . . . . 38.2 1983 . . . . . . . . 39.5 1985 (est) . . . . . . . 40.0 SOURCE National Restaurant Association, 1985, cited in Leisure Lifestyles, The Wa//Street Jourrra/, Apr.21, 1988, p. 180. 4.9 percent Mexican food, 4.5 percent seafood, and 2.4 percent pancakes. Nutritional Consequences General Trends Has all this led to a more healthy diet? The question is vexing because of uncertainties about the links between diet and health. Nutritional factors have been linked to a number of chronic diseases, such as cancer, diabetes, heart disease, and osteoporosis. Researchers are generally agreed that over-consumption of foods high in fat, salt, or cholesterol significantly increases the risk of heart disease. Some researchers have estimated that 30 to 60 percent of cancers stem from dietary factors. ]3 Cancer appears to need both an agent that creates a mutation in DNA l)N~t]Ona] ACademv Of Sciences, National Research Council, DieL Nutritiorr, and Cancer (Washington, DCNationalAcademy Press, 1982~ pp. 1-14. and a promoter agent that leads to tumor growth. Some foods may actually be antipromotersand inhibit tumor formation.14 But since it is difficult to link the disease with specific diets or foods, nutritionists remain tentative in recommending preventive steps. The risk of giving misleading information must be weighed against the likelihood that the information can help improve peoples health.l5 The risks of contamination in food from pesticides and other sources remain controversial. Standards for processed and fresh foods differ greatly. Many foods have naturally occurring chemicals that are 1lIA. coher-1, op. cit., footnote 2. IsThe National Research Counci]s Committee on Diet, Nutrition, and Cancer concludedin 1982 that cancers of most major sites are influenced by dietar-ypatterns. But the committee further concluded that the data are not sufficient to quantify the contribution of diet to the overall cancer risk or to determine the percent reduction in risk that might be achieved by dietary modifications. See Jean-Pierre Habicht, The Role of Nutrition Research in Policy and Program Planning,in What Is America Eating? op. cit. footnote4, pp. 144-147.

PAGE 98

92 highly carcinogenic. Net risk proves difficult to establish.l6 Adherence to emerging dietary guidelines designed to lower the risk of specific diseases may lead to nutrient intakes below the Recommended Daily Allowances (RDAs) established by the National Research Councils Food and Nutrition Board for the nutrient needs of healthy individuals in the United States. Which guideline is then correct? One expert has observed that: lowering of recommended total fat and salt intakes for disease prevention is almost always incompatible with American palates and pocketbooks if the diet must also fulfill all the RDAs. If one also wants to consume a diet with fewer potential carcinogens and more protective factors against cancer, attaining the RDAs is jeopardized even more.17 Conversely, adherence to some diets that satisfy the RDAs could increase health risks.18 Increasing variety of foods can also improve diet. Consumption of a greater variety of foods has been associated with a higher degree of compliance with established dietary guidelines; greater variety has been positively correlated to higher intakes of each of 11 nutrients.l g Americans have enjoyed a steady increase in average per capita consumption of virtually all critical minerals and nutrients. In recent years, consumption of products high in saturated fat and cholesterol has declinedanimal fats by 20 percent, eggs by 19 percent, and whole milk by 58 percent. Consumption of other milk beverages, which include lowfat and skim milk, increased a spectacular 300 percent in 20 years. 20 Consumption of cholesterol declined from a peak of 570 mg/capita in the late 1950s to 480 mg/capita in 1985 (there has been no measISf)uring 1987, two studies conducted by the U.S. General Accounting Office found that less than 0.1 percent of imported food was checked for illegal pesticides or high levels of legal chemicals like lead. Two studies by the National Academy of Sciences criticized the U.S. Department of Agricultures poultry inspection and techniques for determining carcinogens in tomatoes, potatoes, oranges, apples, pork, and a variety of other common products. Cited in Keith Schneider, Congress Looks to the American Table Amid Questions on Food Safety, The New York Times, Jan. 22, 1987, p, A14. 17J. p, f-fabicht, op. cit., footnote 15. ISJohanna T. Dwyer, Nutrition Education, What/s America Eating? op. cit., footnote 4; see also J.P. Habicht, op. cit., footnote 15. 19J. Kreb~Smith, et al., op. cit., footnote 4, pp. 126-140. ZOfU. Bunch, op. cit., footnote 3. urable decline since 1975). The American Heart Association diet recommends 300 mg/day of cholesterol mg/day for those with elevated cholesterol levels in their blood.2l A recent study indicates that 81 percent of Americans knew about the health effects of cholesterol in 1986; the same survey found that while only 23 percent claim to have actually changed their eating habits as a result of this information, this percentage is nearly double that of 1983. 22 High fat diets have been linked with breast cancers, and there is weaker evidence linking high fat/Iow fiber diets to cancers of the colon, pancreas, ovaries, and prostate. The fact that different kinds of fats and fibers appear to have different effects has greatly complicated efforts to construct links. For example, Eskimos in Greenland and people living in Greece and Spain have comparatively low levels of breast cancer even though they have high fat intake possibly because the fat was derived from fish and olive oil.23 While per capita consumption of animal and vegetable fats rose rapidly between 1900 and the 1960s, between 1965 and 1985 per capita consumption of animal fats declined 20 percent while consumption of vegetable fats and oils increased 60 percent. 24 As refrigeration replaced salting, pickling, and smoking as a form of food preservation, deaths from gastric cancer fell sharply. The Japanese, who still eat a significant amount of food preserved in traditional ways, suffer far higher rates of stomach cancer. After increasing 56 percent between 1965 and 1981, alcohol consumption per capita has since fallen steadily, declining 8.6 percent between 1981 and 1985. Beer consumption also began declining in 1981, falling to 34.5 gallons per person per year in 1985, as has consumption of distilled spirits. Wine consumption has increased steadily since 1965 (to 3.8 gallons per person per year in 1985). In 1985, more than 14 percent of the total consumption of ZID M f-fegsted and R.J. Nicolosi, Individual Variation in Serum Chc. Iesterol Levels, Proceedings of the /Vationaf Academy of Sciences vol. 84, No. 17, Sept. 2, 1987, p. 6259; L. Roberts, Measuring Cholesterol Is As Tricky As Lowering It, Science, vol. 238, No. 4826, Oct. 23, 1987, pp. 482-83. z Cholesterol Awareness Survey, The National Heart, Lung and Blood Institute of the National Institute of Health, Bethesda, MD, Dec. 4, 1986. Z3L.A. Cohen, op. cit., footnote 2, P 46. Z4K. L Bunch, op. cit., footnote 3, p. 39.

PAGE 99

93 wine was in the form of coolers, a mixture of wine and fruit juice. 25 Averages, of course, can hide a number of ills. Serious defects in diets remain for special populations such as the elderly, the homeless, the poor, migrant laborers, female-headed households, pregnant women, and psychologically debilitated people. Unfortunately, there is no reliable information on the diets and nutritional status of the population at risk. 2G There are other important gaps in our knowledge because existing surveys are not comprehensive or are out of date. By one estimate, however, there may be as many as 12 million children and 8 million adults who are chronically short of nutrients necessary for growth and good health. The National Household Food Consumption Survey of 1982 found that more than 80 percent of households where spending on food was similar to that of food stamp-recipients failed to obtain the RDAs. 27 An examination of households at risk of poor nutrition indicates that in 1977, 12 percent of Americans were not getting enough food energy and nearly onequarter failed to get enough vitamin B-6 (see table 3-2). Given a continuation of present trends, the analysis suggests that the dietary status of the population will actually worsen in terms of five out of six components examined. About one-quarter of U.S. households would continue to have availability levels of calcium and vitamin B-6 below 75 percent of the RDA in 1995. The GNP would have to increase much faster than 3 percent annually to make a marked improvement on the dietary availability of some nutrients for the groups receiving inadequate nutrition. Claims that Americans are going hungry have been disputed using statistics showing no increase in childhood anemia. Given the poverty of data it is difficult to resolve the issue with precision. For example, not much is known about the relationship between the supply of food and its utilization by specific individuals and various types of households. The per capita food consumption series maintained by USDA provides only crude consumption estimates for the average American. Only one national survey, the 1977 to 1978 National Food Consumption Survey (NFCS), is currently available for assessing major variations in food consumption among individuals and households within the U.S. population. Limited annual surveys initiated by USDA in 1985 (Continuing Survey of Food Intakes by Individuals) will improve this database. Even assuming that data on nutritional status were much improved, scientific understanding of the health consequences of diet remains tentative in many respects. Probably the most salient example concerns the RDAs. Simply put, the specific health risks associated with nutrient intakes below the RDAs remain largely unknown. 28 Eating Away From Home The increase in eating out is complicating dietary assessments. According to one estimate, the 16 percent of all eating and drinking occasions that occur away from home supply 18 percent of total food energy and about 16 percent of most vitamins and minerals. Men and children consume more of their food away from home than do women, and so obtain more of their food energy and nutrients from those occasions. 29 s Ibid., p. 5. ZbJ,p, Habicht, op. cit., footnote 15. Z7J, L Brown, HUr-lger in the U.S., Scientific American, vol. 2S3, No. 2, February 1987, pp. 37-41. Table 3-2.Percentage of Households With Selected Dietary Availability Levels Below 75 Percent of the Recommended Daily Allowance (1977/78 and 1995, assuming 3 percent annual growth in GNP) Year Food energy Calcium Iron Magnesium Vitamin A Vitamin B-6 1977 11.9 24.4 8. 0 13.6 16.7 23.4 1995. 12.6 25.2 8. 4 14.0 15.8 24.0 SOURCE: Future Food Patterns of the U.S. Population, cooperative agreement between the University of Missouri at Columbia and the Office of Technology Assessment, No. 433-7130.0, October 1985. One recent study of snacking and away-from-home eating indicated that where and when foods were consumed had very limited impact on their nutritional status. The researchers noted, however, that future investigations might profitably focus on those 28 Betty B. Peterkin, Assessment of Diet Quality and the U.S. Department of Agricultures Nutrition Policy and Research, What/s America Eating? op. cit., footnote 4. 29 Robert B. Reese and Sharon J. Mickle, Where to EatAt Home or Away? 1982 Yearbook of Agriculture (Washington DC: U.S. Government Printing Office, 1983). sOKaren J. Morgan and Basile Goungetas, Snacking and Eating Away from Home, What IS America Eating? op. cit., footnote 4, p. 123.

PAGE 100

94 who are nutritionally at risk. Furthermore, the away from home share of food expenditures has increased since 1977 and will probably continue to do so over the next decade. The nutritional consequences of this important trend warrant more attention. Choices and Consequences The Trend scenarios for expenditure on the Food amenity (as defined in ch. 2) are based on present patterns of household expenditure, and no significant changes in regulatory programs affecting food prices. The Alternative scenarios differ in that they assume that all Americans have access to a healthy diet and that all income groups pay more attention to the health consequences of their diet. Figure 3-1 indicates the changes that would occur in American eating habits if the recommendations of the National Research Council and other recent studies of diet and health were followed. While they do not correspond precisely to the recommendations shown in figure 3-1, USDA has developed estimates of nutritious diets that could be purchased by households with different incomes. These diets contain recommended intake of calories and nutrients and would reduce fat, cholesterol, sweeteners, and sodium below current levels. They contain more grain products, vegetables, and fruits, and less cheese, eggs, fats, oils, sugar, and soft drinks, than are now eaten in most U.S. households. These diets were used to explore the implications of a shift to a different pattern of food consumption in America. The liberal USDA diet was used as an estimate of the way Americans might eat if the U.S. gross national product (GNP) grew 3 percent per year and the moderate cost diet was used for the 1.5 percent case. These plans are based on the eating habits of the upper half of all households ranked by expenditure on food. Food expenditures at home and away from home for the different scenarios are given in table 3-3. The total value of purchases of the Food amenity (including alcohol) is actually lower in the Alternatives than in the Trend cases, even though they explicitly provide healthy, balanced diets for all Americans and assume a continued increase in purchases of food away from home. The Alternative cases assume that expenditures on cigarettes and alcohol fall below the Trend cases. These estimates are based on a continuation of past trends in per capita consumption expenditures roughly, a 1.6 percent annual increase in alcohol consumption between 1960 and 1983, and a .9 percent decrease in tobacco between 1973 and 1983. Figure 3-1.-Current and Recommended Diet 400 300 200 100 n Grams per person per day ~ Fat Carbohydrate Protein Fiber ~ Current Diet m Recommended Diet SOURCE: L.A. Cohen, Diet and Cancer, Sclerrtiflc American, November 1987, p. 48; and National Cancer Institute, Division of Cancer Prevention and Control, Annual Cancer Statistics Review, January 1988. HEALTH Societys Recipe for Health l The debate over the proper recipe for good health must be one of mankinds oldest. Although health and life expectancy depend partially on hereditary factors, many aspects of health can be influenced or controlled by individual or social actions. These include the following: l Time and energy spent in promoting ones own health or avoiding illness, including the time needed to remain informed about appropriate health strategies and any loss of amenity associated with changes in lifestyle required to promote good health (e.g., avoiding fast driving, smoking, drinking, and inappropriate diets). Disease prevention through vaccination and

PAGE 101

95 Table 3-3.Consumption Scenarios for Food (billions of 1983 dollars) 2005 1983 Trend 3% a Trend 1.5%a ALT 3%bc c ALT 1.5% c Food and alcohol At home . . . . . . . . . . .....298.4 433.0 402.0 387.0 321.0 Away from home . . . . . ..................123.6 235.0 146.8 259.0 164.0 Tobacco . . . . . . . . . . . . 28.0 31.2 35.5 22.0 25.0 Total Food in personal consumption expenditures ... ... ...450.O 699.2 584.3 668.0 510.0 Government purchases of food . . . . . . . 1.7 3.3 2.4 3.3 2.4 Liberal Moderate Single member family 34.6 28.1 2 member family . . 60.8 50.3 4 member family . ... 90.7 80.3 The resulting totals are then calculated on an annual basis. The annual totals, which are based on the assumption that all meals are taken at home, are then adjusted to take into account meals outside the home. It is assumed that in the moderate cost plan, one meal in five is taken out of the home. One-fifth of the estimated Food expenditure is therefore deducted from the total and doubled in value (approximately 20% of all Food and 40/0 of all spending for food was away from home in 1983-see text). In the liberal food plan it is assumed that one meal in four is taken away from home. Alcohol is estimated independently, and then divided between food at home and food away from home categories at the 1983 levels (60 and 40A respectively). For 30/. case, (real) per capita expenditure on alcohol is increased by annual average of 1.60/0 (as during 1960-83 period, multiplied by 2005 population). For 1.5%. case, consumption is assumed to be one-half of 30/. ALT. Per capita tobacco consumption in the 3/0 growth case is assumed to decline 22/0, roughly following 1973-83 trends projected to 2005. Consumption in the ALT 1.5% case is reduced by the ratio between low growth and high growth tobacco consumption computed for the trend cases. SOURCE Office of Technology Assessment, 1988. l l l l early detection of disease, by routine testing for such things as high blood pressure and cancer. Public investments in such things as air and water pollution abatement (public spending for pollution control and other health and safety objectives increased from approximately $7 billion in 1972 to a peak of $9.7 billion in 1978, then fell to $8 billion in 1983). 31 Investments in occupational health and safety. Aspects of homeor work-life that contribute to stress. Purchases of treatment to remedy problems once they have occurred. In principle, it should be possible to compare the relative benefits of investments in each of these categories. 32 Given current state of knowledge about ~lAndrew Maflin, public Health and Safety, contract repOIl prepared for the Office of Technology Assessment, 1985. ~zA formal ~heoW of the production function of health is suggested in several sources, See, for example, Michael Grossman, The Demand /or Hea/th (New York, NY National Bureau of Economic Research, 1972); H.L. Blum, P)arming for l-/ea/th (New York, NY: Human Sciences Press, 1974); and Jack Hadley, More Medical Care, Better Health? (Washington, DC: Urban Institute Press, 1982). the determinants of health, this proves impossible in practice. The situation is made all the more confusing by the rules governing health spending. Only the patient has a clear incentive to optimize spending to maximize real health, yet the patient may not be well informed of alternatives, and may not be in control of the choices made in her name. The economics of health involve a complex combination of public funding, public regulation, and private decisionmaking. Health markets differ from ordinary markets in several important ways: 1. The flow of information between provider and client in the market for health services is unlike that in other markets, depending more heavily on trust among the clients and professional integrity among the providers. Furthermore, the parties are frequently on an unequal footing. 33 There are even cases where it may jsFrances H, Mi]ler, secondary Income From Recommended Treatment: Should Fiduciary Principles Constrain Physician Behavior, in Bradford H. Gray (cd.), The New Hea/(h Care for Profit (Washington, DC: National Academy Press, 1983),

PAGE 102

be ethically (and legally) possible to withhold information from the individual being treated. 2. Many kinds of competitive behavior that would be celebrated in ordinary commercial competition cannot be permitted in the practice of medicine. While it is difficult to raise complaints about an enterprise that manages to sell individuals more shoes than they need (indeed, much of the U.S. economy depends on consumption that goes beyond need), there is general consensus that it is unethical to encourage a patient to purchase more medical treatment than is needed. The medical community now accepts the patenting of drugs, devices, and techniques, 34 but controversy continues over whether it is appropriate for physicians to sell drugs and other products directly to patients. 3. Perhaps most importantly, health care involves a complex mixture of personal, corporate, and government decisions. In 1950, two-thirds of medical spending came directly from patients. By 1983, 73 percent of all medical spending came from insurance companies or government programs such as Medicare and Medicaid. 35 The Record American spending on health has grown rapidly since the 1970s, because of the availability of new and increasingly expensive technology, public programs providing the elderly and the poor with greater access to health care, the difficulty of increasing productivity in a business dominated by personnel costs, and a variety of other factors. Health care costs have risen at an average rate of 1.75 percent faster than all personal expenditure costs since the 1950s, and in spite of efforts to contain costs, health care costs increased more than 2.5 percent a year between 1980 and 1986. 36 These statistics must be treated with considerable caution since our ability to meas~Robert M. Veatch, EthiCd Dilemma of For-Profit Enterprise in Health Care, B.H. Gray, op. cit., footnote 33. MU.S. Department of Health and Human Services, ~ea~(h, United States: 1985 (Washington, DC: U.S. Government Printing Office, December 1985). sGMore precisely, the ratio of the deflator for medical care tO the deflator for all personal consumption expenditures increased by 1.75 percent per year from 1955 to 1986, based on a regression of data pro vialed in U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, Survey O( Current Business, table ?.10. ure inflation in health care in a period of rapid change is plainly limited. For example, an appendectomy conducted in 1986 may appear identical to one performed in 1972, but advances in medical practice are likely to make the 1986 procedure safer. While the relative role of the different factors contributing to health costs proves difficult to disentangle (a modest effort is made inch. 6), the result has been a spectacular growth in per capita health care spending. Figure 3-2 shows changes in per capita spending for health care. Real per capita spending (that is spending adjusted for inflation) increased an average of 5 percent per year between 1972 and 1986. The growth in spending is much higher than would be expected by the aging of the U.S. population. Explaining the links between spending and expected outcomes in U.S. health care is a treacherous undertaking. It proves virtually impossible to use available statistics to disentangle decisions made in response to competent medical advice (e.g., decisions to smoke, drive without seat belts, and engage in lifestyles likely to decrease health status), decisions made in ignorance of the proper way to use the medical enterprise, and decisions made in ignorance about the health implications of behavior. The problem is made all the more difficult by the extent to which a unique set of regulations strongly affects the way consumer desires for good health translate into market incentives in the business community. CirFigure 3-2.-Per Capita Spending for Medical Care Constant 1982$ per capita (thousands) t 1955 1960 1965 1970 1975 1980 1985 SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, Nation. al Income and Product Accounts, historical diskettes, table 2.5.

PAGE 103

97 cumstantial evidence suggests, however, that the performance of the U.S. health care system could be dramatically improved without any changes in medical technology. Recent Trends While it is difficult to determine which factor deserves credit, the health status of the average American has clearly improved during the past few decades. An individual born in the United States in 1984 could expect to live 74.7 years.5 years longer than an individual born in 1950. Someone aged 65 in 1950 could expect to live an additional 13.9 years, while a person reaching 65 in 1984 could expect to live an additional 16.8 years. 37 Infant mortality rates fell from 29.2 per thousand in 1950 38 to 11.2 per thousand in 1983 (rates of improvement have fallen somewhat since 1981). Average improvements in life expectancy have not eliminated the striking difference between life expectancies of different American social groups. Age adjusted death rates for black males, for example, were nearly 50 percent higher than those of white males in 1983. 39 As figure 3-3 indicates, improved rates of infant mortality may be the most important factor behind increases in average life expectancies; the figure suggests that a decline in deaths at an early age has played a comparatively larger role in lengthening average U.S. life spans than has the fact that older people are living longer. The factors leading to the improvements in health status are extremely difficult to disentangle. It does appear that as a result both of improved clinical techniques and increased funding of health care for the poor, death from diseases where medical treatment can save lives declined sharply between 1968 and 1980. Deaths from childbirth declined 72 percent, from influenza and pneumonia 53 percent, from tuberculosis 52 percent, and from diabetes 31 perWJ.s. f)epartrnent of Health and Human Services, ~ea]th, Un;ted States, 1984 (Washington, DC: U.S. Government Printing Office, December 1984). 3SU.S Department of Health, Education, and Welfare, Public Health Service, Health Status of Minorities and Low-income Groups (Washington, DC: U.S. Government Printing Office, 1979). 39uS. Center for Hea]th statistics, Vita/ ,$tatjstjcs of (he united States, cited in U.S. Bureau of the Census, Statistical Abstract of the United States: 1987 (107th cd.), Washington, DC, 1986. Figure 3-3.-Changes in Patterns of Mortality 80 60 1900 ~ 40200 1 I 0 2 10 20 30 40 50 60 70 80 90 100110 120 Age (In years) How to Read This Figure: Given the mortality rates of 1900, 54/0 of all people would reach age 60. Using mortality rates of 1950, 75/0 of all people reach age 60. SOURCE: U.S. Department of Health and Human Services, Social Security Administration, Life Tables for the United States, 1900-2050, Actuarial Study No. 89, Baltimore, MD, December 1983. cent. 40 Death rates from heart disease and stroke have been declining since 1940. Between 1970 and 1983, age-adjusted death rates from heart disease fell 26% while death rates from stroke were cut in half. There is some evidence that this fall is associated with a decline in cigarette smoking, high serum cholesterol, elevated blood pressure, and lack of exercise. Improvements in medical care have also played an important role. 41 It proves to be nearly impossible, however, to allocate credit for reduced heart disease directly .42 In contrast, death rates from the other major killer, cancer, have risen slightly, growing from 157 deaths per 100,000 in 1950 to 171 per 100,000 in 1985. This is largely because of increasing incidence of lung cancer resulting from smoking. Death rates from stom401-J.s, Center for Health Statistics, unpublished data for 1982, cited in D.E. Rogers, R.J. Blendon, and T.W. Moloney, Who Needs Medicaid? New Eng/and ./ourna/ of Medicine, vol. 307, No. 1, July 1, 1982, pp. 13-18. 41U.S. f)epartment of Health, Education, and Welfare, Public Health Service, Healthy People: The Surgeon Generals Report on Health Promotion and Disease Prevention, DHEW (PHS) Publication No. 79-55071, 1979; U.S. Department of Health, Education, and Welfare, Public Health Service, Proceedings of the Conference on the Decline in Coronary Heart Disease Mortality, National Heart, Lung, and Blood institute, National Institutes of Health, NIH Publication No. 79-1610, 1979. 4z.s, pelt and W.E. Fayerweather, Trends in the Incidence of Myocardial Infarction and in Associated Mortality and Morbidity in a Large Employed Population, 1957-1983, New Eng/and Journa/ of Medicine, vol. 312, No. 16, Apr. 18, 1985, pp. 1005-1011.

PAGE 104

9 8 ach, uterine, and some other cancers have declined slightly during the past 20 years. Deaths from all cancers except lung cancer fell from 144 per 100,000 in 1950 to 125 per 100,000 in 1985. 43 Risk factors in the development of other cancers include: alcohol, smoking, certain viral infections, and exposure to cancer-causing substances in the workplace, in the general environment, or in food and drugs. Research is underway to determine whether high fat/low fiber diets also contribute. Heredity may also play a role. While remarkable advances in treatment for certain types of cancer have occurred, it is not clear that the increase in survival for the more common cancers is due to improved medical care. 44 As life expectancy rates increase, other health indicators appear to be changing as well. Since 1973, the percentage of the population reporting some limitation of activity has remained essentially the same, while the average number of restricted-activity days per person per year increased from 14.6 days in 1970 to a peak of 19.1 in 1980 and 1981 and fell to 14.5 days in 1983. Blacks have more restricted-activity than whites (16.6 v. 14.3 in 1983). Bed-disability days remained roughly constant for people under 65 (5.3 days/year in 1970 v. 5.4 days/year in 1983) but rose steadily for those over 65 (13.8 days/year v. 16.7 days/year in 1983). 45 The changes (particularly the reversal of reported disability since 1981) are difficult to interpret. The number of days each American is likely to spend in poor health is in part symptomatic of improved medical technology, such as the development of diagnostic techniques resulting in earlier awareness of disease. Changing public attitudes toward disability, improved social supports for the disabled, and changes in employer-granted sick leave may also increase the reporting of disability. The changing figures may also be a result of medical progress, which is able to save 4SU,S. Department of Health and Human Services, National Cancer Institute, Annual Cancer Statistics Review, summary, Washington, DC, January 1988. up M Boffey, cancer progress : Are the Statistics Telling the Truth? The New York Times, Sept. 18, 1984; A.R. Feinstein, D.M. Sosin, and C.K. Wells, The Will Rogers Phenomenon: Stage Migration and New Diagnostic Techniques as a Source of Misleading Statistics for Survival in Cancer, New England Journal of Medicine, vol. 312, No. 25, June 20, 1985, pp. 1604-1608. 45us. Center for Health Statistics, Vita/ and Hea/(h Sta/jstjcS, Series 10, cited in Statktka/ Abstract of the United States: 1987, op. cit., footnote 39, Table 160. lives but not necessarily prevent lengthy periods of illness. 4G Factors Influencing American Health The United States spends as much or more for health care per person than any nation in the world, yet U.S. life expectancy rates are lower and infant mortality rates higher than they are in many other advanced economies. Moreover, different States within the United States exhibit a range of life expectancies that have little obvious relation to spending on health care. Average male life expectancy in Hawaii, for example, was higher than that recorded by any nation in the world in 1981, while life expectancy in Louisiana was lower than that in Cuba, Singapore, and Bulgaria. 47 The life expectancy of Seventh Day Adventists living in California is more than 11 years longer than the U.S. average. U.S. life expectancy could increase by 0.7 years if rates of mortality in 1984 were equal to that of the white population in the United States, and 5.1 years longer if average U.S. mortality rates were raised to that of the healthiest geographic regions in the United States. This exceeds the 4.4 year gain in life expectancy achievable by eliminating cancer. 48 Achieving these gains would require both changes in access to high quality medical care and changes in the social conditions and lifestyles that may create higher incidence of disease in the least healthy population groups. While there are many unanswered questions, differences in health status are obviously linked to the following factors: l l l l l hereditary factors; diet, driving habits, stress, and other aspects of lifestyle; preventive medicine; access to health care for treatment of illness and appropriate use of health care; and the quality of the care delivered, including technical advances. WL M Verbrugge, Longer Life but Worsening Health?: Trends in Health and Mortality of Middle-aged and Older Persons, Milbank Memorial Fund Quarterly, Health and%ciety, vol. 62, 1984, pp. 475-519. tTThe Japanese, one of the worlds longest-lived peoples, take far less interest in smoking cessation than does the United States. og~ U S Congres, Office of Technology Assessment, Health, sector study, Washington, DC, 1987.

PAGE 105

99 Social class, measured by levels of education, income, or occupation, affect many of these variables and correlate strongly with mortality and morbidity rates. 49 Even psychological illnesses, like depression and schizophrenia, are strongly linked to social class. 50 It is, of course, often difficult to distinguish cause from effect when examining correlations, since poor health is often a direct cause of low income. Education seems to be strongly correlated with good health if only because education is correlated with income. 5 1 Education appears to be connected with knowledge about health effects of lifestyles and the way health care facilities can be used. Figure 34, for example, shows the strong connections between use of seat belts and education. Similarly, the mortality rates of well educated white males aged 25-63 are significantly lower than those in the same group with comparatively poor educations. 52 Women with low levels of education have a much higher number of babies with low birthweights. 53 WE M Kitegawa and P.M. Hauser, Differential Mortality In the united States (Cambridge, MA: Harvard University Press, 1973). S5L, S yme and L,F, Berkman, Social Class, Susceptibility and Sickness American Journal of Epidemiology, vol. 104, No. 1, July 1976, pp. 1-8. SIBased on data by region in the 1970 U.S. Census, cited in J. HadIey, op. cit., footnote 32, pp. 59-61. 5WM. Kitegawa and P.M. Hauser, op. cit., fOOtnOte 49. Sj[nstitute of Medicine, Preventing L OW Birth weight (Washington DC: National Academy Press, 1985). See also U.S. Congress, Office of Technology Assessment, Neonatal Intensive Care for Low Birthweight infants: Costs and Effectiveness, Health Technology Case Study 38 (Washington, DC: Government Printing Office, December 1987). Figure 3-4.-Seat Belt Use in 1983 No H S Degree H S Graduate College Graduate Graduate School under $10,000 $10-20,000 $20-30,000 $30-40,000 over $40,000 Strong social ties, measured by numbers of social contacts, church membership, and membership in social organizations, appear to be correlated with low mortality rates. 54 Evidence of the strong correlations between behavior and health is also evidenced by the fact that marriage is strongly correlated with longer life expectancy. Men appear to benefit more from marriage than women. 55 Hereditary Factors. Hereditary factors may explain some of the differences in health status of nations and States that appear to have access to identical medical technology. Japanese living in America assume most American patterns of disease, and black Americans have patterns of disease that more closely resemble those of Caucasians than those of West Africa. sG In fact, Japanese living in California considered acculturated to American lifestyles had coronary heart diseased 2.5 to 5 times higher than a group that lived in California but had not adopted Western lifestyles. 57 Environmental and Lifestyle Factors.-The potential benefits of some measures to prevent disease and promote good health are well established. Smoking, failure to wear seat belts, alcohol abuse, improper diet, dangerous working conditions, stress

PAGE 106

100 resulting from unemployment or other causes, exposure to air or water pollution, and a variety of other factors contribute to poor health status. There is reason to believe that most of the improvements in life expectancy achieved between 1900 and 1970 resulted from improved income and living conditions rather than improvements in clinical medicine. 58 It is estimated, for example, that 25 to 40 percent of cancer deaths result from smoking and 10 to 70 percent from diet. sg A significant part of the credit for the reduction in mortality due to cardiovascular disease during the 1970s resulted from changes in diet, exercise habits, and other factors independent of clinical medicine. Control of hypertension through therapeutic pharmaceuticals obviously played a major role in addition to advances in heart surgery and other spectacular therapeutic improvements that occurred during the same period. 60 Infant mortality and birth defects are strongly linked to prenatal care for the mothers. Two-thirds of all infant deaths occur when the baby is born weighing less than 5.5 pounds. G1 Poor nutrition, smoking, exposure to toxic substances, alcoholism, and a variety of other conditions contribute to low birthweights and a variety of medical conditions before and after birth. G2 Accidents, which account for 6.5 percent of all deaths and are the leading cause of death for those below the age of 44, provide a further illustration of lifestyle effects. Over half of all accidental deaths are due to motor vehicle accidents, and alcohol abuse is the predominant risk factor in these accidents. 63 Other factors influencing the accidental death rate WJ. B fyfcKeoWin and S.M. McKinlay, The Questionable Contribution of Medical Measure to the Decline of Mortality in the United States in the Twentieth Century, ffealth and Society, summer 1977, pp. 405428; T. McKeown, The Role of Medicine: Dream, Mirage or Nemesis (London: Nuffield Provincial Hospital Trust, 1976) SgDo]] and peto, op. cit., footnote 56, P 1256. 60R.F Gil]um, et a]., sudden Death and Acute Myocardial Infraction in a Metropolitan Area, 1970-1980, New Eng/and Journa/ of Medicine, vol. 309, No. 22, Dec. 1, 1983, pp. 1353-1358; U.S. Department of Health and Human Services, Public Health Service, Smoking and Cardiovascular Disease, Morbidity and Mortality Weekly Report, No. 32, 1984, pp. 677-679; S. Pelt and W.E. Fayerweather, op. cit., footnote 42. GIS. Shapiro, M.c. tvf&Ormkk, M.C. Starfield, et al., Relevance of Correlates of Infant Deaths for Significant Morbidity at 1 Year of Age, American .lournaf of Obstetrics and Gynecology, vol. 136, No. 3, Feb. 1, 1980, pp. 363-373. Gz]nstitute of Medicine, op. cit., foOtnOte 53. GSH.J. Malin, et a]., Alcohol-Related Highway Fatalities Among Young DriversUnited States, Morbidity and Mortality Weekly Report, No. 31, 1982, ~P. 641-644. include road conditions, vehicle safety design, and the use of safety devices such as seat belts, air bags, and child safety seats. 64 Improvements in the speed with which casualties can be treated, the emergence of hospital-based casualty centers, and better rehabilitation programs are also important in reducing mortality and injury rates resulting from accidents. 65 Nearly 60 percent of accidental deaths can be prevented by adequate and timely medical care. The remaining 40 percent, however, occur immediately after the accident. The only way to reduce the mortality rate in these situations is by preventing the accident. 66 U.S. medicine tends to leave the often complex problem of prevention in the hands of individuals, providing only the techniques for fixing problems once they occur. Hurried physicians, for example, may be more likely to prescribe drugs to control high levels of blood cholesterol than to undertake the time-consuming and difficult task of counseling changes in a patients diet. 67 In principle, some of the new mechanisms for financing clinical medicine, such as the self-insuring health maintenance organizations (HMOS), can go part of the way toward reconciling the two kinds of health investment. To date, however, the HMOS do not appear to have taken steps to encourage prevention that are consistently different from those taken by other medical practices. The United States has made enormous investments to improve or at least slow the deterioration of air and water quality. Unfortunately, little is known about total exposure to potentially hazardous substances or substances that may have beneficial effects. It proves extremely difficult to link environmental hazard to specific health effects even when good data are available, because so many variables are involved. Numerous investigations have resulted in few clear correlations between typical exposure to pollutants and health effects. G8 Some studies sugSds.p. Baker, B. ONeil, and R.S. Karpf, The Injury Fact Book (LeXington, MA: Lexington Books, 1984). GsNationa] Research Council and the Institute of Medicine, injury in America (Washington, DC: National Academy Press, 1985). GGDD, Trunkey, rrauma, scientific American, VOI. 249, No. z, August 1983, pp. 28-35. NL.A. Cohen, op. cit., foOtnote 2. ~For a review of these studies, see A.M. Freeman Ill, Air and Water Po//ution Contro/ (New York, NY: J. Wiley & Sons, 1982).

PAGE 107

101 gest that a 1 percent reduction in pollution from stationary source air pollution (pollution from sources other than cars and trucks) results in a 0.1 percent reduction in mortality rates and 0.1 to 0.4 percent reduction in illness. While 1 to 3 percent of the Nations groundwater may be polluted by hazardous waste dumps, pesticide and fertilizer runoff, and other materials, G9 little is known about long-term health effects. Improvements in drinking water at the turn of the century played a decisive role in improving health, and outbreaks of disease are only occasionally traced to poor water supplies today. It is possible, however, that low-level exposure to heavy metals and other contaminants in water has long-term health effects not easily measured. 70 While data on outdoor pollution are reasonably complete, data are extremely poor on exposures received indoors where people spend 90 percent of their time. Indoor concentrations of radon, formaldehyde, tobacco smoke, asbestos, carbon monoxide, and nitrogen dioxide are almost always greater than outdoor concentrations. 7] Pollution from office machines, household chemicals, and a host of other sources usually makes interior air much less safe than external air. Preventive Medicine.-A variety of medical technologies have the potential to reduce mortality and sickness by early detection of disease, or by treating conditions like high blood pressure that are linked with disease. These measures, however, can be costly. There is considerable disagreement about how best to compare the costs and benefits of prevention and cure. 72 The fundamental dilemma is determining how much it is worth to prolong life, and whether different values should be attached to the quality of the life preserved. One method for resolving this issue involves the use of discounted years IWU. S Congress, Office of Technology Assessment, Protect@ the Nation Groundwater from Contamination (Washington, DC: U.S. Government Printing Office, 1984). ToAmerican Medica] Association, Drinking Water and Human Health, Chicago, IL, 1984. TINatjona] Academy of Sciences, Indoor Polutants (Washington, DC: National Academy Press, 1981); H. Levin, Indoor Air Pollution Research and Its Applications in Office Building Development and Operation, The Changing Off;ce Workplace, J. T. Black, et al. (eds.) (Washington, DC: The Urban Land Institute, 1986). 7ZL. B Russel, /S Preventjm Better Than Cure? (Washington, DC: The Brookings Institution, 1986). of healthy life. 73 In this approach, costs include the costs of all preventive measures and any side effects less the savings from treatment costs avoided. The costs of medical care used in survivors longer expected life have sometimes been included but are more controversial. Benefits include discounted years of added life adjusted to reflect the quality of the added years of life. (e.g., Are the years spent in pain or activity severely limited by the methods used to prolong life?) 74 Using such measures, it has been shown that the cost of a discounted year of healthy life saved due to hypertension screening ranges from $3,300 (1975 dollars) for 20-year-old men to $16,300 for 60-yearold men. Costs would average $7,000 per year of life saved if the costs of a national screening program to detect hypertensives were included. 75 A study conducted on the effectiveness of Pap tests recommended one test every 3 years instead of one test every year. 7G Clinical Treatment.A spectacular variety of new and often expensive technologies have reduced suffering and prolonged life in the past few decades. The impact of these technologies, and the way they have reshaped the structure of the Nations health delivery system, are discussed in chapter 6. Technical progress has been so dramatic and changes in management and regulations so rapid that it is difficultoften impossibleeven for experts to provide statistically sound analyses of costs and outcomes. The obvious links between access to care and good health are clouded not only by the difficulty of separating environmental and behavioral factors from the income available for clinical medicine, but also by the facility by which available resources are al73 MC, Weinstein and W.B. Stason, Foundations of Cost Effectiveness Analysis for Health and Medical Practices, New Eng/and .fourna/ of Medicine, vol. 296, No. 13, Mar. 31, 1977, pp. 716-721. T4L.B. Russel, op. cit., footnote 72, pp. 71-72. TS~e M.C. Weinstein and W.B. Stason, Hypertension, A Policy perspective (Cambridge, MA: Harvard University Press, 1976). The hypertension study reveals some of the painful choices that must be made. The discounting technique estimated that a 50-year-old man would be paying $6,900 for each discounted year of life saved ($2,300 if he is lucky enough to have complete success with the treatment). is this too much for society to spend extending the mans life by about four years when these additional years will occur when the individual is no longer in the work force? 76D, M. Eddy, The Economics Of Cancer prevention and Detection: Getting More for Less, Cancer, Mar. 1, 1981, p. 1200.

PAGE 108

102 located and the way they are allocated. An attempt to disentangle data from the 1970 census showed a strong positive return to investments in clinical medicine. A 10-percent increase in medicaid spending per enrollee, for example, was found to decrease mortality by 1.6 percent. 77 Adequate use of the health care system is, of course, not simply a matter of providing adequate funding. There remains the question of how patients choose to use the care. It appears, for example, that 50 percent of patients in the United States do not take prescribed medications in accordance with instructions, and some 20 to 40 percent of recommended immunizations are not obtained. 78 Scheduled appointments for treatments are missed 20 to TTJ. Had]ey, op. cit., footnote 32, PP 5$M1. 713D.L. Sackett, The M~nitude of Compliance and Noncompliance, D.L. Sackett and B.B. Haynes, eds., Compliance with Therapeutic Regimens (Baltimore, MD: Johns Hopkins University, 1976), p. 16. 50 percent of the time. 79 Again, it is not obvious whether these responses reflect informed market choices or other factors. Explanations include ignorance, fear, distraction, and many others factors, in addition to an inability to afford the treatment. Many nations with comparatively advanced economies (Sweden, the United Kingdom, Japan, Canada, and West Germany) have achieved comparable or even better national health status than the United States while devoting a smaller percentage of their GNP to health (see table 3-4). These nations all have more ambitious programs for providing universal health care coverage than the United States. Sweden and Japan, the two countries with the best health records, provide both universal health wJ. p. Kirscht and I.M. Rosenstock, Patients Problems in Following Recommendations of Health Experts, G.C. Stone, et al. (eds.), Health Psychology: A Handbook (San Francisco, CA: Jossey-Bass, 1979), pp. 189-215, cited op. cit., footnote 50. Table 3-4.international Comparisons Us Canada U.K. Sweden Japan 1) Life expectancy at birth (1981): Male. . . . . . . . . . . . . . . . 70.4 70. 8 a 71. 2 73.1 73.8 Female . . . . . . . . . . . . . . 77,8 78. 3 a 77. 2 79.1 79.1 2) Life expectancy at age 65 (1980): Male. . . . . . . . . . . . . . . . 14.1 14. 4 a 13. 1 14,4 14.6 b Female . . . . . . . . . . . . . . 18.3 18. 7 a 17. 2 18.2 17.7b 3) Infant mortality rate (1981) deaths per 1,000 live births . . . . . 11.9 9,6 11.1 6.8 C 7.1 4) Health care spending as percent of gross national product . . . . 10.7 8.4 5.3 10 5.2 5) Population (1982) (in millions). . . . . . . . . . . 232 25 56 8 118 6) Percent of population 65 years and older (1981) . . . . . . 11.2 9.7 15.3 16.5 9.3 7) Crude birth rate per 1,000 population (1982) . . . . . . . 16 15 13 11 13 8) National per capita income (U.S.$) (1981) . . . . . ...........10,094 9,133 8,222 13,146 7,672 9) Physicians per 100,000 population (1980) . . . . . . . . 192 182 154 204 128 10) Hospital beds per 100,000 population (1980) . . . . . . . 629 877 1,136 l,49 2 d 1,06 4 a Data from 1978. b Data from 1961. c Data from 1962. dFigure includes nursing home beds. SOURCES: 1) U.S. data from Health, UrrMd States: 1985 (Washington, DC: U.S. Government Printing Office, December 1965). Canadian data from R. Wilkins and O,B. Adams, Health Expectancy in Canada, Late 1970s: Demographic, Regional, and Social Dimensions, American Journal of Public Health, vol. 73, No. 9, September 1963, p. 1077. Data for U.K. from World Health Statistics Annual, 1983, World Health Organization, Geneva, Switzerland. Data for Japan and Sweden from Hea/tlr, Urrlted States: 1985, op. cit. 2) U.S. data from Health, Urr/ted States: 1985, op. cit. Canadian data from Wildins and Adams, op. cit. Data for Sweden and U.K. from World Health Statistics Annual, 1963, op. cit. Japanese data furnished by Ministry of Health and Welfare, Embassy of Japan, Washington, DC. 3) U.S. data from Health, Un/@d States: 1985, op. cit. Data from Canada, Japan, and Sweden from Demographic yearbook, 1982, United Nations, N.Y. U.K. data from World Health Statistics Annual, 1863, op. cit. 4) 1962 data for Japan from Health and Welfare Statistics in Japan, Health and Welfare Statistics Association, Japan, 1965.1963 data for Sweden from Fact Sheets on Sweden: The Health Care System in Sweden, The Swedish Institute, FS 761Vpb, Sweden, 1963. U.S. 1963 data is from Hea/th, United States: 198S, U.S. DHHS Pub No. 68-1232, 1965. Data from the United Kingdom for 1960 from International Financial Statistics Yearbook, 1961 international Monetary Fund, 1861. Data for Canada (1962) from the Canadian Embassy, Washington, DC, personal communication, April 1966. 5) World Deve/oprnent Repoti, The World Bank, Washington, DC, 1964. 6) World Health Statistics Annual, op. cit., 1963 and 1964. Data for the U.S. is from Sociai Security Area Population Projections, 1963, U.S. Department of Health and Human Sewices, SSA Publication No. 11-11535, 1963. 7) Wortd Development Report, op. cit. 8) Stat/st/cu/ Yearbook, 1981, United Nations, New York, 1983. 9) Estimates from Wor/d Development Report, op. cit. 10) Data from 1960 World Health Organization Statistics reported in World Health Systems: Lessons for the United States, report presented by the Chairman of the Select Committee on Aging, U.S. House of Representatives (Washington, DC: U.S. Government Printing Office, 1964).

PAGE 109

103 insurance. The United States, which provides neither comprehensive health insurance nor guaranteed access, has the lowest health status. (See box 3-A for a discussion of the limits of existing U.S. funding programs.) The United Kingdom and Canada, with comprehensive health insurance but varying degrees of access, have life expectancies and infant mortality rates in the intermediate range. These facts may suggest that health care strategies exist in other countries that are economically preferable to the U.S. model. They may also simply mean that Americans are prepared to sacrifice 5 to 10 years of life in order to enjoy aspects of U.S. culture not enjoyed by other cultures that produce less disease. If nothing else, they suggest that significant changes in the cost and success of the U.S. health care system are possible during the next two decades. Choices and Consequences Estimates of the future performance of the Nations medical system require answers to the following types of questions of health care finance and management: l l l l l l Will systems be developed that provide employers of all kinds a greater incentive to provide employees with health programs that include prevention and adequate care? Will systems be developed that provide access to high-quality health care for those lacking access to well-managed corporate or government health programs? Will there continue to be large differences in the medical benefits received by people in different income groups? Will health-related information available to consumers be improved? Will consumers change their behavior as a result? Can the incentive systems of the health industry be changed to give patients and providers similar interests in the support of cost-effective health care that balances prevention, health promotion, and other forms of health care? Will medical consensus converge at the lower end of the range of health care utilization (implying at least a 40-percent reduction in hospital costs), or will the lower rates prove to be ill advised? Will social policy (and private insurance programs) set standards demanding comparatively large numbers of years saved per dollar spent on health care? Box 3-A.Medical Coverage for the Poor Americas lower income groups are poorly served by the health care system: c Medicaid provides coverage to a decreasing number of the poor. In 1985, Title X1X served only 40 percent of the population with incomes below the poverty line; in 1976, it served 65 percent. ] l Many States have not adjusted Aid to Families with Dependent Children (AFDC) income levels to account for inflation. From 1970 to 1984, State AFDC benefit levels for a family of four, in constant dollars, failed to keep up with inflation in all but two StatesWisconsin and California; across the country, the median decline in benefit levels, adjusted for inflation, was 33 percent. A study by the U.S. General Accounting Office estimated that 493,000 families lost their AFDC coverage as a result of changes brought about by the Omnibus Budget Reconciliation Act. In 1981, 1 million families had at least 1 member who needed medical care during the year but did not receive it for financial reasons. 2 l Among the under-65 population, the percentage that is uninsured grew from 14.4 percent in March 1980 to 16.0 percent in March 1983. Between 1981 and 1982, the number of uninsured adults living with an employed, insured spouse almost doubled. Three-fourths of the 35 to 37 million uninsured are either employed or dependents of employed persons. 3 l As many as 35 million people (17 percent of the population) were uninsured in 1984, including 12 million persons with income below the poverty level who did not qualify for Medicaid. 4 l About one-third of all uninsured individuals are under the age of eighteen. s IJim Sarsgard, Actuary, Division of Medicaid Cost Estimates, Office of the Actuary, Department of Health and Human Serwces, based on Census Bureau Statistics and the Current Population Survey 2M, E, Lewin, Financing Care for the Pm and Underinsured An Overview, in ME. Lewin (cd.), The Health Policy Agenda Some Cnt/ca/ Questions (Washington, DC: American Enterprise Institute, 1985) ~M, E. Lewin and Lawrence Lewin, Financing Charity Care in an Era of Competition, Hea/th Affairs, spring 1987, p. 51 4 Jim Sarsgard, op. cit. 5 Uwe Reinhardt, Health Insurance for the Nations Poor, Health Affairs, spring 1987, p. 101

PAGE 110

104 l Will large medical systems such as integrated hospital systems be able to manage significant savings? Other issues include: Will cures be found for major illnesses such as heart disease, cancer, and senile dementia? If so, will they be cheap (i.e., vaccination), or will they require lengthy and expensive treatment? Will acquired immunodeficiency syndrome (AIDS) or a health menace of equivalent proportions impose new burdens on the system? Will government support of medical education and medical research be vigorous? Will public and private regulation of the quality of health care prove adequate? Considering the Possibilities Any one of the factors listed above could lead to considerable changes in both the cost of national health care and health of the average American. Few of the factors can be predicted simply by looking at trendsmany are matters of public and private choice. Table 3-5 explores some alternatives in quanTable 3-5.Hypothetical Scenarios titative terms. It is based on a 1980 survey linking spending on health care to a patients age and the condition for which treatment was received.80 The increase in spending shown in table 3-5 assumes that the cost of intervention remains unchanged. Changes result entirely from assumptions about the success of the intervention. The basic assumption of the calculations is that a person whose death has been avoided by an assumed improvement in some kind of treatment (e.g., a cure for cancer) would have the same risk of dying of other kinds of disease in later years as the average person of the same sex and age. The results of a calculation like the one displayed in table 3-5 can be difficult to interpret since it includes a variety of offsetting factors. The analysis is not intended to provide a precise forecast, but rather to exhibit the striking range of costs and health out~he National Medical Care Utilization and Expenditure Survey, conducted by the National Center for Health Statistics. Corrections for underreporting were made following the suggestions of a study of this data conducted by J.M Anderson and E. Thorn, Estimates of Aggregate Personal Health Care Expenditures in 1980, ICF, 1985. See D. Gillman, Documentation on the Population Projection Program (PROPOP), paper prepared for the Office of Technology Assessment, 1985. for the U.S. Health Care System Lifetime National Life expectancy disability health care (years at birth) days (1983= 100) costs Year Male Female Male Female (1983= 100) 1983 . . . . . . . . . . . . 70.8 78.0 100 100 100 2005: AIDS epidemic a . . . . . . . . . 70.7 78.0 126 100 140 1983 mortality rates. . . . . . . . . 70.8 78.0 126 100 133 Half infant mortality . . . . . . . . 71.8 79.2 91 89 134 Social Security Administration pessimistic scenario c . 73.1 80.5 125 100 137 No smoking d . . . . . . . . . . 74.3 81.4 108 85 119 Social Security Administration baseline scenario c ., . 75.3 82.9 137 110 142 Cancer cure e . . . . . . . . . . 75.8 83.0 136 104 135 Modest prevention . . . . . . . . 76.0 84.2 105 82 116 Cure heart disease g. . . . . . . . . 77.0 84.0 148 117 144 Extensive prevention . . . . . . . . 78.4 84.5 111 86 115 Social Security Administration optimistic scenario c . 78.6 66.4 156 125 146 optimistic prevention . . . . . . . . 78.7 64.8 91 89 135 aAIDs (acquired immun~eficiency Syndrome) case aaumes no change from 1963 mortality patterns except that death frOm other CaU9eS inrXeSSes 21 0/0. bcuts moflality and morbidity for infant mortality and congenital diseases in half. cMofiality ~sumptions used by the Office of the Actuary, U.S. Social Security Administration in their forecast of U.S. population growth. See U.S. Social Security Administration, Population Projections 1963; Wodd Development Repoti, The World Bank, Washington, DC, 1964. dAss ume s that cigarette smoking is responsible for 150/0 of head diseaae deaths, 32Y0 of cancer deaths, 100/. of deaths from vascular disease, 43~0 Of deaths frOm chronic obstructive lung disease, 160/0 of deaths from diseases of the digestive system, and 16A of deaths from all other causes. ecancer moflality and morbidity falls to zero in lm. fNo Smokingj 50./0 reduction in accident rate from iowered alcohol use and Seat belts and workplace safetY. gReduces heart disease mortality to zero and heart-related illness by 300/. starting in 1990. hsame as (e) except that improved diets are assumed to reduce cancer deaths an additional 67Y0 and deaths frOm digestive diseases 50/0. isame as (h) only cuts mortality and morbidity for diseases of infancy and c0n9enital diseases in half. For more details, see U.S. Congress, Office of Technology Assessment, Health, sector study, 1987; and D. Gillman, Health Cost Forecasting for the U.S., working paper prepared for the Office of Technology Assessment, 1985.

PAGE 111

105 comes possible given a plausible set of assumptions. Total national health care costs hinge on whether comparatively expensive or comparatively inexpensive causes of illness and death are reduced. The costs in the year 2005 depend on the population in that year. This is in part a function of mortality patterns prior to 2005. A comparatively inexpensive magic bullet cure for cancer in the year 2000 (the equivalent of a vaccine or penicillin) would reduce costs in 2005, but would increase them in later years because the individuals living longer as a result of the cure would begin to incur other health care costs. The costs are also extremely sensitive to whether the factors that decrease mortality also lead to a corresponding decline in sickness requiring medical attention. The large range of variation in days of disability translate into differences in national health care costs. Constructing Scenarios The hypotheses just generated can be combined with assumptions about growth in the intensity of treatment, and from these combinations can be developed scenarios of spending on the Health amenity. In all cases it is assumed that the relationship of government purchases to consumer spending remains the same. The results are summarized in table 3-6. In the 3 percent Trend scenario, these assumptions imply that government purchases of Health as a share of all government purchases rise substantially. The rise is less pronounced in the 3 percent Alternative scenario, following the lower cost prevent scenario. However, the share rises 1.5 percent under the Alternative scenario, due to the assumed provision of comprehensive medical care even at low rates of economic growth. The Trend scenarios are based on an assumption that growth in health care spending will follow 1960Table 3-6.Consumption Scenarios for Health (billions of 1983 dollars) 2005 1983 Trend 3% a Trend 1.5%b ALT 3% C ALT 1.5%d Household spending ., 268 650 420 500 418 Government spending 60 147 95 113 94 Total, . 328 797 515 613 512 Percent share of GNP . 9.6 12.2 10.9 9.4 10.9 ABBREVIATIONS: ALT = alternative scenarios, GNP = gross national product aAssumes U S Social Security Administration (SSA) baseline scenario hypotheses illustrated in table 3-5, and Der-caDita intensitv increases at 2.5/0/ yearroughly half the average rate of \he 1972-66 period. bSSA pessimistic scenario from table 3-5, and an intensity growth of 0.60/. Per year. c(Extensive prevention case from table 3-5, and intensity grows at 116 historic rates. df4Modest prevention case from table 3.5, and intensity grows at 1/6 historic rates, SOURCE: Office of Technology Assessment, 1966 83 patterns, and increase 38 percent faster than total personal consumption expenditures (PCE). This implies that PCE Health expenditures will increase by just over 4 percent annually in the 3 percent growth case, and by about 2.1 percent annually in the 1.5 percent growth scenario. 8] The Alternative scenarios are based on the modest and extensive prevention scenarios shown in table 3-5, and on an assumption that health care is more equitably allocated among income groups. It is assumed that the rate of growth in intensity of care (in this case measured in constant dollar spending per treatment) is approximately one-sixth the 1972-86 rates because a better match between health care spending and outcome is achieved. ElThis expenditure trajectory is consistent with projections fOr 1990; Ross H. Arnett Ill, et al., Health Spending Trends in the 1980s: Adjusting to Financial Incentives, Health Care Financing Review, vol. 6, No. 3, spring 1985. HOUSING The Housing Recipe Unlike health, it is difficult to develop tricity, and other fuels. Other includes spending for home furnishings, cookware, water and sewer, and the other products associated statistics with household operation. It does not include products associated with giving an unambiguous measure of housing as an personal business or entertainment such as telephone, stationary, and home electronics. Spending for new housing units is not counted as amenity .82 For most, the American Dream house a personal consumption expenditure in the National Income and Prod-

PAGE 112

106 apparently remains a detached residence, with pleasant grounds, security, and a comfortable interior. A general definition of quality in housing includes minimal maintenance (e.g., plumbing that works, roofs that dont leak, and windows that open), good lighting, and a heating and cooling system that provides desired levels of comfort. Recent discoveries about the quality of indoor air are now making clean and safe air a part of quality in homes, as well. A good location is also an important, if somewhat ambiguous concept. Definitions of desirable locations include assets such as physical attractiveness, safety, and good public schools; and convenient access to jobs, shopping, and recreational areas. This is becoming an increasingly challenging problem as both husbands and wives commute to work. Changes in tastes and demographics have a powerful effect on demand for housing. For example, the growth of the elderly population and increases in home health care place new demands on housing for people with mobility limitations. Nearly 40 percent of all persons aged 65 to 74 have some activity limitation, and 63.2 percent of individuals over the age of 74 have some limitation.83 More than five percent of the total population are blind or visually handicapped, 7.4 percent are deaf or hearing impaired, 3.2 percent have some form of lower extremity impairment, and 1.2 percent have upper extremity impairment. 84 The growth of single-person households and the other household categories translates into complex patterns of demand for housing. The decline in leisure time caused by an increase in work-related activities translates not only into demands for shorter, more intense vacations (see discussion of recreation and leisure later in this chapter), but also into greater demand for recreational amenities within a house.85 New technologies are redefining the definition of interior amenity. Home entertainment involves an increasingly elaborate and powerful array of video and sound equipment, for example. Home offices are becoming common for professionals and there has been SW.S congress, Office of Technology Assessment, Technoioa afld Aging in America, OTA-BA-264 (Washington, DC: U.S. Government Printing Office, June 1985), pp. 291-292. ~U.S Congress, office of Technology Assessment, TeChr?Oio~ and Handicapped Peopie, OTA-H-179 (Washington, DC: U.S. Government Printing Office, May 1982), p. 22. SSC. Vogel, clustered for Leisure: The Changing Home, New York Times Magazine, June 28, 1987, p. 13. a small (but limited) increase in home-based work. Broadband communications wiring in houses is likely to be as common as telephone wiring today. The continuing mobility of Americans means that for many people housing is not a permanent investment. 8G Many make purchases with no intention of sinking deep roots in a community. In 1985, only 58 percent of Americans lived in their 1980 residence. Movers were about equally divided between those moving to a new house in the same county, many of whom were undoubtedly trading up to improved residences,87 and those moving out of the county. While two-thirds of the residents of the northeast remained in their 1980 homes, in the west movers outnumbered nonmovers. Nearly one-quarter of the people living in the west moved outside their county and 15 percent moved from another State or another country. Two-thirds of all young adults (age 20 to 34) moved during the first half of the 1980s; nearly one-third moved out of the county. While many elderly persons may move from their residences to retirement communities, 84 percent did not move at all and only 7 percent moved out of the county where they lived in 1980 (see table 3-7).88 SGSee the last section of ch. 5 for a discussion of the ways changing economic structure are moving jobs around urban areas and among regions. INH.J. Brown and J. Yinger, Horneownership and Housing Alfordabi/itYin the United States: 1963-1985 (Cambridge, MA: Joint Center for Ho&ing Studies, 1986). U.S. Bureau of Census, Current Population Survey, Series P-20, 1985. Table 3-7.Mobility of the U.S. Population Asked in 1985 where they lived in 1980, the responses were as follows: Now live in Lived in a the same unit different unit Lived in another they occupied in same county county or abroad in 1980 in 1980 in 1980 Total . . 58.3 22.1 19.6 By age: Age 5-19 . 57.0 24.1 18.9 Age 20-34 .., 35.3 33.5 31.2 Age 35-64 ....., 68.9 16.2 15.0 Age 65+ . 83.5 9.2 7.3 By 1985 location: Northeast . 66.9 18.8 14.3 Midwest . 60.9 22.9 16.2 South . . 56.0 21.5 22.5 West . . 49.6 26.0 24.4 NOTE: Does not include members of the armed forces except those living off post or with their families on post. SOURCE: U.S. Department of Commerce, Bureau of Census, CPS Series P-20, 1985.

PAGE 113

107 While much of the discussion in this chapter focuses on the ability of new production technology to tailor products to niche markets and changing tastes, housing will be an exception. The mobility of the U.S. population means that housing is seldom tailored to the tastes of individuals but rather to a homogeneous resale market. The fact that homeowning has been a lucrative form of savings for many makes it difficult to link spending patterns with levels of amenity achieved. Programs designed to make housing affordable to the middle class have had the effect of making it an attractive investment. Housing investments are unique because the owner is free from tax on the effective income enjoyed by renting a property to himself. Many households are, as a result, overhoused, in the sense that their spending on housing exceeds the spending that would have occurred in the absence of this investment incentive. 89 Equity in homes represents 64 percent of the median net worth of homeowners today. go Since 1950 there has been a significant increase in the share of spending for housing that goes to financial institutions, landlords, and insurance companies (see figure 3-5). In contrast, the fraction that goes to an increase in net equity in housing (the WU,S, Congress, Congressional Research Service, Housing programs Affecting the Elderly: A History and Alternatives for the Future, Report No. 82-1 19E, Washington, DC, June 1982. 90U,S, Bureau of the Census, Current population Reports, Series P70. No. 7. Figure 3-5.-Spending on Housing by Major Category Percent of all spending on housing 50 1 I /. / .-. / 40 ~.. -----------. -, ---------..1 304 I 20 I 10 I 1 I 1 I I T 1 1 I 1 1 1 1 I 1 1 1 1 1 1 1 1 I 1 1 1 T 1955 1960 1965 1970 1975 1980 1985 Energ y --Rent 8 Interest New Constructio n Othe r SOURCE: US. Department of Commerce, Bureau of Economic Analysis, ( Nation. al Income and Product Accounts, historical diskettes, table 2.5. new construction portion shown in figure 3-5), has fallen steadily since 1950, and has become much more cyclical. The increase in other housing costs has been so great that they make the dramatic increase in energy prices during the 1970s all but invisible. Lower prices for home furnishings (other in the figure), and perhaps some saturation in demand for such goods, have resulted in a steady but slow decline in the fraction of spending devoted to this area. A variety of costs are not directly reflected in the statistics of figure 3-5. For example, residential electricity and gas customers often do not pay their real share of costs because regulatory commissions elect to subsidize residences by charging higher rates to commercial and industrial customers. Few rates reflect the marginal cost of producing energy from new sources. Water and sewerage lead to enormous costs which are typically hidden because much of the cost is provided through the tax base rather than through direct user charges. A survey in Irvine, CA, conducted by its funding task force, found that the cost of providing streets, parks, schools, flood control, civic and performing arts buildings, and libraries was between $16,500 and $23,800 per dwelling. Trends in U.S. Housing Quality and Affordability There is reason to believe that on average American housing has improved in virtually all of the areas mentioned. The average housing unit is larger than it was a decade ago even though household sizes are smaller. g] The number of units with more than one person per room fell from 3.8 percent of all units to 2.6 percent. 92 There has been a significant increase in the number of units equipped with such amenities as air conditioning, garages, and full city water and plumbing, 93 Correspondingly, a smaller glln I g68, the median size of new privately owned singlefamily homes completed in the United States was 1,385 square feet. In 1980, it was 1,595 sq. ft., and in 1985, it was 1,785 sq. ft. See Statistical Abstract of the United States 1987, op. cit., footnote 39, p. 706. gz[rendia lrby, Housing Problems in the United States, U.S. Department of Housing and Urban Development, Division of Housing and Demographic Analysis (Washington, DC: U.S. Government Printing Office, June 1985). gzstatjstlcaj Abstract of the United States 1987, op. cit., foOtnOte 39, p, 710.

PAGE 114

108 gqBetween ]975 and 1985, the percentage of all units defined to be inadequate (e.g., they lacked or shared some or all plumbing, lacked or shared some or all kitchen facilities, were inadequately maintained, or had public hall deficiencies, inadequate heating equipment, electrical defects, or inadequate or broken sewer) fell from 10.6 percent to 8.9 percent. Those judged to be severely inadequate fell from 4.3 percent to 2.6 percent. See Irendia Irby, op. cit., footnote 92. gsstat;st;cal Abstract of the United States 1987, op. cit., footnote 39, p. 451. Figure 3-6.-Cost of Home Purchasing for New Homeowners and Families Who Purchased a Home in 1968 Percent of median Income 0 4 Purchased a Home In the Year Indicated Below 035 03 025 0.20.15Purchased a Home in 1968 I I 1 0 1 I r I I 1 I { 1971 1973 1975 1977 1979 1981 1983 1985 How To Read This Figure: In 1985, a family earning the median income that purchased a home in 1968 would be paying about 13 percent of its income for housing (limited hereto mortgage payments, maintenance, and energy). The same family would be paying about 33/0 of its income for housing if it purchased a house in 1985. Some of the difference results from closing and costs associated with new home purchasing. Onethird of these costs are assumed to be paid in the first year of home ownership, SOURCE: H. James Brown and John Yinger, Home Ownership and Housing Affordability in the United States: 1983-1985, report for the Joint Center for Housing Studies of the Massachusetts Institute of Technology and Harvard University, Cambridge, MA, 1988. graphic groups. Families with only a single wage earner face the most pressing difficulties except where the single earner is in a high-income occupation (e.g., a manager or professional; see table 38). A new home is far beyond the reach of most female-headed households. Table 3.8.After Tax Cash Costs of a New Home Purchased in 1985 (as a percent of annual income a ) SOURCES: Housing costs taken from Home Ownership and Housing Afforda. bility in the United States: 1983-1985, Joint Center for Housing Studies of the Massachusetts Institute of Technology and Harvard University, Cambridge, MA, 1988. Income data from U.S. Department of Labor, Bureau of Labor Statistics, Emp/oyrnent and Earnings, 1988,

PAGE 115

109 How can the fraction of income spent on housing remain constant (again see figure 3-5) while housing stretches the resources of the median household? How can the increase in the size of new housing units be consistent with the shrinking fraction of the population able to afford housing? The answer is a growing gap between haves and have nets in housing. Figure 3-6 compares the cost burden of a new homebuyer with the price paid by a household still living in a home purchased in 1968. Housing actually consumes a declining share of the income of the 1968 purchaser in spite of the rapid increase in energy costs that has occurred since 1968. Accounting for Changing Costs Three factors account for the sharp rise in housing costs documented in figure 3-6: 1. increasing land costs near areas of employment growth (spurred in part by restrictive zoning and other local policies for limiting growth),96 2. increased cost of financing for housing (resulting directly from deregulation in thrift institutions), and 3. rising energy costs. While there has been some increase in the cost of construction materials and labor, and some decline in the productivity of labor in residential construction, these do not appear to have had a dominant effect on overall housing costs. Between 1970 and 1980, the cost of a new single-family home actually fell 16 percent in constant dollars while the cost of the materials and labor in the home fell 30 to 40 percent. During the same period the cost of financing increased nearly 80 percent.97Given that many developers make most of their money from financing and land development, it is easy to see how they are likely to add amenities and a few hundred square feet to a property if this will give them a competitive edge. Between 1975 and 1985, the real cost of a 10,000 square foot lot of improved land increased in price by 66 percent; the price of an unimproved lot increased by 78 percent. Prices were particularly high R. Babcock, The Zoning Game (Madison, Wl: The University of Wisconsin, 1964). 97 Report of the Presidents Commission on Housing, Washington, DC, 1982. Figures cited converted to constant dollars using the consumer price index. in areas of high growth like California and Florida. While a 10,000 square foot lot cost $2,500 in Chattanooga in 1980, it cost $187,000 in San Jose, CA.98 Some of this was the result of scarcity in growing areas, some from speculation based on high inflation rates, and some resulted from extensive growth in restrictive zoning.99 Several parts of the San Francisco Bay area have instituted regulations that have increased housing prices by at least 20 to 30 percent. 100 Many other cities have taken measures to limit growth, including Montgomery County, MD and Fairfax County, VA in the Washington, DC area; and Fairfield County, CT, and White Plains, NY in the New York City area. Fairfield County and many other areas have adopted ordinances designed to limit both residential and commercial construction. Zoning in over one-third of 75 municipalities covered in a 1976 New Jersey survey required minimum lot sizes greater than 1 acre. Most communities squeeze manufactured housing into disadvantaged locations; many do not permit such units at all. ]o] Apart from driving up property values, one curious feature of these policies is to decrease population density. This results from the fact that housing density remains fixed in heavily regulated areas while household size declines.102 The rapid increase in financing costs was due to a combination of general inflation with regulatory changes in the banking industry undertaken when short-term rates rose rapidly during the 1970s. investors turned increasingly to money market and other instruments instead of traditional thrift institutions. In effect, homebuyers are competing for gsResjdentja/ Land Price Inflation Survey (Washington, DC: The urban Land Institute, 1986). ~Robert Cewero, Jobs-Housing Imbalances in Suburban Employment Markets: An Empirical Investigation, Department of City and Regional Planning, University of California at Berkeley, May 1987. IOOD, Dowa]], The Suburban Squeeze (Berkeley, CA: University Of California, 1985). IOISWs. Seidel, f-fousjn~ Costs and Regulations: Confronting the Regulatory Maze (New Brunswick, NJ: Center for Urban Policy Research, 1978), p. 174; and the (Report of the Presidents Commission on Housing, op. cit., footnote 97. lmM, Gellen, Accessory Apartments in Sing&amijy Housing (piSCataway, NJ: Center for Urban Policy Research, 1985); D. Schoup, Curb Parking as a Commons Prob/em (Los Angeles, CA: University of California, Graduate School of Architecture and Urban Planning, 1983).

PAGE 116

110 money with other borrowers throughout the economy. 03 The 1986 changes in tax law affect housing costs in complex ways.104 The number of new rental units is likely to continue the decline that began before the law was changed, partially offset by non-new units converted from nonresidential use. Reacting to Higher Costs Buyers reacted to the rising cost of housing in a variety of ways. Lending rules were stretched to permit buyers to take higher risks. But there is plainly a limit. The increased cost of housing has led to a decline in ownership rates, particularly for younger families most likely to need new housing (see table 3-9). The need for mobility on the part of younger individuals also contributes to their housing dilemma. Declines in home ownership are naturally matched by increases in the number of individuals renting. In 1987,55 percent of young adults (aged 25 to 34) lived in rental unitsup from 48 percent in 1974. In 1983, 62.5 percent of female single parents were in rental units; the figure was 88.7 percent if they had a child younger than 6. Both statistics represent I03A, DOwnes, The Revolution in Real Estate Finance (Washington, DC: The Brookings Institution, 1985). 04 Description of Provisions of the Tax Reform Act of 1986 Relating to Housing, paper prepared by the National Association of Homebuilders, Government Affairs Division, September 1986. Table 3.9.Household Ownership by Age, 1980=87 Percentage owning Increase/decrease their own homes in percent Age 1987 1980-87 Under 25 . . 16.1 .2 25-29 . . . 35.9 .4 30-34 . . . 53.2 .9 35-39 . . . 63.8 .0 40-44 . . . 70.6 .6 45-54 . . . 75.8 1.9 55-64 . . . 80.8 1.5 65-74 . . . 78.1 2.9 Over 75 . . 70.7 2.9 Total . . 64.0 1.6 SOURCE: WC. Apgar, Jr. and H.J. Brown, The Slate of the Nations Housing 1988, Joint Center for Housing Studies of Harvard University, Cambridge, MA, 1988, p. 12. an increase from 1974. Renters avoided the high cost of home purchasing but were still affected by growing housing costs. In 1987, contract rents averaged 25.5 percent of the income of renting families, up from 20.5 percent in 1974. In 1983, 55 percent of households earning less than 50 percent of the median household income paid more than 30 percent (up from 48 percent in 1974). 105 Some renters have accepted a decline in housing quality. While the percentage of rented units with defects has declined, the 19-percent increase in the total number of rented units meant that there was a 7-percent increase in the number of families living in inadequate rented housing. Single adults and individuals between the age of 25 and 35 were most likely to live in inadequate rental units. One index of housing inadequacy indicates that inadequate housing declined from 15.5 to 12.8 percent between 1974 and 1983. 106 Other responses to increased housing costs include living with relatives, forming households of unrelated individuals, or taking in borders. Between 1980 and 1986, the number of total U.S. households increased 9.5 percent. During the same period, the number of households comprised of related subfamilies increased 96 percent, unrelated subfamilies increased 40.3 percent, and unrelated individuals increased 19.2 percent. Single females with children clearly faced the greatest problem finding housing; the number of mother-child families living with relatives increased 173 percent between 1980 and 1986. 107 For those on the bottom of the ladder, housing has become a disaster. The Nations mayors report that demand for emergency shelter increased by an average of 20 percent in 25 major U.S. cities during 1986. ]08 While changes in cost burdens are relatively visible, other changes in housing quality are more difficult to measure. There is compelling evidence that other amenities have been sacrificed in order to keep housing costs acceptable. One traditional cure for high land prices is making a trade-off between com1051, [rby, op. cit., footnote 9Z. IOGH..I. Brown and J. Yinger, op. cit., footnote 87. IOTCurrenl population Reports, op. cit., footnote 88, No. 412. 1The Continued Growth of Hunger, Hopelessness and Poverty in Americas Cities: 1986, The U.S. Conference of Mayors, Washington DC, 1986.

PAGE 117

111 muting time and housing costs in suburban and exurban regions. ]o9 The trade-off has become much more complex as jobs begin to drift toward suburban locations and as husband and wife must compromise on commuting time. Twenty percent of all trips to work are now from suburbs into central cities. The share of people who both live and work in suburbs has risen from 30.5 percent in 1960 to 41.9 percent in 1980. It is 60 percent in greater Boston, Detroit, St. Louis, and Pittsburgh. l10 While middle class families may have paid a price to combine job access and housing amenity, low income groups, particularly blacks, paid an even greater price since their mobility was highly constrained. Ironically, subsidized low income housing may have the effect of anchoring low income groups to declining areas. i ] ] Zoning, housing costs, and multiple-worker families mean that a shrinking fraction of individuals can live in the same community where they work. Only half of the people in the San Francisco area work in the community where they live. There were only 35 jobs for every 100 people housed in bedroom communities like Daly City. San Jose, Fremont, Concord, and Alameda averaged 50 to 67 jobs per 100 residents. In other areas, high property costs make residence virtually impossible for middle income families. As a result, there were 203 jobs for every 100 residents of Palo Alto, 172 for every 100 residents in Santa Clara (Silicon Valley), 150 per 100 for Sunnyvale, and 132 per 100 for Mountainview The effect of sprawl, coupled with the separation of work and housing, has meant that commuting time has increased for many. While there is considIOgFor more on this subject, see W. Alonso, Location and Land Use (Cambridge, MA: Harvard University Press, 1964); J.D. Carroll, The Relation of Home to Work Places and the Spatial Patterns of Cities, Social Forces, vol. 30, No. 1052, pp. 271-282; H.J. Brown, Changes in Workplace and Residential Locations, Journal of the American institute of Planners, No. 41, 1975, pp. 32-39; W.A.V. Clark and J.E. Burt, The Impact of Workplace on Residential Location, Armafs of the Association o/American Geographers, vol. 70, No. 1, Mar. 1980, pp. 59-67; L. Orr, /ncome, Employment and Urban Residential Location (New York, NY: Academic Press, 1975); and J. Quiggley and D. Weinberg, ilntraurban Residential Mobility: A Review and Synthesis, /nterr?aticma/ Regional Science Review, vol. 2, No. 1, Fall 1977, pp. 41-66. I IORobert Cervero, op. Cit., fOOtnO[e 99. I I IJ. D Kasardij and J. Friedrichs, Comparative Demographic Employment Mismatches in U.S. and West German Cities, Research in the Sociology of Work, No, 3, 1985, pp. 1-30. I IZR. Cervero, 1987, op. cit., footnote 99. erable variation, travel distance has increased while travel time has reportedly declined (implying an increase in commuting speed). Curiously, it is the upper income managers who have elected to travel further to achieve their housing amenity (see table 3-10), in spite of the fact that their time is presumably more valuable. Also, since men commute longer distances than women it appears that women may be taking jobs closer to home, possibly sacrificing income in order to combine lives as a homemaker and paid employee. Construction Technology New building technologies (discussed inch. 6) and new materials can reduce the cost and improve the quality of housing. Factory construction can permit reliability and brand-name guarantees for major building components and entire structures because sections are produced consistently under factory conditions. And where markets exist, it should be possible for customers to design houses to their unique needs. 1]3 It may even be possible to construct homes from modular units that permit relatively easy changes of floor plans. This possibility might permit more flexibility in the design of structures that are purchased both as an amenity and for their po113u.s. congress, Office of Technology Assessment, 7echnolo~, Trade and the U.S. Residential Construction Industry-Special Report, OTATET-315 (Washington, DC: U.S. Government Printing Office, September 1986). Table 3-10.Commuting Distances and Times 1983 1969 Men Women Total Total Trip length (miles) ..................11.2 8. 3 9.9 9.7 Trip time (minutes) .................21.6 18,8 20,4 23.1 Average speed (mph) ., ..............31.1 26,5 29, 1 25,2 By occupation: Distance (miles) Time (minutes) Commute Increase from Commute Increase from 1983 1977 to 1983 1983 1977 to 1983 Professionals . 9.8 .9 20,0 1.8 Managers 12.5 2.4 24,2 3.5 Clerks. 9.1 1.0 19.8 0.8 Laborers 11,8 0.3 22,8 1.0 Unskille d 11.0 1,0 21.4 1.1 Services . . 6.7 0.2 16.1 0.0 Tota l 9.9 0.6 20.4 0.6 SOURCE U S Department of Transpodation, Federal Highway Admmistratlon, 1983-1984 Nakmvide Personal TrampoflaOon Sfudy, volume 1, Washington, OC, August 1986. pp 7-6 tO 7-10

PAGE 118

112 tential resale value. A variety of gadgets ranging from microwave ovens to improved security systems are available and may add to the value of housing in the future. New materials, clever use of microelectronic control technologies, and application of competent architectural and engineering in housing design are becoming more common and can reduce energy consumption in homes by factors of two or more. The efficiency of most home appliances can be more than doubled using technology already on the market or under advanced development (see table 3-1 1). Advances in lighting technology permit graceful integration of day lighting and artificial lighting, better color quality and lack of flicker with high frequency fluorescent, and 200to 500-percent reduction in energy use by lighting. 14 Even the window is undergoing radical changes in design. It is possible to develop windows with thermal insulation equivalent to 3 inches of fiberglass.ll s Other designs enviI14s. ~rman, Energy and Lighting, D. Hafemeister, H. Kelly and B. Levi (eds.), EnergySources: Conservation and Renewabks (New York, NY: American Institute of Physics, 1985). I15s. Selkowi&, window Performance and Building Energy Use: some Technical Options for Increasing Energy Efficiency, D. Hafemeister, et al., op. cit., footnote 114. sion windows whose transparency can be altered under direct control, to let solar energy pass in the winter, and to reflect heat in the summer. The combination of enlightened energy pricing policy with new communications technologies can also help. A 200-house experiment in Roswell, Georgia, uses new packet switching technology, along with a small computer in each house, to run air conditioning in a way that minimizes energy costs. The customer simply selects the room temperature desired when electricity rates reach different levels; she may decide, for example, that during the 4 hours a day when prices reach 25 cents/kWh, temperatures should remain at 85 degrees, but during the long periods when electricity costs 2 cents/kWh, temperatures of 65 degrees are preferable. The thermostat adjusts automatically, using signals received through standard telephone lines. The customer is free to override the utility setting at any time, but must pay the going rate for electricity for the privilege. The system saves the utility enough for it to finance the full cost of the system. A number of studies have tried to determine the number of these technologies likely to make economic sense given perfectly rational decisionmaking by consumers, and the probable investments that will be made under markets likely to exist in the Table 3-il.Energy Consumption and Conservation Potential With Residential Appliances (KWh/yr or therms/yr) Fraction of Advanced Primary energy residential 1985 stock 1985 new 1985 best technology use (Q) a total ( 0 /0) UEC b UEC C UEC d for 1990s* Refrigerator . . . . . . . 1,17 7.1 1,500 1,100 750 Freezer . . . . . . . . 0.44 2.7 1,100 800 500 Electric space heating . . . . 1.58 9.6 Central air conditioning. . . . . 1.47 9.0 3,600 2,900 1,800 Room air conditioning . . . . 0,38 2.3 900 750 500 Electric water heating . . . . 1.35 8.2 4,000 3,500 1600 Electric range . . . . . ., . 0.54 3.3 800 750 700 Electric clothes dryer. . . . . 0.45 2.7 1,000 900 800 Lighting . . . . . . . 1.00 6.1 1,000 1,000 650 Electric other. . . . . . . 0.87 5.3 Gas space heating . . . . . 3.36 20.5 730 620 500 Gas water heating . . . . . 0.85 5.2 270 250 200 Gas range . . . . . . . 0.31 1.9 70 50 40 Gas clothes dryer. . . . . . 0.07 0.4 50 40 35 Gas other. . . . . . . . 0.41 2.5 Total . . . . . . . . 14.25 86.8 NOTES: a Applies to 1980. b unit energy consumption per installation in the 1985 housing stock. c unit energy consumption for the typical model sold in 1985. d unit energy consumption for the best available modei sold in 1985 e unit energy consumption possible in new models by the mid-1990's if further cost-effective advances in energy efficiency are made, 200-400 150-250 900-1,200 300-400 1,000-1,500 400-500 250-500 350-500 300-500 100-150 25-30 30-35 SOURCE: Howard Geiler, American Councii for an Energy-Efficient Economy, January 1986.

PAGE 119

113 United States during the next two decades.ll G While techniques differ, all show the potential for reducing energy use in housing by large factors. Studies conducted in 1987 provide an example. A recent study conducted by the State of Michigan indicated that residential electricity use could be cut in half by the year 2005 even accounting for expected growth in the housing stock. 117 An analysis of the technical potential of housing design and equipment available on the U.S. market (and the international market) in 1982 suggested that energy use could be one-fifth of the energy used by an average U.S. house in 1980. ]18 Choices and Consequences Housing costs in the future will hinge critically on the following factors: I ISJ.H. cibbons and W.U. Chandler, Energy. The Conservation Revolution (New York, NY: Plenum Press, 1981); Solar Energy Research Institute, A New Prosperity, Building a Sustainable Energy Future (Andover, MA: Brickhouse, 1981); M.H. Ross and R.H. Williams, Our Energy: Regaining Contro/ (New York, NY: McGraw Hill, 1980); National Audubon Society, The Audubon Energy Plan 1984 (New York, NY: National Audubon Society, 1984); A. Meyer, J. Wright, and A.H. Rosenfeld, Supplying Energy Through Greater Efficiency (Berkeley, CA: University of California Press, 1983); E. Hirst, et al., Energy Efficiency in Bui/dings: Progress& Promise (Washington, DC: American Council for and Energy Efficient Economy, 1986); J. Goldenberg, et al., Ener~ for Development (Washington, DC: World Resources Institute, 1987). ] 17 State of Michigan, Michigan Energy Options Study, 1987. The study defined technical potential to be investments that competed with the short-run marginal costs of existing Michigan electric generation.27 cents per kWhassuming a 3 percent discount rate on incremental capital. The reductions consisted entirely of improved efficiency in air conditioning, lighting, water and space heating equipment, refrigerators and freezers. llscoldenberg, et d., op. cit., footnote 116, p. 59. l l l l l l It policy affecting speculation in real estate (principally tax policy); changes in the geography of economic development, which either encourage geographic concentration (thereby increasing competition for scarce land resources) or result in greater decentralization; policy affecting the cost of mortgage financing; technical improvements in structures that reduce maintenance and operating costs (principally energy costs); policy affecting the housing available for those lacking adequate income; and technical improvements in the construction process that reduce construction costs (improvements that could be made more rapid by increased investment in research). is impossible to adequately reflect all of these ., variables in the scenarios selected. An attempt has been made to show what might happen to aggregate demand if no major changes are made in the way Americans invest in housing, and what might happen if policies succeed in arresting the rapid growth in spending in this area. The Trend scenarios, shown in table 3-12, 119 are based on the extrapolation techniques described in chapter 2, which link spending to demographic factors, prices, and income. Independent estimates were made for household utilities to incorporate improvements in efficiency (see I lgThree components of household expenditures are distinguished: 1) rents, which includes not only rents paid but also imputed rents received by ownemccupiers; 2) household operations, which include such items as insurance, furnishings, and appliances; and 3) utilities, which include expenditures on energy and water. Table 3-12.Consumption Scenarios for Housing (billions of 1983 dollars) 2005 Mortgages & rents a . . . . . . . . . . .. .331 46 5 b 56 5 C 42 9 d Household operations. . . . . . . . . . . 139 307b 221b 30 2 f 2ol f Utilities . . . . . . . . . . . . . 111 16 4 13 8 9 5 8 0 All PCE housing . . . . . . . . . . . 581 1,05 8 82 4 96 2 71 0 Residential structures . . . . . . . . . . 152 22 4 20 2 22 4 20 2 Government . . . . . . . . . . .......14 2 6 1 9 26 19e Total . . . . . . . . . . . . . 747 1,33 6 1,06 3 1,24 0 94 9 Percent share of GNP . . . . . . . . . . 21.9 20.5 22.5 19.0 20.1 ABBREVIATIONS: PCE = personal consumption expenditure, ALT = alternative scenario, GNp = gross national product aDoe~ not include increase in houg~ng equity, Real increases in total U.S. housing appears in the rr)w labeled residential Structures. bc~culated using extrapolative techniques described in ch. 2. CASSumes that th e growth i n mofigages and rents as a fract[on of all consumption can be reversed and the ratio of spending in this CategOfy to tOtal PCE spending falls to 1970 levels. dsame as (c) except that the ratio remains frozen at 1983 ievels. elncludes maintenance se~ices and commodities, tenants insurance, house furnishings, and appliances. fsame as spending i n the Trend cases with an increase for assumed additional spending fOr high efficiency aPPlianCe5 (5W table 3-13). glJtilities include natural gas, electricity, household fuels, and water and sewer. hsee discussion of spending for gross private fixed investment. SOURCE: Office of Technology Assessment, 1987.

PAGE 120

114 table 3-1 3). The Trend projections for energy use are The Alternative scenarios assume rapid adoption of based on a U.S. Department of Energy forecast.1 20 technology in building design and appliance design. IZOu.s. Department of Energy, Nationa/ Energy Policy Pfan projection 2010 (Washington, DC: U.S. Government Printing Office, 1985). Table 3-13.Consumption Scenarios for Household Utilities 2005 1983 Trend 3% Trend 1.5% ALT 3% ALT 1.5% Electricity 51 87 a 69 b 44 e 36 f Natural Gas 29 26 a 26 C 14 e 1lf Other Fuels 18 26 a 18 C 14 e 8 f Water & Sewer. 13 25 d 23 d 25 d 25 d Total ., ., .,111 164 136 97 81 ABBREVIATIONS: ALT = alternative scenario a Based on National Energy Policy Plan Projections to 2010, U.S. Department of Energy (DOE), Washington, DC, 1985. Midrange forecast for the year 2005 increased to reflect the higher GNP growth rates used In the 30/0 Trend scenario. Electricity consumption was increased by 30/. to allow for an assumed increase in the size of housing units and an increase In appliance purchases and fuel switching. bl~ consumption increased in proportion to 2005/1983 increase in consumer units. Electric use Der household assumed to decline by 8A as the result of efficiency improvements (far below optimal levels given availabie technology). Electricity use increases because of an assumed shift from other fuels. See next note. cl~ natural ga9 and other fuel use increased by 2005/1983 increSSe in number of consumer units. No efficiency improvements assumed. 25V0 of other fuel use shifted to electricity. dprojected using CES consumption data (see ch. 2). elge3 consumption SCSIad by growth in consumer units. Efficiency improvements computed separately for units in 2005 that were standing in 1983 and units built between 1983 and 2005. The proportion of new and old units is calculated using the assumption that 1.5% of the 1983 stock is removed annually (following recent trends). For existing units, It is assumed that fuel use per unit can be reduced by 35/0 (primarily through appliance replacement). A combination of improved construction and better appliances is assumed to reduce consumption per new unit by 500/.. fsame a9 (e) except that new units are assumed to have consumption reduced to.3 average 1983 levels and existing units 0.58. It Is further assumed that 25/0 of the other fuel consumption is shifted to electricic demand with a net improvement of 0.75. SOURCE: Office of Technology Assessment, 1987. TRANSPORTATION Americas need for transportation is as diverse as the U.S. population. There can be no objective measl ures of progress in transportation, but most Americans are likely to agree that a successful system is one l l l that: provides as much freedom of movement as l possibleallowing individuals to go where they want, when they want, at the lowest possible cost in time and money; provides mobility to the widest possible range of individualsincluding the young, the elderly, services consumed by households; increases the number of attractive areas for locating homes and businesses, thereby reducing pressures to increase the price of scarce land while increasing access to needed outdoor recreation areas; and has few unattractive externalities, such as making the Nation vulnerable to foreign oil suppliers, degrading air quality, contributing to deaths and injuries, or creating unattractive landscapes. and the physically handicapped; For three generations, American criteria for qualenables efficient supply of the varied goods and ity transportation have translated into a love affair

PAGE 121

115 with the automobile. In 1983, 94 percent of all personal spending for transportation and 87 percent of all government spending for transportation was spent on automobile travel and highways. About 77 percent of all trips (and person-miles) are taken in automobiles or light trucks.1 2] An ability to drive is as necessary for a comfortable life in most suburban areas as an ability to walk. Women, particularly those needing to work, are now comparable to men in the proportion of licensed drivers. While only 74 percent of working women had licenses in 1969, 91 percent had them in 1983 (see table 3-14). The spending estimates actually underestimate the real cost of automobiles, if only because the opportunity cost of extensive free parking spaces provided by businesses is not counted in this total; nor are the costs of garages that are included in home prices. In some areas, these parking costs may be nearly as high as the total cost of purchasing automobiles. Transportation needs are dictated by neighborhood and city design, and by the physical relationships between residential areas, areas of employment, schools, and shopping areasas well as by the details of transportation hardware. A well-designed community may provide access to a multitude of goods and services with few or no vehicle IZI u s Energy Information Administration, Residential Transportation Energy Consumption Survey, Consumption Patterns of Household Vehicles. 1983, Table 3-14.Personal Transportation, 1969.83 Categories 1969 1977 1983 Licensed drivers/persorr . 0.52 Licensed drivers/employed adults: Females ., . 0.741 Male s 0.935 Licensed drivers/unernployed adults: Females ., 0.549 Males, : : : : . 0.648 Miles driven per driver (thousands of milw/year) Females 5.41 Males. ., ., : : : : : 11.35 Vehicles/licensed driver 0.70 Vehicles/person 0.37 Person-trips/person 740.00 Person-miles/person (000s) 7,12 Miles/trip ., : : 9,67 0.60 0.64 0.875 0.911 0.954 0.958 0.629 0.642 0.736 0.760 5.94 6.38 13.40 13.96 0.94 0.98 0.56 0.63 990.00 980.00 8.82 8.48 8.87 8.68 SOURCE: U.S Department of Transportation, Federal Highway Administration, 1983-1984 Personal Transporlatiort Study Nationwide, Volume 1, August 19aa, p 11 trips. Yet it is obvious that the great appeal of the American system of transportation is that it allows an individual mobility even if she lives in such a community. Attempts to develop self-contained, planned communities lacking this type of freedom and mobility are not likely to do well in American markets. Retirement communities are an exception created by necessity. One of the challenges of transportation technology should be to find a way to maintain the greatest possible freedom of mobility for the elderly and other disadvantaged groups. The Nations personal transportation system is mature and surprisingly stable. In spite of radical swings in the price of gasoline, the real cost of operating an automobile has not changed significantly in a generation. At the same time, increased vehicle ownership has meant that auto travel has captured a growing fraction of personal income. The system obviously serves the country well. There have, of course, been many improvements. Compared with the system that operated in the 1950s, the existing automobile fleet is more differentiated, more reliable, more efficient, and safer. While emissions per vehicle have been reduced by law, the rapid growth in driving has made it difficult for many regions to meet goals established for air quality. Sharp increases in the percentage of adults licensed to drive, and in the number of vehicles per driver (the ratio is now 98 vehicles per 100 drivers) have allowed Americans enormous freedom in personal movementused to adapt to changing job and housing locations. The personalization of transport services has increased to a point where twothirds of all trips (and 86 percent of all commuting trips) are made alone. Many of the externalities associated with conventional automobile transportation remain unresolved. A system so completely dominated by automobile travel can leave the elderly, the handicapped, and other groups without acceptable transportation alternatives. More than 13 percent of all U.S. households still own no personal vehicle, greatly limiting their access to an economy dependent on automobile transport. Heavy dependence on petroleum could place the entire economy at risk in the mid1990s and beyond. In spite of the fact that the energy efficiency of transport has increased sharply since the early 1970s,

PAGE 122

116 other parts of the economy have either done a better job of improving efficiency or have managed to shift to other energy sources. Transport consumed about half of all petroleum used in the United States in 1960. In 1986, transportation used 63 percent of all U.S. petroleum consumedmore oil than the United States produced in that year. Automobiles and light trucks alone are responsible for 40 percent of U.S. petroleum consumption. 122 Without some fundamental change in automobile petroleum use, demand can be met only by increasing imports to levels higher than they were during the peak period of the 1970s, at a time when Middle East producers will have a much higher fraction of all producing capacity than they did when OPEC (Organization of Petroleum Exporting Countries) was formed. Progress will need to begin soon if any reform is to have an effect in 15 years. Apart from dangerous dependence on foreign oil imports, transportation presents the economy with a series of problems that have proven difficult to resolve. In most areas, automobiles and other vehicles are responsible for the bulk of air quality problems, At least 30 major cities are not likely to meet 1988 air quality goals established in a 1977 congressional amendment to the Clean Air Act. Fatalities and injuries resulting from automobile and other forms of transport have fallen, but not as rapidly as other forms of-accidental death. Motor vehicle accidents accounted for more than 45,000 fatalities in 1985, and are the third largest cause of death in the United States and the leading cause of death for young males. The greater differentiation of vehicles on the road may have contributed to the problem of safety. Rule changes permit increasingly large trucks on public highways at the same time that automobiles are decreasing in size. Trends in Personal Transport The average American (man, woman, and child) now travels about 13,500 miles a year (see table 315). The average adult male spends 90 minutes a day traveling and women spend more than an hour a day (see table 3-16). The amount of travel depends IZZp.D. patterwn, Ana]ysis of Future Transportation Petroleum Demand and Efficiency Improvements, paper delivered at the IEA Energy Demand Analysis Symposium, Paris, Oct. 12-14, 1987. Table 3.15.U.S. Per Capita Passenger Travel in 1984 Mode/vehicle type Miles per person All modes . . . . . ......13,566 Personal transport . . . .....10,815 Automobiles . . . . . 8,884 Motorcycles . . . . . 58 Personal light trucks . . . . 1,873 Buses. . . . . . . . 517 Transit . . . . . . . 79 Intercity . . . . . . 113 School . . . . . . . 325 Air. . . . . . . . . 1,065 Certified route air carrier . . . 1,013 Other . . . . . . . 52 Rail . . . . . . . . 61 lntercity . . . . . . 18 Transit . . . . . . . 43 SOURCE: Calculated from U.S. Department of Energy, Office of Transportation Systems, Tfanspcmatlon Energy Data 8ook, Oak Ridge National Laboratory, ORNL%325, edition, April 1967, table 1.16. Table 3-16.-Minutes Per Day Spent in Travel Men Women 1975 1985 1975 1985 Work Travel . . . . 25 31 9 17 Family Travel . . . . 33 31 33 33 Leisure Travel . . . . 27 33 21 23 Total . . . . . 85 94 63 73 SOURCE: John P. Robinson, Trends in Americans Use of Time: Some Prelimi. nary 1975.1965 Comparisons, Survey Research Center, University of Maryland, December 1966. on the region and the type of household. Distances and time in travel are increasing slowly.123 Measured as a percentage of nondefense purchases, public and private spending for transportation has remained surprisingly constant for more than a generation. The mix of spending for transportation services has, however, changed slowly over time (see table 3-17). A slow increase in spending for personal transport has offset a decline in government spending resulting primarily from the steady drop in new State and local highway construction. But do these changes indicate progress or the lack there of? While transport can be an end in itself, it is mostly a means to an end. Does increasing per capita spending for transportation mean that additional or qualitatively new services are being provided and that Americans are enjoying more mobility, or does it mean that more time and money are being spent to procure the same set of transportation services? 123 L.11. Burns, Transportation, Temporal, and Spatial Components o~ Accessibility (Lexington, MA: Lexington Books, 1979).

PAGE 123

117 Changing demographics, such as the decline in family size, the growth of multiple-worker families, the suburbanization of job location, and changing lifestyles all translate into different transportation Table 3-17.U.S. Consumption of Transportation (in biiiions of current doiiars) Type of purchase 1955 1965 1985 Personai transport . . Household spending . New autos . . . Used autos . . . Other motor vehicles. . Tires, accessories . Repair, washing . . Gas and oil . . . Bridge, tunnel fees . Insurance . . . Public spending . . Highways a . . . Urban public transport . Household spending . Transit . . . . Taxicab. . . . . Commuter rail . . Public spending . . Transit . . . . Other transport. . . . Household spending . Other rail . . . Bus . . . . . Airline . . . . Travel agents, airport, bus, etc. . . . . Public spending . . Water & air a . . . Rail a . . . . Total transportation . . Household . . . . Public . . . . . 35.8 31.9 13.8 1.9 0.5 1.6 3.9 8.6 0.2 1.4 3.9 3.9 1.9 1.9 1.3 0.5 0.1 0.0 0.0 1.1 1.0 0.4 0.3 0.3 0.0 0.1 0.1 38.9 34.8 4.1 Percentage distribution of spending: Personal transport . . 92.O% Household . . . . 82.0 Public . . . . . 10.0 Urban public transport . 4.9 Household . . . . 4.9 Public . . . . . 0.0 Other transport . . . 2.8 Household . . . . 2.6 Public . . . . . 0.2 Total . . . . 100.0 62.8 55.3 21.4 3.8 1.3 3.5 7.6 14.8 0.5 2.4 7.5 7.5 2.1 2.0 1.3 0,6 0.1 0.1 0.1 2.5 2.0 0.3 0.4 1.3 0.1 0.5 0.5 0.0 67.4 59.3 8.1 93.2% 82.0 11.1 3.1 3.0 0.1 3.7 3.0 0.7 100.0 357.7 329.8 86.9 34.5 31.3 25.0 48.3 92.6 1.3 9.9 27.9 27.9 8.7 7.1 3.5 3.1 0.5 1.6 1.6 26.2 20.1 0.6 1.2 18.3 1.9 6.1 5.9 0.2 392.6 357.0 35.6 91.l% 84.0 2.2 1.8 0.4 6.7 5.1 1.6 100.0 aSWnding i n theWcategofles haabeenallocatad batween flnaIdemand(shown here)andspendlng bygovernment that supports lntermediateuse oftransportation. The support of intermediate transportation iacounted in the governmerit not elsewhere cleasified amenity category.Spending Is partitioned by the ratlobetween finalcommodlty demandandtotal commodity output shown foralr,water, and rail in the 1977input/output tables (U.S. DeparfrnentofCommerce, Bureauof Economic Analysis). Highway spending attributed toflnal demandwaa calculated by taldng total highway spending and multiplying bythe ratio between user fees collected from private automobiles to all highway user fees, SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Survey of Current Business, histor. Icai diskettes, tables 2.4, 3.15 and 3.16. needs. Shrinking household sizes and an increase in the percentage of women in the work force (women tend to work closer to their homes than men) has actually led to a decline in miles of commuting per household (see table 3-18). The harried life of many families is reflected in changed use of personal transportation. Table 3-18 shows a sharp increase in family travel for shopping and other purposes and a decline in recreational travel distances. Continued suburban development also shapes travel needs. Even in 1970, less than 4 percent of all commuting went from suburban areas to the central business district. 124 By 1980, 60 percent of the 31 million commuters in the Nations 25 largest urban areas lived in suburban areas and traveled to jobs outside the central business district. About half of all commutes in other metropolitan areas involved such trips. There are, of course, significant regional differences. While older cities such as San Francisco and Minneapolis still have vigorous downtown areas, many cities that have expanded rapidly in the past decade have several centers rather than one. lZ4Urban Mass Transportation Administration Technical Assistance Program, Joint Center for Policy Studies, Demographic Change and Recent Work(rip Travel Trends (UMTA-DC4Y3-7009), Washington, DC, Feb ruary 1985. Table 3-18.Vehicle Miles of Travel Per Household in Personal Vehicles Miles traveled in 1983 Percent of Percent change (thousands) 1983 total 1969-1983 Work. ... . . . 4.03 34,4 To or from work 3.54 30.1 Work related business . 0.50 4.2 Family and personal 3.56 30.3 Shopping ., 1.57 13.3 Other. ., 1.99 16.9 Civic, educational, and religious . . . . 0.48 4,1 Social and recreational. . 3.53 30.1 Vacation . . 0.25 2.1 Visit friends/relatives. 1.59 13.5 Pleasure driving . . 0.13 1.1 Other. . 1.56 13.3 Other . . 0.14 1.2 Total . . . . 11.74 100.0 ,0 .4 .8 47.9 68.7 34.9 ,9 13.7 .1 6.1 .4 17,4 7 5 SOURCE: U.S. Department of Transportation, Federal Highway Administration, 1983-1964 Nationwide Personal Transpocfatlon Study, vol. 1, Washing. ton, DC, August 1966, table 5-2.

PAGE 124

118 Table 3-19 shows how different commuting needs translate into commuting times. The time people are willing to invest in commuting seems to have changed relatively little in a hundred years. Commuters appear to put a comparatively low value on commuting time for the first 20 minutes and approach psychological limits after 45 minutes. The availability of car telephones, sound equipment, and other amenities in vehicles does not appear to have changed the basic formula. In 1980, only 10 percent of all workers commuted more than 44 minutes. 12 5 Ironically, higher income workers have slightly greater average commuting times and travel longer distances to work. On average, commuting distances have increased somewhat, while commuting times have decreased. Severe congestion problems plague many major cities. Commuting times increased in metropolitan areas with populations over 3 million between 1969 and 1983. Congestion was identified as a major problem by residents polled in San Francisco, Atlanta, Phoenix, Washington, DC, and a dozen other urbanized areas.l26 Undoubtedly, some of the difficulty results from the growing reluctance of Federal or State governments to pay for highway improvements. 125u.s. Departments of Commerce and of Housing and Urban Development, Annual Housing Survey: 1980, Part FEnergy Related Housing Characteristics (Washington, DC: U.S. Government Printing Office, 1983), p. 76, in Anthony Downs, Impacts of Rising Traffic Congestion on the Location of Activities within Metropolitan Areas, April 1985. IZSR. Cewero, Op. cit., footnote 99; Robert Dunphy, Urban Traffic Congestion: A National Crisis? Urban Land, vol. 44, No. 10, 1985, pp. 2-7. Table 3-19.Commuting Times by Home and Job Location in 1980 Average travel Type of trip time (minutes) Living in a central city and commuting to: CBD a of central city . . . . 24.9 Central city outside the CBD . ., ... 20,0 Outside central city . . ., . . . 26.4 Living in the urban fringe and commuting to: CBD of central city. . ., ... ., 35.1 Central city outside the CBD ... .., ..., ,., 27.2 Outside central city . ., . 18,8 a CBD-Central Business District, SOURCE: U.S. Department of Transportatlon, Urban Mass Transportation Administration, Demographic Change and Recent Worktrip Travel Trends (UMTA-DC-09-7009), Washington, DC, February, 1985, p. 36. New freeway construction has all but disappeared. User fees no longer cover highway construction and maintenance costs. 127 Between 1930 and 1950, the length of paved highways per person in the United States climbed from 30 feet to 70 feet, but highway length has scarcely matched population growth since the 1960s.128 Between 1970 and 1985, the number of automobiles in the United States increased 54 percent but the number of miles of roads and highways increased only 4 percent.129 Not surprisingly, the longest commutes were made by those living in an urban fringe and working in a central business district (35.1 minutes in 1980) while the shortest trips involved the growing number of individuals living in the urban fringe and commuting outside a central city (see again see table 319). The longer trips associated with shopping and other personal business (again see figure 3-18) are undoubtedly also traceable to the longer trips required for such purposes in suburban areas. Denver, for example, has six shopping centers that have more retail sales than the downtown area. Commuting patterns are also becoming complex as women enter the work force. Many of the differences between male and female driving habits are beginning to disappear. Women, particularly younger women, are getting drivers licenses in nearly the same percentages as men (again see table 3-14) and are driving greater distances. Working women make chained trips, involving a trip to a child care center and perhaps a shopping area before and after work.130 Women who both work and live in suburbs have a 75 percent higher probability of handling shopping and child care than men. In spite of the increased travel for working women, women still drive half as many miles as men. 27 Automobile Facts and Figures, in U.S. Congress, Office of Technology Assessment, Transportation, sector study, Washington, DC, 1987. IWJalculated from data in the U.S. Bureau of the Census, Hisforicai Statistics of the United States, 1776-1976 and the Statistical Abstract of the United States 1987, op. cit., footnote 39. IzgStafistica/ Absfract of the United States 1987, op. cit., footnote 39. Isowi]]iam M. Miche]son, The impact of Changing Womens Roles on Translation Needs and Usage, prepared for the U.S. Department of Transportation, Washington, DC, September 1983; Julio Perez-Cerezo, Women Commuting to Work in the Suburbs (Berkeley, CA: University of California, Department of City and Regional Planning, 1985).

PAGE 125

119 With the need for individual transport increasing, the number of cars added in the United States has far exceeded population growth for nearly a generation. There is nearly one personal vehicle for every licensed driver in the United States, and 0.63 personal vehicles for every American. The average American goes on nearly 1,000 trips and travels 8,480 miles in some kind of personal vehicle each year. There does, however, seem to be some saturation. Between 1969 and 1977, trips became somewhat shorter even though both trips per person and miles per person increased. Some of this was attributable to an increase in the number of vehicles available to each family. Since 1977, however, there has not been a significant increase in either trips or miles traveled per person. The preeminence of the automobile is challenged only by air travel for very long trips; interestingly, however, trips by air have grown largely as a substitute for rail and bus journeys. Nearly 84 percent of all trips longer than 100 miles were still made in cars in 1982down surprisingly little from a 90 percent share in 1960. Air travel (including business as well as private passenger travel) has grown steadily, from about 4 percent of all miles traveled in 1960 to nearly 15 percent in 1984. 131 The growth in share of trips has come largely at the expense of trips made by rail, bus, and other forms of public transport. The share of land-based public transport for inter-city and intracity trips has declined steadily in recent decades; in some cases, travel by these modes has declined in absolute terms. Deregulation of air travel has reduced services to many smaller towns and cities, with the effect that some trips by air are now taken by other means or, at a minimum, involve a highway trip to a neighboring airport. The effect is important since many of these relatively small centers have experienced rapid population growth in recent years. Future growth in air travel may depend on techniques to reduce total travel time, including the time needed to reach the airport. While air travel costs have declined and productivity has increased, the air transport system also appears to be headed for stagnation. The performance of the air transport system cannot be decoupled from 131 u ,.s, BUreaU of the Census, Statistic/ Abstract of the United States 1986 (106th cd.), Washington, DC, pp. 26 and 591. that of the highway system, since the efficiency of air travel is reduced significantly by delays and congestion encountered reaching an airport by automobile or other means. Including travel time to an airport, travel time between cities less than 500 miles apart has not changed significantly in a generation. Travel time between New York and Philadelphia, for example, required an hour and fifty minutes by train and an hour and thirty minutes by air in 1986 (including travel time to the airport). Studies indicate that 6 to 10 major U.S. airports either already face severe bottlenecks due to lack of curbside for connecting highway to air travel networks, or will face such problems in the future. This lack of progress is ironic, given the extremely high value that air travelers apparently place on their time. Other reports suggest that a large group of air travelers are willing to pay $30 to $60 per hour of time saved in bus travel to or from an airport. 132 Looking to the Future Virtually all attempts to make improvements in the performance of the transportation system have met with failure. Public transportation has lost ground in spite of massive subsidies and a growth of urban workers during the past decade. New technologies can help in a variety of ways. The near doubling of automobile fuel economy between 1974 and 1984 had a major effect on world oil markets. With fuel economy at 1974 levels, the United States would now be importing approximately $40 billion more oil each year. Innovations in vehicle designs could triple fuel economy by the end of the century. A variety of other technologies could work to improve real system performance. Methanol, made from natural gas, coal, urban waste, or biological materials, could provide an acceptable substitute for petroleum by the next century if steps to plan for the conversion are taken in the near future. Significant improvements in the net performance of the system, however, require attention not just to the technology of vehicles but to the entire transportation network. Fuel efficiency improvements can ljzGreg W. Harvey, A Study of Airport Access Mode Choice, .lourna/ of Transportation Engineering, vol. 112, No. 5, September 1986, pp. 525-545.

PAGE 126

720 easily be offset by heavy congestion and long commutes. New vehicles may require new kinds of guideways. Real productivity changes in transportation may require zoning permits and other mechanisms influencing the design of communities. Communication technologies could provide better traffic control, and possibly give drivers better guidance about which routes are least congested. Progress in these areas depends on a skillful mix of public and private decisions. Conventional public transit systems do not appear to offer much hope as an alternative to private vehicles. Even in the best of circumstances, the systems work well only during periods when there is a high demand for movement along a well defined corridor. But with the complex patterns of living and working emerging in todays suburban society, such corridors are increasingly rare. Uncertainty about patterns of economic growth increases the risk of investments in relatively inflexible systems along fixed routes. The American distaste for the comparative inflexibilities and inconveniences of public transit is obvious. Between 1970 and 1980, real family income declined, the number of workers living in urban areas increased by 15 million, large new public transit investments were made in Washington DC, Atlanta, and San Francisco, and large operating subsidies meant that cost of public transit rose only 44 percent while the cost of owning and operating an auto increased 250 percent (gasoline prices tripled). Nonetheless, public transit ridership fell from 8.9 percent to 6.4 percent of journey to work, while the use of personal vehicles for the journey to work increased from 80.2 to 85.7 percent. With the exception of systems like BART (the Bay Area Rapid Transit system in and around San Francisco), designed primarily to relieve congestion in areas where incomes are high, mass transit remains the option of last resort, and its use is virtually a measure of the extent to which groups are disadvantaged by the U.S. transportation system. The exceptions are those using the system to commute to a central business district, and people living in areas of the northeast (particularly New York) traditionally served with good transit. The average one-way travel time for workers using public transportation (42.2 minutes in 1980) was more than twice that of workers using automobiles (20.8 minutes). Nationwide, in 1980: l l l l l 36 percent of transit riders lived in households with no auto available; blacks were three times as likely as whites to use public transit; 31 percent commuted into a central business district; 25 percent of workers living in a building with 50 or more housing units used public transport, while only 7,1 percent of workers living in detached single-family dwellings used public transport; and minorities, older workers, women, and the working poor were more likely to use public transit than other groups. People with other options leave the transit system rapidly. Ridership has fallen as even the poor are able to purchase vehicles. The number of households with no vehicles fell from 20.6 to 13.5 percent between 1969 and 1983 (see table 3-20). Teenagers (sometimes considered a group disadvantaged by the auto-based transport system) actually reduced usage of public transport between 1970 and 1980. Moreover, the labor problems associated with operating a system with morning and evening peaks, separated by very low densities, makes traditional transit systems uneconomical in many areasparticularly those experiencing the most rapid growth. Between 1960 and 1983, private spending on public transport other than air travel actually fell by 1.3 percent per year. Government spending for public transportation projects, however, increased rapidly during the period, even though the use of most public transit systems declined, In many areas the bulk of these subsidies have benefited middle and upper income groups rather than the poor, though the poor tend to pay a greater Table 3-20. -Vehicie Ownership by Households Percent owning: 1969 1983 Percent Change No vehicle. . ., ........20.6 13.5 5 One vehicle . . . . .48.4 33.7 3 Two vehicles. . ., ., .. ..26.4 33.5 26.9 Three or more vehicles. . . 4.6 19.3 319.5 SOURCE: AutomobNe Facts and Figures, U.S. Congress, Office of Technology Assessment, Transportation, sector study, Washington, DC, 1987.

PAGE 127

121 fraction of their incomes to subsidize the systems since most systems are paid for through non-progressive taxes. 133 Subsidies for transit systems have been justified by a desire to generate growth at suburban sub-centers. While this has worked in some areas, in others desired growth either has not occurred or has actually been opposed by local residents. 134 There has, however, been a dramatic if poorly documented increase in alternative forms of public transport. Some are private systems serving specialized markets, such as limousines and vans serving hotels and airports. Others are supported at public expense. The State of California provides coupons to lowincome and elderly people that can be used to purchase trips to shopping areas or health centers in their community; taxi companies offer the State volume discounts. It must be recognized, of course, that one of the advantages of para-transit alternatives to conventional public systems is the difference in wage rates. Traditional bus and transit jobs, at least in larger cities, tend to be unionized and pay an average of $11 to $13 per hour ($18,000 to $25,000 per year). Para-transit operators average $4 to $6 per hour ($12,000 to $15,000 per year). 135 Choices and Consequences Is it then possible to envision a transport system, based primarily on personal vehicles, that could offer greater flexibility, diversity, and freedom of choice to the American public? The answer appears to be yes, but . Changes permitting real improvement in mobility through greater differentiation, cost reduction, or speed will require basic changes in the design of guideways, control systems, and parking strategies. Where feasible, they may require new strategies of community design to minimize travel needs and integrate pedestrian and non-pedestrian travel. The emergence of such systems will obviously require a mixture of public and private investments. They will also require coordinated planning and a ISSM, Weber, The BART ExperienceWhat Have We Learned, Monograph No. 26, Institute of Urban and Regional Development, University of California, Berkeley, CA, 1976. 1341bid, (Top Hourly Wage Rate Summaries Update, American Public Transit Association, various years; reports of various State public utilities commissions. long time to implement. Highway and air travel are particularly dependent on Federal, State, and local investments in infrastructure, signaling and traffic control, and terminals; fundamental change can be undertaken only through a combination of public and private decisions. The future seems to belong to a personal transportation system capable of providing the personal and flexible service demanded by a rapidly shifting economy. Table 3-21 illustrates the difference between car ownership and car usage patterns. While there has been some differentiation in car size in recent years, a large mismatch between vehicle capacity and vehicle use remains. Less than 3 percent of all cars on the road are designed for two passengers (mostly Corvettes, Fieros, and other sports cars). More than 96 percent of all work trips, 87 percent of car trips of all kinds, and 83 percent of all vehicle miles traveled could have been taken in a twopassenger vehicle in 1983. Six-passenger cars are full on 0.5 percent of all trips. The increased individual mobility resulting from greater vehicle ownership has lead to a steady decline in vehicle ocTable 3-21 .Use and Ownership Patterns of Personal Vehicles in the U.S. Fleet, 1983 (in percent) Vehicle miles Number of trips traveled Number of Earning a living All purposes All purposes occupants 1977 1983 1977 1983 1977 1983 One ., 81 .2% 86.4% 59.6% 65.7% 51 .7% 57.4% Tw O . . 94.6 96.2 84.3 87 2 79.4 83.1 Three ., . 98.0 98,8 92.6 94.6 89.1 91.4 Four 99.2 99.5 97.0 98.0 95.3 96.6 Five. . 99.6 99.7 98.8 99.2 98.1 98.7 Six & up. .. .100,0 100.0 100.0 100.0 100.0 100,0 How to Read the Above: Of all U.S. automobile trips to work made in 1977, 81.2 percent were made with one person in the car, 94.6 percent were made with two people, etc. In the same year, 51.7 percent of all vehicle miles traveled were made with one person in the car. Size of U.S. automobiles by size (in percent): 1985 1985 1986 Size class stock sales sales Two-seater . . . . . 2.1 % 3.3% 2.5% Minicompact . . . . . . 4.8 1.0 1.7 Subcompact . . . . . .. ...22.7 22,0 22.4 Compact . . . . . ...........17.1 32.7 33.2 Midsize. . . . . . .. ..28.4 28.0 26.9 Large ., . . . . . .. .24.8 13.0 13.2 SOURCE: For type of trip, see U.S. Congress, Office of Technology Assessment, Transportation, sector study, Washington, DC, 1987. For stock and sales, see U.S. Department of Energy, Office of Transportation Systems, Trartspoflatlon Energy Data Book, Oak Ridge National Laboratory (ORNL-6325), edition 9, April 1987, pp. xvii and 2-29. 52-635 0 88 5

PAGE 128

122 cupancy. On average, more than two people ride in vehicles only for trips longer than 20 miles, when families with two adults and young children travel for religious, social, or recreational purposes, for families with young children, or for family business. There appears to be a large potential market for vehicles designed to serve the needs of those traveling alone or in pairs. Actual purchasing decisions, of course, are based on the assumption that the vehicle may need to be used occasionally to carry large luggage loads or a large number of people. With many cars available, however, large families use large cars primarily for surge capacity. Given alternatives, or faced with problems resulting from fuel costs or congestion, many individuals might elect to purchase vehicles better matched to their dominant transportation needs and simply rent trucks or larger automobiles for the rare occasions when they are needed. Travel in pick-up trucks is an extreme example of poor capacity utilization, since most are driven as personal vehicles with no loads. Pick-ups accounted for 14 percent of all personal vehicles in 1983 and slightly more than 14 percent of all vehicle miles.1 37 Nearly 57 percent of the 33.8 million trucks on the road in 1982 were used principally for personal trips. 138 Measured in vehicle miles traveled, use of light trucks is growing 3.4 times faster than use of automobiles.lsg Considering the Possibilities Scenarios for the future presented here are all built on the assumption that autos will continue to dominate personal transportation markets in the United States. The Trend scenarios differ from the Alternative scenarios principally by assuming that in the Alternatives, there will be greater product differentiation and higher fuel efficiencies, and there will be a rationalization of the location of airports and systems for transferring from aircraft to other forms of transport. Automobile travel today depends exclusively on a large, general purpose vehicle operated on highsTransportation, op. cit., footnote 127. ltTlbid. pickups traveled an average of 10,550 miles per year, while automobiles average 10,055 miles per year. WY.J.S. Depa~ment of Commerce, Bureau of the Census, 1982 Census of Transportation, Truck Inventory and Use Survey. Itgpatterson, op. cit., footnote 122. way lanes up to 12 feet in width and requiring parking spaces sized to the vehicle at all destinations. A future system might seek to tailor vehicles more closely to trip functions. Work trips could be undertaken in a small, high velocity vehicle designed for one person but with room for one other person or some baggage. The vehicle could be inexpensive (as little as $2,000), run at 100 miles per gallon, and be parked in a small space. The vehicle could be designed to have good ride characteristics and handling. General Motors has such a vehicle in an advanced stage of development and several foreign producers are considering alternatives. While the vehicle could operate on standard highways with special lanes and parking spaces carved out of existing facilitiesincreasing capacity because of the smaller vehicle sizeit would be preferable to develop special roads for these small vehicles. Separate lanes and fly-overs could be much less expensive to construct than standard highways, since the lanes could be half the width of lanes built for conventional traffic and would require significantly less structural strength if they did not need to carry heavy trucks. Even with the conventional car fleet, fuel efficiency can be improved substantially within acceptable cost ranges if a market for efficiency develops. Table 322 indicates the kinds of vehicles in testing. Fuel economies as high as 98 miles per gallon are possible even for comparatively roomy vehicles capable of carrying 4 to 5 passengers. A neighborhood car with associated infrastructure could be developed for trips that did not require high speed travel. An inexpensive vehicle, designed for low-speed operation over relatively short distances, could improve the attractiveness of neighborhoods and could significantly increase the mobility of the elderly, the very young, and households not able to afford a conventional vehicle. Protected operating environments could be fitted into local street networks to permit safe operation. The vehicles could be highly efficient, use a variety of nonpetroleum based fuels, and be relatively non-polluting. An illustration of some alternatives for future car transport is given in table 3-23. This table explores lIOsee DebbY BleViS, Preparing for the 1990s: The Worfd Automotive Industry and Prospects for Future Fuel Economy Innovation in Light Vehicles j Federation of American Scientists, January 1987,

PAGE 129

123 Table 3=22.Fuel Economy of Test and Prototype Vehicles Maximum Curb Fuel economy power weight Capacity (miles/gallon) (HP) (pounds) (persons) Commercial: 1986 Honda CRX 54 60 1,700 2 1986 Chevy/Suzuki Sprint . . . 57 48 1,500 4 1985 Ford Escort d 55 52 2,100 5 Prototype: VW Auto 2,000 d ..... 6 6 60 1,700 4-5 Volvo LCP 2,000 M 69 60/90 1,600 2-4 Renault EVE d . 70 50 1,900 4-5 Toyota Compact d 98 56 1,400 4-5 NOTES: For U.S. vehicles, efficiencies use EPA combined fuel economy. European and Japanese prototype data were converted to EPA test values using conversion factors recommended by the International Energy Agency. Unless otherwise indicated, the vehicles use gasoline as a fuel, (D) indicates a diesel vehicle, (M) indicates a multi-fuel vehicle. SOURCE: Robert H. Williams, A Major Role for Developing Countries in Promoting Super Efficient Cars, paper presented at San Palo Workshop, November 1985. the cost implications of different types of car fleets and car ownership rates, and the implications for gasoline consumption. Alternatives to conventional public transit can be developed that are more suited to the practicalities of modern commuting, and that provide improved mobility for transportation of the handicapped. It is likely that many of these alternatives would require lower social subsidies than systems built along traditional rail and bus systems. A key to the system would be the use of modern communication systems to dispatch a variety of vehicles to neighborhood stops, homes, or businesses on demand. Customers could choose from a variety of vehicles, ranging from demand-responsive taxis to para-transit services. Some of these options would offer subsidized fares to those needing assistance. Jet aircraft are not well designed for short flights, and service for long-haul flights is becoming increasingly concentrated at hubs not easily accessible to many towns. Systemic performance could be improved with a better system of hubs, designed to provide service to a wide region through improved land vehicle and high-speed rail links as well as shorthaul aircraft. The systems performance can be optimized only by considering all elements of the transportation link. While these strategies could radically improve the performance of the transportation system, few of them depend on the development of radically new technologies. However, their introduction requires solutions to political and institutional problems that can be much more difficult to resolve than sophistiTable 3-23.Background Assumptions for Transportation Scenarios A. Use and Distribution of Vehicles Two Four Six Cars per Miles per person person person adult a adult 1983 . . 2 38 60 0.65 8.6 2005: Baseline c . 2 38 60 0.65 l0.0 b Case #l d . 18 32 50 0.65 8*6C Case #2 e . 44 28 28 o.75f 10.09 Case #3 h . 35 25 40 0.85 i 10.0 B. Cost of Vehicles Total Average Total vehicle price/ spending purchases vehicle on new vehicles (millions) ($1,000) (1983=1) 1983 . . 7.4 9.6 1.00 2005: Baseline . 11 .3j 14. 6 k 2.32 Case #l . . 8.5 1 8. 5 m 1.02 Case #2 . . 10.8 1 9. 8 m 1.49 Case #3 . . 17. 0 7. 4 k 1.77 C. Fuel Efficiency of Vehicles Total Average vehicle miles Spending miles per on gasoline (billions) gallon (1983= 1) 1983 . . 1,465 16.5 1.00 2005: Baseline . 2,140 26.7 0.90 Case #1 . . 1,840 44 0.47 Case #2 . . 2,140 59 0.40 Case #3 . . 2,140 53 0.45 aFOr these calculations, an adult is a person aged 16 to 75. bAss ume s that car miles per adult will increase at the rate prevailing from 1%0 to 1984. cunchanged from 1983. dMi x changed to reflect the extensive downsizing case used by Melvyn Ches. low, The Effect of Changing Household Composition on the Size Mix of New Automobile Sales: 1979-2LW0, Evaluation Research Corp., Vienna, VA, May 1980. eFollowing Chestow, it is assumed that half of all single-person households and half of all children of driving age living at home use two-passenger cars, while the rest drive four-passenger cars. fAs~umes that cars per adult will increase at half the average rate of the period 1960 through 1975. gAssumes that car miles per adult will increase at the rate prevailing from 1960 to 1984. hsame mix assumptions as case #1, but assumes that 30 percent of households will also purchase a two-passenger car. iAssumes that cars per adult will increase at the rate of the lg60s. jcalculated assuming cars per adult shown above, and using mid-range forecast showing 214 adults and a 7 percent annual scrap rate per year. kAssumes that new car prices increase at about 1970-1980 rate (recent priCO increases have been much higher). calculated as in (j) above, only using a 5 percent scrap rate. m&.sumes that the real price of fourand six-passenger CarS do not chan9e, and that a two-passenger car can be purchased for S3,000. nAssumes that a two-passenger Car can be built to acll~ew ~ mpg, a fOUrpassenger car to achieve 40 mpg, and a six-passenger car to achieve 30 mpg SOURCE: Office of Technology Assessment.

PAGE 130

124 cated technical problems. Admittedly, building any of these systems would require fundamental changes in traditional design protocols and standards as well as extensive public investment. New technologies could make significant contributions to conventional transportation systems, as well as facilitate the emergence of changes in transport strategies such as those outlined above. Information technology could play a critical role by optimizing the routing and dispatch of aircraft, public vehicles, and perhaps even personal vehicles, and by optimizing the performance of the vehicles themselves. Advanced control technologies could contrib ute to improved intercity automobile transportation, and many of the monotonous tasks of driving could thereby be simplified. The driver could be given information on speed limits, distances to exit ramps, and details about road conditions. Car status systems could warn the driver when there are problems in the car that need attention. Electronic safety systems could include night vision equipment, automatic braking, and collision avoidance systems. Traffic data systems could provide information on traffic situations such as construction and accidents. Local highway departments could also benefit from such services since they could anticipate traffic patterns and demand. The data could be used to improve traffic signaling, and to suggest alternate routes that would distribute traffic more evenly. More effective integration of vehicles and the roadway would make driving easier and safer, and would also reduce travel times and costs. Technology could also play a key role by improving the energy efficiency of vehicles, increasing the safety of vehicles of all sizes, facilitating the development of low-cost and reliable sources of methanol or other non-petroleum fuels, and reducing emissions, Constructing Scenarios Expenditures on the Transportation amenity for several different scenarios are illustrated in table 3-24. In most cases, the Trends are derived using methods described in the previous chapter. A departure from the projection of existing patterns is made for gasoline expenditures, to take into account the increasing share of the post-1973 higher fuel efficiency automobiles in the total vehicle fleet. For the 3 percent Trend scenario (the 2005 baseline Table 3-24.Consumption Scenarios for Transportation (billions of 1983 dollars) 2005 1983 Trend 3% Trend 1,5% ALT 3% ALT 1.5% Vehicles .......109 192 a 14oa 192 162 f Vehicle maintenance 72 134 a 98 a 135 b 114b Gas and oil ... 90 81 C 75 d 41 8 36 1 Air fares . 15 42 a 22 a 52g 23g Other public transport 9 13 a 10 a 10h 7 h Total ........295 462 346 430 346 Government purchases . 48 91 66 130 94 Total ........343 553 412 560 440 Percent share of GNP. . 10.1 8.5 8.7 8.6 9.3 ABBREVIATIONS: ALT = alternative scenarlo, GNP gross national product a computed using extrapolative techniques described In ch. 2. b scaled to vehicle purchase amount using the ratios calculated for the Trend case. c Baseline cage in table 3-23 dBmellne efficiency from table 3-23 with an assumption that ITIlleS/@ult lncreSSes at half the rate shown in the case where economic growth doubled. eBased on case #3 in table 3-23 fBm~ on cue #2 in table 3-23 gAssumed increase In expenditure due to a If) percent fall in alrl[ne prices, in contraat to stable prices assumed for Trend scenarios. h E xpen dit ure on public transport arbitrarily reduced by 10 Percent because of greater private mobility. iAgsumed to k 10 percent of total government purchases fOllOwing trend of P~t decade. SOURCE: Off Ice of Technology Assessment, 1987. case), it was assumed that miles per gallon rose to 27 compared with 17 in 1983, and that miles driven per adult rose by 16 percent over 1983 levels roughly a continuation of an historical trend. The same fuel efficiencies are assumed for the 1.5 percent Trend case. Because of the lower growth rate, miles per adult are assumed to remain at the 1983 level. The main difference between the Alternative high and low growth scenarios is the extent to which personal vehicles are purchased as a replacement for existing vehicles or in addition to them. These two scenarios are constructed from cases 2 and 3 of table 3-23, respectively. That is, the 3 percent Alternative scenario provides for a higher vehicle ownership, a car fleet with more large cars, and rather lower fuel efficiencies than the 1.5 percent case. Vehicle miles are similar in both cases. A comparison between the Trend and the Alternative scenarios suggests that increased mobility could be achieved in both the 1.5 and 3 percent cases at a minimal increase in vehicle expenditure and a reduction in gasoline costs.

PAGE 131

125 For air fares, it is assumed in the Alternative case that fares decline by 10 percent in real terms instead of staying constant as in the Trend scenarios, and that expenditures on air travel consequently increase. This higher level of expenditure in the 3 percent Alternative case is in line with historical trends. The relatively small expenditure on other public transport is assumed to be 10 percent lower in the Alternative case, in recognition of the greater private mobility that this scenario envisages. Government purchases of transportroads, waterways, etc.are assumed to rise in all cases due to the need to provide infrastructure for increased mobility. The increase in the Alternative cases is sharper, as special guideways for the new vehicles would be needed. Total expenditures on Transportation under the 3 percent Alternative would therefore be about the same as in the 3 percent Trend scenario, but the Alternative allows for increased mobility. Greater m~ bility is also incorporated in the 1.5 percent Alternative case, but here expenditure would be some 7 percent higher than in the 1.5 percent Trend scenario. The income distribution scenarios suggest that people place a high priority on increased mobility as incomes rise. CLOTHING AND PERSONAL CARE The category Clothing and Personal Care includes clothing and shoes, jewelry and toiletries, and the services associated with these commodities. This section focuses primarily on apparel, which represents roughly three-quarters of consumer expenditures in this area. Both men and women demonstrate strong interest in the communicative quality of their clothing, and its ability to influence image, career advancement, and self-esteem. Personal spending on highpriced fashion productswhich generally remain popular for only ten weeks before being replaced by another styleis increasing rapidly. }4] Apparel retailers face the challenging opportunity of the coming of age of those born during the baby boom. With the oldest of this group now approaching 40 and the youngest just leaving college, this generation is entering its prime years of earning and spending, and it will have an enormous influence on apparel markets for the next 20 to 40 years.1 42 Households headed by individuals between the ages of 35 and 54 have, on average, the highest household income, and spend more on apparel (and textiles) as a percentage of total expenditures than other households. As figure 3-7 demonstrates, the portion Figure 3-7.-Apparel Consumption Percent of all consumption by household ~ 6 i 5 5 { I 5 4 5 4 L 1 2 3 4 5 6 7 Household income cohort SOURCE: U.S. Department of Labor, Bureau of Labor Statistics, Consumer Expenditure Survey, 1982/83, unpublished data, 1988. of household spending that goes to apparel increases with rising household income.1 43 In constructing the different scenarios, two developments were taken into account. The first is price. Past prices fell in real terms, largely due to competition from low-cost producers abroad. Thus, import penetration ratios 144 rose from 8 to 14 percent between 1974 and 1982. If this trend continues, prices can be expected to remain low. If, on the other hand, 141u.s. Congress, Office of Technology Assessment, The U.S. Textile and Apparel Industry: A Revolution in ProgressSpecial Report, OTATET-332 (Washington, DC: U.S. Government Printing Office, April 1987), p. 16. 14zu,s. Bureau d the Census, Current Population Reports, Series p25, No. 986, Projections of the Number of Households and Families: 1986 to 2000 (Washington, DC: U.S. Government Printing Office, 1986), Table 2, Series B. WOTA, collected from Consumer Expenditure Survey, 1982/83, U.S. Department of Labor, Bureau of Labor Statistics, unpublished data, 1986. lddThe ratio of imports to new supply (domestic product shipments plus imports); Statistical Abstract of the United States 1987, op. cit., footnote 39, Table 1350.

PAGE 132

126 additional protection measures are introduced, a powerful stimulus to low prices will be removed. The second development is technology, which could a have significant impact on clothing in two ways (see ch. 6 for more on this subject): 1. 2. Information technologies in the clothing distribution chain could lower costs by reducing inventories and paperwork, improving productivity of the retail end and cutting the time from order to delivery. This is particularly relevant for domestic producers, whose comparative advantage lies in rapid responsiveness to market changes. Production technologiesincluding computercontrolled apparel assembly, computer-assisted design and manufacturing techniques, and auto mated transfer of fabricscould improve quality, provide a wider range of choice in fabrics and sizes at no extra cost, and enhance flexibility in response to consumer demand. It may soon be possible to use computer-generated images of customers wearing a wide variety of styles and fabrics and have the article made to exact individual measurements (in effect a return to individual tailoring) without a significant increase in cost or production time. Table 3-25.Consumption Scenarios for Clothing and Personal Care (billions of 1983 dollars) 2005 Clothing .......167.4 378.9 238.7 446.0 281.0 Personal care commodities and services. . 34.4 72.2 51.5 72.0 52.0 Total . ..201,8 451. 1 290.2 518.0 333.0 Percent share of GNP . 5.9 6.9 6.2 8.0 7.1 ABBREVIATIONS: ALT = Alternative scenario, GNP = gross national product. NOTES: Trend scenarios use extrapolative techniques described in ch. 2. In the alternative scenarios, personal care goods and services are as in the Trend scenarios; for clothing, an increase in expenditure is assumed due to declining prices of about 15 percent. SOURCE: Office of Technology Assessment, 1987. Defining Demand for Education 1983 Trend 3% Trend 1.5% ALT 3% ALT 1.5% For these reasons, there is considerable uncertainty about future trends in prices. Scenarios are described in table 3-25. In the Trend scenarios, spending was assumed to follow extrapolative techniques discussed in the previous chapter. In the Alternative scenarios, it is assumed that prices fall by about 15 percent, which is still less than in the past. As a result, expenditures on clothing rise substantially above the trend levels. Total Clothing and Personal Care expenditures in the Alternatives would reach 7 and 8 percent of GNP, compared with 6 and 7 percent in the Trend scenarios. EDUCATION If there is anything certain about the future of the U.S. economy, it is that the demands on the Nations educational system will increase. While it is flatly impossible to predict precisely which skills will be needed in a successful future economy, all evidence suggests that a nation whose wealth depends increasingly on the ability to develop and exploit technical innovations, and on its ability to respond quickly to changing market conditions, will need a workforce that is itself adaptable and flexible. This flexibility requires workers capable of identifying what it is that they need to know in confusing circumstances, and quickly teaching themselves what needs to be learned. Continuous learning is already a major part of the job description of many Americans, and the demand for learning can only increase. Ironically, increased use of information technology is also likely to increase rather than decrease the need for people that can work as a member of a team assembled from individuals with diverse backgrounds and skills. A system of education designed to provide the economy with the skills needed to prosper in the emerging world economy can be perfectly consistent with a broader set of objectives for the Nations educational system: the development of a system that could allow all citizens to discover and enjoy the potential of their own intelligence, to have practical access to all knowledge, and to understand and celebrate the accomplishments of the human mind and spirit. This means much more than being able to design a better robot. It means being able to learn about how things work. It means being able to have

PAGE 133

127 fun with the wit, the music, the images, or the ideas of people they will never meet. It means having an opportunity to understand the struggles and the compromises that produce free governments. It means increasing each persons freedom and capacity to grow and change. It means increasing each persons capacity for enjoying leisure and retirement. The correlation between education and an ability to prosper in modern American society go far beyond an ability to get a good job. The strong correlation between education and measures taken to promote good health was noted earlier. Education is strongly correlated with an ability to avoid personal disasters (such as divorce, major unemployment, disability, eviction, or a sharp decline in family income) and an ability to recover from disasters once they occur. The effect is measurable even when adjustments are made to account for the effects of income, IQ, age, region, and race. 145 America has, of course, always understood the links between education and national wealth and security. Concern about the quality of public education has always been a key part of the long-standing consensus about equal opportunity. Education translates an individuals theoretical right to political and economic freedom into practical powers. Changes in the structure of the national economy are likely to increase the burden on the Nations educational system for all of these purposes. The burst of industrial expansion following the Civil War was paralleled by unprecedented national legislation establishing the Land Grant College system and setting aside school sections for homesteaders. In the 1950s, the Nation responded to the growing technological prowess of the Soviet Union with a massive National Defense Education program. The challenge faced today may be no less critical. The terms of international trade hinge increasingly on an ability to generate and capitalize on ideas, and on a work force trained well enough to adapt quickly to new requirements. A fundamental difficulty in measuring progress or the lack of progressin education is the extraordinary difficulty of measuring the output in any useful way. The complex list of objectives stated above defy precise measurement. If anything, the 14sG.J, Duncan, years of poverty Years of Plenty (Ann Arbor, Ml: institute for Social Research, The University of Michigan, 1984), p. 26. problem is compounded as the kinds of skills required by the workforce become more abstract. A skill today means an ability to translate complex problems into solvable ones, an ability to find out what needs to be learned and to learn it, and an ability to absorb complex and often inconsistent information quickly. These skills are much more difficult to measure than basic bookkeeping, arithmetic, or memory skills. The perpetual problem of management in education is that the system tends to reward results that can be measured (and therefore make progress in what can be measured) while the most important products may go unmeasured. Developing adequate measurement techniques therefore becomes a critical priority for making progress in an educational system. Developing a consensus on priorities is a difficult process and will require care and leadership. At present, there appears to be a significant gap between the expectations of the teaching profession and that of their clients. Schoois are expected to cure social problems ranging from drug abuse to the shortage of babysittersroles that professional educators do not always find comfortable. Statistics reveal a sharp increase in student interest in practical, job-related skills, and a declining interest in training not directly related to employment objectives. Job-related bachelors degrees increased from 50 percent of all degrees granted in 1970/71 to 64 percent of degrees granted in 1982/83, while there were sharp declines in degrees granted in the humanities and basic physical and biological sciences. 14G Only 6 percent of the elementary and secondary teachers polled in 1984 thought that helping students get good\ high-paying jobs deserved the highest rating when given a list of objectives for education, while 46 percent of parents surveyed gave it the highest rating. Trends in Inputs and Outputs In 1985, the United States spent between 8 percent of its GNP on education (13 percent if corporate training costs are included). One American in three engaged in some kind of training or educational ex14GU.S. f)epartment Of Education, Center for Education Statistics, Digesl of Education StatMcs 1987 (Washington, DC: U.S. Government Printing Office, May 1987). 147phi Delta Kappan, The Gallup Poll of Teachers Attitudes Toward the Public Schools cited in U.S. Department of Education, The Condition of Education (Washington, DC: U.S. Government Printing Office, 1986).

PAGE 134

128 perience during the year (see table 3-26). Education is not considered to be a form of investment or savings, yet it plainly serves a role as important as investment in plant and equipment. Table 3-26 also shows that spending for education may represent 40 to 80 percent of gross national investment in new plant and equipment, and perhaps twice an large an investment after depreciation (it is difficult to know how to depreciate an investment in education). Spending patterns in education have shifted rapidly during the past few decades. A large increase in public spending for education during the 1960s and early 1970s was reversed in the mid 1970s as the baby boom generation left public schools. While statistics are poor, it appears that expenditure on other forms of education increased rapidly. 148 Approximately 23 million people over the age of 17, or 13.5 percent of all adults, took some kind of parttime adult education in 1983nearly double the rate reported in 1957, Sixty percent of the courses were related to employment. Most adult students were relatively affluent, white-collar workers who had already received a good education. Adult education is paid for by individuals, the Federal Government, and private industry; government and industry expenditures could rise significantly, since more Americans will need to be retrained as new technologies enter the marketplace. Despite these large expenditures, the record of the U.S. educational system is mixed. While it is undoubtedly possible to obtain a better education in the United States than virtually any other place in the world, the United States appears to let a larger fraction of its population fall through cracks in the system than many of its key trading partners. There are three different ways of measuring progress. The first involves examining changes in achievement over time. The past 20 years have witnessed a steady increase in time spent in school. Thus the Ids%, for example, ch. 6 of U.S. Congress, Office of Technology ASSESment, Information Technology and Its Impact on American Education, OTA-CIT-187 (Washington, DC: U.S. Government Printing Office, NO vember 1982). ldgln 1980, On]y ] 7 percent of Americans 65-70 years of age had completed a year of college, compared with nearly 45 percent of the 25-29 year olds. The increased numbers participating in higher education in recent years is, however, the result of a rapid growth in the numbers of people entering tw~year (rather than four-year) courses of study, with the aim of upgrading job skills rather than receiving a broad liberal education. fraction of the school age population that received a high school degree grew steadily during the 1970s, and the number of people seeking higher education also grew quite rapidly. 149 With the decline in available financial support, the correlation between income and education is likely to grow. The percentage of young blacks enrolling in institutions of higher education actually fell between 1984 and 1985.150 It is much more difficult to measure the qualify of the education provided by the American system. Scores in the Scholastic Aptitude Test, designed to predict performance in college, suggest a decline in the educational achievement of high school graduates. Verbal scores fell steadily from 478 in 1963 to a low of 424 in 1981, but rose to 431 in 1985. Mathematical scores fell from 502 in 1963 to 466 in 1981, and are now at 475. 151 Part of the decline and subsequent increase in test scores may be due to changes in the number and type of students taking the tests. Attempts to measure the quality of literacy among U.S. graduates has constantly been frustrated by an inability to define the term with any precision. The number of illiterate Americans is estimated to be between 20 and 60 million. This number increases by 1 to 2 million per year. }52 The growth results primarily from people dropping out of school (15 percent of Americans aged 20 to 24 had not completed high school in 1985), and from large numbers of immigrants. A recent survey of 3,600 young American adults aged 21 to 25 showed a striking range of practical competence (see table 3-27). Approximately 80 percent of all whites and 40 percent of all blacks were abIe to do simple sums if the problem was presented in a form familiar from school arithmetic tests. Only about 30 of the whites and 2 percent of the blacks taking the test, however, succeeded in solving problems that required taking a simple percentage. Very few people were able to translate a practical problem into quantitative terms.

PAGE 135

129 Another sign of defects in literacy can be found 1983 to 1984, remedial math courses were taken by in the growing need for remedial programs con25 percent of all students entering colleges and uniducted both by employers and by universities. In versities, remedial writing by 21 percent, and remeTable 3-26.The U.S. Education System in 1985 Spending Enrollment (billions) (millions) Primary & secondary . . . . . . . . . 157 46.6 b Household . . . . . . . . . . . 14 Government. . . . . . . . . . . . 143 Higher education . . . . . . . . . . 65 12.3 b Household . . . . . . . . . . . 16 Government . . . . . . . . . . . 49 Other . . . . . . . . . . . . . 41 23.0 Personal spending on education & research . . . . 14 Public libraries & other . . . . . . . . 9 Labor training & services . . . . . . . . 5 Department of Defense d . . . . . . . . 1 8 e Business & government training f . . . . . . . 30-210 Formal government training . . . . . . . 59 Formal corporate training . . . . . . . . 259 Informal training . . . . . . . . . . 50-180 h Total . . . . . . . . . . . . . 293-472 Household . . . . . . . . . . . 43 Government . . . . . . . . . . . 232-254 Corporate . . . . . . . . . . . 66-175 For comparison: Gross national product . . . . . . . . . $4,010 Gross private domestic investment . . . . . . 642 Net private domestic investment . . . . . . . 205 U.S. population . . . . . . . . . . . 239 a U.S.Department of Commerce, Bureau of Economic Analysis, Nationallncome and Product Accounts, Surveyot Currerrt Busirress, JulY 1987. bU.S. Depaflme nt of Education, Center for Education Statistics, Di9est of Educatiorr Statistics 1967 (Washington, DC: US Government Printing Office, May 1987). cEnrollment statistics f or pr imary, secondary, and higher education fromadulttrainlng Participation from Office of Educational Research and improvement, U.S. Census Bureau, cited in U.S. Congress, Office of Technology Assessment, Technology and Struclura/ Urremp/oyrnent-Rf3emPbYin9 DlsplacedAdults, OTA-lTE-250 (Waahlngton, DC: U.S.Government Printing Office, Februa~ 1986) See also Training Maga.zlne, October 1984. The figure includes corporate training. There are probably 8milIion studentsin formal Iearning programs incorporations; see N.P. Eurich, Corporate Classrooms (New York, NY:The Carnegie Foundation for the Advancement of Teaching, 1985). Counted as defense spending in the amenity accounts used in this analysis. eTh e statistic shown here, b~ed on Orlansky (see beiow), is higher than the$14 billion shown in the National Income and Product Accounts presumably because OrlanskywasabletO include training costs otherwise hlddenin otherDoDaccounts. Orlanskysdatafor 1985 show 469,000 recruitsin initial training ($l.7billionh 1.4 million students obtaining specialized skills including flight training (S6.7 billion, of which flight training Is $2.3 billion), and 69,000 assmiated with officer acquisition and professional development ($1.3 billion); support and other costs account for most of the remaining costs. J. Orlansky, The Cost Effectiveness of Military Training, Proceedings of the Symposium on the A.Wtary Value and Cost% ffecfiveness of Training (DS/A/DR(85)167), Brusseis, Belgium, NATO headquarters, Defense Research Group on the Defense Applications of Research, January 1985, p. 4. Education and training costs in these categories include tralnlng of government and private employees undertaken at the employers expense. They include formal classroom instruction and informai training that is often difficult to disentangle from routine worki.e., learning a new word-processing software, or learning a new bureaucratic protocol. They are not considered as a part of final demand because they are treated as a part of government spending in other areas, or because they are purchased by private businesses and therefore considered to be intermediate input. Data in this large and important part of education and training are very poor. gR, L, Craig and C. Evers, Employers as Educators: The Shedow Education System, G. Gold, cd., Business and Higher Education: Toward New Alliances (San Francisco, CA: Jossey-Bass, 1981). The S30 billion estimate for formal trainin9 iS approximately the same as that derived by applying a cost model to the Survey of Participation in Adult Education or SPAE (US Census), adjusting the results for underreporting that can be estimated when the SPAE participation rates for enrollment are compared with known enrollment statistics in institutions where enrollment is well known (e.9., universities). See A.P. Carnevale and H. Goldstein, Employee Training: its Changing Role and An Anaiysis of New Data, American Society for Training and Development Press, Washington, DC, 1963, in A.P. Carnevaie, testimony before the Subcommittee on Taxation and Debt Management of the U.S. Senate Finance Committee, Washington, DC, Nov. 30, 1987. Eurlch (see note C) suggests that formai corporate training costs at least $40 blilion. A more recent estimate places the corPorate trainino budget in the range of $50-70 billion (R. Neff, Videos are Starring in More and More Training, Business Week, Sept. 7, 1987, p. 110. hTh e amount paid for training that does not occur in a formal setting kI Simpty not known. Even definitions are difficult ince it is often difficult to distinguish between work and iearning. (See Carnevale testimony, cited in above note). In 1985, U.S. employers paid approximately $2.5 trillion for employee compensation. Anecdotal evidence suggests that 2-4/0 of an average employees time is devoted to some kind of informai learning. This would yield $50-100 billion in informal training costs. Government and corporate shares of total spending in this category are allocated in proportion to spending to formal training. 2.0 e 81.9

PAGE 136

130 Table 3=27.Percent of Young U.S. Adults Able to Perform Tasks Requiring Different Levels of Literacy Percent a Prose tasks Document tasks Quantitative tasks 90-100 Locate information in a Enter date on a deposit Total bank deposit entry sports article slip 50-75 Locate information in a Follow directions to travEnter and calculate news article el from one location to checkbook balance another using a map 20-30 State in writing b Determine tip given as argument made in 10% of bill lengthy news article 5-15 Generate unfamiliar Use bus schedule to Estimate cost using theme from a short select appropriate bus grocery unit price poem for given departures & labels arrivals How To Read This Table: Only 10/0 of the young adults (aged 21-25) taking the test were able to generate an unfamiliar theme from a short poem. 90/0 were able to total a bank deposit entry. a See the source for precise percentages. b No document task had a % in this range. SOURCE: National Assessment of Educational Progress, Literacy: Profiles of Americas Young Adults, Educational Testing Service, Princeton NJ, 1986. dial reading by 16 percent; 82 percent of all colleges and universities are forced to offer such courses. 153 It is reasonable to ask why such resources could not have been spent teaching the students better in the first place. Another way of gauging quality in American education is to compare U.S. and foreign systems. The U.S. education system no longer has a clear lead over systems elsewhere in the world. The quality of the K-12 education received in the United States may be lower than that delivered by some of our major trading partners. The United States of course has a much more heterogeneous population than most nations, U.S. primary and secondary students regularly score below average in standardized tests. But even U.S. elementary students in a middle class suburb of Minneapolis scored far lower on standardized tests than comparable classes in Japan and Taiwan. 155 The United States does remain ahead of most nations in the fraction of its population receiving some college education. Still another way to measure quality is to see how efficiently the traditional educational system provides the skills needed for a growing and changing econISsDigest of Education Statistics, op. cit., footnote 146, P 48. lS4The Condit;on 01 Education, op. cit., footnote 151, 1985 edition; Digest 01 Educatjonaj Statistics, op. cit., footnote 146, p. 303. lssHaro\d W. Stevenson, Shin-ying Lee, James W. Stigier, Mathematics Achievement of Chinese, Japanese, and American Children, Science, vol. 231, No. 4739, Feb. 14, 1986, p. 693. omy. It is not clear that the priorities of the primary and secondary school system are well tuned to the needs of U.S. businesses. While school officials have often cited vocational skills as the most important factor in youth employability, the business view has been that if the schools provided adequately educated youth, business would provideindeed, overwhelmingly does providetechnical training. What business decidedly indicated it did not want to do, but is in fact doing, is to educate its employees in ninth and tenth grade skills. Of greater importance, there are a number of reports that suggest that the educational system is not geared to producing the skills that will be needed in the future if the Nation is to become fully competitive in a new-technology world. ]5G In sum, what is known about the productivity of the traditional education system is not reassuring. Choices and Consequences Spending levels for education hinge on a variety of factors: l the number of people in school age population groups, l the fraction of these people who will be enrolled, l the fraction of the adult population requiring retraining, ISGRepOTt of conference on Potential Funders of Education in Mathematics, Science and Technology, Carnegie Corporation, New York, 1985.

PAGE 137

131 l l the fraction of adult training that will be purchased by individuals and government agencies and the fraction that will be purchased by private businesses (and therefore not counted as final consumption using standard accounting procedures), and the techniques employed in teaching at different levels. The last question is clearly critical. The discussions presented in chapter 6 suggest that it is possible to use innovations in pedagogy, innovations in instructional technology, and improved management to enhance the rate of learning and to make learning more useful and fun. There are two principle innovations: 1. tailoring instruction to the learning styles, interests, and abilities of individuals rather than attempting to force a comparatively homogeneous system on a diverse population, and 2. freeing teachers of routine work so that they can spend more time with individuals. It is also possible, of course, that pressures to reduce costs will result in more uniform curricula and reduced student access to instructors. The contrasts are stark. The demographic model described earlier indicates that in the year 2005, the population aged 5 to 24 will be roughly the same as it is today. Many more Americans will be middle-aged by the year 2005, and these individuals are likely to demand more of the Nationss educational system than middle-aged peopIe today. Table 3-28 provides some basic statistics on the number of people in each age group participating in education (not including those engaged in trade schools or correspondence courses). Rates increased for both the youngest and the oldest age groups during the period from 1965 to 1985. A hypothesis about future participation rates is also shown. Estimates of the number of people requiring education and training are combined with estimates of teaching methods (measured by intensity or spending per student) in table 3-29. The assumptions are shown in the notes to the table. The Trend case assumes that teacher salaries will be increased and that increased training will be provided for older age groups. It also assumes that teaching methods remain comparatively unchanged. The Alternative cases, on the other hand, are built around an assumption that teaching strategies are changed in fundamental ways. In K-12 teaching, instruction is divided into three kinds of activities: 1. 2. 3. working with computers and other new technology in away that requires a minimum amount of direct teacher supervision, participating in tutorial sessions where a single teacher works full time with a small number of students, and standard lectures in which students would listen to an instructor talk and lead group discussions. Table 3=28.-Factors Affecting Enrollment in Schools Pormulation (in millions) by age cohort Year o-4 5-17 18-21 22-24 25-29 30-34 35-44 1983 . . . . .........18.3 45.4 16.7 13.6 21.7 19.6 30.1 1985 . . . . .........18.6 45.4 15.9 13.0 22.5 20.6 32.9 2005 . . . . .........18.7 49.9 16.2 12.2 19.1 19.4 43.9 Percent change (1983-2005) . 2.2 9.9 .1 10.3 .9 1.0 45.8 Participation rates by age cohort Year 3&4 5-6 7-13 14-17 18-19 20-21 22-24 25-29 30-44 1965 . . . . ..0.106 0.849 0.994 0.932 0.463 0.276 0.132 0.061 1985 . . . . ..0.389 0.961 0.992 0.949 0.516 0.353 0.169 0.092 0.025 2005 . . . . ..0.389 0.961 0.992 0.949 0.543 0.391 0.188 0.130 0.100 SOURCE: Participation rates for 1985 and 1985 from U.S. Department of Education, Center for Education Statistics, Digest of Education StafkWcs 1987 (Washington, DC: U.S. Government Printing Office, May 1987), p. 12; population computed from demographic model described inch. 2 using Social Security Administration midrange assumptionsfor ages 3-17 participation rates are assumed to be the same in the year 2W5 as they were in 1985, for ages 17-24 participation rates grow at approximately half historic rates (with some adjustments), and assumptions for older age groups are speculations based on a rapid growth in demand for retraining.

PAGE 138

132 Table 3-29.Consumption Scenarios for Education (billions of 1983 dollars, except where noted) 2005 1963 Trend 3% Trend 1.5% ALT 3% ALT 1.5 % Primary & secondary . . . 134 16 4 a 15 5 a 20 6 f 206 f Private . . . . . . 12 14b 14b 1 8 b 1 8 b Government . . . . . 123 150b 14 2 b 188b 188b $/student . . . . . .. .2,963 3,31 1 3,13 2 4,15 0 4,15 0 Higher education . . . . 56 5 9 C 5 9 C 68g 5g 9 Private . . . . . . 13 14b 14b 1 6 b 1 4 b Government . . . . . 42 45 b 45 b 5 2 b 45 b $/student ......................4,46 1 5,13 1 4,66 4 5,13 1 4,46 1 Other . . . . . . . 23 4 2 3 2 5 2 3 7 Private education & research . 11 14d 1 5 C 14d 1 5 C Public libraries & other.. . . 8 1 5 C 1l C 15C 1l C Labor training & services . . 4 1 3 e 6 C 2 3 h l l h Total . . . . . . . 213 265 246 326 302 Private . . . . . . 36 43 43 49 47 Government . . . . . 177 222 203 277 255 Percent share of gross national Product . . . . . . 6.2 4.1 5.2 5.0 7.1 ABBREVIATIONS: ALT alternative scenario, GNP gross national product aAssumesnOchange i n enrol[mentrate9 fOragess.lTand a16/0 Increase lnthe intensityofeducation which could result from a23%lncreese In teechlngsalarles (salaries are47% of total costs) The populatlonln age cohort 5-17 years old lncreaeesby 9.9% between lSS3and 2005. bAssumes prlvatelpubllc ratios remain unchanged following Iong-standing trends. cAssumes n. changes in pafllclpatlon rates frx cohorts aged 18-2$. Spending assumed to change in Proportion to enrolled students aged 16-28 (a decllne of 0.750A) with Intensity Increased 15/0. dA99ume9 no decline (equivalent to an asaumptlon that spending in this area increases as the paIliCipatin9 school Population aged 16-30 increases, offset by a 1O O /. increase in intensity). eAss ume s new participation rates for cohorts aged 30-44. Spending is assumed proportional to enrolied Population with a 15?/0 increase in intensity. fAssumes a 9.9 0 / 0 growth ~cause of increase in schooi age population. intensity Increases to 4,000 per student. This could be achieved with a S50tYyear per student charge for capital equipment, a student teacher ratio of 30 and students spending 25% of their time on computers or other equipment. It could aiso be achieved with a student teacher ratio of 20 and students spending 50A of their time on computers or other equipment. See U.S. Congress, Office of Technology Assessment, Education, sector study, 1967. gspending scales with population aged 16-29 with assumption of higher participation rates. lntenSity inCreSSeS ls~o as the resuit of growing use of computer beeed equipment in the 3% growth case but doas not increase in the case of 1.5/0 growth. hA99um~ t. Scaie with population aged 30-44 with high participation rates. kr the 3V0 growth case, intensity is assumed to doubie. SOURCES: 1963 data on spending In major categories from U.S. department of Commerce, Bureau of Economic Anaiysis, National income and Product Accounts, Survey of Current Bus/ness, Juiy 1967; teaching costs for 1963 are computed by dividing the totei expenditures for 1963 shown in the National income end Product Accounts by total enrollment in elementary and secondary schoois and higher education provided by the Office of Educational Research and improvement, U.S. Department of Education, in D/gest of Education Statistics 1987, Washington DC, 1987. In addition to these three categories of instructional staff, there is a group of teachers who at present specialize in various activities such as teaching the handicapped, or providing special tutoring in music, art, or library assistance, perhaps in one school but sometimes in a number. These specialized teachers would be integrated in the above categories. It is assumed that information technology can be made available at an amortized cost of $500 per year per work station. 157 Obviously, there area large number of ways to allocate student and teacher time given the alternatives suggested here. These new strategies are consistent with a significant increase in staff salaries. (see the appendix for details). ISTThiS estimate is derived from the informal consensus of the group participating in OTAs workshop; see /n/ormation Technology and /ts hnpact on American Education, op. cit., footnote 148. PERSONAL BUSINESS AND COMMUNICATION The category Personal Business and Communication combines two related activities. The first, communication, involves the transfer of messages now accomplished through telephone and mail. The second, personal business, covers insurance, legal assistance, banking, and a variety of other activities that rely heavily on these message services. Both activities center on the processing of information,

PAGE 139

133 and improvements in information technology are having a profound effect on both. Demand for Communications and Information In general, new communication technologies like the telephone have come in addition to, and not as substitutes for, existing communication channels. News and information received on television complements rather than replaces information in magazines and newspapers. New technologies, however, are confusing the situation by creating a number of intermediate steps between broadcasting a message to a broad audience, and communicating a specially tailored message point-to-point. Anticipating the way consumers will react to the myriad of new communication products and services likely to be offered during the next few years is virtually impossible, if only because very little is actually known about what individuals need information for or how much they are willing to pay for it. This is partly because much of the information available to individuals is provided by advertisers at little or no monetary cost, although the cost in terms of the time needed to absorb such information can be high. The uncertainties have led some very sophisticated companies into some disastrous investments. The following paragraphs sketch out a few basic facts about residential information markets. Table 3-30, for example, shows how first class mail is used. Of all 1986-87 first class mail received by households, 80 percent involved a transaction of some kind, and nearly 50 percent involved bills, invoices, and payments. The remaining 20 percent involved personal correspondence. First class mail is responsible for roughly two-thirds of U.S. Post Office revenues. Demand for first class mail is quite inelastic. A 10 percent price increase in the first class rate is likely to have a negligible effect on volume, and therefore is likely to produce a 10 percent increase in revenue. The other mail classes, even though subsidized by first class, are subject to competition and sensitive to price increases. Telephone prices have fallen in real terms for more than 30 years. Between 1950 and 1982, for examTable 3.30.The Distribution of First Class Mail, 1986-87 (received by households) Mail by type Percent Personal correspondence from a friend or relative. . . . . . . . . 17.9 Personal letter . . . . . . . (6.3) Holiday/seasons greeting card . . . (6.5) Other greeting card, invitation, announcement (5.2) Business . . . . . . . . . 80.9 Bills, invoices, receipts. . . . . . (46.4) Advertisement, notices, etc. . . . . (18.1) Other . . . . . . . . . (16.4) Dont know/no answer . . . . . . 1.2 Total . . . . . . . . . 100.0 NOTE: The category of other includes such items received from private businesses, government, or social, charitable, political, or nonprofit groups that are not elsewhere classified (as well as educational acceptances and report cards). Annual totals compiled by adding totals for the first three quarters of postal year 1986/87, and approximating data from the fourth quarter (6/8/87-9/28487) by doubling the totals of the third (3/16/67-617/87). Sample size over three quarters-20,713 weighted pieces. SOURCE: US. Postal Service, Demand Research Division, September 1967, unpublished. pie, the cost of telephone and telegraph services measured in current dollars increased by 213 percent while the cost of all products in the economy increased by more than 350 percent. There is no reason to expect that this pattern will be broken, since a variety of new technologies are available for reducing communication costs. And because the elasticity of demand for telephone service is high, new technologies can be expected to lead to significant increases in the volume of telephone traffic. Enhanced services, however, are likely to change the nature of telephone service in qualitative ways, and prices may increase to cover these enhancements. The question of whether the cost of the basic service will also increase depends on the extent to which regulatory authorities permit regional holding companies to include the cost of enhanced service in their basic rates. Market fragmentation is clearly having a major effect on telephone service. At the end of the 1970s, virtually every household in the country had the same plain old telephone service (known as POTS). Today there is a proliferation of home telephone devices and a slowly growing range of services (most now marketed to the home through 900or 979services, providing everything from stock market information to dirty jokes). Portable telephones, redial features, and cellular telephones are entering markets once limited to the standard handsets available from AT&T. At least 10 percent of American homes

PAGE 140

134 now have answering machines, sales of which increased at a compound rate of 34 percent between 1980 and 1984. 158 Future competition for the point-to-point message service is likely to be dominated more by improvements in quality than by reduction in the price of basic services. There is no good vocabulary for measuring the quality of communication, but at least four categories suggest themselves: timeliness, accessibility and convenience, presence, and the maintenance of an archival record. The situation will become more complex as the new technologies introduce a range of intermediate services between point-to-point communication and broadcasting. Consumers are already being besieged by computer-generated broadcast telephone calls. Catalogues and even the editorial content of magazines are being tailored to specific regions, demographic groups, and marketing groups. Many businesses and universities are experimenting with narrowcasting training sessions, announcements about policy changes, and other business communications. Cable television and VCR rentals are fragmenting the broadcast of television programming, allowing access to more specialized and narrower markets. Countering this trend toward geographic and demographic targeting is the growth of national newspapers. The Wall Street Journal and Gannetts USA Today, transmitted by satellite to distribution points throughout the country, reached a circulation of two million in 1985. 159 Direct marketing is thriving on a system in which data about products is broadcast through television, and catalogues are sent by third class mail with customers making purchases through 800-number telephone calls. Rather than broadcast data through publications, database systems permit users access to highly specific data. There are a large number of potential applications for new kinds of point-to~point communication services. However, few of them can be realized with the existing state of technology available in the home, and few can by themselves justify the introduction ISspacific Te]esis Group Voice Storage and Retrieval (VSR) hformation Services, July 11, 1986, cited in P.W. Huber, lhe Geodesic Network (Washington, DC: U.S. Government Printing Office, 1987). ISgArlene K. F]eming and Robert S. November of LINK, Inc., The Impact of Technology on Home Information, Transactions, and Entertainment, contract report prepared for the Office of Technology Assessment, June 1985. of advanced technology-although increased use of home satellite dishes and cellular telephones has in part resulted from demand for such services. Highspeed communications will continue to enter individual residences because of a market for broadcast or narrowcast entertainment and news programming. If a significant fraction of the cost of the system can be attributed to this function, point-to~point services available to the home can be delivered at a lower marginal cost than that of services of equivalent quality, and growing markets can be discovered. Financial, technical, regulatory, and marketing problems have prevented these systems from offering anything but the most basic re-broadcast services, and it is unlikely that existing cable networks can evolve into a multi-purpose information system capable of delivering point-to~point communication services in the form of voice or data communications. Sophisticated residential communication services will almost certainly be provided by an organization associated with existing telephone companies. Applications might include: l l Home Banking and Bill Paying. To date, there has been more talk than action in home banking. At the end of 1986, there were about 100,000 home banking installations. Most of them are operated by individuals who have already invested the time and money needed to become familiar with home computers, and who are willing to pay $5 to $20 a month for the privilege of instant access to transactions. But demand could increase rapidly. In addition to simple checkbook accounting, the systems could be given a variety of features such as sorting and documentation for home records and tax purposes. Still more sophisticated systems could offer complex financial services and ready access to financial information. A recent survey by the insurance industry concluded that most households would be pleased to order insurance at home over computer lines and to make claims through electronic mechanisms, instead of filling out tedious forms. Home Health Care. Increasing pressures to get patients out of hospitals lead to a growing need for communication between homes and health care facilities. Microprocessors can provide assistance in scheduling the administration

PAGE 141

135 of drugs at home, and can also monitor dosage to eliminate undesired combinations of medication. Physicians, acting with the assistance of visiting nurses, nursing home staff, and the patients themselves, can use equipment located in a patients home and even devices implanted in the patients body, such as pacemakers and implantable pumps, to monitor the progress of a patient. l Residential Controls. Home energy systems capable of scheduling residential electric demands to minimize system costs have already been described. Other systems that could be coupled with the network include security systems, pay-per-view television ordering, and other applications. l Information Services. Systems creating an easily accessible, nationwide database have already been instituted in France, Japan, and several other nations. The absence of a government supported program, coupled with a ban preventing the Bell Operating companies from offering information services, is at least partly responsible for the fact that home information systems have moved much more slowly in U.S. markets than they have elsewhere. l Home Shopping Services. The growing volume of catalogue shopping resulting in part from a decline in available time, suggests that a market may exist for electronic shopping at home. A variety of projects are under consideration including the ambitious Trintex system conducted as a consortium of Sears, CBS, and IBM. The French national telephone company appears to have succeeded in introducing a nationwide information network in that country by distributing inexpensive terminals to their customers. The system provides access to the 23 million listings of the French telephone system as well as more than 4,000 privately offered information services billed through the telephone system. lGO Services include an electronic newspaper, classified ads, bulletin boards, home shopping, a dating service, etc. About half of the subscribers reportedly use home banking and 24 percent use the home shopping service on a regular basis. lG l The information services are billed at l~OTe/e~a(j~ue News, Paris, spring 1987. I~Ip,w, Huber, The Geodesic fVetwork, report prepared for the U.S. Department of Justice, Antitrust Division (Washington, DC: U.S. Government Printing Office, January 1987). rates varying from 5 to 10 cents per minute of use. There were 2.24 million Minitel terminals operating at the end of 1986, with an average rate of use of 97 minutes per year per terminal. 1G2 The total number of terminals has doubled in the past year. The French claim that the system is now profitable. Telephone information services provide approximately 17 percent of total revenue. The remainder coming from private service firms using the Minitel system as a vehicle for reaching residences. Japan, Germany, Canada, and England have had similar experiments with much less spectacular results. The Japanese are experimenting with a 3.5inch floppy disc now available with the entire fourvolume Tokyo telephone directory. In one second, this disc can provide listings for 9,170 sushi restaurants in Tokyo, and can then narrow the choice based on geographic vicinity. Keytron, Viewtron, and other efforts to introduce videotext services in American markets resulted in expensive failures. In contrast to the centrally planned experiences abroad, however, the regionalization of American industry makes it likely that locally arranged formulas, which accord to the specifics of local markets, could be more successful than a single national program. Demand for Business Services Demand for personal business services is being reshaped by increasing consumer sophistication and a rapidly changing regulatory environment. Demand is becoming more fragmented as technology and affluence permit greater choice; it is also becoming more homogeneous in areas where technology permits economies of scale in the sale of standardized products. Changes under way in life insurance provide a vivid example. Traditional whole life polices combined three kinds of products: investment, risk coverage, and financial service. In the parlance of the industry, these products have become unbundled. Higherincome individuals interested in investments now look to other financial instruments, which offer greater liquidity, more flexibility, and higher returns. They also look elsewhere for more comprehensive 162J, Grenier and G, Nahon. France Wins Big With h Minitel Videotext System, ie/ephony, July 27, 1987, pp. 46-49.

PAGE 142

136 financial services. Lowand middle-income individuals interested primarily in risk coverage now look increasingly to standardized, simple, and inexpensive insurance that provides no thrift or savings value. As a result of these changes, premiums for whole life policies fell from 76 percent of all premiums in 1950 to 42 percent in 1982, while cheaper term insurance, providing no savings, grew from 41 percent of all life purchases in 1960 to 59 percent in 1981.163 Comparatively inexpensive group products, which represented 30 percent of insurance in force in 1960, grew to 46 percent by 1982. Consumer savings have moved sharply away from life insurance, commanding only 3.4 percent of individual assets in 1980 as opposed to 5.9 percent in 1960.164 Demand for property/casualty insurance grows roughly in proportion to the number of things to be coveredprimarily houses and automobiles. These businesses have become highly competitive and offer highly standardized products inexpensively. Some are sold through financial supermarkets, offered by companies as diverse as Citicorp and Sears. The I FS American counci] of Life Insurance, Life Insurance Fact Book, 1983. Wbid. conversion of a product to a commodity does not necessarily reduce the quality of services offered. Clerks in local offices now have access to semiautomatic underwriting and claim review systems, allowing them to enter information directly from a local office or even the field, where a portable computer can be used in claims examination. The time required to process applications and process claims can be greatly reduced. The price of policies can be expected to decrease as automation substitutes for the current inefficiencies of the insurance system, which are being eliminated rapidly as intense competition forces an unprecedented examination of costs. Choices and Consequences Spending on communications for home use will be highly sensitive to technical developments throughout the economy and to programs designed to facilitate high-quality communication to the home. Table 3-31 explores some scenarios. The Alternatives assume a continuation of the rapid decline in prices and improvements in the range of service quality and variety. Table 3-31.Consumption Scenarios for Personal Business and Communication (billions of 1983 dollars) 2005 1983 Trend 3% Trend 1.5/0 ALT 3/0 ALT 1.5 % Telephone . . . . . . 37.9 78. 7 a 66. 2 a 111b 9 2 b Stationery . . . . . . 5.8 1I.l c 8. 0 C 1 I c 8 C Personal Business ................132. 5 282. 9 a 171 .3 a 333d 20l d Total PCE . . . . . .. ..176.2 372, 7 245. 5 45 5 301 Government. . . . . . 0.7 1. 3 1. 0 1.3b l b Total . . . . . ......176.9 374.0 246.5 456 302.0 Percent Share of gross national product . . . . . . 5.2 5.7 5.2 7.0 6.4 ABBREVIATIONS: ALT = alternative scenario, PCE = personal consumption expenditure a Follows extrapolative technique described in Chapter 2. hhe trend scenario assumed a 20% decline in the real cost of telephone communication and a price elasticity of -1.12. The ALT 30/0 case assumes a 40/0 decline in costs following a trend in communication equipment. N is also taken to be the reduction in cost of Information processing equipment likely to constitute a major part of spending in the telephone category during the next 20 years. cAssumes that s~nding in this category is the same proportion of totaJ Personal Ckmsumptlon SS it wSS in 1~ for ~1 scenarios. dArb]tradly incresged by 150/o to include charges of a variety of home information and business Services rangin9 from consumer marketing to financial services. SOURCE: Office of Technology Assessment, 1988.

PAGE 143

137 RECREATION AND LEISURE Increased household income, coupled with decreasing amounts of free time, have reshaped American leisure habits. Many of the changes are difficult to document with precision. For the purposes of this discussion, recreational and leisure spending include: 1. use of the media (now largely within the home) this includes reading books, newspapers, and magazines; listening to the radio, records, and tapes; and watching television; 2. participatory sports, spectator sports, and club functions (including welfare and religious activities), which usually take place away from home; 3. traveling for pleasure and vacationsspending for domestic transportation is included in the transportation amenity discussion, but all other expenses (such as hotels and amusements) as well as foreign travel are included in these accounts. The definitions are thus guided by expenditure categories. But it is obviously a mistake to conclude that demand for recreation can be reduced to spending on related products and services. Recreation and Leisure also includes considerable non-market activityparticipation in certain religious, professional, political, or athletic activities, for example, and informal social visiting among family and friends. Changes in the way Americans use their leisure time were discussed in the previous chapter. Having more money to spend on recreational activities does not necessarily mean that free time is being better used; and it certainly does not mean that there is more free time available. Recent Trend# A number of often paradoxical trends are notable in this complex economic sector, one in which changing demographic factors and social trends makes the past an unreliable guide to the future. These include: l a squeeze between decreasing usable leisure time and increasing costs of recreational activities; l l l l l l l l l partial displacement of mass-production leisure products and services by more differentiated (and more expensive) specialized ones, resulting in parallel mass and differentiated economies in some fields; a lack of clear movement from outdoor to indoor recreation; in fact, media and participatory activities in some ways reinforce each other as well as competing for time and money; a growing diversity of activities with a high degree of specialization, each involving its own media, equipment, clothing, and social organizations; burgeoning demand for convenient, close-tohome, or short-term recreations, which fit into increasingly complex personal schedules; growing competition for mass media as targeted advertising and pay-per-view media drain revenues that currently subsidize inexpensive or free media like newspapers and broadcast televisionalternatively, the new media could complement the old in another parallel mass/ differentiated dichotomy; a potential split between overcrowded, deteriorating public recreational facilities and expensive, specialized private ones. (The pressure for user fees to replace public funding of facilitiesincluding parks and public libraries could widen the existing economic/educational gaps in enjoyment of recreational variety.); keen competition in some recreational fields, producing superior products and service with better employment opportunities at only slightly increased cost to the consumerfor instance, formerly exotic cuisines are increasingly available to all, just as more customized travel opportunities abound; growth in recreational businesses catering largely to childless couples and young singles as well as to seniors who are healthier, more affluent, and more knowledgeable than in the past, and fuller participation by minorities and handicapped Americans in the whole range of recreation available; the difficulty of single-parent families and less

PAGE 144

138 educated people to have full access to recreational opportunities. The entry of more and more women into the labor market particularly alters the nature of recreational demand; a need for greater leisure literacy to enable the whole population to take best advantage of multiplying opportunities the way many of the most affluent and educated do now, in spite of tightly scheduled personal lives; and opportunities for the United States to exploit its domestic advantage in the Recreation and Leisure sector while increasing its export revenues. There has been a rapid increase-more than 5 percent per annumin spending on Recreation and Leisure over the past 23 years. This general rise conceals several contrasting developments. Expenditure on books, newspapers and magazines, and spectator amusements (such as the movies, theater, and sports events) rose moderately (between 1 and 2 percent a year), while expenditures on sports equipment and other entertainment services (such as sports clubs and golf courses) rose rapidly (between 6 and 7 percent annually). There was a particularly sharp rise in purchases of radio, television, and other electronic devices (more than 10 percent), whereas expenditures on repairs rose slowlytestifying to improved quality and reliability. The two vacation itemshotels and overseas travelrose at 4 percent annually, faster than total personal consumption expenditures but not so fast as some other elements of Recreation and Leisure (see figure 3-8 for more detail on these trends). The trends suggest that demand in this sector is primarily for goods rather than services, Americans are devoting more of their Recreation and Leisure dollars to new media technologies. Much of this development has been driven by electronics technology. To the degree that most Americans enjoy spectator sports, theater, concerts, and other forms of art and entertainment, they are increasingly able to do so by means of the mass media. This is confirmed by patterns of expenditures on spectator amusements. The share of these amusements in the total, while never large, dropped from 10 percent in 1960 to about 4 percent in 1983. Expenditures on movie theaters actually declined in certain years, reaching a low in 1975 that was 65 percent of the 1960 level. While there has been some later recovery-many movie theaters have adapted to rapidly changing consumer demand by increasing the number of films shown at any one time or serving refreshments at ones seatspending at movie theaters grew a miniscule 5.1 percent between 1960 and 1983. In contrast, there has been spectacular growth in spending on television and home electronics. According to one survey, the percentage of consumer media spending allocated to new electronic media cable and pay TV, video cassette recorders (VCRs), video games, and home computersrose from 7.5 percent in 1978 to 30.8 percent in 1982.165 In the late 1970s, growth was concentrated in cable television; by 1984, cable television revenues had reached $7.5 billion.166 The early 1980s were also the era of the video game. In 1981, video game arcade users spent $5 billionequal to the combined revenues of the Las Vegas gambling industry and the U.S. film industry, or the total television revenues and gate receipts of major league baseball, football, and basketball. 167 In 1982, this figure rose to $7 billiongreater than the combined revenues of the movie and record industries. Demand for video games began to slacken in 1983, but electronic media spending was buoyed by yet another enormously popular productthe VCR. So far, the VCR has followed the classic growth curve of innovative home electronic products; the first eight years of VCR sales parallel almost exactly the growth years of color television.168 In 1985, consumers spent between $2.3 and $4.5 billion to rent and purchase video tapes.169 By 1986, movie industry revenues from video-cassette sales equaled revenues 165 Ronald Rice, Development of New Media Research, in Ronald Rice and Associates, The New Media: Communication, Research, and Technology (Beverly Hills, CA: Sage Publications, 1984), p. 16. lbsRushworth M. Kidder, Videoculture: TV-for Better or Worse, a Window on the World, Christian Science Monitor, June 10, 1985, p. 20. ldTRona]d Rice, New Media Technolo~: Growth and Integration, in Rice and Associates, op. cit., footnote 165, p. 66. l~David Lachenbruch, Home Video: Home IS Where the Action 15, Channels of Communjcatjon Field Guide To the Electronic Media, p. 42. lb9A]jean Harme[z, studios Woo Cassette Mass Market, New york Times, Feb. 27, 1986, p. C26.

PAGE 145

139 40 1 38 J 36 34 3 2 30 1 Figure 3-8.-The Leisure Industry, 1986 Recreational Book Spending 1982$ Per capita /1974 1976 1978 I I 1980 1982 1984 Market Share of Music Storage Technologies Percent of sales in current dollars 75% 50% 25% nw L m 1975 1980 1981 1982 1983 1984 1985 1986 = LPs = Singles n 8-Track m Cassette ~ CD S Motion Picture Audiences Index where 1980=100 120 115110105\ \ 95.. 90 r ( r F r 1 1980 1981 1982 1983 1984 1985 1986 1987 Admission s --Revenues (1982$) Per-Capita Consumption of Alcohol 1982$ dollars per person 150 \ 125100-------75 ---------/~--- .-. ----_---~-------. 50 1 1 1 I 1 ( 1 r t 1 r r 1 r I I f 1 I I 1 I 1 1 T 1 1955 1960 1965 1970 1975 1980 1985 Off Premise s -On Premises NOTE: Deflator for admissions to specified spectator events used to convert motion picture admissions to $1982. SOURCES: U.S. Bureau of the Census, Statlstlcal Abstract of the Urrlted States. 1987 (107th edition), Washington, DC, 1988, tables 388 (recreational books) and 372 (music technologies); U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, tables 2.4 (motion picture revenues) and 2.5 (alcohol consumption); Motion Picture Association of America, New York, NY for admissions. Motion Picture revenues and recreational book spending converted to 1982 dollars using table 7.10 of the National Income and Product Accounts. Categorization based on Leisure Statistics, The Wa// Street Journal, Apr. 21, 1988, p. 13D. from box office movie sales. Approximately one-third for part of the particularly rapid increase in purchases of American households already own a VCR, and of these items. estimates suggest that VCR penetration will reach 50 percent by 1990. These developments owe much to increases in income coupled with falling prices. Over the past 23 years, prices for television and audio entertainment have declined 60 percent in constant dollars. Moreover, the most recent forms of electronic equipment have entered the market at relatively low prices, ensuring rapid market penetration. Prices of sporting equipment have also declined sharply, accounting Changes in disposable time have interacted with these developments. Studies of the way people use their time show a sharp decline in the amount of free time available, especially in the past 10 years. This places a premium on home entertainment, such as running a movie on the VCR or seeing a sporting event on a large screen, rather than on spending extra hours in transportation to entertainment centers outside the home. This, in turn, further diversifies demand for recreational activities that can serve

PAGE 146

140 the particular tastes and needs of individual consumersactivities being improved by technologies that permit greater flexibility in response to consumer demand. Choices and Consequences In the area of electronic entertainment, there seems every reason to believe that market penetration by some of the more recent technologies, such as compact discs, cable television, and VCRs, will continue. The development of such new technologies as video discs and high-definition television will also improve the quality and range of entertainment (and other information activities) within the home. New developments will combine the telephone and personal computerat present, mainly designed for social contact and personal businessto provide a variety of information and entertainment services. The development of such new technologies as video discs and high-definition television will also improve the quality and range of entertainment (and other information activities) within the home entertainment centers; there is every reason to believe that these innovations will eventually become available to consumers at reasonable cost. The French ( Minitel hookup, originally designed to give consumers more efficient telephone service and buy them time through efficient transaction of business, has won widespread acceptance through its adaptation as a recreational device. Despite regulatory differences, an American system or systems could take advantage of similar consumer demand. A substantial part of Recreation and Leisure activities, such as ski centers, golf courses, sightseeing buses, and hotels are still labor intensive. This fact has accounted for a past increase in relative costs. If new information technologies permit improvements in labor productivity, these costs could decline in the future. Yet demand for better services could produce a generation of recreation or hospitality professionals to enhance the consumers experience. Two related factors, demographic trends and available time, will also influence expenditure. The ability to enjoy leisure and recreation needs a combination of time and money. If time is in short supply, leisure may be curtailed regardless of money, or at least recreation expenditures may be channeled into directions that save time. The growing complexity of family schedules fuels demand for convenient, close-to-home recreation and shorterthough perhaps more frequentvacations. The time budget data referred to in chapter 2 serve as a basis for a rough estimate of future trends in leisure time. It is first assumed that disposition of time for both men and women in 2005 will be the same as in 1985. In this case, free timetime which is not spent on work, traveling to work, housework, personal care, and eatingwill increase by about 25 percent between 1985 and 2005. The largest portion of this increase, 21 percent, is due to the rise in population, and the balance to an older population that will include a greater share of retired people. Individuals, however, may find themselves more pressed for time than ever. On the other hand, there may be changes in the way time is spent. Time budgets of 2005 may well be different than those of 1986. A IO-percent reduction in the work week occasioned by improved technology could increase the hours of free time available by 6 percent, while a halving of the time devoted to shopping due to improved information technology could increase free time by 5 percent. If, in contrast, the rising participation in the labor force which caused the reduction in leisure time between 1975 to 1985continued, the amount of time available for leisure and recreation would fall. Household formation is also of importance. The share of young single people, who spend more of their income (and higher absolute amounts) on recreation, will decline, thus tending to depress Recreation and Leisure expenditures. A countervailing factor, however, is the expected increase in childless households, traditionally high spenders on recreation. Furthermore, the new elderly, a rising share of the population, may spend more on Recreation and Leisure than their present day equivalent. The future, therefore, offers mixed prospects, On the one hand, the abundance of leisure activities and cultural opportunities potentially available is the greatest in history. There will be an unprecedented selection of destinations, books, entertainment, sports, media, and cuisine. As new technologies enhance flexibility in consumer choice while helping to reduce prices, the information necessary to per-

PAGE 147

141 sonalize fulfillment of leisure needs and opportunifor example, spend virtually nothing on travel or enties may soon be within the reach of all. tertainment. Emerging fee-for-service media may On the other hand, not all Americans maybe able drain advertising revenue from traditional publishto share in these riches. Household expenditure data ing and broadcasting, depriving the less affluent of suggest that there are many families too poor to take much affordable or free information and entertainment. advantage of recreational activities. Single parents, Scenarios are described in table 3-32. Table 3.32.Consumption Scenarios for Recreation and Leisure (billions of 1983 dollars) 2005 1983 Trend 3% Trend 1.5/0 ALT 3 ALT 1.5/ 0 Entertainment services . . . 58.2 1 54.3 a 81.1 a 183 b 97b Entertainment commodities . . 62.4 149.8 a 104.9 a 165 b 115b TV and sound . . . . . 31.4 87.6 a 61 a 106 b 78 b Lodging and foreign travel c . . 22.2 64.9 37.7 84 b 51b Religious and welfare d . . . 47.6 91.2 65 119 86 Total PCE on recreation and leisure . . . . . ....221.8 547.8 349.7 657 427 Government purchases of recreation and Ieisure e . . . . . 8.6 16.5 11.9 20 13 Total . . . . . ......230.4 564.3 361.6 677.0 440.0 Percent share of GNP . . . 6.8 8.7 7.7 10.4 8.3 ABBREVIATIONS: ALT = alternative scenario, GNP = gross national product, PCE = personal consumption expenditure aFollows Extrapolative technique described in ch. 2. bAlternative estimates of demand for enteflainment services, entefiainment commodities, and tv and sound eqUipment are above those computed in the trend analysis for reasons discussed in the text. The increases are constructed by assuming a price reduction in the items that is more rapid than the one used in the trend case. The price of entertainment services is assumed to fall 150/., entertainment commodities 300A, and TV and sound equipment 400/.. Cln th e period 1980.72 real expenditures on foreign travel incre~ed 480/0 f@er than all PCE. After 1972 there were wide variations in spending reflecting swings in exchange rates. Extrapolation is clearly difficult. In the estimates shown it has been assumed that foreign travel will continue to grow at roughly the 1980-72 rates. It has therefore been assumed that spending in this category would increase 4.44/Olyear if GNP grows 3/Olyear and that spending would increase 2.2/Olyear if GNP grows 1 .5 V01year. dAs s ume s that spending in this category IS the same proporflon of total PCE ss it was in 1983 for all Scenarios, eGovernment purchases assumed to be in proportion to personal consumption. SOURCE: Office of Technology Assessment, 1987

PAGE 148

Part 11 The Networks of Production Understanding how changing patterns of consumer and government purchases translate into demand for output from specific businesses is a heroic challenge in a modern economy. Products once brought to consumers through direct and obvious channels are now provided through networks of extraordinary complexity. In the 19th century many Americans lived on the farm. They produced most of their own food and, with the exception of an occasional purchase from a blacksmith, were essentially self-sufficient. Today, not only have farms become specialized, highly capitalized, and technologically sophisticated, they have become comparatively minor elements in a sophisticated network of businesses delivering the Food amenity to consumers. This section describes the operation and performance of these production systems in a modern economy, and speculates about how they may change in the future. The growth of linkages connecting diverse parts of the economy helps explain the explosive growth of sectors other than traditional manufacturing. A packaged frozen pizza may combine produce picked and processed in California, sophisticated packaging materials made in Michigan, and a paper label manufactured in Washington and printed in Illinois, as well as financing from Chicago, advertising and legal services from New York, trucking from independent truckers throughout the Nation, inventory control and billing software from Palo Alto, and communication systems from New Jersey. The service businesses play a growing role in getting pizzas to American plates. Among other things, such businesses facilitate the billions of transactions that connect elements of production networks, make it possible for firms to redesign their operations quickly in response to new market opportunities and changes, and allow production networks to make and deliver products and services more precisely targeted for the increasingly diverse markets described in Part I. Increasingly, networks of production also involve foreign producers and consumers (the topic of Part 111). Tightly integrated networks create a situation where the effects of international trade are felt far beyond the sectors actually engaged in trade. Imports that displace U.S. manufacturing output, for example, also reduce output in a constellation of nonmanufacturing enterprises linked to manufacturing such as banking, insurance, and business services. The complex networks that connect producers to consumers in the American economy have become so elaborate that it is difficult even to determine how and where value is added in the process of delivering useful products and services to final consumers. Yet most of the important changes taking place in the American economy can only be understood by observing the integrated performance of these networks operating as a whole. In such a situation it is easy to be misled by asking the wrong questions. It is entirely possible, for example, that firms appearing to enjoy productivity gains may perform poorly within a dynamic and flexible network. A firm that depends on mass production and great certainty about future markets to ensure low costs may have difficulty in a world where products and production strategies are in constant flux. Certainly, any attempt to improve network performance through public policy now requires a system-wide perspective to ensure that changes actually help, rather than frustrate, the emergence of efficient networks. Driven by new technologies, the pressures of foreign trade, new patterns of demand, and a changed regulatory environment, profound changes are sweeping through virtually every major business network. Even those that have resisted change, such as residential construction and education, may be on the brink of major change. The next three chapters provide a description of trends in the performance of the eight major networks that deliver amenities to final consumers and suggest possibilities for future patterns of change. Chapter 4 provides some basic tools for describing the integrated performance of these networks. Chapter 5 uses these tools to examine recent patterns of change. Three kinds of change are examined: 1. Changes in the production recipe, or the mix of goods and services a business uses to make its product. Virtually every such recipe involves proportionately less natural resources and more transactional serv143

PAGE 149

144 ices like banking and legal help. And nearly all production recipes are growing in complexity, increasing the links connecting one part of the economy with another. 2. Changes in the internal management of busine~, and in the way businesses connect to form production networks. Generalizations are difficult. It appears that highly fragmented structures (elements like small farms, homebuilders, or family physicians) are increasingly becoming parts of larger enterprises, while firms that had concentrated production in a few establishments are turning to smaller facilities and purchasing more inputs from other businesses. In an increasing number of cases, a few dominant firms set de-facto standards and serve, indirectly, to organize activity. 3. Changes in the geography of production. The growth of activity outside of traditional manufacturing and the vast increase in communications technology offers the potential to decentralize production in ways never before considered possible. In general, however, this potential has not been realized, and economic activity increasingly accumulates in and around major urban centersparticularly those on the east and west coasts. Finally, chapter 6 pursues these themes through the eight major amenity networks discussed in chapter 3. Each, of course, has a unique story to tell. Taken together, they can provide a perspective on structural change in production that cannot be obtained from an examination of economy-wide statistics.

PAGE 150

m (l) d

PAGE 151

CONTENTS Page Defining a Ten-Sector Economy . . . . . . . . . . .. ..148 Margins . . . . . . . . . ...........................150 The Production Recipe . . . . . . . . . . .........151 Intermediate Inputs and Direct Linkages . . . . . .............154 Indirect Linkages. . . . . . . . . . ................155 Networks That Provide Amenities . . . . . . . ....,......159 A Concluding Note . . . . . . ...............................161

PAGE 152

Chapter 4 Defining the Production Recipe The first chapter of this volume proposed a strategy for understanding how complex networks of activity in a modern economy connect work with amenity. The first step in this process (undertaken in Part 1) was to convert demand for amenity into demand for output from specific industries. The second step (the task of the present chapter) is to describe the production recipe that U.S. businesses use to create these goods and services. The accounting method employed to perform this calculation is commonly called input/output analysis. Part IV will examine how these outputs translate into jobs. A concrete example can help to introduce the basic concept and the somewhat baroque vocabulary used in the accounts. One way to begin is by asking where a customers money goes when he buys a frozen pizza. The pizza was made in a factory that ships products to a grocery outlet through a series of intermediaries. The sales price can be divided into two categories: one that generates business for the pizza factory; and another that goes to pay for the insurance, transportation, wholesale, and retail businesses involved in getting the pizza from the factory to the grocery store shelf. These latter costs are called margin s. The question then becomes one of determining the business recipes used by the pizza factory and by the enterprises that provide the margins. These recipes consist of two parts: intermediate inputs and value-added. The intermediate inputs are purchased by a firm from other businesses. For the pizza company this might include ingredients like flour and tomatoes, as well as the advertising, accounting, and legal services needed to conduct business. The difference between the revenue received by selling pizzas and the cost of the intermediate inputs is the value-added generated by the pizza firm. This value-added consists of salaries paid to employees, indirect business taxes (like sales taxes), and profits.1 A business recipe of this sort can be constructed for the grocery store where the pizza was ] More precisely, value-added shown on the input/output tables includes the following categories: employee compensation (a category that includes wages, salaries, and employer contributions for such purposes as pensions and health plans), property type income, and indirect busipurchased, the wholesale operations that supplied the store, and the insurance company and trucking firms that provided other needed services. This is, however, not the end of the story. The intermediate inputs purchased by the pizza firm in turn generate demand for other products. Purchased flour, for example, generates demand for farm production. Farming generates demand for fertilizer. The purchase of a pizza therefore generates indirect demand for the products of the chemical firms that produce fertilizer. Since both the farms and the chemical fertilizer firms have intermediate input expenses as well, they retain only a portion of the consumers dollar as value-added. The challenge, then, is to trace the consumers dollar through this elaborate network of activities in a way that avoids double-counting. A $5 frozen pizza can only generate 5 dollars worth of value-added in the economy. A complete set of production recipes (one that covers all parts of the economy) can be used to estimate the output in each business created by different patterns of consumer and government purchasing using input/output analysis (see box 4-A). The remainder of this chapter uses this method to show how consumer purchases result in activity throughout the U.S. economy. It proceeds in the following steps. First, it assigns each business operation to one of 10 production sectors. This is done so that broad shifts in production recipes can be illustrated. Second, it shows how consumer purchases in different amenity categories translate into demand for different types of business activity (including an estimate of the margins). It then turns to a discussion of the production recipe for each major business type and explores recent trends in the intermediate inputs required. Finally, it uses input/ output methods to calculate the direct and indirect consequences of different types of purchasing. The techniques can be used to show how spending for each amenity category generates value-added directly and indirectly in all areas of the economy, and how the economy is becoming increasingly interconnected. ness taxes (a category that does not include corporate income tax). For more detail on employee compensation, see ch. 11. 747

PAGE 153

148 Box 4=A.Input-Output Analysis The logic of input-output accounts has been recognized since 1758, when they were published as a Tableau Economique by Francois Quesnay, the French economist. Refined and applied to the U.S. economy by Wassily Leontief in the late 1930s, input-output accounts (1-O) form the foundation of most modern econometric models. Leontief was later awarded the Nobel prize in economics. Input-output accounts are used by approximately 70 countries. They incorporate data from all Federal industry censuses and nearly 100 other data sources.1 I-0 accounts are not economic models in the common sense of the term. Rather, they provide a mechanism for displaying and manipulating a large amount of data that has been forced into a consistent format. The central feature of the accounts is a table in which each column represents the production recipe for an industry. In effect, this table represents a series of linear equations that can be solved simultaneously to convert a pattern of final demand to industry output. While I-O accounts have the invaluable feature of making the technology of production an explicit part of an analysis, they do suffer from a number of limitations. Because the model is based on observed data, there tends to be a long lag time between the collection of data and the availability of I-O tables. A benchmark table for 537 business categories is published following publication of the industrial censuses, which are conducted every 5 years. The benchmark table for the 19771-0 tables became available only in 1985. A 1980 summary and revision of this benchmark, published for 85 business categories and available from the U.S. Department of Commerce, is used throughout this analysis. 2 Fortunately, later discussions will indicate that with very few exceptions, the interindustry relationships change extremely slowly. Factors that change more rapidly, such as patterns of consumer and government demand, imports, exports, and investments, are updated each year, and are incorporated in the analysis presented here. I-O data are much more detailed for manufacturing than for service industries, which tend to be lumped into highly aggregated categories. At the detail published in the 1980 I-O accounts there is a separate category for the manufacture of metal boxes, while health, education, and social services are lumped into a single category. The most important assumption made in I-O analysis is that of linear, or constant, economies of scale. Unlike the myriad of assumptions implicit in elaborate econometric analyses, this assumption has the virtue of being simple and clear. Since 1-0 essentially represents an accounting technique, the I-O accounts force the user to construct dynamic characteristics of the economy-assumptions about how production recipes will respond to price changes, to new technologies, and to changes in the scale and scope of industrial organization. The burden of making changes, such as those that will appear chapter 13, thus falls squarely on the user. The links between I-O accounts and the primary data are transparent. See W. Leontief, /nput-Output Economics (New York: Oxford University Press, 1966); and R.E. Miller and P.D. Blair, /rrpuf-Output Anafysis: Foundations and Extensions (Englewood Cliffs, NJ: Prentic&Hall, Inc., 1985). For a description of data sources, see Definitions and Conventions of the 1972 Input-output Study, Bureau of Economic Analysis, U.S. Department of Commerce, pp. B3-B7. U.S. Department of Commerce, Bureau of Economic Analysis, Input-output Tables, 1980, unpublished. A 1981 matrix was released in the Survey of Current Business of January 1987, See Appendix for the algebra of input-output analysis. DEFINING A TEN-SECTOR ECONOMY The analysis used throughout Part 11 can be earful for keeping broad patterns of change in view. ried out by grouping industries into as many or few These 10 sectors have been selected as ones likely sectors as needed. The original data that forms the to be affected in similar ways by changes in techbasis of this report actually used 85 sectors (see the nology, trade patterns, and regulation. The 10 secappendix), but only 9 or 10 will be displayed in the tors, and their relative shares of the U.S. gross naexamples that follow. Summary categories are usetional product (GNP), are shown in box 4-B.

PAGE 154

149 Box 4-B.The 10 Production Sectors % share Description of GNP 1 1. 2. 3-5 3. 4. 5. 6. 7. 8. 9. 10. Natural Resources includes the production of raw materials and energy of all kinds, including the generation of electricity. These industries were singled out to measure the impact of different kinds of economic activity on depletable natural resources, many of which are imported, and at the same time to trace the impact of substitutes for strategic raw materials. Construction is given its own category because of the unique nature and large size of construction activities, and in view of the critical role construction plays in renewing infrastructure and improving productivity throughout the economy. The highly cyclical nature of construction activities also sets this category aside from other business activities. Manufacturing activities have been selected because of growing concern about the future role of these industries in the U.S. economy. Significant direct and indirect linkages exist between manufacturing and the other parts of the economy. Manufacturing has traditionally been the major source of U.S. productivity growth, increasing at twice the rate of the economy as a whole between 1960 and 1983. It is also likely that wage increases in other industries can be traced to productivity growth in manufacturing. Activities within manufacturing have been subdivided into low, medium, and high wage sectors, based on the average level of annual compensation per full-time equivalent employee in 1984. This was done due to concern for the quality of jobs offered by growing and shrinking enterprises; more trade, for example, is attractive if the transactions result in the net substitution of high wage employment for low wage employment. The division by wage levels also provides groupings roughly commensurate with other areas of policy interest. Low Wage Manufacturing is clustered in the traditional apparel, footwear, and furniture industries. Medium Wage Manufacturing contains most enterprises recently tagged as high technology, because these firms conduct significant amounts of research and employ relatively larger numbers of engineers and scientists. It includes industries such as electrical equipment, communications equipment, scientific instruments, and computers, and less technology-intensive industries such as food & kindred products. High Wage Manufacturing is dominated by traditional smokestack industries, such as those that produce motor vehicles, iron and steel, construction machinery, and glass. However, the high wage category also includes such technologically sophisticated industries as chemical production and aircraft manufacturing. Transportation & Trade are clustered because together they form much of the overhead associated with the physical movement of products. These activities are increasingly tied to manufacturing through sophisticated inventory control and dispatching networks. New technologies in transportation will be essential to system-wide improvements in efficiencynot so much from innovations in specific kinds of transportation or retailing equipment, but through advances in information flows that connect production with the marketplace more closely. While many of these technologies are difficult to trace, it appears that dramatic changes may occur in the near future. Transactional Activities deliver financial and information services to businesses. In 1984, the sector generated more value-added than the whole of manufacturing. The activities are clustered because, taken together, they are the most rapidly growing sector in the U.S. economy in terms of output and employment, and are associated with activities in which productivity improvements due to new information technologies could be enormous. Persona/ Services are selected because, with the exception of retailing, they contain most activities traditionally associated with the service sector of the economy: hotels, beauty parlors, and dry cleaning, for example. They also contain most activities associated with recreation and leisurea sector that has grown rapidly in response to rising affluence among many consumers. Social Services follow a unique logic because of the involvement of government. With the exception of government overhead functionsthe sector includes the salaries of both the President of the United States and the authors of this document-most of the activities in this sector, such as public and private health care and education, are delivered directly to consumers. In effect, social services support a human infrastructure. National Defense was separated from the social services of government because it is plainly affected by a unique set of factors. 9% 6% 22% 3% 10% 9% 19% 23% 4% 14% 2% NOTES: The divisions of service functions are modifications of the categories used in J. Singlemann, The %ctoraf Transformation of the Labor Force in %ven Industrialized Countries, 192@1970, American Journal 01 Sociology, vol. 83, No. 3, 1978, pp. 1224-1234. See also the discussion of alternative taxonomies in J 1, Gershuny and I.D Miles, The New Service Economy (New York: Praeger, 1983). The appendix shows the industrial composition of each of these sectors in greater detail. The percentages represent the fraction of all value-added in the economy generated by the sector in 1984. They are calculated in 1980 dollars.

PAGE 155

Finding a taxonomy that reveals rather than obscures deep structural change proves to be a significant challenge. Perceptions of change are often guided by what one chooses to measure. Take the distinction between services and manufacturing. The printing & publishing industry is conventionally counted as a manufacturing enterprise, while television & radio broadcasting is not (presumably because this industry does not produce anything tangible). An individual writing news is considered a manufacturing employee if employed by a newspaper, but is not a manufacturing employee if employed by a television station. Technology has begun to blur many traditional distinctions. Ten years ago, four-fifths of the value of a computer was embodied in its hardware, the remainder being associated with the software. Today, these ratios are reversed.2 Complex patterns of merger and acquisition have further confused the situation. General Motors earned more than onequarter of its profits in 1985 from its finance division, GMAC, the Nations single largest holder of consumer debt. Does that make GM a service industry?3 Chapter 6 will provide many other examples. Zoffice of the U.S. Trzide Representative, U.S. National Study on Trade in Services, Washington, DC, December 1983. 3JameS B. Quinn and Christopher E. &@lOn, will ~rvices Follow Manufacturing into Decline? I-larvard Business Review, November/December 1986, p. 101. MARGINS The cost of wholesale and retail trade, transportation, and insurance can add a significant amount to prices paid by consumers. In some areas, such as clothing, such margins can reach more than 50 percent of the total consumer priced Moreover, technology may rapidly change the role played by margins. Later discussions will show how new technologies and management systems are reshaping the connections between producers and retail outlets. Not only will changing patterns of consumer purchasing reshape the nature of demand for services provided by retail trade, wholesale trade, and transportation, but the net productivity of these systems may change radically. Meeting demand for a series of niche markets expected to change quickly, for example, clearly requires a different retail and delivery network than a system designed to meet demand for a comparatively undifferentiated product competing entirely on the basis of prices 4 Based on the 1977 margins published in U.S. Department of Commerce, Bureau of Economic Analysis, Survey o/Current Business, vol. 64, No. 5, May 1984, p. 46. SFor [he purposes of the analysis presented throughout this document, consumer purchases expressed in the categories of the National income and Product Accounts of the U.S. Department of Commerce, Bureau of Economic Analysis (historic diskettes), are converted to demand expressed in input~utput categories through use of a bridge provided by the U.S. Bureau of Economic Analysis (The 1977 Input-Output Structure of the United States, Survey of Current Business, vol. 64, No. 5, May 1984, pp. 46-49). A similar bridge between government purchases and input-output categories was also made available by the Bureau of Economic Analysis. Table 4-1 indicates the changes in margins that occurred between 1972 and 1984 for the amenity Food. In 1972, for example, groceries (food purchased for off-premise consumption) required a commodity mix with more transportation & warehousing and food & kindred products, but less wholesale & retail trade than it did in 1984.6 The conversion from goods and services needed to satisfy an amenity to commodities results in a rearrangement of the consumption recipe into a consistent set of consumer demands to which industries can respond. Commonly referred to as final demand, this demand stimulates a second round of interindustry intermediate demand for various commodities needed as inputs in the production process. The inputs, which consist of physical materials, services, and the capital and labor required to produce output, are referred to here as the production recipe. 6 The 1984 commodi~ mix is based on the hIPUt-OUtPUt COITlmodity Composition of Personal Consumer Expenditure, published in the Survey of Current Business, vol. 64, No. 5, May 1984.

PAGE 156

151 Table 4-1.The Commodity Mix for Food in 1972 and 1984 (current dollars, in percent) Food purchased for off-premise Purchased meals Other consumer consumption & beverages food purchases Total Industry 1972 1984 1972 1984 1972 1984 1972 1984 Transportation & warehousing . 3.60/o 2.40/o 0.0% 0.1 %0 0.90/0 0.7% 2.60/o 1.7% Wholesale & retail trade . . 29.5 32.7 0.0 0.0 41.3 40.6 23.9 25.1 Finance & insurance . . . 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Livestock and livestock products 0.9 1.1 0.0 0.0 1.3 1.2 0.8 0.8 Other agricultural products. . 3.1 3.3 0.0 0.0 2.0 1.4 2.3 2.3 Food & kindred products . . 62.1 60.1 0.0 0.0 13.9 17.4 43.8 41.6 Tobacco manufacturers. . . 0.0 0.0 0.0 0.0 40.6 38.7 3.6 3.1 Eating & drinking places . . 0.0 0.0 100.0 98.4 0.0 0.0 22.4 24.7 Amusements . . . . . 0.0 0.0 0.0 1.4 0.0 0.0 0.0 0.4 Other. . . . . . . 0.8 0.4 0.0 0.0 0.0 0.0 0.5 0.3 Total . . . . . . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Total (billions of dollars) . . 115 183 37 69 15 22 167 273 HOW T O Read This Table: On average, $100 in groceries purchased in 1984 (food purchased for off-Premise consumption) was allocated as follows: $2.40 went to transportation and warehousing businesses, $32.70 went to grocery stores and other wholesale and retail businesses, and $60.10 went to firms that supply food to the system. A significant fraction of the $60.10 received by these firms, of course, went to pay their suppliers. For an estimate of these intermediate inputs, see table 4-2. NOTES: Totals may not equal 100 percent due to rounding. aThe 1~ Commodity Mix is estimated using the 1977 composition Updatecl with 1984 demand fOr these products. SOURCE U.S. Department of Commerce, Bureau of Economic Analysis, The Input-Output Structure of the U.S. Economy, Survey of Current Business, April 1979 for 1972 data, and May 1984 for 1977 data; National Income and Product Accounts, Survey of Current Business, table 2.4, July 1987. THE PRODUCTION RECIPE The production recipes for 9 of the 10 sectors listed earlier are summarized in table 4-2 for the years 1972 and 1980. The table indicates that the 1980 recipe for making $100 of output in the Construction sector involved: l total intermediate input purchases of $56.80: $1.20 in products from the Natural Resource industries, $0.10 from other Construction businesses, $7.30 from manufacturing enterprises paying low wages, $20.40 from Medium Wage Manufacturing businesses, etc.; and l $43.20 in the form of wages paid and returns to capital in the Construction sector. Table 4-2 exposes expected differences in the production recipes of different business categories. The value-added generated by manufacturing firms in 1980 accounted for between 29.6 and 36.5 percent of total manufacturing output, while intermediate purchases-those goods and services purchased from other businesses that become part of a final manufactured productaccount for the rest. On the other hand, in the relatively labor-intensive fields of Transportation & Trade, Transactional Activities, and Personal Services, value-added accounts for 59.0 to 72.8 percent of total sectoral output. Table 4-2 includes only direct intermediate purchases with expected lifetimes of less than a year. It does not include purchases of buildings or capital equipment needed to expand operations or replace old machinery. 7 But capital equipment is clearly an essential ingredient in any production recipe. The National Accounts refer to items that last more than a year as Gross Private Fixed Investment (GPFI). This category includes purchases of residential and nonresidential structures, and of capital equipment such as machine tools, computers, and tractors. Components of GPFI include accounts for items bought as both replacements for older equipment and equipment purchased for expansion. Purchases of producers durable equipment (PDE) are nearly half of the 1984 GPFI total. 8 The remainder consists primarily of non-residential and residential structures, which each represent about 7 A more detailed examination of the role of capital flows in a dynamic model of the economy has been developed. See W. Leontief and F. Duchin, The Impacts of Automation on Employment, 1963-2000, Contract #PRA-801 2844 to the National Science Foundation, Washington, DC, April 1984, p. 2.1. 8 Unless otherwise noted, data about PDE comes from table 5.7, Private Purchases of Producers Durable Equipment by Type in Constant Dollars, in National Income and Product Accounts, op. cit., footnote 5.

PAGE 157

Table 4-2.Recipes for the Production Sectors I Natural Low Wage Medium Wage High Wage Transport Transactional Personal Social Production sector Resources Construction Manufacturing Manufacturing Manufacturing & Trade Activities Services Services The 1980 Direct Requirements Table Natural Resources . . . . . . . 27.2% 1 .2% 4.4% 12.6% 20.4% 2.4% 1 .2% 3.4% 2.6% Construction . . . . . . . . 2.5 0.1 0.5 0.6 0.7 1.2 3.6 1.4 2.2 Low Wage Manufacturing . . . . . 0.4 7.3 27,3 2.0 2.3 0.7 0.2 1.9 0.7 Medium Wage Manufacturing . . . . . 3.9 20.4 2.8 18.6 5.9 5.1 1.6 5.0 2.6 High Wage Manufacturing . . . . . 8.8 6.8 15.4 15.0 29.0 5.5 1.1 6.7 3.2 Transportation & Trade . . . . . . 4,1 10.5 7.5 8.6 7.7 8.8 2.0 7.1 2.5 Transactional Activities . . 5.7 9.9 4,6 5.2 3.7 12.6 15.6 10.1 5.9 Personal Services . . . . . . . 0.5 0.5 0.6 0.7 0.5 2.2 1.1 4.8 0.8 Social Services . . . . 0.3 0.1 0.4 0.4 0.2 0.5 0.9 0.6 1.2 Total intermediate inputs... . . . . 53.3 56.8 63.5 63.7 70.4 39.2 27.2 41.0 21.7 + Value-added . . . . . . . . 46.7 43.2 36.5 36.3 29.6 60.8 72.8 59.0 78.3 Total . . . . . . . . . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 The 1972 Direct Requirements Table (1972 I-0 in 1980 dollar in percent) Natural Resources . . . . . 26.l% 1.1% 6.0% 14.2% 18.2% 2.9% 1.4% 2.9% 2.5% Construction . . . . . . . 1.7 0.0 0.3 0.3 0.5 1,2 4.2 1.1 1.8 Low Wage Manufacturing . . . . 0.4 6.8 30.4 1.9 2.4 0.5 0.4 2.9 0.5 Medium Wage Manufacturing . . . 3.0 19.3 3.7 16.7 6.2 5.5 1.6 4.7 1.5 High Wage Manufacturing . . . . 6.4 10.8 15.2 17.0 26.8 5.4 2.1 11.2 2.1 Transportation&Trade . . . . 2.4 8.9 6.9 6.5 5.4 6.9 2.2 4.5 1.7 Transactional Activities . . . 4.8 5.2 4.4 4.9 3.5 10.2 16.1 10.1 4.4 Personal Services . . . 0.3 0.4 0.7 0.8 0.5 2,1 1.6 4.6 0.6 Social Services . . . . . 0.2 0.1 0.4 0.4 0.2 0.6 1.1 0.3 0.8 Total intermediate input . . . 45.3 52.5 68.1 62.6 63.6 35.4 30.6 42.2 16.0 + Value-added . . . . . . 54.7 47.5 31.9 37.4 36.4 64.6 69.4 57.8 84.0 Total ., . . . . . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 How To Read This Table:On average, each $100 in sales made by a Natural Resource business in 1980 resulted in $46.70 paid as compensation to employees in Natural Resource businesses or payments to the owners of these businesses, or indirect business taxes paid by these businesses, collectively called value-added. The remainder of the $100 in sales went to pay firms that directly supplied Natural Resource industries. $2.50 went to purchase goods and services from construction firms, $0.40 went for purchases from Low Wage Manufacturing firms, $3.90 from Medium Wage Manufacturing, etc. These estimates reflect only the direct inputs, not the secondary or indirect contributions which are estimated in table 4-4. NOTE: Totals may not equal 100 percent due to rounding. SOURCE: Office of Technology Assessment, from data provided by the U.S. Department of Commerce, Bureau of Economic Analysis, The Input-Output Structure of the U.S. Economy, 1972, Survey of Current Business, April 1979 and Input-Output Tables, 1980, unpublished; U.S. Department of Labor, Bureau of Labor Statistics, Time Series Data Base for Input-Output Industries, unpublished.

PAGE 158

153 one-quarter of GPFI. Residential structures are not considered to be part of the production recipe and expenditures on non-residential structures have remained relatively constant, leaving PDE as the component of capital investment that has experienced the most change. Most of this change has been in the form of purchases of information processing equipment, which increased from 6 percent of PDE in 1950 to over 40 percent of all such investment in 1986, with most of the growth occurring since 1973 (see figure 4-l). Two-thirds of the increase was attributable to two categories of equipment: office, computing, & accounting machines; and communications equipment. Of the growth in office, computing, & accounting machines, 93 percent occurred during the last 10 years. Recent estimates attribute 77 percent of all of the office, computing, & accounting machines and 95 percent of the communication equipment expenditures to purchases made by the service sector. g The other components of PDEindustriaI, transportation, and other equipmentare of roughly Stephen S. Roach, The Information Economy Comes of Age, /formation Management Review, vol. 1, No. 1, summer 1985. Figure 4-1.-Share of Producers Durable Equipment (constant 1982 dollars) Percent 50 ] 1 40 30 20 10 Information Equi p Industrial Equip Transportation Equip Other Equipment How To Read This Figurs: Information equipment (computers, photocopiers, communication equipment, instruments, and related equipment) were only 6/0 of all business investment in producers durable equipment in 1950, but were 40/0 in 1986. The percentages are computed after converting all spending to constant dollars. Producer durables were almost exactly half of all private investment in 1986, with the remaini ng investment going to buildings and other structures. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 5,7. equal size and all have lost about the same share of total expenditures, falling from about one-third in 1950 to about 22 percent in 1985. In standard accounts, GPFI is treated in the same manner as consumer purchases. Unless otherwise stated, however, the calculations presented in this analysis will treat both durable and non-durable elements of production as being part of the production recipe, Purchases of residential structures were examined in chapter 2 and are not included in the production recipe because they are not considered an input to production. There are, of course, powerful links that cannot be exhibited in a table such as 4-2. The health of U.S. research and development efforts, for example, may be badly hurt if manufacturing capacity in an industry moves offshore, since there is evidence that commercially useful research thrives when it is integrally connected with practical manufacturing prob lems.l Likewise, every business relies on educated workers as an input, but schools are not formally linked as an input in the production process. Such connections must be recognized using tools other than the ones presented here. The table also reports all value-added as a single statistic, lumping labor costs of managers, production workers, and scientists together with the cost of capital. In fact, some of the most interesting changes in production recipes are occurring within these value-added categories. Chapter 10 will explore the value-added recipe in much greater detail. As prices, technologies, regulations, and other factors change, so will the recipe used for production. Generally, the process of changing this recipe is slow and gradual; even a large shock such as the quadrupling of oil prices between 1972 and 1980 had a long lag period before an adjustment was incorporated into the production process.l ] Nevertheless, comparing the 1972 production recipes with those of 1980 reveals that in almost every sector, the level 10 For a di~ussion of this link, see Stephen S. Cohen and John ZYSman, The Myth of a Post-Industrial Economy, Technology Review, February/March 1987; or Charles F. Sabel, et al., How To Keep Mature Industries Innovative, 7%chno~ogy Review, April 1987. I Isee Anne Carter, changes in Input-Output Structure Since 1972, Interindush-y Review, Data Resources Inc., summer 1980, p. 1.16; and Stanley J. Feldman and Karen Palmer, Structural Change in the United States: Changing Input-Output Coefficients, Business Economics, January 1985, p. 39.

PAGE 159

154 of inputsparticularly service sector products increased between 1972 and 1980. Chapter 6 provides a detailed discussion of the ways production recipes are changing. Major changes can For l l l l be found in virtually every business network. example: Heating oil dealers have been able to cut down on the number of trucks and drivers, and on their inventories, by maintaining data on the capacity of customers tanks, consumption rates, and the weather. The recipe has been changed, substituting information for vehicles, people, and storage facilities. Levi Strauss & Co. uses 3-D computer imaging equipment to simulate how different fabrics and styles will look and wear, instead of producing samples. Software and computers have replaced sewing machines and material. Advances in polymer technology and a desire to boost fuel efficiency have altered the recipe for producing a car, as high strength plastics and aluminum are substituted for iron and steel. Computer-Aided-Design and robotics have optimized designs so that less steel is used. Banks have used automated tellers and communication networks to substitute for many hand operations. In all of these cases, changes in the recipe of production, whether as a result of price changes or technological innovation, have altered the respective positions of industries in the U.S. economy.l2 Intermediate Inputs and Direct Linkages As table 4-2 indicates, the manufacturing sectors make comparatively heavy use of intermediate inputs purchased from other businesses. Roughly twothirds of the value of manufacturing sales must go to pay for intermediate inputs supplied by other sectors, while the three sectors retain significantly less than half of the sales price of their goods. In comparative terms, the Social Services sector is only weakly linked to other parts of the economy. Of the price of Social Services sold, 78.3 percent results in direct value-added (returns to capital and labor) to the Social Service industries themselves; not much leaks out to the other sectors of the economy. Indeed, this insular quality holds true for all the sectors that are characterized as services. Taken as a whole, the economy became more interconnected between 1972 and 1980, as the share of goods and services produced for use as intermediate inputs rose by 1.2 percent. This translates into more than $60 billion (1980 dollars) if applied to the 1980 economy, or more than all the value-added generated by the eating& drinking industry in that year. 13 Some sectors became much more highly linked during the 1972-1980 period. The Natural Resource and High Wage Manufacturing sectors made much heavier use of intermediate inputs, increasing their intermediate inputs by 8 and 7 percentage points, respectively. The bulk of the increase for both of these sectors occurred between 1972 and 1977, and was in the form of energy-related commodities: crude & refined petroleum, chemicals, and electric, gas, and water services (utilities). Nevertheless, the single biggest increase in an input for the High Wage Manufacturing sector was the service provided by the wholesale & retail trade industry. The change in use of intermediate inputs is not uniform. For example, the Natural Resource sector greatly reduced inputs of livestock and agricultural products inputs while High Wage Manufacturing reduced its purchases of iron ore, steel products, and metal containers. Transactional Activities and Low Wage Manufacturing actually reduced their intermediate inputs between 1972 and 1980, becoming less tightly linked to the rest of the economy. In the case of Low Wage Manufacturing, the bulk of the decrease was attributable to a decline in textile and apparel inputs; for Transactional Activities, the decrease was more evenly spread between paper products, real estate, and maintenance & repair construction. In many cases, relatively small total changes mask significant offsetting changes in production recipes. Izsince many Of the sectors reflect the combination of many industrial processes, changes in demand, which necessarily change the share of a particular product or industry within the broader sector, may appear to be a change in recipe at this higher level of aggregation. lsThiS is ().()12 times $5,210 billion (total gross output in 1980 in 1980 dollars).

PAGE 160

155 For example, intermediate inputs for Construction grew only 4 percentage points. Construction, however, greatly increased its purchases of business service inputs while reducing demand for refined petroleum. The Transportation & Trade sector also increased its use of business services, as well as communication services. Overall, Medium Wage Manufacturing increased its use of inputs only slightly, but its inputs of wholesale& retail trade, electronic components, and office & computer equipment increased dramatically, while its use of primary iron & steel, livestock products, and food& kindred products decreased. When these sectors are combined and their relative size is taken into account, it becomes clear that the Nations production recipe has undergone a significant realignment, using more service inputs and fewer raw or semi-finished materials.1 4 Table 4-3 lists the 10 industries which contributed most to an increase in intermediate inputs and the 10 that contributed most to declines between 1972 and 1980. llsee Andrew G. C]em and William P. Thomas, New Weight StrUcture Being Used in Producer Price Index, Monfh/y Labor Review, August 1987, pp. 12-21 for a similar analysis. Table 4-3.Changes in Production Recipe Inputs From 1972 to 1980 (ranked by greatest gain and loss) Industries gaining share 1, Wholesale & retail trade 2. Business services 3. Communications 4. Electronic components & accessories 5. Maintenance & repair construction 6. Electric, gas, water, & sanitary services 7. Transportation & warehousing 8. Plastic & synthetic materials 9. Office, computing, & accounting machines 10. Finance & insurance Industries losing share Primary iron & steel manufacturing Livestock & livestock products Other agricultural products Primary nonferrous metals manufacturing Motor vehicles & equipment Stone & clay products Forestry & fishery products Paper & allied products Lumber & wood products Broad & narrow fabrics How To Read This Table: Intermediate inputs from wholesale and retail trade were responsible for the largest share of the total increase in intermediate inputs occurring between 1972 and 1980. Changes in intermediate inputs of primary iron & steel manufacturing contributed most to counteract this increase. SOURCE: US. Department of Commerce, Bureau of Economic Analysis, The Input-Output Structure of the U S. Economy, 1972, Survey of Current Business, April 1979 and Input-Output Tab[es, 1980, unpublished; converted to a constant dollar basis using Bureau of Labor Statistics Output Deflators, unpublished; U.S. Department of Labor, Bureau of Labor Statistics, Time Series Data Base for Input-Output Industries, unpublished The overall increase in the use of intermediate inputs and the pronounced rise of service sector inputs suggests that the economy has become more specialized. This specialization is apparent in the fact that more interindustry transactions are taking place, requiring significant increases in wholesale & retail trade inputs as well as inputs that tend to facilitate transactions: communications and business services. Much of this specialization is fueled by technological developments and competitive pressures that make it nearly impossible, both technically and financially, for any one firm to conduct all facets of production. Specialized contractors fill this void, providing the contracting firm with additional flexibility because costs are shared, but also making the firm more dependent because of the strategic position of the contractor in the production process. Chrysler cites a 50-percent decrease in engineering costs to its increased use of suppliers. ]G Much of the increase in interindustry connections is also due to new technologies, particularly information processing technologies, that allow coordination of complex production processes which increasingly span the globe. As a result of this growth in subcontracting, and the wider geographical dispersion it entails, service sector businesses have thrived because of the increased need for legal contracts, consulting services, transportation, communication, and wholesale & retail trade. 17 Indirect Linkages Although the analysis of direct input requirements shows how recipes have changed, it does not include the indirect economic activity generated by an industrys output. As in the frozen pizza example, grain is needed to make flour, fertilizer is used to produce grain, chemicals are needed to make fertilizers, and so on. Numerous upstream and downstream linkages are associated with almost every commodity, although some are more tightly linked than ls~e Anne cafler, s~rUC~U~a/ Change in the American J%or?orny (cMIIbridge, MA: Harvard University Press, 1970), for a more detailed analysis of specialization. lcElizabeth A. Haas, Breakthrough Manufacturing, ~arvard ~USjrres.s Review, March/April 1987, p. 79. lTChar]es F. Sabe], et al., op. cit., footnote 10, p. 32; and Richard McKenzie, The Emergence of a Service Economy: Fact or Artifact? Policy Analysis No. 93, Cato Institute, Washington, DC, October 1987.

PAGE 161

156 others. By tracking these connections, the impact of buying a new domestically produced car on the auto industrys suppliers (like the steel industry) becomes apparent, as does the secondary impact of those increased steel purchases on the steel industrys suppliers (like coal). The indirect connections between different parts of the economy, implicit in table 4-2, can be illustrated using mathematical techniques described in the appendix to estimate how $1 of output in one industry sector generates value-added in other sectors, incorporating both direct and indirect effects. Table 4-4 reconfirms that the Social Services sector is the most insular, even when indirect effects are taken into consideration. Of the total value-added generated throughout the U.S. economy by demand for Social Services, 77 percent is retained within the Social Service sector-only 23 percent spills over into the rest of the economy (e.g., for purchase of hospital or school supplies). As was the case when only direct effects were examined, this insular quality holds true for all the sectors characterized as services when indirect effects are included. On average, almost two-thirds of the value-added generated from demand for services stays in the service sectors. Their linkage to other sectors is relatively weak; of every $1 of value-added generated by demand for services, only around 15 cents ends up in manufacturing. The biggest spillover to manufacturing, more than 16 cents on each dollar, comes from demand for Personal Services; the smallest, 8 cents, comes from demand for Social Services. A very different situation prevails in manufacturing, where the three sectorsas beforereap under half of the value-added generated from the sale of their goods, passing the other half on to other sectors of the economy. Low Wage Manufacturing is the most insular of the three, but even in this case, nearly one-quarter of every dollar spent on goods produced by the Low Wage sector, such as apparel, ends up in the coffers of a service industry-especially Transportation & Trade (12 cents per dollar) and Transactional Activities (9 cents). This strong link to services is also evident in the Medium Wage sector; about one-quarter of the value added generated by demand for the products of Medium Wage Manufacturing is captured by services. But in contrast to the Low Wage sector, the demand for $1 of Medium Wage products, such as electronics or processed food, also translates into 12 cents of value-added for the Natural Resource sector. The other major link to Natural Resource industries other than the Natural Resource sector itselfis High Wage Manufacturing. Only about 43 cents of every dollar of value-added resulting from demand for High Wage manufactured goods is retained by the High Wage sector. The bulk of the remaining 57 cents is divided between Natural Resources, reaping 17 cents, and Medium Wage Manufacturing, Transportation & Trade, and Transactional Activities, each gaining more than 8 cents worth of value-added. The Natural Resource and Construction sectors are the least insular of all. Only 36 cents of $1 in value-added generated by demand for Construction goes to construction firms, while five of the eight other sectors gain at least a nickel in value-added from such purchases. About one-quarter of the value-added that is not retained by Construction from demand for its own products is split between Transportation & Trade and Transactional Activities 4 and 12 cents, respectively. The Natural Resource sector is the least insular sector, reaping only 25 cents of the dollar of value-added that results from purchases of Natural Resources. Major beneficiaries from demand for this sectors products are Transportation & Trade (16 cents) and Transactional Activities(11 cents). Table 4-5 provides a summary view of the way that each of the sectors is linked to the rest of the economy. The table estimates the total industry output resulting from $1 of demand in each of the 10 sectors. This statistic is conventionally called the output multiplier. laone do]]ar of sales can, by definition, generate oniy $1 of value-added. If an industry purchases no items from other industries, there would be no direct or indirect links to the rest of the economy, and $1 of its output would generate only $1 of total output (namely its own). All value-added would be captured by the producing enterprise. If a large portion of the value in an industrys sales represents the cost of goods and services purchased from other enterprises, only a fraction of the total value-added in the economy generated by the value of its sales will remain in the industry itself. The total output generated by $1 of the industrys output will be larger than 1.0 because the intermediate outputs are counted at least twice (once as a part of the output of the industry itself, and once as the output of the supplying industries). A chain of connections thus leads to multiple counting when an industry is not completely insular. The extent of this multiple counting provides a good measure of the degree to which the industry is linked

PAGE 162

Table 4-4.Sectoral Linkages: Value-Added Derived by Production Sector From the Purchase of $100 of a Sectors Product a (1980 dollars, in percent) Natural Low Wage Medium Wage High Wage Transport Transactional Personal Social Production sector Resources Construction Manufacturing Manufacturing Manufacturing &Trade Activities Services Services Natural Resources . . . . . . . 25.3% 6.1% 6.8% 11.8% 16.9% 5.6% 3.6% 4.9% 2.5% Construction . . . . . . . . 10.5 35.9 3.3 3.9 4.3 3.8 12.7 3.9 2.7 Low Wage Manufacturing . . . . . 2.3 4.7 47.6 2.4 2.7 1.6 2.1 2.5 1.0 Medium Wage Manufacturing . . . . . 13.2 14.8 7.7 44.5 12.2 6.9 8.9 7.4 4.2 High Wage Manufacturing . . . . . 12.9 10.5 11.8 12.4 43.4 5.7 5.3 6.5 3.0 Transportation&Trade . . . . . . 15.9 13.8 11.6 13.2 10.3 62.1 7.4 8.8 3.8 Transactional Activities . . . . . 11.4 12.2 8.8 9.4 8.2 11.0 57.5 9,7 5.3 Personal Services . . . . . . . 1.3 1.1 1.1 1.3 1.0 1.7 1.0 55.0 0.7 Social Services. . . . . . . . 7.1 0.9 1.2 1.2 0.9 1.6 1.4 1.3 76.9 Total . . . . . . . . . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 How To Read This Table: Following all the direct and indirect effects, $lOO of products purchased from Natural Resource businesses in 1980 created $25.30 invalue-added in the Natural Resource businesses themselves, $10.50 in Value-added for the Construction industry,etc. The $loo purchase can only generate $l00 in value-added throughout the econorny. it can, however create much more than $100 in output. Table 4-2 looked at direct effects of $100 in output produced by the Natural Resource industries. Table 4-4 looks at the chain of events created by $100 in final demand for commodities produced by Natural Resource industries. aBased on the distribution of demand as it existed in 19S4, includes 1977 capital flows table updated to 1984 levels. NOTE: Totals may not equal 100 percent due to rounding. SOURCE: Office of Technology Assessment, from data provided by the U.S. Department of Commerce, Input-Output Tables, 1980, unpublished, Bureau of Labor Statistics, Time Series Data Base for lnputOutput Industries, unpublished, and 1977 Capital Flows Table.

PAGE 163

. 158 Table 4-5.Changing Sectoral Linkage: Output Multipliers for 1972 and 1980 (includes capital equipment) Production sector 1972 1980 High Wage Manufacturing . . . . 2.6 3.0 Natural Resources . . . . . . 2.3 2.8 Low Wage Manufacturing .................2.9 2. 8 Medium Wage Manufacturing . . . .2.6 2.8 Construction . . . . . . . 2.5 2.6 Transportation&Trade ...................2.0 2. 1 Personal Services . . . . . .. 2.3 2.1 Transactional Activities . . . . . 1.8 1.7 social Services . . . . . . 1.5 1.6 Total . . . . . . . . 2.3 2.4 How To Read This Table: In 1972, $100 in final demand for commodities produced by Natural Resource industries created $260 in total output throughout the economy. This had risen to $300 in 1980 because the sector had become more highly linked with the rest of the economy. A completely insular business purchasing nothing from the outside would create $100 in output and $100 in value-added for $100 in sales. A highly linked business might buy a $90 product from suppliers and sell it for $100. This creates at least $190 in total output for the economy. Adding output inherently involves double-counting because each output contains the value of intermediate inputs. SOURCE: Office of Technology Assessment, from data provided by the U.S. Department of Commerce, Bureau of Economic Analysis, The lnputOutput Structure of the U.S. Economy, 1972, Survey of Current Business, April 1979 and Input-Output Tables, 1980, unpublished; U.S. Department of Labor, Bureau of Labor Statistics, f Time Series Data Base for Input-Output Industries) unpublished, 1972 and 1977 Capi. tal Flows Table. It is again evident that service enterprises are comparatively independent of the rest of the economy, while the Natural Resource and manufacturing sectors are highly linked. The changes in output multipliers between 1972 and 1980 tend to parallel the changes in production recipe: Natural Resources and High Wage Manufacturing had the largest increases, while Low Wage Manufacturing, Transactional Activities, and Personal Services experienced a decrease. In other words, Transactional Activities has altered its operations so that this sector had less interaction with other sectors in 1980 than in 1972. Presumably, industries becoming more tightly linked with the rest of the economy require more connections to other business or more expensive inputs. They may also be contracting for work or buying inputs that were previously generated incontinued from previous page directly and indirectly to other enterprises. See J.M. Szyrmer, Measuring Connectedness of Input-Output Models: Survey of Measures, Erwiror?ment and P/arming, vol. 17, 1985, pp. 1591-1612, for a discussion of the use of this measure. house,l9 The analysis of High Wage Manufacturings production recipe indicates that the bulk of these additional links are with tradeand energyintensive sectors, suggesting that the increase in connectivity is due to the higher cost of energy and the increased number of transactions needed for productionpossibly due to the use of imported inputs.2o Table 4-5 also shows that the output multiplier for the whole economy increased between 1972 and 1980, The U.S. economy has become more interconnected in spite of the fact that sharp growth has been experienced in service sectors, which exhibit comparatively weak links to the rest of the economy. These seemingly contradictory findings are explained by the fact that businesses are increasingly using services as an input into their production recipes, both directly and indirectlythus increasing services role in the economy and creating a higher level of interindustry linkage. Increased linkages have both attractive and unattractive effects. A more tightly linked economy allows a greater degree of specialization, flexibility, and efficiency. Nevertheless, interdependent sectors also mean that an economic downturn in one sector will quickly spread to other sectors of the economy. Unlike the situation in manufacturing, where growth in demand for its products has also meant growth in the Natural Resource and service sectors, the growth of services-especially that of Transactional Activitiesresults in little growth outside transactional businesses. On the other hand, the health of transactional service businesses may depend heavily on a healthy manufacturing sector. z The concept of linkage takes on greater importance when viewed in terms of international trade, an issue that will be addressed in chapter 8. Trade has a major effect on manufacturing, which is highly linked to the rest of the economy. Therefore, trade problems in manufacturing ripple throughout the U.S. economy. IgSee John Tschetter, Producer Services Industries: Why Are They Growing So Rapidly? U.S. Bureau of Labor Statistics, Monthfy Labor I?eview, December 1987, pp. 31-41. ZOFor example, the big three U.S. auto companies now draw a Significant amount of their parts from foreign producers. See Kevin Flaherty, Foreign Sourcing by the U.S. Automobile Industry, U.S. Congressional Research Service, Nov. 8, 1985. Z] Stephen S. Cohen and John Zysman, The Myth of a Post-Industrial Economy, op. cit., footnote 10.

PAGE 164

159 NETWORKS THAT PROVIDE AMENITIES The methods just described can be used to show how the set of consumer and government purchases needed to satisfy specific amenities described in Part I connect to direct and indirect demand for industry output. The chain of analysis proceeds as follows: 1. consumers purchase a variety of goods and services to achieve an amenity, 2. these purchases translate into direct demand for output from a variety of different industries (including businesses providing margins such as transportation, insurance, and trade), and 3. the value of the products and services sold to consumers is distributed across a complex network of businesses because of direct and indirect connections. It is possible, therefore, to connect purchases for an amenity such as Food with economic activity distributed throughout the U.S. economy. The distribution of value-added generated in each industry by purchases needed to serve 11 major amenity groups are summarized in table 4-6 for 1984 and 1972. This table shows how all value-added flowed through the economy from producing sectors to amenity categories in each of the 2 years. The sum of all value-added in each year is the entire U.S. gross national product. Table 4-6 indicates, for example, that including all direct and indirect effects, approximately 15 percent of the U.S. Food bill goes to purchase value-added from farms and other Natural Resource operationsdown from 17.6 percent in 1972. Similar decreases in share came from the Medium Wage and High Wage Manufacturing sectors. The bulk of the value-added required for the Food amenity, increasingly true since 1972, comes from the Transportation & Trade sector. Although Natural Resource inputs were above average for the Housing, Transportation, and Export amenities both in 1972 and 1984, the share of value-added contributed by the Natural Resource sector has declined in 9 of the 11 amenities since 1972. Interestingly, the largest share of value-added required for Housing, both in 1972 and 1984, does not go to the Construction sector. Rather, Transactional Activities accounts for nearly 45 percent, due to the enormous impact of the real estate industry on homebuying. 22 This is up dramatically from 1972. Exports also require a larger share of inputs from the Transactional sector as does Personal Business and Communication. Between 1972 and 1984, the increased spending for Personal Business and Communication translated directly into value-added for Transactional Activities, while every other production sector lost shareparticularly Construction and High Wage Manufacturing. Indeed, the importance of Transactional Activities has increased sharply since 1972 for every amenity except Defense, which is sensitive to the choice of 1972 as an endpoint due to the influence of the Viet Nam War. Not surprisingly, value~added from the Social Service sector plays a large role in creating the Health, Education, Government, and Defense amenities, and this role has increased in size compared to 1972. All amenity groups except Defense and Personal Business and Communication registered increased demand for inputs from Transportation & Trade. The Food, Transportation, Clothing and Personal Care, Recreation and Leisure, Federal Defense, and Export amenities make the heaviest use of manufactured inputs. With the exception of Defense, however, the share of value-added contributed by manufacturing declined significantly between 1972 and 1984. Purchases from High Wage Manufacturing fell sharply, especially for Transportation, Clothing and Personal Care, and Exports. IZThe Nationa] Income and Product Accounts also impute a rentalequivalence value for homeowners, which is allocated to the real estate industry.

PAGE 165

Table 4-6Networks That Provide Amenity: Contributions From Sectors Needed To Satisfy Amenity categories, 1964 and 1972 (1960 dollars, in percent) Amenity categories Personal Clothing and Business and Recreation Government Federal Production sectors Food Housing Transportation Health Personal Care Education Communication & Leisure n .e.c. Defense Exports Total 1984~ Natural Resources . . . 15.0% 9.7% 14.5% 4.3% 4,9% 4.0% 2.6% 6.0% 5.2% 4.4% 16.4% 9.1 % Construction . . . . 3.3 12.9 6.1 3.7 2.4 5.2 2.8 3.7 11.0 3.8 3.4 6.2 Low Wage Manufacturing . 1.5 3.2 2.7 1.5 17.0 1.2 1.1 3.5 1.9 1.4 3.8 3.2 Medium Wage Manufacturing 16.8 6.8 7.8 6.6 5.2 4.8 6.7 12.1 6.1 10.9 19.4 9.7 High Wage Manufacturing . 8.1 5.7 16.1 5.9 7.6 3.3 2.9 7.1 5.0 17.6 19.5 8.7 Transportation & Trade ., . 39.1 12.8 30.1 10.3 39.1 4.1 6.0 21.7 8.0 8.1 18.8 19.3 Transactional Activities . . 12.7 44.7 12.3 15.6 12.8 7.0 70.9 15.4 12.4 9.0 16.1 23.5 Personal Services. . . . 1.6 2.3 8.0 1.4 10.0 0.6 3.2 14.4 1.4 1.2 1,5 3.7 Social Services. . . . 1.8 2.0 2.3 50.8 1.1 69.9 3.7 16.2 49.1 43.4 1.1 16.4 Total . . . . . .. 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Total (billions of 1980$) . . 428 672 264 285 171 180 161 200 120 173 238 2,892 1972 b Natural Resources . . . 17.6% 13.8% 13.1 % 5.2% 5.7% 6.1 % 3.7% 6.7% 7,4% 20.1 % 15.9% 12.5% Construction . . . . 3.7 12.2 7.1 5.6 4.1 6.1 5.1 5.5 12.8 9.8 3.9 7.6 Low Wage Manufacturing . 1.5 3.8 2.9 1.7 18.3 1,3 1.6 4.1 2.0 3.5 4.1 3.8 Medium Wage Manufacturing 19.8 7.0 8.0 6.3 5.5 4.8 8.1 12.3 6.6 10.2 17.2 10.0 High Wage Manufacturing . 10,7 8.4 22.7 8.0 11.3 5.1 5.4 9.6 8.0 14.6 27.9 12.0 Transportation & Trade . . 34.4 12.1 27.0 10.1 33.9 3.8 6.7 21.2 7.3 14.4 17.0 18.3 Transactional Activities . . 9.6 37.8 10.2 13.9 10.3 5.7 61.6 11.7 11.0 11.7 11.6 18.1 Personal Services. . . . 1.2 2.9 6.3 1.2 9.8 0.5 4.0 12.4 1.3 2.0 1.3 3.5 Social Services, . . 1.5 2.1 2.6 48.1 1.1 66.5 3.8 16.3 43.6 13.8 1.1 14.3 Total . ., .............100.0 100.0 100,0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Total (billions of 1980$) . . 374 509 263 191 161 168 104 160 126 377 131 2,564

PAGE 166

161 A CONCLUDING NOTE This chapter provides a set of tools for viewing years, and are the basis for speculating about future the economy as a series of interconnected networks, changes. This will be the task of chapter 5. The tools where the product of one sector works in conjuncalso provide a way of describing the operation of tion with the products of another sector to satis~ complex networks of business activities that must the needs of a consumerwhether that consumer combine operations to deliver goods and services is a person, a business, or a government agency. that consumers require for different amenities. This These tools can be used to explore many of the will be the task of chapter 6. structural changes that have occurred in the past few

PAGE 167

Chapter 5 Three Perspectives on Structural Change

PAGE 168

Constant Selecting Sectoral Contributions to GNP. . . and Current Dollars . the Categories --- -Recent Trends in Sector Explaining the Changes Tracing the Significance for Measuring Shares. . . . . . . Sectoral . . . of Structural Change Changes in Scale and Scope . . . . Describing Changes in Scale and Scope . Forces of Change . . . . . Measuring Changes in Scale and Scope . . . Consequences of Changes in Scale and Scope. The Geography of Economic Activity . . Regional Movements. . . . . . The Move to the Urban Fringes . . . . . . . . . . . . . . . . . . Share . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Page 65 66 67 68 72 ,77 .77 .79 80 82 186 :90 ,91 96

PAGE 169

Chapter 5 Three Perspectives on Structural Change Structural changes in the U.S. economy fall into three major categories: 1. changes in the contribution that each major business sector makes to the total value added in the economy, or the gross national product (GNP); 2. changes in the organization of firms measured in terms of their scale of operation and scope of production; and 3. and changes in the location of production. Changes in these three areas have been underway for some time. Structural change is a dynamic process which reflects the interaction of many forcesprimarily new technologies, shifts in consumer demand, the way producers satisfy that demand, increasing competition in international trade, and the introduction of new rules and regulations that govern business enterprises. The discussion that follows attempts to isolate the effects of these principle forces. Understanding changes in the structure of the U.S. economy is critical for understanding changes in job generation, pay, and regional growth, and for estimating the likely direction of the economy in the future. The dynamics of an economy heavily dependent on natural resources are likely to be very different from one primarily dependent on intellectual resources. An economy consisting of tightly integrated networks of small establishments will behave differently in response to shifting domestic and international markets than one characterized by large, centralized establishments and firms. An economic structure that spreads wealth unevenly around the country, or that places the burden of structural adjustment primarily on a few regions, presents a unique set of problems. In all three cases, policies designed to facilitate growth and mitigate the pain of structural change depend on a clear grasp of the patterns of change and the forces shaping them. SECTORAL CONTRIBUTIONS TO GNP Structural change is measured first by the relative contribution each major business sector makes to GNP. l Since the 1950s, Natural Resource industries, Construction, and manufacturing (particularly High Wage Manufacturing), have contributed less to GNP, while Transportation & Trade and Transactional Activities have increased their relative contributions. Taken as a whole, the share of service businesses increased from 52 percent of GNP in 1950 to 63 percent in 1984; at the same time, the share of goods defined as all manufacturing, Natural Resources, and Constructionfell from 48 to 37 percent. The rising share of the service sectors has been due to the strong growth in demand for services and, in the case of Transactional Activities (the fastest growing sector), a sharp increase in the use of this sectors Unless otherwise noted, all measurements of sectoral share are in constant 1980 dollars. products as inputs in the production recipe. 2 The opposite holds true for the manufacturing sectors, where demand has risen relatively slowly and production recipes have called for reduced use of manufactured products. Trade has tended to reinforce these trends. Many of these patterns have been underway for some time. There has been considerable debate over whether the U.S. economy is deindustrializing and how such a phenomenon would be defined and 2For a discussion of the impact of technology and the growing complexity of production in the increased use of services, see James B. Quinn and Christopher E. Gagnon, Will Services Follow Manufacturing Into Decline? Harvard Business Review, November-December 1986: for a description of the role of tastes, incomes, and economic growth in promoting the growth of the service sector, see Robert Lawrence, Can America Compete? (Washington, DC: The Brookings institution, 1984). 165

PAGE 170

166 measured.3 Some evidence suggests that rates of decline in some manufacturing industries have accelerated since 1972,4 and that some of this accelerated structural change can be attributed to slower economic growths Whether or not rates of change have increased, there is little doubt that the cumulative effect of change over the past two decades has been the creation of an economy where the origins of value are differentin some cases dramatically differentthan they were two decades earlier. In the midst of such a process it is difficult to point to a particular moment when a sudden change in structure took place; such things are easier to see in retrospect. Any attempt to measure changes in the origin of value by business type faces two major dilemmas. First, how should a sectors share be measuredin constant dollarsa proxy for share of physical outputor in current dollarsthe amount of money in which the economic activity actually occurred? Second, do the sectors selected for analysis accurately measure changes in the underlying structure of production, or is change masked by the categories chosen? Constant and Current Dollars The distinction between the type of measure used is critical, since measurement in constant dollars indicates that manufacturings share of value-added qThe dec]ine in the current dollar share of manufacturing has M to a controversy over whether the United States is deindustrializing. Bennett Harrison of the Massachusetts Institute of Technology and Barry Bluestone of Boston College, who popularized the term deindustrialization, argue that the manufacturing sectors in the United States are in decline; see The Deindustrialization of America (Boston, MA: Basic Books, 1982). Charles Schultze and Robert Lawrence of The Brookings Institution argue that deindustrialization has not occurred (see Robert Lawrence, op. cit., footnote 2, and also Robert Lawrence,The Myth of Deindustrialization, Chaf/errge, November/December 1983). For a review of the subject, see R.D. Norton, Industrial Policy and American Renewal, Journaf of Economic Literature, vol. XXIV, March 1986, pp. 1-40. dFOr an ana]ysis which supports this conclusion, SEW Nicholas S. Pertla, The Shift from Manufacturing to Services: A Concerned View, New Eng/and Economic Review, January/February 1987; for an opposing view, see U.S. Department of Commerce, Office of Economic Affairs, Changes in the Structure of the U.S. Economy Since 1960: A Primer, working paper, Washington, DC, January 1986. 5Jarn~ l-f. crossing and Arye L. Hillman, Shifting Comparative Advantage and Senescent Industry Collapse, The American Economic Review, June 1986, p. 516; and Ronald E. Kutscher and Valerie A. Personick, Deindustrialization and the Shift to Services, Monthly Labor Review, vol. 109, No. 6, June 1986, p. 10. has remained relatively constant for many years while this sectors current dollar share has fallen sharply (see figure 5-l). Both current and constant dollar measures provide useful views of the changes underway. The difference is largely due to productivity increases, which allow the same amount of goods and services to be produced for a relatively lower price (see box 5-A). Measurement of output in current dollars, or the dollars of the day, has strong intuitive appeal, since Figure 5-1.-Manufacturings Share of GNP and Jobs (current and constant 1982 dollars and persons) Percent 30 25 -- --- ---- ----- ---20-.. 1 1 1 1 1 1 15 I 1 1 I 1 1 1 I I I 1 I 1 1 r I I I 1 1 1 I 1 1 / 1955 1960 1965 1970 1975 1980 1985 --Job s Current dollar s Constant 1982 dollars SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts. historical diskettes. tables 6,1,6.2, and 6.10. t Box 5-A.Productivity and Constant Dollars Imagine an island that produced only pizzas and poems and sold equal numbers of both. Suppose that in 1950, 2 minutes of work were needed to produce either a pizza or a poem and that each cost $2. Suppose that thanks to surging innovation in pizza production, by 1980 poems still took 2 minutes of work and cost $2 dollars but pizza required only 1 minute and sold for $1 because prices were set in highly competitive markets. If demand did not change with price, and if the capital investments needed to increase pizza productivity were negligible, the constant dollar percentage of production in the island would remain the same at 50-50, but the current dollar share of poetry would increase from one-half to two-thirds of the islands GNP.

PAGE 171

167 it uses the prices at which transactions take place. G Measured in current dollars, manufacturings share of GNP fell from one-third in 1950 to less than onequarter in 1984. The constant dollar measure reflected in the U.S. Department of Commerces Gross National Product by Industry data series, however, reveals no discernible trendindeed, it indicates a relatively stable share at about 22 percent of total output. 7 Manufacturings share of GNP measured in current dollars has fallen largely because productivity increases in this sector have been passed on to consumers in the form of comparatively lower prices and demand has not risen enough to offset the price decline. Because increases in productivity frequently result in lower labor requirements, the current dollar measure tends to track employment in an industry (again see figure 5-l), making this a useful tool in analyzing policy considerations focusing on labor. Constant dollar measures attempt to remove changes attributable to prices, by tracking actual changes in quantities. In theory, this measure provides a more accurate indication of contributions to GNP and thus structural change. 8 In practice, the process hinges on the arcane business of creating deflators, which convert current dollars to the dollars of a particular yearconstant dollars. This task has become increasingly difficult as the economy generates more value in specialized high-technology products and in services. It is particularly difficult during periods of high inflation like the late 1970s. Accurate conversion from current to constant dollars is only possible where indexes can be based on measurable, fungible commodities like steel and oil. g SFO~ ~ description of structural change using a current dollar meas ure, see U.S. Congressional Budget Office, The lndustria~ Poficy Debate, Washington, DC, December 1983, p. 11; and Nicholas S. Perna, op. cit., footnote 4. TAna]ysts using different data series and methodologies, including the one presented later in this section, find results which conflict with the finding that the manufacturing sector retains a stable share of GNP when measured in constant dollars. See Larry Mishel, Trends in Manufacturings Level and Share of Output, Economic Policy Institute, Washington, DC, November 1987, unpublished. 8For a more detailed examination of how to define and measure struc tural change, the strengths and weaknesses of constant versus current measurement, and what role employment measures can or cannot play, see Changes in the Structure of the U.S. Economy Since 1960: A Primer, op. cit., footnote 4, or Nicholas S. Perna, op. cit., footnote 4. %%s A.G. Clem and W.D, Thomas, New Weight Structure Being Used in Producer Price Index, Monthly Labor Review, vol. 110, No. 8, August 1987, pp. 12-21, for a description of how a new price basis is estab lished. But the bulk of GNP is composed of goods and services whose various characteristics and specifications do not stay constant over time, making direct comparisons without some type of quality adjustment inappropriate. The question of adjusting for quality becomes extraordinarily complex when diverse products are entering the market and many products are new. Is a computer purchased in 1988 so radically different from one bought in 1960 that it is effectively a different product? This problem is magnified in the case of services, which are in many cases even harder to quantify. Heart surgery might be more expensive today than it was 10 years ago, but the probability of surviving may be much higher. Even in places where it is possible to develop a good estimate of the constant doIlar value of an industrys product (e.g., a deflator for automobiles), developing a deflator for the value-added portion of that product proves to be difficult. Since value-added is primarily composed of compensation paid to workers and income retained by the business, there is no clear and intuitive interpretation for a value-added deflator. A discussion of various techniques used to deflate value-added appears in box 5-B. Given the limitations outlined in the box, it is clear that constant dollar value-added estimates should be interpreted with care. Selecting the Categories for Measuring Sectoral Share Chapter 4 pointed to the difficulty of selecting business sectors in a way that revealed rather than obscured the structural shifts underway in the U.S. economy. This problem haunts all attempts to express structural change in a limited number of categories. Observed shifts of economic activity from one sector to another can merely be an artifact of the measurement process, and not a true change. Spurious change can occur because different business types are combined in a single category, definitions have changed, or the measurement process has improved. This problem applies not just to the 9 business categories used here to summarize structural change; it can also apply to the 85 industries that were combined to make the 9. ]0 For example, the IOMore than 500 sectors appear in the U.S. Department of Commerces input/output tables available for the benchmark years. Unfortunately, the most recent year for which data at this level exists is 1977.

PAGE 172

168 Box 5-B.Deflating Value-Added Under ideal circumstances, deflators for valueFigure 5-B.-Price Deflators for Various added inputs are calculated using input-output techSectors (1982=1.0) niques. The inputs purchased by each business are 1,2 deflated separately and the total is subtracted from a ----deflated level of industry sales (total output), creating ---1a residual which is deflated value-added. This technique, called double-deflation, is recognized as a pre~ 0.8ferred method by the Commerce Department because 0.6 ~ of its use of a consistent set of price indexes. However, ) double deflation is only used on the 29 percent of the 0.4, :-:1 1 I 1986 GNP generated by the manufacturing, farm, and 1972 1974 1976 1978 1980 1982 1984 1986 construction sectors; a variety of scaling techniques and other methods are used for the rest of the econ --GN P Manufacturin g omy.1 But even the preferred method is subject to se Machinery-non ele c vere problems which limit its usefulness. Errors in the SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, Naconstruction of the deflators or the calculation of the tional Income and Product Accounts, historical diskettes, tables 6.1 and 6.2. inputs or outputs appear in the remainder, which is value-added. tifact of deflation techniquesnot real structural change The problems encountered in creating a constant in the economy. dollar value-added series, even in manufacturing, are The rise of imports as intermediary inputs in the made obvious by figure 5-B, where data are particularly good. The figure shows the deflator for the production process presents another problem, since most deflation techniques use domestically based deflanonelectrical machinery industry (an industty categoty tors. 3 For example, suppose that General Motors buys that includes computers), the deflator series for all its steel for autos from South Korea in both 1972 and manufacturing industries, and the deflator for GNP. 1980. If the price of imported steel grew more slowly It is apparent that a new deflator for computers was than the price of domestic steel, a calculation of the introduced in 1982.2 Without this change, manufacgrowth in auto value-added using domestic deflators turing deflators would have increased between 1982 would overestimate real growth in value-added. and 1986. With the change, there was no significant difference. Changes in constant dollar manufacturing The importance of these factors increased during output for manufacturing are, therefore, partly an arthe 1970s and early 1980s, a period of rising imports, huge increases in the price of oil coupled with tremendous decreases in the prices of computers, and sharp growth in services-an activity which is inherently difficult to deflate. ] Milo Peterson, (loss Product by Industry, Survey of Current Business, VOI 67, No. 4, April 1987, 3A.G, Clem and W.D. Thomas, New Weight Structure Being Used in See David W. Cartwright, Improved Deflation of Purchases of Comthe Producer Price Index, Morrfh/y Labor Review, vol. 110, No. 8, August puters, Survey of Currenf Business, vol. 66, No. 3, March 1986 1987, pp. 12-21. mix of businesses clustered as the steel industry merely be due to changes in the industrial compomay change as small mini-mills producing spesition of a broadly defined sector. cialty products increase and the traditional large-scale mills producing bulk products decline. Recent Trends in Sector Shares Figures 5-2a&b, 5-3a&b, and 5-4a&b summarize Aggregation can also affect the analysis of what changes in relative shares of GNP held by the profactors cause a shift in the share of a sector. Shifts duction sectors that have occurred in the U.S. econthat appear to result from changes in the producomy over the past 35 years. A discussion of trends tion recipe induced by technological innovation may appears in box 5-C.

PAGE 173

169 Figure 5-2a. -Current Dollar Shares of GNP for Natural Resources, Construction, and Transportation & Trade 0!lt:[[l~lll~l :~~~~ 1950 1955 1960 1965 1970 1975 1980 1985 Natural Resources Constructio n -Trans & Trade SOURCE: U.S. Department of Commerce, Bureau of Economic Analyses, National Income and Product Accounts, historical diskettes, table 6.1. Figure 5-3a. -Current Dollar Shares of GNP for Low Wage, Medium Wage, and High Wage Manufacturing Percent of GNP 20 ~ 15J 10 -.-.-------. 5 1950 1955 1960 1965 1970 1975 1980 1985 Low Wage Manuf --Medium Wage Manuf. High Wage Manuf SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 6.1. Figure 5-4a. -Current Dollar Shares of GNP for Transactional Activities, Personal Services, and Social Services Percent of GNP 30 25 20 15l0 5 I 1 1 1 1 0 1 1 1 1 I 1 r I 1 I 1 1 I 1 1 T I I I [ 1 1 I I 1 I 1 1 I 1 1950 1955 1960 1965 1970 1975 1980 1985 Transactiona l Personal Service s --Social Services SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 6.1. Figure 5-2 b.-Constant Dollar Shares of GNP for Natural Resources, Construction, and Transportation & Trade Percent of GNP 30 25 20 ----_ -------------,--------------------------------15l0 5 Natural Resources Constructio n --Trans, & Trade SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 6,2, Figure 5-3 b.-Constant Dollar Shares of GNP for Low Wage, Medium Wage, and High Wage Manufacturing Percent of GNP 15l0 ~ ---------------------------------5 0LT1lIIIII:IIIT: 1~1l~~ f] 1950 1955 1960 1965 1970 1975 1980 1985 Low Wage Manu f -Medium Wage Manuf High Wage Manuf SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 6.2. Figure 5-4 b.-Constant Dollar Shares of GNP for Transactional Activities, Personal Services, and Social Services Percent of GNP 25 2015~ 10 5 1 I 1 0 1 1 1 I I 1 1 1 1 1 1 1 1 1 1 I 1 I 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 ( 1950 1955 1960 1965 1970 1975 1980 1985 Transactiona l Personal Services --Social Services SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 6.2.

PAGE 174

170 Box 5-CShifting Shares of Value-Added in the U.S. Economy Natural Resource Industries One of the clearest structural trends in the U.S. economy has been the relative decline of the Natural Resource sector. This decline is reflected in both current and constant dollar shares of GNP. The current dollar share has fallen from more than 12 percent to less than 9 percenta 29 percent loss since 1950 (see figure 5-2a). This loss would probably have been much greater had oil and gas prices not quadrupled in the 1970s. The constant dollar share also declined steadily. Over three-quarters of the loss from 1950 to 1984 is attributable to the farm industry. Between 1972 and 1984, however, the constant dollar share of the farm industry remained unchanged; nearly 90 percent of the sectoral loss was due to declines in crude petroleum & natural gas and electric, gas, water, & sanitary services, reflecting the conservation and substitution of energy products. Construction The constant dollar share of Construction increased slightly more than 1 percent during the 1950s (see figure 5-2 b). Because this was matched by gains in productivity, the current dollar share held steady. Since the mid 1960s, productivity growth has been small, and in some cases even negative. As a result, the current share has remained relatively stable while the constant share has declined, mainly in line with recent economic downturns; about one-fifth of the 30-year de crease occurred from 1973 to 1975, and another quarter occurred from 1979 to 1982. Low Wage Manufacturing Low Wage Manufacturing held a relatively stable share in constant dollars between 1950 and 1984, but lost half its share in current dollars (see figures 5-3a and 5-3 b). The difference between constant and current share is attributable to the fact that Low Wage Manufacturing had strong productivity gainsgains driven in part by intense foreign competition. Between 1972 and 1984, Low Wage Manufacturings share of constant dollar GNP remained largely unchanged because of offsetting trends in different industries. Footwear and miscellaneous manufacturing (mainly composed of the jewelry and toy industries) declined, while lumber & wood products and apparel grew. The relatively large rubber & plastics and furniture & fixture industries kept a constant share. Medium Wage Manufacturing Medium Wage Manufacturing has lost nearly a quarter of its share measured in current dollars since 1950, but gained over a quarter in constant dollars (see figures 5-3a and 5-3b). As with the low wage sector, this discrepancy is due to steady gains in productivity. The three leading industries of this sector-non-electrical machinery, electric industrial equipment, and food & kindred productsexperienced respective price increases that were 10, 28, and 27 percent below the 1970 to 1983 U.S. average. ] During the 1972-84 period, the increase in the constant dollar share of this sector resulted from industries characterized as high-tech: electronic components, optical equipment, scientific instruments, and office, computing, & accounting machines. These four accounted for 73 percent of the growth of all industries in this sector that rose in share; office, computing, & accounting enterprises increased their share by a factor of almost 6 over the 12-year time period. High Wage Manufacturing The fate of High Wage Manufacturing is particularly important for the U.S. economy since, as the name suggests, it has long been a source of well-paid jobs. Its importance also lies in its strong links to the rest of the economy (see ch. 4). Nevertheless, the sector lost both constant and current dollar share of GNP (see figures 5-3a and 5-3 b). The bulk of High Wage Manufacturings loss in constant dollars occurred during the 1970s, a period of weak productivity growth for this sector and of increasing competition from foreign firms. The petroleum refining, primary iron & steel manufacturing, and motor vehicle industries were the major sources of this decline. Of the 20 industries included in this sector, only one, drugs, cleaning, & toilet preparations, witnessed any substantial growth, with most of that occurring between 1972 and 1977. Productivity in many high wage sectors has increased in the past few years, but average growth during the 1972-1984 period was sIo W 2 The discrepancy between the shares reported in figure 5-3b and the share shifts calculated using the input/output methods described in chapter 4 is particularly great for this sector. U.S. Bureau of the Census, Statistical Abstract of United States, 1985 (106th cd.), Washington, OC, 1984, table 783. Z lbid.

PAGE 175

171 Transportation & Trade Transportation & Trade is one of the largest sectors in the economy, responsible for approximately onefifth of the Nations GNP over the past 35 years. The sectors current dollar share has declined gradually since the mid 1950s, while the constant share has tended to rise (see figures 5-2a and 5-2 b). Individual industries, however, have experienced significant changes. Of the sectors 5-percent gross decline in current dollar share from 1950 to 1984, more than half was attributable to railroad transportation enterprises. On the other hand, air transportation enterprises were responsible for half of the gross increase. From 1972 to 1984 there was little change, although nearly all of this sectors growth in constant dollar share has come from the wholesale & retail trade industry. Wholesale & retail trades contribution to GNP has historically been the largest of any industry; by 1984, its share was more than that of all industries contained in the Low and High Wage Manufacturing sectors combined. Transactional Activities No other sector has experienced the dramatic growth generated by the group of industries characterized as transactional. The growth of this sector has been rapid and continuous since 1950, with both the current and constant share rising in unison until 1980 (see figures 5-4a and 5-4 b). After 1980, however, the constant share leveled off while the current share continued to climb. This has probably been due to the escalation in real estate prices since 1979. 3 Since 1972, the dominant growth industry within this sector has been business services, responsible for more than 40 percent of the increase in sectoral share. Real estate & rental also grew, contributing more than one-quarter of the constant share increase; this development made real estate the second largest industry, after wholesale & retail trade, in the 1984 U.S. economy. 4 Personal Services Losing share in both constant and current dollar terms, Personal Services are the exception to an otherwise dramatic shift towards a service-oriented economy (see figures 5-4a and 5-4 b). The largest decline is located in the household services industry (domestic services), much of which has fallen prey to the introduction of labor-saving household appliances. 5 Since 1972, the sector has maintained a stable constant dollar share. This balance was achieved through gains in the automobile repair& services and amusements industries, offsetting a continued decline in household services. Social Services The Social Service sector is comprised of the public and private health and education industries, as well as Federal and State government. This sector has grown dramatically over the past 35 years; in fact, it has the fastest current dollar growth rate of any sector, increasing by 70 percent between 1950 and 1984 (see figure 5-4a). State and local government enterprises were the main contributors to this current dollar growth, generating about half of the increase since 1950. The constant dollar share, however, only grew at the same rate as the economy. Part of this dichotomy can be attributed to weak gains in productivity; only the Personal Services and Construction sectors have shown less productivity growth than Social Services. The other factor involved is a sharp increase in prices. Between 1972 to 1984, the combination of Federal, State, and local government decreased their share in constant dollars. Most of this loss was compensated for by a gain in share from private health, education, & social services. tlbid,, table 790. 4The National Income Accounts also impute a rental-equivalence value for home owners that is allocated to the real estate industry. 5 Joe Schwartz, The @ods Life, Amt+can Demographics, December 1987 can be isolated from other factors, by using 1980 protechnology of production. Table 5-2 details these duction recipes to estimate industry output given patchanges. terns of final demand occurring in 1972, 1977, 1980, and 1984. In other words, the effects of final demand Changes in the composition of final demand from can be separated from those brought about by new 1972 to 1984 had a significant impact on the output recipe patterns, which tend to reflect changes in the produced by industries in several sectors, especially

PAGE 176

172 The changes may seem small at first. The share of Natural Resources, for example, fell from 13 percent in 1950 to about 9 percent in 1984hardly a dramatic decline. It must be recognized, however, that given the size of the U.S. economy, even a one percent shift in GNP share is significant. In 1984, a 1 percent change in share meant $38 billion gained or lost. For comparison, the total compensation paid to al] of the employees of the motor vehicle industry was $35 billion, and businesses spent $34 billion on computers, in 1984. 11 Explaining the Changes Which of the factors listed at the beginning of this chapter were responsible for the sectoral shifts described in figures 5-2, 5-3, and 5-4? Methods introduced in chapter 4 can be used to distinguish between changes resulting from new patterns of demand, new production recipes, and new patterns of international trade. 12 In brief, changes in demand, both domestic and international, and changes in production recipes have had almost equal effects on the overall shift in share during the past decade, but the impact on individual sectors and industries varies widely. Not surprisingly, demand changes tend to have their greatest effect on industries that sell final, finished products, such as motor vehicles and real estate. Recipe changes affect industries such as oil, steel, chemicals, and business services, which typically supply goods or services that are then used as ingredients in a finished product. International trade has reinforced the changes generated by domestic demand and production recipes, but its effects were not pronounced until after 1980. The period 1972 to 1984 was chosen for in-depth analysis. This is the most recent time frame for which I IU.S. f)epafirnent of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, Survey of Currerrt Business, vol. 67, No. 7, July 1987, tables 5.6 and 6.4B. Demand refers to final demand, or the purchase of goods and services for final consumption (see chs. 2 and 3). Recipe represents the intermediate demand for goods and services (ingredients) that will be combined to create a final product (see ch. 4). For examples of using demand as a factor of change, see Robert Lawrence, Can America Compete? op. cit., footnote 2, and U.S. Department of Commerce, Office of Economic Affairs, Trade Ripples Across U.S. Industries, Washington, DC, 1986. For analyses that focus on the role of recipe, or technical coefficients in the process of structural change, see Anne Carter, Structura/ Change in the American Economy (Cambridge, MA: Harvard University Press, 1970) and Wassily Leontief and Faye Duchin, The Impacts of Automation on Employment, 1963-2000, contract report for the National Science Foundation, Washington, DC, April 1984. consistent and detailed information could be assembled. It was an interesting period, characterized by major structural movements. Between 1972 and 1984, for example, oil and gas prices more than tripled, steel production fell by 30 percent, information processing equipment grew from 13 percent to over 40 percent of all durable equipment expenditures, and the ratio of imports and exports (gross) to GNP nearly doubled. Table 5-1 summarizes the analysis of shifts in the share of constant dollar value-added contributed by ten sectors between 1972 and 1984, showing the extent to which the shifts were attributable to changes in production recipes as opposed to foreign and domestic demand. 13 The table is not intended to describe exact movements between sectors resulting from different factors, but to exhibit the relative magnitude of the different factors as part of sectoral change. In this sense, the table provides a sensitivity analysis and not a precise mapping of cause and effect. Of the 8.9 percentage points of GNP that was exchanged between sectors gaining share and sectors losing share from 1972 to 1984, roughly half was attributable to changes in the production recipe and half was connected to changes in final demand. Within the broad category of final demand, the impact of domestic demand on changes in economic structure was-for the economy as a wholemore than twice as large as the impact of trade. Nevertheless, the impact of trade was still impressive, because most trade effects occurred over a relatively brief period-after 1980while the effects of domestic demand have been relatively constant since 1972. The Impact of Trade and Domestic Demand on Economic Structure The effects of changing patterns of final demand on the changing shares of different business sectors ItThe selection of any one year for a comparison point is fraught with the peculiarities of that particular year. This analysis uses 1972 and 1984 because they offer the widest time range given the available data, and because they are close to being the end points of their respective business cycles, effectively eliminating cyclical factors. Different endpoints can yield different effects attributable to demand and recipe. See Stanley J. Feldman, David McClain, and Karen Palmer, Sources of Structural Change in the United States, 1963-1978: An Input-Output Perspective, The Review o/Economics and Statkks, 1987. The 1984 endpoint uses 1984 demand in conjunction with the 1980 Input-output tables, and therefore does not completely reflect the 1984 economy.

PAGE 177

.Table 5.1 .The Sources of Change in Percentage of Constant Dollar Share of Structural Change: GNP From 1972 to 1984 by Sector (1980$) (numbers will not necessarily add due to rounding and interactive effects) Natural Low Wage Medium Wage High Wage TransportIon Transactional Personal Social Federal Resources ConstructIon Manufacturing Manufacturing Manufacturing & Trade Actlvltles Services Services Defense Other a All .3 3 0.2 06 .3 2.0 55 0.1 0.4 .9 1 Production reclpe b o 9 o 5 04 03 .3 06 26 0.2 .4 .0 0.0 Final demand 1,4 o 7 .3 05 o 9 12 2.9 1 0.8 .9 ,1 Domestic demand c ,1 8 0.1 0.7 .5 0.7 24 .2 0.6 .9 o Trade c 4 01 ,4 .3 .4 0.5 0.5 00 0.3 0 0 0 0 Interactive .0 01 .0 0.2 0.1 .2 .0 .0 .1 0.0 0.0 How To Read This Table: Between 1972 and 1984, the percentage of all value-added in the economy generated by Natural Resource businesses fell 2.3 percentage points. This is equivalent to saying that Natural Resource contribution to GNP fell 2.3 percentage points. The decline resulted from new patterns of domestic demand (responsible for a 1.1 percentage point decline in the Natural Resource share), trade (0.4 percentage point), and production recipes (0.9 percentage point). a Includes non-production, accounting industries such as non-comparable imports, scrap, rest of the world industry, and inventory valuation adjustment. b Production recipe refers only to 1972 to 1980 change. c Estimated from the 1984 National Income and Product Accounts (U.S. Department of Commerce) d 1984 Trade data are estimated and rebased into 1980 dollars from data provided by the U.S. Bureau of Labor Statistics. NOTE: All = Recipe + Final Demand + Interactive: Final Demand = Domestic Demand + Trade SOURCE Office of Technology Assessment. 1988

PAGE 178

174 Table 5-2.U.S. Structural Change Resulting From Changes in Final Demand (constant dollar GNPshare for selected years of final demand; value-added shown by production sector and calculated using the 1980 1.0) Production sectors 1972 1977 1980 1984 Natural Resources 10.7 9.4 8.8 9.3 Construction . 6.9 6.4 5.9 6.2 Low Wage Manufacturing . 3.6 4.2 3.4 3.3 Medium Wage Manufacturing . 9.5 10.0 9.8 10.0 High Wage Manufacturing . 9.9 9.9 9.3 9.0 Transportation & Trade . . . 18.2 18.9 19.5 19.4 Transactional Activities . . 20.7 23.1 23.3 23.6 Personal Services 3.8 4.0 3.7 3.7 Social Services . 13.5 15.0 14.5 14.2 Federal Defense . 3.0 2.3 1.9 2.1 Other. . . . 0.2 .2 0.0 .9 Total 100.0 100.0 100.0 100.0 %lutput derived from demand estimated from the National Income and Product Accounts NOTE: Totals may not add to 100 due to rounding. SOURCE Office of Technology Assessment, 1988. Natural Resources, High Wage Manufacturing, and Transactional Activities. In the case of the Natural Resource sector, most of the shift due to final demand occurred between 1972 and 1977. Not surprisingly, the two industries affected were crude petroleum & natural gas and electric, gas, water, & sanitary services. For crude petroleum, most of the decrease was attributable to trade as imported oil replaced the output originating from domestic companies. In the case of electric, gas, & water services, nearly all the decrease was traced to lower domestic demand from domestic consumers. The High Wage Manufacturing sector also experienced a significant reduction in share due to final demand. Half of this sectors loss in GNP share due to shifts in final demand came from the motor vehicle, iron & steel, and petroleum refining industries. The motor vehicle industry didnt start to lose share until after 1977, at which point the decline was primarily due to slack domestic demand from consumers and business. Trade was a relatively small factor for this industry until after 1980, when its negative effect almost eliminated what could have been a relatively large gain in share achieved through strong domestic demand. The iron & steel industry has suffered a fate similar to that of motor vehicles: the decline in share from 1977 to 1980 was of domestic origin; after 1980, domestic demand would have slightly increased this industrys share of GNP if not for a detrimental trade balance. Although these are the major declining sectors, another sector, Low Wage Manufacturing, registered a small decline and is of interest because of the dominant effect trade has had on this sector. In a notrade environment, five out of the twelve Low Wage Manufacturing industries would have gained in GNP share, resulting in a net gain for the sector over the period from 1972 to 1984. But the addition of trade reverses the trend, resulting in a decline in GNP for three-quarters of the industries. The industries that experienced the most dramatic turnaround in share due to trade were fabrics, textile goods, apparel, footwear, and miscellaneous manufacturing (which includes the production of jewelry, toys, and writing implements). By and large, these trade impacts have been felt since 1972, but they are most pronounced in the period from 1980 to 1984. Constructions share of GNP fell over the 197284 period, due to declines in domestic demand for new construction. The defense sector also lost GNP share between 1972 and 1977, in part because of the end of the Viet Nam war. (The value-added in defense consists entirely of compensation paid to military employees.) Balancing the decline of these sectors are four sectors that gained in GNP share as a result of changes in final demand: Transactional Activities, Transportation & Trade, Social Services, and Medium Wage Manufacturing. Of the four, Transactional Activities was responsible for gaining over 50 percent of the total share that shifted between 1972 and 1984 due to final demand. Four-fifths of this sectors gain can be credited to domestic demand, primarily consumer demand, the bulk of which has been fueled by consumer demand for the products of the real estate & rental and finance & insurance industries.l4 Transactional Activities also benefited the most from international trade; about one-fifth of the sectors total gain due to demand came from trade. The industries that increased their GNP share because of trade were similar to the industries that benefited 14The rea] estate & rental industry includes a Commerce Department imputation for owner-occupied housing.

PAGE 179

175 from domestic demand: real estate & rental and business services. The Transportation & Trade sector gained share because of both trade and comparatively rapid growth in domestic demand. The wholesale & retail trade industry was responsible for most of the growth. Within this sector, eating & drinking places increased share particularly rapidly, primarily because of rapid growth in final demand. Social Services also gained share because of changes in final demand. Most of the increase resulted from growing demand for the health, education, & social services industry. These increases offset a large decline in demand for the services of Federal, State, and local activities that occurred between 1972 and 1984. Ninety percent of this loss occurred after 1980. Most of Medium Wage Manufacturings increase derived from final demand came from industries that can be characterized as high technology: office, computing, & accounting machines; radio, TV, & communication equipment; and optical, ophthalmic, & photographic equipment. Domestic demand for these products, particularly from the business sector, was strong, but the negative effect of trade dampened what could have been impressive gains. Trade also contributed to a loss in GNP share for less technologically oriented industries like metalworking machinery and electronic components. Nearly all of this loss of GNP share due to trade occurred between 1980 and 1984, reversing a positive impact from trade achieved from 1972 to 1977. The Impact of Production Recipes on Economic Structure If changes in demand, including trade, account for roughly half of the total structural change in the economy, the other half can presumably be explained by changes in intermediate demandthe demand for goods and services used as ingredients in making a product for final consumption, a major component in the production recipe. Chapter 4 explained how this recipe changed between 1972 and 1980, revealing that service sector inputs had increased while inputs from the natural resource and manufacturing sectors decreased. How has this change in recipe affected GNP share? The impact of changes in the production process can be seen by estimating how the output from different industries would change when production recipes vary-using those of 1972, 1977, and 1980 but leaving demand as it was in 1984. What would GNP look like if the production recipe of 1972 were used to satisfy demand for a car, compared to the recipe used in 1980? It must be recognized, of course, that the process of producing 1984 demand using a 1972 production recipe is highly artificial. For example, most industries would have used a dramatically different amount of oil in 1972 if the the price of oil in 1972 was as high as in 1980. The procedure reveals that the impact of recipe on structural change was roughly the same as that of demand. l5 As table 5-3 shows, three of the five sectors with increasing share due to changes in recipe were the same three sectors that benefited from changes in demandTransactional Activities, Transportation & Trade, and Medium Wage Manufacturingalthough within these sectors, the individual industries that benefited differed widely from those gaining share due to final demand. Sectors losing share due to changes in both recipe and demand were High Wage Manufacturing, Natural Resources, and Construction. Most changes were therefore reinforcing rather than offsetting, although Low Wage Manufacturing and Personal Services proved to be exceptions; these sectors experienced a gain in share due to recipe, but lost as a result of demand. Social 15 The time span is restricted to an 8-year period, 1972 to 1980, because of the availability of data. Table 5.3.U.S. Structural Change Resulting From Changes in the Production Recipe (constant dollar GNP share derived from 1984 total final demand; by production sectors, using the 1972, 1977 and 19801-0 adjusted to 1980$) Production sectors 1972 1977 1980 Natural Resources . . . 10.1 9.7 Construction . . . . 6.8 6.6 Low Wage Manufacturing. . 2.9 3.3 Medium Wage Manufacturing 9.7 10.0 High Wage Manufacturing . 11.3 9.7 Transportation & Trade . . 18.8 19.4 Transactional Activities . . 21.0 22.0 Personal Services . . . 3.5 3.6 Social Services. . . . 14.7 14,4 Federal Defense . . . 2.1 2.1 Other . . . . . . .9 .9 Total . . . . . .. .100.0 100.0 9.3 6.2 3.3 10.0 9.0 19.4 23.6 3.7 14,2 .9 100.0 NOTE: Totals may not add to 100 due to rounding. SOURCE: Office of Technology Assessment, 1988.

PAGE 180

176 Services lost share because of recipe shifts, but gained share because of demand. The sector which lost the most due to changes in recipe was High Wage Manufacturing, Approximately 70 percent of the decline occurred between 1972 and 1977, and can be traced to three industries: petroleum refining, chemicals, and primary iron & steel manufacturing. These industries account for more than four-fifths of the 1972 to 1980 drop in High Wage Manufacturings GNP share due to recipe changes. Unlike the situation resulting from changing demand, the motor vehicle industry experienced only a small reduction in share because of changing production recipes. The changes in High Wage Manufacturings share of GNP seem to have resulted from the increased availability of new materials, such as plastics, which can act as substitutes for steel; 16 the sharp increase in crude oil prices, which pushed up the price of chemicals and refined petroleum products and thus forced a recipe change; and the influx of foreign intermediate inputs as ingredients in the production recipe. The share lost by High Wage Manufacturing due to recipe changes was primarily absorbed by Transactional Activities. The Transactional sector picked up nearly two-thirds of the shift in share that occurred between 1972 and 1980 due to recipe change. Growth was strong from 1972 to 1977, but even stronger from 1977 to 1980; presumably, a 1984 recipe would indicate that this sector gained even more. The business services industry, which includes activities such as advertising, legal services, and computer and data processing services, was responsible for nearly 70 percent of the sectoral increase; 17 the finance & insurance industry came in a distant second. Production recipe changes resulted in a sharp increase in demand for Transactional Activities business services in particular. As discussed in chapIsStan]ey J. F&fman and Karen Palmer, Structural Change in the United States: Changing Input-Output Coefficients, Business Economics, January 1985, p. 43. ITFor a more detailed analysis of the changing role of the business services industry in the recipe of production, see Feldman and Palmer, op. cit., footnote 13; Bobbie H. McCrackin, Why Are Business and Professional Services Growing so Rapidly? Economic Review, Federal Reserve Bank of Atlanta, August 1985; and John Tschetter, Producer Services: Why Are They Growing So Rapidly? A40nthly Labor Review, vol. 110, No. 12, December 1987, pp. 31-41. ter 4, more complex business networks seem to require larger numbers of transactions, resulting in more demand for services like those provided by lawyers and consultants. Decisions to contract out services that were previously done in-house and increasing geographic dispersion of production also result in an increase in Transactional Activities. l8 For example, many manufacturing firms now contract out for janitorial services that had been performed internally. Although the actual amount of this work has not necessarily increased, it appears to grow because it is now counted as an arms-length market transaction. l9 The impact of changed production recipes on sectors other than High Wage Manufacturing and Transactional Activities is rather small. Most of the decline among Natural Resource industries was located in agricultural products (non-livestock), and in electric, gas, water, & sanitary services. Changes in recipe also had a negative effect on the share of Construction, particularly in the new construction industry. The last sector to experience a loss of share due to recipe changes was Social Services, with most of the loss attributable to the health, education, & social service industry. After Transactional Activities, the sector that gained the most in terms of share due to recipe changes was Transportation.& Trade. The wholesale & retail trade industry was responsible for most of the increase gained by this sector. This is not surprising, considering the increased role that marketing now plays in the delivery of a product, and the fact that constant dollar per capita retail sales grew by 23 percent between 1972 and 1984. 20 Unlike the situation in High Wage Manufacturing, both Low and Medium Wage Manufacturing benefited from changes in the recipe of production. In the low wage sector, the industries devoted to apparel and lumber & wood products achieved the greatest gains. For Medium Wage Manufacturing, the principle industries behind the gain in sectoral share were office, computing, & accounting machine~ and electronic components (mostly semiconductors). lgM&rackin, op. cit., footnote 17, and Tschetter, op. cit., footnote 17. lgRichard B. M~enzie, The Emergence of the Service Economy: Fact or Artifact? policy analysis No. 93, Cato Institute, Washington, DC, Oct. 27, 1987. 20U.S, Bureau of th e Census, statjs~jca] Abstract of the United States, 1986 (106th cd.), Washington, DC, 1985, table 1388.

PAGE 181

177 In the case of computers and electronic components, the reasons for an increase in share due to recipe change are apparent. During the 1970s, the computer became an accepted element in the production process; as time goes on, adoption, adaptation, and implementation of computer technologies in the workplace is certain to increase. The reason behind the growth of apparel as an intermediate input is less obvious, except perhaps for the fact that the Multifiber Arrangement, which allows for trade agreements that restricted fiber imports into the United States (and other developed nations), took effect in 1974; some researchers have found that these trade restrictions had a positive effect on domestic output. 21 The increase in the output of the lumber industry is probably due to the role lumber plays as an input in housing production, and the fact that new housing starts for the 1972-84 period peaked in 1972-73 and 1977-78. 22 Tracing the Significance of Structural Change It is difficult to pass judgement about whether changes in sectoral share of GNP are desirable or 21 U.S. Congressional Budget office, Has Trade Protection Revitalized Domestic Industries? (Washington, DC: U.S. Government Printing Office, November 1986), p. 32. 22 Statistical Abstract of the United States 1986, op. cit., footnote 20, table 1297. undesirable. Much of the remainder of this document is devoted to tracing the way shifting production recipes effect U.S. responses to trade, opportunities for employment, and the net productivity with which the economy delivers amenity to consumers. Chapter 7 traces the relationship between trade and sector contributions. Chapter 10 shows how changing production recipes affect demand for different kinds of jobs. Patterns of change in sector shares (including the surprisingly constant share of manufacturing) have not changed radically for many decades. On the other hand, a prolonged period of slow change can have a major cumulative effect. After reaching a threshold, what seemed like a gradual process may suddenly be seen as a transition. This threshold is more likely to be perceived in periods of slow economic growth than in periods of strong growth, where downward fluctuations are masked by general prosperity .23 The issue is not whether structural change is occurring at any greater or lesser rate that it has in the past. In some areas it has and in others it has not. The critical point is that the collective effect of three decades of change has left the United States with a much different economy. 23 James H. Crossing and Arye L. Hillman, Op. cit., footnote 5. CHANGES IN SCALE AND SCOPE The second kind of structural change to be examined involves the way business networks are owned and managed, in terms of the size, or scale, and the product mix, or scope, of their operations. This analysis is critical because different patterns of scale and scope shape the ability of a business to compete in international markets, dictate the quality and stability of the jobs offered, and determine the success with which new technologies can be exploited. One of the difficulties in any discussion of this kind is that there is no good vocabulary for describing the variety of ways that business networks in an economic sector are organized. One source of confusion is the distinction between an establishment and an enterprise. The word enterprise (sometimes also called a firm) indicates an independently owned corporate entity. An establishment is a specific plant, branch, or subsidiary within an enterprise. A large enterprise may consist of many small establishments. Beyond this distinction, there are a variety of patterns in which business networks can be organized. The following is a partial list: l Sectors dominated by a small number of large firms that either: concentrate their activities in a single plant or a comparatively small number of plants (e.g., the old Ford Motor Co. Rouge River Plant); maintain tightly managed subsidiaries and branch offices, where managers of distinct components are not given appreciable flexibility (e.g., branches of major insurance companies); or

PAGE 182

178 have the dominant form of organization through loosely controlled establishments, subsidiaries, or franchises, in which the establishments management is given considerable autonomy and is compensated primarily on the basis of decisions made as an independent agent (e.g., a branch bank manager with freedom to establish his or her own loan criteria). l Sectors dominated by one or more large firms that effectively regulate a market for a large collection of small, independent firms (e.g., IBM sets de-facto standards for a wide variety of personal computer equipment and software systems produced by many different firms; similarly, the firms formerly part of the old Bell system set standards for a host of other communications suppliers). l Sectors dominated by small firms operating independently, which are either: nominally independent, but constrained by their product or by a lack of research to behave as though they were producing massproduced commodities (e.g., farms, teaching, and home construction); or independent entrepreneurs, providing imaginative responses to new markets and new production technologies. Such firms may pool research or marketing through trade associations (an example might be the semiconductor industry in Silicon Valley). Given the difficulty of developing an acceptable taxonomy, there is no easy way to trace patterns of change. Moreover, firms in traditionally fragmented sectors (e.g., farms, physicians, and home builders) are amalgamating into larger units, while sectors that were traditionally highly concentrated (e.g., automobile production) are turning to networks of small and medium-size suppliers for a greater share of parts and engineering services. 24 It appears extremely unlikely that the new form of industrial organization will bear much resemblance to the new industrial state that seemed so inevitable a generation earlier. 25 While small firms Z4Michae] S. flynn, out-sourcing Rediscovered, L!XESpectrum, vol. 24, No. 10, October 1987, pp. 46-49. zs~e John Kenneth Ga]braith, The New /ndustria/ State (New York, NY: New American Library, 1971). may provide invaluable sources of innovation, it appears equally unlikely that significant growth will originate from small firms unless they are able to work together in a way that at least approximates the economies of larger enterprises-small firms now command a declining share of output in virtually every sector of the economy. However, they play an important role by providing employment when few other alternatives are available, and provide a large share of jobs added during economic recessions (e.g., as carpenters and machinists become selfemployed repairmen or work in retail stores). 26 Under current circumstances, small firms are able to provide growing employment opportunities by offering comparatively low wages, few if any nonwage benefits, comparatively poor working conditions, and weak job security. While parts of Italy and West Germany may have created an environment where networks of small firms can avoid many of these liabilities, the growing share of U.S. employment in smaller businesses, coupled with a shrinking share of assets, profits, and sales, may lead to the growth of a two-tier system in the United States. Workers finding stable employment in comparatively productive large firms may do increasingly well, while their counterparts in small firms do increasingly poorly. Economic success now appears to depend increasingly on technical innovation, but little is known about how the scale and scope of businesses influence rates of innovation. Entrepreneurs may alone be able to grasp a truly revolutionary concept; the merits of semiconductors were almost universally ignored by firms with a large stake in the manufacture of vacuum tubes, and the merits of microcomputers were not initially recognized by firms with large interests in mainframes. But groups of entrepreneurs are able to manage large markets only in exceptional cases. In the semiconductor and microcomputer shake-out, survivors have typically grown to become large enterprises or have been absorbed by large firms. Industrial structure in major sectors clearly results from private management decisions, but these deZsB~ce E. Kirchhoff and Bruce D. Phillips, Examining Entrepreneurships Role in Economic Growth, paper delivered at the Seventh Annual Babson Entrepreneurship Research Conference, Malibu, CA, April 30, 1987.

PAGE 183

179 cisions are strongly influenced by public policy and in many cases are the conscious result of public policy. Private decisions about mergers, pooling of research, foreign marketing, adoption of communications and other standards, contributions to retraining programs, and other actions that shape the way large business networks are managed, are all strongly influenced by Federal and State policydirectly through formal regulation, and indirectly through tax laws and other macroeconomic decisions. Describing Changes in Scale and Scope Changes in scale and scope affect both individual firms and the establishments of which they are composed. For example, insurance firms are delegating more authority to small sales offices located throughout a region, while they are consolidating record keeping and other functions. Large construction firms are combining property development, mortgage financing, factory construction of components, and site erection. General Motors has a finance division, GMAC, that provides loans for buyers of GM cars. Even though the creation of this division is an expansion of scope, because it is directly tied to new car sales, it does little to insulate GM from falling sales. To counter this problem, GM has proposed another expansion of scope by which it will begin to offer home equity loansa business divorced from cars. This diversification of products can lower costs, as fixed investments are amortized over a broader base of products. There is clear evidence that the growth of large firms is increasingly built around the aggregation of many comparatively specialized small establishments. As a result of these developments, many large firms claim that they are reorganizing operations to encourage more entrepreneurial behavior on the part of individuals and establishments. Techniques range from large rewards for inventors and patents (IBM can award $10,000 or more for an important innovation) to the AT&T system for encouraging venture business, in which employees can earn both salary and profits from new projects and can even invest personal funds or defer part of salaries. The firm reports that it has already received 2,000 proposals. 27 ZTR.M, Kanter, The Attack on Pay, Harvard Business Review, vol. 65, No. 2, March-April 1987, pp. 60-67. Obviously, a radical change in corporate philosophy is needed before established, hierarchical firms will be willing to tolerate real entrepreneurial behavior. While GM claims to have encouraged its production establishments to shop for price and quality, in practice the assembly operations appear to be forced to give preference to Central Foundry Division for castings, Rochester Products Division for fuel systems, and Delco Moraine Division for brakes. These GM subsidiaries regularly win long-term contracts because they are allowed to bid without including overhead, capital, and other fixed expenses. GMs inability to escape rigid vertical integration, and the comparative flexibility of Ford and Chrysler, has been cited as an explanation for GMs recent loss of market share. 28 A large firm, or a set of firms, can also provide a de facto set of standards and regulations to organize the activity of a large number of smaller satellite enterprises. 29 The dominating firm can operate at a variety of places in the system, tying resources to final markets, and can be a producer, such as IBM, a retailer, such as Safeway, or a mixture of both, such as health maintenance organizations. Other examples include large hospital chains and group practices, which are capturing markets once enjoyed by fragmented private practices and local unaffiliated hospitals; these firms are often innovators in medical practices designed to reduce costs. Small community banks, caught in a changing regulatory environment that has spurred the need to broaden their product mix and enhance their technical expertise, are forming partnerships with larger regional banks. 30 Small, independently owned firms can form effective networks through a series of free market contracts, but their performance can often be enhanced by cooperating in areas of mutual interest such as research and development, training, and overseas marketing. The ability to create some of these consortia is difficult for U.S. firms because of antitrust laws and entrenched business cultures, but they have ZEW. f-farnplon and J.R. Norman, General Motors: What Went Wrong, Business Week, No. 2990, Mar. 16, 1987, p. 102. ZgThe term is used by Michael Piore and Charles Sable, The Second Industrial Divide (New York, NY: Basic Books, 1984). 30 Dwight B. Crane and Robert C. Eccles, Commercial Banks: Taking Shape for Turbulent Times, Harvard Business Review, vol. 65, No. 6, November/December 1987, pp. 94-100.

PAGE 184

180 been used effectively in Japan, West Germany, and Northern Italy 31 The problem of defining large and small firms or concentrated and competitive sectors has become increasingly difficult given the growth of national and international production networks. While antitrust considerations were formerly needed to consider whether a firm monopolized regional markets, should they now consider the scale of firms with respect to international markets? Forces of Change Changes in scale and scope are driven by the same four forces that lead to changes in sectoral contributions to GNP: technology, rules and regulations, patterns of domestic demand, and international competition. Technology Changes in industrial organization are made possible by radical improvements in the way information can be communicated among firms and establishments, and by changes in technology that greatly reduce the size at which economies of scale become important. The availability of inexpensive computerassisted design (CAD) and computer-assisted manufacturing (CAM) technologies operating on personal computers has allowed even small machine shops to adopt these innovations, reducing waste, errors, and down time. The time to convert a customers drawing to a cutting die, for example, has fallen by a factor of seven. 32 Even small metal fabrication facilities can make use of technologies such as Flexible Manufacturing Systems, which operate at a minimum scale of six machines and a half dozen people. 33 Large data entry and typing pools are disappearing as data entry is integrated into other functions distributed throughout the firm. Communications technologies have rewritten many rules, providing unprecedented opportunities 31 Danie]le M~onjs and Marjo Pjanta, An innovation strategy fOr Traditional Industries, draft report prepared for the industrial association of Prado, Italy, September 1986. sZMechanica/ Engineering, January 1987, P. 6. wsee R. Jajkumar, Postindustrial Manufacturing, Harvard Business Review, vol. 64, No. 6, November/December 1986, p. 76; and Mechanical Engineering, op. cit., footnote 32. This issue is discussed at greater length in the manufacturing section of ch. 6. for uniting small enterprises into a dynamic framework. 34 Production technologies capable of tailoring products to specialized markets without a significant sacrifice in productivity or increase in cost can vastly diminish the value of economies of scale. 35 Of course, small subsidiaries of large firms may still be favored over independent firms that lack access to sophisticated communication networks. But there are instances where smaller enterprises, tied to an independently operated information network, can enjoy new life because of a close relationship with national and international markets. A national data network, for example, may help small-scale U.S. farmers who produce high value products, such as fruits or specialized vegetables, to bid on the international market. Advances in telecommunications allow companies like American Airlines to take advantage of cheap labor in Barbados for data entry, saving the company about $3.5 million per year. 36 Regulation Formal changes in regulations have had a profound effect on patterns of business organization throughout the economy, particularly in transportation, health, and communications. Changed regulation in transportation created an explosion (possibly of short duration) of independent owner-operators, as well as the formation of integrated transportation firms such as Federal Expresswhich combines truck pickup and delivery, air freight, and telecommunications. Health regulations have encouraged the formation of health maintenance organizations and consortia of physicians. The breakup of AT&T has, of course, fragmented a sector once organized exclusively under a regulated monopoly. The interpretation of antitrust regulation (the Sherman Act, the Clayton Act, and the Federal Trade Commission Act) is also in flux (see box 5-D). These rulings plainly have a strong influence on business structure. Rapid growth in conglomerate as opposed to horizontal mergers undoubtedly resulted in part 34 Werner Neu, Karl-Heinze Neuman, and Thomas Schnoring, Trade Patterns, Industry Structure, and Industrial Policy in Telecommunications, Telecommunications Policy, March 1987. 35 The advent of mini-mills in the steel industry is a good eXa3T3ple of this phenomenon. See Bela Cold, Technological Change and Vertical Integration, Manageria/ and Decision Economics, vol. 7, 1986. 3bBruce Stokes, Beaming Jobs Overseas, Nationa/ ~OUrfld, vol. 17, No. 30, July 27, 1985, p. 1726.

PAGE 185

181 Box 5-D.Anti-Trust Regulation Affecting Mergers and Acquisitions That no person engaged in commerce or any activity affecting commerce shall acquire, directly or indirectly, the whole or any part of the stock or other share capital [where] the effect of such acquisition may be substantially to lessen competition, or to tend to create a monopoly. The Clayton Act, Sec. 7 (15 U. S. C., Sec. 18 (1980)) Every contract, combination in the form of trust or otherwise, or conspiracy in restraint of trade or commerce among the several states, or with foreign nations, is hereby declared to be illegal. The Sherman Act, Sec. 1 (15 U. S. C., Sec. 1 (1975)) Unfair methods of competition in or affecting commerce, and unfair or deceptive acts or practices in or affecting commerce are hereby declared illegal. Federal Trade Commission Act, Sec. 5 (15. U.S. C., Sec. 435 (1982)) SOURCE: Compiled in W. D. Appler, Mergers m the Food Industry The impact on FTC Regulation, paper delivered at the The Food Update Conference of the Food and Drug Law Institute, Mega-Mergers The Impact of Consolidation Tampa FL, April 1987 from the Cellar-Kefauver Act. 37 Antitrust law has been amended to permit research consortia designed to promote U.S. competitiveness, though the act has failed to stimulate much real collaborative research. Liberal interpretations of anti-trust statutes have virtually eliminated formal barriers to mergers during the Reagan administration. In 1986, for example, 2,406 pre-merger filings were made pursuant to the Hart-Scott-Rodino (HSR) Act, three times the rate of applications in 1979. Of these, 2,108 were approved without further inquiry by granting requests for early termination of the statutory waiting period, The rest were issued letters requesting further information, and 25 HSR filings were made. The Federal Trade Commission filed court papers in only three cases between 1979 and 1985Pepsicos proposed acquisition of Seven-Up, Kiddes proposed acquisition of Horneschfeger (mobile hydrolic cranes), and Conocos proposed acquisition of Asamera. And 37 John M. Connor, Mergers in the Food Industry: Trends, Effects, and Policies, staff paper No. 87-9, Department of Agricultural Economies, Purdue University, March 1987, p. 2. in one of these cases, a consent order settled the issue before a suit was filed. 38 Demand If changes in scale and scope are made possible by new regulations and new technologies, they have also been made necessary by both dramatic changes in domestic demand and a massive invasion of imported products. There may once have been a period when large U.S. manufacturers could in effect ignore market dynamics by creating markets for their products, through advertising in a product market (like automobiles) dominated by a small number of domestic firms. This pattern of performance was encouraged by a long period of post-war business successes and government programs designed to create a climate favorable for growth built around oligopolistic activity. To some extent, the emergence of large manufacturing firms meeting relatively homogeneous markets resulted from the limitations of production technology; they may also have been encouraged by a political process that equated this form of industrial organization with progress and growth. 39 Organized labor shared this vision, since large, centrally managed firms and predictable markets provided a sound basis for stable employment. However, high levels of concentration can be justified by the economies of scale derived from large-scale operations in only a few industries. 40 Times have changed. Growing affluence and new technologies make it necessary to substitute niche markets for relatively homogeneous ones for products ranging from automobiles to health care (see ch. 3). Firms that ignore how their products connect with the consumer are increasingly at peril. If nothing else, the demand for dynamic performance (an ability to react quickly to both opportunity and disaster) has outstripped the importance of static performance (low cost in a predictable environment) .41 3 8 W.D. Appler, Mergers in the Food Industry: The Impact on FTC Regulation, paper delivered at the The Food Update Conference of the Food and Drug Law Institute, "Mega-Mergers. The Impact of Consolidation, Tampa, FL, April 1987. 39 Piore and Sable, op. cit., footnote 29. q~Bruce E. Kaufman, scale of Plant Relative to Market Size in U.S. Manufacturing, Southern Economics Journal, October 1979, vol. 46, No. 2. dlBurton Klein equates dynamic productivity with an ability to take advantage of good luck and recover from bad luck. See Burton Klein, Dynamic Economics (Cambridge, MA: Harvard University Press, 1977).

PAGE 186

182 Radically changed perceptions about the desirable scale of electric power-generating facilities provides a particularly clear example. Large 1,000-MWe plants may have been able to produce power less expensively when they were commissioned in the 1970s, but their economic success depended critically on levels of demand for electricity projected over a 20-year period. Massive mistakes in estimates have left many regions with large, unused generating capacity. Few utilities are now willing to gamble on 20-year forecasts, and are willing to pay a premium for small units that can be added to as needed. At the same time, advances in electric generating technology like steam-injected gas turbines may make it possible for 100to 200-MWe plants to compete with larger plants without incurring the risks associated with uncertainty. 42 Foreign firms have called into question traditional patterns of domestic business management never seriously questioned by domestic competition. Dynamic, entrepreneurial firms making semiconductors in Californias Silicon Valley find themselves overwhelmed by oligopolies managed by greying bureaucrats in Japan. Large U.S. manufacturers of automobiles find themselves with lower productivity than Japanese production facilities one-third their size. Large U.S. textile machinery firms find their markets picked to pieces by an invasion of aggressive producers from West Germany that work harder to identify market niches. US. farmers find their export surplus overwhelmed by imports of such products as Italian tomatoes and Brazilian orange juice. With the exception of declines in farming (a sector dominated by small businesses) each of the forces at work would tend to increase the role of comparatively small establishments if past trends continue. There are now more points of entry for smaller, specialized enterprises; small firms have moved alertly into businesses such as software development, maintenance and repair of communication and computational equipment, and computer-generated graphic design. Such firms traditionally flourish during periods of rapid transition, since the bureaucratic inertia of large firms may blind them to opportunities qZEric D. ~rson and Robert H. Williams, Steam-Injected @ Turbines, Journal of Engineering for Gas Turbines and Power, January 1987; Eric D. Larson and Robert H. Williams, Steam-Injected Gas Turbines and Electric Utility Planning, Technology and Society, March 1986. in places where none were expected. Who could have imagined a business built around graphics for personal computers a decade ago? The rules of the game are changing rapidly. Measuring Changes in Scale and Scope Scale Small employers appear to be providing an increasing fraction of jobs but a declining fraction of output. 43 The largest firms, however, appear to be expanding mainly by aggregating large numbers of comparatively small branches, or subsidiaries, which are typically involved in different kinds of businesses, Small businesses, defined here as employing 100 people or less, accounted for 98 percent of all U.S. business enterprises and 34 percent of all employment in 1984. 44 The average firm size dropped from 26.2 employees in 1976 to 21.7 in 1982. Although large firms created 1 million more jobs than small firms between 1976 and 1984, table 5-4 indicates that in relative terms most growth occurred in midsize firms, those employing between 20 and 500 people. The smallest firms (less than 20 employees) maintained their share of U.S. employment while the largest firms (more than 500 employees) lost share. The role of small business depends heavily on the sector examined. An auto plant employing only several hundred workers would be considered small, whereas several hundred lawyers would represent 43 Measuring the scale of a firm involves relying on empirical measures like jobs, assets, or outputall of which are very sensitive to definitions of small or large, the data source used and its coverage, and the methodology employed. In particular, some controversy has arisen over the role small business plays in the job generation process and the accurate classification of a business as a small establishment or a small enterprise. See Catherine Armington and Marjorie Odle, Sources of Job Growth: A New Look at the Small Business Role, Economic Development Commentary vol. 6, No. 3, fall 1982; David L. Birch, The Job Generation Process, MIT Program on Neighborhoal and Regional Change, Cambridge, MA, 1979; Candee S. Harris, Small Business and Job Generation: A Changing Economy or Differing Methodologies, working paper prepared for the Brookings Institution, February 1983; Sue Birley, Finding the New Firm, 1984 Proceedings of the Academy of Management, Boston, MA, August 1984; A.L. Walton, How Small Businesses Contribute to Job GenerationThe Pitfalls of a Seemingly Simple Question, paper presented at the 1983 Conference on Industrial Science and Technological Innovation, Evanston, IL, May 1983. 44 Nearly all of the data in this discussion come from various editions of U.S. Small Business Administration, The State of Small Business (Washington, DC: U.S. Government Printing Office, 1984 and 1986 editions).

PAGE 187

183 Table 5-4.U.S. Employment by Firm Size: 1975-84 (percent of all jobs) Firm size (number employed) 1975 1978 1980 1982 1984 Under 20....... . . . 27.l% 26.6% 26.0% 26.8% 27.1% 20 99 . . . . . 26.9 27.9 28.3 28.5 28.8 100 499 . . . . 22.6 23.2 23.8 23.5 23.5 500 999 . . . . 8.0 7.9 7.6 7.3 7.2 Over 1,000 . . . . 15.4 14.4 14.3 14.0 13.4 Total . . . . . 100.0 100.0 100.0 100.0 100.0 Number of jobs . . 60,565 70,289 74,836 74,297 77,996a NOTE: Totals may not add to 100 due to rounding. SOURCE: U.S. Bureau of the Census, Statistical Abstract of the United States, 1987 (107thed), Washington, DC, 1986, table 858. a very large law firm. Figure 5-5 indicates, for example, that manufacturing and transportation industries have a comparatively large share of firms with more than 500 employees while natural resource businesses, retail &wholesale trade, and other services have a comparatively large share of business with fewer than 20 employees. With the exception of the retail industry, the rate of job growth in firms with fewer than 20 workers was faster than average during the period when the economy was entering the recession of the early 1980s, while growth was slower in firms with more than 500 (see table 5-5). Indeed, 56 percent of all manufacturing jobs added between 1976 and 1982 were found in firms with fewer than 20 workers, even though such firms represented only 7 percent of all employment in 1976 (the average number of employees per establishment, however, remained roughly the same for all sizes). And more than 46 percent of jobs added in the finance industry were Figure 5-5.-Employment by Establishment Size and Industry in 1984 All Industries 1 Natural Resource 1 Manufacturing Transportation Trade Other Services 0 .20 40 60 80 100 Percent of employment NOTE: Natural Resources includes Construction. SOURCE: U.S. Bureau of the Census, Statistical Abstract of the United States: 1987 (107th ed.), Washington, DC, 1986, table 859. in firms with less than 20 employees, despite the fact that these firms represented only 19 percent of all employees in 1976. There were two important exceptions. The market share of builders with volumes greater than 100 housing units per year grew from 6.9 percent in 1959 to 67 percent in 1986, while the share of builders with volumes less than 25 units per year fell from 70 to 11 percent during the same period. 45 Similarly, more than 50 percent of all jobs added in retailing between 1976 and 1982 appeared in firms with more than 500 employees, even though such firms employed less than one-third of all retail employees in 1976. Changes in patterns of self-employment also illustrate the changing scale of American enterprises. Table 5-6 documents a steady decline in self-employment, with the most significant changes occurring between 1950 and 1970. This drop was led by sharp losses in self-employment in retail trade. Selfemployment in agriculture remains relatively high, but is steadily declining. The creation of owneroperator trucking and bus companies, made possible by deregulation, has presumably been responsible for the recent increases in self-employed transportation workers. The decline of self-employment in business and social services seems to have halted, but recent gains must be viewed with caution since they may represent a response to the most recent business cycle and not a long-term trend. Surprisingly, self-employment has increased in manufacturing since 1970, although the share of selfemployed persons in manufacturing remains low. In terms of assets, sales, and creation of GNP, large businesses continue to dominate and gain share. In 45 National Association of Home Builders, Profile of the Budder, Washington, DC, 1979; and Housing Focus, November 1987.

PAGE 188

184 Table 5-5.Percent Change in U.S. Employment (by size of firm, 1976.82) Change by number employed: Total change 1-19 20-99 100-499 500 + Agriculture, Forestry, & Fishing . . . 4.9%0 21 .80/0 .70/0 10.90/0 13.80/o Mining . . . . . . . . 37.6 72.1 52.3 59.2 24.1 Construction. . . . . . . . 7.9 24.8 .1 14.1 1.4 Manufacturing . . . . . . 5.3 42.7 10.7 2.1 1.1 Transport, Communication & Utilities . 13.0 33,9 11.4 8.1 10.3 Wholesale trade . . . . . . 15.2 28.8 8.2 12.7 4.7 Retail trade . . . . . . . 15.6 9.5 10.7 20.4 24.6 Finance, Insurance & Real Estate. . . 19.3 46.6 14.3 7.6 13.9 Services. . . . . . . . . 29.3 52.6 26.2 19.6 26.2 All Industries . . . . . . 15.6 29.3 13.1 10.7 12.2 SOURCE: U.S. Small Business Administration, The State of Small Business (Washington, DC: U.S. Government Printing Office, 1985), table Al .21. Table 5-6.U.S. Self-Employment by Industry (percent of all full-time equivalent employment in that Industry) Industry 1950 1960 1970 1975 1980 1986 Agriculture, Forestry, & Fishing . . . . . . . 67.7 61.3 58.9 54.9 52.4 50.1 Mining . . . . . . . . . . . . . 3.4 4.0 2.2 2.1 2.7 3.3 Construction . . . . . . . . . . . . 29.4 19.7 16.7 19.9 22.0 22.8 Manufacturing . . . . . . . . . . . 2.6 1.9 1.4 1.5 1.8 2.0 Durable goods . . . . . . . . . . . 2.8 2.0 1.5 1.7 1.8 2.0 Nondurable goods . . . . . . . . . . 2.3 1.7 1.2 1.4 1.8 2.0 Transport, Communications & Utilities . . . . . . 4.8 4.5 4.2 5.0 5.5 6.0 Wholesale trade . . . . . . . . . . . 9.2 8.4 6.3 5.8 5.4 5.1 Retail trade . . . . . . . . . . . . 25.5 21.8 13.3 12.3 11.6 9.2 Finance, Insurance, & Real Estate . . . . . . . 10.3 8.8 6.9 7.7 8.4 8.4 Services .. .. .. .. . . . . . . . . . . 19.5 18.5 15.9 14.7 14.0 13.9 Total . . . . . . . . . . . . . 19.6 15.3 11.1 11.1 10.9 10.7 SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National lncome and Product Accounts;historical diskettes, tables 6.7b and 6.10b. 1977, firms employing more than 500 people represented only 0.4 percent of all businesses but were responsible for 53 percent of all value-added in the economy. 46 In 1985, manufacturing firms with more than $l billion in assets held nearly 70 percent of all corporate assets and firms with more than one billion dollars in annual profits enjoyed nearly 70 percent of all U.S. corporate profits. 47 Moreover, the Nations largest businesses had sales that grew faster than the sales of small businesses between 1976 and 1982 in spite of the increase in small business employment. Firms with more than 10,000 employees produced sales growth of 16 percent annually from 1976 to 1982, 48 while firms with between 10 and 19 employees saw their sales grow an average of14 percent and other businesses with fewer than 500 employees averaged between 4.3 and 9.0 percent sales 46 See Thecha@ng lndlrstria] and Size Composition of U.S. Business, The State of Smal/ Bus;ness, op. cit., footnote 44, 1984. p. 118. iTStatistica] Abstract of the United States 1987, op. cit., footnote 4gThe State of Small Business, op. cit., footnote 44, 1984, p. 67. growth, This continued a trend already evident between 1958 and 1977.49 Most of the expansion of large businesses, however, appears to have occurred through net creation of new establishments rather than expansions of old ones. Seventy percent of the new jobs emanating from large businesses originated from net additions of establishment startups minus closings, and not from expanding existing establishments. 50 From 1976 to 1984, small businesses (under 100 employees), commonly thought of as the source of new establishments, derived only 60 percent of their (net) new jobs through this source (see figure 5-6). 51 Figure 5-7 shows the dramatic growth in the number of establishments owned by large firms. This 49 lbid., p. 134, 50 U.S. Smal] Business Administration, Office of Advocacy, Small Business Data Base, USEEM File, table 4, unpublished data, April 1987. 51 Bruce E. Kirchhoff and Bruce D. Phillips, op. cit., footnote 26.

PAGE 189

185 addition of establishments corresponds to the fact that the largest increase in the share of jobs was not in small establishments owned by small firms, but in small establishments owned by large firms. 52 s@oWth rates were calculated for the 1978 to 1982 time period. The State of Small Business, op. cit., footnote 44, 1984, p. 15. Figure 5-6.-Composition of 1976 to 1984 Employment Changes by Type of Change for Firms With Fewer Than 100 Employees 0 10 20 30 4 0 50 Percent = <100 employees How To Read This Figure: Of all the jobs created by new firms that started businesses between 1976 and 1984, about 40/0 were created in firms with less than 100 employees. SOURCE: U.S. Small Business Administration, Office of Advocacy, Small Business Data Base, USEEM File, unpublished data, 1987. Figure 5-7.-Average Number of Establishments per Firm by Firm Size Number of establishments 70 4 I ;; 50 40 30 20 100 1-19 20-99 100-249 250-499 500+ Number of employees per firm 195 8 = 1967 m 1972 El 1977 = 1982 How To Read This Figure: In 1958, the average firm with 500 or more employees owned 37 establishments (firms indicate ownership, establishments are separate plants). By 1972, firms with more than 500 employees owned more than 60 establishments, and owned approximately the same number in 1982. SOURCE: U.S. Small Business Administration, The State of Small Business: A Report of the President (Washington, DC: U.S. Government Printing Office, March 1984), table A2.20. scope While large firms may be producing many different products, individual establishments within these firms, as well as small independent establishments, appear to be specializing rather than diversifying. Independent manufacturing establishments showed particularly strong movement toward specialization. 53 It is difficult to determine whether large firms are assembling specialized operations in order to achieve economies of scale, for reasons having to do with the investment interests of owners, or for the way the component establishments work together. Most recent mergers in manufacturing and mining are classified as conglomerates (see table 5-7); the heterogeneity of the firms being assembled into large businesses could suggest that mergers are not being undertaken to improve integrated efficiencies. 54 On the other hand, firms might be diversifying into different components of their respective systems or networks (buying an airline as well as holding a rental-car company and a hotel chain) while retaining economies of scale at the establishment level .55 Several manufacturers have begun to sell intermediate products like machine tools and software, in addition to using them in their own operations. Rather than fitting into the traditional notion of forward or backward integration, this spinning-off of new products reflects a diagonal movement that crosses traditional technological and industry boundaries. 56 It has been noted, for example, that per dollar spent, airlines make more money on reservation systems than they do on flying passengers. 57 An analysis of the food industry, which has experienced more than 5,000 mergers and acquisitions from 1975 to 1985, concluded that most of the mergers were undertaken to market complementary products, create new products from combined know-how, and distribute the weaker merger partners products through a stronger sales network. 58 Given the com53 Frank M. Gollop and James L. Monohan, From Homogeneity to Heterogeneity: A Proper Index of Diversification, Technicaf Notes, U.S. Department of Commerce, Bureau of the Census, Washington, DC, October 1986, p. 22. Wbid., p. 2 s Ibid., p. 29. sGBela Gold, op. cit., footnote 35. 57 Helen Wheeler, Air Reservations: New Savvy in the Skies, High Technology Business, vol. 7, No. 11, November 1987, pp. 36-40. 58 William D, Appler, op. cit., footnote 38.

PAGE 190

186 Table 5.7.Merger Trends in Manufacturing and Mining, 1948-79 1948-53 1956-63 1963-72 1973-77 1978 1979 Horizontal . . . . 36.8 19,2 12.4 15.1 28.5 2.3 Vertical . . . . 12.8 22.2 7.8 5.8 15.1 7.3 Conglomerate. . . . 50.4 58.6 79.8 79.1 56.4 90.4 Total . . . . 100.0 100.0 100.0 100.0 100.0 100.0 SOURCE: Frank M. Gollop and James L. Monohan, From Homogeneity to Heterogeneity: A Proper Index of Diversification, Bureau of Census Technical Notes, U.S. Department of Commerce, Washington, DC, September 1984. plex assembly of service and production operations that must be combined to deliver products to markets, traditional definitions of vertical integration are difficult to apply and may be inappropriate. In manufacturing enterprises, it is clear that firms are diversifying production through holding a more varied portfolio of establishments that are producing a more homogeneous mix of products. Using an index that measures the number of different products produced by an establishment, it can be shown that manufacturing establishment diversification fell by a factor of nearly 3 from 1963 to 1982. 59 The diversity of products made by individual establishments fell in 17 of 20 major manufacturing categories examined. 60 In contrast, diversification of manufacturing enterprises increased during the period in 14 of 20 industries. Since 1963, this level of diversification has increased by 15 percent. 61 By 1984, 12 percent of manufacturing employment was in establishments not classified as manufacturing. These non-manufacturing establishments owned by manufacturing firms had an employment growth rate of 6 percent between 1982 and 1984, compared to a loss of 1 percent in the manufacturing establishments. 62 Consequences of Changes in Scale and Scope There is a rich literature describing the relationship between business size, management strategies, and dynamic performance, and a small theoretical literature covering economies of scope is beginning 59 Gollop and Monohan, op. cit., footnote 53, P 29. 60 Ibid., p. 25 61 lbid., pp. 25, 26, 29. 62 Marjorie Odle and Catherine Armington, "is American Manufacturing Creating Jobs Again? unpublished working paper, Applied Systems Institute, p. 3. to appear. 63 Unfortunately, most of this developing work deals with manufacturing firms, not with enterprises whose primary output is information or other non-products. The following discussion will concentrate on how new technologies may act to challenge the rules that have governed forms of business organization. Three classes of implications are considered: l l l Will the emerging structures be more aggressive in their pursuit of invention and innovation, and will they be more likely to adopt innovations when they occur? Will they be more or less likely to efficiently match resources to amenity needs? Will the new systems be more attractive to employees in that they permit greater stability, higher pay, or more unpaid benefits? Are Smaller Firms More Inventive? Are smaller firms more likely to exploit the advantages of innovation and force changes in market structures because larger firms are too sluggish or are immobilized by institutional inertia and complacent management? Or are large firms more likely to take advantage, since they are better able to un~!jee C. Freeman, J. Clark, and L.L.G. Soete, Unemployment and Technicaf Innovation (Westport ~: Greenwood Press, 1982); J.K. Galbraith, op. cit., footnote 25; Morton L Kaimen and Nancy L. Schwartz, Market Sh-ucture and innovation (Cambridge, England: Cambridge University Press, 1982); F.M. Scherer, Firm Size, Market Structure, Opportunity, and the Output of Patented Inventions, American Economic Review, vol. 55, No. 5, 1965, p. 1097; and L.L.C. Soete, Firm Size and Inventive Activity: The Evidence Reconsidered, European Economic Review, No. 12, p. 319; Woodward, Lawrence, and Lorsch, Organization and Environment (Boston, MA: Harvard University Press, 1967); T. Burns and G.M. Stalker, The Management of Innovation (London: Tavistock, 1961); E.F. Fama, Agency Problem and the Theory of the Firm, Journal of Law and Economics, 1980, vol. 88, pp. 288-307. A useful review of all this literature is provided by Ronald W. Cotterill, The Economic Efficiency of Alternative Forms of Business Enterprise, Storrs Agricultural Experiment Station, Staff Paper No. 8510, U.S. Department of Agriculture, 1985.

PAGE 191

187 dertake long-term research and transform the payoff from innovation into a cycle that produces more innovation? Both may be correct. Periods dominated by entrepreneurial activity may alternate with periods dominated by large firms. 64 It is also possible that both trends occur simultaneously in different parts of the economy. The critical question for this analysis is whether the U.S. economy is moving into a period where research is unusually beneficial to large, bureaucratic firms, or unusually likely to depend on and encourage the emergence of small, innovative ones. Michael Piore and Charles Sable argue that the pendulum has swung back to a situation favorable to the entrepreneurial model. 65 This argument suggests that the conventions which formed the basis for prosperity under the Fordist model of industrial structureand the entire structure of private and public regulation that supported ithave been fatally eroded by new technology and the challenge of foreign trade. As a result, the system must move either in the direction of smaller, more flexible, more entrepreneurial institutions that are regulated by the market, or toward greater international regulation of concentrated markets. A number of factors can act as barriers to innovation in large-scale operations: top management isolation, intolerance of eccentric ideas, short-time horizons, narrowly defined accounting objectives, and inappropriate incentives. 66 Large U.S. firms have also been accused of lethargy, inattentiveness to the market, and a bureaucratic structure that prevents flexibility and innovation. 67 Large firms may go through periods of stagnation and renewal without the challenge of entrepreneurial firms. The U.S. automobile industry, for example, was challenged not by small entrepreneurial competitors but by foreign competition. After a fAA. Phi]]ips, Technology and Market Structure (Lexington, MA: Lex ington Books, 1971); Christopher Freeman, The Economics 01 lndustria/ innovation (Cambridge, MA: The MIT Press, 1982), 2d cd., p. 210. Even schurnpeter seems to have changed his mind on the subject. Drawing on the experience of 19th century capitalists, his 1912 book emphasized the importance of the entrepreneur for bringing inventions to the market. His revised view published in 1943 emphasized the importance of large firms. Sspiore and Sable, op. cit., footnote 29. S.lames Brian Quinn, Managing Innovation, Harvard Business Review, vol. 63, No. 3, May-June 1985, pp. 76-77. 67 SeeIinterview with Malcolm Baldrige, Despite Barbs, Baldrige Hopeful on U.S. Business, Washington Post, Dec. 7, 1986; Quinn, op. cit., footnote 66; and Charles Sabel, How To Keep Mature industries Innovative, Technology Review, vol. 90, No. 3, April 1987, pp. 27-35. period of relative complacency, this industry has recently been forced to re-examine an enormous range of accepted conventions, and may soon enjoy a renaissance unshaken by domestic competitors. 68 Another complication is posed by several qualitative factors that defy precise measurement. Technologies developed by small firms tend to move rapidly into the hands of larger enterprises, whose ongoing research establishmentsand possibly better management, marketing, and production engineering capabilitiesare able to translate innovation into business success. And as chapter 9 will show, many of the large firms capitalizing recently on the ideas of small, innovative U.S. businesses have been foreign-owned. Small companies face particularly serious problems during periods of generally poor business performance. As Piore and Sable argue, while smaller firms may be in a better position to respond quickly to new market conditions by adjusting output and wagessometimes because they face little or no union opposition they also face formidable obstacles. Small manufacturing firms tend to have relatively high capital costs; 69 the U.S. Small Business Administration estimates that on average, small businesses pay a premium that is 2 to 3 percent higher than capital rates charged to larger firms. 70 Combined with external debt, high capital costs often make operating a small enterprise particularly difficult during bad times. Entrepreneurs may be tempted to react by selling intellectual assets, which can constitute the core of their comparative advantage. As has been noted, a large firm covers an enormous variety of management styles and structures. Some are collections that are in essence financial portfolios, others are combinations with a functional theme. Networks of specialized enterprises may provide the basis for real flexibility, 71 and enhanced flexibility is a crucial ingredient in fostering innovation. 72 68 William Abernathy, Industrial Renaissance: producing a Competitive Future in America (New York, NY: Basic Books, 1983). 69 W. M. Cohen, R.C. Levin, and D.C. Mowery, Firm Size and R&D Intensity: A Re-examination, National Bureau of Economic Research, Cambridge, MA, working paper No. 2205. 70 The State of Small Business, op. cit., footnote 44, 1986. 71 See Peter B. Doeringer, Make Way for Mature Industries, unpublished paper delivered at the 1987 IRRA Spring Meeting; and Quinn, op. cit., footnote 66, p. 4. 72 L. Balcerowicz, Organization] Structure of the National Economy and Technological Innovations, Acta Oeconomica, No. 24., 1980.

PAGE 192

188 On the other hand, a boutique structure may prove too fragmented to be efficient. 73 An argument can also be made that the continual process of reshuffling national assets through mergers can drain talent and capital from more productive investments. 74 From 1980 to 1985, the amount spent on mergers and acquisitions was more than double the level of industry financed research and development. In the last 6 years, General Electric has spent more than $12 billion acquiring 325 new businesses while divesting itself of 225 businesses for a sale price of $8 billion. 75 There is a danger that firms with a comparatively long-term planning horizonlarge investments in research and development, and comparatively heavy capital investmentmay be the targets of take-overs by firms attempting to realize short-term gains. If so, such mergers could be disastrous to any national strategy built around long-term commitment to innovation and worker training. 76 The key question is whether the transactional costs associated with forming new business structures lead to long-term savings, in the form of both lower transaction costs as a result of greater economies of scale and scope and more productive use of capital. Empirical support for any of these theories has been difficult to develop. Some data suggest an inverse relationship between innovation and formalization (the number of rules and specified procedures). 77 Clearly, larger firms are responsible for the lions share of all research conducted in industrial nations. Some experts argue that there are irreversible dynamic scale economiesonce a technology begins to return significant income, firms that benefit initially can then re-invest their profits and multiply their initial advantage, set de facto standards, and effectively dominate the market. 78 Nearly 60 percent of all the research reported in the United States during the 1970s was conducted by the Nations 20 largest firms; more than 90 percent was conducted by firms with more than 5,000 employees. 79 On the other hand, there is only a weak correlation between research intensity and firm size as measured in employment or sales. Small firms apparently generate more patents per R&D dollar than larger ones, and small entrepreneurs are more likely to patent an idea than large enterprises. 80 And a study looking for correlations between firm size and R&D intensity determined that most differences between firms can be explained by factors unrelated to size. Some manufacturing sectors pursue research aggressively while others failed to do so. 81 Studies of the correlation between productivity and firm size yield similarly ambiguous results. An extensive analysis of labor productivity and multi-factor productivity gains between 1947 and 1972 indicates that larger firms productivity growth has increased more rapidly that smaller firms. 82 Another study concludes that the rate of earnings per dollar of assets decreases as the size of the firm increases. 83 Moreover, empirical explanations of correlations between firm size, market concentration, and productivity growth provide few insights into how each of these variables affects the ability of firms to conduct research, invest in new plant and equipment, and train employees. Table 5-8 suggests that if innovation is measured by the number of new products produced in a specific year per million dollars worth of sales or R&D, then small firms are the most innovative. However, this measure assumes that all new products are equally important, and fails to adjust for the fact that many small business innovations (particularly hightech developments) are spin-offs from large businesses. 84 The National Science Foundation states that the popular notion of the solo inventor operating in a basement is largely fiction; 85 that those innovations having a significant social impact will 73Can America Compete? Business Week, No. 2995, Apr. 20, 1987. Ti~e James W. Brock, Bigness I S the Problem, Not the SOhNiOn, Chal/enge, vol. 30, No. 3, July-August 1987, pp. 11-25. TsTom peters, Business in the Future Sense, Washington fosf, Oct. 4, 1987. 76ebate Between T. Boone Pickens and Lester C. Thurow, The Energy Journal, vol. 8, No. 2, 1986. 77J, ROthrnan, P/arm@ and Organizing forsociid Change: Action pnnciples from Social Science Research (New York, NY: Columbia University Press, 1974). 78 Paul David, Presentation at the Symposium on Economics and Technology, Mar. 17-19, National Academy of Sciences, Palo Alto, CA, 1985. 79 Freeman, op. cit., footnote 64, p. 132. 80 Ibid., p. 136. 81 W. M. Cohen, R.C. Levin, and D.C. Mowery, op. cit., footnote 69. 82 Steven Lustgarten, Final Report to U.S. Small Business Administration on Firm Size and Productivity (Washington, DC: U.S. Government Printing Office, September 1982), pp. 16-17. 83 Stahrl W. Edmunds, organizational Size and Efficiency in the U.S., The Antitrust Bulletin, fall 1981. 84 L. Tornatzky, et al., The Process of Technological Innovation: Reviewing the Literature (Washington, DC: U.S. Government Printing Office, 1983), p. 178. 85 Ibid.

PAGE 193

189 Table 5-8.New Products First Marketed in 1982 by Size of Firm Firm size Number of products per Number of products ($ millions of net sales) $ million of net sales $ million of R&D Less than 100 . . . 0.113 3.76 100 350. . . . . 0.067 2.17 350-1,000 . . . . 0.027 1.49 1,000 4,000 . . . . 0.010 0.66 4,000 and more . . . 0.007 0.59 All firms . . . . 0.045 1.75 SOURCE: National Science Foundation, Science Indicators: 1985 (Washington, DC: U.S. Government Printing Office, 1987), p 262 probably require a concentration of experts; and that the implementation of that technology will require a centralized means of production. It is virtually impossible to measure gains resulting from mergers, or to distinguish mergers motivated by tax advantages or managerial hubris from those designed to provide real gains in productivity or production flexibility. 86 A review of recent literature concluded that there was not sufficient empirical evidence to support or refute the notion that mergers result in efficiency gains. 87 Nevertheless, the number and value of both friendly and hostile mergers have increased dramatically in the 1980s (see figure 1-16 of ch. 1). While the effect of friendly mergers on employment and wages appears to be benign, 88 and while mergers can play a useful role as a catalyst for restructuring an industry, forcing managers to be responsive to shareholder interests, 89 the effect of hostile takeovers can also shift wealth from stakeholders (communities and employees) to shareholders with little net gain to society. 90 A study of recent takeovers could not demonstrate that the takeover targets were more likely to be those with a long-term planning horizon, high cash flows, or low debt. In fact, the statistics indicated that com86 J.M. Connor, op. cit., footnote 37. 87 David J. Ravenscraft and E.M. Scherer, Life After Takeover, Federal Trade Commission Working Paper No. 139, Washington, DC, February 1986; Julian Allen, Mergers and Their Impact on Todays Economy: A Survey, U.S. Congressional Research Service, Report No. 82-118E, Washington, DC, 1982; and Julian Allen, Corporate Takeovers: A Survey of Recent Developments and Issues, U.S. Congressional Research Service, Report No. 87-726E, Washington, DC, 1987. *Gail McCallion, Mergers and Acquisitions: The Impact on Labor, U.S. Congressional Research Service, Report No. 87-705E, Washington, DC, 1987. BgMichael C. Jensen, TakeOVerS: Their Causes and Consequences, working paper, Harvard Business School, Cambridge, MA, August 1987. ~Andrei Sh]eifer and Lawrence H. Summers, Breach of Trust in HOStile Takeovers, National Bureau of Economic Research, working paper No. 32, Cambridge, MA, August 1987. pared with average firms, takeover targets had low investments in R&D, low capital/earnings ratios, and were virtually indistinguishable from the average firm in terms of cash flows and debt/equity ratios. 91 Meeting Amenities The relationship between the size and scope of enterprises in a sector and the facility with which firms can identify and reach new markets with new products in some ways mirrors the relationship between structure and innovation. But innovation in production technology by individual firms neither guarantees that the firms in a sector will work effectively together in marketing products, nor assures that they will be effective in discovering new consumer markets. Chapter 6 will address this issue in detail. Employment The discussion of large versus small firms also overlooks the quality of jobs associated with different levels of scale. Large businesses typically provide better wages, better non-wage compensation, and greater job stability. 92 Sectors characterized by a few dominant firms supported by a series of satellite firms can achieve great flexibility by forcing smaller firms to absorb gains and losses. Ways of achieving flexibility without such a high human cost are discussed in chapters 11 and 12. Firms employing more than 500 workers pay 41 percent of their workers $10.00 or more per hour, compared to the U.S. average of 30 percent and the 91 J. Pound, K. Lehn, and G. Jarrell, Are Takeovers Hostile to E CO nomic Performance? Regulation, vol. 10, No. 1, September/October 1986, p. 23. gzsee Edward M. Miller, hrge Firms Are Good for Their Workers: Manufacturing Wages as Function of Firm Size and Concentration, The Antitrust Bulletin, spring 1981.

PAGE 194

small business (25 to 99 employees) rate of only 23 large business (more than 500 employees) rate of percent. 93 In respect to providing health benefits to 85 percent. The probability of a worker becoming workers, firms with less than 25 workers provided unemployed and remaining unemployed from a this benefit to only 39 percent of workers, compared small firm (less than 100 employees) is higher than to the economy-wide average of 67 percent and the that of an employee of a large firm. 94 gsThe Stafe Ofsma]j Business, op. cit., footnote 44, 1986, Table C 18, 94Ibid., table C.17, p. 247. p. 248. THE GEOGRAPHY OF ECONOMIC ACTIVITY The growth of complex production networks is also reshaping the location of Americas economic activity. In the 1930s, it was still possible to predict the location of most economic activity by looking at a map of the United States showing only major geological and topographic features. Manufacturing was clustered around sources of raw materials, energy, and access to heavy transportation. Service activity was centered at the transfer points connecting transportation systemsriver junctions and ports. Very little economic activity in the United States is now limited by such constraints. Improved communications technology and the declining significance of natural resources increases options for both manufacturing and service establishments. Instead of spreading wealth and economic activity more evenly around the country, however, the changes appear to have resulted in a greater concentration of economic activity. Growth has been particularly rapid in regions immediately surrounding major coastal cities. Only scattered empirical evidence exists to explain these phenomena. It appears that l l l l concentration results from: a sustained need for personal (rather than electronic) communication; relatively rapid transportation within metropolitan areas, and comparatively easy access to other regions if needed or desired; a continuing interest in the comparatively good educational institutions often associated with major cities; and the fact that people seem to prefer the cultural opportunities and variety of living near a major metropolitan center. 95 95 M.L. Moss, "TeleCOrnrnUnj ca tj on S : Shaping the Future, paper prepared for the Conference on Americas New Economic Geography, Washington, DC, Apr. 29-30, 1987. The forces that would lead to greater concentration in an economy where services play a large role have been known for some time. Writing in 1960, Raymond Vernon noticed that: The most probable outcome of the increased freedom offered by swifter air travel will be the further concentration of the office elite at a few headquarters cities. This tendency will be fortified by the use of high-speed electronic data-processing machines. For these machines will contribute to the centralization of data-processing and decision-making at fewer points in the structure of the giant company. 96 Some service activities, such as health, education, and food service, are necessarily spread in rough proportion to population density, though chapter 6 documents a number of cases where even this tradition is changing: health care industries may centralize the sophisticated tertiary care hospitals while decentralizing out-patient services; technology may also permit greater geographic decentralization of large university campuses. Transactional services and manufacturing, however, are now able to become more footloose. Transactional service industries, which provide a growing share of all employment, appear to have used the potential of communications technology to decentralize most of their activities, giving rise to a complex structure of enterprises. Insurance firms are centralizing functions ranging from record-keeping to processing such relatively undifferentiated products as home and auto insurance, but are decentralizing other work to local sales offices capable of conducting sophisticated analysis and underwriting with 96 Raymond Vernon, Metropolis 1985 (Cambridge, MA: Harvard University Press, 1960), p. 84, cited in M.L. Moss, op. cit., footnote 95. See also J. Gottman, Megalopolis and Antipolis: The Telephone and the Structure of the City, in Ithia de Sola Pool, (ed.), The Social Impact of the Telephone (Cambridge, MA: The MIT Press, 1977).

PAGE 195

191 computer terminals. The dispatching of trucks may be centralized, while the ability to serve geographically dispersed freight customers increases. In manufacturing, the emergence of comparatively small batch production, just-in-time inventory systems, and a need to participate in tightly linked networks supplying goods for retail outlets may encourage greater decentralization of activity. At the same time, these factors can lead to more centralized control over production, since sophisticated communications systems and comparatively lightweight products make it possible to operate a geographically dispersed network with increasing efficiency. The ability of multinational firms to build responsive production networks around the world is a case in point. (See discussion in ch. 7.) Changes in the geography of economic activity have a number of indirect effects on the operation of the U.S. economy, including: l l l l influencing the price and quality of housing, since decentralization of employment can lower the cost of housing within a reasonable commuting range of work; changing options for improving productivity in transportation-with few exceptions, job growth has been highest in suburbs and geographic regions poorly served by public transportation; affecting opportunities for career growth, to the extent that back office functions are geographically segregated in suburbs while opportunities for advancement are most likely to be found in central offices located elsewhere; and undermining job stability and making union organization-more difficult, as more and more jobs appear in service and manufacturing firms able to move locations with comparative ease in search of attractive labor pools. Two kinds of change are discussed in the following section: first, relative changes in population, employment, and income growth in different parts of the country; and second, a continued economic movement to a constellation of population centers roughly associated with major cities, though not to urban centers themselves. Regional Movements Defining the Change Measured in terms of population growth rates, table 5-9 indicates that cities on or near the east and west coast and in the southwest dominated U.S. urban growth during the period 1970-86 as well as 1985-86. Indeed, coastal metropolitan areas appear to be drawing the greatest benefit from the increasing importance of transactional services to total U.S. output (outlined earlier in this chapter), a development that has been reinforced by the relative rise of the transactional sector as a source of U.S. jobs (see ch. 10). 97 New York City, for example, continues to act as an international center for finance, tising, and the arts, while Los Angeles is also coming a major center for finance and trade its easy access to Asia and Latin America. At the the same time, sharp population decl continued in the major manufacturing centers o midwest, particularly those affected by autom and steel manufacturing. This development m have been expected since the share of U.S. out generated by manufacturing industries payin g wageswhich contribute a large share of empl ment throughout the midwesthas dropped co erably over the past 15 years. (Again, see discu of changes in value-added share by production tor earlier in this chapter.) Looking at developments throughout the Un States in terms of personal income, the converg of regional income that characterized the 1 9 period appears to have stopped, and in some has been reversed, during the 1980-85 period. the east and west coasts continue to dominate jor transactional services, States along the At coast have regained their advantage in incomes 1980a reversal of the trend during the prev decade (see figure 5-8). Moreover, the south Atla tic region has joined the northeast and Califor a relative growth region; it appears that the Bo Washington corridor is spreading south Personal income per capita in New England the middle eastern Stateswhich fell relative to gTSee discussion of the impact of both trade and transactional services on metropolitan areas in Thierry Noyelle, A New Economic Order, Forum for Applied Research and Public Policy, vol. 2, No. 1, winter 1987, pp. 97-105.

PAGE 196

192 I I I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. 5 I . . . . . . . . . . . . . :$ :Z .0 .< :0 . . . . . . . . . . . . . . . . . . . . . . . . . ,... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . :4 .Co .= s . .$: :Z: J. :LL: . a

PAGE 197

193 Figure 5-8.-Per Capita Income of Coastal States (percent above national average) Percent 16 I 14 12 10 8 6 4 2 0 1970 1975 1980 1985 1986 ~ East coast B West coast How To Read This Figure: People living in States on the U.S. east coast in-l 970 averaged nearly 6/0 more income than the national average per capita income. By 1980 per capita income in these States was only 1.8/0 above the national average, but by 1986 it had increased so that the east coast averaged 8% more income per capita than the national average. SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Information System, tables SA1, SA2, and SA3, unpublished, August 1987 national average during the 1970srose sharply in the early 1980s and is now significantly above the national average (see figure 5-9). The far west has maintained a comparatively high level of per capita income through the period. Incomes in the major manufacturing centers of Ohio, Illinois, Indiana, Michigan, and Wisconsin (the Great Lakes region), which were above the national average in 1970, fell below the national average in 1985 for the first time. Incomes in Texas, Oklahoma, Arkansas, and Louisiana (the southwest), which had begun to approach the national average, fell sharply during the 1980s largely as a result of the declining fortunes of the domestic oil industry-while those of Kentucky, Tennessee, Alabama, and Mississippi remain 20 to 25 percent below national averages. The only region to show continuous growth between 1970 and 1985 is the southeastern coast, stretching from Maryland to Florida. 98 In citing these figures, of course, it should be noted that personal income as an indicator of economic 98 U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Information System, tables SA1, SA2, SA3, Washington, DC, unpublished. 125 100 Figure 5-9.-Convergence and Divergence of Regional Income (per capita personal income as percent of U.S. average) Percent of U. S. average 1 ---------+ 0 U S average w 1925 1935 1945 1955 1965 1975 1985 Mideast + Far West + New England Great Lakes Rocky M l + Plains + Southwest Southeast SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, Regional Differences in Per Capita Personal Income Widen in the 1980s, news release No. BEA 87-39, Washington, DC, Aug. 20, 1987. vitality can be misleading when separated from other indicators. Costs of living can vary according to region, and regional aggregations mask important movements on a local level. 99 Nonetheless, the trends revealed in figures 5-8 and 5-9 are reinforced by trends in earnings (compensation paid to employees) over the past decade. Earnings in the far west remain far above the national average, while States along the Atlantic coast have enjoyed rising relative earnings. 100 Indeed, the one-third of all 50 States located along both coasts have generated roughly three-quarters of real U.S. growth in wages and partnership income during the 1980s. 101 Explaining the Change The changes now underway in the geography of U.S. economic activity are being driven by several forces: l Information and transportation technologies ranging from nationwide computer networks to inexpensive, fast air freight services-allow easier movement of the goods and information flows needed for efficient management of activ99 Linda LeGrande and Mark Jickling, Earnings as a Measure of Regional Economic Performance, U.S. Congressional Research Service, Report No. 87-377E, Apr. 27, 1987, p. 2. 100 Ibid, p. 8. See also Lynn E. Browne, Too Much of a Good Thing? Higher Wages in New England, New England Economic Review, January/February 1987, pp. 39-51. 101 Daniel Bell, The World and the United States in 2013, Daeda/us, vol. 16, No. 3, summer 1987, p. 20.

PAGE 198

194 ities in diverse areas. l The growing contribution of transactional service industries, where many activities do not require physical proximity to clients or raw materials, permits a mobility not available in heavy manufacturing. l Within manufacturing, growth has been largely in goods with a high value per unit weight, particularly the products of many of the newer enterprises located in the northeast. 102 Such firms can transfer production facilities more rapidly than those involved in traditional heavy manufacturing, which still provide a comparatively large share of the economic base in midwestern States. l Declines in natural resource inputs throughout the economy increase the comparative importance of market access, transportation networks, economic base, labor quality and cost, tax environment, and even climate and quality of life in the selection of location. l Air-conditioning and enlightened governance in the sun belt (improving schools, roads, and other infrastructure) have made the region increasingly attractive for retirement, recreation, and enterprise location. The population shifts that ensued have carried a second wave of economic activity with them. l Geographic movement can also be strongly affected by Federal, regional, and State policies. Subsidized transportation and water helped accelerate development in many regions. State efforts to attract industry through attractive tax or regulatory programs can also play a role. While they once concentrated on efforts to attract manufacturing enterprises, States are now competing for transactional service businesses that may be comparatively footloose, Citicorp, for example, agreed to build a back office facility in Hagerstown, Maryland in order to earn the right to enter the Washington/Baltimore market. Delaware offers attractive locations for data processing and credit card centers; Marine Midland and Morgan Guaranty have established computer centers in that State. 103 102 See Daniel Bell, Op. cit., footnote 101; and Benjamin Chinitz, The Regional Transformation of the American Economy, paper delivered to the American Economic Association, May 1986. 103 M. Moss and A. Danau, Will the Cities Lose Their Back offices? Real Estate Review, vol. 17, No. 1, spring 1987. Many States are also attempting to establish environments conducive to high-technology enterprises, attracting private firms with promises of additional public assistance; 35 such State efforts were operating in 1985, compared to 4 in 1979. 104 A recent program in Pennsylvania involving public allocation of $26.4 million over a 3-year period is credited with encouraging 500 business startups and expansions in that State. 105 When all factors are accounted for, however, it appears that a region capable of providing well educated people has an advantage that is difficult to overcome through other incentives. Southern and western metropolitan areas attracted many new firms, expansions, and regional transplants during the post-war era. The west continues to grow, while the southwest has suffered somewhat due to changing conditions in the oil industry. An improved interstate highway system, the maturing of the U.S. trucking industry, and innovations in air transportation provided new forms of access to regions with lower labor costs, playing a major role in attracting business to the south and west. 106 Rapid business growth in these regions created jobs that were filled quicklyeven in tight labor marketsas workers from depressed areas took advantage of faster and cheaper means of transportation in order to arrive at areas with employment opportunities. Population growth followed. This, in turn, was followed by yet more employment growth. To support the larger numbers of enterprises and workersas well as the influx of retirees to the southern regions, who were now able to move to warm climates without breaking ties to their former places of residencedemand for service industries rose dramatically. 107 Of course, services were also a significant part of the initial movement away from the in104 Edward J. Malecki, Hope or Hyperbole? High Tech and Economic Development, Technology Review, vol. 90, No. 7, October 1987, p. 45. 105 W. John Moore, High-Tech Hopes, National Journal, vol. 18, No. 46, N OV 15, 1986, pp. 2769-2773. 106 Mark Perlman, Patterns of Regional Economic Decline and Growth (Washington, DC: American Enterprise Institute, 1982). For more on this subject, see R.D. Norton and J. Rees, The Product Cycle and the Spatial Decentralization of American Manufacturing, Regional Studies, vol. 13, 1979, pp. 141-151. IOTSee Lynn E. Browne, op. cit., footnote 100; and Richard J. Rosen, Regional Variations in Employment and Unemployment During 197082, Monthly Labor Review, vol. 107, No. 2, Feuary 1984, pp. 34-35.

PAGE 199

195 dustrial north, due to the ability of service enterprises to transfer operations with comparative ease. 108 The pattern of decline that characterized many northeastern cities during the 1970s has been slowed and in some cases reversed, buoyed by sharp growth in transactional service employment and high technology (largely medium-wage) manufacturing. The northeast, and the middle Atlantic States to a lesser extent, now enjoy economic growth rates as good or better than the U.S. average. In some cities, this has been reflected through recent population gains; as table 5-9 shows, the Philadelphia area grew by 57,000 people and the New York City area grew by 50,000 between 1985 and 1986, while the Washington, DC area continues to rank among the Nations population growth leaders. On the other hand, job growth in much of New England has been greater than the rise in population. 109 More than three-quarters of new jobs result not from moves, but from startups or expansions. 110 New England enjoyed rapid rates of growth of new business and comparatively low rates of business failures during the 1980s. 111 For many companies, the theoretical advantages of moving a facility to a low-wage region appears to have been offset by the advantages of a good educational system and other critical infrastructure. New Englands stubborn maintenance of high-tax and extensive business regulation proved less of a barrier, given highly rated State and local educational systems and other infrastructures purchased from these taxes 112 -as well as highly respected private educational institutions enjoying indirect public support through tax exemptions. Ironically, the comparative resurgence of northeastern States over the past decade may have been spurred in part by the recessions of the early 1980s. 113 New England, which had been in relative decline prior to that time, was affected less severely than other regions. Moreover, these recessions brought real interest rates to historically high levels, leading to a sharp increase in the cost of housing in the south and westregions which depended heavily on new construction and suffered from spiraling land costs. 114 Companies making high-technology products, contacted in a 1982 survey, did not feel particularly constrained by access to raw materials, energy, or climatenone of which are at a premium in the northeast-in selecting a site (see table 5-10). Access to raw materials was consistently at the bottom of the priority list, with only 27.5 percent of the respondents saying that such access was significant or very significant in their choice of a site. Instead, companies tended to choose locations on the basis of skills, labor costs, tax climate, costs of living, and several categoriesacademic institutions, transportation, and access to marketsgenerally associated with metropolitan areas. 115 High-technology firms can generally move with comparative freedom from. one region to another. Given the greater relative growth of the U.S. east and west coasts during the 1980s, this has led to a preponderance of high-technology activity around urban centers in these regions. Such a trend is reinforced by recent patterns of concentration among U.S. research and development (R&D) facilities, which tend to require higher levels of capital investment than do high-technology manufacturing enterprises. 116 Ranking the number of leading R&D centers by metropolitan area in 1987, one study has found that only one of the top eight (Chicago, ranked fifth) was not located on or near either coast. 117 The same study concludes that the prominence of the California and East Coast core areas in R&D facilities was even more pronounced in terms of hightech and microcomputer firms." 118 While job-related moves have not been the only reason for these regional shifts, table 5-11 suggests that nearly half of all interregional moves are job108R, D. Norton, op. cit., footnote 3. I@This has led to some concern that New Englands economy maY soon be constricted by a shortage in the supply of labor. See New England Warned of a Labor Shortage, The New York Times, sec. 1, Dec. 2, 1987; and Lynn E. Browne, op. cit., footnote 100. 110 Development Report Card for the States (Washington, DC: Corporation for Economic Development, March 1987). 111 The State of Small Business, op. cit., footnote 44, 1984, PP 16-17, 20-21. 112 Development Report Card for the States, op. cit., footnote 110. 113 Mark Perlman, op. cit., footnote 106. 1 14 1 bid., p. 302. 115u.s. congress, Joint Economic Committee, Location of High Technology Firms and Regional Economic Development, Washington, DC, 1982. llsEdward J. Malecki, op. cit., footnote 104, p. 46. I ITMichae] () Condor, Many Prominent R&.D Centers Favored by Facility Planners Continue 25-Year Dominance, Site Sehdorr Handbook, vol 32, No. 3, June 1987, pp. 564-572. 118 Ibid.

PAGE 200

Table 5-10.Factors that Influence Location Decisions of High-Technology Companies (percent responding that the factor was significant or very significant) Choices among different regions Percent 1. 2. 3. 4. 5. 6. 9. 10. 11. 12. Labor skills/availability . . . . 89.3 Labor costs. . . . . . . 72.2 Tax climate in the region . . . 67.2 Academic institutions. . . . . 58.7 Cost of living . . . . . . 58.5 Transportation . . . . . 58.4 Access to markets . . . . 58.1 Regional regulatory practices . . 49.0 Energy costs/availability. . . . 41.4 Cultural amenities . . . . . 36.8 Climate . . . . . . . 35.8 Access to raw materials. . . . 27.6 NOTE: High-technology Industries were taken from the following Standard Industrial Classifications: drugs, ordinance and machine, electrical and electronic machinery, equipment parts, miscellaneous transportation equipment, and measuring, analyzing and controlling instrument, photographic, medical, and optical goods, watches and clocks. Most respondents were semiconductor or telecommunication firms located in Massachusetts and California. SOURCE: U.S. Congress, Joint Economic Committee Location of High Technology Firms and Regional Economic Development, Washington, DC, 1982. related; 60 percent are either directly job-related or are associated with retirement, the armed services, or education. 119 Within this overall structure, priorities for moving differ widely according to age. Younger people tend to move to attend school or look for work more than other age groups; middle-age Americans move as a result of job transfer more than their younger and older counterparts; and the vast majority of older people relocate for retirement, climate, or family considerations. It is interesting to note that this pattern may change somewhat as the babyboom generation moves through middle age. Since two-earner families are more likely to be found among baby-boom households, more two-earner families will move into middle age. Two-earner families are less likely to relocate for employment reasons than other household types, 120 suggesting that the rate of job-related movement among middle-age Americans may decline decades. The Move to the Defining the Change over the next several Urban Fringes A move of population and employment from central cities to nearby suburbs has been underway for 119 Larry H. Long, Migration and Residential Mobility in the United States (New York, NY: The Russell Sage Foundation, 1988). 120 "Job Seekers Stay Put, The New York Times, p. D1, Oct. 4, 1987. many years. Increasing incomes have allowed Americans to escape the congestion and high cost of urban centers and to search for the amenities of suburban living, while remaining close to the cultural and economic opportunities associated with cities. In the northeast and west, much suburban growth has occurred in small areas adjacent to large metropolitan centers, while exurban growth in the south and north central regions has occurred near small and intermediate-sized metropolitan areas. 121 In the 1960s, all regions exhibited more rapid metropolitan growth than non-metropolitan growth (see table 5-12), and much of the non-metropolitan movement came from population increases in counties adjacent to metropolitan areas. The differences narrowed during the 1970s, although non-adjacent counties still grew less rapidly than adjacent ones. Nonetheless, non-adjacent counties were growing more rapidly than urbanized counties for the first time, in all regions except the south. Indeed, the 1970s saw non-adjacent rural areas of less than 2,500 people grow by 14.6 percent, after such areas had decreased in population by 4.2 percent during the 1960s. On a regional basis, these smallest-sized areas grew more rapidly than did larger non-adjacent counties in the northeast, south, and west. Such a remarkable change in population growth suggested further non-metropolitan population reconcentration, representing a break with past trends. 122 While the population of all central cities grew by only 0.1 percent during the 1970s, central cities in SMSAs (standard metropolitan statistical areas) with populations of less than half a million grew 10.6 percent. Central cities in SMSAs with a population of more than 1 million declined by 4.2 percent or more. 123 IZILarry Long and Diana DeAre, The Economic Base of Recent population Growth in Nonmetropolitan Settings, U.S. Department of Commerce, Bureau of the Census, Washington, DC, 1982. IZZD. R vining, Jr. and A. Strauss, A Demonstration That the Current Reconcentration of Population in the United States is a Clean Break With the Past, Environment and P/arming A, vol. 9, 1977, p. 751. IZsJOhn F. ~ng, Population Reconcentration in the United States, U.S. Department of Commerce, Bureau of the Census, Washington, DC, 1981.

PAGE 201

Table 5-11 .Reasons for Inter-Regional Moving, 1979-1981 (in percent, by age of household reference person) Age of reference person: All ages Under 25 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70+ Job transfer . . 22.2% 14.8% 25.0% 28.4% 32.6% 30.2% 24.7% 23.7% 15.4% 9.7% 0.0% O.O O /O Look for work . . 6.3 9.7 5.9 5.6 6.7 4.4 5.0 6.2 6.6 1.5 0.0 2.9 Take new job . . 18.7 21.8 21.7 21.6 19.3 18.7 19.1 16.0 9.0 6.5 0.5 0.8 Armed Forces. . . 3.4 6.9 5.2 2.5 0.8 3.5 1.1 0.0 0.0 0.0 0.0 0.0 Retirement . . 2.4 0.0 0.0 0.1 1.4 2.9 3.8 2.8 10.7 23.3 14.9 6.2 Attend school . . 5.6 15.4 6.9 2.9 2.4 0.0 0.6 0.0 0.7 0.0 0.0 0.0 Closer to relatives . 8.6 4.2 6.9 7.2 6.0 5.7 8.6 12.8 13.7 14.9 21.5 44.7 Change climate . 6.0 2.9 4.7 3.2 5.3 4.6 3.6 13.2 15.0 14.1 30.2 13.4 All other. . . . 26.8 24.4 23.8 28.5 25.5 30.0 33.5 25.3 29.0 30.0 32.8 32.0 Total (percent) . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Total (000s). . . 6,250 1,362 1,352 988 694 471 338 252 235 203 143 211 NOTE: Total percent may not add to 100 due to rounding. SOURCE: Larry H. Long, Migration and Residential Mobility in the United States (New York, NY: The Russell Sage Foundation, 1988), table 7-2.

PAGE 202

198 Table 5-12.Population Growth by Region by Metropolitan and Non-Metropolitan Counties (1960=70 and 1970-80, in percent) County population change Total Northeast North-central South West 1980-70: Total. . . . . . . . . 13.4% 9.80/o 9.60/o 14.3% 24.20/o Metropolitan . . . . . . . 17.1 10.0 13.1 22.2 28.4 Non-metropolitan Total . . . . . . . . 3.9 8.4 2.0 2.6 9.0 Adjacent to metro area. . . . . 6.2 10.0 4.3 13.3 Nonadjacent to metro area . . . 1.0 3.5 0.8 5.0 1970-80: Total. . . . . . . . . 11.4 0.2 4.0 20.0 23.9 Metropolitan . . . . . . . 10.0 1.8 2.6 21.5 22.1 Non-metropolitan Total . . . . . . . . 15.4 12.8 7.4 17.3 31.6 Adjacent to metro area. . . . . 16.7 13.1 8.6 19.6 34.4 Not adjacent to metro area . . . 13.8 11.8 6.0 14.7 28.8 NOTE: SMSA/NECMA boundaries as of January 1, 1980. SOURCE: Larry Long and Diana DeAre, The Economic Base of Recent Population Growth in Nonmetropolitan Settings, U.S. Bureau of the Census, 1982. The trend toward suburbanization among U.S. metropolitan areas as a whole slowed between 1980 and 1984, while central cores actually gained population share. 124 Nonetheless, it still appears that urban fringe areas are alive and well, expanding in every region of the country-from Tysons Corner, Virginia to the South Coast Metro Center in Orange County, California. 125 Jobs have followed the population to the suburbs and smaller cities. Between 1967 and 1977, suburban employment growth accounted for 81.7 percent of all job growth in the 50 largest SMSAs, while central city shares of total metropolitan employment in the 50 largest SMSAs fell from 63.6 to 53.4 percent. 126 Total central city employment grew 7 percent over this period, while suburban employment grew 59.2 percent. And between 1970 and 1980, the most pronounced declines in the number of people commuting from homes in central cities were found in the 25 largest cities (see table 5-13). In smaller SMSAs, though most suburban growth rates again exceeded those of central cities, average IziCharles L. Leven, Post-Industrialism, Regional Change and the New Economic Geography, paper prepared for the Conference on Americas New Economic Geography, Washington, DC, Apr. 29-30,1987. IzsWil]iam K. Stevens, Defining the Outer City: For NOW, call It a Hybrid, The New York Times, sec. 1, Oct. 12, 1987, reporting a conference of the Urban Land Institute, lzGRobyn S. phi]]ips and Avis C. Vidal, The Growth and Restructuring of Metropolitan Economies, Journal of the American Family Planning Association, summer 1983, p. 295. growth rates were less divergent. Central city employment for 50 selected small and midsized SMSAs increased by 23.7 percent between 1967 and 1977, as opposed to the 7 percent growth of the 50 largest SMSA S 127 Moreover, between 1970 and 1980 there was a sharp increase in the number of people living in suburbs of major cities and commuting to jobs outside the city center. (Again, see table 5-13; for more on this subject, see discussion of transportation in ch. 3.) Metropolitan areas gained much of their job growth between 1975 and 1979 as a result of increases in service employment (see table 5-14). The dependency on services was even more pronounced in the largest metropolitan areas. For those with 3 million or more in population, services accounted for 41 percent of all jobs generated between 1975 and 1979. While Citicorp moved some of its back office functions to the Dakotas, this appears to be an exception rather than a rule. Of the top nine banks contacted in a regional survey, seven had located operations for high-volume check processing and credit card operations in suburban or satellite cities but most were within 60 miles of a city center. 128 Securities firms originally concentrated in downtown Manhattan because of the need to physically 1 27 Ibid. 128 M. Moss and A. Danau, op. cit., footnote 103.

PAGE 203

199 Table 5.13.Changes in the Location of U.S. Jobs and Homes Workers living in the Workers living in largest 25 urbanized areas other urban areas 1970 1980 1970 1980 Living in a central city and working in: CBD a . . . . . . . . . . . 6.6 /0 4.7 /0 5.60/o 6.50/o Non-CBD portion of central city . . . . . 30.0 24.7 45.9 40.1 Outside of a central city . . . . . . 8.9 6.8 12.9 10.7 Living in the suburbs and working in: CBD . . . . . . . . . . . 3.2 3.7 1.7 2.8 Non-CBD portion of central city . . . . . 13.1 14.3 13.4 16.4 Outside of a central city . . . . . . 30.0 45.7 20.4 23.4 Total (percent) . . . . . . . . . 100.0 100.0 100.0 100.0 Total (millions) . . . . . . . . . 25.5 30.7 17.0 25.3 a CBD = Central Business District. SOURCE: U. S. Department of Energy, Urban Mass Transportation Administration, Demographic Change and Recent Worktrip Travel Trends (UMTA-DC-09-7009), Washington, DC, February, 1985 deliver certificates. Although only one of the Nations top 10 securities firms has back offices and headquarters in same buildings, and only 2 have data processing in their corporate headquarters complex, most of these firms are not moving out of the New York metropolitan area. 129 Rather, they are moving to New Jersey or other close sites in New York metropolitan area Morgan Stanley moving to Brooklyn, or Paine Webber to Weehauken, New Jersey. Dean Witter was an exception, moving to Dallas. Significant differences exist in geographic employment shifts within industries (again see table 5-14). Much of the decentralization of manufacturing, such as IBM to rural Vermont, has resulted from decentralized management (described earlier in this chapter). Branch plants are often located in suburban areas distant from a companys main plants or headquarters; one study found that most branch plants were spawned from corporate headquarters in the manufacturing belt, even in the southwest. 130 Thus, part of non-metropolitan industrial growth seems to occur with the decentralization of production processes into peripheral areas of the manufacturing belt. 131 Similarly, data-processing and client-aid services, which can be transferred from central offices in order to save on labor and real estate costs, can move easily because of advanced telecommunications networks. 129 lbid. IJORodneV A, Erickson and Thomas R. Leinbach. Characteristics Of Branch Plants Attracted to Nonmetropolitan Areas, Nonmetropolitan Industrialization, R.E. Lonsdale and H.L, Seyler (eds.) (Washington, DC: V.H. Winston & Sons, 1579). 131R.D, Norton and J. Rees, op. cit., footnote 106. 13zM, MOSS and ,4. Dtinau, op. cit., footnote 103. Explaining the Change At the turn of the century, elevators, telephones, indoor plumbing, and other technologies made highrise office buildings and apartments possible, thereby opening possibilities for highly concentrated urban centers. The national highway system, built during the 1950s, literally paved the way to suburban development, just as railroads had opened the West to development two generations earlier. In each case, regulations and public action played a major role. Western development was a conscious act of public policy encouraged by subsidies to transportation systems, free land, rural electrification, and a variety of other programs. Suburban development was shaped by highway policy and zoning. Undoubtedly, the desire to escape the problems of urban centers also played a major role. One of the features of the emerging U.S. economy is that the rules governing the shape of American cities and towns may be changing. An economy increasingly dependent on transactional services, and a manufacturing system where rapid growth can occur in relatively small facilities or facilities with relatively modest freight requirements, allows greater flexibility in locating businesses close to areas where employees can find attractive housing, schools, and recreational facilities. (Again, see ch. 3 for a discussion of changing patterns of transportation between work and living.) Relocation along these patterns, however, may be contributing to a dilemma of major proportions (discussed further in ch. 11). Suburban movement can

PAGE 204

200 Table 5.14.Change in Employment for Metropolitan and Non-Metropoiitan Counties and Types of Non.Metropolitan Settings, 1975.79 (In percent, by major industry group) Industry Metro Non-metro Non-adjacent Adjacent Agriculture, forestry, and fisheries . . . 45.7 28.1 35.9 18.5 Mining . . . . . . . . . 21,8 13.5 14.0 13.1 Contract construction . . . . . . 34.8 44.2 48.1 40.1 Manufacturing . . . . . . . . 15.8 20.7 20.1 21.5 Transportation and public utilities . . . 14.5 30.9 31.2 30.5 Wholesale trade . . . . . . . 19.4 21.4 21.0 21.8 Retail trade . . . . . . . . 22.2 26.6 26.7 26.6 Finance, insurance, and real estate . . . 20.0 29.6 31.0 28.0 Services . . . . . . . . . 32.9 32.6 33.2 31.9 Total employment. . . . . . . 22.5 25,2 25.8 25.6 NOTES: SMSA/NEGMA boundaries as of January 1, 1980; employment data from County Business Patterns; non-adjacent means not adjacent to SMSA; adjacent means adjacent to SMSA. SOURCE: Larry Long and Diana DeAre, The Economic Base of Recent Population Growth in Nonmetropolitan Settings, U.S. Bureau of the Census, 1982. leave behind the significant share of a regions population living in central cities, where attaining a job often requires the mobility provided by a car due to the limited nature of public transportation from cities to suburbs. Moreover, in the sense that the new employment opportunities within coastal cities are largely related to transactional services, and require higher levels of education than traditional manufacturing jobs, the comparatively low educational standing of inner-city residentsmany of whom are minoritiesmay preclude them from these positions. 133 Nonetheless, much suburban activity continues to be tied to the cultural and economic opportunities afforded by cities. And it appears that a significant amount of the movement away from urban centers has resulted from the increased merger activity discussed earlier in this chapter. Following the opportunities for growth in suburban and exurban areas, many firms have not only expanded the scale of their operations and the scope of their productsthey have branched out physically as well. 134 Recent patterns of suburban investment may arise from a variety of other factors: 135 l l l l l the availability of low-cost labor; employers perceptions that worker productivity and dependability are greater outside urban areas; lower unionization levels, and a perception that workers are less likely to unionize; non-economic reasons, such as the simple desire to relocate, that are often given as preferences for less urbanized areas; and policy decisionsrural areas enjoy a variety of subsidies relative to more densely populated regions, as highways, telephone service, electric service, police, fire, and other services in rural areas are often heavily subsidized; prices would rise sharply if low-density areas were forced to pay the real marginal costs of these services. Itswi]liam J. Wi150n, The Truly Disadvantaged: The Inner City, the Underclass, and Pubiic Poky (Chicago, IL: University of Chicago Press, 1987). See also Truman A. Hartshorn and Peter O. Muller, Suburban Business Centers: Employment Implications, U.S. Department of Commerce, Economic Development Administration, Washington, DC, 1987. ltqMi]ford B. Green, Corporate-Merger-Defined Core-Periphery Relations for the United States, Growth and Change, vol. 18, summer 1987, pp. 12-35. 135 Some of these factors are discussed in L.Steven R. Kale and Richard E. Lonsdale, Factors Encouraging and Discouraging Plant Location in Nonmetropolitan Areas, Nonmetroplitan Industrialization, op. cit. 130.

PAGE 205

Chapter 6 The Networks That Produce Amenity

PAGE 206

. CONTENTS Page Food. ........ . . . . . . . . . . . . ....203 prospects . . . . . . . . ..........................203 Structure and Performance. . . . . . . . . .............205 New Technologies and System Integration. ................................211 Health . . . . . . . . ..............................212 Prospects .. .. .. .. .. .. .. .. .. $ . . . ..........................212 Structure and Performance. . . . . . . . . .............212 New Technologies and System Integration . . . . . .............220 Housing . . . . . . . . . . . ....................222 Prospects .. .. .. .. .. .. .. $ . . . . ............................222 Structure and Performance. . . . . . . . . . .........222 New Technologies and System Integration . . . . . . . . ..226 Transportation . . . . . . . . . . ..................229 Prospects .. .. .. .. .. .. .. .. .. .. .. $ . . ...........................229 Structure and Performance. . . . . . .........................229 New Technologies and System Integration . . . . . . . . ..233 Clothing and Personal Care . . . . . . . . ................236 Prospects . . . . . . . . ............................236 Structure and Performance. . . . . . . . . ...........237 New Technologies and System Integration. ................................237 Education . . . . . . . . ............................240 Prospects . . . . . . . . ............................240 Structure and Performance. . . . . . . . . . ........241 New Technologies and System Integration . . . . . . . . ..242 Personal Business and Communication . . . . . . . . ....251 Structure and Performance. . . . . . . . . ...........251 New Technologies and System integration . ...........................263 Recreation and Leisure . . . . . ..............................264 Prospects . . . . . . . . . . . . . .. ....264 Structure and Performance . . . . . . . . . . .. ....265 Manufacturing . . . . . . . . . .......................273 Factors Forcing Change . . . . . . . . ................273 Network Components and Productivity Change. ............................277

PAGE 207

Chapter 6 The Networks That Produce Amenity The previous two chapters described broad patterns of change in the national economy. While there are themes common to all business networks, each is changing in unique ways. A close examination of each network is needed to understand many of the changes taking place. These examinations provide a basis for hypotheses about the future that could not be constructed by extrapolation. They also offer a different perspective on the practical choices facing public policy makers. In some cases, the discussions reveal the need for reforms that apply only to specific business sectors. In others, they show how programs designed to facilitate national economic growth will help or hinder productive restructuring in a given sector. The sector discussions in this chapter test national themes to see how they apply in practical cases. Each addresses the following questions: l l l l l How does the network operate as an integrated system, combining goods and services to connect primary resources to final markets? How are the roles of the different component businesses within the network changing? How are the connections between them changing the performance of the whole? How does the scale and scope of enterprises within the network affect the production process? What is the significance of new developments in the geographic location of production facilities? What is the relative importance of technology, trade, regulation, and other forces shaping change in the networks? l What shape may these networks take in the future? l What practical choices affect the direction the networks take? Each discussion represents, in effect, one column of table 4-6 in chapter 4. Organization of the chapter by amenity group maintains a clear view of the integrated performance of networks. Examinations of specific business sectors comprising the larger amenity network appear as follows: l l l Most amenity networks make heavy use of inputs from one production sector. An analysis of the current and possible future performance of such business sectors therefore appears in the section covering the related amenity group. The discussions of Health and Education, for example, cover many issues relevant to the Social Service sector. Transactional Activities are treated in the discussion of Personal Business and Communication. Where appropriate, the amenity discussions also contain a review of intermediate demand for the products of the primary business sectors analyzed. While manufacturing issues appear throughout the amenity discussions (e.g., the manufacture of textiles and apparel is discussed in the Clothing and Personal Care section, and food manufacturing in Food), a final section pulls the pieces together and provides an overview of changes in manufacturing. Prospects fraction of all American spending on the Food amenity goes to add value after food Products leave Most farm products are produced by businesses farming operations. This value comes in the form managed more like manufacturing facilities than the of increased variety in fresh foods as well as varyfarm operations of an earlier generation. A growing ing degrees of processing. Processing alternatives now include such things as tablecloth restaurants, fast food, and a range of frozen and fresh entrees IMuCh of this di~ussion is drawn from U.S. COngreSS, Office of Tech nology Assessment, Food, sector study, Washington, DC, 1987. in grocery stores. 203

PAGE 208

204 The American food production and distribution system has been reshaped by changing consumer incomes and expectations (see ch. 2). Parts of the industry never before exposed to foreign competition find themselves struggling to maintain market share. Imports of processed foods (such as confections, alcoholic beverages, and processed foods with an ethnic appeal) have grown so rapidly that they now approximately equal U.S. food exports (which depend almost entirely on bulk commodities and products like meat and meals that require little processingsee ch. 9). Imports of food processing equipment and food processing licenses have also grown. Technology has had an uneven effect on the businesses that comprise the food network. Farm productivity and the productivity of many food manufacturing activities has grown rapidly, while productivity growth in wholesale and retail enterprises has been sluggish. Technologies likely to have the most profound effect on the operation of the American Food system may not appear as measured productivity gains in any component business (as will be the case with the provision of Clothing and Personal Care). Information technology can tie retail, wholesale, transportation, food processing, and farm operations together in ways that facilitate system-wide flexibility and productivity. New packaging and preservation techniques can improve the quality of products reaching consumers and can reshape the structure of the businesses moving farm products to forks. Several different outcomes are possible during the next 20 years: l The food production system could grow in away that would make it a tightly integrated, hightechnology production network. Improved management systems can be implemented through information technology and lowered costs made possible by new, safe packaging and preservation technologies, which allow a greater variety of fresh products to be available at comparatively low costs. Foods can be tailored to regional or ethnic tastes, for special dietary needs (e.g., low-sodium or low calorie) and for other specialized markets (e.g., elderly individuals looking for a high-quality menu that can be prepared at home with little effort)-all without a significant increase in cost. Competition on the l l l l basis of quality could increase consumer knowledge of the health implications of their diets. The flexibility of the new communication system could allow a variety of comparatively small producers of farm products and small manufactured food processors to enter networks once reserved for very large firms. Improved on-farm technology, coupled with improved packaging, communication, and transportation networks, could lead to increased productivity in specialized food products like cultivated fish and many varieties of fruits and vegetables. Capital equipment could replace many clerical and low-skill tasks in retail, and wholesale operations could be automated. With the possible exception of meat processing, the productivity of food processing facilities could increase to the point where dangerous and low paid occupations would be largely replaced by machinery. Lack of effective competition could decrease product quality and the potential of new efficiencies may not rebound to the advantage of consumers. Foreign products could provide much of the variety available on store shelves. Large oligopolies, investing little in research or new equipment, could compete largely on the basis of advertising rather than product quality. Low-income and rural areas could find themselves facing a sharp decline in real choice and an increase in real food prices. Productivity and food preservation could be achieved by methods with adverse health effects. Oligopolies in food services could replace regional and ethnic diversity with homogeneous national products. Small farms could be reduced to little more than hobbies with virtually all food value produced on very large farms (many of which may be owned by a single family). The post-harvest businesses could be divided between sophisticated manufacturing enterprises on the one hand, and meat-packing, wholesale, grocery, and restaurant operationswhich show virtually no productivity gains and provide large numbers of poorly paid jobson the other. These alternatives are obviously not mutually exclusive. Signs of each can be found in existing trends. The role of government in the Food network varies greatly from sector to sector. The farming industry

PAGE 209

205 is heavily influenced by government programs through both heavy Federal funding of research and massive price support programs. Federal farm programs totaled $60 billion between 1981 and 1986. 2 Federal regulations provide extensive if uneven coverage of food-product safety. Structure and Performance The system that brings food to American forks provides business throughout the U.S. economy. Of the approximately $427 billion of value-added in the economy generated directly and indirectly by Food consumption in 1984, nearly 40 percent ended up in the transportation and wholesaling industry, grocery stores, and restaurants. Approximately 15 percent resulted from farm and other resource inputs. Food manufacturing contributed approximately 17 percent of value-added. During the past decade, farming, other natural resource inputs, and manufacturing have lost their share of the American food dollar (measured in constant dollars), while transportation, trade, and transactional enterprises have gained (see figure 6-l). The Farm Sector Technology and Productivity. -Enormous productivity gains in farming were among the most obvious symbols of economic transformation at the turn of the century. During the past 60 years, the U.S. agricultural sector has undergone two technological revolutions the mechanization of farming, and the widespread use of chemical inputs. Less than 3 percent of the U.S. work force feeds the U.S. population and provides a large surplus for export. A century ago, the majority of American workers were employed on farms. The system is far from running out of ideas. Rates of productivity growth are expected to be rapid for the next two decades. 3 Agriculture is probably on the verge of a third technological revolution, stemming from advances in biotechnology, informationprocessing technology, and a variety of other innovations. 2 Economjc Report of the President, 1987 (Washington, DC: U.S. Government Printing Office, 1987), p. 154. 3 U.S. Congress, office of TechnoloW Assessment, Technolo~, Pub/ic Policy, and the Changing Structure of American Agriculture, OTAF-285 (Washington, DC: U.S. Government Printing Office, March 1986). Figure 6-l. -Value-Added To Meet Demand for Food ($428 billion* in 1984) Natural Resources Construction Low Wage Manuf. Medium Wage Manuf. High Wage Manuf. Transport & Trade Transactions Personal Services Social Services o 10 20 30 40 50 Percent of $428 billion l Constant 1980 dollars. SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4). A study analyzing the potential impacts of 150 emerging technologies on future agricultural production predicts strong productivity growth throughout the next two decades. 4 In the near term, the largest increases are forecast for the dairy sector, with milk productivity projected to increase at an annual rate of 3.9 percentcompared with 2.6 percent over the past two decades. The application of biotechnology to crop production is expected to proceed more slowly, with widespread adoption forecast after the year 2000. Scale and Scope. It is not surprising that enormous growth in farm productivity led to massive structural changes in American farming. The new production systems changed activities on the farm, requiring sophisticated management practices, new technical expertise, and heavy investment in capital equipment. They have also reshaped the economics of large and small farming operations. The question of whether new agricultural technologies have selectively benefited large or small farms is a matter of dispute and recent litigations What cannot be disputed is that the number of farms in the United States has decreased by nearly two-thirds since the pre-World War II peak of 6.5 million. Of the 2.3 million farms that remained in 1985, 5 percentabout 100,000 farmsaccounted 4 Ibid. 5 Decision of the Superior Court of Alameda County, CA, California Agrarian Action Project, Inc. et al., v. Regents of the University of California, Nov. 17, 1987.

PAGE 210

206 for half of the Nations agricultural production, up from 42 percent in 1960. In terms of net farm income, the largest farmsthose with annual gross revenues of more than $500,000, or 1.2 percent of all farmsclaimed two-thirds of total income in 1982. In contrast, farms with annual revenues of less than $99,000, which constituted 86.5 percent of all farms in 1982, accounted for less than 2 percent of net farm income. Moderate-sized farmsthose with annual sales of between $100,000 and $199,000, and perhaps the closest match to the traditional notion of the family farmrepresented 8.1 percent of all farms in 1982, and claimed a 14.6 percent share of net farm income. Consolidation of farm enterprises is expected to continue through the remainder of the century. By the year 2000, small and part-time farms with annual sales less than $100,000 are likely to represent 80 percent of all U.S. farms but may earn only 4 percent of cash receipts. Large-scale farms, on the other hand, with annual sales greater than $250,000 only 14 percent of all farmsmay earn 80 to 90 percent of all cash receipts. 6 Geography.-The geography of farm activity is heavily influenced by public programs. Farm product costs vary widely around the nation. In 1982, for example, wheat grown in the northeast cost $2.26 per bushel to produce, while producers in the northern Plains States had costs as low as $1.25. Small farming operations in some areas were able to produce at lower costs than large farms in others. 7 It is possible for farms in several regions to be profitable because public programs regulate prices. Subsidized irrigation and policies allowing the mining of groundwater also have the effect of encouraging farm activity in areas which would otherwise find profitable agriculture difficult. Food Manufacturing The food manufacturing sector plays a critical role in the Food system, purchasing 70 percent of U.S. agricultural production and largely determining the mix of products available to consumers. It is also the most concentrated sector in the Food network. &Technojo~, pub];c Policy, and the Changing Structure of American Agriculture, op. cit., footnote 3 7 Ibid; and U.S. Congress, Office of Technology Assessment, A Review of U.S. Competitiveness in Agricultural Trade, OTA-TM-TET-29 (Washington, DC: U.S. Government Printing Office, October 1986), p. 46. Technology and Productivity. -Labor productivity in the food manufacturing business has grown so rapidly during the past several decades that it has outstripped demand growth. Employment in the industry actually declined between 1950 and 1985. 8 With an appropriate research program in food process engineering, annual productivity increases of 2.5 to 3.5 percent are easily possible in food manufacturing during the next decade. 9 Innovations in food manufacturing have had the effect of reducing a variety of factor costs in addition to improving labor productivity, and are likely to continue to do so. Table 6-1 summarizes a study that covered a variety of innovations under development in the food manufacturing industries. Most of the innovations resulted in a faster process that made more efficient use of energy and raw materials, and most required more capital equipment. Almost without exception they resulted in both improved quality and lower costs. The rapid growth in imports of manufactured food products and the poor performance of U.S. exports in this area (see ch. 9) do not result from an imbalance in food processing technology. [n fact, the productivity of the U.S. industry is probably higher than that of most U.S. trade competitorsincluding those in Western Europe. 10 New technologies are expected to have a major effect on the physical processing of food products and the way the products are packaged. Most result from straightforward advances in production engineering. Bioengineering may result in a curious variety of radically new technologies. Artificial citrus juice, for example, has already been produced in lab oratories. 11 Intensive efforts are being made to develop calorie-free cake and ice cream. 12 8 U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, Table 6.7b. 9 U.S. Congress, Office of Technology Assessment, Agricultural Postharvest Technology and Marketing Economics ResearchA Technical Memorandum, OTA-TM-F-21 (Washington, DC: U.S. Government Printing Office, April 1983), p. 47. 10 Center on Transnational Corporations, Transnational Corporations in Food and Beverage Processing (New York: United Nations, 1982), cited in John M. Connor et al., The Food Manufacturing Industries (Lexington, MA: D.C. Heath & Co., 1985), pp. 22-23. 11 J. Flynn, Want Some O.J.? Its Fresh from the Test Tube, Business Week, No. 2973, Nov. 17, 1986, pp. 160-161. 12 "Will Foods of the Future Be Safe? The New York Times, Nov. 19, 1986, p. C15.

PAGE 211

207 Table 6-1.The Impact of Innovation in Food Manufacturing (simple counts of primary and secondary impacts) Reduced use Increased use of Impacts on production inputs of factor factor Labor . . . . . . 325 8 Energy . . . . . . 128 9 Packaging materials . . . 126 23 Plant time . . . . . 238 10 Transportation . . . . 26 0 Equipment . . . . . 96 269 Impacts on product quality Higher quality Same quality Lower quality Lower price . . . . . 28 14 2 Same price . . . . . 26 4 0 Higher price . . . . . 77 5 2 a a Both cases occurred in packaged fluid milk production. SOURCE: John M. ConnoR "Market Structure and Technological Opportunities in the U.S. Food Manufacturing lndustry," contract report prepared for the Office of Technology Assessment, March 1986. New packaging and preservation techniques present a major opportunity for savings. Losses of fruits and vegetables during transportation and storage have been estimated to be as high as 30 percent of the total supply 13 Packaging may also contribute as much as 30 percent of the value-added to processed foods. 14 Packaging costs exceed the cost of food products in beer, soft drinks, breakfast cereals, frozen specialties, canned soups, baby foods, and pet food. 15 Packaging weight and bulk contributes significantly to transportation and warehousing costs. If nothing else, the lack of standardization in packaging results in a situation where transportation, wholesaling, and retailing must handle from 2,500 to 5,000 different container sizes. 16 A major shift in containers is already underway. Glass and metal containers contain a shrinking fraction of the fruit and vegetables purchased by American shoppers. 17 Plastics and other polymeric materials will continue to displace glass and metal containers. New technology has the potential to in13 A~r;CultUra] postharvest Technology and Marketing Economics ResearchA Technical Memorandum, op. cit., footnote 9, p. 47. 14 T.W. Dowries et al., The Impact of New Technologies on the Food Packaging and Preservation Industries, contract report prepared for the Office of Technology Assessment, 1985. IS John h4. connor et al., op. cit., footnote 10, P. 38. ISC,W. Abdul]a, potential Effects of Standardized Packaging $Wterns on Grocery Manufacturing and Distribution, contract report prepared for the Office of Technology Assessment, 1985, p. 64. ITcanned fruits ancl vegetables declined from 20 percent of per caPita consumption of 1979 to 18 percent in 1984. Canned fruits (not including juice) declined from 13 percent in 1979 to 11 percent in 1984. U.S. Department of Commerce, International Trade Administration, 1986 U.S. Industrial OutLook, Washington, DC, January 1986. crease food quality, prolong shelf-life, and reduce shipping weight and volume. Multi-layer packaging can provide long shelf-life without refrigeration for many products. Aseptic processing and packaging, which involves fast heating and sealing, is now used in fruit juice and is being introduced for wines, fruit purees, dairy products, tomato products, and edible oils. Packaging with precise mixtures of oxygen and carbon dioxide can allow fresh fruit to ripen to, but not beyond, desired levels. A freshness indicator can provide visible warning to merchants and consumers when fish products are not safe. Irradiation of food to preserve freshness has promise in some areas but has encountered consumer resistance. Economies of scale will be pronounced for relatively small, free-standing irradiation facilities. Low-unit costs, however, can be realized only with fairly high processing levels. Consequently, only large firms may be able to justify investment in irradiation. But jointly owned facilities, centrally located in important agricultural areas, could be an economic option for smalland medium-sized processors. 18 Many if not most of the innovations that will lead to productivity growth in food manufacturing are unlikely to be developed by the food manufacturing industries themselves. The industry spends only about 0.4 percent of sales on research and developmentfar lower than the average for all manufac18 R, M, Morrison and T. Roberts, Food irradiation: New ferSpeCtiVeS on a Controversial/ Technology, contract report prepared for the Office of Technology Assessment, December 1985.

PAGE 212

208 turing. Between 1969 and 1977, 90 percent of patents applicable to the six major food manufacturing industries were granted to government laboratories, educational institutions, individuals, foreign firms, or domestic companies outside food manufacturing. 19 Between 1971 and 1977, only 12 percent of the Putnam awards for innovations that increase efficiency of food processing were given to U.S. food manufacturers; 45 percent of awards were given to small firms. A study of these awards shows clearly that food processing firms use mergers as a substitute for in-house researchone-quarter of the Putnam awards went to firms that were acquired shortly before or after getting the award. 20 Scale and Scope. The number of food manufacturing firms has been decreasing at an average annual rate of 2.5 percent since 1947. Concentration varies widely. Only 9 makers of chewing gum exist today, but there are nearly 2,000 wholesale bakers. 21 Much of the growth of large firms has been achieved through acquisitions rather than internal expansion. With $14 billion spent on purchasing, 1985 was a boom year for mergers; tobacco companies began to diversify. R.J. Reynolds spent $5 billion to acquire Nabisco. Phillip Morris acquired General Foods for $5.6 billion. Nestle S.A. purchased Carnation for $3 billion. Beatrice purchased Esmark for $2.7 billion in 1984. 22 The discussion in chapter 5 showed the difficulty of linking firm size to rates of innovation. A careful study of innovation in food processing suggests that innovation increases with firm size until total assets reach $125 million to $150 million, when they begin to decline. 23 Another study indicates that rates of innovation also increase with market concentration levels, until four firms have captured 50 to 60 percent of the market, but decline thereafter. 24 One concern about concentration in food processing is that firms in food manufacturing invest more 19W. F. Muel]er, J. (u]be~wn, and B, Pwkham, Market s~ruc~ure a~~ Technological Performance in the Food Manufacturing Industries, University of Wisconsin, Monograph 11, NC-1 17, Madison, Wl, 1982. Z Ibid. ZIJ. M Connor et a]., The Organization and Performance of the Food Manufacturing Industries, in B.W. Marion, cd., The Organizatjon and Performance of the U.S. Food System (Lexington, MA: D.C. Heath and Co., 1986), p. 21]. ZZ19g6 U.S. /ndustria/ Outfook, op. cit., footnote 1 T, P. 40-1. 2JMUeller et al., op. cit., fOOtnOte 19. Z4J. M connor et al., The Food Manufacturing hrdustries, Op. Cit., fOOtnote 10, pp. 322-323. heavily in advertising to differentiate products than in research to produce real innovation. Food manufacturers typically spend at least ten times as much on advertising as on research and development. 25 In 1979, firms in food and tobacco manufacturing accounted for 32 percent of spending for advertising, but only 12 percent of manufacturing receipts. The largest four firms accounted for 21 percent of advertising in 1982. 26 The meatpacking industry deserves special attention, since it has been an exception to many trends in the industry. With the exception of beef, most meat processing has moved out of the retail stores and into specialized processing centers. Meatcutters preparing beef products are the last skilled trade to have a presence in grocery stores. The prospect of moving these jobs to factories would complete a separation between retail and food processing businesses (see box 6-A). Distribution sectors Wholesale and retail trade in food has grown rapidly. Grocery stores and restaurants gained share of retail sales between 1975 and 1985, while most other retail sectors lost share. 27 While some productivity gains have been measured in these businesses, technology plays a comparatively minor direct role. The indirect effects of the information technology entering these businesses, however, may be large. Wholesalers have moved slowly to use computers to control inventories, monitor shipments, schedule work, plan storage layouts, and dispatch trucks. A variety of technologies, such as automatic handling equipment, have the potential to improve productivity of food wholesale operations by as much as 50 percent. 28 Explanations for the slow rate of adoption vary. Is it lack of competition, inefficient management, or the shortage of trained personnel in wholesale firms capable of managing a transition to new technology? Do the large number of products and container types make automation impractical? The debate cannot be resolved easily. 25 Ibid., pp. 23, 87-89. Zblnterna] Revenue Sewice data, cited in J.M. Connor et al., The Food Manufacture@ industries, op. cit., footnote 10, p. 81. zT]9g6 U.S. /ndus/rjal Outlook, op. Cit., footnote 17, p. 57-1. 213G. Grinnell, and L. Friedman, Productivity Potential in Dry Grocery Centers, U.S. Department of Agriculture, Economic Research Service, AER 484, Washington, DC, 1982.

PAGE 213

209 Box 6-A.Boxed Beef In 1920, the four largest beef packers were vertically integrated and commanded 49 percent of all sales. An anti-trust settlement in 1920 dictated a long process of change. By 1970, the combined share of four largest packers fell to 15.8 percent. In the 1950s, improvements in refrigeration technology and other changes led to greater centralization of cattle processing in feedlots. Retailers like Safeway began consolidating meat cutting into local processing centers. In 1967, the independent (and non-union) Iowa Beef Packing Co. (now IBP, Inc.) opened a semi-automated slaughtering and processing plant in Dakota City, Nebraska. Boxed beefs share of total beef production from all sources is now probably more than 40 percent. Rapidly expanding fast food chains have provided eager markets for their products. The number of meatcutters and butchers employed in retailing has dropped significantly between 1970 and 1980. 2 Largely as a result of centralized facilities which allowed more sophisticated equipment, meatcutting productivity nearly tripled between 1950 and 1981, measured in beef and pork packed per worker. Partial automation reduced skills and heavy lifting, although the tasks were made much more specific. Injury rates and turnover remain high, while the industry no longer requires a large pool of skilled butchers and meatcutters. 3 1 J. McCoy, Livestock and Meal Marketing (Westport, CT: AVI Publishing, 1979), p. 180. U S. Department of Labor, Bureau of Labor Statistics, Employment Projections for 1995, Bulletin 2197, Washington, DC, March 1984, p. 48. 3S.W. Hiemstra, Technological and Organizational Changes in the U.S. Beef Packing Industry, contract report prepared for the Office of Technology Assessment, 1984 Food reaches consumers through two channels: food retailers (primarily grocery stores), and food service businesses (restaurants and institutional food outlets). Concentration is increasing in both sectors. However, some convergence exists between the sectors as the variety and extent of processing offered by grocery superstores increases. Many of these stores now offer a variety of prepared salads and entrees ready for eating. Management changes have played a major role in the food service industries. Fast food had the effect of substituting customer labor for service jobs. Technical improvements led to some productivity gains in restaurants between 1958 and 1976, with the introduction of microwaves, deep-fat friers, and other equipment, but productivity has actually fallen since 1976. 29 In food retailing, advanced electronic cash registers and scanners have improved productivity in check-out, thoughas the later discussion will showproductivity changes are difficult to measure in individual stores. Scale and Scope.-The number of wholesale establishments was trimmed by nearly half between 1950 and 1982, and consolidation shows no sign of abating. The eight largest general wholesalers accounted for 26.5 percent of sales in 1982, as compared with 16.2 percent in 1972. Yet the consolidation of warehouse ownership may actually increase competition if larger, sophisticated firms compete in the same region. A recent U.S. Department of Agriculture study concluded that there are more local suppliers now than in the mid seventies. 30 The total number of grocery stores has fallen after peaking in the mid 1970s. While conventional supermarkets still claim the majority of grocery sales (59 percent in 1984), their role is rapidly declining. 31 Superstores, which account for only 3.7 percent of all grocery stores, account for 28 percent of grocery sales. 32 These stores may have 30,000 to 200,000 square feet of sales area and as much as $1 million per week in sales. At the other end of the spectrum, the number of convenience stores, where the average sale is $1 to $3, have tripled between 1963 and 1984, and now account for 12 percent of all grocery sales. 33 Specialty stores like bakeries have steadily lost share. In 1982 they commanded only 6 percent of grocery sales. 34 Smaller grocery stores are owned increasingly by smaller firms rather than national chains. Z9R.B. Carries and f-f. Bran, Productivity and New Technology in Eating and Drinking Places, in A BLS Reader on Productivity, U.S. Department of Labor, Bureau of Labor Statistics, Bulletin 2171, June 1983, pp. 67-72. 30w.B, Epps, IIF~ Who]esaling, Food Marketing Review, 1985, U.S. Department of Agriculture, Economic Research Service, AER 549, Washington, DC, March 1986, p. 21. tlpR Kaufman, F~ Retailing, Food Marketing Review, 1985, U.S. Department of Agriculture, Economic Research Service, AER 549, Washington, DC, March 1986. tz19g6 Industrial Outlook, op. cit., footnote 17, P. 57-7. jtKaufman, op. cit., footnote 31, P 26. 34 B.W. Marion, cd., op. cit., footnote 21, p. 295.

PAGE 214

210 The larger stores are located almost exclusively in suburban areas. Urban and rural areas may suffer declining choices. Inner-city residents often pay substantially more for similar items than shoppers in suburban supermarkets. A study in Hartford, Connecticut found that a family of four forced to shop at local stores paid $1,500 more annually than if they had gone to suburban markets. 35 Many inner city residents do not have automobiles and are forced to shop in local stores. The nature of products provided in new, larger grocery facilities has changed significantly. Many have become multi-purpose retail outlets providing pharmacies, teller machines, and a variety of food specialties as well as standard grocery products. A recent survey found that 75 percent of all supermarkets offer health/natural food and gourmet food. At least one-third have low-calorie selections. There is sharply declining interest in price brand generic products, the share of which fell from 17 percent in 1982 to 14.6 percent in 1985. 36 The new grocery stores may also serve as local meeting places and social centers. Many command customer loyalties if only because the time saved by going to a convenience store outweighs the savings that could be achieved by careful shopping. A recent survey indicated that only 1 in 10 consumers said they would shop at a store other than their principal supermarket to get advertised specials. 37 Grocery stores also serve as major financial centers. Supermarkets cash one-third of all non-government checks. 38 Food service is an enormous enterprise in the United States, providing nearly 1 out of 13 U.S. jobs. The activities fall into the categories shown in table 6-2. Approximately 70 percent of sales are commercial, the rest in institutions. Because of demographics, there has been a decline in food served in educational facilities and an increase in food sold through ssHartford Citizens Research Educational Network, The poor pay More: Food Shopping in Hartford, Hartford, CT, 1984. sg]gg6 U.S. Industrial Outlook, op. Cit., footnote 17, P 57-7. sT~uis Harris and Associates, Inc., Trends: Consumer Attitudes and the Supermarket, survey conducted for the Food Marketing Institute, Washington DC, 1985. 3SR.E. ONei]], whats New in ElT, Progressive Grocer, August 1985, pp. 59-66. Table 6-2.Sales of Meals and Snacks in 1977 and 1982 (percent of all sales in year) Type of sale 1977 1982 Commercial . . . . 69.1 70.6 Restaurants, lunchrooms . 30.2 30.7 Fast food . . . . 24.9 26.4 Other . . . . . 14.0 13.5 Institutional . . . . 30.9 29.4 Educational institutions & day care . . . . . 10.3 8.6 Hospitals & care facilities . 7.7 8.2 Plants, office buildings . 4.4 4.5 Military. . . . . . 2.0 1.7 Vending machines . . 3.1 2.8 Other . . . . . 3.4 3.6 SOURCE: Michael Van Dress, U.S. Department of Agriculture, private communication, U.S. Congress, Office of Technology Assessment, Food, sector study, Washington, DC, 1987, hospitals and long-term care facilities. Fast food is rapidly gaining share. Table 6-3 shows that there has also been a strong movement toward restaurant chains. If anything, the table understates the power of chains, since many independently owned establishments are franchises. Franchises account for 35 percent of all commercial sales. About two-thirds of these franchises are owned by individuals. 39 As a result, nearly half of the sales shown for -unit firms in table 6-3 may be sales by franchised operations. The gap between annual sales of franchises and annual sales of independent restaurants is growing rapidly. 40 39 U. S. Department of Commerce, Bureau of Industrial Economics, Franchising in the Economy 1982 -1984, Washington, DC, January 1984. 40 J. R. Schmelzer, The Commercial Foodservice Sector: Trends in Growth and Market Structures, Working Paper Series WP-56, Michigan State University, East Lansing, Ml, September 1981. Table 6-3.Distribution of Eating Places (sales, in billions of dollars) Number of units 1963 1967 1972 1977 1982 1 a . . 80.4 77.4 65.9 59.5 52.3 2-3. . . 5.8 5.3 6.3 6.8 7.1 4-10. . . 2.9 3.7 4.7 6.1 7.2 2-10. . . 8.7 9.0 11.0 12.9 14.3 11 or more 11.0 13.6 23.2 27.6 33,4 Fast food. . 14.6 19.0 30.3 37.6 39,4 a lncludes franchises (see text). SOURCE: J.J. Putnam, H.R. Linstrom, and M.G. Van Dress, Food Service, in Food Marketing Review, 1985, U.S. Department of Agriculture, Economic Research Service, AER 549, Washington, DC, March 1986, p. 44,

PAGE 215

211 New Technologies and System Integration A variety of new technologies promise to improve the integrated performance of the American Food network in ways that are difficult to demonstrate by examining productivity in individual business sectors (as with Clothing and Personal Care). The two classes of technologies most likely to affect the integrated performance of the food production and distribution system are new packing and preservation technologies (already discussed), and the use of modern communications and information processing systems. Computers, optical scanners, and other electronic information technologies are now common, but only recently have firms directed their attention toward uses that could pay system-wide benefits. Foodsystem firms were among the first to take a cooperative approach toward extending the applications of computers. The most visible product of their efforts is the Uniform Price Code (UPC), the now-ubiquitous bar code that is read by point-of-sale scanners and portable scanning guns or wands that record incoming shipments and are used to perform shelf inventories. Retailers, through their new ability to analyze costs and profits in previously unattainable detail, have benefited most directly from these innovations. But the biggest windfall is the large volume of highprecision data it generates, and this information may serve as the basis for improving the entire food marketing system. Electronic price scanners using the universal product code were first introduced in 1974. It is likely that half of all grocery items were scanned in 1986. 4] For ing l l l food retailers, the system promises the followtypes of advantages: direct savings through reduced checkout time (by 40 percent) and error rates; an ability to make rapid price changes (some stores are moving to electronic price indicators on shelf); improved coupon management (in 1983 customers redeemed 5.6 billion coupons worth $2 billion; an estimated 20 percent of coupons are 41 P.R. Kaufman, op. cit., footnote 31, pp. 26-27. processed in error 42 ); l reduced error rates in billing and ordering (at $11 to $18 per adjustment, price-related errors on about half of the invoices issued by food manufacturers impose annual costs conservatively estimated to total $100 million 43 ); and l precise identification of the effect of advertising campaigns and changes in store format and product placement. The systems may eventually make it easier for firms to provide the government with requested information. If customers begin to pay for food with credit cards or direct debit cards, sales can also be linked with demographic and income characteristics with great (some would say Orwellian) precision. For food manufacturers, many of whom purchase scanning data from supermarkets, the information is used to guide marketing strategies, identify new market niches, and evaluate potential changes in products that will boost sales. And for food wholesalers, some of whom use in-house computers to evaluate the efficiency of their operations, feedback on storage and distribution costs may spawn measures to improve efficiency. Perhaps most importantly, the systems allow a detailed performance evaluation of the thousands of products on their shelves. These evaluations can provide estimates on the real cost of selling each item, preferred stocking patterns, and even information about which warehouse to use as a supplier. At present, few retailers take advantage of the potential to guide merchandise strategies. 44 A more recent development is the Universal Communication Standard (UCS), with which food marketing firms can substitute computer-to-computer communication (electronic data interchange) for paper-based exchanges of invoices, purchase orders, and reams of other information generated throughout the system. 4zf-f. Mona[, Misredemption of Coupons: A New Solution, Trim, ]nc, Los Angeles, CA, 1984. qsRona]d Cotteri][, Effects of Electronic Information Technology on Employment in the Food Manufacturing and Food Distribution Industries, contract report prepared for the Office of Technology Assessment, 1985. ldFood Marketing ]nstitute, Retailer Applications of Scanning Dafa, 1985, cited in C.E. Morris, Supermarkets: Super Data, Food Engineering, May 1986, p. 120; and Food Marketing Institute, Statement of Food Marketing Institute before the Bureau of Competition, Federal Trade Commission. Milwaukee, WI, Sept. 11, 1984, cited in B. Marion, cd., op. cit., footnote 21.

PAGE 216

212 Now in the early stages of adoption, the UCS, which is a set of protocols that allow one firms computer to send information to the computer of another, should streamline distribution and marketing functions. It should also eliminate redundant clerical tasks, and some of the duties of the manufacturers sales force. In all, the communication network is estimated to offer annual savings of $196 million to $324 million (1979 dollars), assuming the technology is used by the 10,000 largest firms in the food marketing system. Most of the savings would result from reductions in clerical staffs, inventories, and waste and spoilage. Moreover, the network can accelerate recalls of tainted food items, as firms now have the capability to track items throughout the marketing system and, sometimes, back to the farm. The impact of the communication network on competition is not clear at present. It is clearly possible that large, sophisticated firms will be able to make more effective use of the analytical power of the equipment because of greater access to national data bases and because they can afford to invest in market analysis. They could also carefully monitor regional price competition and quickly cross-subsidize products to beat smaller competitors. Electronic versions of price fixing are possible. On the other hand, the networks make it possible for large firms to keep track of larger numbers of products and suppliers. This may make it easier for small producers to sell through large stores. HEALTH 45 Prospects The high cost and complexity of new health care technologies have been largely responsible for the massive reorganization of the Nations health care network that has occurred since the end of the Second World War. Experiments in management and cost containment have proliferated. During the next two decades, the system operating to provide health care could take any of several different courses. They include: A system where institutional incentives throughout the economy are clearly linked to providing the best possible health outcomes at the lowest cost. This means a system capable of making an even-handed assessment of investments in different kinds of measures to promote health or prevent illness (including investments in air and water quality and occupational safety, and advice on diet, exercise, and other aspects of lifestyle), and clinical treatment of illness when it occurs. l A system where a patients care is diverse and flexible, and can be integrated in a way that ensures optimum health care. This involves a mix45 Much of this discussion is drawn from U.S. Congress, Office of Technology Assessment, ( Health, sector study, Washington, DC, 1987. l l l ture of home care, treatment by specialized clinics, and local out-patient services with highly sophisticated medical centers in a way that is sensitive to the needs and desires of individual patients. A system that provides timely and accurate records on the effectiveness of alternative treatment strategies. A system which attempts to contain increases in health costs resulting from a growing elderly population needing medical care, and a growing array of expensive new medical producures and equipment with increasingly baroque regulations. A system that provides an enormous range in health care quality depending on the patients ability to pay, permitting only the most affluent to enjoy real choice about courses of treatment. Again, these alternatives are not mutually exclusive. They depend heavily on public decisions. Structure and Performance About 1 dollar in 10 in the U.S. economy is spent on the Health amenity, either by individuals or the government. This figure does not include spending aimed at improving health, such as investment in environmental protection or safety, healthy food, or

PAGE 217

213 fitness. More than half of all value-added in health care enterprises remains in the hands of the Social Service sector, with the rest being spent on such things as hospital construction, purchases of medical equipment, and supplies from High and Medium Wage Manufacturing firms. Purchased Transactional Activities (such as insurance and legal fees) account for nearly 16 percent of Health costs (see figure 62). Within the health care enterprises themselves, however, production recipes are heavily dependent on labor; nearly three-quarters of the value of the output of such firms is derived from value-added, most of which is in the form of employee compensation. Factors Forcing Change Until the 1950s, the structure of the Nations health care system was relatively easy to understand. The system was highly fragmented, highly individualistic, and free of anything but self-regulation by professional societies like the American Medical Association (AMA), which guarded their prerogatives jealously. The courts ruled that anti-trust regulation could not be applied to these organizations since they were professional societies. 46 Individual physicians treated most patients in a private practice and resorted to relatively small com4Klark Havinghurst, The Contributions of Antitrust Law to a Procompetitive Health Policy, in Jack A. Meyer, cd., Market Reforms in ffeafth Care (Washington, DC: American Enterprise Institute, 1983). Figure 6-2.-Value-Added To Meet Demand for Health ($285 billion* in 1984) Natural Resources Construction Low Wage Manuf. Medium Wage Manuf. High Wage Manuf Transport & Trade Transactions Personal Services Social Services o 10 20 30 40 50 60 Percent of $285 billion = 1972 ~ 1984 l Constant 1980 dollars SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4) munity hospitals (nearly always operated as not-forprofit organizations) for difficult surgery and some specialized assistance. In extreme cases, they might have turned to a larger hospital in a major urban center. The range of treatment was limited; the family physician played a central role. Most people paid for health care in cash. In 1950, two-thirds of the bill was paid directly by the recipient of the service. Explosive growth in the scope and cost of medical services occurred following World War 11, and a complex tangle of programs was put in place to regulate them. Rapidly advancing medical technology opened a wide range of therapeutic options with proven value. The high cost of these opportunities required new approaches to financing. The HillBurton program initiated in 1946 encouraged hospital construction. Health care for the needy and the elderly was greatly expanded under the Medicare and Medicaid Acts of 1965 (although gaping holes remain in the coverage). Private insurance programs grew rapidly, encouraged in part by Internal Revenue Service (IRS) rulings that employer contributions for health insurance were not taxable as employee income. 47 The public and private insurance systems operated in a way that almost completely insulated the payer and provider from estimates of cost, since they reimbursed the provider for virtually any measure taken. By 1982, 73 percent of all health bills were paid by government agencies or private insurance companies. 48 There has also been rapid growth in both the number of cases treated and the intensity of care provided. Between 1960 and 1982, the number of physicians per capita grew 46 percent; the annual number of graduates in medicine grew 128 percent; the annual number of nurses graduated grew 146 percent; and the number of community hospital beds per capita grew 22 percent. 49 Hospital costs grew from 25 percent of health expenditures in 1940 to 40 percent 47 Alan Enthoven estimates that revenue loss to government because of non-taxation of employer contributions to health insurance and medical expense deduction is as large as the Federal contribution to Medicaid and almost half the size of Medicare. See Alexander M. Capron, Allocating Finite Resources: Questions of Equity and Access, in M.E. Lewin, ed., The Health Policy Agenda: Some Critical Questions, (Washington, DC: American Enterprise Institute, 1985). 48 U.S. Department of Health and Human Services, Health: United States (Washington, DC: U.S. Government Printing Office, December 1985). 49 Ibid.

PAGE 218

214 in 1980. Using a relatively narrow definition of health care spending, health costs grew from 6.1 percent of the U.S. gross national product (GNP) in 1965 to 9.8 percent in 1981. The reasons for this growth are difficult to disentangle. Costs increased both because more people had access to the health care system and because advances in medical science increased the range of treatment. Rapid technical advances continued to drive the price of recommended practice to ever higher levels. Decisions about how much is enough health care became critical. Professional review organizations presided over the processes used to designate treatments as prevailing professional custom and practice. 50 As the complexity of the system grew, so too did the range of possible mistakes. Among other things, it was no longer possible to assume that a physician, competent at the time of original certification, would be competent in the revolutionary changes sweeping the medical field. 51 Consumers, increasingly aware of the range of their choices, began to question decisions made in their name. The result was regulation by tort law in the form of malpractice litigation. The issue of prevailing professional custom and practice came under courtroom examination, with the perverse effect that physicians were forced to include costly procedures even when there was only a small chance that the patient would benefit. Though the courtroom was a poor place to negotiate the issue of how much is enough, it took on this role by default. This helped transform a decentralized and diverse system. 52 During the 1970s, renewed interest in public health issues accompanied the explosive growth in clinical medicine. The National Highway Traffic Safety Administration and the Environmental Protection Agency began operation in 1970; the Consumer Product Safety Commission in 1972; the Mining Enforcement and Safety Administration, the Drug 50 C. Havinghurst, Decentralizing Decision Making: Private Contract vs. Professional Norms, in Jack A. Meyer, cd., op. cit., footnote 46, p. 24. 51 Regulation of ones collegues requires enormous fortitude, and there is every indication that it has not proven very effective. A peer review undertaken by the American Medical Association indicated that as many as 10 percent of active physicians were so impaired that they presented a danger to their patients. See The New York Times, June 4, 1984, p. A17. 52 C. Havinghurst, op. cit., footnote 50. Enforcement Administration, and the Occupational Safety and Health Administration in 1973; and the Nuclear Regulatory Administration in 1975. 53 Capital spending for air and water pollution abatement rose from $700 million in 1965 (about 1 percent of total plant and equipment investment) to a peak of nearly $5 billion in 1975(4 percent of total new plant and equipment) (all in 1972 dollars). The Surgeon Generals office became more aggressive in forcing the country to recognize the dangers of smoking. As medical costs reached 10 percent of GNP and health care enterprises grew to the size of major national corporations, it was no longer possible to maintain the fiction that the industry could be governed by the camaraderie of professional organizations. Taxpayers complained about the growing costs of Federal and State programs. Employers found that their contributions for health insurance co-payments had risen from 2.2 percent of wages in 1970 to 5.3 percent in 1984. 54 There was concern that an inefficient national health system was placing an intolerable burden on the costs of American products. There was also a reaction to the enormous costs of pollution abatement. Attempts to make the health care delivery behave more like an ordinary market by creating a more perfect match between the interests of patients and the interests of privately owned firms in freely operating markets have proven to be extraordinarily difficult. For reasons discussed in the Health section of chapter 3, individuals face enormous problems in making informed choices about appropriate levels of health care. Even experts disagree about appropriate levels. Attempts to introduce greater market freedom have come to be coupled with extensive and complex regulations governing both price and practices. The result has been a system that is at once more free and more heavily regulated than the one it replaced. Federal payments, for example, are now governed by a precise schedule of hospital fees for diagnostic groups and plans are underway to regulate physician fees. Private insurers are negotiating with preferred providers for health care. At the same time, a 1984 survey of 1,115 firms indi53 S. Breyer, Regulation and Its Reform (Cambridge, MA: Harvard University Press, 1982). 54 Emp]oy& Benefit Research Institute, Private Expenditures to Con tain Health Care Expenditures, EBRI Issue Brief No. 55, Washington, DC, June 1986.

PAGE 219

215 cated that 97 percent were taking specific steps to reduce health care costs. 55 Network Components The structure of the health industry has become increasingly complex in recent years, responding to the explosion of new technologies, to new professional specialties, and to the radical changes in the financial rules under which the health system operates. In general, the effect has been a proliferation of specialized treatment facilities, reduced use of hospitals, and increased use of out-patient treatment. Use of capital equipment has steadily increased. As in other service-oriented sectors, however, human skills are increasingly accompanied by sophisticated equipment. Figure 6-3 illustrates how labor costs have declined as a fraction of all hospital costs during the past 15 years. The decline is surprisingly unaffected by the radical changes in health care financing. The drive toward efficiency has forced increases in the size of both hospital establishments and groups of physicians surrounded by a complex constellation of supporting organizations like testing laboratories, dialysis centers, hospices, nursing homes, and home health care providers. Patterns of ownership are in flux, as some of these specialized facilities are owned by large firms while many are independent. 55 The Wyatt Co., Group Benefits Survey, Washington, DC, 1984. Figure 6-3.-Direct and Indirect Labor Costs as a Percent of All Community Hospital Costs Percent of all costs 69 67 63 -062 percent per year trend + 61 + + + + 59 + 57 1 1 1 I 5 5 1 T r 1 I 1 1972 1974 1976 1978 1980 1982 1984 SOURCE: American Hospital Association, Hospital Statistics, Chicago, IL, 1986 edition. Changes in the financial system have also resulted in a movement toward private rather than public ownership. Even publicly owned facilities are managed more like private firms. While there are differences in the quality of health care delivered in different parts of the country, the opportunities for health care businesses obviously follow population movements and provide employment and investment opportunities around the country. The rapidly growing areas of the south and west have furnished particularly attractive opportunities to health care enterprises well matched to the new financial rules. The increasing size of hospitals has reduced easy access to these facilities, but a variety of outpatient clinics, urgi-centers, doc-in-a-box, and other health facilities have proliferated in suburban areas. Rural areas are most likely to suffer a decline in health care access. At issue is whether the benefits of economies of scale in large facilities outweigh the benefits of short travel times to smaller facilities. Although the effect of the changes just described on the quality of care delivered has proven difficult to measure over the entire Health network, the following examination of individual components reveals some interesting trends. Physicians.-Cost containment by insurance programs, an actual decline in use of physician services resulting in part from rising prices, 56 and a potential surplus of physicians are in the process of restructuring the series of independent enterprises that were once at the center of medical practice in the United States. As a result, competition has begun to play a major role. The AMAs resistance to advertising has faded to a prohibition against deceptive practices and creation of unjustified expectations. 57 Physician offices have extended their office hours, and some even make house calls. The most dramatic effect of competition, however, has been to force independent physicians into group 56 Between 1976 and 1981, the average number of yearly physician visits per person declined from 4.9 to 4.6 while physicians in officebased practice rose 33.3 percent. See U.S. Department of Health and Human Services, National Center for Health Services Research, Contacts with Physicians in Ambulatory Settings: Rates of Use Expenditures and Sources of Payment, Data Preview 16, DHHS Publication PHS 833361, Washington, DC, 1983. 57 Robert M. Veatch, Ethical Dilemma of For-Profit Enterprises in Health Care, in Bradford H. Gray, cd., The New Health Care for Profit (Washington, DC: National Academy Press, 1983), p. 128.

PAGE 220

216 practices. Instead of moving from medical school into lucrative private practices, an increasing number of certified physicians are finding that they must accept corporate employment paying $35,000 to $40,000 per year. 58 Aside from private practice, the different ways that physicians are now organized include: l Preferred Provider Organizations (PPOs) and Individual Group Practices. The PPOs are loosely knit groups of private practitioners and hospitals that contract with employers and insurers to provide care at a discount, ranging from 5 to 30 percent on physician fees and 7 to 15 percent on hospital fees. 59 Physicians are typically paid on a salary basis. It is difficult to determine whether cost savings are real, 60 but the introduction of price competition undoubtedly puts pressure on prices. Between December 1984 and June 1985, enrollment in PPOs was estimated to have increased from 1.3 to 5.8 million. 61 l Health Maintenance Organizations (HMOs). HMOs began with the encouragement of the HMO Act of 1973, which required businesses with 25 or more employees to offer an option for HMO care. HMOs have managed to replace professionally dictated standards with their own. Although Federal subsidies have stopped, HMOs have grown rapidly. Between 1981 and 1985, enrollees increased from 10.3 to 18.9 million, and membership may reach 40 to 50 million by 1990. The geographic pattern of their growth has, however, been uneven, and they have begun to dominate the practice of medicine in some parts of the country-in some large metropolitan areas 75 percent or more of all doctors participate in HMOs. HMOs are most ac58 Eli Ginzberg, The Destabilization of Health Care, New England Journal of Medicine, vol. 315, No. 12, Sep. 18, 1986, pp. 757-760. 59 D. Ermann and J. Gabel, Preferred Provider Organizations: performance Problems and Promise, Health Affairs, vol. 4, No. 1, spring 1985, pp. 24-40. 60U.S. congress, General Accounting Office, Constraining National Health Expenditures: Achieving Quality Care at an Affordable Cost, GAO/HRD-85-105, Sept. 30, 1985; and U.S. Congress, Office of Technology Assessment, Payment for Physician Services: Strategies for Medicine, OTA-H-294 (Washington, DC: U.S. Government Printing Office, February 1986). 61 "PPO Enrollment Jumps Dramatically in 1985, Hospital Week, vol. 21, No. 40, October 1985, p. 1. tive in urbanized areas of Hawaii, Minnesota, California, and Oregon, and are least important in rural areas in Mississippi, Montana, and Wyoming. 62 Most are new; in 1983, only half had been in business for more than 5 years. 63 HMOs are organized using an entirely new principle in medical finance: they are self insuring. Patients pay a flat fee and rely on the HMO bureaucracy to provide needed treatment. This principle, coupled with the fact that the organizations can employ professional business managers to replace physician-dominated management of traditional health facilities, allows them to consider costs and benefits from a unique perspective. Not only do they create a unique regulatory environment from which to view alternative procedures, they create a unique bureaucratic environment for the physicians employed. There have been suggestions that one of the explanations for the lower rates of aggressive treatment administered at HMOs is that they do not attract the kind of personalities that practice aggressive medicine, and do not provide the kind of peer support for such actions that might be available in other kinds of practices. 64 While data is ambiguous, HMOs appear to be able to provide health care at lower cost. Estimates indicate that HMOs can offer a wide range of services at prices 10 to 40 percent lower than fee-for-service expenses, even though the HMO patients receive as much ambulatory service (4.42 physician visits v. 4.19 fee-for-service). 65 HMO hospital admission rates were 40 percent lower than fee-for-service rates, while hospital stays averaged the same length. In 1981, average number of hospital days per thousand population for all payers was 1,316for HMO members, it was 458. This could not entirely be 62 Grady Wells, Healthy Growth for HMOS, American Demographics, vol. 6, No. 3, March 1984, pp. 34-37, 46-47. 63 N. Baker, J. McGee, and M. Shadle, HMO Status Report, 1982-1983, Interstudy, Exelsiour, MN, August 1984. GIC. Havinghurst, op. cit., fOOtnOte 50. 65 Emp]oy& Benefit Research Institute, Op. cit., fOOtnOte 54; H.S. Luft, How Do Health Maintenance Organizations Achieve their Savings? New England Journal of Medicine, vol. 298, 1978, pp. 1336-1343; and W.G. Manning et al., A Controlled Trial of the Effect of a Prepaid Group Practice on Use of Services, New England Journal of Medicine, vol. 310, 1984, pp. 1050-1510.

PAGE 221

217 explained by the difference in ages between HMO enrollees and other patients. 66 Not surprisingly, HMOs initially faced great skepticism from both the public and physicians, but this resistance appears to be fading in the face of a growing consensus about the need for cost containment. 67 Hospitals. -Spending for hospital care has increased rapidly, because (1) hospital use was encouraged (sometimes required) by most public and private insurance programs, and (2) hospital construction was encouraged by the Hill-Burton Act. There is evidence that these factors led to overuse of hospitals. For example, a study which reviewed 1,132 records of patients admitted to hospitals concluded that only 60 percent of the admissions were appropriate; 23 percent were judged to be inappropriate in that the examiners could determine no positive benefit from treatment in the hospital setting, and 17 percent of these could have been avoided with ambulatory surgery. 68 An independent study found that 19 percent of hospital admissions and 27 percent of hospital days were inappropriate. 69 A study of laboratory tests indicated that 47 percent of the tests performed in a teaching hospital could have been eliminated without any apparent loss in the quality of patient care. 70 While such conclusions are controversial, cost reduction programs have put unrelenting pressure on hospital systems to reduce expenses, by increasing efficiency andto an extent that is difficult to determineby eliminating implicit cross-subsidies that had the effect of supporting teaching, research, and care for the indigent by overcharging patients able to pay. MR.J, Arnold, L.W. Debrock, and J.W. Pollard, Do HMOS Produce Specific Services More Efficiently? Inquiry, No. 21, 1984, pp. 243-253. GTC. HaVinghUrSt, op. cit., foOtnOte 50, p. 30. A.L. Siu et al., Inappropriate Use of Hospitals in a Randomized Trial of Health Insurance Plans, New England Journal of Medicine, vol. 315, No. 20, Nov. 13, 1986, pp. 1259-1266. This study examined records from Seattle, Dayton, Fichburg and Franklin Counties in Massachusetts, and Charleston and Georgetown Counties in South Carolina. Upper income (> $61,000 family income) families and families on Medicare were excluded. 69 J.D. Restuccia et a]., The Appropriateness of Hospital Use, ~ealfh Affairs, vol. 3, No. 2, summer 1984, pp. 130-138. 70A.R. Maflin et al., A Trial of Two Strategies to Modify the Test Ordering Behavior of Medical Residents, New England Journal of Medicine, vol. 303, 1980, pp. 1330-1336. Recent trends in hospital care include: l l l The numbers of full-time in-patients, the rate of hospital admissions, and admission rates for the elderly all fell sharply in 1985. Average hospital stays were 7.7 days in 1975 and 7.1 days in 1985; 71 hospital occupancy rates fell from 80 percent in 1970 to 69 percent in 1985. 72 This alone led to great pressures to close inefficient facilities. Industry studies assert that changes in insurance coverage provided by Deere & Co. of Illinois resulted in a 21 percent decline in hospital days per 1,000 beneficiaries over a period of 36 months; a major insurance company reported savings of $523 per patient in 1981 all because of ambulatory surgery and pre-admission testing. 73 Although in-patient costs are still growing, the rate of growth has slowed considerably and the average length of a hospital stay has been reduced. 74 In-patient expenses per capita increased 12.6 percent between 1976 and 1982, but only 4.5 percent per year after 1983. Competitive pressure has forced hospitals into horizontal integration (in the form of hospital chains) and other services, such as clinics, nursing homes, and ambulance companies, into vertical integration. Some have even diversified into real estate and resort management. 75 An increasing fraction of hospitals are being operated as parts of chains rather than as independent community-based facilities. Between 1975 and 1982, multi-hospital systems increased their share of all community hospitals from 25 to 33 percent. 76 Ownership of multi-hospital chains is dominated by five large companies that own 6 percent of all U.S. acutecare beds, and nearly 50 percent of all for-profit TIAmerican Hospital Association, Hospital Statistics: 1986 (Chicago, IL: American Hospital Association, 1986), p. xvii. There were actually 8 percent fewer people in hospitals in 1985 than there were in 1975. 72 lbid., p. 2. 73 R. Kauer, Deer & Company: Utilization Review, Health Systems Management Center, Case Western Reserve University, under contract with the Business Round Table Health Initiative Task Force, 1983; and Employee Benefit Research Institute, op. cit., footnote 54. 74 Medicare Prospective Payment and the American Health Care System: Report to the Congress, February 1986. 75 L. Punch, Contract Management Companies Manage Growth Rate of 13.3 percent, Modern Health Care, vol. 14, August 1984, pp. 45-52. TGD, Ermann, and J. Gabel, Multihospital Systems: issues and Empirical Findings, Health Affairs, vol. 3, No. 1, spring 1984, pp. 50-64.

PAGE 222

218 l l l beds. The number of beds owned by these firms grew from 29,000 in 1976 to 62,000 in 1982. Their holdings are concentrated in the South and West, where hospital construction was most rapid during the past two decades. 77 While scale economies have encouraged horizontal integration, there has been some vertical disintegration as hospitals have elected to contract for specialized services (housekeeping, food services, emergency services, plant operations and maintenance, records keeping, billing, and collection). 78 The average numbers of beds per hospital has also risen. In some cases this has led to economies of scale. The number of hospitals with fewer than 100 beds fell 8 percent between 1975 and 1985, while hospitals with more than 300 beds increased their share from 45.8 percent of all beds to 47.8 percent. 79 An increasing number of hospital patients are now treated privately, and more hospitals are privately owned. Public facilities faced enormous difficulties because of competitive pressures and the increasing difficulty of obtaining funding from philanthropic or public funds. 80 In many cases, costs may have been difficult to contain because of the large number of indigent patients. Between 1975 and 1985, the number of beds in investor-owned facilities increased from 7.7 to 10.3 percent of non-Federal hospitals. 81 Even not-for-profit hospitals are hiring private hospital firms to manage their operations, From 1970 to 1981, hospitals operated under contract management increased from 14 to 497. 82 Cost containment in standard hospitals, coupled with growing consumer interest, has spawned rapid growth in specialty hospitals. More than one-third of all private hospitals now specialize in psychiatric care, alcohol and drug abuse, physical rehabilitation, or the care of women. 83 77 R. B. Siegrist, Jr., Wall Street and the For-Profit Hospital Management Companies, in Bradford H. Gray, cd., op. cit., footnote 57. Tsln ]983, growth rates in contractual services ranged from 18 to 200 percent. See L. Punch, op. cit., footnote 75. 79HOSpjra] s~a(jsrics: 1986, op. cit., footnote 71, P. xvii. Ibid. 81 Ibid., pp. 3-4. 82J.W, Salmon, organizing Medical Care for Profit, in J.B. McKinIay, cd., Issues in the Politica/ Economy of Health Care (New York, NY: Tavistock Publications, 1984). 83fv1. Freudenheim, Specialty Health Care Booms, The New york Times, Nov. 25, 1987, p. D1. Clinics.Free-standing clinics and sophisticated services offered in physician group practices are introducing new competition for hospitals. There were 1,800 to 2,000 such centers in 1983, 84 approximately 23 percent of which were controlled by three companics. 85 Technology now allows many procedures to be conducted out of hospitals and many insurers encourage the use of the less expensive clinics, where for some procedures rates average 55 percent less than hospital in-patient rates and 18 percent less than hospital out-patient rates. 86 Many of these centers are owned by individual physicians or partnerships, but in some cases hospital chains are diversifying to own smaller facilities themselves. Support Facilities. Expanded options for care have produced many specialized organizations designed to serve specific market niches. For example, there are now 1,200 centers that provide dialysis and many clinical laboratories that conduct medical tests. Because medical science has improved diagnosis faster than treatment there has been a growing demand for hospices, where patients with known but essentially untreatable illnesses can be made comfortable. There were 1,345 hospices in 1984, increasing at a rate of 1 per day. Similarly, the rapid growth in the ranks of the elderly, and the failure of medicine to provide cures for the chronic illnesses of old age, has led to an expansion of nursing homes. This trend has been accentuated by the number of women wishing to remain in the work force, who are therefore unprepared to care for an aging parent. The costs of these facilities, about half of which are paid by Medicaid funds, already represent nearly 10 percent of all health expenditures. This expense has led a number of States to take draconian actions to block further cost increases. The number of nursing home beds grew 3.7 percent annually between 1971 and 1976 but growth slowed to an annual rate of 2.4 percent between 1976 and 1982 in spite of a 4.5 percent annual increase in the number of persons over 85. By 1984, nine States ordered a halt to nursing home construction, and others were limit84 G. Richards, FECs Pose Competition for Hospital EDs, Hospitals, Mar. 16, 1984, pp. 72-82; and National FEC Group Changes Name, oct. 1, 1984, p. 22. 8SH.J. Henderson, surgery Centers Double, Modern Heahh Care, vol. 15, June 7, 1985, pp. 148-150. SW. Ansberry, outpatient Surgety Rises as Firms Push tO Reduce Health-Care Costs, The Wall Street Journal, Feb. 20, 1985.

PAGE 223

219 ing approvals to retirement housing and life care. 87 Levels of service vary enormously depending on the policies in individual States. There were 94 nursing home beds per 1,000 elderly in Wisconsin in 1980, and 22 per 1,000 in Florida. Most nursing homes are now in small private chains of 2 to 7 facilities with interlocking ownerships, but concentration is increasing. The five largest investor-owned chains increased their share of beds from 64 to 73 percent between 1982 and 1983. 88 While most facilities do not provide much more than routine convalescent care (two-thirds of all jobs are for orderlies, therapy assistance, food service, laundry, and housekeeping), a number have begun to offer a series of diversified services. Some are specializing in areas like physical and respiratory therapy, home health care, adult day care, residential retirement, and life care. 89 All are seeking to increase service for patients covered by private rather than public insurance. Home Health Care. Home health care has increased dramatically as regulatory programs reduce the frequency and length of hospital stays, transferring health care costs to the home (usually to women, see note 111, p. 391). A growing number of home health care services are now covered under public and private insurance. Medicare reimbursements for home health care grew 376 percent between 1976 and 1983. 90 In response to growing demand, a variety of new technologies have been developed that permit an increasing range of treatment at home. For example, technology now permits oxygen therapy for emphysema and IV-nutritional therapy to be conducted at home more cheaply than in the hospital. The implantable pump is an example of a highly sophisticated device that, like the implantable heart pacemaker, makes it possible for the patient to carry a 87K.A. f7aCke]mann, NUrSing f-fome crunch tO f+it Hospitals SOOn, Modern Health Care, vol. 14, Nov. 15, 1984, pp. 42-46. E8B.C. v]ad~k, fJn/ovjng care (New York, NY: Basic Books, 1980), and L. Punch, Chains Expand Their Operations, Expecting Prospective Pay Boom, Modern Health Care, vol. 14, May 1984, pp. 131-140. 89 A. Renschler, Testimony Before the Subcommittee on Aging, U.S. Senate, Committee on Labor and Human Resources, pp. 79-86, 1983; and P.G. Hollie, Nursing Homes Seek Affluent, The New York Times, PP. 31, 33, Sept. 15, 1984. XIu.s, ~nate, SpWia] CommittW on Aging, Medicare and the Health Costs of Older Americans: The Extent and Effects of Cost Sharing, Washington DC, April 1984. sophisticated piece of equipment inside the body cavity and to become completely ambulatory. There will soon be a version that can be reprogrammed by a physician, and eventually through sensors in the body itself. About 15,000 implantable pumps are now in place, primarily for delivering chemotherapy for liver cancer, morphine for intractable pain, and insulin for diabetics. The manufacturer estimates that costs, including the initial cost for surgical implantation, would be $17,000 for one year of treatment with the implantable pumpas opposed to $65,000 for one year of conventional therapy. 91 Technology also makes it possible for comparatively small firms to maintain the sophisticated record keeping needed in modern medicine. Comparatively inexpensive systems can maintain financial records, personnel and payroll records, records of billable activities, and the voluminous forms required by the Health Care Finance Agency 92 Less complex equipment is available without prescription at retail stores. There has been a sharp increase in sales of incontinence products, cardiac monitoring equipment, diabetes therapy products, home nutrition products, kits for colon disease, and infections. 93 Monoclinal antibodies have made it possible to offer a wide variety of sophisticated diagnostic kits to the home market and there is a large market for herpes virus testing and home screening for pregnancy, ovulation, venereal disease, and strep infections. If they are properly used, these systems can also reduce health care costs. A number of small agencies have grown up to serve the exploding need for home health care services. What was once a not-for-profit service has become a profitable enterprise. Investor-owned agencies increased share of home health care from 5.0 percent in 1978 to 26.7 percent in 1984. Most firms are small (averaging 45 employees) and depend heavily on Medicare. Services provided range from housekeeping to training family members in the operation of complex home care equipment. gllmplantable Drug Infusion Pump, Issues in Health Care Technology, VOI 5, No. 2, pp. 1-3, 1982; and Johns Hopkins/NASA/IndustryDeveloped Implantable Infusion Pump, Blue Sheet, Johns Hopkins University, Baltimore, MD, Mar. 7, 1984. 92M.A. Beach]ei, what a Computer Can f)o for a Home Health Agency, Caring, vol. Vl, No. 6, June 1987, p. 5. 93B. ~mondWn, *The Market for Medic~ self~re, American Demographics, No. 51, pp. 35-37.

PAGE 224

220 Drug stores and other retail outlets have become more efficient in their ability to serve home health care needs. There has been rapid growth in overthe-counter sales of drugs (9 percent annually since 1972). Computers operated by retail pharmacists can now keep track of patients records, and help to identify possible adverse drug combinations. Retail operations have also become more efficient in inventory control, ordering, billing, and invoicing. Measuring Productivity How well does each component of the complex Health network do its job, and how well does the system operate as a whole? The increasing complexity in the health care system has made the second question more important than at anytime in the past. Answering this question involves addressing several other issues: l l l What is known about the factors contributing to the incidence of disease, the methods available for reducing risk factors, and the way investments in prevention compare with investments in treatment? What is known about the costs and benefits of alternative methods of delivering medical care? Is society allocating its resources to different classes of individuals equitably and in a way most likely to minimize costs? The retrospective system of payment that prevailed in public and private insurance until recently encouraged procedures that exceeded any reasonable estimate of benefit, and in some cases may actually have encouraged practices that entailed more risk than benefit. These procedures continue even under the current payment system, as witnessed by the case of fetal monitoringa technique shown to be of no benefit and some risk in most cases. 94 Disagreements, or lack of information about the relative effectiveness and costs of different courses of medical treatment, lead to wide variations in the way medicine is practiced around the country. 95 For example: 94 R. Neutra et af., Effect of FetaJ Monitoring on Neonatal Death Rates, New England lournti of Medicine, vol. 299, Aug. 17,1978, pp. 324-326. 95 J. E Wennberg, J.P. Bunker, and B. Barnes, The Need for Assessing the Outcome of Common Clinical Practices, Annual Review of Public Health, vol. 1, 1980, pp. 277-295. l l In Maine, the probability that a woman has a hysterectomy is 20 percent in some markets and 70 percent in others. 96 Table 6-4 summarizes a recent study that revealed enormous differences in the rates at which different procedures were performed in hospitals around the country. The differences are extremely large even for costly and dangerous procedures, like coronary-artery bypass surgery where rates differed by a factor of 3.1 in the facilities studied. Unfortunately the authors of the study had no basis on which to estimate whether the high rates were too high or the low rates too low: the available data do not allow us to explain the wide variations we have observed. In addition, we cannot establish the correct use rates from these data. 97 The rules governing the financing of a health care organization may influence the kind of care provided. Patients of physicians who owned X-ray equipment, for example, are more likely to receive an Xray and less likely to see a radiologist than those of physicians who did not. 98 In a recent experiment, a series of case histories involving heart problems was reviewed by cardiologists with no personal interest in the cases. Physicians in independent feefor-service practices were significantly more likely to recommend tests and surgery than those working for prepaid group practice. 99 However, the rules governing the allocation of costs between patient and insurance company do not seem to have a significant effect either on the total cost of health care, or on the priority with which funds are spent. New Technologies and System Integration Emerging information technologies offer great potential for achieving a new kind of integration in the 96 J E Wennberg, v ar i a ti on s : A Proposal for Action, Health Affairs, pp. 7-32. 97 M.R. Chassin et al., Variations in the Use of Medical and Surgical Services by the Medicare Population, New England Journal of Medicine, vol. 314, 1986, pp. 285-290. 98 A.W. Childs and E.D. Hunter, Non-medical Factors Influencing the Use of Diagnostic X-ray by Physicians, Medical Care, vol. 10, No. 4, July/August 1972, pp. 323-335. 99 M.A. Hlatky, E. Botvinick, and B. Brundage, A Controlled Comparison of Cardiac Diagnostic Test Use in a Health Maintenance Organization, paper presented at annual meeting of the Robert Wood Johnson Clinical Scholars, San Antonio, TX, Nov. 11-14, 1981.

PAGE 225

221 Table 6-4.Variations in the Use of Medical and Surgical Services Coefficient Highest/lowest Procedure of variation rate ratio Injection of hemorrhoids . . . . . . . Hip arthroplasty. . . . . . . . . . Destruction of benign skin lesion . . . . . Skin biopsy . . . . . . . . . . Humeral fracture repair . . . . . . . Total knee replacement . . . . . . . Lumbar sympathectomy. . . . . . . . Coronary-artery bypass surgery . . . . . Carotid endarterectomy . . . . . . . Hiatus hernia repair . . . . . . . . Excision of malignant skin lesion . . . . . Coronary angiography . . . . . . . Excision of benign breast lesion . . . . . Craniotomy . . . . . . . . . . Total hip replacement . . . . . . . Arterial grafts of lower extremities . . . . . Cones fracture repair. . . . . . . . Bronchoscopy . . . . . . . . . Appendectomy. . . . . . . . . . Abdominal aortic aneurysm repair . . . . . Mastectomy . . . . . . . . . . Diagnostic upper gastrointestinal endoscopy . . Colectomy . . . . . . . . . . Prostatectomy . . . . . . . . . Lens extraction . . . . . . . . . 0.79 0.69 0.67 0.58 0.51 0.47 0.44 0.41 0.39 0.38 0.37 0.32 0.31 0.31 0.31 0.28 0.25 0,21 0.19 0,19 0.17 0.16 0.15 0.12 0.11 26.0 11.4 8.0 4.8 7.9 6.0 4.0 3.1 4.0 5.9 3.3 2.3 2.2 2.6 3.0 3.5 2.3 2.2 2.2 2.2 2.7 1.6 1.6 1.7 1.5 SOURCE: M.R. Chaasin et. al., Variations in the Use of Medical and Surgical Services by the Medicare Population:" New England Journal of Medicine, vol. 314, 1988, pp. 285-290. Nations health care system that would reduce costs and improve the quality of services rendered. In many cases, new devices can contribute a qualitatively new service; in others, they can so improve the speed and accuracy of procedures that their costs maybe reduced to a point where they are offered routinely. For example, computer aided diagnostic and treatment systems can provide systematic seeond opinions for complex decisions. Such devices are already available for general abdominal pain, lung and thyroid diseases, glaucoma, cancer, and even neurological and psychiatric problems. These systems are unlikely to displace a physicians judgmerit, but at a minimum they can play a role in eliminating simple mistakes. Computers and communication equipment can also play a powerful role in rationalizing the management of different parts of the health care system, as they have in other business enterprises. For hospitals, record keeping is a major expense that has grown rapidly with the introduction of complex cost containment reporting requirements. New information technology can have a dramatic impact on the efficiency with which records are maintained. Automated ordering and invoicing systems can control overhead while ensuring that adequate supplies of a large number of items are in stock without maintaining large inventories. Information technologies can also improve the quality of clinical medicine by coordinating information about individual cases and maintaining patient records in away that makes it easy to review alternative programs of therapy. In this way, physicians can benefit from the collective knowledge of case histories instead of relying on their own experience and that of their immediate colleagues, as is the case today. Hospital investment in information-processing equipment has been spectacular. Information system costs have now grown to 2 percent of the overall hospital budget, second only to investments in the hospital building itself. At present 60 percent of system capacity is used for billing and collections, 22 percent for in-patient care, 13 percent for ordering tests and drugs, and 5 percent for management information. 100 Many more applications are expected soon and equipment purchases are expected to in w HospitalSystems Expenditures Skyrocket Hospitals, April 1, 1984, p.39.

PAGE 226

222 crease by a factor of four between 1984 and 1990. thetics (like artificial joints and limbs); improved reaApplications range from improved software for inord keeping that may combine efficient bookkeeping terpreting data from CAT, MRI (magnetic resonance with an intelligence capable of detecting obvious erimaging), X-ray, and other diagnostic equipment; adrors; and improved telemetry for connecting homes vanced systems for designing and operating prosand ambulances to professional facilities. HOUSING 101 Prospects The U.S. housing construction industry has been called the industry that capitalism forgot. Compared with other manufacturing enterprises, the industry is fragmented and decentralized, with little investment in plant improvement, new equipment, or permanent staffing. It conducts virtually no research on either its products or the techniques by which products are assembled. Innovation has resulted almost entirely from component and equipment suppliers. Productivity in the construction industry as a whole appears to be falling. Few builders have adopted the optimum engineering techniques, innovations in basic materials and structural designs, or performance analyses which have been applied to nearly all major products manufactured by other U.S. industries. A typical home, for example, has about 15,000 parts-approximately the same number as an automobilebut is assembled almost entirely by hand at the construction site. The construction industry may be on the brink of a major change as modern production technology and the threat of imports force basic changes in the industrys products, in its production process, and in the way its goods are marketed: The industry could turn increasingly to factory assembly of products and components. These production facilities could be efficient, flexible, and enjoy the productivity increases that are being obtained in other manufacturing enterprises. Field assembly could be rapid, with entire homes assembled in a day. Increased investment in research could lead to major improvements in housing and other construction products. HouslolMuCh of this discussion is drawn from U.S. Congress, office of Technology Assessment, Construction, sector study, 1987. ing could be made more comfortable and less expensive to operate, using new technologies ranging from advanced windows to efficient appliances. Homes could be sold through outlets that would permit prospective clients to participate in the design of a new house, or to be given a video tour of existing structures. The industry could remain largely unchanged, with productivity declining. Housing prices could rise with respect to other amenities. The Nation could be burdened with poorly constructed products that would become an increasing burden if energy prices rise during the next few decades. Foreign suppliers could provide increasing amounts of the value-added through components (e.g., major appliances and fittings), construction equipment (everything from hand tools to heavy construction equipment for tunneling), and licenses. Americans could find themselves assembling foreign products with foreign tools and adding little value other than site preparation. Housing and other types of construction operate with relatively few direct government regulations, other than the highly fragmented set of building codes typically operated at the local level. Compared with the extensive public support for research in the farming industries, total public spending on housing research is negligible. With 9 million people, Sweden supports more research on home construction than the United States; virtually all U.S. housing research is supported by the U.S. Department of Energy. Structure and Performance A diverse collection of enterprises combines to provide Americans with comfortable homes. Of the funds spent on the Housing amenity, however, onethird of the value-added ends up in the real estate

PAGE 227

223 Figure 6-4.-Value-Added To Meet Demand for Housing ($672 billion* in 1984) Natural Resources Construction Low Wage Manuf Medium Wage Manuf. High Wage Manuf Transport & Trade Real Estate & Rental Other Transactional Personal Services Social Services I 1 0 10 20 30 40 50 Percent of $672 billion m 1972 ~ 1984 l Constant 1980 dollars SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4). industry (see figure 6-4), most in the form of direct returns to capital invested by homeowners or other investors. Since these costs result from factors largely unrelated to the structure of production, they are not examined here. There is little opportunity to change land costs, or the cost of home financing, through changes in production recipes in construction. Price changes in these areas result from a variety of factors discussed at some length in chapter 3. Approximately 16 percent of the spending for Housing results in value-added in the manufacturing firms that produce appliances, furniture, and housing components for housing (as well as value-added generated indirectly by suppliers of these firms; the system for producing these products is discussed more fully in the last section of this chapter). About 10 percent of the value-added in Housing goes for energy or raw materials. Only about 13 percent of all the value-added in the U.S. economy resulting from Housing ends up in the Construction sector. Most of this results directly from home construction, but some arises from the need to construct facilities that provide energy for housing and from other indirect effects. Housing provides by far the largest market for Construction activities, accounting for almost 48 percent of value-added in that sector. For convenience, the industry that meets intermediate demands for construction (dams, harbors, bridges, roads, sewers, hospitals, and office buildings) will also be examined in this section. 102 The network of industries associated with this kind of construction includes architecture, engineering, field erection, component production, maintenance, and repair. Factors Forcing Change There are several reasons for the lack of progress in the U.S. housing construction industry. The most important is undoubtedly the enormous uncertainty ties in demand. Demand for construction business can change by as much as 50 percent in one year. Most firms achieve the flexibility required under these circumstances by minimizing use of permanent staff and long-term capital investments. The ratio of capital to labor in construction, after increasing by 4.2 percent per year between 1950 and 1968, declined by 0.8 percent per year during the early 1970s. 103 The industry is highly fragmented and decentralized. While some vertical integration exists, most projects involve teams assembled from different parts of a network. Most construction firms are relatively small; of the 558,000 U.S. construction firms in 1982, only 10 percent had more than 19 employees. 104 There are few barriers to entry, particularly for homebuilders. Less than 8 percent of homebuilders have annual sales of over $1 million. This fragmentation is encouraged in part by the fragmented form of regulation governing the industry. The homebuilding industry is regulated almost entirely at the State and local level. Currently, local governments apply several thousand major and minor variations of four model codes. There are at least as many inspection systems, accounting for differences in building code interpretations and varying degrees of enforcement. The various fire safety codes 102 As defined by the U.S. Office of Management and Budget (OMB), the construction industry actually consists of many different subindustries: Standard industrial Classification (SIC) code numbers 15, general building contractors and operative builders; 16, heavy construction general contractors; and 17, s+wcial trade contractors. In addition, subdividers and developers, a subdivision of real estate (SIC code 65), is included in statistics on the construction industry. lo3f-f. Kemble Stokes, Jr., An Examination of Productivity Decline in the Construction lndust~, The Review of Economics and Statistics, vol. 63, No. 4, November 1981, pp. 495-502. I04(J.s. small BUSineSS Administration, The State of Sfd Business 1985 (Washington, DC: U.S. Government Printing Office, 1985), Table A1.21.

PAGE 228

224 and inspections systems that relate to buildings compound this regulatory complexity. Enforcement systems also vary, both among and within States. This complex system presents problems for large U.S. homebuilders. The producer must satisfy hundreds if not thousands of building codes and inspection systems in order to serve the national market and still abide by the law. In addition to creating difficulties for large firms attempting to ship products to many different locations, 105 this array of State and local building codes and differing inspection practices can present obstacles to technological innovation. Many of the features that have led to the existing industrial structure are changing. The construction industries in Japan and Sweden have moved rapidly to substitute factory construction for on-site hand labor. In Sweden, over 90 percent of new homes are made in factories. 106 In Japan, homebuilding is moving into the hands of large firms (e.g., Matsushita and Sekisui Chemicals) with a large capital base and a large investment in research engineering. 107 Factory assembly offers a number of potential advantages. The factory permits greater quality control, greater use of modern fabrication equipment, and better working conditions. A factory-assembled home can be erected quickly on a site, minimizing construction loans and disruptions due to poor weather and offers the potential for greater flexibility of design. All of these advantages are moot, of course, if fluctuating demand leaves the equipment idle. In the United States there has been a partial movement toward factory construction. Most U.S. factory construction, however, is essentially hand-assembly indoors. Very little capital equipment is used. In addition to the problem faced by fluctuating demand, factory-based homebuilding firms face an important marketing barrier in the United States. While factory construction is associated with brand name 105 U S Congress, Office of Technology Assessment, Technology, Trade, and the U.S. Residential Construction Industry-Special Report, OTA-TET-315 (Washington, DC: U.S. Government Printing Office, September 1986), p. 70. IWL. Schipper, A. Meyers, and H. Kelly, Coming in From the Cold: Energy Ef7icient Housing in Swden (Cabin John, MD: Seven Locks Press, 1986). lo7Jarnes MCKe]]er, ]nclUStriafized Housing: The Japanese Experience, Alberta Department of Housing, Alberta, Canada, December 1985, p. 81. reliability and quality in places like Japan and Sweden, it is associated with cheap pre-fab and mobile homes in the United States. Intermediate Demand for Construction The rapid growth of service-related industries and the comparative decline of employment and investment in manufacturing and natural resource activity are both reflected in shifting patterns of intermediate demand for construction. Buildings taken as a whole have retained a surprisingly constant share of all investment in structures during the past 35 years, capturing approximately 70 percent of all pub lic and private investment in structures. The rapid fluctuations in demand for new residential construction are compensated partly by activity associated with renovation and remodeling, and partly by the fact that changing demand for non-residential building construction does not respond as rapidly to the changes in interest rates and other factors that lead to collapsing markets for new homes. There have, however, been significant shifts in the composition of non-residential building construction. Rapid growth in service employment has meant that between 1976 and 1985, commercial building space and government offices grew from about 56 to 70 percent of all investment in buildings. During the same period, hospitals and educational facilities fell from 21 to 12 percent of building investment. Investment in industrial buildings fell from a peak of nearly 25 percent of all building activity in 1979 to 15 percent in 1985. There has been a striking decline in the share of construction work slated for highways, sewers, water projects, and other infrastructure investments (see figure 6-5). This decline has continued in spite of a significant rise in investment in highway construction between 1982 and 1986 (outlined in the Transportation discussion of this chapter). Declining investments in new electric generating facilities, new pipelines, and other energy-related projects counted as structures led to a sharp decline in energyrelated construction work in the 1980s. Network Components Heavy Construction. Most major civil engineering projects, power plants, and large office complexes are constructed by relatively large builder/

PAGE 229

225 Figure 6-5.-lnvestments in Infrastructure (percent of all structures in constant 1982 dollars) Percent of all investment in structures i / 10 8 6 1955 1960 1965 1970 1975 1980 1985 Highways, Sewers, Public Utilities Water Supply SOURCE: U.S. Department of Commerce, Bureau of Economic Analysis, National Income and Product Accounts, historical diskettes, table 5.S. developers or general contractors. The top firms generate more than $1 billion in annual revenue and account for 40 percent of domestic contracts. However, many of these large firms have faced serious financial difficulties in recent years, due to the decline of large domestic projects; receipts for the largest companies fell by 30 percent between 1981 and 1983. Although some firms have offset losses through foreign contracts, competition has grown for major international projects. Foreign firms have captured markets that were once the domain of U.S. giants. The problems of the U.S. nuclear industry, for example, have eroded the position of such major American companies like Bechtel in international markets. At the same time, Japanese, French, and West German firmsmany of which began with American licensescapture increasing shares. Many foreign firms offer advantages that the U.S. firms cannot match such as discounts on materials such as steel and concrete as part of a government-sponsored package that includes design and engineering. Home Builders.-The homebuilding industry includes many small firms, but larger companies have begun to dominate housing output. In 1986, for example, homebuilders with annual volume greater than 100 units produced 67 percent of domestic unitsup from 7 percent in 1959. 108 Moreover, the 108 National Association of Home Builders, Housing F OCUS news release, Washington, DC, November 1987. share of major firms has grown over the past two decades, reflecting trends in the industry as a whole. Most homebuilding enterprises in the United States continue to operate on a small scale. A National Association of Home Builders (NAHB) survey conducted in 1979 showed that 46.7 percent of NAHBs member firms were run by either the principal or the sole owner, while only 8.1 percent had a salaried executive. The average U.S. builder produces 5 to 10 units per year. Most homebuilders serve specific local markets; less than 15 percent of all homebuilders do business outside of their home State. 109 These firms move rapidly into other lines of work during periods of low housing demand. In the bad years of 1967, 1974, and 1975, nearly 20 percent of all NAHB members switched to other businesses. Even the comparatively large firms that build homes or home components in factorieswhich average 320 homes per year 110 are not large by the standards of most manufacturing facilities. Subcontracting to specialized firms accounts for a growing share of construction activity. In 1976, 58 percent of U.S. homebuilders contracted over threequarters of their construction activities, as opposed to one-third in 1969. Firms reporting that subcontracts accounted for less than one-quarter of their work fell from 19 to 10 percent during the same period. Over 90 percent of all industry receipts were paid to subcontractors in 1977. 111 Market concentration has increased in both conventional and factory-based construction. The top 100 production builders now control about 25 percent of the market, up from 17.3 percent in 1978. 112 The top 25 manufactured (mobile) home producers controlled 74 percent of the total market in 1983. Concentration in the metal buildings industry is particularly high, with the top 30 companies controlling over 95 percent of the market. NWF.T. Ventre, on the Blackness of Kettles: Inter-Industry Comparisons in Rates of Technological Innovation, Policy Sciences, vol. 11, 1980, pp. 309-328. llORenee MathieU, me Prefabricated Housing Industries in the United States, Sweden and Japan, Construction Review, vol. 33, July-August, 1987, p. 20. 111 U.S. Department of Commerce, Bureau of the Census, 1977 Census of Construction ]ndustries, Washington, DC, April 1980. I IZAutomatjon jn Housing and Manufactured Home Dealer, various issues.

PAGE 230

226 Remodelers.The remodeling and reconstruction industry has even greater diversity and flexibility than producers of new structures. Traditionally, many firms have worked in both areas, partly to protect themselves against market fluctuations, Newer firms, however, often specialize in renovation, even going so far as to specify types of structures. NAHB finds that the largest and smallest construction firms tend to limit their work to new structures, while many intermediate size firms engage in both new construction and remodeling. The renovation industry is becoming an increasingly distinct business specialty. Over 90 percent of all NAHB members that do remodeling have ceased all other operations. The use of specialized subcontractors has also increased, particularly when renovation is undertaken to improve energy efficiency or safety; this generally involves electrical and plumbing experts. Renovation places a high premium on worker skills, since unexpected developments often accompany the opening of older structures for repair. Investment in building improvements, additions, alterations, and replacements represent a growing share of construction activity. A survey by the 600 largest U.S. construction firms put reconstruction at approximately 40 percent of total construction activity in 1984 and 1985; that figure was 26 percent in 1979. Disagreement about the potential future growth of the renovation market arises from a poor understanding of its framework. Traditionally, rehabilitation rates have been highest during slack periods in the new construction market, but this pattern may be changing. 113 Differences between the cost of new construction and the cost of extensive renovations have decreased. One cause of high renovation and rehabilitation costs is the inherent uncertainty in entering an older structure. As a result, most renovation contracts are negotiated, and are not subject to competitive bidding. This reduces pressure to reduce costs or improve efficiency. Accelerated depreciation of new structures often favors new construction over renovation. Nevertheless, tax credits for historic preservation and other indirect subsidies for renovation can encourage renovations in certain areas. Renovated commercial properties may rent for less than new structures. Measuring Productivity Inadequate funding for construction research and the inability of industry firms to make permanent investments in people or equipment explain much of the stagnant or falling productivity in the U.S. construction industry. A number of other explanations have been offered as well: deflators may fail to account for changes in construction quality (like the obvious improvement in energy efficiency offered by new homes); repair and maintenance work may be underreported; off-site assembly of components may shift highproductivity activity into manufacturing industries and out of the construction industry, leaving comparatively unproductive site-work in construction; and there has been comparatively rapid growth in young, non-union construction workers during the past decade. 114 The pace at which innovation is accepted by local codes provides a good measure of rates of change in the industry. A study of 14 technologies (including non-metalic electric cable and wood roof trusses) showed that 20 years typically passed between the time when 10 percent of all codes accepted the change and 90 percent accepted the change. 115 Builders, whose financial position makes it difficult for them to tolerate mistakes, are extremely sensitive to possible negative consumer reaction. Unions have also resisted change; their effect has been to slow the introduction of innovations. New Technologies and System Integration New construction technologies and higher energy costs have already had a noticeable effect in changing the nature of the structures themselveshow buildings are manufactured, and how they may be adapted for differing conditions and uses. Often 113 Duane r. McGough, AdditkmS to the Housing SuPPIY by Means other than New Construction, U.S. Department of Housing and Urban Development, Office of Policy Development and Research, Washington, DC, December 1982. 114 T~bnO/op, Trade, and the U.S. Residential Construction industry, op. cit., footnote 105, pp. 31-33. Ilsventre, op. cit., footnote 109, p. 319.

PAGE 231

227 buildings can only be made to operate efficiently when the construction and future operating costs are considered as an integrated design problem. Institutional problems and the lack of research targeted at net building design and performance (particularly for residences) has limited progress. Often the cheapest way to reduce the heating or air-conditioning costs of a building involves good basic design. New computer-based systems can improve the productivity of building design and analysis in a number of ways. They can rapidly convert concepts to drawings, drawings to analysis, and can combine these steps in order to estimate initial and operating costs. Such systems might be used to prepare both working sketches and detailed drawings. For example, a draftsman could call a computer file with routine building components at a moments notice, thus bypassing the tedious aspects of drafting. Price lists could be included, allowing for instantaneous projection of the cost of various design alternatives. Computer assisted design systems could aid the designer by removing many of the barriers between inspiration and execution, and could improve clientdesigner communication, allowing for more design and investment flexibility. New information systems can also influence the relationship between consumers and designers. Advanced systems can allow for a computer-based tour of building interiors and exteriors, serving as the basis for a computer-based structural analysis, a study of lighting, or an assessment of energy consumption. Japanese housing producers sell factory built homes through retail outlets and show rooms, much like automobile dealerships in downtown locations. Some make use of computer assisted displays that permit customers to design their own products. 116 While there has not been extensive use of such technology in the United States, important suppliers of housing products have begun to experiment with new communication technologies. Sears Roebuck &Co. has developed an interactive video system which allows customers to browse through 11,000 types of curtains, blinds, and shutters and to receive decorating tips. 117 11615W 7~ChnOjO~, Trade, and the IJ.,$. /+sj&ntja/ Construction Industry, op. cit., footnote 105, p. 45. I17L. Therrjen, Bifih of a Sa]esman: How Vidm is Rewing Up Retailing, Business Week, No. 3015, Sept. 7, 1987, p. 109. New technologies could reduce the cost of modifying designs while preventing errors in areas unaffected by the change. The risks of trying a new idea would be reduced, because a concept would be transferred to drawings and analyzed without a major investment in time or money. Automated design systems coupled with communication systems could facilitate the performance of geographically dispersed teams, allowing clients, architects, engineers, and construction firms to cooperate during the entire implementation process. Computer based technologies could also improve the system of competitive bidding; they might quicken the initial bid application, reduce the uncertainties associated with bidding, and decrease the burdens associated with analysis. Design flexibility is not limited to commercial structures. Prospective homebuyers can now plot their own floor spaces, or compare the appearance of different interior and exterior wall coverings. 118 If computer assisted design represents the first major revolution in building design, factory production of structures and components is the first in assembly. History is littered with predictions of an end to site construction and the development of an industry that would resemble more conventional manufacturing enterprises, 119 making it easy to be cynical about new claims. However, there have been recent gains in factory-based construction of homes and small commercial structures. Though data sources differ, it appears that as many as one-half of the homes built in the United States today involve at least some factory construction. The other half employ such factory-built components as roof trusses, pre-hung windows and doors, and wet-cores (bathroom and kitchen modules). Assembly techniques for building components have also changed. Various new products are used I16u.s. Congr=, Office of Technolo~ Assessment, Technolo~ and the Future of the U.S. Construction Industry-Proceedings of the Panel on Technicaf Change and the U.S. Building Construction Industry (Washington, DC: American Institute of Architects Press, 1986). 119A Ig30S Cornrnissiorl organized for President Roosevelt made this claim, which led to significant investment in factory-based home construction by a number of steel and aluminum companies; however, these groups never captured more than 2 percent of the market. President Truman appointed a housing expediter, who was to solve the postwar housing shortage through increased factory production. Only a fraction of the goal was achieved. Under the Johnson Administration, George Romney rekindled the dream with Operation Breakthrough, but this also failed to reach its initial goal. See Technology, Trade, and the U.S. Residential Construction Indushy, op. cit., footnote 105, pp. 23-26.

PAGE 232

228 for insulating materials, floor coverings, exterior wall surfaces, glazing, and floors. Technology has challenged conventional notions about how to provide basic structural support, and optimum design engineering has refined conventional designs. Truss systems can vastly reduce the cost of large, unsupported spans. There is even a possibility that active controls can be installed in structures to adjust for the dynamics of wind loadings. And new adhesive materials have been introduced to facilitate a wide range of construction from decorative paneling to exterior sheathing. Roofing has always necessitated much site labor, and most client complaints result from imperfections in roof assembly. New techniques can produce roofing sections with standing-seam rib panels, joined mechanically by semi-automated equipment. Some U.S. firms have developed a self-propelled roof runner to do the mechanical crimping. And over 90 percent of all roof trusses are now built in factories; inexpensive computers can design trusses in all but the smallest facilities. Software serves most common applications, allowing for flexibility in design at no additional cost to the builder. In addition, concrete structural units can be manufactured in factories and trucked to building sites. Several firms now sell low-rise office buildings, which can be erected in a single day, to clients ranging from retail operations and banks to schools, These structures can be disassembled on the completion of their useful life. Structures using metal frames and sheathing offer clear opportunities for mechanization, since conventional metal fabrication tools can serve this purpose. Metal-framed housing has never made significant inroads into U.S. markets, due more to market conservatism than to factors of cost or performance. These structures are designed, engineered, and manufactured under computer control. Computers also monitor inventories and material flows, which is critical to an industry where success depends on the ability to deliver products quickly. Technical change has also affected field erection techniques, particularly for commercial structures. Computer assisted operations have been introduced into equipment for uses ranging from earth-handling to erection, and fully robotic systems will arrive soon. Computer assisted equipment serves to replace people in high-risk situations while improving precision. For example, crane operations account for a significant fraction of all construction accidents. Automated systems can alleviate this factor. Control equipment remembers both critical lift heights and swing envelope restrictions. New technology has not facilitated a large increase in the productivity of renovation or retrofit operations. This work often involves gut renovations, where interior work resembles new construction and new materials and assembly techniques. Renovation costs have been reduced by surface wiring for wall mounting, flat-wiring that can run under rugs, modular furniture, modular wall partitions, lighting systems attached to movable partitions, and new types of scaffolding. There is controversy over whether the result is attractive office space. Diagnostic equipment has been developed to help pinpoint defects in older buildings. Infrared cameras can locate heat leaks, and pressurization devices can indicate sources of unwanted air infiltration. Because facades of certain buildingsespecially in older cities like Chicago and New Yorkhave fallen to the street, renovation must now address issues of public health and safety. For example, X-ray, video, and other diagnostic equipment are employed to spot defects; nuclear scanning, capacitance testing, and other techniques are now used to locate sources of roof cladding defects. It is entirely possible that the forces that reshaped the construction industry abroad will also revolutionize the U.S. construction industry. The construction firm of the future may resemble that of conventional manufacturing industries in terms of method and application of research, product assembly, and worker skill levels. Large firms with a high level of technical sophistication are capturing a larger share of markets in areas that were once the domain of smaller companies, creating an industry with a national rather than regional scope. And technology can alter the seasonal nature of construction employment: flexible factories that serve changing markets can replace a flexible workforce,

PAGE 233

229 TRANSPORTATION 120 Prospects Emerging production networks and American lifestyles depend on a flexibile and responsive transportation system, tailored to individuals and comparatively small shipments. These demands have the potential to reshape the network of businesses that provides transportation for U.S. consumers and businesses: l Personal transportation systems could be rebuilt around personal vehicles, and guideways (highways, and rail lines) could be tailored to provide fast, safe transportation. Pedestrian and mass transportation systems could play small but critical roles in high density areas. Freight movement could exploit the power of new communication systems to ensure fast, reliable delivery of comparatively small shipments. l Congestion and a deteriorating highway infrastructure could limit growth of mobility and limit the growth of dynamic production networks based on just in time integration of large and small producers, in different parts of the country or different parts of a given metropolitan area. Destructive competition could lead to a decline in safety. Public regulation of transportation has sharply declined in the past decade. Government influence over the future structure of the system can be critical. Public decisions about research priorities, construction of new highways and other facilities, and regulation of safety will play a decisive role. Structure and Performance The production recipe for the transportation system involves a complex amalgam of private and public spending. With the exception of rail, most guideways are purchased at public expense (primarily through user fees) while equipment is privately owned. Moreover, a significant amount of transportation is purchased as an intermediate good by bus120 Much of this section is drawn from U.S. Congress, Office of Technology Assessment, Transportation, sector study, Washington, DC, 1987. inesses, rather than being purchased directly by individuals or the government. The purchases involved in transportation include: l l l l l l l equipment automobiles, trucks, railroad rolling stock, and aircraft (the manufacture of vehicles falls almost entirely into the High Wage Manufacturing sector, and is discussed in the final section of this chapter); guideways (the businesses that build these facilities are included with the Construction sector, and are addressed in the Construction analysis of the Housing discussion); control and communication networks-dispatch systems and air traffic control (some of these are discussed in the section on Personal Business and Communication); terminals and transfer facilitiesparking spaces, airports, rail yards, ports, harbors, and warehouses; vehicle maintenance facilities; fuel purchases, insurance, and associated supplies (the production of automobile insurance is also examined as part of Personal Business and Communication); and packaging and containersstandardized bulk containers for rail and truck or shipping as well as product packaging (packaging technology is discussed in the section on Food). The discussion in this chapter will deal primarily with the provision of transportation services. The network of value-added generated by consumer and government purchases directly related to the Transportation amenity is dominated by transportation services (30 percent), purchases of vehicles (High Wage Manufacturing captures 16 percent), and Natural Resource enterprises providing energy for vehicles and materials needed for vehicle and highway construction (see figure 6-6). Factors Forcing Change The Nations recipe for delivering transportation services is comparatively stable, in part because the enormous capital investment in equipment and guideways makes change difficult. Technology has a com-

PAGE 234

230 Figure 6-6.-Value-Added To Meet Demand for Transportation ($264 biiiion in 1984) Natural Resources Construction Low Wage Manuf. Medium Wage Manuf. High Wage Manuf. Transport & Trade Transactions Personal Services Social Services 0 5 10 15 20 25 30 35 Percent of $264 billion m 1972 = 198 4 l Constant 1980 dollars. SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4). paratively small direct effect on transportation recipes. Nonetheless, rapid changes in the nature of demand for personal and business transport, the potential for rapid increases in fuel prices, and growing congestion in critical bottleneck areas (such as busy airports and crowded suburban highways) have created significant pressure for change. Public decisions about highway construction, airports, air-traffic control systems, and the implementation of regulations will necessarily play a key role in the performance of the transportation system emerging over the next two decades. The Nations transportation system depends almost entirely on liquid fuelsprimarily fuels derived from petroleum. There is almost universal agreement that the price of petroleum will rise sharply during the next two decades. Optimists hope that the rise will be gradual enough to allow private markets to make a graceful adjustment to a new fuel source. Others fear that the rise may come rapidly in response to a crisis in the Middle East or elsewhere. The price of petroleum in the United States now reflects the operation of comparatively free world markets. The United States has opted to use regulations rather than fuel prices as an inducement to increase fuel economy in automobiles. These regulations proved impossible to maintain when low fuel prices gave consumers a message that fuel was inexpensive and plentiful while regulations demanded parsimony. Many other developed nations, however, place a $1 to $2 per-gallon tax on petroleum fuel, giving consumers a strong incentive to find substitutes. This was originally done because driving was taxed as a luxury like alcohol and cigarettes; it now reflects the high national cost associated with heavy petroleum dependence that is not included in market-based fuel prices. Changing regulations in the past decade have begun to reshape the Nations transportation system. Air, truck, rail, and bus transportation systems were all once heavily controlled by Federal regulation. Most of these regulations have been either eliminated or substantially modified in the past decade. The full impact of these changes has yet to be felt, but they appear to have increased the overall efficiency of transportation in the sense that the new system offers more choice in quality and cost of service. There is, of course, continuing controversy about whether deregulated airlines are safe, whether drivers for small trucking firms are as safe as those working for larger firms, and whether consolidation and mergers occurring throughout the industry will lead to lower prices and more choice or the converse. The critical issue of licensing truck drivers remains one of State discretion. Concern about safety has led to proposals that more classes of driver licenses be introduced, and that State licensing be either coordinated or replaced by a national framework in cases where driver responsibility is critical. Transportation of hazardous materials is an example. 121 One trend that deserves special attention is greater use of intermodalism, in which different transportation modes are combined to provide a transport (rather than road or rail) service in a way that minimizes costs. Examples are the park-and-ride lots that allow commuters to transfer from private autos to a bus or train, joint airfare-car rental discount packages, and the sale of airport-to-downtown limousine service tickets on the aircraft. Intermediate Demand for Transportation Business Travel.As the number of white collar jobs has increased, business demand for the transportation of people has grown more rapidly than that for freight. While statistics are poor, it appears that 121 U.S. Congress, office of Technology Assessment, Transportation of Hazardous Materials, OTA-SET-304 (Washington, DC: U.S. Government Printing Office, July 1986),

PAGE 235

231 about one-third of all intra-city travel is for business and therefore must be considered as intermediate and not final demand. 122 There are, of course, many ambiguities in this accounting since most people pay for their own trip to work while the employer typically picks up the price of parking (a not inconsiderable expense hidden in the cost of commercial and industrial structures). Since the businesses that require significant personal travel are growing rapidly, personal travel would seem to be a growing part of the recipe for American production. On the other hand, modern communications can substitute electronic for personal contact. Telephones, computer communications, and eventually high-quality video conferencing could replace much business travel. But will this occur? The signals are mixed. The introduction of the telephone put more people in contact with one another and actually increased, rather than decreased, the need to travel. 123 Newer forms of communication may have the same effect. Freight.Changes in economic structure are reshaping demand for freight services in a variety of ways. Perhaps most fundamentally, the economy is becoming less intensive in its use of materials. The consumption of steel, cement, paper, and a variety of chemicals appears to have peaked. A variety of factors are involved: l l l l l more efficient use of materials in manufacturing, a shift toward light-weight materials (e.g., substitution of plastic for steel in automobiles and of light packaging material for metal cans), recycling of existing materials, development of materials with longer useful lives, and a movement toward an economy dependent on high-value information and products-with a high value per unit weight 124 As a result of these changes, freight traffic (measured in tons shipped per dollar of GNP) has fallen IZZU.S. Depaflment of Commerce, Bureau of Economic Analysis, Input-Output accounts, Survey of Current Business, various issues. ~zJHerbert S. Dordick, What Home Work Means to f.-arge @animations, Home Workers and Home Builders, Land Development Studies, vol. 3, 1986; and 1. de Sola Pool (cd.), The Social /rnpact of the Telephone (Cambridge, MA: The MIT Press, 1977). IZ4R. D ~rwn, M.H. R OSS and R.H. Williams, Beyond the Era of Materials, Scientific American, vol. 254, No. 6, June 1986, p. 34. by 40 percent since 1950. At the same time, however, bulk commodities are being shipped further. Manufacturing centers appear to have moved away from sources of raw materials, sources of bulk materials close to population centers have been exhausted, and some regions appear to have specialized in the production of raw materials. More than 20 percent of all domestic U.S. energy resources, for example, are now shipped from Alaska. Rapid growth in exports of bulk farm materials and coal have added to demands for long-distance freight movement. Intercity tonnage increased 60 percent between 1950 and 1983 while intercity ton-miles increased 100 percentthe same rate as GNPimplying an increase in average haul length of over 30 percent. Increasing interest in better inventory control and integration of geographically dispersed production centers has placed a premium on fast, reliable delivery of relatively small shipments. While there may bean upper limit to the tons of material per person that an economy needs to move, there is no apparent limit to the amount of value per pound that can be added by sophisticated production. Increasing the value per unit weight of goods, coupled with production systems that are paying close attention to inventory controls, is requiring higher quality from transportation services. Demand for freight transport has also changed, as manufacturing in the United States is becoming more geographically dispersed and is moving from urban to suburban and rural areas. Changes in the location of population centers, retail outlets, and production centers have increased the market for relatively short-haul collector/distributor freight shipments within urban areas. All these effects translate into a growing demand for qualityspeed, reliability, and securityand for batch rather than volume shipments. In the past decade, this has resulted in rapid expansion of truck and air shipments at the expense of rail shipment and Great Lakes waterways. Trucks appear to be the vehicle of choice for most high-value shipments, Trucking accounted for 37 percent of total intercity tonnage in 1982, up from 26 percent in 1950. Railroads have countered by introducing containers and vehicles that can operate on both rails and highways, which offer greater flexibility and reliability. While only about 4 million tons of material are shipped

PAGE 236

232 by air, air freight grew by a factor of 10 between 1950 and 1982. Measuring Productivity Given the qualitative changes occurring in the nature of demand for transportation, it is increasingly difficult to develop an acceptable measure of performance. Measured in terms of the average price of a ton-mile and a passenger-mile, the Nations transportation system has shown steady improvement. Some of the price reductions may be the result of the changing mix of services demanded, since the price of public transportation and freight services has grown more rapidly than average. Such price indices can be misleading, however, because they span a period when deregulation radically changed the price structure of many transportation services particularly air fares, as survey techniques may not accurately reflect the real impact of discount fare structures that have become the norm rather than the exception. Price per pound is, however, only one element of service competition. If efficiency is measured in terms of average speed, the record is not particularly distinguished. Although highway speed restrictions limit trucking performance, problems of track quality and yard efficiencies have limited the average speed of the U.S. rail system. Increased competition has made the U.S. transportation system more flexible and dynamic. Some of this flexibility has been achieved at the cost of stable, well paid jobs in the industry (see ch. 12). At the same time, outmoded work rules continue to present problems in established industries like railroads. A journey from Chicago to Los Angeles, for example, can require nearly 20 crew changes. Real efficiency improvements will require innovations in labor-management relations as well as new technology. The indications are that the U.S. freight transportation system can adjust to changing economic conditions, patterns of demand, and input costs more rapidly than in the past. Some of this new flexibility has been achieved by improved communication and computation technology, equipment changes, and new management strategies made possible by deregulation. Scale and Scope.-Major changes have occurred in the way that industry firms are organized. The deregulation of air travel following the 1978 Airline Deregulation Act has introduced much confusion as the industry struggles to rebuild itself around a new set of rules. There has been much controversy over the extent to which regulation of safety can be separated from regulation of fares and other features of air travel. The efficiency of hub systems seems to have been revealed by new regulatory freedom. The new public rule scheme has also revealed the subsidies inherent in regulation, resulting in higher fares to certain locations (typically smaller towns) and lower fares on heavily traveled routes. Nonetheless, shadows of former regulations are still visible in the air transport system: l l l l The two airline reservation systems whose development was, in effect, subsidized by regulated fares on all flights are privately owned. Development of a competing system is virtually impossible. Many airports are largely owned by one or two carriers. American Airlines controls 63 percent of the traffic from Dallas-Fort Worth International Airport. Northwest controls 81 percent of the flights from Minneapolis-St Paul and 64 percent of the flights leaving Detroit. U.S. Air controls 84 percent of the Pittsburgh market. 125 Landing and takeoff rights during peak times at congested airports are not available on the open market. They can be bought only by purchasing an entire airline. A series of mergers and acquisitions has significantly increased the concentration of the airline industry. In 1986, 12 major carriers controlled 85.5 percent of air travel while in September 1987, 8 airlines account for 94 percent. 126 All of this suggests that deregulation may result in less competition and not more. Deregulation has had a major impact on the rail industry as well. Many feel that the rule changes helped improve efficiency by concentrating traffic on a smaller number of routes, and by permitting greater economies of scale in functions such as billIZ5A. Sa]pukas, Air Fare Increase Worry Regulators, The Nerw york Tjmes, Sept. 9, 1987, p. D5. 1261bid.

PAGE 237

233 ing, invoicing, and routine management. 127 The Railroad Revitalization and Recovery Act of 1976 and the Staggers Rail Act of 1980 liberalized merger policy and made it easier for railroads to demonstrate unprofitability of a line. The result was a dramatic decrease in the number of miles in the system. A new element of flexibility in pricing was also introduced by deregulation. While average annual rail freight costs outstripped the overall U.S. GNP price index (or deflator) between 1975 and 1982, growth in freight costs since that time has been less than two-thirds that of GNP. 128 Special prices can now be arranged to reduce the need for empty cars in backshipment. For example, Union Pacific has introduced a reduced rate for hauling fertilizer for Cargill north along the Mississippi in grain cars that would otherwise have traveled empty. The rates are designed to be competitive with barges. Such a strategy would have been virtually impossible without regulatory changes. The Motor Carrier Act of 1980 greatly eased entry into trucking, and relaxed regulations governing rates and service offerings. The effect on the industry was dramatic. The number of trucking firms increased from 18,000 in 1980 to nearly 30,000 in 1983. Profit margins and returns on equity have fallen rapidly; indeed, the rapid rise in the number of firms has been partially offset by the nearly 6,500 business failures occurring since 1978. 129 While average profitability for the industry has fallen below traditional levels, profits have improved for some specialized carriers, particularly larger firms. 130 The market for trucking seems to be bifurcating. Economies of scale and scope appear important in the less-than-truckload (LTL) market, and this business is increasingly dominated by comparatively few carriersthe top 10 LTL shippers control 60 percent of all U.S. LTL shipments and 90 percent of all LTL profits. 131 These economies result in part because firms large enough to operate sophisticated, com127 Transportation, op. cit., footnote 120. 128BU.S. Bureau of the Census, Statistical Abstract of the United States: 1987 (107th cd.), Washington, DC, 1987, table 699, p. 1053. 129 Chris Welles, Is Deregulation Working? Business Week, No. 2078, Dec. 22, 1986, p. 52. 130 W. Legg, Fimmia] Analysis of the Motor Carrier Industry, 1985 (Baltimore, MD: Alex Brown, 1986). IslChris We]les, op. cit., footnote 129. puter-based communication and dispatch systems have more flexibility and operate more economically. On the other hand, there appears to be cut-throat competition among small shippers and owner-operators in the truckload market; most of the industry turnover has occurred among truckload shippers. 132 A competitive market may favor larger shippers who have greater market power, higher volume, and more continuity. Some case studies indicate that small firms lost ground in service and rates in Minnesota and Florida. 133 On the other hand, volume discounts are generally not large, and can be given to brokers and shippers agents who can pass the savings through to smaller shippers. The problem of captive shippers also appears to have grown since deregulation. A growing fraction of bulk shippers are tied to a single railroad for at least a part of their route. This presents real dangers of monopoly pricing and reduces incentives for improvements. 134 Geographic Effects. Small cities with air fares implicitly subsidized by more lucrative air routes have faced increasing prices or less frequent air service. Bus service and short-haul air service has filled many of the gaps. While statistics are not available to confirm the point, it appears that deregulation has not led to a decline in the freight transportation services available for most communities; service from small shippers actually appears to have improved. Competition opened a number of options for remote shippers and often reduced prices. On the other hand, over 240 locations had their scheduled bus service reduced by more than 50 percent in the three months following bus deregulation in 1982 (the Bus Regulatory Reform Act). New Technologies and System Integration A number of trends in the Nations transportation system are almost certain to continue for a genera132 I`bid. 133 "Transportation, op. cit., footnote 120. 134 See Mary H. Cooper, Economic Deregulation, Congn%siond Quartedys Editoria/ Research Reports, vol. 2, No. 3, July 24, 1987; Theodore Keeler, Railroads, Freight and Public Policy (Washington DC: The Brookings Institution, 1983).

PAGE 238

234 tion. They include: increasing geographic integration of production within the United States; increasing demand for fast, reliable shipments of relatively small batches; l no significant growth in demand for commodity shipments; and continued increases in energy efficiency and labor productivity in rail and air transport. Personal Travel The key question for the future of business travel is whether new forms of communication will substitute for physical presence. The answer depends more on psychological attitudes than economics, since it is less expensive to use the telephone than to travel. It is difficult to explain why the physical presence of another individual is felt to be critical to certain kinds of discussions. The need to read gestures, expressions, and other nonverbal communications must have deep psychological roots. Activities that require sensitive discussions, such as personnel evaluations, student tutorials, or delicate contract negotiations, appear to require physical presence while contact such as the exchange of routine business information, is acceptably conducted by mail or telephone. There are, on the other hand, situations where individuals prefer to avoid the presence of another individual, prefering the relative anonymity of mail or a message service. 135 The range of alternatives between meetings and mail will continue to increase-the result of new communication technologies that can create a series of niche markets offering different levels of contact. Unfortunately, little is known about the demand for these kinds of services. Such information may develop only after the technology has been in place long enough for society to adjust to its capabilities. It took a generation for individuals to feel comfortable with the telephone. A number of studies of office location have examined the reasons behind decisions to locate offices in city centers. 136 According to some observers, the 135 R. Johansen, J. Vallee, and K. Spangler, Electroric Meetings: Technical Alternatives and Social Choices (Reading, MA: Addison Wesley, 1979). IWS* J. Thomas Black et al. (eds.), The Changing office workplace (Washington, DC: The Urban Land Institute, 1986). need for physical presence in communication is highest for headquarters operations and lowest for back offices that undertake functions like billing and payroll. Aircraft efficiency has been improved, through both increased load factors and the addition of highly efficient wide-bodied jet aircraft. The gains following energy price increases were dramatic. Commercial aircraft achieved 17 passenger miles per gallon in 1973 and approximately 30 in 1986.* 37 The new wide-bodied jets are capable of significant improvements; the Boeing 757 averaged 40 passenger miles per gallon in 1985, and the 767 averaged just over 35 in 1984. 138 Improved air traffic control can also reduce fuel use, by optimizing descent and climbout procedures and by reducing delays. In 1986, the Airline Transport Association estimated that $2 billion in operating costs were lost due to landing and take-off delays resulting from air traffic control problems. These costs do not include the value of the 1 million passenger hours per year lost due to the delays. 139 Video conferences have been on the horizon since the 1960s, but their cost (still about $1,000 per hour for a coast-to-coast communication) has been prohibitive. In addition, the conferences can usually only be held in specialized facilities. Few individuals value the characteristic of talking to a moving image enough to pay much for the privilege. 140 A significant market could appear if the price of video transmission was drastically reduced and the process made more convenient; 84 percent of Fortune 500 companies have either installed a videoconference facility or plan to do so in the near future. There were 575 installed teleconferencing rooms in place in 1981 4,000 may now be in place. 141 Freight The costs and system-wide performance of the transportation system have been improved by a valsTUnpub]ished data provided by the Federal Aviation Administration. IWbid. lsgAirline Transport Association, Discount Fares Save Airline Passengers $3 billion in 1986, press release, Washington, DC, 1987. 140R. Johansen, J. Vd]ee, and K. Spang]er, op. cit., footnote 138; and A. Reid, Comparing Telephone with Face-to-Face Contact, in 1. de Sola Pool (cd.), op. cit., footnote 123, pp. 386-415. 141 For more on r~ent developments in teleconferencing by U.S. businesses, see Fortune, Aug. 5, 1985, p. 63, and May 2, 1983, p. 295.

PAGE 239

235 riety of new technologies, many of which were encouraged by changes in regulation. Railroads have been able to achieve significant cost reductions through the use of integral unit trains for coal, farm products, phosphates, and other chemicals. Performance was further improved using more specialized cars, consolidated yards, mechanized hump yards, and weigh-in-motion systems. The integral trains can move more quickly through yards and loading facilities. The tare, or empty weight of freight cars, has been reduced through greater use of aluminum. A variety of improvements in truckscenter sills, load dumping systems, drawbars, couplers, wheel slip controls, and multiple unit controlshave improved system performance. Automated sensors now inspect for wheel cracks, and car identification is facilitated with bar-coding. Computer-based dispatching, scheduling, and control systems can improve reliability and timeliness. The efficiencies of bulk truck hauling have been improved largely through a steady increase in the size and weight of trucks permitted on the highways. In 1974, Congress raised maximum permissible gross vehicle weight on the interstate highway system from 73,280 pounds to 80,000 pounds, though States were allowed to retain lower, grandfathered weight limits. The Surface Transportation Act of 1982 made Federal standards preempt State limitations on truck sizes and permitted double trailers on all interstates. These changes may have increased labor productivity in freight transport along some corridors, but the trade-offs remain highly controversial. The efficiency of barge traffic has also improved due to economies of scale. There has been a steady growth in the size of barges and Great Lakes freighters, made possible by lock enlargements funded primarily by the Federal government. Specialized barges have been introduced that permit high maneuverability in congested areas like the upper Mississippi River. Pipelines offer great economies for bulk movement. New, larger diameter pipes are now possible because of new technologies. The 36 -diameter Alaska oil pipeline is an example. The performance of pipelines is further improved using better compression systems, pumping systems, and sensors tied to computers that increase effective capacity. A variety of innovations has improved the system for moving standardized containers from ships to specialized flat-cars and trucks. Entire truck trailers are also carried on freight cars. The transfer of containers is being steadily improved with specialized terminals, automated loading systems, and standardization. Another class of innovation helps different branches of transportation offer a joint service-intermodalism. For example, the Railroader and Railmaster systems use vehicles with both rubber and metal wheels that can operate on either rails or highways. The vehicles are about 1,500 pounds heavier than standard truck trailers, but are much lighter than standard rail cars. The small performance sacrifice in relatively short highway hauls is more than offset by the advantages offered when these vehicles are used on rails. Specialized truck dispatching firms have sprung up to provide coordination and status information on material being trucked within the large, fragmented network of intramodel and intermodel carriers. This service would not have been possible without the widespread implementation of computers necessary for controlling, tracking, and Scheduling. 142 Only through such a system can the small truckers compete on a nationwide basis with the large corporations. Information services have, as a result, become a critical part of the production recipe for freight transport. Energy price increases during the 1970s led to a variety of innovations in the transportation system. If the transportation system of 1981 had operated with 1974 energy efficiency, fuel consumption would have been 15.3 percent higher. But energy efficiency is affected as much by the choice of transportation mode and system efficiency as it is by the performance of vehicles. The shift from rail to truck transport, for example, significantly reduced the energy efficiency of the system, since rail transport is three to four times more energy efficient than trucking. A variety of technical improvements can improve the energy efficiency of freight transport. In trucks, use of radial tires, aerodynamic body shapes (including the now familiar flanges on the cab roofs of many 142 Thomas S. Gray, The Effects of Trucking Deregulation on the U.S. Economy: 1980-1985, presented at the Third Creativity, Innovation, and Entrepreneurship Symposium, Framington, MA, May 1986.

PAGE 240

236 intercity trucks), low-weight materials, and diesel engines can improve energy efficiency. Further improvements include the use of both double-wide tires and engine additions taking advantage of exhaust energy. The average efficiency of trucks can probably be improved by at least 30 percent using technology now on the shelf. It offends common sense to believe that a system using a single person to move a truck trailer across the country can compete with a system using two people to move hundreds of trailers via rail, and rail enjoys markets for commodity and container shipments greater than 600 miles. Trucks tend to compete in long-haul markets primarily when high quality (e.g., timely) service is required or when rail service is not otherwise adequate. Some analysts believe that railroads can never cover the real opportunity costs of their real estate and operating costs carrying anything but commodities hauled on relatively high density routes. It may well be that rail lines and rail yards are a poor use of a unique resource: long rights of way extending into the hearts of virtually every major city in the United States. To some extent, of course, the balance between rail and trucking will be a public and not a private decision. Rail lines must be maintained entirely by private funds, while some of the cost of public highways is subsidized by taxpayers rather than user fees. CLOTHING AND PERSONAL CARE 143 The system of businesses that combine to bring clothing to American markets provides a particularly clear example of the way changes in the warp and the weft of business networks can affect the costs and quality of products delivered, as well as the fate of individual industries that form parts of the network. This discussion will cover the network of businesses that convert fiber to finished apparel offered for retail sale. Clothing and related products represents the bulk of the economic activity in this amenity group. Prospects The apparel production system could follow one of several radically different paths during the next two decades: l The industry could be transformed by technology and innovative management strategies. It could develop an ability to respond quickly to emerging tastes and new developments in production. It could rely heavily on skilled labor and programmable equipment instead of large numbers of low-paid workers. It could tie producers and retail operations together in a tightly integrated network. An ability to tailor products ltsMUCh of this di~ussion is drawn from U.S. Congress, Office of Technology Assessment, The U.S. Textile and Apparel Industry: A Revolution in Progress-Special Report, OTA-TET-332 (Washington, DC: U.S. Government Printing Office, April 1987). to specific tastes and sizes without significant increases in price would, in effect, complete a cycle that began when the industrial revolution replaced tailor-made apparel with mass production. The manufacturing part of the network could be gutted by imports, leaving a domestic industry consisting primarily of transportation, whole sale, and retail operations. It is likely that a small fashion industry would survive and that domestic producers of high volume commodities (like bed sheets) would continue, although the equipment to produce the materials might be imported. Markets for industrial fabrics could be tied heavily to the fate of manufacturing elsewhere in the economy. Without enforcement of wage and hour regulations, there could be rapid growth of underground assembly through employment of illegal aliens at subminimum wages. There may be no middle ground between these scenarios. Given the pace of change in global textile and apparel production, the direction that the U.S. industry will take should soon become clear. At present, signs can be read that point in both directions. The makers of textile machinery in the United States have failed to keep pace with the world state-of the-art. Imports are cutting deeply into domestic apparel businesses. A decline in these industries cuts deeply into prospects for growth in textile and fiber production.

PAGE 241

237 On the other hand, there has been active movement in the development of tightly integrated production networks, use of flexible production equipment by textile firms (though much of it is purchased from foreign suppliers), and a renaissance in the development of technology for apparel assembly. With the exception of comparatively heavy trade protection, the rules governing the domestic operation of textile and apparel production are almost entirely those of freely operating markets. Structural changes in this ancient business, however, can be heavily influenced by custom, and tradition. Public support for research has been minimal, but Federal regulations may have inadvertently had a major effect on industry structure through rules restricting worker exposure to fibers that caused brown lung. These rules may have played a critical role in stimulating private investment in new, highly productive textile equipment. 144 Many of these basic rules are in a state of flux. A small Federal investment in new apparel production equipment has proven highly successful. Disparate elements of textile and apparel networks have formed a new series of trade councils to develop standards, communication protocols, and universal product codes. Structure and Performance The production and sale of clothing has been dividied into sharply defined operations for generations. Firms that produce man-made or natural fibers supply material to yarn producers, which in turn sell to weaving or knitting facilities. There has also been rapid growth in production of nonwoven fabrics like felt or medical gowns. Fabric is sold to a finisher and then to a garment manufacturer. There, jobbers oversee the movement of products from one processor to another, supplying a finished product to cutters or retailers and maintaining product supplies for spot markets. The cultures of each of these sectors are different, and almost tribal. Producers of man-made fibers are typically enormous, sophisticated chemical indus144Ruth Ruttenberg, compliance with the OSHA Cotton Dust Rule: The Role of Productivity Improving Technology, contract report prepared for the Office of Technology Assessment, March 1983. tries like E.I. du Pont de Nemours & Co. Textile production typically occurs in the southeastern United States, in comparatively small businesses often owned by old families. Apparel production is generally an elaborate network of small businesses in the northeast, and more recently by networks of firms using inexpensive American and Mexican workers in southern California. Retail merchants are almost a completely independent tribe. Many rely heavily on foreign supply, and their interests may diverge sharply from many domestic manufacturers. While there has been some horizontal and vertical integration, movement has been slow and uneven. Distinct patterns are difficult to identify. In some areas, a period of expansion and mergers was reversed when firms that had diversified returned to narrower specialties. Some concentration has occured in cotton weaving and man-made fiber weavingthe two largest sectors of the textile industry. In the early 1950s the leaders of the textile industry, particularly Burlington Industries and Milliken, moved to vertically integrate the textile product from fiber to finishing; apparel production, however, remains highly competitive and fragmented. Of the more than 200 apparel companies in the United States, less than 1 percent have sales over $100 million per year. While there has been some vertical integration in apparel retailing, most efforts to combine functions have proven unsuccessful. The 1980s saw growth in both large discount and relatively small and expensive outlets. The system that starts with raw materials and ends with apparel offered for sale in retail stores involves virtually every part of the economy. In 1984, only about 17 percent of value-added generated by spending on the Clothing and Personal Care amenity ended up in businesses directly associated with the production of cloth or apparel. Nearly 40 percent appeared as value-added in transportation, wholesale, or retail activity; 13 percent wound up in the hands of transactional businesses (see figure 6-7). New Technologies and System Integration Technologies are improving both individual steps in the process used to convert fiber into retail products, and the efficiency of the overall system.

PAGE 242

23a Figure 6-7.-Value-Added To Meet Demand for Clothing and Personal Care ($171 billion* in 1984) Natural Resources Construction Low Wage Manuf. Medium Wage Manuf. High Wage Manuf. Transport & Trade Transactions Personal Services Social Services o 10 20 30 40 50 Percent of $171 billion m 1972 ~ 198 4 l Constant 1980 dollars. SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4). Technical improvements in individual processes include: l l l l l l the substitution of synthetic for natural fibers, and an increase in the strength and consistent quality of yarns of all types; the introduction of programmable, shuttle-less looms and other devices that improve the efficiency of winding and weaving, and of new equipment for producing non-woven fabric; computer-assisted garment design; numerically directed cutters, including laser cutters, that reduce waste when patterns are cut from cloth and make it possible to produce comparatively small batches of patterns economically; automatic sewing systems that can assemble components (like sleeves or trouser legs) using robotic handlers; price code scanners that enable retailers to improve inventory control and reordering. While each of these innovations can, and have, increased productivity in individual businesses, their most important contribution may appear in the way they can be tied together to build a system for delivering apparel to consumers that is more efficient taken as a whole. Improved information flows and responsive batch production are key to what the industry calls a Quick Response strategy. The key is holding inventories low and avoiding overstocking while still ensuring that retailers stock what customers want to buy. Accomplishing this will require fundamental changes in the way information flows between the different components of the fiber-textile-apparelretail chain, and an associated revolution in the style of production. In many ways, the institutional difficulties that must be confronted in implementing such a system pose a greater barrier than the technical problems. Efficient transportation will also play a key role. Currently, it can take up to 2 months for a product to get from an apparel plant to the sales floor. Initial experiments with Quick Response have shortened this time span considerably, using United Parcel Service for rapid deliveries. 145 Textile suppliers are now able to communicate with large apparel companies with such precision that apparel firms can obtain reliable information about the time and size of delivery, as well as the color and location of fabric within a truck. This permits apparel producers to integrate deliveries closely into their plans, and allows them to avoid costly and lengthy inventories of materials delivered. The Levi Strauss Co. estimates that their new communication system by itself saves as much as 10 cents per square yard of material. 146 The system-wide productivity gains made possible by new communication technologies and flexible production technologies are difficult to measure using conventional methods. There is a certain irony in this, since the consumer obviously measures the performance directly in the price and characteristics of products offered for sale. The dilemma of measurement lies partly in the fact that the productivity of individual businesses may actually appear to decline while the productivity of the system as a whole increases. It is possible to describe the areas where system-wide benefits can be expected even when it is difficult to develop precise measurements: l The enormous inventories carried by textile mills, apparel manufacturers, and retailers can be reduced. On average, it takes roughly 65 weeks for fiber to move from a manufacturing plant to the customers hand. The material is in processing for only 15 out of these 65 weeks; the remaining 50 weeks are spent in inventory. 147 The cost of this inventory alone represents 145 Kurt Salmon Associates, Inc., Crafted With Pride in U.S.A. council, Inc., Quick Response Program Report, June 11, 1986. 146 R.E. Cotton, QRs Bottom Line, Apparel Industry Magazine, JUIY 1986, pp. 26-31. 147Sig Scheier, IIQR to Consumer Demand Vital, Hinerfeld Warns, Daily News Record, Oct. 9, 1985, p. 11.

PAGE 243

239 6.4 percent of retail sales. With good management, it should be possible to reduce inventories by 25 to 50 percent. 148 Proper inventory control can increase sales per square foot, ensuring that the assortment on the selling floor matches proven market demand for styles, colors, and sizes. l It should be possible to reduce incidents of forced markdowns that result from orders for goods that fail to sell as expected. Forced markdowns have grown by 50 percent during the past decade, and the National Mass Retail Institute estimates that total losses may be as high as 15 percent of retail sales. 149 Forecasting failures are due in large part to the long, 65 week cycle (requiring that most initial orders for seasonal products be placed more than a year before the products are sold). With Quick Response it may be possible to reduce initial order times to 2 or 3 months, and reorder cycle times to a few weeks. Accordingly, the need for long range, imprecise forecasting is greatly reduced. l Tightly integrated networks can result in cost savings in situations where business is lost because a customer cannot find apparel in the desired style or size when the store is out of stock. Quick Response systems permit smaller initial orders, allowing stores to reorder more of a product that proves popular. The product can then be in stock at full price during the selling season. Estimating the magnitude of stock out losses is a difficult task, since many consumers who fail to find what they want simply leave a retail store without registering their disappointment. Industry estimates suggest that losses from stockouts are about 8 percent of apparel sales, 150 though field experiments with Quick Response systems suggest that this may be an underestimate. Quick Response reordering systems for sales of basic mens slacks have increased inventory turnover at the astonishing rate of 30 percent, with a comparable increase in gross margins on inventory. 151 A stock count in an 48 Robert M. Frazier, Quick Response, presentation made at DUPAATCH, Sept. 13, 1985. 149 lbid. 150 Ibid. 151 "Quick Response Pilot Program Update, Crafted Wisth Pride, Jan. 1987. l l A experimental installation indicated that while 29 percent of items checked were out of stock before the program began, only 17 percent were out of stock after the Quick Response system was initiated. Retail stores can offer a greater variety of products without a significant increase in inventory through the ability to replenish stocks quickly. Responsive networks can also reduce costs and paperwork associated with such overhead operations as billing, invoicing, and inventory controls. Improved information flows and standardized reporting systems can greatly reduce handling and processing costs, such as quality control audits, hanging and premarking of merchandise, and time spent handling and counting deliveries. Perhaps most importantly, four networks that link different parts of the fiber-to-finished product chain more effectively have been created within the last 2 years: the Fabric and Supplier Linkage Council (FASLINC), the Textile and Apparel Linkage Council (TALC), the Sundries and Apparel Findings Council (SAFLINC), and The Voluntary Interindustry Communications Standards (WCS). Finally, productivity gains can be realized within the apparel production facility through the use of off-the-shelf equipment, and better management practices can facilitate integration with the overall system. Moving away from the progressive bundle systema process driven by repetition of standardized tasks, which may have been cost-effective in an environment where response time and inventory control was not criticalto a modern unit production system can reduce processing times of 4 to 6 weeks to 1 or 2 days. 152 Computer-controlled cutting techniques can reduce material losses by 2 to 3 percent and can take 1 to 2 weeks out of planning, while reducing the number of parts that are cut simultaneously by 30 to 50 percent. 153 conservative estimate of the savings that can be realized from a relatively straightforward implementation of Quick Response technologies indicates that the industry could have saved $12.5 billion in 1984. 154 The cost of apparel could be reduced by one152 R.E. Cotton, op. cit., footnote 146. 153 Frazier, op. cit., footnote 148. 154 Peter N. Butenhoff, U.S. Apparel Competitiveness, paper Presented to OTA by E.I. du Pont de Nemours & Co., June 1986.

PAGE 244

. 240 eighth. These savings are realized by the system acting as a whole, and would not be possible through improvements occurring only in component businesses. Most of the savings result from better matching of production to patterns of consumer purchasing because it would allow re-ordering popular items. Inventory costs could be reduced or merchants could maintain a wider selection of items by simply keeping fewer of each style and size in stock at any given time. All of this, of course, depends on close integration of all parts of the design and planning cycle. It requires flexibility in the production of cloth, flexibility in apparel assembly, and flexibility in the transportation network. It also requires relatively shortbatch runs. Where could the technology lead? There appears to be no technical reason why the planning/production cycle could not be reduced to a few days for a wide range of fabrics and designs. Computer displays capable of combining real images of an individual with a synthetic image of different kinds of clothing are already in experimental use in expensive retail outlets. In principle, these systems could be combined with a file maintained on the individuals measurements, producing clothing tailor-made to individual specifications. Assuming rapid delivery services, the garment could be available in 2 to 3 days. Highly responsive networks are likely to be built largely of comparatively small establishments connected together by a well managed communication and transportation system. It is not at all obvious that these systems would operate more or less efficiently given greater horizontal or vertical integration of management. The standards, agreed protocols, and flexible contracting arrangements may well be preferred to a formal bureaucratic management scheme. A considerable amount of experimentation will be needed before a preferred management strategy emerges. EDUCATION 155 Prospects The Nations educational enterprises appear to be on the edge of a major reform. There may be no area where changing patterns of demand and the challenges of new technology are creating greater pressure for change. Real economic growth depends as never before on the skills of people in a wide range of occupations. The essential elements of flexibility and adaptability depend not only on knowledge of a particular skill, but on the ability to identify opportunities, to work together, and to acquire new skills efficiently. At the same time, new technology makes it possible to think about major changes in the productivity of learning. Technology has created an unprecedented power to bring practical problems into the classroom environment, adapt instruction to individual needs, and integrate teaching and learning into the work environment so that information and instruction is available when and where it is most essential. 155 Much of this discussion is drawn from U.S. Congress, Office of Technology Assessment, Education, sector study, Washington, DC, 1987. Taken together, the need for change and the power of the new technology can reshape the way instruction is delivered, where it is delivered, and when it is delivered in a persons career. As in all of the other cases examined in this analysis, the network of activities that produce education in the United States are unlikely to show real productivity gains unless basic changes are made in the way the system operates as a whole. The choices in education are particularly critical and l l stark: A system could be developed that makes learning more efficient and more fun, and that allows teachers to spend more time as tutors and coaches and less in routine tasks. Education could be tailored to individual styles of learning and adapted to strengths and weaknesses in a persons talents and background. instruction could be available throughout a persons career, empowering individuals to adapt and grow in a dynamic economy. A system could be developed that would lead to greater centralization, mechanical and imper-

PAGE 245

. 241 sonal instruction, and uniformity imposed by a perception of scale economies. National regulations could choke individualism, and the dominance of a small number of product suppliers could limit real choice and flexibility. Rigid control could be maintained over a students course of instruction. Increased use of capital equipment could lead to major changes in the structure of the Nations educational system. The process of education could become separated into more differentiated components, coming to resemble other information-intensive enterprises. The development and maintenance of hardware and software for education and training could become an important specialty, and a variety of other specialties could develop. The size and location of institutions could change if technology allows decentralization without sacrificing quality. Instructional activities could be integrated more formally into all business enterprises. Unlike many other production networks, the future structure of the Nations educational establishment is primarily the responsibility of public decisions. Some way must be found to combine the need for enterprise and diversity with the need to find resources adequate to the major development efforts that will be needed. Structure and Performance The network of operations leading to educational services remains astonishingly unchanged by the forces that have reshaped other parts of the economy engaged in the transfer of information. Personnel costsprimarily for teachersand the cost of buying and operating buildings constitute about three-quarters of all costs related to business activity associated with the Education amenity (see figure 6-8). Intermediate Demand for Education The value of a well-educated and well-trained work force is probably the most important input purchased by most businesses, but this input is virtually impossible to measure. 156 During the 1970s, when com156 Several attempts have been made in this area. Among the first were T.W. Schultz, Investing in Human Capital (New York, NY: The Free Press, 1971); G. Becker, Human Capita/ (Chicago, IL: University of Chicago Press, 1975); and T.W. Schultz, Investing in People: The Economics of Population Quality (Berkeley, CA: University of California Press, 1981). Figure 6-8.-Value-Added To Meet Demand for Education ($180 billion* in 1984) Natural Resources Construction Low Wage Manuf. Medium Wage Manuf, High Wage Manuf. Transport & Trade Transactions Personal Services Social Services O 10 20 30 40 50 60 70 80 Percent of $180 billion = 1972 = 1984 l Constant 1980 dollars. SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4). paratively well educated members of the baby boom generation were entering the work force and replacing people with lower levels of education, there was some speculation that America had overinvested in education. This argument held that education served primarily to help get preference in hiring but did little to actually contribute to performance once on the job. 157 Given the problems of measurement, such disputes are difficult to resolve with precision. Recent data demonstrate strong links between education, income, and an ability to avoid unemployment (see ch. 11). American businesses are making heavy investments in education. The total cost of corporate training is probably in the range of $30 billion to $200 billion (see notes to table 3-26). Precise levels are impossible to gauge, since the vast majority of employee learning comes from watching peers, reading manuals, and other informal routes. 158 Unfortunately, as with investment in research, investment in training is often reduced when it may be needed mostwhen business cycles turn down. 159 157 I. Berg, Education and Jobs: The Great Training Robbery (New York, NY: Prager, 1970); A. Weiss, A Sorting-Cum Learning Model of Education, Journal of Political Economy, June 1983, pp. 420-442; and R.B. Freeman, The Over-Educated American (New York, NY: Academic Press, 1976). 158 one estimate suggests that informal training represents 80 to 90 percent of job related training. See S. Lusterman, Trends in Corporate Education and Training(New York, NY: The Conference Board, 1985), 159M. MeyerSon and R. Zemsky, Trainin#s Policies: Public and pr;vate Reinforcement for the American Economy, (Philadelphia, PA: University of Pennsylvania Press, 1985), cited in F.D. Fisher, Redefining Training: An American Perspective, presented to the Computer Assisted Approaches to Training, Foundations of Industrys Future conference, May 25, 1987, Lugano, Switzerland.

PAGE 246

242 Few firms have any clear view of what their real training costs are. The costs of staff training maybe included as general administration expenses, rather than as training or education. Even where training costs are identified separately they may include only teaching and equipment costs, rather than the costs of an employees time and other factors that may be much greater than training costs measured directly. The life insurance industry was recently astonished to find that it costs an average of $125,000 to place an agent in the fieldthe Massachusetts Mutual Life Insurance Company for example, pays trainees during the full 2 to 3 years it took them to be self-supporting through commissions even though, on average, only one out of four entrants managed to become an agent. 160 The Federal government also invests heavily in training. About 20 percent of the active military is either being taught or engaged in teaching. 161 The U.S. Department of Defense spent $13.4 billion on training in 1984, providing the equivalent of 250,000 full-time students. The Office of Personnel Management estimates that more than 500,000 students per year are trained elsewhere in the Federal government (not a full-time equivalent estimate) at a cost of about $371 million, excluding the wages of the employees being instructed. 162 While the costs of training are difficult to measure, it is even more difficult to measure the losses resulting from inadequate training. There is obviously a high social cost when adults are unemployed or underemployed because of inadequate training. 163 Direct costs are also large. Chrysler Corporation estimates that it saved $1.3 million by training workers about the dangers of hazardous substances in the workplace. 164 Poor training may have been a con160 Life insurance Management Research Association, Hartford, CT, 1986, reported by Jane Curtis, Director of Field Development, Massachusetts Mutual Life Insurance Co., private communication, November 1987. 161 Sar A. Levitan and Karen Alderman, Warriors at Work: The Volunteer Armed Forces (Beverly Hills, CA: Sage, 1977). 162 Training Magazine, October 1984. 163 See U.S. Congress, Office of Technology Assessment, Techno~om and Structural Unemployment: Reemploying Displaced Adults, OTAITE-250 (Washington, DC: U.S. Government Printing Office, February 1986). 164 R. Neff, Videos are Starring in More and More Training programs, Business Week, No. 3015, Sept. 7, 1987, p. 108. tributing factor in the Three Mile Island disaster. 165 Nearly one-quarter of Army and Air Force budgets go for maintenance. These two services probably spend as much to maintain as to purchase electronic equipment. Largely because of poor training, between 9 and 32 percent of all person-hours spent in maintenance was wasted (depending on the type of maintenance problem), and between 4 and 43 percent of all parts removed as defective were not defective. 166 An even greater implied cost of poor training, of course, can result if a companys employees are not able to adapt to new production systems or new management strategies. Defining the kinds of skills needed for this ability poses a tremendous challenge. Without greater effort to define them, however, there is a danger that the Nations educational system will have difficulty aiming at the proper target. Any system forced to consider productivity gains is guided in powerful ways by what can be measured. A basic grasp of skills in language, mathematics, and science is clearly essential; box 6-B provides a hypothetical list of other qualitative skills likely to be critical for future employment opportunities. New Technologies and System Integration A number of new information technologies are being developed with the potential to change the nature of human communication in fundamental ways by vastly reducing the cost of storing, manipulating, and communicating information. It may soon be possible to manipulate images and sounds as easily as the printed word. This means being able to break some of the barriers of abstraction that separate scholarship from the world people see, hear, and understand. Modern information technology also has the power to give participants more active control and greater choice without necessarily adding significantly to costs. These innovations can radically change the performance and structure of the educational system. 165 The Presidents Commission on the Accident at Three Mile Island, The Role of the Managing Utility and its Suppliers, staff report, Washington, DC, October 1979. 166 J. OrIansky and J. String, The Performance of Maintenance Technicians on the Job, IDA paper P-1597, Institute for Defense Analysis, Alexandria, VA, 1981.

PAGE 247

243 Box 6-B.A New Set of Job Skills Skills of Problem Recognition and Definition: recognizing a problem that is not clearly presented defining the problem in a way that permits clear analysis and action tolerating ambiguity Handling Evidence: collecting and evaluating evidence working with insufficient information working with excessive information Analytical Skills: brainstorming hypothesizing counter-arguments using analogies Skills of Implementation: recognizing the limitation of available resources recognizing the feed-back of a proposed solution to the system the ability to recover from mistakes Human Relations: negotiation and conflict resolution colaboration in problem solving Learning Skills: the ability to identify the limits of your own knowledge the ability to ask pertinent questions the ability to penetrate poor documentation the ability to identify sources of information (documents and people) SOURCE: Based largely on Francis D. Fisher, Redefining Training: An American Perspective, presented to the Computer Assisted Approaches to Training, Foundations of Industrys Future conference, May 25, 1987, Lugano, Switzerland. The new generation of technologies, built on relatively inexpensive equipment, are qualitatively different from the film strips, television shows, and other techniques that have been used in limited ways to augment instruction in the past. They represent something fundamentally new. In principle, new information technologies can actively engage students or groups of students, and can shape instructional programs to idiosyncratic styles of learning, individual interests, and individual assets and strengths. The Potential for Change In order to understand the potential for change in a sector where productivity and structure have been static, it is necessary to begin by demonstrating that real change is attainable. The following discussion identifies several areas where progress is clearly possible. Improved Understanding of the Learning Process.-There is much room for improvement in what we know about how learning takes place and about the efficiencies of different learning techniques for different people. This is true both for formal classroom instruction and for skills acquired informally at home or on the job. Informal learning on the job can take a variety of forms. It can mean a worker on a shop floor struggling with the documentation for a new machine tool, a secretary puzzling over new word processing software, or an employee in the back office of an insurance firm learning how to underwrite a new type of policy. Clearly, sailors, physicians, and carpenters approach problems in different ways. Each profession, however, combines formal knowledge and formal reasoning from abstract principles with experience and a reservoir of anecdotal evidence. Failure to understand the learning process obviously means that there are few good guidelines for improving its productivity. Most formal training by corporations is practically identical to instruction delivered in standard classroom settings. But a sensitive understanding of learning pathologies is required to teach adults who enter retraining with an enormous variety of backgrounds. The efficiency of instruction is obviously improved where it is possible to discover what people do not know, and why they fail to understand important topics. The new field of learning pathology is beginning to discover why students fail to understand some topics. Analyzing errors made by elementary school students, John Seeley Brown of Xerox found 150 distinguishable patterns of errors in simple subtraction problems. There is obvious merit in finding a way to focus on a students real problem (perhaps a failure to understand the meaning of the number zero) rather than repeating an entire lesson. Many basic issues in pedagogy would have needed to be addressed even if new demands had not been placed on the Nations educational system. New technologies make the issues impossible to avoid.

PAGE 248

244 Research on advanced computer designs may provide some serendipitous help. There are powerful connections between research on human learning and research on productive designs for information technology. The study of artificial intelligence, for example, addresses basic problems in inference and tries to understand how experts think about complex problems. The results of such efforts are likely to influence strategies in teaching before they lead to significant changes in the design of computers. 167 John Henry Martins Writing to Read program, developed with IBM support, is a basic innovation in pedagogy stimulated in part by an analysis of how best to integrate computers into classrooms. 168 The basic idea is to shrink the distance between the rich tools of spoken language of pre-school children and the limits of written language that are available to students entering kindergarten or first grade. It is done by teaching students to decode spoken words using 44 phonemes, and then asking then to use these codes to write stories that use their entire vocabulary. The computer helps primarily by making it easier for students to form letters and by making the task more fun. The Ability To Teach Individuals, Not Aver= ages.-Earlier discussions have provided examples of the way modern production equipment makes it easier to serve niche rather than mass markets. The American system of education, however, continues to deliver a relatively uniform level of instruction with a rigidly proscribed conception of what is to be learned. New technology can help teachers design programs of instruction suited to individual tastes and talents. Some students learn faster from text, some from lectures, some from repetition, and some from visualizations or examples drawn from areas of personal interest. Technology would not, of course, sub stitute for a good teacherbut it could provide an invaluable tool for helping teachers tailor instruction to the needs and talents of many individual students. This is obviously difficult in a standard classroom setting. It is possible to imagine an expert sys167 W. K. Estes and A. Newell, Research Briefings, 1983, briefing panel on Cognitive Science and Artificial Intelligence, Committee on Science, Engineering and Public Policy, National Academy of Sciences (Washington, DC: National Academy Press, 1983). 168 John Henry Martin, Writing to Read (New York, NY: Warner Books, 1986), Alice Kelly and Sophia Kelly, private communication. tern of the kind available to physicians, designed to diagnose the source of a specific learning problem or to construct a program of instruction most likely to interest a specific student. Technology also has the potential to monitor the progress of individual students in a way that rivals the attention of an individual tutor. It can ensure that the pace of instruction does not outpace a students understanding. Using a series of built-in tests, it can ensure that each student has mastered a subject before being forced onward. It can also permit the student to take curiosity excursions into areas that may not be of general interest. And it can facilitate a detailed diagnosis of the sources of a students misunderstanding and habitual errors. From the students perspective, the power inherent in new technology means more control over the pace of learning and less repetition of things already understood. It means not being embarrassed to acknowledge what is not understood. It means the freedom to have fun by building a program of instruction around a set of interesting problems. It means the security of building new knowledge on a base of information that is familiar. The Ability To Simulate.There is also much to be learned about the role that games play in learning. In traditional societies, where children could see much of the work done by adults, play mimicked life. Skills learned in playing at hunting or at household tasks translated gracefully into the practical world. In contrast, contemporary society places enormous barriers between the world of adults and the world of children. Few children have a clear sense of their parents work life; classrooms seldom provide good tools for understanding the work environment. In principle, simulations can mix work with instruction and instruction with play in entirely new ways. The interest and entertainment of an interactive spy adventure can be used to teach map reading, geometry, or chemistry. After centuries without practical alternatives, people have become accustomed to think about communication in terms of words and books. New visual technologies, however, can provide a means of communication that is in many areas more powerful than language. Leaving aside the question of whether language and grammar are an integral part of human intelligence, it is obvious that words in a geography

PAGE 249

245 text are a poor substitute for a visit to Brazil and that a flight manual is a poor substitute for a turn at the controls of an F-16. The emerging generation of information technology has been used to create flight simulators capable of producing an experience so faithful to reality that even seasoned pilots are able to benefit. This power to imitate reality can have a helpful role in education. Much of modern history is available to us in the form of film or television images and sound recordings. These images and sound recordings are arguably responsible for some of our history. The radio clearly helped Franklin D. Roosevelt gain the presidency and strongly influenced the way he communicated with the public as president. Television helped Ronald Reagan in much the same way. Yet the capacity for managing images (particularly moving images) has never been developed with the same facility as that for sorting and analyzing words on paper. Pictures are a second class resource in any library. Technology now emerging can change this status by making it nearly as easy to retrieve, make, and modify images as words. Simulations can also help where student motivation depends on a belief that knowledge will provide power for solving real problems. Simulations can take a wide variety of forms: l l l l The Defense Department has a long and successful record of using simulations to train pilots, tank drivers, TOW missile crews, maintenance personnel, and a variety of other specialists. 169 Goodyear successfully used simulators to train repair crews in the use of multimeters and oscilloscopes. 170 Simulators have been used to train welders who use light-pens instead of welding rods. This equipment can determine such factors as the beginning and ending points of the weld made and whether the simulated rod is moved at the proper rate. Medical technicians can be trained to identify 169 J. Orlansky, The Cost Effectiveness of Military Training, in Proceedings of the Symposium on the Military Value and CostEffectiveness of Training, NATO headquarters, Defense Research Group on the Defense Applications of Research (DS/A/DR(85)167), Brussels, Belgium, January 1985 (evaluations of the program are reported later in this section). 170 S. Holzberger, Goodyear Tire Compares ITC ACTIV With Traditional Training Methods, The Videodisc Monitor, March 1987, pp. 14-15. l l l tumors on displays of actual Magnetic Resonance Imagery (MRI) or Computer Assisted Tomography (CAT) images, Simulations can reproduce medical emergencies demanding immediate responses and analysis for physicians in emergency rooms. Insurance sales personnel can be confronted with a variety of potential customers using video representations of consumer types. The agents responses are videotaped and played back so that students can judge their own responses and ask for expert advice. This experience is so compelling that many students elect to drop out rather than continue training. There are obvious advantages for both the student and the company if the students incompatibility with the job is identified early. 171 A new data compression technology developed by the David Sarnoff Laboratory allows a student to simulate a walk through the Mayan city of Palenque using data stored on a standard 5 compact disk. The system permits the student to go virtually anywhere at the site, stop, and look around a 360 degree vista. It is possible to take pictures of sites for further study. 172 A variety of chemistry laboratory experiments have been simulated using video representations which could be controlled by computer. 173 These simulations are obviously safer than actual experiments. Communication v. Isolation in the Learning Process. -In the abstract, computer-based instruction often evokes an image of passive, hollow-eyed students staring at flickering screens in great windowless halls, or small cubicles. This would be a disastrous outcome given the desirability of allowing students to learn from each other and work as parts of dynamic groups. While it is obviously possible to use computer-based instruction in nightmarish ways, new instructional technology appears to open opITIJane Cufiis, op. cit., footnote 160. lT2The system was deve]oped with the assistance of the Bank Street College. Mike Tinker, David Sarnoff Laboratory, private communication, December 1987. 173 Oone test found no measurable difference between the test results of students working with the computer system alone and students in real laboratories. See S.G. Smith, L.L Jones, and M.L Waugh, Production and Evaluation of Interactive Videodisc Lessons in Laboratory Instruction, Journal of Cornputer-f?ased Instruction, vol. 13, No. 4, pp. 117-121, autumn 1986.

PAGE 250

246 portunities for mixing isolated study with work as a part of a group. It seems possible to use the equipment to improve communication on many levels: between teachers and students, among students, between schools and the body of knowledge and specialized expertise outside of the school system, between schools and parents, and among the teachers within the system. Giving students an avenue to participate in different kinds of social groups is an important contribution of the school system. 174 The ability to collaborate with individuals having different kinds of skills and different viewpoints, to perform effectively as part of a heterogeneous society in a variety of circumstances, and to debate effectively about a disputed point are all important skills both on and off the job. The goal may be one of having students learning around the computer as opposed to learning with the computer: when several students work together at a computer display two things happen: learning becomes more of a social activity than in the usual school environment and an impressive amount of peer learning takes place. The potential for teaching groups has obvious advantages in corporate and government training programs, but documentation is rare. The U.S. Army has found that one of the most valuable uses of its new M-1 tank simulator was the way it permitted a team of tanks to practice working together in difficult tactical situations. It is critical that members of a tank battalion learn to work together, to trust each other, and to adapt quickly in unexpected situations, since group maneuvers are extraordinarily difficult to undertake in the field. Informal observation suggests that adults can be made more comfortable in reentering a classroom environment if they are paired with another person of similar skills. The partners can help each other, and if baffled are less restrained in asking the instructor a question. They assume that if their confusion is shared, they should not be ashamed. 176 It is fair to ask whether the new strategies proposed for learning around a computer represent an im174 George Leonard, Education and Ecstasy (Berkeley, CA: North Atlantic, 1987). 175Margaret COX, cited in )hCiitiOn, Op. Cit., fOOtnOk 155. 17~i%inia Nelms,IfBM Corp., private communication, December 1987. provement over the existing system. Appearances to the contrary, the standard classroom instructional techniques prove to be highly impersonal. Collaboration is discouraged in virtually all cases-it is often considered cheating. Companies have complained that formal instruction methods result in students who want to solve problems in isolation, unable to benefit from group discussions. 177 One observer has found that percent of all instructional time in the high school involves only two methods: telling, lecturing, and questioning the total group and monitoring some form of seatwork." 178 Similarly, about 70 percent of the students time in elementary school was spent listening to lectures; the figures were 77 percent in junior high and 76 percent in high school. The lectures in this investigation typically presumed student passivity. Barely 5 percent of instructional time was designed to create student anticipation of a need to respond, and when a response was required it was often a simple recall of fact. The overall strategy of interaction was summarized as follows: First, the vehicle for teaching and learning is the total group. Second, the teacher is the strategic, pivotal figure in this group. Third, the norms governing the group derive primarily from what is required to maintain the teachers strategic role. Fourth, the emotional tone is neither harsh and punitive nor warm and joyful; it might be described more accurately as flat. The passivity just described scarcely portrays a situation characterized by a great deal of communication within the schools themselves, either between students and teachers or among peer groups of students or teachers. There is also is little communication with the world beyond the school door, or between schools and parents. And with the exception of major universities, most teaching material does not reflect recent progress in research. Available mechanisms for upgrading teachers skills are seldom adequate; textbooks remain the main source of substantive communication with external research, yet texts for courses taught in primary and 177 W. Glasser,Control Theory (New York, NY: Harper & R OW 1985); and Judith Resnick, Learning in School and Out, The Educational Researcher, vol. 16, No. 9, December 1987. 178 John I. Goodlad, A Place Called School (New York, NY: McGraw Hill, 1984), p. 12. 179 Ibid.

PAGE 251

247 secondary schools are infrequently revised. School texts in Texas, for example, despite being updated every eight years, have become out-dated in important areas. 180 While universities are designed to forge good links between instructors and advanced research, these links are growing weaker. Some faculty members are forced to take on heavy teaching loads in order to keep college costs down, while others are engaging in proprietary research that cannot be shared with students. Improving Productivity The challenge of measuring productivity in education was discussed at some length in chapter 3. The difficulty results from both an inability to measure net investments in education from the many contributing sources, and an inability to develop acceptable measures of educational quality. Problems of measurement are, if anything, becoming more difficult. The flexible skills now needed by the U.S. work force defy easy documentation. Many attempts have been made to link student performance to teaching salaries or per-capita spending. Correlations are difficult to establish when adjustments are made to correct for the income and educational levels of families. 181 Students with strong family support for education have tended to do well in virtually all circumstances; 182 strong correlations have also been found between the performance of a teachers students and subjective evaluations of the teacher made by supervisors. 183 The failure to find correlations between spending and progress obviously does not mean that greater investment cannot lead to greater results. It does suggest that much remains to be learned about the efficiency of different strategies for teaching and learning. Against this background, estimates of improvements in educational productivity must be treated with caution. Quantitative estimates of productivity change resulting from new technology are particu180 See Billy Reagan and Patricia Sturdivant in Education, Op. cit., footnote 155. ME. Hanushek, Throwing Money at Schools, Journai Of PO@ Anafysis Management, fall 1981, pp. 19-41. 182 J.S. Coleman et al., Equali& of Education/ Opportunity (Washington, DC: U.S. Government Printing Office, 1966). 183 D. Armor et al., Analysis of the School Preferred Reading program in Selected Los Angeles Schools (Santa Monica, CA: The Rand Corporation, 1976). larly difficult to document because of the poverty of information on experimental programs. Experiments with documented results have typically used primitive computer-based teaching technologies that were available only for short periods of timenone involved the kinds of fundamental reform outlined above. Many systems entering the market have not been well designed or tested. As a result, it is not clear whether tests of effectiveness are measuring the quality of the software or the inherent power of the new technology. In spite of these limitations, significant improvements in teaching and learning can be documented. 184 An exhaustive survey of Defense Department experience with training is summarized in table 6-5. Information on flight training is well documented because of its long history. The Defense Department has spent over $1.2 billion on flight trainers; the cost of operating a trainer ranges from 2 percent of flight costs for a B-52G to 50 percent for an AH-ls. The amount of savings depends heavily on what is being learned. For some types of training, simulators have proven useless or worse. In other cases, an hour on a trainer was twice as effective as an hour in actual flight. 185 On average, each hour spent on a simulator was equal to 0.48 hours of actual flying experience. The simulator, of course, could provide experiences that would be dangerous or even impossible to rehearse in a real aircraft. Extensive studies are also available for computer assisted training in maintenance. In 12 of 13 examinations, students trained with a simulator achieved the same or better test scores than those trained with actual equipment. Students using simulators, however, required 20 to 50 percent less time to learn their tasks, and 90 percent of the students questioned preferred simulators to conventional training (instructors gave mixed reviews). 186 A key question, of course, is whether the quality of the training received is as good as conventional methods. Are the topics learned retained longer? Are they more easily transferred to practical applications? Such questions are virtually impossible to address 184 See Technology and Structural Unemployment, op. cit., footnote 163; and Richard E. Clark, Reconsidering Research on Learning from Media, Review of Educational Research, winter 1983, vol. 53, No. 4, pp. 445-459. 185 J. Orlansky, op. cit., footnote 169, P. 9. 186 Ibid., p. 28.

PAGE 252

-. .. 248 Table 6=5.Measured Effects of Defense Department Flight Simulators, Computer-Based Instruction, and Maintenance Simulators Flight Maintenance Computer-based Factor simulators simulators instruction Student time savings. . . . . 50 percent 20-50 percent 30 percent Acquisition cost savings. . . . 30-65 percent 20-60 percent NA Operating cost savings . . . . 8 percent 50 percent NA Life-cycle cost savings . . . . 65 percent 40 percent NA Amortization . . . . . . 2 years 4 years NA NA = not applicable. SOURCE: J. Orlansky, "The Cost Effectiveness of Military Train ing, in Proceedings of the Symposium on the Military Value and Cost-Effectiveness of Training, NATO headquarters, Defense Research Group on the Defense Applications of Research, (DS/A/DR(65)167), Brussels, Belgium, January 1965. with existing data. Evaluations conducted in the past decade necessarily combine results of well designed and poorly designed programs (by comparison, imagine an evaluation of automobiles conducted in 1910). The available evidence, however, seems to suggest that computer based instruction is at least as good, if not slightly better, than conventional methods ranked on the basis of quality (most evidence is anecdotal, and is limited by the fact that the state-of-the-art in instructional software and hardware is changing at breathtaking rates): l l l A review of 24 evaluations of educational technology found that new technologies could cut learning time by 25 percent, as compared to conventional instruction; 5 of the 24 were in basic education and the rest in specialized technical subjects. 187 The Adult Basic Literacy Education (ABLE) in Central Piedmont Community College in Charlotte, NC, found that the average student needed only 21 hours to gain one grade level in reading and math, whereas the average time had been 150 hours. A review of 28 studies of newly introduced videodisk training systems used by corporations and the military found that achievement was improved in over 60 percent of the cases reported, user attitudes were improved in 56 percent, and training time was reduced in 80 percent (20 percent of the cases reported mixed results). 188 187 Chen-Lin, C. Kulik, James A. Kulick, and Barbara J. Shwalb, Effectiveness of Computer-Based Adult Education, Center for Research on Learning and Teaching, University of Michigan, presentation to the American Educational Research Association meeting, Chicago, IL, March 1985. 188 J. Bosco, An Analysis of Evaluations of Interactive Video, Educational Technology, May 1986, pp. 7-17. l l IBMs Writing to Read program, described earlier, has consistently shown an ability to increase learning rates of kindergartners and first graders. An advanced system designed to teach literacy skills to adults has consistently shown an ability to increase measured grade levels by 2 to 3 years during a 20-week course. 189 A number of specific applications have proven highly effective: Federal Express has reported a 50 percent increase in the learning rates of agents in training; McDonnell Aircraft has reported a 33 percent increase in efficiency in teaching computer-aided design drafting; and the University of West Florida has reported a 24 percent increase in speed in teaching new health workers. 190 The Massachusetts Mutual Life Insurance Companys system claims to have reduced learning time by as much as 50 percent. 191 There is also clear evidence that students enjoy the experience of working with technology in the course of learning. For example, in a basic education program in Great Neck, NY, 98 percent of the students said that they enjoyed studying with a computer and were particularly impressed by the clear explanations and immediate feedback provided. Three-quarters said that they would like to spend more time on a computer while only 5 percent asked for less. 192 189 Literacy Program is a Revelation for Non-Reading Adults, Technological Horizons in Education, vol. 15, No. 2, September 1987. pp. 81-82. 190 J.W. Nelson, Evaluation Data on Successful Applications of Technology Based Training Systems, U.S. Department of Commerce, Washington, DC, October 1987. 191 V. Nelms, op. cit., footnote 176. 192 Norman D. Kurland & Associates, The Role of Technology in the Education, Training, and Retraining of Adult Workers, contract report prepared for the Office of Technology Assessment, Oct. 5, 1984.

PAGE 253

249 Structural Effects Most enterprises where information technology has the potential to make significant changes in productivity have found that the potential cannot be achieved without basic changes in management. Education is unlikely to be an exception. A real improvement in the productivity of the Nations educational system will require asking basic, blunt questions about what is being learned, how it is being learned, when it is being learned, where it is being learned, and why it is being learned. There is growing sentiment that a fundamental review is in order (see box 6-C). At issue is whether institutional mechanisms needed to conduct research and to test results are adequate. Few educational institutions are organized in a way that allows real attention to productivity improvement. School decisions are often highly fragmented and politicized. School systems typically resist proposals for reform; university instructors are among the most recalcitrant. This has both healthy and unhealthy results. Teachers tend rightly to be skeptical about any new grand scheme for reformparticularly so in reacting to claims that technology can provide good pedagogy. They require convincing proof that schemes proposed for using technology can be an advantage for their students. 193 Skepticism is justified since, as might be expected, some of the computer based instruction packages offered initially were poorly designed. The incentive systems governing educational investment in private industry are fundamentally different from those governing education in traditional school settings. In a corporate setting or the military, where students are paid for their time as well as the teacher, it is obvious that the productivity of a students time is worth more than teacher productivity by a factor roughly equal to the student-teacher ratio. In most public school settings, however, the perceived cost of a students time is zero. Organizations that pay the salaries of both teachers and students appear to be far more likely to investigate innovations in teaching techniques than those for whom the cost of a students time is zero. In fact, privately funded training has changed the structure of education more rapidly than public inBox 6-C.Some Views on Technology and Structural Change in Education The introduction of a substantial amount of information technology into conventional classrooms as they operate today will, in my belief, produce only marginal improvement in educational effectiveness. It will take substantial institutional changes to bring about the improvement that we are seeking; the school environment must be altered to provide motivation, self-paced progress, and integration of out-of-school learning experiences. It will also be necessary for parents to develop new expectations for their childrens education. Lewis Branscomb Harvard University (formerly of IBM) am persuaded that the existing K-12 school bureaucracy is having its last hurrah, and that designing new models for education that serve all Americans is of vital importance. A fundamental shift in the nature of the education system is not only possible, but essential. Theodore Sizer Chairman, Department of Education Brown University If technology cannot be used to bring about a radical restructuring of how we teach, then its effect will be nil, We should not flinch from the fact that we are talking about a revolution in education. Richard M. Cyert, President Carnegie Mellon University SOURCE: U .% Congress, Office of Technology Assessment, Education, sector study, Washington, DC, 1987 stitutions. Interviews with 218 large U.S. companies have found that 60 percent of these firms have made significant changes in training methods and training technology over the past decade. 194 They reported extensive use of television, computers, live videotape playback, and satellite television networks. One company used satellites to tie 67 of its locations together so that training in sales and servicing could take place without extensive travel. The Defense De193 See Education, op. cit., footnote 155. 194 Seymour Lusterman, op. cit., footnote 158, p. v. 52-635 0 88 9

PAGE 254

250 partment has also invested heavily in advanced instructional technology. 195 Measured by the zeal with which they are purchasing computers, public school systems appear to be moving rapidly toward greater use of instructional technology. While barely 10 percent of U.S. elementary schools had computers in 1981, more than 90 percent had them in 1986. Nearly all American high schools had computers in 1986. The schools now average one computer for every 37 students. 196 It is more difficult to determine whether the new information technology is leading to significant changes in the productivity of learning and teaching. Most equipment appears to be used in relatively straightforward ways. Less than one-fifth of all applications are used for purposes other than drill and practice, teaching computer programming, or word processing. 197 Scale and Scope. -Will the increased use of capital equipment make schools and training more like an undifferentiated commodity, or more closely adapted to individual needs? Both directions are technically possible. Properly used, the technology could be used to encourage questioning and to explore alternatives. A system that frees teachers to spend more time with individual students could make teaching and learning less and not more mechanical. The high cost of producing software systems for new technologies could, however, lead to the replacement of individual authors and teaching styles with teaching approaches developed by bureaucracies or large teams of people. 198 There is also a risk that inflexible and unimaginative software could crush the spirit of inquiry and reward unquestioning obedience to the one right way accepted by a machine. 195 See J. Orlansky, op. cit., footnote 169. Defense Department budget documents cited in this report, which appeared in January of 1985, show $39 million spent for education and training, $87 million for human factors, $38 million for manpower and training, and $227 million for simulation and training devices. 196 U.S. Congress, Office of Technology Assessment, Trends and Status of Computers in Schools: Use in Chapter 1 Programs and Use with Limited English Proficient Students, staff paper, Washington, DC, Mar. 13, 1987. 1971985 National survey of ]nstructionai Uses of School computers, Center for the Social Organization of Schools, Johns Hopkins University, cited in Ibid. l~Arthur Wirth, The Violation of People at Work in Schools, working paper, Department of Education, Washington University, St. Louis, MO, 1987. Judgments about whether technology will increase or decrease flexibility and change the desirable scale of educational operations are made difficult because so little is known about the flexibility of the existing system of education. At first blush, the system appears to be highly decentralized and capable of tailoring instruction to individual needs. But closer examination seems to indicate that decentralization has created stunning uniformity. A survey of 1,000 classrooms across the United States found unexpected uniformity in what was being taught, in how it was being taught, and in the texts from which it was taught. 199 From most points of view, there has been relatively little change over time in the nature or content of instruction in elementary and secondary schools, or in the time spent in school. This uniformity is encouraged by the use of textbooks written to satisfy a broad market. If the decentralized system is to be effective in providing instruction tailored to individuals, teachers must have some way to keep track of the progress of individual students, and to identify individual sources of confusion and individual receptivity to different pedagogical strategies. Teachers report that it is virtually impossible to accomplish this for 20 to 40 students at a time in a classroom, or for 100 or more students during a day. The pacing of instruction must be keyed to the teachers sense of the average progress of a classsomething that is inevitably a crude compromise even for the most talented instructor. One observer has argued that most teachers stop when from 30 to 50 percent of a class understand about 80 percent of what they need to know to master a subject. 200 As a result, few students completely understand the subject before they are forced to go forward. Similarly, another study has indicated that an average of 20 percent of the students experienced difficulty understanding the teachers comments or directions at any given time. 201 No comparable study has been conducted in colleges and universities. College and university education in the United States is also highly decentralized, But few institutions give any systematic thought to the productivity of teaching and learning on their own campuses, and the use of technol199 John I. Goodlad, op. cit., footnote 178, p. 112. 200 Benjamin Bloom, cited in John 1. Goodlad, op. cit., footnote 178. 201 John I. Goodlad, op. cit., footnote 178.

PAGE 255

251 ogy is idiosyncratic. Many colleges employ technology not to individualize instruction but to permit students in mass lecture courses to see the instructor. Geography. New technology has the potential to reshape the landscape of education, primarily combining geographic decentralization with an increase in the scale or scope of management. The effect is more likely to be felt by universities offering courses for large numbers of students on one campus, but it could also reduce the size of large high schools or other facilities in urban areas. The key is an ability to deliver specialized courses without having a specialized instructor, library, and experimental apparatus at each location. It is clearly easier for people to continue to receive training and instruction throughout their careers if they do not need to physically move to where instruction is available. Closed broadcast technology is already being used to deliver instruction to people watching in rooms supplied by their employers. Stanford University has a program where students in a corporate facility watch a videotaped lecture in the presence of another employee of the firm with professional credentials in the subject area, who answers questions during and after the viewing. The TAGER system in Dallas/ Fort Worth ties together 17 universities with a closedcircuit network for graduate level engineering business and computer science. Rio Salado Community College in Phoenix, Arizona has no classrooms; the college uses public and commercial television, cable television, audio teleconferencing, slow-scan television, and videotape. 202 The emerging generation of technology could obviously do more. It might even permit major universities to offer courses and degrees throughout the Nation, through what amounted to small franchise operations tied to the central campus by advanced communication systems. 203 While the prestige associated with ivy covered campuses will undoubtedly always play a major role in helping people make important contacts and obtain attractive jobs, the role of colleges in delivering practical training may well change. It is possible that improved teaching will make it possible to give people what amounts to a good junior-level training by the time they reach 18. It is also possible that employers will be increasingly interested in specialized training, with the understanding that new employees will need to be retrained every few years. Under these circumstances, is it the role of a university to give the student an initial specialty with the assumption that subsequent specialties will be taught by employers? Should universities screen out individuals likely to fail in different occupations? Is the universitys central role one of providing students with basic intellectual tools and a capacity for understanding their culture? ZOZR.J. Lewis, Research Questions on the Impact of Computers in the Classroom, The Ontario Institute for Studies in Education, University of Toronto, Toronto, Canada, 1983. Z03F. D. Fisher, Higher Education Circa 2005, Change, January/Feb ruary 1987, pp. 40-45. PERSONAL BUSINESS AND COMMUNICATION 204 Structure and Performance wave systems, satellite links, fiber optics, high-speed The telephone, banking, insurance, legal, accountswitches, the ability to store massive volumes of data in ways that permit easy access, and a variety of ing, and other businesses that deliver communication other new kinds of equipment have the potential to and business services to individuals and businesses fundamentally reshape industries whose primary are at the core of the much discussed information function was routine paper-pushing. They also have economy, and are most obviously affected by the the potential to create entirely new lines of business capabilities of new information technology. Powerful and inexpensive computers, high-speed commuin areas where services need to be tailored to specific applications. Indeed, a rise in consumer, govnication made possible by relatively low-cost microernment, and business demand for these transactional products has been responsible for a significant ZOqMUCh of this discussion is drawn from U.S. Congress, Office of Technology Assessment, Communications and Information, sector study, share of U.S. economic growth during the past 15 Washington, DC, 1987, years.

PAGE 256

252 This technical revolution has, in turn, undermined (some would say overwhelmed) the regulatory apparatus that evolved over decades around older systems. Such pressures, combined with new attitudes toward market freedom, have weakened or eliminated many ancient regulatory constraints on telephone services, insurance, and banking. Intense competition for long-distance telephone service has replaced the heavily regulated AT&T monopoly. Insurance regulations implemented over generations for a wide range of purposes (primarily at the State level) are being reviewed and modified. 205 Taken together, changes in technology, demand, and regulation, and the intense competition engendered by these changes are in the process of revolutionizing some of the stodgiest enterprises in Americaand some of the Nations leading employers. The rules governing the telecommunications and broadcasting industries have been radically reshaped by changes in regulatory policy during the past few years. The Carterfone decision, the Modification of Final Judgment (disbanding the Bell System), the Federal Communication Commissions deregulatory initiatives established in Computer II, Specialized Common Carrier, Computer III, and other proceedings, have facilitated competition, innovation, and entry. Clearly the cast of characters has expanded considerably. In addition to the components of the old Bell system, independent telephone companies, and other common carriers, financial service providers such as credit card companies, banks, brokerage firms, and insurance companies have invested heavily in national and international communication systems. Some of these enterprises have expanded their networks to offer investor and financial information for remote users. Citicorp is a leader in this kind of private networking by financial institutions. It uses two Westar V satellite transponders, is a digital termination system vendor, and plans to offer both information and network services nationally. Railroads, as well as electric and gas utility companies, have long utilized private microwave facilzOSRegulations in insurance cover consumer protection, licensing Of companies, restrictions on the types and quality of company investment, regulation of rates and advertising, limits on grounds for cancellation and non-renewal, provisions for minimum coverage, regulations on underwriting criteria, cost disclosures, unfair trade practices, and transactions between parent companies and affiliates. ities for command and control of their network operations. They own or control extensive rights-of-way, especially railroads with rights-of-way that connect urban centers and have been active in providing rights-of-way, by sale or joint venture for the construction of intercity coaxial or fiber optic transmission systems. GTE Sprint, the third largest interexchange common carrier, developed from Southern Pacifics private microwave network. As the information processing and data transmission industries merge, integration by many manufacturers into at least some transmission markets is taking place. Mitel and American Satellite have formed a joint venture, EMX Telecom, to provide end-to-end telecommunications services. Finally many large firms are assessing the feasibility, for their private networks, of providing their own (Integrated Services Digital Network (ISDN) facilitiesusing T1 carrier circuits obtained from AT&Ts competitors, satellite transponders, or excess capacity of other users internal networks, rather than leasing lines from telephone companies. While growing use of advanced computers and communication equipment has increased the capital invested by the network of businesses that delivers communication and business services such as legal help, financial services, insurance, and real estate, these businesses remain comparatively isolated from the rest of the economy. Of the value-added generated in serving consumer and government needs for the Personal Business and Communication amenity, 70 percent remains in the sector classified as Transactional Activities (see figure 6-9). Information transactions are still dominated by human expenses and not equipment costs. The costs of a telephone call or a letter are irrelevant compared with the time spent preparing the material. The cost of using advanced databases is dominated by the personal costs needed from the producer to prepare the information and from the purchaser who retrieves and interprets it. Prospects Will all of the institutional shuffles just described result in real competition and flexibility in the Nations information industry? The answer depends both on decisions made by regulatory authorities in

PAGE 257

253 Figure 6-9.-Value-Added To Meet Demand for Personal Business and Communication ($161 billion* in 1984) Natural Resources Construction Low Wage Manuf. Medium Wage Manuf. High Wage Manuf. Transport & Trade Transactions Personal Services Social Services o 20 40 60 80 Percent of $161 billion l Constant 1980 dollars SOURCE: Office of Technology Assessment (see table 4-6 of ch. 4) the near future and on how formerly regulated companies survive the transition to competitive markets. Competition will also be heavily influenced by the way explicit and de facto standards are adopted throughout the industry. An open system, such as the architecture of the IBM personal computer and the MS-DOS operating system, clearly led to an explosion of hardware and software developments by firms of all sizes. Regulation Message Service and Information Processing.The creation and maintenance of communication monopolies regulated in the public interest could be justified by arguing that the economies of scale possible under monopoly regulation can lead to prices so much lower than those likely to be available in a competitive environment that they offset any consumer advantage that may exist under competition. But while regulation can avoid a situation where a monopoly exploits its privileged position through excessive prices, it can seldom provide adequate incentives for the adoption of new, cost-saving products or services. The logic of telephone regulation was undermined by technology in two principle ways. First, the clear natural monopoly advantages of economies of scale were undermined by the explosion of competing technologies for the provision of long-distance service. Relatively inexpensive microwave and satellite relay systems can now, for example, be operated by a variety of companies at prices competitive with AT&T. For some markets there is a debate about whether the provision of local area service, now restricted to the Bell Operating Companies, has also lost monopoly advantages. Second, the logic of regulation collapsed because of the growing overlap between the provision of standard telephone service and activities involving the storage and processing of information where regulation was clearly unjustified. It became necessary to make painful distinctions between different types of advanced telephone equipment and a variety of value-added services such as computing and data services, protocol conversion (converting the language of one computer to a form where it can be interpreted by another), electronic mail, database services, electronic publishing, voice store and forward (a sophisticated phone answering machine that can store voice messages and permits intended recipients an opportunity to call into the system at any time and hear sorted messages), and alarm and telemetry services. It proved difficult to find a way to make adequate use of the Bell System Operating Companies (components of the old AT&T system) enormous sophistication without allowing them to take unfair advantage of their regulated access to local telephone markets. At present, the Bell operating companies are allowed to provide some kinds of electronic information services (message storing and electronic advertising), but are still prevented from providing most long-distance services and making telephone equipment. 206 The series of mergers and acquisitions in telecommunications has resulted in part from the decision to dismember the Bell system along market lines, instead of simply separating communication markets into a series of vertically integrated regional firms. It is possible that the latter choice would have resulted in more effective competition in end-to-end services and perhaps better international competition. The point may ultimately be moot, however, as an increasingly unregulated market moves rapidly to create end-to-end competition despite the original regulatory intent, and as restrictions on the Bell Operating Companies are removed. ZWC. Sims, Most flegu]atory Curbs on 7 Baby Bells Kept, The fVew York Times, Sept. 11, 1987, p. D1.

PAGE 258

254 By the end of this decade, the current telecommunications marketplace is likely to consist of select tiers. A central sector will consist of fully integrated end-to-end suppliers who primarily employ their own facilities to offer users a complete package of services. Among the firms most likely to be major players in this sector are AT&T, the seven Bell Regional Holding Companies, and perhaps some of their larger competitors (MCI, GTE, United, and Continental). A second tier would consist of equipment and carriers carriers. For instance, the fiber optic networks will provide strong competition to AT&T in the carriers carrier sector, while Northern Telecom and IBM/Rolm will be similarly successful in the network equipment and large private branch exchange (PBX) area. Industry shakeouts are currently underway in long-distance markets as well as in large switching equipment (especially digital), PBXs, and key sets. Even within such an oligopolistic market structure, there will be many opportunities for firms to apply organizational advantages of large scale production and vertical integration. However, in each of the supply tiers, there may be problems of discrimination, cross-subsidy, and access to technical information. In sub-markets, particularly those of limited population and traffic density, supply will retain many monopolistic attributes. In central core areas, supply bottlenecks in local exchanges for basic service will remain and provide the most attractive area for application of structural separation. In other words, while the new regulatory environment for communications has opened opportunities for competition in areas traditionally controlled as natural monopolies, it is likely that the industry will continue to be dominated by enormous firms that yield only a small fraction of their total business to companies serving relatively small niche markets. In spite of rapid automation of many activities, the costs of human time and talent should continue to dominate the production recipe in communication and data processing. Given that a way is found to restructure management to take advantage of new information technology and overcome cultural barriers to new communication modes, this blunt fact will remain a significant economic constraint on the way new communication technologies are used. For example, the cost of message delivery primarily involves the time invested by the individuals sending and receiving the information. 207 Labor costs associated with management and the time spent reviewing data, for example, represent nearly 90 percent of the cost of database services, with capital costs of communication and computer equipment and the cost of communication services representing the rest (see table 6-6). The cost of preparing a first-class letter in an ordinary business may be on the order of $20 to $30, while the cost of postage is only 25 cents. The cost of an average 10-minute long-distance business telephone call is $16.24, of which $8.97 is for communication services and the rest is the result of employee time. Figure 6-10 illustrates the relationship between price and demand for different kinds of point-to-point communication services. Demand for mail has proven to be extremely inelastic, since mail delivered per person has continued to increase sharply even though real prices have increased. 208 The elasticity of demand for telephone service is negative. Long distance telephone demand appears to increase by 10 percent when prices fall by 10 percent (a price elasticity of-1 .0), while local telephone usage seems to increase by only 2 percent when prices fall by 10 percent (a price elasticity of -0.2). 209 Broadcasting and Print Media.Regulation of broadcasting and print media are discussed in the next section, which describes the network of busi207 1. de Sola Pool et al., Communications FlowsA Census in the United States and Japan, North Holland, Amsterdam, 1984. 208Communications and Information, op. cit., footnote 204. Z@~ster D. Taylor, Te/ecornrnurrjcations Demand: A Survey afrd a Critique (Cambridge, MA: Ballinger Publishing), 1980. Table 6-6.Production Costs for a Database Percent Cost area of costs Sales, marketing, and administration . 45 Content acquisition . . . . . 40-45 Data communications equipment, frontend processors, database storage. . 6 Communication services (local exchange, long-distance value added net or public switched) . . . . . 7-8 SOURCE: Peter W. Huber, The Geodesic Network: 1987 Report on Competition in the Telephone industry, prepared for the U.S. Department of Justice, Antitrust Division (Washington, DC: U.S. Government Printing Office, January 1987), based on Link Resources Corp., Pricing and Marketing On-line Information Services, 1988, pp. 17-18; Dun & Bradstreet; and other sources.

PAGE 259

255 Figure 6-10.-HypotheticaI Relationship Between Price and Quality of Message Services (e=price elasticity) Long Distance Telephone Service e=065 to 1 Local Telephone Usage e=-0 2 \ SOURCE: U.S. Congress, Office of Technology Assessment Communications and Information, sector study, Washington, DC, 1987. ness that provide the Recreation and Leisure amenity for Americans. Information Processing. -Telecommunications and computers have traditionally been considered separate industries, partly because of legal constraints. AT&T was not permitted to sell or produce computers for anything other than use within the Bell system. 210 As telecommunications networks are, in an important sense, computers (or at least a primary example of the application of computers), this ban did not stop AT&T from emerging as a major presence in the industry. The regulation of telecommunications also prevented computer companies from competing in the telecommunications arena. Nevertheless, the growth 210 This decree was issued in an antitrust case brought by the Justice Department to constrain this communications giant from dominating the nascent computer industry. of distributed processing and computer terminals in the 1970s made computer communication dependent on facilities supplied by the regulated public telephone network. Frustration with certain parts of this system led to the development of alternative special networks for data even as the proliferation of computers in the office and factory led to the creation, by computer companies and others, of Local Area Networks. Computer and telecommunications firms began to overlap and compete. There is growing demand for systems that combine voice and data communication. The rules have changed. AT&T and IBM will compete in each others traditional markets because regulators removed restrictions on AT&Ts computer activities after the court-ordered Modification of Final Judgment, 211 and because several decades of regulatory changes have allowed more competition in the provision of communications equipment and services. Intermediate Demand Table 6-7 provides a crude review of the relationship between intermediate and final demand for information services in 1977. While the data are old (1977 was the last year for which detailed statistics are available on intermediate demand in the categories shown), the table provides some indication of the relationship between intermediate and final demand for information. Some communication services have actually decreased as a fraction of all intermediate demand in the U.S. economy during the past few decades, in part because prices have fallen in many areas. Paper and publishing was 5.1 percent of all intermediate inputs in 1963 and 4.1 percent in 1977. Communication services decreased from 1.8 percent to 1.6 percent of intermediate demand during the same period. Network Components Insurance. 2]2 -Three features of the emerging U.S. insurance system deserve attention: zllThiS ~CUr~ed as of Jan. 1, 1984, and technically, modified the 1956 Consent Decree. 212Much of this discussion is drawn from Barbara Baran, T ec h no logical Innovation and Regulation: The Transformation of Regulation in the Insurance Industry, contract report prepared for the Office of Technology Assessment, Washington, DC, January 1985.

PAGE 260

256 Table 6=7.Demand for Information and Data Processing by Major User Group Percentages Intermediate Personal Government Total use demand consumption demand (1977 $ millions) Communication services: Communications (except radio and