PAGE 1
Automation of America's Offices, 1985-2000 December 1985 NTIS order #PB86-185055
PAGE 2
Recommended Citation : U.S. Congress, Office of Technology Assessment, Automation of America Offices (Washington, DC: U.S. Government Printing Office, OTA-CIT-287, December 1985). Library of Congress Catalog Card Number 85-600623 For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, DC 20402
PAGE 3
. Foreword Automation of America Offices, 1985-2000, assesses the consequences of the continuing and rapid introduction of information and telecommunications technologies in offices: the workplace of about 45 million Americans. The use of computers and new communication systems in offices is bringing about fundamental changes in employment patterns, the skills needed for white-collar occupations, and the quality of worklife and the office environment. These changes will affect all industry sectors, since office work is a growing component of every industry as well as all public sector organizations. The study, requested by the Senate Committee on Labor and Human Resources and the House Committee on Education and Labor, will also be of interest to many other congressional committees because it addresses a wide range of subjects of concern to industry, government, and educational institutions, and to employers, employees, and their organizations. OTA wishes to thank the many people and organizations that contributed to this assessment through advisory panels, workshops, interviews, and other means of sharing their information and experience with us. The final responsibility for the study, however, rests with OTA. Director ///
PAGE 4
Automation of Americas Offices Advisory Panel Charles E. Branscomb Vice President, Telecommunications Communication Products Division IBM Corp. Dennis Chamot Associate Director Department for Professional Employees AFL-CIO Robert L. Chartrand Senior Specialist Congressional Research Service Library of Congress Marvin Dainoff Professor of Psychology Miami University Rosalyn L. Feldberg Visiting Research Scholar Henry A. Murray Research Center Radcliffe College Thomas G, Hermann Chairman, Law Office Technology Committee American Bar Association Squire, Sanders, and Dempsey Robert C. Hughes Vice President and Group Manager Business and Office Systems Marketing Digital Equipment Corp. Barbara B. Hutchinson Director, Womens Division The American Federation of Government Employees Henry C. Lucas Chairman Department of Computer Applications and Information Systems Graduate School of Business New York University Lois Martin Processing Services Director FBS Information Services Karen Nussbaum Executive Director Nine-to-Five: National Association of Working Women Robert M. Peabody Assistant Vice President and Director of Office Automation Mutual of Omaha Randy J. Pile, Jr. Department Head AT&T Information Services Robert Ellis Smith Editor and Owner Privacy Journal Vernell K. Munson Sutherland President Knowledge Systems Ralph E. Upton, Jr. Director St. Augustine Technical Center iv
PAGE 5
Automation of Americas Offices: OTA John Andelin, Assistant Assessment Staff Director, OTA Science, Information, and Natural Resources Division Fred W. Weingarten, Program Manager Communication and Information Technologies Program Project Staff Vary T. Coates, Project Director Benjamin C. Amick III, Analyst Marjory S, Blumenthal, Analyst Janet DeMott, Detailee-DHHS M. Karen Gamble, Analyst Mary Ann Madison, Research Analyst Zalman Shaven, Senior Analyst Administrative Staff Liz Emanuel Shirley Gayheart Patricia Keville Renee Lloyd Audrey Newman Contractors and Consultants Eileen Appelbaum Christopher P. Astriab Alan Porter Georgia Tech Research Corp. Larry Hirschorn University of Pennsylvania Joan Greenbaum Institute for Labor Education and Research, Inc. Larry McClure Northwest Regional Educational Laboratory William Neufeld Jon Turner New York University Anne Posthuma University of Sussex, United Kingdom Tora Bikson Rand Corp. Kathleen Christensen Research Foundation of the City University of New York Leslie Schneider Harvard University, Kennedy School of Government Marlene Thorn IMT Associates Alan Westin Educational Fund for Individual Rights
PAGE 6
Office Automation in Federal Agencies Lewis B. Arnold Systems Policy Staff, Office of Information Technology Justice Management Division U.S. Department of Justice K, C. Bacher HQ/USAF/DAX The Pentagon Ross Bainbridge Information Resources U.S. General Accounting Office Roger Bullock Director, Information Resources Federal Maritime Commission Eliot Christian Office of the Associate Administrator for Information Resources Management Veterans Administration Claire Dolan Human Resource Management Analyst Internal Revenue Service PM/HR/HRT Barry L. Freedman Manager, OMB Systems Automated Systems Division, Office of Administration Executive Office of the President Esther Georgatos Office of Data Management and Telecommunications Veterans Administration Michael J. Gilbride Chief of the Office Automation Division (office of the Managing Director Federal Communications Commission Carolyn Hahn U.S. Department of Transportation Terrell Hicks Director of Management Systems Tennessee Valley Authority Charles Hudnall Information Processing Staff, Office of the Assistant Director for Administration National Science Foundation David Johnson Office of Information Resources Management Agency for International Development H. Kasprzak Department of the Army HqDA, DAIM-PSP Tom Kurihara U.S. Department of Transportation Christos Kyriazi Management Information Services U.S. Department of Commerce Alan Kotok Chief of Planning and Development Staff U.S. Information Agency Coyeen Lawton Office of the Assistant Secretary for Administration and Management U.S. Department of Labor Barry Leonard Acting Director, Foreign Affairs Data Processing Center U.S. Department of State Howard E. Lewis Director of Information Systems U.S. Department of Energy Steven Malphrus Federal Reserve Board Hal Niebel Information Systems Office U.S. Department of State Charles B. Newton Office of Information Resources Management Federal Emergency Management Agency Ern Reynolds Special Assistant to the Deputy Undersecretary for Intergovernmental Affairs, Office of the Secretary U.S. Department of Health and Human Services Jack J. Sharkey Director, Office of Data Management and Telecommunications Veterans Administration John Strain Office of the Assistant Secretary, Program and Resources Management U.S. Department of the Treasury Wally Velander Office of the Associate Administrator for Management National Aeronautics and Space Administration Lydelle Wertheimer Human Resources Technology Group Internal Revenue Service vi
PAGE 7
Office Automation Quality of Worklife Workshop Nicholas Ashford Director, Center for Policy Alternatives Massachusetts Institute of Technology Dean Baker Associate Professor, UCLA School of Public Health Center for Health Sciences Tora Bikson Senior Scientist Rand Corp. David Celentano Department of Behavioral Science and Health Education School of Hygiene and Public Health The Johns Hopkins University Michael Delarco Program Manager, Air, Toxics, and Radiation Monitoring Research Agency Ray Donnelly Occupational Safety and Health Administration U.S. Department of Labor Charles E. Grantham Human-Technology Specialist, Local/Office Systems Honeywell, Inc. Judy Gregory Research Associate, Department for Professional Employees AFL-C1O Mary Haan University of California, Berkeley Bonnie Johnson Corporate Strategic Staff Intel Susan Klitzman Division of Environmental Sciences, School of Public Health Columbia University Philip Kraft Center for Survey Research University of Massachusetts Andrea LaCroix Post-Doctoral Fellow, Department of Epidemiology School of Hygiene and Public Health The Johns HQpkins University Charlotte LeGates Director of Communications Computer and Business Equipment Manufacturers Association Mary Murphree Regional Administrator, New York Womens Bureau U.S. Department of Labor Diana Roose Research Director Nine-to-Five Jan Rowland Epidemiologic Consultant Art Rubin Research Psychologist, Center for Building Technology National Bureau of Standards Steven Sauter Section Chief, DHHS PHS CDC NIOSH Applied Psychology and Ergonomics Branch Lawrence Schleifer Stress & Motivation Research Section, DHHS PHS CDC NIOSH Applied Psychology and Ergonomics Branch Tapas Sen Division Manager, Human Resources AT&T Richard P. Shore Bureau of Labor Management Relations and Cooperative Programs U.S. Department of Labor Michael Smith Associate Professor, Department of Industrial Engineering University of Wisconsin Jeanne Stellman Associate Professor, School of Public Health Columbia University Jon Turner Department of Computer Applications and Information Systems Graduate School of Business New York University Hal Vreeland Center for Preventive Research National Institute of Mental Health vii
PAGE 8
Reviewers and Other Contributors Eileen Appelbaum Temple University Walt Baker IBM Corp. Tora Bikson Rand Corp. Robert Bednarzik U.S. Department of Labor Sharon Canter Manpower Temporary Services Kathleen Christensen City University of New York Steve Coil Inc. Magazine Keith Cooley MSA Gerald Davis Harbinger Group, Inc. Jim Day Council of Vocational Educators William J. Dennis, Jr, National Federation of Independent Businesses Steven Deutsch University of Oregon Claire Dolan Internal Revenue Service Colin Drury SUNY, Buffalo Claudia Goldin University of Pennsylvania Joan Greenbaum LaGuardia College Bill Grenawalt Optical Coating Laboratory, Inc. Heidi Hartmann National Research Council Ron Hertzfeld National Council on Compensation Insurance Margaret Hilton Communications Workers of America Timothy L. Hunt W. E. Upjohn Institute for Employment Research Jim Jackson Prime Computer Laura Johnson Social Planning Council of Metro Toronto Judith Karnm Bentley College Kenneth Kraemer University of Southern California, Irvine Charlotte LeGates Computer and Business Equipment Manufacturers Association Dave LeGrande Communications Workers of America Dennis Little Merit Systems Protection Board Andrea Long University of Michigan Donald Marchand University of South Carolina Robert Mason Metrics Research Corp. James McInnerney IBM Corp. Charles McMillion House Committee on Small Business U.S. House of Representatives Jack Mileski Digital Equipment Corp. Mark Mueller AT&T Information Systems Mary Murphree U.S. Department of Labor Keith Nelms Georgia Institute of Technology William Neufeld Consultant Gregory Nicklas Communications Workers of America Norman Nissenoff Thierry J. Noyelle Columbia University Margrethe Olson New York University Olov Ostberg Swedish Telecommunications Administration Bruce Phillips Small Business Administration Bonnie Johnson Intel .,. VIII
PAGE 9
Joanne Pratt Joanne Pratt Associates David Roessner Georgia Institute of Technology Carol Romero National Commission for Employment Policy Diana Roose Nine-to-Five: National Association of Working Women Fred Rossini Georgia Institute of Technology Mike Roush National Federation of Independent Businesses Arthur Rubin National Bureau of Standards Peter Sassone Georgia Institute of Technology Steven Sauter Robert A. Taft Laboratories William Scheirer Small Business Administration Perry Schwartz Georgia Institute of Technology Tapas Sen AT&T Phil Shelhaas IBM Corp. Richard P. Shore U.S. Department of Labor OTA Reviewers Michael Smith University of Wisconsin Wanda Smith Hewlett-Packard Roberta Spalter-Roth George Washington University Ronnie Straw Communication Workers of America Sharon Szymanski The Labor Institute Jim Taylor Sociotechnicd Systems Thomas Taylor Mountain Bell Maureen Tierny AT&T Jon Turner New York University Hal Vreeland National Institute of Mental Health Steve Weyl Syntelligence Frank White Human Systems Incorp. Robert Yellowlees American Telesystems Corp. John Alic, Senior Analyst Eugene Frankel, Senior Analyst Gretchen Kolsrud, Program Manager Audrey Buyrn, Program Manager Linda Garcia, Analyst Karl Kronebusch, Analyst Wendell Fletcher, Senior Analyst Julie Gorte, Analyst Linda Roberts, Senior Analyst
PAGE 10
Contents Chapter Page 1. 2. 3. 4. 5. 6. i. 8. 9. 10. 11. 12. The Outlook for Office Automation: 1985-2000 . . . . . . 3 Productivity and Employment . . . . . . . . . 33 Training and Education for Office Automation . . . . . 75 The Changing Nature of Office Work . . . . . . . 95 Office Automation and the Quality of Worklife . . . . . ..125 Confidentiality and Security Issues With Office Automation. ...........171 Home-Based Automated Office Work . . . ...................189 Off-Shore Office Work . . . . . . . . . . .. ...211 The Automation of Federal Government Offices . . . . ,233 Office Automation in State and Local Governments. ..................265 Office Automation in Small Business . . . . . . . .283 Office Automation and Differentially Affected Groups: Women and Minorities. . . . . . . . ...297 Appendix AThe Technology of Office Automation. . . . .307 Appendix BOTA Case Studies ..., . . . . ..., . ....330
PAGE 11
, Chapter 1 The Outlook for Office Automation Technology, 1985-2000
PAGE 12
Contents Introduction . . . . . . . . . . . . . . . Organization of Report . . . . . . . . . . . . . Looking to the Future . . . . . . . . . . . . . What is Office Automation? . . . . . . . . . . . . Understanding the Impacts of Office Automation . . . . . . . . The Technology of Office AutomationPresent and Future . . . . . Distributed Information-Handling and Networking . . . . . . . Proliferating Options . . . . . . . . . 1 . . . The Capture of Data. . . . . . . . . . . . . . Communication Between Organizations . . . . . . . . . How Rapidly Will Office Automation Occur? . . . . . . . . The Possible Consequences of Office Automation, 1985-2000 . . . . . Economic and Employment Effects . . . . . . . . . . Training and Education . . . . . . . . . . . . . Organizations and Jobs . . . . . . . . . . . . . Office Workers and Their Workplace . . . . . . . . . . Data Security and Confidentiality . . . . . . . . . . Home-Based Office Work . . . . . . . . . . . . Off-Shore Performance of Office Work.... . . . . . . . . . Federal Government Office Automation . . . . . . . . . State and Local Government Offices . . . . . . . . . . Small Business and Office Automation . . . . . . . . . Working Women and Minorities . . . . . . . . . . . Policy Issues for Congress . . . . . . . . . . . . Further Discussion: Policy Issues and Questions . . . . . . . . Employment. . . . . . . . . . . . . . . . Conversion to Part-Time, Temporary, and Contractor Status . . . . . Training . . . . . . . . .. .. $ . . . Labor/Management . . . . . . . . . . . . . Health and Safety . . . . . . . . . . . . . Data Security and Confidentiality . . . . . . . . . . Home-Based Clerical Work . . . . . . . . . . . . Off-Shore Office Work By or For U.S. Firms . . . . . . . . Federal Procurement Policy . . . . . . . . . . . . Federal Personnel Policy . . . . . . . . . . . . State and Local Governments . . . . . . . . . . . Small Businesses. . . . . . . . . . . . . . Women and Minorities . . . . . . . . .. ... ... ... ...~. . Figures 3 4 5 6 8 11 11 12 13 13 14 15 15 17 18 19 20 21 21 22 23 23 23 24 24 24 26 26 26 26 27 27 28 28 28 28 28 29 Figure No. Page l-1. Changing Structure of the Work Force . . . . . . . . . 3 l-2. History of Technology Used in the Office . . . . . . . . 9 l-3. User Institution Model of Technological Change . . . . . . . 10
PAGE 13
Chapter The Outlook for Office Automation Technology, 1985-2000 INTRODUCTION America has become an information society. Our economy is driven as surely by the incessant demand for information as it is by the continuing necessity of converting raw materials into finished products. The majority of American workers are now white-collar workers, and about 45 percent of all American workers work in offices. (See figure l-l. ) Office work is rapidly being automated, or computerized. What does this mean for the productivity of office workers and the number of office jobs that will be available in the future? What skills will white-collar workers need? What job ladders will be open to them? Will the quality of their working life be better, or not as good? What new opportunities and new problems can we expect, as a result of sweeping technological change in Americas offices? These were some of the questions that led the Senate Committee on Labor and Human Resources and the House Committee on Education and Labor to ask the Office of Technology Assessment (OTA) to conduct a study of the growing use of microelectronic information and communication technologies in office work. With less than 3 percent of the labor force now employed in agriculture and the proportion of blue-collar workers steadily declining, the automation of white-collar work inevitably raises concerns about the number of jobs that will be available for the still growing labor force in the long-range future. This report deals with those white-collar workers who primarily work in offices, although information technology is also affecting others, for example, department store clerks and supermarket cashiers. 3
PAGE 14
4 Automation of Americas Offices But American industry is now participating in a global economy. Competition for both world markets and domestic markets is a powerful incentive for seeking higher productivity. Microelectronic technology has enormous potential for increasing productivity in whitecollar work, which is a large and growing part of every industry sector. The office is the primary workplace for many industries, such as banking, insurance, and real estate, but the office is also a vital element of every industry from manufacturing to farming. An OTA assessment of factory automation, for example, found that. the salaried or white-collar work force will constitute a larger proportion of manufacturing employment in the future. Increased productivity in office work thus would contribute to productivity and growth in all sectors of the economy. The demand for information will continue to grow. With computerization, the unit cost of collecting, processing, distributing, and using information will decline. More kinds of information will be gathered and used for new purposes, and many new information services and products will be created. Demand for information and increased productivity are two major factors in the employment equation; the consequences of office automation for office U.S. Congress, Office of Technology Assessment, Computerized Manufacturing Automation: Employment, Education, and the Workplace, OTA-C IT-235 (Washington, DC: U.S. Government Printing Office, April 1984). employment will depend in large part on the interactions between them. Just as the successive waves of mechanization of farm and factory work have changed U.S. society and economy, so will the automation of white-collar work. Social and economic transitions in the past have raised policy issues that had to be addressed, and in many cases are still being addressed, by Congress. The new wave of automation will also create both opportunities and problems that demand the attention of Congress. Technological change is also related to questions already on our political agenda, such as comparable worth and pay equity, international trade, and health and safety in the workplace. The effects will eventually be felt by everyoneproducers and consumers; managers, professionals, and clerical workers; large and small organizations; the private sector and the public sector. The role that information and communications technologies will play in offices in the future, and the opportunities and problems they present, thus concern all Americans and all of their representatives in Congress. This report puts before its readers a broad range of likely consequences of office automation, and calls attention to large areas of uncertainty. It points to some public policy issues that are emerging or may arise. Many other questions must and can be resolved only by the informed choices and cooperative efforts of individuals and organizations. ORGANIZATION OF REPORT After a brief look at the context of office ceptual framework or model that guided the automation from the perspective of history, assessment. It then summarizes the findings this chapter highlights some expectations of the assessment, identifying policy issues about the technologies and their development that are likely to concern Congress over the over the next 15 years. It introduces a connext decade.
PAGE 15
Ch. lThe Outlook for Office Automation Technology, 1985-2000 l 5 Many of these issues are just coming to public attention and have not yet been widely discussed; specific proposals for dealing with them have not been put forward. A few of the issues, however, are already before Congress or are apt to be the subject of congressional consideration in the near future. Chapters 2 through 6 discuss the possible effects of office automation in more detail. They deal with potential effects on employment levels; the kind of training and education needed for office work; changes in work content, jobs, occupations, and organizations; the quality of work life, the office environment and labor management relations; and the security and confidentiality of information. Chapters 7 and 8 consider two alternatives to conventional offices, made feasible and economically attractive by office automation. The first of these is home-based work, especially the use of the workers home as the primary or sole site for clerical work. The second is offshore performance of data-entry operations, in which work is sent off to be done in countries with lower paid workers. Chapters 9 and 10 look at office automation in the public sector Federal agencies and State and local governments. Chapter 11 is a brief survey of the limited information currently known about office automation and small businesses, an important sector of the economy that is just beginning to automate its offices. Chapter 12 considers the implications of office automation for two groups that are likely to be particularly strongly affected: working women and minority white-collar workers. Appendix A describes office automation technology as it is now and as it is likely to develop between 1985 and 2000. Appendix B summarizes case studies of the automation of several offices to provide some examples of the changes that occur when offices are automated. In particular these examples illustrate the variety of offices affected and the difficult transition stage that occurs as offices automate their work. The summary discussion in this first chapter is keyed at appropriate points to later chapters, where the reader will find more lengthy discussions. LOOKING TO THE FUTURE During the present transitional stage of office automation, there are many problems that are real but do not require congressional action. Some of the fears that people have in anticipating technological change later prove groundless. Many problems are resolved by ingenuity, trial and error, and negotiation between groups that have competing interests but a shared motivation to benefit from technology. Structural changes in the economy, on the other hand, can create lasting inequities and conflicts. They can also open up new opportunities to resolve old issues and realize new social benefits. OTA chose a 15-year perspective because Congress will be concerned less with ephemeral effects and transitional problems than with long-range structural changes. These structural changes are likely to become clearly visible only after office automation has been widely adopted and organizations learn to use its full capabilities. Information and communication technology is itself rapidly evolving and expanding its capabilities. The range of technological choices that an organization has for accomplishing any given information-related objective is wide. The number of manufacturers and vendors of office automation equipment is large, and the competition between them is strong. This suggests that the technology of office automation will be strongly influenced over the next dec-
PAGE 16
6 l Automation of Americas Offices Photo credit L/brary of Congress Photo credit. Michae/ J SnJIfh The evolution of the office environment is shown in these two photos from 1897 and 1980. Perhaps in the future there will ade or more, by the needs and wishes of the users. The as-yet-undetermined characteristics of future office automation technology will strongly influence the social consequences of the automation of white-collar work. But changes in the U.S. economy and society are not, or need not be, entirely technologically determined. They depend in part on choices and decisions made by individuals and organizations, and they also can be guided by public policy. The conclusions of this assessment are there fore conditioned by unavoidable uncertainties about the choices that users will make, about economic growth, and about future public policies. Major directions in the evolution of the technology can be discerned, but when these technical improvements will occur is more uncertain. It is most prudent to assume that some technical breakthroughs may come sooner than be a truly paperless office. now projected, rather than later, since this has happened repeatedly in the last few years. Also uncertain is the speed with which offices will adopt new technologies. This will be influenced by general economic conditions, but in the last decade office automation has been less sensitive to these factors than many expected. Because it can be accomplished incrementally and with relatively small investments, adoption of office automation may be much wider, more general, and more rapid than has often been the case with new technologies. This makes transitional problems more visible and structural problems more important-e. g., displaced workers will have fewer options for adjustment if their occupations are affected in most regions and most industries in a relatively brief period. It also indicates that decisionmakers should now begin to attentively monitor the changes that are occurring, in order to be prepared to deal with problems that may arise. WHAT IS OFFICE AUTOMATION? Almost any place where information handling is the main activity is called an office, whether it is one person at a desk or a complex hierarchy of executives, professionals, and clerical workers. For the purposes of this report, the office is wherever office work is done, and office work is the processing and use of information for the purpose of track-
PAGE 17
Ch. lThe Outlook for Office Automation Technology, 1985-2000 7 ing, monitoring, recording, directing, and supporting complex human activities. One of the striking consequences of information and communication technologies is that together they make much office work independent of the place where it has usually been performed; that is, they allow it to be done in the home, in airplanes and trains, and in other countries. For the purposes of this report, the term office automation is used broadly to mean the application of microelectronic information technology and communication technology to office work. It includes large mainframe computers, smaller minicomputers, personal computers or microcomputers, stand-alone word processors, and the many diverse comBy extension, people also speak of places where other professional work is done as an officee. g., a dentists office. In this report we attempt no rigorous definition of what is or is not office work, but use a commonsense approach. We have generally excluded from consideration such peripheral or specialized places of white-collar work as the dentists office, the scientific laboratory, and the draftsmans office. munication devices and systems that can link them together. The first offices may have been in the homes of Babylonian merchants or Phoenician traders, or perhaps they were construction project offices in the palaces of Egyptian Pharaohs. Almost certainly something like an office came into existence as soon as records could be kept of the exchange of goods, on clay tablets, chisled stones, papyrus, and quipus. 3 Office work is inseparable from commerce because it is concerned with gathering, keeping, and using information about human activities, and particularly those activities that have to do with the production and exchange of goods and services. If office work began in the homes of merchants and traders, it has nevertheless for most of history been done in central locations close to the production of goods and services. 4 The office plays the same role for an organization that the brain plays in a living organism. It receives information flowing in from all parts of the organization (or organism) and from the external environment, processes that information and sends back responses, instructions, and commands through an extended nervous systemestablished channels of communication. At the dawn of written history, clerks and scribes were the first office workers. Information handling work has always carried with it a degree of respect or status (even when many scribes were slaves) because it requires skills and education that, through most of history, few people had. The tools used in this work were for thousands of years very simple, basically writing implements and something to write on, and some means of storing the records written or received. As the scale of human activities increased, the information about them became more voluminous and more diversified. The tools became slightly more sophisticated (a 3Quipus were knotted cords used by the Peruvian Indians (who did not develop writing) to keep records. One of the effects of new communication and information technology may be to allow office work to again be done in homes; see ch. 7, Home-Based Automated Office Work.
PAGE 18
bookkeeping ledger rather than a stone tablet or roll of papyrus) and office workers became more numerous and more specialized. When clocks were invented, the work became more subject to measurement, pacing, and management control. When electric communications were invented, it became less sensitive to proximity but more sensitive to time. 5 But not until the advent of the telephone, the typewriter, and the adding machine, near the end of the 19th century, was a significant part of the work automated. (See figure 1-2.) The mechanical stage of office automation was followed by the electromechanical stage, with electric typewriters and calculating machines. The present age of computers began a little more than 15 years ago, and this assessment looks forward for another 15 years together, only about one generation in human terms, and less than the working lifetime of a white-collar worker. Through the mechanical and electromechanical eras of office automation, whitecollar work continued to be labor intensive. Capital investment in office work has always been low compared to capitalization in other economic sectors such as manufacturing and agriculture. Note that until the 1840s, and the invention of the telegraph, it might take days or weeks to communicate with another office in a different part of the country. Until the transatlantic cable of 1866, it took at least 4 weeks for a merchant to send a message to his field agent in Europe and get a response, and it might take months to exchange messages with the ship carrying his goods, since he could not know when and where it would make port. About 85 percent of office operating costs are labor costs. Now capitalization is occurring rapidly, in the form of information and communication technologies that are transforming the nature of white-collar work. Many other forces have affected office work in recent decades. The scale, geographical scope, and concentration of economic enterprise, and therefore of offices, has increased. Some sectors of industry and commerce have expanded and some have shrunk in importance. The growing pervasiveness of science and technology as components of the economy has both increased the demand for data and changed the nature of the information handled in offices. There has been a strong tendency to professionalize and credentialize many occupations. The office work force has also changed. The average educational attainment of office workers has increased, yet the educational gap between office workers and the general population has narrowed or disappeared. Women, once a very small part of the office work force, now make up the larger part of it. Changes in values, lifestyles-and some claim, the work ethic-have affected office workers along with all others in society. These trends are all important. But technology or the tools people use, have a primary affect on their work, how it is done, and how it is rewarded by society. UNDERSTANDING THE IMPACTS OF OFFICE AUTOMATION In carrying out this assessment, OTA used of effects: substitution, adaptation, and transa simple conceptual model as an aid in lookformation. ing for the possible effects. 8 (See figure 1-3.) It suggests that when organizations adopt new The new technology usually replaces an older technology, there are likely to be three kinds technology or human labor, or both. There are direct substitution effects, both at the task This model was developed by J.F. Coates and V.T. Coates; level and at the organizational level. As word for an example of other as~essrnents in which it has been used, processors and computers replace the typewriter, see V.T. Coates, et al., A Retrospective Technology Assessbookkeeping ledgers, and payroll systems, and rnent: Submarine Telegraphy (San Francisco: San Francisco Press, 1980); and V.T. Coates, The Potential Impacts of Roother communication systems augment telebotics, The Futurist, February 1982. phones, there are effects on productivity, size
PAGE 19
. Ch. lThe Outtook for Office Automation Technology, 1985-2000 l 9 Figure 1-2. History of Technology Used in the Office I (Third phase off!ce automation) 1 Local area networks, integratad systems. Non-impact printers. (Second phase otflce automation) Software packages for microcomputers. Microcomputers l Optical scannlng and recognittcm equipment l Video display terminals for data/text processing l [ (First phase offtce automation) 1 Facsimiie transmission l Etectronic(solid state) calculating machines. Microchip computera l Magnetic tape "seletric typawritars. Magnetic tape (replacing punched cards). Magrtetic ink character reoognit@rr (cfteck proof lngfsorfin@ l Electronic digitat computers (trsdstor$}. Efeotronic digital computers (vacuum tubes). I Electrification Era (1920-1960) 1 Data processing telewriters c Data processing computypers l Data processing paper tape or card: l Xerographic duplication l Mechanical listing printing calculators, 4 functions Punched card syatems (e.g., PayrOfl) l Oict@ting/stenographic machines with plastic belts l Common language concept for business machines l Bank check sorting/proofing machines Dialing Telephones l Elect r Machine accounting systems(central records, control, payroll) Multilifh duplicating (offset printinfj Addressograph/multigraph with automatic fl Adding/subtracting calculating machines [Ditto machines (gelatin duplcating) l Power statistical accounting machine l Bookkeeping and billing machines (combinations of typewriting and computing machlnesj l Loose-leaf leadger Sheet$ I 1 l Muttigrapff (Mechanical Era 1800-1920) l Two-color typewriter rfbbon Addressograph l Adding machine, Ilstlng and non-listing Hotlerith machtnes {card punch, tabulating and sorting machines) Cash qister l ComPtometar Oakutattrrg maohirm M\mtqr@f maoftina @wmil Wttln$ duWMiW Pnaumatie tufnis l Qlotating and $tanOgmpW faaohifraa l Wa@One l Catbon paper Typawritar # stia~ cwpming machine l Fountain pen 1 I I 1 I I I t I I I 1 I 1 I I I I I 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 SOURCE Off Ice of Technology Assessment
PAGE 20
10 Automation Of Americas Offices Figure 1-3. User Institution Model of Technological Change I Innovative technology technology may be modified to suit needs of users better Institution adopts technology to carry out existing functions more efficiently J lnstitution changes internally to take better advantage of new efficiencies J \ Other institutions I may be created, or old ones changed, to utilize new technology Competition between institutions develops new functions and activities made possibly by additional capabilities of the technology I institution Step I Step II Step Ill Step IV SOURCE Joseph T Coates, Aspects of Innovation Public Policy Issues in Telecommunications Development, Te/ecornmunications Po/icy, vol. 1, No. 3, June 1977. of work force, job content, the skill required of workers, etc. These effects are perceived as good or bad depending on ones perspective and interests. The institutional structures, culture, operating procedures, and management expectations at this point still reflect the old workflow and work process. Tension is created because the characteristics of the new technology are different, and the requirements for effectively using it are different. Until this tension is resolved, the full benefits of the substitution are not realized and productivity may even fall. Many organizations are still at this stage in office automation. The institution deliberately or unconsciously, by plan or by trial and error, begins to modify itself to suit new ways of doing things. This is the adaptation stage. The adaptations may include for example formal reorganizations, shifts in power relationships, adjustments in responsibilities, or changes in the way workers are recruited and compensated. Two kinds of problems may arise in this phase. If significant changes are made quickly and by explicit decisions, they may evoke resistance and resentment from those who lose power or who are uncomfortable with any change in the status quo, especially when people do not understand the reasoning behind
PAGE 21
Ch. lThe Outlook for Office Automation Technology, 1985-2000 1 1 the changes or have not participated in the decisions. On the other hand, if adaptations are not planned, there may be a long period of frustration and inefficiency before common sense indicates just what changes are necessary. The third kind of effects, transformations, come about because new technologies are likely to have entirely new capabilities not offered by the older technologies. The organization may develop new activities, products, or services, using these capabilities. For example, computers offer not only a more efficient way to do bookkeeping, but also the capability for continuous inventory control, not possible before. They make it possible to target mailings to special customers, and track the results. Some organizations began, as soon as they computerized their own data processing, to offer these services to others. This round of effects may bring about the restructuring of an industry or of the mix of industries within the economy. The financial services industry, 7 for example, used the new technologies to avoid legal boundaries between banks, insurance companies, brokers, and other elements of the industry. Some organizations fail to adopt new technology, even when it becomes the norm among competitors. They risk eventual obsolescence and failure. For example, mail order businesses that have not automated customer services are in serious trouble. The feedback loop in the model is important; further development of the technology is shaped by the market and by the demands of users. New businesses may be spawned that specialize in innovative use of the technology, or specialize in helping other firms use it. See U.S. Congress, Office of Technology Assessment, Effects of Information Technology on Financial Services Systems, OTA-CIT-202 (Washington, DC: U.S. Government Printing Office, September 1984). THE TECHNOLOGY OF OFFICE AUTOMATION PRESENT AND FUTURE The dominant trends in office automation, from 1985 to 2000, are likely to be: l l l l l l a continuing strong movement toward microcomputers and toward distributed data access and data handling, usually superimposed on rather than superseding centralized automatic data processing; more powerful, easier to use, software; a strong trend toward linking and networking of microcomputers, minicomputers, mainframes, and peripheral and supporting systems; increasing choice among technological options for accomplishing information handling objectives; more and more capture of data at the point of origin, decreasing the need for repeated keyboarding and centralized data entry; and growing capability for communication, be tween devices, between organizations, and between locations. Distributed Information-Handling and Networking In about three decades, there have been three overlapping phases of computer-based office automation: centralized computing, decentralized or end-user computing, and networking. 8 Before the last decade, large organizations were preoccupied with computerizing their mass data handling and typing, and were developing large systems and a corps of computer specialists to run them. This gave rise to the familiar EDP (electronic data processing) or ADP (automated data processing) center, staffed by computer specialists, primarily for batch processing of data and the development of large corporate databases. The reader who wants a more detailed description of office automation technology (but one still intended for the lay reader) and a discussion of the outlook for its development over the next 15 years, may go directly to app. A.
PAGE 22
72 Automation of Americas offices In industries that deliver customer services, such as insurance, this phase is sometimes called back office automation, since it chiefly affected the part of the office characterized by large numbers of clerical workers doing the kind of paper processing seldom seen by the consumer. They entered data into the computer using dumb terminals -i.e., using keyboards that fed data to a large central computer that did the processing. Many organizations established central word processing departments or pools, in which specialized clerical workers took over the typing (or at least the keyboarding of all lengthy documents) for a department or for the entire organization. The second phase of computerization began in most organizations about 1978 to 1980, with the introduction of small stand-alone word processors and microcomputers or personal computers (PCs), used by people who are not computer specialists. Software packages allow people who know little about computers or the arcane skills of programming to draw on databases or add to them, to manipulate text and quantitative data, to generate tables and graphic displays, and to exchange information with other computer users, without the direct mediation of computer specialists. PCs are increasingly used by managers and professionals as well as by support staff. Many executives who would not ever have typed now use word processing to draft letters, memos, or reports, or generate reports using spreadsheet software. This has come to be known by the awkward term end-user computing. Today, the use of computers by nonspecialists-end-user computingis a highly visible trend. End-user computing is not replacing central computers, but is often added to or superimposed over a centralized EDP process within an organization. The third phase is already beginning-the linking together of microcomputers, and the linking of microcomputers to mainframes or minicomputers so that they can act (compute) either independently or as an extension of the larger central processing unit. For the next decade, networking will be a major trend in office automation. Such linked systems can also connect computers with printers and copiers, and with outside communications systems (telex, telephone lines, cables, etc.) to create integrated office systems. Networking is not easily implemented. Because of the wide diversity in hardware, software, and interface mechanisms pro tided by vendors, it is often difficult to connect devices and systems so that they can talk to each other or work together as an effective system. In spite of these problems, in the last 2 years many organizations have developed networks or linked systems, and the trend is rapidly gaining momentum. Differences between centralized EDP computing systems and end-user computing will gradually blur. The first and second phases of office automation, considered separately, often appear quite different, but it will become less important whether a worker is using a dumb terminal, or a PC that is networked to other computers, because they will be able to access the same databases and perform roughly the same functions. Proliferating Options Broad choices among vendors, devices, systems, software packages, connecting devices, communications technologies, and service pro viders now characterize office automation. This range of choice, and the rapid evolution of the technology, creates problems for organizations that want to plan their automation rationally over a long period. However, it also allows offices to automate their work a few tasks at a time, if they so choose. Microelectronic office equipment is a highly competitive industry, and this has contributed to both declining prices and expanding capabilities. The computer industry has become a consumer industry. Large volume buyers (e.g., the Federal Government) have a less dominant influence over the direction of technological development than the cumulative choices made by the great number of middle and small size organizations. Even within large organizations, the purchase of microcomputers and word processors has often been relatively un-
PAGE 23
Ch. 1The Outlook for office Automation Technology, 1985-2000 l 13 controlled and the authority to choose between competitive brands has often been decentralized and dispersed. Thus, considerations such as the overworked phrase user friendliness have become important in the design of office automation. A striking feature of the market so far has been that only a few older office technologies have been eliminated completely. g Many have preserved their special niche, often by incorporating microelectronic componentse.g., the typewriter and microfiche. Technologies that were once separate are converging-typewriters become like word processors that in turn become almost indistinguishable from personal computers. Telephones incorporate small computers, and computers serve as communication devices. No one piece of equipment does everything but nearly all do more than one thing. As a result, users can put together devices and components to meet specific needs, and there are few obvious limits to what information-handling tasks and functions can be automated in the long run. Few critical technical barriers exist to future higher processing speeds, larger memories, much improved input and output technologies, and full communication between systems and devices without regard to distance. The Capture of Data A large proportion of the work in todays offices involves putting data into computers. Whether data is generated within an organization or drawn from outside the organization, it usually must be keyboarded into computers for further processing. This is especially true for data that is collected on a disaggregated basisorders placed or received, ticket stubs, invoices, checks, transport forms, vouchers, customer complaints, etc. When organizations exchange information either directly or through a client (e.g., payments and receipts, or a patient sending hospital bills to a health insurance provider) the information often must be Reprography has just about eliminated the multilith, mimeograph machine, and carbon paperbut not entirely. As yet, electric typewriters have not completely eliminated mechanical typewriters. rekey boarded even if it came from one computer and goes into another computer. But much of this work is being eliminated, or is likely to be eliminated in the future. Increasingly, computers are able to communicate directly with each other, through modems or other technological means. A second way of eliminating data-entry work is to allow (or require) a consumer or client to enter information directly into the organizations computer. This happens, for example, when a banks customer uses an automated teller machine (ATM) to deposit or withdraw funds or shift funds between accounts, or when he/she uses a home computer to instruct the bank to pay his/her monthly utility bill. A third way of eliminating data-entry work is to have the computer directly read typed or printed information with optical scanning technology; or to enable the computer to hear and store information conveyed by voice (i.e., speech recognition). Optical scanning devices are improving rapidly and increasingly in use; speech recognition technology is in an earlier stage of development, but is already being used in a limited way. A critical determinant of the results of office automation over the next 15 years is the outlook for computer input technology. Data entry, including word processing, is probably the largest single computer-related category of clerical employment today. Organizations are seeking and finding ways to avoid the necessity of keyboarding data for a second, third, or fourth time. Beyond that is the possibility of never having to keyboard it. If this happens, then both the number of jobs dedicated to data entry, and the costs of data handling will decline dramatically-at least, for a given volume of data. Communication Between Organizations As already noted, one way to avoid secondary entry of data is for interacting organizations to exchange data directly from computer to computer. For example, a hospital computer may send bills directly to the health insurers computer, which instructs the banks computer to transfer funds; the banks computer
PAGE 24
74 l Automation of Americas Offices Photo credit: Bell Labs One characteristic of office automation is the integration of computing and communication technologies then notifies the hospitals and the insurers computer that the payment has been accomplished. As another example, the computer of a manufacturer may receive an order from the computer of a customer, then order and coordinate the shipment, with the payment handled and recorded by direct communication between the computers of the buyer, seller, and bank. A few pairs of organizations are said to have such linkages operating at present; more are likely to do so in the future. How Rapidly Will Office Automation Occur? Office automation may proceed more rapidly and penetrate economic activities more thoroughly than have other waves of automation. The pace of technological change has, in the past, repeatedly confounded expectations; sometimes it has been slower than expected, sometimes more rapid. Much depends on the resources required to adopt new technology, the time required to recover the necessary investment, and the economic conditions that prevail. Particular factors of importance are the costs of capital and labor; the availability of people to use and manage the new technology; the structure of the adopting industry or industries, and their competitive environment; and a variety of social and behavioral factors that can be collectively called organizational culture. 10 An examination of these factors in relation to office automation points to a relatively broad and speedy adoption. The role of offices and office work is tending to increase in every sector (i.e., there is more paper work or information-handling in the production of all goods and services), and automation can be adopted by and adapted for offices in every industry sector. Many kinds of office activities are similar in all organizations-generating text, keeping records, circulating memos, preparing payrolls, filing, etc. Because many kinds of office automation can be implemented incrementally, and at relatively low cost, it can be adopted by small as well as large offices. There are a variety of reasons for automating information-handling, from reduction of labor costs, to improving the quality or variety of services, to reaching a larger market in a given time. If an organization adopts automation for the purpose of providing new services or products, or reaching a new market, its competitors may feel forced to do likewise. The effects of technological change often depend on how rapidly that change occurs. Problems may solve themselves if there is much time to adjust. On the other hand, opportunities not quickly grasped may be forever lost, or problems left to drag on may fester. Later chapters of this report will point to organizational and behavioral problems that could slow the adoption of office automation. At present however, it appears that these are largely transition problems for which many solutions are being developed. The analysis here and in ch. 2 draws on an OTA contractor report, J.D. Roessner, Market Penetration of Office Automation Equipment, Trends and Forecasts, November 1984. This report is part of an OTA contractor report, A.L. Porter, et al., Office Automation Outlook: 1985 -2000, February 1985.
PAGE 25
Ch. lThe Outlook for Office Automation Technology, 1985-2000 l 1 5 THE POSSIBLE CONSEQUENCES OF OFFICE AUTOMATION, 1985-2000 Some emerging and potential policy issues identified by OTA are of interest because contending parties have already voiced their concerns, and in some cases, related proposals are before Congress. Most of the policy issues, however, are of more long-range concern or are contingent on conditions anticipated but still in the future. In relation to most of the policy issues, some attempt has been made to indicate potential congressional responses. This is done to indicate the range of conceivable policy interventions. OTA has not fully evaluated all of these policy options in terms of their own possible effects, or the pros and cons of adopting them, either in this chapter or in the course of the assessment. Economic and Employment Effects Increased Productivity By the mid-1990s nearly every office will have at least one computer, just as nearly all offices now have telephones. There will probably be a terminal of some kind for at least every two or three office workers. Since many organizations will by then have adapted their work process and work environment to the new technology-although restructuring and change will surely continuemany solutions to current transitional problems should be available. Office productivity should increase significantly. Productivity in white-collar work is difficult to define and measure, as is discussed in chapter 2. But however productivity is perceived by a specific office or industry, significant increases as a result of office automation will affect employment levels, at least for some office occupations. Employment There will certainly be a significant reduction in the hours of labor associated with a given volume of information-handling. The magnitude of the reduction will depend in part on the technological trends noted above and in part on management strategies. The reduction in labor will be most significant in the clerical/support occupations, especially those that predominantly involve data entry. Fewer lower level workers would in itself indicate a need for fewer first line supervisors and managers. But the span of management control can also be broadened by automation of the work process. Thus, fewer managers may be needed (again, for a given volume of information-handling). Some of the tasks of lower level managers can be automated, or be taken over by clerical workers who are lower paid, so this tier of jobs is again likely to shrink. Some of the tasks of paraprofessional or technical workers may follow the same route. Professional occupations are less vulnerable, but not immune to the substitution and adaptation effects of automation. Whether or not organizations have as their primary motivation for adopting office automation a reduction (or constraint on growth) in the work force, relatively few can yet demonstrate that they have achieved that result. Many have hired more workers. OTA case studies, internal corporate studies, interviews with business executives, and reports in trade literature have repeatedly shown such shortterm inefficiencies resulting when a new technology is introduced into a workplace, process, working group, and organizational structure not yet designed to use it to best advantage. Yet most evidence suggests that in a given task, time and labor saving from automation is significant. Some tasks or steps are eliminated entirely. Declining costs and the proliferation of new uses and needs for information argue for a strong and continuing growth in the volume of information-handling and thus for a steady increase in the office workload. The growth in demand for information and the reduction in labor associated with information-handling
PAGE 26
16 l Automation of Americas offices are two competing trendsboth associated with information and communication technologies that will affect the level of future office employment. Which force outweighs the other is surely a judgment call; it may ultimately depend on broad economic conditions, other unforeseen technological developments, and choices and decisions made in industry, government, and society at large. Because in the past, employment growth as a whole has continued through periods of technological change, many economists believe that the number of office jobs will continue to grow strongly in spite of office automation. Others doubt this because of the high potential efficiency of microelectronic technology in office work (especially clerical work), which has always been very labor intensive, especially since the economy is generally not expected to grow as rapidly in the future as it has in some past decades. For example, a recent scholarly article suggests reasons for the common misconception that future economic growth spurred by high technology will necessarily provide a net increase in jobs; the authors said, especially speaking of computers that: ,, the impact of (new) technologies is likely to be more widespread than that of past technologies because their cost has declined so sharply relative to their capability and relative to the costs of labor (and) the economic context has changed considerably from the past. The present capability for economic and employment forecasting is not good enough to resolve this very large uncertainty. Thus, it is possible that there will be continuing strong demand for additional office workers. The most likely outcome, however, is slowing growth in the number of office jobs and even eventually an absolute reduction in the number of jobs in offices. While the latter outcome is by no means certain, there is sufficient evidence pointing in this direction to justify Russell W. Rumberger and Henry M. Levin, i Forecasting the Impact of New Technologies on the Future Job Market, Technology Forecasting and Social Change 27, 1985, pp. 399-419. watchful concern by Congress, and to merit efforts to improve the monitoring of employment trends so that corrective or compensating actions can be taken when and if they are needed. Recent employment forecasts should not lull policymakers into complacency. This is an area fraught with uncertainty. If the number of office jobs continues to grow, no congressional action will be called for; if that growth falters or reverses, there will immediately be strong demands for Federal intervention. However, early signs of trouble may be missed. At present, the Federal Government is poorly equipped to detect or understand early signals of problems arising from structural changes in the economy related to technology. Labor force and employment data collection does not support analysis of emerging trends because it is necessarily collected, aggregated or disaggregated, and analyzed in categories that reflect the occupations and jobs framed around old technologies or occupational/disciplinary tools. Census data is necessarily old (in terms of todays rapid change) by the time it is available for analysis. Econometric models are insensitive to realistic expectations about new technology and the changing substitutability of capital for labor. There is a great need for improved databases, for greater attention to advancing the state of the art in both technological and employment forecasting, and for building and institutionalizing a capability that combines and integrates those two disciplines. No agency now has a strong capability of this kind, nor the resources and specific mandate to develop such a capability; moreover, there is very little Federal funding available for research into crucial areas of economics, social science, and organizational behavior related to this area. The critical question for Congress at present is not one of immediate interventions, but how to improve the capability of the Federal Government to understand, and thus be prepared Hearings of the Science Policy Task Force of the House Committee on Science and Technology, Sept. 17-19, 1985.
PAGE 27
Ch. IThe Outlook for Office Automation Technology, 1985-2000 l 1 7 to respond to, technological and structural changes that are occ urring in the United States and other advanced industrial nations. Conversion of Employee Status Office automation may enable employers to convert more employees into part-time and temporary workers or independent contractors. This allows offices to adjust their labor costs to a fluctuating workload, to reduce some overhead costs, and to eliminate or externalize some secondary labor costs (benefits packages). While some workers prefer and seek opportunities for part-time employment, the number of involuntary part-time or temporary workers has been slowly and steadily increasing. Significant growth in the use of part-time and temporary work forces could ultimately reduce the number of full-time jobs for those who need or desire them. It may also shift responsibility for benefits and protective mechanisms such as health insurance, unemployment and disability insurance, and retirement income to the worker, costs that may ultimately have to be assumed in part by the taxpayer. Job Opportunities In some occupations and industries, those tasks that have historically been at the crossover point between clerical jobs and lower level management or professional jobs have been automated. In other occupations or industries, office automation has been used to narrow and specialize clerical tasks or to deprofessionalize tasks. In this way, some important channels of career mobility (job ladders) have been truncated. These effects are of special importance to women and minorities. In some industries or organizations those channels have only been opened to women and minorities within the last decade, and are looked to as a primary opportunity for escaping from poorly paid occupations and segregated womens work Office automation also creates some new categories of jobs, or new specialties within existing occupations. Proficiency in solving prob lems related to office automation has given many people the opportunity to create special roles for themselves and has led to new job ladders and career paths. Men and women with the capability, imagination, daring, and opportunity to rise to this challenge are moving into the upper levels of organizational hierarchies. Whether or not this career path will create anew pool of potential senior managers or chief executive officers, only time will tell. Some of the likely effects on employment are not inevitable results of office automation, but result from management/institutional choices about the combination of technologies adopted, the way they are implemented, and the way the work is restructured. But to the extent that the technology may encourage business decisions to be made only on the basis of narrowly framed, near-term cost considerations, this could have a negative effect on both long-term productivity and future opportunities for productive and rewarding employment for many Americans. Training and Education A major factor in achieving the full productivity benefits of office automation is the availability of workers skilled in its use. From the standpoint of organizations and their managers, an adequate supply of trained workers to meet future needs is critical. From the standpoint of office workers, access to training and the ability to master new skills is critical. In the short term, the training need not be lengthy or excessively expensive. The trend in both hardware and software is toward greater simplicity for the user, and more opportunisty for self-teaching based on training procedures built into the technology. However, because office automation often brings about a drastic redesign and resequencing of tasks, both initial and continuing tr aining is essential. Training provided by vendors and by some employers is highly systemand task-specific. For the user, this makes training easier, but limits its future usefulness. Economic pressures, however, motivate some employers to provide no more than minimum trainin g. Many
PAGE 28
78 Automation of Americas offices organizations, in deciding to automate, have underestimated the continuing training and support that is necessary, and these costs can dwarf the original investment. Some employers also see it as in their interest to make training as narrow, specific, and brief as possible. In that way they have minimum investment in workers who could then be lured away, and replacement workers can be trained rapidly. But this practice can also trap workers in dead-end jobs and limit their ability to move up to jobs of wider scope and higher skill requirements. Many organizations are providing their workers with ample training both in their own interests and from a sense of responsibility to their workers and to the larger society. This is a significant cost to the employer. While in the long run it may increase the productivity benefits that the organization seeks, it may also defer the realization of sought for cost reductions. Workers just entering the labor force also need training, at least in specific tasks or with specific systems. Those aspiring to management positions increasingly seek advanced training in business and business technology before entering the labor market. In recent decades, there has been a strong trend toward externalizing training costs and lateral recruitment-i.e., depending on colleges and commercial schools to train workers, and hiring from the outside at the management level rather than advancing people from lower levels within an organization. Enrollment in business schools, vocational schools, and adult education courses has been rising. The students, rather than an employer, bear the cost of this training; but since it is often out of reach for lower income families, this raises questions about equitable access. Public schools and community colleges are to some degree handicapped by the costs of providing instructors and a wide range of equipment for students to learn on, especially since the technology is constantly changing and instructors themselves must be continually learning. Commercial schools may be in some ways better able to respond quickly to market changes-i. e., to be able to invest in up-to-date technology and retraining for instructors. Again, the costs to workers of commercial courses, in terms of both money and time, mean that the most disadvantaged populations are put at a further disadvantage. Organizations and Jobs There is much controversy at present around the question of whether office automation enhances or de-skills office work. De-skilling means the standardization and routinization of tasks, in such away that human knowledge, judgment, and decisions are minimized and the technology (in this case a computer) directs, controls, and paces what a worker does. Office automation can be used to de-skill tasks. It is, and has been, so used in many organizations that process huge volumes of standardized data. More data can be processed in a given time, by less highly trained (and lower paid, more easily replaced) workers, with less variability in outcome and-supposedly but not always in practicewith lower error rates. Professional tasks can also be de-skilled, by building sorting rules, decision trees, and analytical processes into computer processes and software packages. For the worker, this may mean less job interest and satisfaction and increased stress. Factory-like offices could repeat the worst mistakes and problems created by manufacturing assembly lines. However, rationalizing work (i.e., simplifying and routinizing it) can also create new white-collar jobs for people with less ability and training than is required for some other office work. These jobs are highly valued by people whose employment opportunities are limited, or who are just entering the job market. Office automation can also be used to enhance jobs, by relieving people of routine repetitive steps. Jobs can be designed to integrate simple tasks into fewer broader tasks so that the worker has a better sense of the purpose and outcome of the work. Informa-
PAGE 29
Ch. 1The Outlook for Office Automation Technology, 1985-2000 19 tion systems can also give a worker access to knowledge that could previously be obtained only through advanced professional education, and thus allow a nonprofessional worker to take over interesting tasks previously considered the privileged province of the professional. For example, clerks can use computerized databases to search for information that formerly only a lawyer, medical doctor, biologist, or Ph.D. in history or anthropology would have known about. Some organizations are deliberately using office automation to upgrade and enhance work at all levels. As a result, they can sometimes develop a flatter institutional structure and reduce the costs of management. Other kinds of organizational restructuring are likely to occur. There is evidence that adoption of large computers tended to lead to centralization of control. Adoption of end-user computers could lead toward some decentralization of decisionmaking. These shifts in power depend less on the characteristics of the technology than on the characteristics of the organization and its management strategy. Almost always some redesign and restructuring of the flow of work is necessary, leading to shifts in responsibilities, jurisdictions, allocation of resources, and relationships between coworkers and between working groups. Communication patterns within the organizational hierarchy are likely to change. Typically problems arise over ownership of information and responsibility for assuring the integrity of organizational data banks. There is more flexibility in implementing office automation than in most changes in basic technology; thus management has an even greater responsibility for the outcome. The ways in which organizations are implementing office automation and redesigning their work processes are rich in innovation and diversity. Results in terms of productivity will be a matter of debate for some time to come. Some organizations will be disappointed in their expectations, and this may make others more cautious about moving into office automation. But a great deal of cross-learning and shared learning is occurring, and many organizations may learn from the experience of industry leaders and thus shorten their own troublesome transition period. Both in terms of productivity (because of the importance of motivation and job satisfaction) and in terms of equity and quality of work life, problems of work de-skilling, occupational downgrading, and disappearing job ladders could result in serious social problems. The handling of such problems is, in our system, usually a management prerogative or a matter for labor-management negotiations. Experience shows that the most fruitful strategy, both in terms of productivity and job satisfaction, is usually some form of worker participation or representation in the search for solutions. Less than 20 percent of office workers in the United States are represented by unions (although the growth of white-collar unions is said to be a high priority for labor organizations at this time). Organizations are now using or experimenting with other mechanisms for worker participation in decisions related to technological change and job design. Office Workers and Their Workplace With more and more Americans working in offices, there is growing attention to the quality of work life in offices, to job satisfaction, and to the effects of office work on physical and mental well-being. Office automation has aroused some added concerns in terms of the long-range effects of physical and mental stress, and fears related to work with computer video display terminals (VDTs). Workers using VDTs have increased complaints about eyestrain and musculoskeletal problems. Better workstation design, improved human/computer interfaces, and work breaks from long periods of VDT work can greatly alleviate these problems. There is no evidence as yet that such problems, while serious in terms of day-to-day discomfort, lead to any organic deterioration or chronic disease or illness. However, evidence from other occupations with heavy workload and repetitive tasks suggests musculoskeletal strain in VDT work may lead to chronic health effects.
PAGE 30
There has been serious concern over reports of clusters of reproductive failures and accidents among clerical VDT workers. Epidemiological investigation has so far not explained these scattered clusters; but scientific research has failed to find any possible cause related to VDT technology. OTA has reviewed current information and ongoing research in this area and concludes that there is at this time no good basis for fear of VDT effects on reproductive processes. Scientific research should continue to be monitored for new findings. However, office automation can increase stress on users. Stress is not always bad; it can be viewed as increased challenge. But some forms of continuing, unrelieved stress are clearly harmful. Concern in the past has emphasized the so-called executive heart attack, but there is growing evidence that clerical workers are most likely to suffer from continued stress. Computer pacing and computer monitoring, which result in reduced autonomy or control over ones work and performance level, or which induce continuing visual and musculoskeletal strain, significantly add to stress. Fear and anxiety over ones ability to learn and perform effectively or to change ones management and supervisory strategies, as well as perceived job insecurity, create stress. There is increasing evidence that long-term high levels of stress are conducive to several kinds of chronic illnesses. Visual and musculoskeletal problems currently can be alleviated through appropriate office and workstation design; stress-related illness or disease is more difficult to control and may emerge as the greatest public health problem among office workers in the future. Involvement of workers in decisionmaking about office automation and work redesign appears to reduce stress levels. Organizations that have successfully sought the participation of users at all levels of the organizational hierarchy have reported that this led to higher productivity. In some countries, there is already legislation or regulation related to office automation and quality of work life; for example, West Germany and France have regulations requiring regular eye examinations for VDT workers, and in Sweden there are standards calling for periodic work breaks. Data Security and Confidentiality Data security and the confidentiality of data in computers and data banks is a continuing concern. End-use computing adds to these concerns primarily because there is wider access to data and to means for manipulating it; data disks are easily lost, stolen, destroyed, or copied; computers are linked to other computers providing greater access; and small computers are not usually monitored or physically guarded. Control of databases is often separated from responsibility for their integrity and reliability. Most end-users are not as well informed about the requirements of confidentiality for client and employee data as are computer professionals. In addition, office automation often allows work to be done away from the office, in airplanes, at home, or in client offices, which further increases vulnerability. Data can also be destroyed or made temporarily inaccessible by accidents, electrical outages, or natural disasters. As more and more data exists only in electronic form, and as day-to-day operations become more dependent on technology, there is greater need for secure back up. Federal agencies, like private sector organizations, are faced with new problems in protection of data security and confidentiality as a result of decentralized office automation. But government offices have, in many cases, been more sensitive to this problem than has industry in general. Personal data about Federal employees or agency clients, however, may be inadequately protected. There are both technological and procedural means for protecting data. Most organizations
PAGE 31
Ch. lThe Outlook for Office Automation Technology, 1985-2000 l 2 1 have established these with regard to centralized EDP operations, but often overlook the vulnerability y of data processed by end-user devices. Organizations moving into office automation for the first time may not realize the threats until serious problems arise. Home-Based Office Work A special and important capability of office automation technology, resulting from the convergence or combination of information and communications technology, is to allow work to be done at a distance from the office. The workers home can become the primary work site. Home-based clerical work, with the work performed on a piece-rate or hourly basis, could increase significantly in the next few years. Already controversial, it would then become much more so. The opportunity to work at home is highly valued by many workers. Many professionals now do part of their work at home, by their own preference and at their own and their employers convenience. This flexibility is mutually beneficial and should be protected. A second kind of home-based, computer-mediated work is the performance of office services (e.g., word processing, data entry, analytical studies, computer programming, or consulting) by entrepreneurs who establish small businesses in their own homes. They sometimes find barriers to operating their businesses in the form of zoning laws or confusing tax provisions, but such ventures are proliferating in spite of this. Many women, in particular, are gaining business skills and experience in this way. A more controversial form of home-based office work is the farming out of clerical office work (e.g., data entry and word processing) to be done in homes, usually by women and especially by mothers of small children. Homebased workers are often paid at piece-rates or hourly rates. They may be former employees converted to the status of independent contractor, thus giving up employee benefits. Critics object to this form of home-based work on the grounds that it eliminates jobs for regular employees and constitutes unfair competition, tending to depress the general level of clerical wages and preserve the segregation of women into the lowest level office jobs. They also argue that it is difficult to assure the workers of safe and healthy working conditions; and that it weakens social pressure for the establishment of child day care centers, accommodations for handicapped workers in the office, and other social services needed by workers employed under more conventional conditions. A work-at-home opportunity at present is almost always sought by the worker, not forced by the employer. The most common reason is the need or the wish to combine paid employment with care of children or other household responsibilities. Some home-based work programs are designed for physically handicapped or retired workers. At present only a few thousand people are engaged in homebased clerical work. However, the technological and economic conditions exist for substantial expansion in the future. If home-based workers begin to compete with office-based workers for a shrinking number of clerical jobs, then this issue will be much more highly visible and controversial. Bills have been introduced before Congress that would encourage home-based computer work by means of tax credits. The AFL-CIO has called for a ban on home-based clerical work. Other groups argue for enforcement of existing occupational safety and health laws and other worker protection laws in homebased work. There are legal questions to be resolved regarding the status of some of those designated as independent contractors; IRS has recently ruled that those accepting work only from one employer/organization must be regarded for at least some purposes as employees, not contractors. Off-shore Performance of Office Wor k Some U.S. firms have relocated their dataentry operations to other countries to take advantage of low labor costs. Intermediary en-
PAGE 32
22 l Automation of Americas Offices trepreneurs also make off-shore data-entry services available to U.S. firms. Recent advances in communications technology are making this activity increasingly cost effective, and it is also being encouraged by the Federal Government as a mechanism for assisting economic development in Caribbean countries. However, if predicted technology developments reduce the need for large-scale data entry, off-shore sourcing of data entry will probably cease to expand. Off-shore clerical work for U.S. firms now involves only a few thousand workers, chiefly in the Caribbean countries but with some in the Far East and India. With further improve ment in communications services, this activity could expand considerably in politically stable developing countries in many regions of the world. The governments of many developing countries, and U.S. economic development experts, perceive valuable potential benefits in encouraging this activity. It does, however, represent a direct loss of U.S. data-entry jobs. Federal Government Office Automatio n Effects of office automation will be felt in public sector as well as private sector offices. In Federal Government offices these effects have implications for procurement policy, personnel policies, and budgetary planning. They may have implications also for the delivery of government services, the access of citizens to government information, and the ability of citizens to participate in public decisionmaking. This report only briefly considers such effects on the quality of governance, primarily in terms of the possibility that loss of accountability could result from erosion of established bureaucratic communication channels. Present procurement and acquisition policies have, in general, allowed the Federal Government to keep pace with the private sector and to adapt office automation technology to the needs of Federal agencies. There are some conspicuous exceptions and some unsolved problems. Chief among these problems are the lack of compatibility among devices, which is hampering networking within agencies and between agencies and field offices; and the lengthy procurement cycle for large systems, which is out of step with the rapid evolution of the technology. Gains in Federal office productivity are obvious, yet hard to measure or document. Some agencies have been able to handle greatly increased workloads without a proportionate increase in the work force. There are, however, built-in disincentives to achieving maximum cost reductions. Government managers are offered few rewards, and may suffer subtle penalties for reducing the number of people they supervise or cutting their annual expenditures. Grade level or promotion may depend on the number of people one supervises, and unexpended funds may encourage further stringent budget cuts. There has been strong pressure by the Administration to reduce Federal employment. There are some indications that office automation may have already contributed to curbing the growth of the Federal work force. A significant reduction in the number of clerical workers and a change in the ratio of clerical to professional and administrative employees seem to have occurred, along with changes in job content and skill requirements that are likely to continue to accelerate. The Federal Government has a responsibility to see that changes in Federal white-collar employment, working conditions, and career opportunities are managed smoothly, equitably, and with due concern for civil service employees. There should be continuing reexamination of job classifications and classification criteria. Uncertainties in future employment levels should affect present recruitment and hiring, and projected changes in future employment needs should also guide the planning for retraining and redeployment of present employees. Such changes are not being systematically tracked, studied, or considered in personnel or budgetary planning. Since 1975 there has been a slight rise in average grade levels, which was strongly criticized by the Grace Commission as overgrad-
PAGE 33
Ch. lThe Outlook for Office Automation Technology, 1985-2000 23 ing. The changes in occupational distribution (i.e., a smaller proportion of clerical workers) may have contributed significantly to this slight rise. Neither the Office of Personnel Management nor Congressional Budget Office has accounted for this factor in their analysis of the grade level rise. If the number of lower level workers has been reduced, and will be further reduced, then efforts to hold constant the average grade level will significantly hinder the ability of the Government to attract and hold highly qualified middle and upper level employees. State and Local Government Offices State and local governments are struggling to manage an increased workload, partly resulting from present Federal policies, without a proportionate increase in staff. The more than 78,000 governmental units in this country show a rich diversity in approaches to office automation, both in large systems for many governmental operations and more recently in end-user computing. Small cities and rural counties are lagging in use of office automation, in spite of the opportunities presented by microcomputers. In part, this probably results from lack of access to expertise to help them in choosing and supporting office equipment and training workers. There is, at this early stage, little information about the aggregate effect of automation on local and State government offices across the country. Evaluations of productivity so far show mixed results. Some researchers, in specific States and localities, have found evidence of increased productivity and of work force reduction, others have not. Some researchers have reported findings of standardization and depersonalization of government services, of strong reinforcement of the existing distribution of bureaucratic and political power, and of increased difference in the relative access of citizens to government information. Small Business and Office Automation The effects of office automation on small businesses are of particular concern to Congress because of the vital role that these organizations play in job creation and in innovation. Small computers and improved software packages appear to make office automation more practical for small organizations, but can still represent a significant capital investment in comparison to their assets. The time required to make informed decisions about equipment, to redesign work procedures, and to train staff, plus the lack of in-house expertise in trouble shooting and problem solving are also significant problems. Although computer vendors identify small business as an active and growing market, there is little empirical evidence about the results of their experience with office automation so far. Optimists hope that automation may allow small firms to expand their markets and successfully compete with larger organizations. Pessimists fear that many will incur capital costs beyond their ability to support. Their experience should be carefully monitored for emerging public policy concerns. Working Women and Minorities The effects of office automation are of particular concern to women. Most clerical jobs are now held by women, and one-third of working women are in clerical occupations; they are vulnerable to displacement. Women now in clerical positions are trying to move into managerial and professional jobs, but some job ladders may be truncated by automation. In managerial and professional occupations that are vulnerable to office automation, women tend to have less seniority than men. On the other hand, some new jobs and occupations, offering good potential for advancement, are being created by office automation; access to education and training for these specialties is vital for women. As skill requirements and training prerequisites for traditional office jobs change, their comparability to other jobs changes, but these shifts may not be reflected in changes in job titles or even in formal job descriptions. Debate over pay equity and comparable worth requires understanding of the changing nature of office work and its changing skill requirements.
PAGE 34
Office automation is also of particular inMinorities, especially Black and Hispanic womterest to minorities, and for similar reasons. en, are disproportionately represented in jobs Clerical jobs have often been the first step in likely to be directly affected by office autowhite-collar work for disadvantaged groups. mation. POLICY ISSUES FOR CONGRESS A number of policy issues have been identified above, and are discussed in more detail in the following chapters. OTA has concluded that: Many of the concerns that have been raised with regard to office automation are transition problems that will be solved by market forces and by the common sense and ingenuity of usersthey do not require Federal action of any kind. Some concernsmost importantly, the possible effects on future white-collar employment levelsare matters of long-term national interests. They merit watchful attention, and Congress should begin now to make sure that the nation is prepared to deal with them if and when conditions warrant action. This however will be difficult to do, because the capability to detect and understand emerging structural change in the economy is poor. Better economic and employment data and development of improved forecasting techniques are urgently needed. There are a few specific issues of immediate concern; some of them are already the subject of legislative proposals before Congress or State legislatures. These include, for example: documented or suspected health problems associated with office automation, l l l proposals to encourage or to ban homebased office work, and largely unexamined relationships between pay equity and comparable worth questions and the changes associated with office automation. There are also specific issues of concern to Congress in its oversight of Federal agencies and activities, including: the implications of Federal office automation for procurement, personnel, and budgetary policies, and adequacy of present provisions to assure data security and confidentiality and continuation of Federal functions under emergency conditions. An emerging issue, which may become more important and more controversial in the near future is the beginning trend toward off-shore sourcing of data-entry work. There are several groups for whom the effects of office automatic are particularly important and whose interests merit special attention: women in clerical, managerial, and professional office positions, minority workers, and small business firms. FURTHER DISCUSSION: POLICY ISSUES AND QUESTIONS Employment fice jobs is already affecting large numbers In the immediate future, the anticipated efof people. It is therefore attractive to think fects on employment are only possible or probof immediate actions that would position Conable. There is little evidence that decline in gress to act in timely fashion when and if the growth in, or reduction of, the number of ofneed becomes real. A prudent preparatory
PAGE 35
Ch. 1The Outlook for Office Automation Technology, 1985-2000 25 strategy would be systematic monitoring and research to clarify the emerging effects of office automation and provide an alarm signal if they seem likely to exceed acceptable limits. This would require actions to improve the collection and collating of occupational and employment data, and improved mechanisms for monitoring and analyzing structural changes in the economy. It is difficult to track and demonstrate even long established employment trends because the kinds of data collected by the Bureau of the Census, the Bureau of Labor Statistics, and other government agencies change over time. The way in which those data are aggregated or disaggregated, and the way they are labeled also change. The various bureaus and agencies differ, at any one time, in the way they define categories, so that their data often cannot be compared and the data from one agency cannot be used to augment and supplement, or explain, data from another agency. For example, cashiers may or may not be included with other clericals; Federal employees may or may not include postal workers and congressional employees; part-time workers may or may not be distinguished from fulltime workers. Sometimes only a highly specialized data expert can disentangle employment data to address a simple question; some questions cannot be answered at all because data does not exist. It maybe impossible, for example, to determine when, and under what conditions, blacks entered office employment in significant numbers because no records of black/white employment were made until recent years. What data is collected and how it is recorded depends, of course, on what questions one expects to address. At the present time, no Federal agency has a clear mandate, nor the available resources, to develop an indepth capability for economic and employment forecasting of the kind Congress and the executive branch will need in a future in which technological change will be continuing and rapid, and will have pervasive effects on all economic activities. Those agencies with primary responsibility for collecting economic and demographic data have little capability, and few resources, for understanding advanced technology and the way it is likely to evolve in the future. The Federal agencies with the most highly developed knowledge of new technology, including computer technology, have no clear mandate, little capability, and few resources for analyzing economic and employment effects. Moreover, the primary government source of forecasts of commercial information technology, the institute of computer Science and Technology of the National Bureau of Standards, has recently suffered a severe cut in support of its planning and forecasting functions. This indicates that the ability of Congress to be apprised of disturbing trends in office employment, or the broader area of whitecollar employment, should they occur, is not likely to improve in the future unless steps are taken to ensure that this capability is being developed within Government or that support is available for such research and analysis outside of the Government. If it should become clear at some future point that white-collar employment is indeed not growing, or if there are signs of increasing structural unemployment, then Congress will need to consider interventions. At a minimum it may be necessary to take steps to assure that changes in employment opportunities do not differentially burden certain groups of workers, for example, women and minority workers. Workers in low-level clerical jobs may need special assistance in finding other employment. Steps may be needed to assure that there are entry level jobs for untrained workers, with on-the-job training opportunities. Sometime within the next 15 years it could become necessary for Congress to take positive steps to deal with a declining number of office jobs or white-collar jobs, whether or not the economy as a whole remains strong. What steps are appropriate will, however, depend in part on whether there are available jobs in other employment sectors for which office workers may be retrained. The many strategies for consideration in responding to high levels of employment have not been a focus
PAGE 36
26 Automation of Americas Offices of this assessment; they have been discussed many times in many places in continuing debates about long-range employment policy. They include share the work or shortened workweek strategies, strategies to stimulate economic growth and the creation of new industries, and strategies to ensure or augment family incomes. Most discussions of such policy actions in the past have focused on blue-collar unemployment. Their use to deal with whitecollar unemployment would require care ful consideration of the economic, social, and political conditions under which they would operate. Conversion to Part-Time, Temporary and Contractor Statu s A large increase in the number of full-time jobs lost to part-time or temporary workers without worker protection mechanisms (perhaps Occurring due to office automation) could ultimately burden taxpayers or lead to degradation in the level of social well-being. If Congress chooses to intervene, it could constrain conversions of employees to other status by legislative confirmation of recent IRS rulings that an independent contractor accepting work from only one organization is in fact an employee; a mandatory minimum ratio between full-time and part-time or temporary employment; or mandatory pro-rating of benefits packages and social security contributions. Alternatively, Congress could alleviate some of the problems of involuntary part-time employment by further actions to encourage or require Voluntary Reduced Work Time arrangements, making involuntary part-timers eligible for partial unemployment benefits. Training Because continuing training and education are essential both to realize the productivity gains that office automation promises, and to alleviate undesirable effects on employment, Congress may want to consider actions to assist public school systems in planning and upgrading white-collar vocational training; encourage school systems and community colleges to provide retraining and continuing training programs; establish accreditation for commercial wh.itecollar vocational schools; or direct the Departments of Labor and Education to develop guidelines for vendors and employers in designing tr aining programs. Should there be signs of developing structural unemployment among clerical workers, Congress might then consider providing loans or other assistance to white-collar workers seeking training or retraining to move into other jobs and occupations. Labor/Managemen t In general, the transitional problems involved in changing the technological base of an organization are best solved by the cooperative efforts of managers and workers in redesigning the work process and organizational structure. The public interest in this process is that of realizing the promise of increased productivity for the economy and encouraging fair play for those whose jobs and occupations will be affected. Because most office workers are not represented by unions, management has a particular need and responsibility to provide opportunities and mechanisms for involving employees in decisions about technology and its implementation. Given a minimal Federal role, Congress may wish to consider the desirability of educational or information programs for employers, employees, and their associations with regard to problems associated with office automation and strategies that organizations have successfully used for alleviating them. Other possibilities to be considered under some circumstances are clarification of worker rights under existing labor-management negotiation procedures, or changes in labor law to require worker involvement in technology-related decisions. Health and Safety In view of the concern about suspected hazards associated with computers and VDTS, options for Congress to consider include: 1) a public information program to inform employers and office employees about what is known
PAGE 37
. Ch. 1 The Outlook for Office Automation Technology, 1985-2000 27 about the effects of office work on health and well-being, and what can be done to reduce visual and musculoskeletal strain and psychological stress; and 2) directing the Occupational Safety and Health Administration to develop ergonomic advisories, guidelines, or standards for use in public and private sector offices and in design of equipment and furniture used in office automation. If these actions and self-initiated organizational management actions to alleviate these concerns are inadequate, then Congress may wish to consider legislation limiting the number of hours that workers can be required to spend continuously in VDT use. In any case, because scientific evidence is still equivocal, and significant risks involved in computerized office work would affect a large proportion of Americans, Congress could support the funding of research on the relation of stress to chronic disease and illness and on the possibility y of reproductive hazards related to VDT use. It may also wish to establish national mechanisms to monitor the health status of American office workers. Data Security and Confidentiality Previous actions to increase data security and ensure confidentiality in the handling of information about citizens have concentrated on risks from large centralized computer operations. Most organizations have been slow to recognize additional risks to confidentiality and security that come with decentralized computers. These risks are largely related to the increased access to organizational data banks; the portability of data storage media; and the lack of knowledge among general users about privacy laws, the principles of confidentiality, and established practices for safeguarding data. Information about clients, employees, and corporate resources and activities is subject to compromise or misuse. Organizations need to develop policies and practices that better protect the confidentiality and security of data that is processed or stored by small computers or other end-user devices. Existing Federal privacy and security laws are largely designed to strengthen the ability of an individual to challenge the use of information about himself or herself. If lack of attention by organizations to data protection in decentralized computing leads to serious abuses, Congress may need to consider more stringent data protection laws, including liability for breaches of privacy and security. Home-Based Clerical Work Because this activity is already controversial, Congress may have to formulate policy regarding it within the next several years. Proposals have been made in Congress to provide tax credits for the purchase of computers for home-based work (and other nonrecreational uses). Other ways to encourage homebased work include legislative actions voiding IRS rulings and court decisions that make many home-based workers employees rather than independent contractors, and removal of other regulatory or tax barriers to home-based work. However, Congress may instead wish to discourage home-based computer-mediated clerical work. It could then consider a ban on paid computer-related employment when the primary or sole work site is the workers home, but would probably wish to distinguish this employment from similar activities designed to establish entrepreneurial small businesses. Congress may, instead, wish to clarify through legislation the conditions under which homebased workers are employees entitled to the normal worker protections afforded by law and equity, and to require pro-rated benefits for workers accepting work to be done in their homes. Instead of banning or strongly discouraging home-based office work, Congress could insist on rigorous application and enforcement of existing worker protection laws and regulations to home-based work, including wage and hour laws, occupational safety and health regulations, and all applicable reporting requirements. This strategy would require
PAGE 38
strengthening the resources and capabilities available to enforcement agencies. Rather than banning or discouraging homebased work, Congress could adopt a strategy of increasing the range of employment options for workers through tax deductions or provision of day care centers; alternative means of caring for children, aged, or disabled dependents; and by further acting to increase mainstream employment opportunities for handicapped workers. Off-Shore Office Work By or For U.S. Firms The issue to be resolved here involves the desire for cost-saving for U.S. firms and economic development in Third World Countries v. the demand for preservation of clerical jobs in the United States. Congress may wish to encourage off-shore sourcing, which could be done by extending further technical assistance and information to interested companies in the United States and to development officials in prospective host countries, or merely by taking no action that would counter presently favorable market forces. If Congress wishes, on the other hand, to discourage off-shore sourcing, it could do so through restrictions on data flow; additional privacy protection laws, requirement for more extensive security measures, etc.; regulations analogous to local content or buy national requirements; imposition of taxes or tariffs, such as value-added taxes or trigger price tariffs; limitations on the availability or use of dedicated telecommunications lines; or an outright ban on offshore sourcing. Federal Procurement Policy The present procurement policy for automating Federal offices has worked well, but there are some problems impeding further progress. Congress may want to consider asking the General Accounting Office and the Congressional Budget Office (CBO) to reexamine the effects of present procurement policies and regulations on the ability of Federal agencies to procure state-of-the-art systems, and to plan toward integration and networking of Federal microcomputers and related devices. Federal Personnel Policy To assure that emerging effects of office automation are effectively managed, with full regard for the rights and interests of civil service employees, Congress may want to consider asking the Office of Management and Budget, the Office of Personnel Management, and CBO to conduct studies and prepare recommendations for changes in personnel recruitment, retention, job classification, and promotion and compensation policies to reflect changes in personnel needs, job content, and skills requirements resulting from office automation. State and Local Governments Federal policy is to shift responsibility for many decisions and programs to the local level. To handle these programs effectively and at the lowest cost to themselves, local governments need to increase office productivity. Small government units are lagging in automation. Congress may want to consider block grants to State governments for this purpose, direct technical assistance to small city and county government units, or other means of assisting small governments in office automation. Small Businesses Although there is a risk that small businesses will create fewer jobs if they are encouraged to automate their offices, this could be offset by increased viability and competitiveness, and enhanced opportunity to grow. Little is known as yet about small business office automation. Congress may wish to consider requesting a study from the Small Business Administration or other Federal agencies of the potential opportunities for and consequences of office automation in small businesses, especially with regard to their overall viability and their role in innovation, local economic development, and job creation. If addi-
PAGE 39
Ch. 1The Ouflook for Office Automation Technology, 1985-2000 l 2 9 tional evidence suggests that there is a public l interest in assisting small business office automation, options to be considered include an information and education program for small organizations considering office automation, l or specific technical assistance or financial loan programs for small businesses for the purpose of office automation. l Women and Minoritie s If office automation reduces the number of l office jobs, women and members of minority groups who have historically been disadvantaged in the job market, are likely to be most negatively affected. Congress may then wish to consider a series of steps to alleviate their disadvantage. It could: take explicit notice of these concerns and factors in discussion of bills concerning pay equity and comparable worth, which are now before committees; provide incentives for maintaining or improving ratios of female to male employment in labor force reductions associated with automation; provide incentives for maintaining the share of employment held by minority groups in office work; and provide subsidized child care facilities or increased tax deductions for child care to increase the employment options open to working parents.
PAGE 40
Chapter 2 Productivity and Employmen t
PAGE 41
Contents Page The Uncertain Outlook. . . . . . . . . . . . . . 33 The Framework for Analysis . . . . . . . . . . . . 34 What Follows.. . . . . . . . . . . . . . . 36 Recent Labor Force Projections . . . . . . . . . . . 37 Productivity and the Nature of White-Collar Work. . . . . . . . 45 How Office Automation Affects Employment . . . . . . . . 47 Emerging Occupational Shifts . . . . . . . . . . . 50 Technology and Recent Trends in White-Collar Employment. . . . . . 56 Part-Time and Temporary Employment . . . . . . . . . 59 Analogies From Past Waves of Automation . . . . . . . . 64 The Future White-Collar Labor Supply . . . . . . . . . 67 Conclusions . . . . . . . . . . . . . . . . 68 Policy Considerations: Labor Market Adjustment Options . . . . . . 68 The Need for Monitoring of Structural Economic Change Related to Information Technologies . . . . . . . . . . . . 68 Longer Range Policy Options . . . . . . . . . . . 69 What Action Is Needed Now? . . . . . . . . . . . 71 Tables Table No. Page 2-1. Growth unselected Occupations as Projected by the Bureau of Labor Statistics, 1982-95 . . . . . . . . . . . . . . . . 38 2-2. Percent of Total Work Force in Occupational Groups by Selected Industry Sectors, 1982 . . . . . . . . . . . . . . 51 2-3. The Growth in the Number of Clerical Jobs, 1950-84 . . . . . . 57 2-4. Changes in the Percent of Total Employment for Involuntary and Voluntary Part-Time Workers as Percent of Total Employment, 1968-84. . . 59 2-5. Shifts unemployment by Industry Sectors, 1900-80 . . . . . . 65 2-6. National Unemployment Rates During Recession Troughs and Recovery Peaks, 1961-84 . . . . . . . . . . . . 66 Figures Figure No. Page 2-l. Framework Describing the Relationship of Office Automation to Changes in the Number of Office Jobs. . . . . . . . . . . . 35 2-2. Framework for Analyzing Long-Range Effects of Office Automationon White-Collar Employment . . . . . . . . . . . . 50 2-3. Changing Percentage of Work Force From 1900-80 White-Collar Compared to Blue-Collar. . . . . . . . . . . . . 56 2-4. Capital Investment Per Production Worker, 1970-80 . . . . . . 57 2-5. Changes in the Percent of Total Employment for Involuntary and Voluntary Part-Time Workers, 1968-84 . . . . . . . . . 59
PAGE 42
Chapter 2 Productivity and Employment Office automation promises an increase in productivity in all sectors of the U.S. economy, because all industry sectors increasingly depend on information as a major component of products and service. Office automation will significantly reduce the number of jobs required for a given volume of informationhandling work. It will also fuel a growing demand for information and information processing. Thus, it is possible that there will be continuing strong demand for additional office workers. The most likely outcome is, however, slowing growth in the number of office jobs, and eventually, an absolute reduction in the number of jobs in offices from some peak in the 1990s. While the latter outcome is by no means certain, there is sufficient evidence pointing in this direction to justify watchful concern by Congress, and to merit efforts to improve the monitoring of employment trends so that corrective or compensating actions can be taken when and if they are needed. Recent employment forecasts should not lull policy makers into complacency. This is an area fraught with uncertain y. THE UNCERTAIN OUTLOOK The first two conclusionsthat work required for handling a given volume of information will decrease, and that demand for information will grow-can be made with considerable certainty. Between these two conclusions and the third, that the number of office jobs may ultimately decline, lie several large questions. The most important is, how much will the total volume of information-handling increase? Second, what are information-handling jobs, and which ones will be affected? Third, if there is a stable or lower demand for white-collar work, are there ways of adjusting without creating structural unemployment ? This assessment indicates that, with the amazing capability inherent in computer and communications technologies, even the needs of a thoroughly information-driven economy do not, in the long term, assure rising levels of white-collar employment equal to growth of the labor force. That labor force by the year 2015 will be about 142 million people or 35 percent larger than it is now. But the number of people entering the work force each year is declining. If the transition to stable or declining office employment is slow enough then the negative effects would be muted. If there is strong economic growth, then office employment could continue to grow, although more slowly than in the past. Strong continuing economic growth like that of the 1950s and 1960s is not, however, certain in the future. The overall employment effects of the first phase of office automation, the large computers installed in the 1960s and 1970s, have been hard to detect amid other changes in the economy. But the second phase of office automation, decentralized computing and advanced communication capability, is spreading rapidly. The third phase, the evolution of integrated office systems or networking, will bring about much restructuring of the flow of work in and between offices. Delayed effects of the first phase and the emerging effects of the second and third phases are overlapping. From this perspective, it is possible to see the latent productivit y enhancing effects of office automation as water building up behind a dam. The dam is made of institutional inertia and the unavoidable transition problems. When that is removed, there could be a flood of work force reductions, unless 33
PAGE 43
34 Automation of America's Offices workers are channeled into productive new jobs and industries. Organizations often resist laying off workers when they adopt labor-saving technology, preferring, when possible, to let attrition solve the adjustment problem. This is one example of institutional inertia. But every year some firms go out of business and some new businesses are started. New organizations are likely to use new technology from the beginning, creating fewer jobs than new starts would otherwise create. This is probably already happening; for example, Dun and Bradstreet reported that in 1984 more new firms were started than in 1983, but they employed nearly 5 percent fewer people. In the financial, insurance, and real estate industries, which are the leaders in office automation, the number of workers hired by new organizations declined by over 9 percent. If new organizations tend to create fewer jobs because they make effective use of new technology, then employment effects would tend to accelerate over time. Which white-collar jobs will be affected? Most directly and strongly, they will be clerical jobs. The number of professional and managerial jobs is apt to be less strongly affected, and professional work may continue to expand indefinitely in an information society. However, even managerial and professional jobs are not immune from the labor-saving effects of office automation. Businesses are now engaged in strong efforts to reduce labor costs and to increase productivity. Forbes Magazine reported that the nations 500 largest publicly owned companies (which account for about a fifth of all civilian employment) expanded their total sales in 1984 by 4 percent and at the same time shrank the number of people on their payrolls by 4 percent, or 840,000, 2 Sales per employee rose by 10 percent, and assets per employee rose by 11 percent; these are two rough measThe 10~,329 new starts in 1984 hired 578,838 workers compared to 607,416 new hires for the 100,868 new starts in 1983. See Dun and 13radstreet Looks at Business, a Dun and l~radstreet subscriber newsletter, vol. 3, No. 3, ill ay-June 1985. -Forbes hJagazine, Apr. 29, 1985, p. 231. ures of increasing productivity. There are other signals of pressure on employment growth despite continuing job creation. During recent business cycles, unemployment rates in business recovery years have remained higher than they were in recession years before the mid-1960s, in spite of the creation of thousands of new jobs. The number of involuntary parttime workers is continuing to grow. These indicators are not tied directly to office automation, yet there is a strong possibility that some of the effects on employment oft wo decades of office automation are now becoming apparent. The Framework for Analysis The long range effects of office automation on office employment are pictured, in figure 2-1, as a dynamic interaction between: l growth in demand for information, and the labor-saving characteristics of office automation technology. Information handling and communication play a larger and larger role in all economic activities. This is what is meant by an information society. Computers whet the appetite for information. More kinds of information can be collected. It can be analyzed in more ways, thereby producing still more information. All of this data can be used in new ways and easily communicated to more people. Organizations are likely to find more internal or intermediate uses for information, and to produce new information-intensive services or products as office automation technology makes it possible to combine, package, and distribute information in innovative ways. Consumer expenditures appear to be shifting from hard goods to soft goods and services, as evidenced by the growth in service industries and in particular by the proliferation of new financial services in recent years. Because information-handling costs have always been primarily labor costs, and because office automation technology is labor-saving, it is assumed that the cost of informationhandling will decrease. This should, accord-
PAGE 44
Ch. 2Productivity and Ernployrnent l 3 5 Figure 2-l. Framework Describing the Relationship of Office Automation to Changes in the Number of Office Jobs cost of information + Demand for information o \ New information products and services SOURCE Off Ice of Technology Assessment Time-saving (conventional tasks) Labor-saving characterlstics of office automation technology + ing to standard economic theory, tend to further increase the market for information. Office automation is adopted for many reasons, and often reduction of the work force is not one of the objectives. But it is, by nature, labor saving, even when that is not a primary motivation for its adoption. This argues for a significant reduction in the labor necessary for handling a given volume of information. The statements made above are hypotheses or arguments often heard in discussions about whether office automation will lead to an increase or a decrease in office jobs. The purpose of this chapter is to examine these propositions in more detail. They provide a basic framework for discussing the long-range effects of office automation on future office employment. But there are several levels of complexity that must be considered. Many of these propositions involve, or conceal, definitional New data input technolo~y or logical difficulties. The interactions between them are not well understood. There are other factors and forces acting on white-collar employmentand intermediate steps in these relationships that are ignored in the simple diagram shown in figure 2-1. Estimates of the relative magnitude of these competing forces are judgment calls. Those who see an increasing demand for information producing an expanding need for informationhandlers, and those who anticipate new industries or new occupations to create jobs for displaced office workers can not, in the nature of things, offer a valid description of what those future products, industries, and jobs might be. They are therefore often accused of optimism based on ungrounded faith. Those who see disturbing signals of future job loss can point only to fragmentary and widely dispersed evidence that is difficult to compare or aggregate, and are equally subject to charges of ungrounded pessimism.
PAGE 45
36 l Automation of Americas Offices Employment data is notoriously difficult to work with because job categories and occupational titles are not standardized and change over time. Projection of trend lines is not very helpful in discussions of significant technological change, and arguments from analogy to automation in the past are not convincing because the surrounding social and economic environment has radically changed. These caveats do not mean that analysis is useless; they only warn of the degree of uncertainty necessarily involved. It is the task of policy analysts to advance conclusions in spite of incomplete evidence, as it is the task of policymakers to make decisions under conditions of uncertainty. What Follows Since OTAs conclusions about the possible employment effects of office automation are less optimistic than some recent long-range employment projections, these will be reviewed to point out forces and factors that are usually not adequately accommodated by econometric models and that strongly affected the conclusions of this assessment. These considerations have to do with the way in which technological change erodes the validity of conventional assumptions, disrupts long-term trend lines, and changes the meaning of established categories of occupations and industry boundaries. The conclusion that the number of office jobs may, between 1985 and 2000, tend to stabilize and possibly begin to decline, rests on the following points: Computers represent a fundamental technological change rather than a marginal improvement; their adoption will resemble that of the telephone and the typewriter rather than that of the industrial robot. Office automation will be pervasive. Recent technology/employment forecasts are flawed and probably understate the employment effects of office automation. (See Recent Labor Force Projections.) The potential productivity benefits of office automation are larger than are generally recognized; they have been and will be masked during a transition period because of technical and managerial problems, and institutional inertia in adapting to technological change. (See Productivity and the Nature of White Collar Work.) Office automation will reduce the need for labor by sharply reducing the need for both primary and secondary data entry; time saving in analytical, computational, and communication-related tasks; direct substitution for labor in many tasks; eliminating intermediate and preparatory steps or tasks; and transferring tasks from highly paid to lower paid employees (see How Office Automation Affects Work). Specialized occupations with narrowly defined tasks will be most directly and immediately affected (see Emerging Occupational Shifts ). Automation will eliminate or reduce the need for tasks characteristic of some generic clerical occupations (i.e., those common to most or all industry sectors). Growth of some categories of management occupations may be sharply constrained or even reversed because of time savings in information collection, analysis, and formatting; time savings in communications; increases in scope of supervisory attention and monitoring capability; and reduction in the number of workers to be supervised. There are already possible indicators of constrained employment growth from office automation, including declining unit labor costs in the heavily automated financial, insurance, and real estate (FIRE) industries (see Technology and Recent Trends in White Collar Employment). Office automation could also encourage conversion of employees to part-time, temporary, or independent contractor status, with a reduction in the number of full-time, permanent jobs (see Part-Time and Temporary Employment). Off-shore sourcing of clerical work, now small in volume, could increase significantly in the next decade (see chapter 8). Office automation may have more significant or more highly visible impacts than
PAGE 46
. Ch. 2Productivity and Employment l 3 7 earlier waves of mechanization and autoThe remainder of this chapter discusses mation in the United States because the these points in more detail, beginning with a economy will not be growing as rapidly, review of recent employment forecasts. It ends it will affect more occupations, and it will with a brief discussion of labor force growth, cut across industry sectors (see Analopossible adjustment mechanisms, and some gies From Past Waves of Automation). policy implications. There are, in many cases, some signs that technology has contributed to rising unemployment rates since the end of World WarII. RECENT LABOR FORCE PROJECTIONS The Bureau of Labor Statistics (BLS) expects employment in managerial, technical and professional, and clerical occupations to increase by about 28 percent by 1995. 3 Most job growth is projected to be in the service producing industries (transportation, communications, utilities, trade, finance, insurance, real estate, government, and other services), which BLS expects to account for 75 percent of all new jobs, or 18.7 million by 1995. Since these industries are the ones in which office automation could have major effects, the BLS reasoning should be examined closely. BLS expects that a quarter of all employment growth, 31 million jobs, through 1995 will be in the category of other services. Within this category the largest industry is business services, which includes a variety of things from business consultants to janitorial services, and is projected to grow by 5.3 percent per year, with employment growth of 3.9 percent per year. Professional services (legal, engineering, accounting, etc. ) are expected to add another 850,000 jobs. Financial and banking services are projected to have strong growth but modest gains of employment of 1.9 percent per year or 21 percent increase in 10 years. The BLS projection assumes: 1) full recovery from the 1982 recession and stable economic growth through 1995, 2) continuing deGeorge T. Silt estri, tJohn hl. I.ukasiewicz, and Marcus E. E:instein, Occupational Employment Projections Through 1995, Emplo~rment Rejections for 199,5, U.S. Department of [.abor, Rureau of I,abor Statistics, Bulletin 219, March 1984, p. 44. cline in the average weekly work hours per worker, 3) productivity growth at the rates of the late 1960s and early 1970s, and 4) some shifts in employment, namely that new laborsaving technologies will cause shifts to occur among industries, with many of the old-line factory jobs giving way to new industries and occupations Factory automation and office automation will displace some workers but they will have an opportunity to move into other jobs; what new industries and occupations will be created is not specified. BLS expects that jobs requiring college or specialized technical training will increase significantly, as will some less skilled jobs; for example, nursing aides and orderlies. The BLS analysts recognize that employment growth in many occupations will be affected by technological change and that among these are typists. Nevertheless, they expect many office occupations to grow, as shown in table 2-1. BLS projections show three alternative growth rates, which are based on alternative assumptions about the overall rate of economic growth. The BLS employment projection process links a labor force model, an aggregate macroeconomic model, an industry activity model, a labor demand model, and an occupational demand model. Under different assumptions Nlanufacturing employment. which has declined from 25 percent of all jobs in 1959 to under 19 percent, is projected to hold this share through 199,5. This means that manufacturing would have to create about 4,3 million additional j ohs, or one-sixth of all new jobs.
PAGE 47
38 l Automation of Americas Offices Table 2-1. Growth in Selected Occupations as Projected by the Bureau of Labor Statistics, 1982-95 Projected growth a Range of jobs added b Occupations Low Moderate Hig h (thousands) Total : all occupations . . . . . . . . . . . 23/0 Total: professional, technical, managerial, and clerical . . . . 26 All professional, technical, and related occupations . . . . . 30 Accountants, auditors . . . . . . . . . . . 38 Computer specialists . . . . . . . . . . . 79 Economists . . . . . . . . . . . . . 29 Managers, officials, and proprietors . . . . . . . . 26 All clerical workers . . . . . . . . . . . . 24 File clerks . . . . . . . . . . . . . O General clerks . . . . . . . . . . . . 7 Office machine operators . . . . . . . . . . 26 Computer operator personnel . . . . . . . . . 25 Data entry operators . . . . . . . . . . Secretaries, stenographers . . . . . . . . 24 Typists . . . . . . . . . . . . . . 15 Telephone operators. . . . . . . . . . . . 6 250/o 28 31 40 81 27 28 26 2 9 28 27 26 16 8 28% 31 35 44 84 30 31 29 5 12 30 30 29 19 10 23,336-28,392 11,921-14,165 4,961-5,741 325-373 414-439 9-9 2,476-2,935 4,484-5,489 21-34 642-765 243-284 147-172 644-787 146-185 19-31 aprojectlons were made for IOW, moderate, and high occupational growth scenanos with associated variation by industry bThe range represents the low tO high estimates SOURCE GeorgeT Sllvestri, JohnM Lukas!ewlcz, and MarcusE Einstein, OccupatlonalE mployment Projections Through 1995; Ernp/oymen/ Pro)ecfions for 1995, US Department of LaboC Bureau of Labor Statistics, Bulletln 2197, March 1984, p 44 about economic growth, industry sectors grow at differentiates. If an industry is expected to grow, occupations concentrated in that industry are projected to grow proportionately, with marginal adjustments made to accommodate anticipated technological changes. The tying of occupational growth to growth in specific industries accounts for some curious outcomes to be noted in table 2-1, which is abstracted from much larger tables in the BLS projection. For example, demand for economists increases less under the moderate growth scenario than under either the low or high growth scenarios, presumably because economists are concentrated in industries that would expand least in a period of moderate economic growth. Some occupations that declined from 1979 to 1982 are nevertheless projected to grow in the future, because the decline is attributed to the business recession and it is assumed that with recovery, past trends in occupational growth will be resumed. The model does not separate change due to business cycles from technological change. For example, drafters decreased by l.6 percent from 1979 to1982, but are projected to increase from 2 to 8 percent. Cost estimators declined 2 percent but growth of 41 to 48 percent is projected. Bookkeepers (hand) declined by 3.9 percent but 13 to 18 percent growth is projected. General clerks declined by l.2 percent but 27 to 33percent growth is projected. This mechanical linking of occupational growth to industry growth suggests that BLS may underestimate the effect of technological change (and office automation in particular) even though the text accompanying its projection acknowledges that: Most office clerical occupations are expected to grow more slowly during 1982-95 than in the 1970s because of office automation. Secretaries will increasingly use advanced office equipment in the future, thereby becoming more productive. This in turn will dampen demand for the occupation. Nevertheless, seeretaries are projected to grow at a rate that is about average because of the growth of industries in which they are concentrated. 5 Anticipated growth in demand for goods and services, generated through the BLS macroeconomic model, is translated into a projection of industry activity (e.g., growth rates) through the use of input/output analysis. As a commentary on the BLS methodology notes: (Input/output modeling) assumes, for example, fixed coefficient production; that is, Silvestri, Lukasiewicz, and Einstein, op. cit., p. 44. John A. Hansen, Bureau of Labor Statistics Methodolo~, for Occupational Forecasts, a contractor report for the Office of Technology Assessment, April 1984. p. 13.
PAGE 48
that there is a single production process that must be used in the production of each good. No substitutions are permitted between the various inputs to the production process (U)nless a new input/output table is constructed for each time period, technological change cannot be incorporated into analyses based on input/output tables. The use of input) output tables therefore naturally makes it difficult to incorporate technological change into occupational forecasting. BLS analysts are acutely aware of this problem. Adjustments to take into account technological change are made at several points in the BLS process. Technological change will effect the forecast output of the economy as a whole, and the allocation of final demand into the 156 categories used in the model. A technological development that affected the distribution of demand for transport of freight between the trucking and air transport industries would require changing coefficients at many points in the model, for example, those governing the relationships between demand for transportation, energy, and perhaps steel and rubber. Industry outputs must then be translated into demand for laborand number of jobs. The econometric model that does this is based on a constant elasticity of substitution production function, that is, a basic formula that interprets the substitutability of one input for another, for example, capital for labor. It includes a technology variable that governs the total output-to-capital ratio for the economy as a whole. This variable is given a different weight for each industry. When the outputto-capital ratio is changing significantly (as it has been in office-oriented industries) this weight must obviously be adjusted. Some adjustments are also made in translating labor demand into number of jobs, to accommodate trends in length of workweek and work year. Such adjustments are made, in the BLS procedures, by the analysts using their own knowledge of and assumptions about current and anticipated technological developments. They are, in other words, made on the basis of judgment and can only be evaluated in that light. Ch. 2Productivity and Employment l 3 9 However, the rationale for making these adjustments, and the precise changes that are made, are difficult for outsiders to track and examine. 7 BLS is, however, notably conservative in making such adjustments, relying heavily on past trends. In a recent General Accounting Office (GAO) publication, R BLS stated its assumptions for some growth occupations. They were very simple, brief, and generalized. For example, for accountants and auditors, the only technological assumption was: It was assumed that there will be no technological changes or developments that will affect employment in this occupation. For bookkeepers (hand), the relevant assumptions were as follows: Computerization has had a significant effect on employment in this occupation by slowing down the rate of growth. This trend will continue, causing the employment of bookkeepers to grow more slowly than average. Computerization will continue to develop, but its adverse effect will be offset by the rapid increase in the volume of business. Computerization will continue to spread between 1982 and 1995. There will be further evolution of labor-saving technologies and continued diffusion throughout the economy, resulting in higher productivity and a slowerIbid., See also imothy I,. I lunt and H, Allen Ilunt. of the W.PJ. Upjohn Institute for l;mplqment Research. .$n :\ssessment of Data Sources to Stud} the F; mplo~. merit l; ffects of Technological Change, a paper submitted to the National A cadem~ of Sciences Committee on Wrornen Flmplolment and Related Social 1 ssues, october 1984. They comment: Technological change actual]> enters into the (f31.S) sy<(tln) in at least three plactls [rl)tcl]nolo~ical change will ha~e speclflc effects on some occupations. it will ha~re an overall impact on t hr product i~it} of workers, and it will affect the demand for good\ anci ser~]ces generally I t IS worthy of note th~it t h]s system in~ {)l~,t~s a con>ld[,rat)l~, amount of judgment, cspeclall~ in anticipating the effects of t ec~v rmlogxal change. There are no sm~ple equations that pr(diet changt~< in staffing pat tt,rns w it hin an lndustr}. I n fa{t, t he 111 ,S ~t tiff }~;]~ found that trends ]n industr} ~ln~plo~n]{,nt It (,1~ (an I)(I pr(dict<)(i more accuratel~ than the changes in occup; it ion.il (In)pl(J~ mt, nt This 1~ due in large part to the difficult} of pro it[t]ng spwiflc (Ncupat ional impacts of tec.hnolo~i[al chan~~, ~l,s, ( ;~~ncral :~~c~unt~ng office, specific 7ec.hnolo@(>:d ..\.ssumptions A fftwtin~~ the l{ureau of I.abor Statistics 1995 b;m pio~ment lroje(tions, Report tot he Hon. Flerkle~ Mdell, (1, S. Ilouse of Represent ati\es, (;, A() OCE-85-2, hla~ 20, 1985.
PAGE 49
4~ Automation of Americas offices than-average rate of growth for this occupation, The occupational forecast is made by disaggregating the labor demand further; that is, the 156 industry sectors used in the other models is spread apart into 360 industries and the total number of jobs is allocated between them. For each of the 360 industries, BLS considers the historical staffing patternobtained from the Occupational Employment Survey and carried out by State Departments of Employment Security. g The jobs that it is assumed the industry will provide are distributed among the occupations according to the ratios that have been obtained in the past, with some qualitative adjustment. In preparing the occupational projections for 1995, the BLS procedure assumes that all workers will be in occupations that not only exist now, but have existed long enough that employers begin listing them as separate categories. 10 BLS uses several kinds of sensitivity analyses in preparing its forecasts; the analysts lay out three different scenarios or trend lines for economic growth, and also make varying assumptions about labor productivity. These, however, would not capture fundamental changes in production processes and capitalto-output ratios. No sensitivity analyses are performed to test the effect on staffing patterns of technological change. In summary, the strong link in BLS forecast methodology between industry growth and projected occupational growth, together with reliance on traditional industry staffing patterns and occupational distributions, tends to greatly underestimate the effects of fundamental technological change. This forecast cannot be assumed to capture reliably the potential effects of office automation. The Leontief-Duchin Forecast, The Impacts of Automation on Employment, 1963-2000, The survey covers 500,000 businesses, one-third of which are covered each year, and 1,500 occupations found within those businesses, Hansen, op. cit., p. 18. Prepared by the Institute for Economic Analysis, New York University, for the National Science Foundation, April 1984, Wassily I.eontief and Faye Duchin, Principal Investifzators. also used input-output analysis. To deal with the problem of incorporating the effects of technological change into the model, however, Leontief and Duchin allowed the coefficients in the matrix to change over time, adjusting these coefficients on the basis of qualitative judgments. Four scenarios were developed. Scenario 1 assumed no further automation or any other technological change after 1980; it is clearly a highly unlikely scenario and is used as a reference or benchmark for the others. Scenarios 2 and 3 project an increasing use of computers in all sectors for specific information processing and machine control tasks and their integration. Of these two, Scenario 3 assumes faster technological progress and more rapid adoption of available technologies, including more powerful software. Scenario 4 is identical to Scenario 3 except that it uses different final demand assumptions. The study concluded that the intensive use of automation: will make it possible to achieve over the next 20 years significant economies in labor relative to the production of the same bills of goods with the mix of technologies currently in use. Over 11 million fewer workers are required in 1990 and over 20 million fewer in 2000, under Scenario 3 as compared to Scenario 1: this represents a saving of 8.5 percent and 11.7 percent respectively of the reference scenario labor requirements. 12 (Emphasis added. ) The differences by 1990 between Scenario 1 (no further automation) and Scenario 3 for major categories of workers are: 13 l 5.5 percent more professionals, l 13.9 percent fewer managers, 32.4 percent fewer clerical workers, and l 8.4 percent fewer jobs in all categories in the national economy. .. I,eontief and Duchin, op. cit., p. 1.15. There is a slight discrepancy between the percentages given in the authors text quoted here and those given below and computed from their tables (p. 1.16 of the report cited); the difference is minor and perhaps was caused by rounding errors. This includes all private sector employment plus public education and health, but does not include public administration, the armed forces, or household employees.
PAGE 50
By 2000, the changes are greater: l 21.5 percent more professionals, 41 percent fewer managers, 45 percent fewer clerical workers, and l 11.4 percent fewer jobs in all categories in the national economy. Note, however, that fewer here means only fewer than there would be without any further automationnot fewer workers than there are now. As compared to the real figures in 1978 (the base year), total employment will grow by 49 percent under Scenario 3 as compared to 53 percent growth under Scenario l no further automation. By 2000, it will increase 76 percent over 1978 employment in Scenario 3, as compared to 98 percent increase in Scenario 1. Under Scenario 3, the proportion of professional jobs to all jobs rises from 15.6 percent in 1978 to 19.8 percent in 2000; managers jobs decline from 9.5 to 7.2 percent, and clerical jobs from 17.8 to 11.4 percent. Sales workers decline slightly relative to total employment, while the proportion of craftsmen, operatives, service workers, laborers, and farmers each increases. The increases in professional jobs, under Scenario 3, are mainly in demand for computer specialists and engineers. The good news in the Leontief-Duchin forecast is that if their projections were correct, the labor requirements for Scenario 3, in 2000, would exceed the projected available labor force of 157.4 million (after allowance is made for public administration, household workers, and some multiple job holders). This means, however, that the rate of growth in final demand that they assumed could not be achieved without still greater technological change. This problem reveals a serious weakness in the model; it ignores the importing of both capital goods and services. Since imports are a major, and growing factor in the economy, the Leontief-Duchin model does not reflect economic reality. Leontief and Duchin developed a fourth scenario to assess the growth in demand that could actually be attained under their technoCh. 2Productivity and Employment l 4 1 logical assumptions. Growth in demand was progressively reduced until the labor force needed fell within the range of the projected labor force available for the years 1990 and 2000, This scenario does not correct the defect of ignoring imports; all demand is again assumed to be supplied by domestic labor. This required a reduction of their projected output demand of 4.4 percent for 1990 and 17.8 percent in 2000, Because overall economic activity is lower in Scenario 4, capital investment is also lower, and the number of jobs related to production of capital goods falls, especially craftsmens jobs. The occupational composition of the work force otherwise remains much the same. The percentage reduction in demand for labor is, however, greater than the reduction in final demand. Scenario 4 therefore represents an estimate of the extent to which real per capita consumption can increase if the entire projected labor force is employed, using computer-based technologies, and demand is met by domestic production. It is important to note that the Leontief-Duchin team did not generate its own projection of consumption (final demand, or delivery of goods and services), but used that generated by BLS for its moderate growth scenario, discussed above. Leontief and Duchin, as noted, incorporated the potential effects of technology into their forecast by changing the input-output coefficients over time. These changes are exogenous to the model; the direction and magnitude of the changes were arrived at by the team of analysts on the basis of extensive review of research results and scholarly and trade literature, and their assumptions are set out within their project report. The factors explicitly taken into consideration were estimates of the rate and extent of capital investment, the amount of time in 1977 that a worker spends performing particular tasks, the amount of automatic equipment that will be applied to those tasks, the amount of time that can be saved per task by using automated equipment, the percent of workers in particular occupations and industry sectors that will use the
PAGE 51
42 l Automation of America's offices new technology in a given year, and the possible increase in demand for certain office occupations (i.e., the effects of increased demand for information-handling). Coefficients were modified accordingly; for example, in Scenario 3, the labor coefficient for stenographers, typists, and secretaries for 1990 was set at 0.65 of the 1977 coefficient, and for 2000 it was set at 0.45 of the 1977 coefficient, meaning that for a given unit of work done by these workers in 1977, only 65 percent as much labor will be needed in 1990 and only 45 percent as much by the year 2000. To look more closely at the narrower category secretaries, the team assumed that by the year in comparison to 1977: l l l l l l 30 percent of secretaries will not be affected by word processing, 20 percent of a secretarys time is spent in tasks not affected by word processing, word processing saves 80 percent of the time required for conventional typing, 35 percent of secretaries are also affected by other office technologies, 45 percent of their time is affected by other office technologies, and 75 percent of secretaries time is saved by new technology relative to old technology. While other analysts may quarrel with one or all of these estimates or assumptions, and may question whether they have foreseen likely technological breakthroughs, their assumptions in general do not appear to be overly conservative. There is no obviously better way available as yet to arrive at such estimates than the one the project usedexpert judgment. The far larger problem has to do with the assumptions about future economic growth. This brings one back to a broader version of the question proposed at the beginning of this chapter: will the demand for information (which is to a large extent a function of the level of economic activity in this country) really grow to an extent that it will more than balance the labor-saving effects of information technology? Both the BLS and Leontief-Duchin forecasts of employment growth depend on an assumption that it will do so. The model formulated by Leontief and Duchin is dynamic in that investment is a function of changes in output in industry sectors. They project significant gains in employment in most sectors because the model projects nearly unlimited increases in the production of capital goods (output). But the model ignores the strong trend toward import of capital goods (and of services); in reality, the income generated by production of capital goods outside of this country does not directly translate into increased consumer demand within this country and can translate instead into a loss of employment. Leontief and Duchin, recognizing this uncertainty, conclude that it is not yet possible to pass a final verdict on the question of technological unemployment by the year 2000. Another reason is, as they acknowledge, that to do so they would have to incorporate into their forecast other kinds of technological change, which will change the nature and level of final demand for goods and services. Leontief and Duchin also postulate on the basis of their model that labor requirements to satisfy a continually but moderately increasing standard of living will number 124 million jobs in 1990 with the required occupational composition reflecting the technologies that will be in place but because of very slow change in the orientation of education, training, guidance, and so on, these individuals skills and occupational expectations will reflect the mix of jobs that corresponded to the technologies that were in place in 1978 Under those conditions, they point out, 744,000 managers and over 5 million clerical workers could be potentially unemployed while there would be about the same number of unfilled jobs in other occupational categories, for which they lacked the necessary skills. The problem for public policy, in other words, could be very serious even if demand for labor is equal to or greater than the supply. A major difficulty with this forecast is doubt about the assumption of great and increasing demand for information-handling, which is ulti-
PAGE 52
Ch. 2Productivity and Employment l 4 3 mately derived from the assumption of steadily increasing production of capital goods. A Georgia Institute of Technology Forecast prepared for the U.S. Department of Labor 14 dealt with clerical employment in the banking and insurance industries only. It concluded that by the year 2000 there would be an absolute reduction in clerical employment of 22 percent (over 1980 levels) in the insurance industry and 10 percent in the banking industry, which together employed in 1980 more than 20 percent of all clerical workers (the clerical employees in these two industries alone constitute nearly 4 percent of the entire U.S. work force). The Georgia Tech research began from the premise that there are weaknesses in techniques used to incorporate technological change in employment forecasts, 5 including the unsystematic and qualitative consideration of emerging technological capabilities, the use of existing occupational descriptions based on fixed technologies, the inability to generate estimates of how particular technologies change the amount of time spent on basic tasks, and the use of an overly short time horizon (usually 10 years). The study began with a technological forecast of office automation hardware and software, organized in terms of information processing functions (descriptions of the content of tasks related to information processing in the two industries). The forecast was designed to identify breakthrough technologies that would have major consequences for clerical employment. Industry officials participated in a Delphi (a forecasting technique that generates by consensus opinion quantified estimates of specific variables). This produced assumptions about different paths that penetration of office automation technology might take in each industry and the different employment/occupational mixes that might result. The analysts Georgia Institute of lechnolo~r, Impact of office A uto mation orI office II orkers, final repor~, April 1984, prepared for the ~~mplo~ment and Training .4dministration, U.S. Department of I.abor; J. I)avid Roessner, Project Director. (;eorgia Institute of Technolog~, op. cit., Iol. 1 I, Technical Report p. 2. then used the technological assumptions to estimate the reduction in clerical time needed to perform specific functions in 1985, 1990, 1995, and 2000. Tying these estimates to 1980 clerical employment, an analytical model calculated the changes in the task/function work distribution over the period 1985-2000 for specific clusters of clerical jobs in each industry. For sensitivity analysis and validation, the results generated by the model were compared with those generated by the industry Delphi to determine the sensitivity of results to different estimates of the extent to which and rapidity with which specific technological changes will affect particular clerical functions. The Georgia Tech team concluded that breakthrough technologies have the potential to displace or otherwise reshape clerical employment in at least two functional areas-data input and data processing. These include technologies such as optical scan, speech recognition, software languages and programs, and artificial intelligence, most of which the OTA Assessment also identifies as critical elements in the employment outlook. There were several strengths in the Georgia Tech approach which lent credibility to the results. First, it went beyond conventional occupational categories to consider what it is that clerical workers actually do, in these two industries, in their daily and hourly work. Secondly, it attempted to estimate the relative amounts of time spent in these tasks, and to relate this to the time-saving potential of existing and emerging office technologies. Thirdly, it considered a number of potential organizational adjustments to the changing nature of the tasks in terms of organizational structural patterns. Finally, this approach permitted examination of future clerical work at several disaggregate levels, e.g., functions, tasks, and job clusters, rather than in terms of formal occupational categories. There are also a number of weaknesses in the approach. The analytical model lacked feedback loopsfor example, it did not account for the way in which the level of clerical employment affects the number of clerical supervisors needed, and did not account for rela-
PAGE 53
44 l Automation of America Offices tionships between clerical and nonclerical work. It did not consider creation of entirely new jobs and had little to say about the conversion of professional work to clerical tasks. It did not consider the possibility of significant restructuring of the industries themselves, and their products and services. More importantly, the forecast does not explicitly deal with growth in workload, except through very broadly stated, arbitrary macroeconomic assumptions, i.e., growth within the two industries of about 3 percent annually. There was not explicit consideration of expansion of the role of information within the two industries, which are of course already highly information intensive. Finally, the methodology used depends on judgmental dataestimates by industry experts and team analysts-even though it is treated quantitatively. In this of course, all employment forecasts are alike insofar as they go beyond mechanical extrapolation of timeseries data. To do otherwise would defeat the purpose of anticipating change brought about by technological or economic trends. All of the above forecasts thus are highly dependent on the assumptions made about the increasing volume of information handling (independently of the credibility of the models used) and the continuation of long established trends in the growth of occupations and their association with specific industries. The BLS projection anticipates increases in office employment in the neighborhood of 25 percent over the next decade but is probably much too conservative in its attention to technological change. The Leontief-Duchin forecast points to continued growth, although highly constrained by the effects of automation, but bases this forecast on flawed economic assumptions. The Georgia Tech forecast, which points to a decline at least in clerical employment, gives more attention to likely technological breakthroughs (in this field more credible than The most important of these were (op. cit., p. 18): 1) periods of growth and recession, but no major depression; 2) modest overall economic growth averaging 3 percent annually; and 3) growth in insurance and banking industries paralleling growth in the general economy. an expectation of smooth technological evolution) but is narrowly limited to two industries. Agreement or disagreement with these forecasts does not turn primarily on the models and methodology used, but on the complex assumptions, estimates, and judgment that generated the numbers fed into those models. All such models address the question of what will happen if. .; if for example, technology changes in certain ways, is adopted at certain expected rates, produces in practice the productivity benefits that in theory should result, and leads to the changes in organizational behavior that can be rationally anticipated. In spite of the fact that vendors and advertisers are often accused of overstating the benefits to be gained by automation, forecasts such as these may underestimate them. Technology and employment forecasts often misfire in this way, for two reasons. First, forecasters have no reliable way of estimating how pervasive a technological change will bethe ultimate rate of penetration. Some technologies are likely to be widely adopted only in a few industry sectors or in certain parts of a given industry, or only by organizations in a certain size range. Other technologies affect a broad range of economic and social sectors or become so pervasive that they fundamentally change social and economic activities for example, the internal combustion engine and electric communications (the telephone and telegraph). Office automation, in particular, computers, are in the latter and more fundamental category of technology; and likely to become as pervasive in office activities as the telephone and the typewriter have become in the last few decades. Secondly, technology forecasts often go wrong either because they assume technological or scientific breakthroughs that fail to come about, or because they assume a smooth evolutionary development of technology rather than breakthroughs that suddenly occur. In computer technology, the latter mistake is more common. Recent employment forecasts appear to assume that computer-based technologies of the next 5 to 15 years will not be signifi-
PAGE 54
Ch. 2Productivity and Employment l 4 5 cantly better than or different from today s. On the contrary, there is no reason to believe that their rapid development has come to an end. For example, input technologies, specifically optical character recognition and speech recognition, are likely to have a significant impact within 15 to 20 years. The rapid development and spread of end-use computing was not anticipated in many technology forecasts of the 1960s and early 1970s. In the field of microelectronics, it appears safer to anticipate quantum leaps in capability than to ignore the probability of their occurring. The chief value in these and other such forecasts is that they force attention to the many and complex uncertainties that will affect future levels of white-collar employment. OTAs analysis differs from those above primarily in suggesting that there are impending technological developments that will have a particularly significant effect on some large categories of office employment; that adoption of office automation will proceed at an increasing rate and be much more widespread than most other kinds of technological change have been; that productivity gains have been and will be for a few years masked by transition problems but will soon become apparent; that economic growth rates could be lower than they have been for most of our history; and that competitive pressures will lead organizations to take full advantage of the labor-saving characteristics of the technology. These conclusions, like those of the forecasts described above, are certainly arguable but merit prudent consideration. PRODUCTIVITY AND THE NATURE OF WHITE-COLLAR WORK Productivity is defined in economic terms as a ratio between quantity of output and quantity of input; increased productivity is achieved by producing the same amount of product or service with fewer resources (technological change), or producing more product or service with the same resources (resource reallocation). Sometimes organizations automate their office work in order to do the same amount of work with less labor cost (which could mean either fewer workers, or the same number of workers doing less skilled work for lower pay). Sometimes organizations foresee or seek an expanding workload and hope to accomplish it without increasing their work force (or their total expenditure for labor). Applied to white-collar work, however, the concept of productivity becomes complicated and troublesome. It is least complicated in the context of routine processing of large amounts of standardized, easily quantified datanumber of units sold, dollars paid or received, ticket stubs returned, keystrokes made. It is most complicated when applied to work where quality is more important than quantity-analysis, decisions, staff support, and policy formulation. If the final output of the organizations activity is an information product or service (an advertisement, a document or research report, a legal brief, an advisory memorandum or set of guidelines), then the number of units processed per hour is often less important than the quality, the timeliness, the fit to a client needs, or the degree of customer satisfaction. Effectiveness is more important than efficiency. But effectiveness is much more difficult to measure, since it refers to the characteristics of the output rather than to a ratio between countable units of input and output. If the organizations final product is a material good such as automobiles, the contribution of office work to overall productivity is difficult to determine because much of it is concerned with the coordination of intermediate steps in the conversion of resources into products. Both effectiveness and efficiency are important and impossible to separate.
PAGE 55
46 Automation of Americas offices There are other general problems in discussing the productivity of service industries. Service outputs are different from material products. They can usually not be stockpiled. Customer behavior intrinsically affects the delivery mechanism. For example, a study of increased productivity in British accounting firms found that most of the productivity gains came not so much from anything the accountants did but rather from the clients computerization of their own accounts, which allowed the accountant to save vast numbers of person-hours in auditing. 17 In fact, the concepts of input and output both become blurred in many kinds of whitecollar work. Hours logged in or hours paid for are often not the same as hours worked. This is true in both blue-collar and white-collar work, but perhaps most obvious for professional and managerial workers who may be processing information or formulating decisions while they read the newspaper or fall asleep at night. The quality of work is often more important than the quantity of work for both blueand white-collar workers, but for white-collar workers quality is more difficult to measure in terms of error rates. The skill that a receptionist uses in soothing irate clients, or that a secretary uses in locating an elusive file, shows up in the companys records only indirectly as an addition to a salespersons accounts or a lawyers clients. Contributions to corporate reputation or employee morale are difficult to measure. In terms of outputs, a good decision or the elimination of erroneous information from a data bank may not be counted, or even be possible to identify as a discrete output. There is a tendency to use worker activity, an input, as surrogate for an output in measuring productivity, with the unwarranted assumption that more is better. Thus, computers and word processors sometimes lead to a proliferation of reports that may or may not be useful and may or may not represent increased productivity. Irving H. Siegal, Productivity Measurement: An Evolving Art, Work in America Institute Studies in Productivity, No. 16 (New York: 1980). It is therefore difficult to measure, to define, and even to discuss the amount of increased productivity that could be gained by office automation even if it is widely adopted by organizations of all kinds and all sizes. Some organizations have been slow to adopt office automation for this reason. However, an organization will usually not decide to automate a specific task or set of tasks if it expects that over the long run the task will therefore require more labor, or will take longer to perform per unit of workload than without automation. Automation usually implies some transference of work from humans to machines, and thus a substitution of capital for labor. If office automation only allowed those organizations that automate (assumed for the present to be more efficient) to take over markets previously served by nonautomating organizations, the result would be a reduction in the total amount of labor used. On the other hand, if office automation only stimulated the creation and sale of products and services that could not previously be offeredthat is, created new markets-the total amount of output and labor used would increase. But office automation will both create new information markets and reduce the amount of labor required for existing and new products and services. The difference-net labor demand-could be either positive or negative. The factor of time is crucially important. For some period of transition, longer for some organizations than for others, there is little or no gain in productivity and often some loss. On the other hand, the new technologies are so powerful that some organizations settle for the short-run labor savings and limited cost reductions that are possible from limited use of automated devices, rather than run the risk of temporarily disruptive restructuring of the work to systematically capture all the potential benefits. Across the economy as a whole, with sectors and industries and organizations at all stages of this transition, it is difficult to anticipate how long this transition will take. Over timemeasured in decadesmore efficient organizations should tend to replace less
PAGE 56
. Ch. 2Productivity and Ernp/oyment l 4 7 efficient ones, and newly created organizations economy, be automated in the next 15 years? should tend to be automated, and more effiWill the demand for information processing cient, from the beginning. create more than enough new jobs to compenThis is the dynamic shift which needs evalusate for the labor reduction effects of office ation in terms of the net labor demand. How automation? thoroughly will office work, throughout the HOW OFFICE AUTOMATION AFFECTS EMPLOYMENT Office automation can substitute for labor, supplement labor, or reorganize work and thereby make labor more efficient. It can allow highly technical, knowledge-intensive work to be done by relatively untrained and unskilled, lower paid workers. (Skill enhancing and job enhancing features of the technology are also important, but are discussed elsewhere. ) It can change the characteristics and skills associated with occupations and alter their role and relative importance to an industry. It can allow office work to be done away from the office and outside of conventional office hours, even outside of the country. The most dramatic potential substitution of technology for labor, in the future, could be the elimination of a large proportion of todays data-entry work (either numeric or text) interorganizational transfer of data, directly from computer to computer; direct input of data by optical scanning technologies, and possibly by speech recognition technology; and capture of data at the point of origin, in a variety of ways ranging from bar code readers to consumer use of terminals, e.g., bank automated teller machines (ATM s). In short, second-time keyboarding of the same data may eventually disappear almost completely (i.e., once data exists in machine readable form it will be endlessly changeable and exchangeable). First-time keyboarding may also be sharply reduced. This may not happen within the next decade, and yet optical scanning technology and computer exchange of data is already reducing data entry in some areas, and all of these trends could develop very quickly. I I I L Cred/f Communications Workers of A~erlca We ve decided to call it the neutron chip. It eliminates jobs but /eaves the work-p/ace intact. Direct substitution for labor is only one possibility. Automation reduces the time required for many tasks. The measured time-saving for specific tasks, in numerous pilot projects and implementation case studies, varies from minutes to days. There is no way to validate, compare, and aggregate such figures across organizations. The estimate of percent time saving for complex procedures and task clusters recurs with great frequency in both private and public sector office automation plans, cost justifications, and cost-benefit studies surveyed by OTA, but again it is difficult to document or generalize this measure. The most that can be concluded is that time-saving is real but its magnitude and overall effect on productivity cannot yet be stated. s App. 13 summarizes se~reral case studies done for this OTA assessment; others from scholarl~ and trade literature are referenced throughout the report,.
PAGE 57
48 Automation of America Offices Automation also can reduce the time that people have unavoidably spent between productive tasks, or in sequencing tasks. One example is the way voice mail and electronic mail can reduce the time spent in communicating by eliminating telephone tag or frequent callbacks, searching for telephone numbers, and getting wrong numbers or busy signals. Electronic filing can eliminate walking to a file cabinet and searching for a particular file; word processing eliminates cutting and pasting or retyping documents to eliminate errors. In some applications a long series of tasks that required several workers is combined into one task. For example, in mail processing, a many-step function that involved several clerks (mail openers, sorters, typists) can be redesigned to require only a mail opening machine and a clerk entering information into a computer terminal. 19 Professional case workers in a New York City Municipal agency formerly sorted mail from clients and selected and typed one of four form letters in response. Now lower paid clerical workers call up the client file and push one computer key to send the appropriate form. 20 These are examples of process change, or reorganization of work and workflow to take advantage of office automation. Some insurance companies have been phasing out jobs that involve responding by individual letters to mail inquiries about claims, policy changes, or policy applications. Policy holders or applicants have a toll free number to call. The employees who answer such calls use terminals to retrieve the information required to answer most queries. By telephone the employee can elicit all the necessary information, some of which is often absent from mail communication, thus requiring more than one exchange of letters. The cost per inquiry is reported to be one-third the cost per mail inquiry. 21 As the conceptual model described in chapter 1 suggests, even if an organization adopts office automation to reduce unit labor costs in standard tasks, managers will subsequently recognize that it also has other capabilities, and will adapt the workflow or production process to take better advantage of those possibilities. This may take considerable time, since problems arising from the initial substitution and from the interface of new technology with old procedures usually get first attention. Initial transaction costs of adapting the production process to new technology are high. Training and system support costs in particular often outweigh hardware and software expenditures. The risks associated with trying new ways to get the work done appear very large and potentially disruptive to many firms. For this reason the full effects on employment are likely to appear only after considerable time and experience. Office automation can also complement and augment labor, making new tasks possible. Computers made it feasible for insurance companies to move from annual billings to quarterly or monthly billings (thus encouraging lower wage earners to buy insurance), and the increased availability of information also led States to impose increased reporting require ments on the industry. In other financial service industries, automation allowed the creation of new consumer services that would not have been cost-effective otherwise. In large law firms, some evidence suggests that the introduction of word processing resulted in support staff employment growing several times faster than legal staff. 22 Some people suggest that once word processing is substituted for typing, a great deal of hidden work appears-work that there was no time to get done before. This is one aspect that leads to an increase in the demand for information processing, and often therefore to a net increase in demand for labor. Matthew P. Drennan, Implications of Computer and Communications Technology for Less Skilled Service Employment Opportunities, Final Report to the U.S. Department of Mary C. Murphree, Brave New Offices: The Changing I.abor, grant No. USDL 21-36-31, Jan. 21, 1983, p. 2. World of the Legal Secretary, Y ~romen To~ls and Triumphs Wee ch. 10. at the Workplace, Karen Sacks and Dorothy Remy (eds. ) (New Matthew P. Drennan, op. cit., p. 64. 13runswick, NJ: Rutgers University Press, 1984).
PAGE 58
The net effect of other, equally important, effects of office automation on work and workflow is even more difficult to assess. Chapter 4 considers these changes in considerable detail. Automation clearly leads to reallocation of tasks, for example, professionals and managers drafting their own letters and documents on the word processor. Analytical work that formerly required specialists with long training can be incorporated in software that can be used by people without extensive professional training. Reason suggests that in the interest of cost-saving organizations will tend to professionalize clerical worki.e., push tasks downward in the hierarchy-and reduce the number of middle-level professionals and managers rather than the number of clerical workers, since the former are paid more. This is the argument for the so-called disappearing middle management or flattened hierarchy phenomenon that some experts anticipate. 23 The clericalization of professional work appeals to many professionals, because it gives them more autonomy and more control over the quality and pacing of their work; they can alter and revise as they go, they need not queue up or compete with others for the typists time. There is some evidence from case studies and general observation that the ratio of support staff to professional staff is tending to decrease in many officesthat is, the pattern of one secretary to one boss has already commonly hlan~ office automation experts and industry planners expect that \arious computer applications will allow managers to extend their scope of supervision and planning, and will thereb~. allow organizations to reduce the number of super\ isory managers required, and flatten the management hierarch~. This thesis has sometimes been addressed, in popular literature, in terms of a related issue, that of the disappearing middle (income class ). Saskia Sassen-Koob and others have put forward an argument that there is, or will be, a growth in emplo~ment in both highand low-income categories in the fastest g-rowing industries {ad\ anced ser~ices), with a decline in middle-income categories, and thus income polarization. See Saskia Sassen-Koob, The New I.abor Demand in I.ocal Cities, Cities in Transformation, Michael Smith (cd. ) (E3e\erly Hills, CA: Sage Publications, 1984), pp. 137-171. The income statistics put forward by Sassen-Koob as e~idence that the disappearing middle is occurring are effecti~el~ challenged b? Neal Ii. Rosenthal, The Shrinking hliddle Class: hlyth or Realit~? Alonthl~ I.abor I/e\iew, Nlarch 1985, pp. 3-10, However the possibilit~~ of a reduced management hierarch~r need not stand or fall on the terms of this controyersjr at the macroeconomic le}el, Ch. changed to one 2Productivity and Employment l 4 9 secretary serving three to five professionals, or a cluster of support staff working for a large group of professionals and managers. Office automation may result in a more varied occupational structure, because of the diverse choices among applications and because automation allows production process and staffing to vary widely with the type of product. For example, three staffing patterns have been observed among insurance carriers. In highly automated personal lines underwriting departments of large property/casualty carriers, rating and risk assessment may be computerized and the functions formerly divided among many clerks, raters, and underwriting assistants, may be consolidated and delegated to highly skilled clericals. Unskilled clerical work is then largely eliminated. Underwriters become exceptions handlers, doing more complex work, and their work may shift to planning and marketing. Skilled clericals become the bulk of the work force. In other insurance firms, where products are standardized and high volume, almost all of the semiskilled tasks are computerized and a bipolar work force is created, with a large number of routine-data entry clerks and a few highly skilled, professional exceptions handlers. In still other cases where product lines are low in volume, specialized or complex, there is likely to be less extensive, more discrete automation. Data entry will be done by operators using dumb terminals, and underwriters review all policy. Thus, a view of office automations effects on employment must take into account not only substitution of capital equipment for labor in specific tasks, and the expansion of workload, but many additional and intervening variables including (see figure 2-2) recombination of tasks, reassignment of tasks across jobs and occupations, changing skills -Barbara Baran, Technology Innovation and Deregulation: The Transformation of the Labor Process in the Insurance Industry, Berkeley Roundtable on International Economy, contract No. 433-3610.0, prepared for Technology and ~~conomic Transition Project, Office of Technology Assessment, Januar~ 1985.
PAGE 59
50 Automation of America offices .Figure 2.Framework for Analyzing LongRange Effects of Office Automation on WhiteCollar Employment Iemployment and unemployment Ievelsl SOURCE Office of Technology Assessment requirements and corresponding job definitions, changes in the definition of occupations, and shifts in the number of jobs per occupation. The way in which these changes occur is further discussed in chapter 4. Emerging Occupational Shifts Other things being equal, the introduction of labor-saving technology is most likely to cause displacement where the task that is auto mated has constituted all, or nearly all, of the responsibilities of a given job. In other words, the more narrow and specialized the job, the more likely the job-holder is to be displaced by automation. This applies, at least potentially, to professional specialists as well as to clerical workers. Highly specialized knowledge is potentially most appropriate for incorporation into an expert system (a special software for decisionmaking) while broad, general knowledge is difficult to incorporate. 25 When word processing is introduced, assuming that the workload does not increase, fewer dedicated typists (keyboarders) will be needed. A general secretary who spends only a part of her time typing, is not likely to be displaced by a word processor; more likely she will have more time for other responsibilities and may take on new ones. Secretarial positions have been increasing throughout the two decades of office automation, while typist jobs are decreasing. -Science, Artificial Experts, Mar. 23, 1984, p. 1281.
PAGE 60
At the most general level, table 2-2 shows the relative shares of major occupational categories in the economy. Clerical workers are the largest employment category, followed closely by professional and technical. 26 In the first stage of automation (large computers), the tendency was to make affected jobs more narrowly definedin other words, to rationalize work. Batch data processors did not learn or do other work. Word processors were set apart in word processing centers. New specialties were created, ranging from computer operators to programmers, and the holders of those jobs typically did nothing else. If dataentry work is completely automated (e.g., by optical scanning technology) those who do only data entry are most likely to be displaced and secretaries are unlikely to be displaced by that development. As discussed in chapter 4, second phase automationend-user computingappears less likely to rationalize tasks or to narrow A major problem in assessing occupational change is a set of difficulties and deficiencies encountered in working with occupational data as it is now collected and aggregated. Occupational data are deri~ed from sur~eys, and the more specific and narrow the occupational category used, ~he smaller the sur~e~ sample is, and therefore the less reliable the estimate of the total. A more important problem is the lack of consistent timeseries data because of frequent changes in occupational classifications, especially between the 1970 and 1980 censuses. hluch of the data used in this chapter is drawn from the occupational industry employment matrix prepared by the Bureau of I.abor Statistics from occupational ~~mployment Surve~ data, modified by 131.S use of a statistical model. Both the survey and the model have some methodological problems; for example, there are \ariations o}er time in the way that sur~ey data was collected, and matrix data for 1980 and 1982 are not fully comparable. For more information on emplo~ment statistics consult the 111.S Iiandbook of Afethods, vol. 1, Bulletin 2134-1, December 1982. Ch. 2Productivity and Employment l 51 jobs. Personal computers can be used to integrate tasks and broaden jobs. Moreover, organizations that rationalized work during their early office automation are in some cases using further automation to reverse that process. Many firms have decided that computer and communication technologies are often most effective in reducing costs when control, communication, and decisionmaking are decentralized and when hierarchic organization and functional specialization of tasks are reduced. They are experimenting with the elimination of both low-skill clerical jobs and routine technical/professional jobs, and with the creation of new multiactivity, skilled clerical positions. 27 In some insurance firms, the result of task reintegration has been a significant reduction in unit labor requirements and an increase in the average skill levels of the remaining clerical, sales, and professional work force. 28 It is important to note that reduction of unit labor requirements has been achieved in both work rationalization and work integration, with office automation being used for both, The networking of computers is likely to eliminate some jobs that have until now provided the link between automated tasksfor example, those concerned with reorganizing -In the insurance industry, management first tended to follow the logic of scientific factory management, rationalizing and fragmenting tasks. Some insurance firms are now using integrated systems to reintegrate tasks, iillowing one person to handle multiple service transactions so that the individual master record for each policy is a complete database. -Eileen Appelbaum, Technolog~ and the Redesign of Mork in the Insurance Industr~, Stanford Uni\ersitJ, I nstitute for Research on Educational Finance and Go\ernance, Project Report No. 84-A22, No\rember 1984, p, 10, Table 2-2. Percent of Total Work Force in Occupational Groups by Selected Industry Sectors, 1982 Industry Sector All All Health Occupations industry manufacture ng Fir e a Services TCU b services Trade Professional, technical, and related occupations . . 16.39\0 10.270/o 9.46\o 33.65 0/0 7,86 0/, 34.47 0/0 3,78 Managers. officials, and proprietors . 8.37 6.69 17.06 6.82 8.99 4.38 9.34 Clerical ... . . ... 20,36 11,75 63,40 19,11 33.29 17,02 20.68 Total . ... 45,12 28.71 89.92 59.58 50.14 55.87 33.80 a FIRE F!nanclal Insurance real estate b TCU Transportation commuqlcatlons ut(lltles SOURCE U S Department of Labor Bureau of Labor Statlstlcs Ernp/oymenf and Earn/ngs, VOI 30, No 1 January 1983
PAGE 61
52 Automation of Americas Oftices or resorting data generated by computers in one department for entering into another system, or the rekeyboarding that has been necessary when one organizations data was passed on to another. Office occupations must be considered in the context of industries and industrial sectors, all of which have some component of office work. In office-oriented industries such as finance, insurance, and real estate (FIRE industries), a very high proportion of the work force in clerical and professional occupations is likely to be directly affected by office automation. In other industries, such as machine tools production, a much smaller proportion of the work force is in office occupations, and these tend to be in less specialized jobs (i.e., general secretary). Some occupations are more or less standard across industries. Typists and payroll clerks do much the same work with only minor variations in whatever industry they find themselves. The same is true of many professional categories tax lawyers, auditors, or certain types of engineers work for many industries that require the same general skills. Other occupations are characteristically found in one industry or a cluster of highly related industries. Both generic and industry-specific occupations may perform narrowly defined tasks and thus be potentially subject to displacement. If office automation is broadly adopted across industries for tasks performed by generic occupational groups, then the people who are displaced will find it difficult to find similar jobs in a different industry. On the other hand, if industry-specific occupations find their work being automated, they may move to smaller firms or less rapidly automated firms within the same industryunless those applications are insensitive to the size of the firm. Labor force adjustment to office automation may require major shifts in employment levels across occupations. Industry sectors differ in the proportion of workers in major occupational categories (see table 2-2). Professional Occupations Professionals here are defined as workers doing cognitive tasks that require specialized knowledge gained through lengthy education, often with a graduate degree, or special certification or licensing. 29 They make up more than 16 percent of the work force and are found in all industry sectors, but there are relatively few in some sectors such as agriculture and trade, whereas in some service industries (health services, business services) they make up more than a third of the work force. Some professional categories are generic; lawyers, accountants, and computer scientists for example are found in every industry. Others are industry-specific, such as insurance investigators, title examiners, and brokers floor representatives; or are heavily concentratedmore than 82 percent of financial analysts, for example, are in the FIRE industries. 30 In an information-driven economy, in which serviceswith high concentrations of professionals-is the most rapidly growing sector, it is likely that the demand for professional workers will increase. This category has indeed been increasing much more rapidly than any other occupational category. Its growth is not, however, unaffected by office automation. Many professionals are not primarily office workers; for example, teachers, medical doctors and laboratory scientists may have or use offices, but the core of their activity is not office work and is not directly susceptible to office automation. However, it usually requires supporting clerical and administrative work that will be affected. Professionals may themselves use office automation in peripheral tasks, and their professional tasks may be
PAGE 62
Ch, 2Productivity and Employment l 5 3 automated by other kinds of computer applications (e.g., diagnostic technology) beyond the scope of this assessment. About half of all professionals are in officeoriented occupations; for example, lawyers, accountants, analysts, and consultants. As already noted, some highly specialized professional tasks are potentially subject to computerization through expert systems. More commonly, information systems and data banks allow part of the professionals responsibilities to be taken over by paraprofessionals with less extensive graduate education and less stringent credentialization and certification requirements. For example, paralegals, engineering technicians, and library assistants do some of the work formerly done by professionals, and are paid less than professionals. Thus, there is strong economic motivation to constrain the increase in professional jobs by combining automation and paraprofessional workers, and the number of paraprofessionals is growing rapidly. Office automation also typically decreases the amount of time that a professional spends in accomplishing a given amount of work, for example, in telephone communications, locating and aggregating information from scattered libraries, or reformatting tables and drawing graphics. Some tasks incidental to, and preparatory to, generating or analyzing information can be eliminated or shortened. From this perspective, it can be argued that the amount of information-handling might increase a great deal before requiring a significant increase in the number of professionals. Once organizations have gained experience in successfully automating basic clerical tasks, office automation for professionals and managers will look increasingly attractive, since their higher wages will increase the cost-saving possible from automation. out this decade, and the National Science Foundation has forecast a shortage between 1982 and 1987. 3] This is however growth from a small base; there were just over half a million computer specialists in 1982, 0.56 percent of total employment. Moreover, computer technology itself may dampen the expected demand; as computers and their software become more usable by nonspecialists, the need for programmers is shifting from user organizations to producers of computer goods and services, and should grow at a slower rate. Software engineering systems will automate some of the work of both programmers and systems analysts. Thus, there is likely to be some leveling off of the need for computer specialists in spite of the growth in number of computers in use. Demand for information scientists may continue to grow, especially with the spread of the relatively new industry of information service, which provides and manages databases, on-line search services, and customized searching and abstracting. 32 It is not clear from occupational statistics how many information scientists there are or how rapidly their number is increasing, since many different job titles are used in this new professional area. There are other specialized professional occupations that are growing because of computers. For example, personal financial advisors and tax advisors were once used only by the very rich, but small computers made it possible for them to offer their services more cheaply to middle-income people who have money to invest but are confused by complicated choices among financial services and investment and tax-sheltering schemes. The ultimate effect of office automation on professional employment is hotly debated. Some argue that there will be a peaking and eventual decline in information-handling jobs, Office automation stimulates the growth of ] National Science Foundation, Projected Response of the some professional categories. Between 1970 Science, Engineering, and Technical Labor Market to Defense and 1978, the number of computer specialist and Nondefense Needs: 1982-1987 (Special report NSF 84-304), jobs grew by 58 percent (compared to 20 perJanuary 1984. -U.S. Department of Commerce, International Trade Adcent growth in all jobs). BLS expects growth ministration, Competiti\e .4ssessment of the U.S. Information in computer specialists employment throughServices Industrj, May 1984.
PAGE 63
54 Automation of America Offices including professional jobs. 33 Others argue that the role of professionals within organizations will change, as the information component of products and services increases, and they will become generalists who include many traditional managerial functions in their everyday work. 34 Possibly the meaning of the term professional will become progressively blurred as specialized knowledge becomes more widely accessible through information technology. Clerical Occupations The great growth in clerical occupations mirrors the broad shift from a manufacturingbased economy to a service-based economy. In manufacturing, clerical jobs are only a small proportion of employment, although that proportion has been increasing. In those service industries that collect and use large volumes of standardized data, such as legal firms, insurance, banking, and credit, more than half of all employees are clerical workers. A large number of clerical occupations are generic, and the work is similar across industries. For example, bookkeepers, accounting clerks, file clerks, general clerks, office machine operators, payroll and timekeeping clerks, personnel clerks, receptionists, telephone operators, order clerks, and shipping and receiving clerks can be found in nearly all industries. There are more than 100,000 workers in each of those categories. Official occupational statistics treat other occupations as industryspecific; thus, according to BLS, all insurance clerks (medical) are employed in health services, all train ticket clerks and freight rate clerks are in transportation, communications, and utilities industries, and all credit authorizers are in wholesale and retail trade. However, the basic office skills used in those industry-specific occupations are in large part transferable to other clerical occupations. For ex~mple, Charles Jonscher, Information Resources and Economic Productivity, Information Economics and Policy 1 (North Holland: Elsevier Science Publishers, 1983). Paul A. Strassman, Information Pa.\off: The Transformation of Ilork in the f;lectronic Age (New York: The Free Press, 1 985). There are of course exceptions to this general rule; general secretaries could probabl~ not become legal secretaries without additional training. although legal secretaries might become ~eneral (or execut i~e) secretaries in another industry, To the extent that office automation reduced employment in one or a few industries, clerical workers should be able to move to similar jobs in other industries. Their mobility might be greater than that of professionals in industry-specific occupations. This rationale sometimes appeared in the 1970s, for example, in explaining why more attention was paid to the displacement of aerospace engineers than to displacement of clerical workers from the same firms. But office automation appears likely to be adopted across industry boundaries in a relatively short time. Thus, if computers sharply reduced the demand for bookkeepers, bookkeeping jobs would be increasingly difficult to find in most industries or locations. Potentially the most dramatic and widespread impacts of office automation on clerical employment are related to data entry, as already discussed. The strong trend toward capture of data at the point of origin, and further development of optical character recognition (OCR) technology and speech recognition technology are likely to greatly reduce the need for primary keyboarding. OCR is already being used to this end. More and more information will be in machine-readable form from the beginning, and computers and telecommunication technologies will increasingly exchange information between organizations without the necessity of rekey boarding. This trend alone will have a strong impact on clerical employment. Advanced communications strongly affect clerical jobs that provide the interface between people, organizations, and activities, for example, messengers, mail clerks, and telephone and switchboard operators. Such displacement has been going on for a long time. It also affects clerical workers not usually considered communications workers. A chain of hotels may have a centralized, computerized worldwide reservation system with local-area networks and microcomputers. When a reservation is made through a toll-free phone number, all appropriate information, including credit references, will be automatically loaded into the central mainframe database. Information about next-day guests will be transferred to the hotels microcomputers in the middle of the night and a reservation clerk will have only
PAGE 64
Ch 2Productivity and Employment l 55 to key in information on last minute walk-in guests. 36 Not all clerical jobs are equally vulnerable. of the ten fastest growth clerical occupations between 1970 and 1978, all but one (computer operator) were occupations that require direct contact with people outside an officee.g., cashiers and receptionists. These jobs are sometimes considered relatively impervious to office automation. However, when the personal transactions can be standardized the job can be automated; again the banks automated teller machine is one example. Supermarket cashier stations are another. Some clerical jobs include both informationprocessing and manual tasks. Shipping and weighing clerks, packing clerks, etc., may be only indirectly affected by office automation, but are also vulnerable to the effects of automated materials handling and storage systems. Many organizations are also beginning to electronically print forms as needed, eliminating the need to buy, handle, and warehouse preprinted forms. In this case, the cost-saving sought is primarily in the costs of space and materials, but labor-reduction is an added benefit. Ten occupations with declining employment from 1970 to 1978 were tabulating, bookkeeping, calculating, keypunching, stenography, postal clerk, telephone operators, mail carriers, and meter readers. Most of these are occupations directly affected by the early phase of office automation and the effects are now showing up in statistics. They are also occupations that are relatively narrowly defined in terms of tasks, again illustrating that the most narrowly defined jobs are those likely to be first eliminated. In a later section of this chapter, the use of part-time and temporary office workers is discussed, and other chapters examine in detail the phenomena of home-based work and off-shore work. Part-time and temporary work distribute employment among more workers, ((~n]r]]unicati(jns Tiwk, Iiolida~ 1 nns to (Jse SN .+1 (;atewa?r to I,ink 1,600 P(s Writ h hl ainframe, N 01. 5, 1984, p, 49. but they tend to depress the number of fulltime-equivalent jobs. They allow employers to hire the minimum number of people necessary for their base workload, relying on part-timers and temporaries to handle short-duration increases or peak loads. Home-based work programs have the same benefit of load-leveling since most of the home-based workers are used on the basis of when work is available. Offshore data entry, on the other hand, simply eliminates jobs in the United States. In short, while office automation will likely lead to a great deal of shifting between clerical occupations, it is also highly likely to result in an absolute reduction. across the board. Managers and Supervisors Table 2-2 showed that managers comprise a relatively small proportion of total employment; the proportion varies considerably across industries. BLS data identify a principal group of managers, officials, and proprietors, and also identify supervisory occupations within the clerical and other personnel categories. within most industries, firms vary considerably in the proportion of managers to other employees, with small firms tending to have proportionately fewer managersi.e., a flatter structure. Organizations are generally using more data, in more systematic ways, in coordinating their operations. As management becomes more information-intensive, it is possible that more managers may be needed. There is some relationship between the information intensiveness of an industry and the industry proportion of employees who are managers. Manufacturing firms have relatively low levels of managerial staffing, while insurance carriers, securities, computer and data processing services, and mailing and reproduction services have relatively high levels. Managerial employment is also above average in accounting and auditing services; engineering, architectural, and surveying services; wholesale and retail trade; transportation, communications, and utilities; and printing and publishing. It is relatively low in health services and legal services, but
PAGE 65
56 l Automation of Americas Offices these industries have large numbers of professionals who also act as managers. There are large numbers of managers whose work is only partly and secondarily office work; for example, bar and cafe managers or automobile repair shop managers. Office automation technology, if adopted, may help them to focus on supervisory responsibilities by reducing the time spent on office chores. In offices, some managers are primarily concerned with the direction and supervision of clerical or production operations and personnel. If fewer clerical personnel are needed with office automation, proportionately fewer supervisory managers should also be needed. In addition, as discussed in chapter 4, office automation can be used to increase the scope of supervision; that is, one manager or supervisor can monitor the performance of more workers or an increased volume of production. The tasks of many lower and middle-level managers largely center on: 1) information collection, processing, and reporting, tasks that office automation either facilitates and enhances, reduces the necessary expenditure of time, or takes over completely; and 2) communication and coordination, where office automation can be very time-saving. Higher level managers concentrate more on decisionmaking. Here computers can be used in many ways. Some decisionmaking can be, and is being, built into computer programs, reducing the need for lower level managers. Management information systems and other kinds of computer programs are designed to help managers make decisions, by integrating and displaying the information they need. In some situations this saves a great amount of time for the manager; in other situations, it takes more time to make a decision because more information is available to be considered. In some organizations, information that was once collected, integrated, and laid out by lower level managers for review by higher level executives, is now aggregated and formatted by computers and accessed directly by the decisionmaker. This points toward a reduced number of lower level managers, and some corporations are reported to have adopted office automation with the explicit objective of flattening the management hierarchy. Heightened competition and pressure to cut labor costs can lead firms to try to keep a lean management staff, since their salaries are high compared to other workers. At present, the job market for managers is strong because of overcutting of managerial ranks during recent recession years. 37 Thus, there are conflicting trends to be considered in the outlook for managerial jobs, but they are not immune to the effects of office automation. Technology and Recent Trends in White-Collar Employment White-collar employment has grown rapidly in recent decades. (See figure 2-3. ) The narrower category of clerical jobs, about 16 percent of all employment, has also grown rapidly even through the first phases of office automation. Table 2-3 appears to indicate, however, that this growth may already have slowed or stopped. 38 -Dun Business Month, Executive .Job Market: Filling the Talent Gap, November 1984. Does not include cashiers (of whom there were about 230,000 in 1950 and 2.2 million in 1984). Beginning in January 1983 BLS reclassified some occupations; cashiers and real estate appraisers were removed from the clerical occupations category. These reclassifications are one of the continuing pitfalls of working with occupational data time-series. Figure 2-3.Changing Percentage of Work Force From 1900-80 White-Collar Compared to Blue-Collar a) ~m : 50 White collar z ~ 40 E Blue collar : 20 & n lo o~ 1900 1910 1920 1930 1940 1950 1960 1970 1980 Year SOURCE U S Department of Commerce, S(at/st/ca/ Absfracfs of the Urr/fed Sfafes 1981, No 673
PAGE 66
l 5 7 Table 2-3.The Growth in the Number of Clerical Jobs, 1950.84 Number of clerlcal Growth In the Total employ-merit Year jobs 4 (mllllonj number of jobs (percent ) 1950 66 11 3V 1960 88 33 134 1970 129 46 164 1980 169 31 169 1982 168 O 6 168 1984 167 O 6 159 a The difference between the 1982 figure for clerical jobs on tables 22 and 23 Ill ustrates the dlff(culty of analyzlng deta!led employment trends because of am blgultles In data sets SOURCES U S Department of Commerce Bureau of the Census Census of Popu/at/on f950 1980 1970 1980 and U S Department of Labor Bureau of Labor Statlstlcs Current fopulaffo~ Suwej Annual Aver ages 1982 and 198 4 If the automation of office work is laborsaving, why did clerical employment increase so rapidly during the 1960s and 1970s? The capital-output ratio, which has traditionally been very low in white-collar work, increased steadily (see figure 2-4), but from these figures would seem to have had little effect on labor demand. A closer look at the financial service industries, which have been at the forefront of office automation, tends to counter this impression. They were expanding rapidly during this period in part because of the new products and services made possible by their office .eontief and Duchin, citing man~ sources, accept that white-collar workers at the beginning of the 1980s worked with an a~erage of $2,000 in equipment compared to the factcq workers ;$25,000 in equipment; op. cit., p. ,5.7. Figure 2-4. Capital Investment Per Production Worker, 1970-80 32 6 30 25 20 19.5 15 10.4 10 91 5 4.8 0 1970 1975 168 III 7.0 1980 Year c1 anufacring clTe mc aThe ratio IS net eaul~ment stock (In bill Ions of dollars) to number of Droduc tlon employees In coristant dollars SOURCE David Roessner Market Penetration of Off Ice Automat Ion Equipment Trends and Forecasts (working paperl School of Social Sciences Georgia Instlfute of Technology 1985 Ch. 2Productivity and Ernp/oyrnent automation. Life insurance carriers, for example saw sales increase over 49 percent from 1970 to 1980. Yet their labor force increased only 9.8 percent. During the 1970s the average annual real value added per full-time employee equivalent (a measure of productivity) for all industries was 1.1 percent; in the insurance industry it was 2.7 percent. From 1975 to 1979 it was 1.3 percent for all industries but 6.7 percent for insurance carriers. Social, economic, and political conditions favored the expansion of the FIRE industries during this period. High levels of employment growth during the 1960s and 1970s and the rapid increase in the number of two-income small families drove the proliferation of checking accounts. The rising level of disposable personal income during much of that period, and the fact that more women were earning income and insuring themselves for the first time, increased insurance expenditures. Fringe benefits also expanded, and workers compensation coverage was extended. A flood insurance program was established. New insurance and financial service products responded to, as well as contributed to, the growth in the market. l Because these industries were growing rapidly, office automation in the 1970s constrained growth of their work force but did not reverse it. The labor-saving effects are nevertheless apparent; between 1970 and 1978, insurance industry professional and technical workers increased by 24 percent and managers by over 21 percent, but clerical workers increased by only 8 percent. During this first phase of automation, technology most directly affected clerical work. In the occupations where automation directly substituted for labor the effect was greater. Key operators declined by 22 percent, bookkeepers by 7 percent, file clerks by 20 percent, mail clerks by 11 percent, and typists by 12 percent (but secretaries increased Baran, op. cit., pp. 100 ff. Inflation and higher interest rates had a mixed effect; insurance carriers, for example, derive more income from investments than from policies but were sometimes locked into older low-interest in~estments and needed new investments to balance these. Inflation left life insurers more vulnerable to disintermediation and made forecasting of future cash flows for investment difficult. Liability settlements were also growing.
PAGE 67
58 l Automation of America's Offices by 8 percent and computer operators by 119 percent). In some parts of the insurance industry, there were also dramatic labor reductions in some areas of professional work; for example, among underwriters in the life and health insurance fields, there was a dramatic decrease in numbers during the 1970s. As a result of the interaction between market growth and office automation, labor displacement in FIRE industries as a whole clearly occurred but resulted in depressed growth, rather than decline, in employment. 4~ In financial services, employment increased by about 2o percent, but growth was lowest in sales and clerical jobs, and highest for managers and professionals. The labor-saving effects of office automation are likely to become more apparent in the near future. The first phase of automation was a more direct substitution of machine for labor than is later office automation, but in the first decade of use, organizational goals placed higher priority on better data collection and reporting and news services. The mass data handling industries were automating pre-existing functions, such as payroll, inventory control, and other basic procedures. But they were dealing with anew technology, with no experience and precedents to guide them; reorganization of the organizations workflow and labor force took time. In the FIRE industries there are strong indicators that the emphasis has more recently shifted to cost-reduction, and much reorganization and labor force reduction is now occurring. Unit labor costs have in fact been dropping since 1969 and the drop began accelerating about 1975. 4 Market conditions can either offset or reinforce employment effects. In the case of the Baran, op. cit. Valerie Personick, The Job Outlook Through 1995: Industry Output and Employ merit, ~lonthl~ Labor lie~iew, No\rember 1983, p. 34; U.S. Department of I.abor, Bureau of I,abor Statistics, fechnolo~ and Laborin Fitw Industries, Bulletin 2033, 1979. Baran, op. cit., p. 105. 13aran, op. cit.; also P:ileen Appelbaum, op. cit. insurance industry, one group of analysts concludes that: some of the present job losses occurring in the industry are directly attributable to computer technology, though more often than not these are the delayed effects of an earlier stage of innovation rather than the latest developments: some of the losses are best regarded as the indirect consequences of technology, which for example might allow reorganization and rationalizations to be made; and some losses are attributable to separate factors such as declining market conditions or out-moded management structures. On the whole one might sum up by saying that computer technology has provided the vehicle which makes it possible to respond efficiently to market conditions, whether this be by expansion of business or by contraction of operating costs. 46 This analysis agrees with other evidence that white-collar employment is becoming more sensitive to cyclical economic conditions. In both future and past shifts in employment, it is difficult to separate the role of technology from the effects of market change and other broad economic factors. But analysts at Bell Canada have studied the effects of technological change on their work force from 1952 to 1972, and concluded that technological change resulted in substitution of capital for low-skill labor, overwhelming any price complementarily with capital. In 1952, Bell Canada was using an average of 48 to 52 million person hours yearly, 23,000 to 25,000 jobs. Over the next two decades output steadily increased by 7 to nearly 10 percent per year, or over 500 percent, but labor demand at the end of the 20 years was less than 15 percent higher than in 1952. The econometric model used in the Bell Canada study was designed to separate technological effects (i.e., automation) from price and Richard 13arras and Julia Swarm, The Adoption and impact of Information Technolo~ in the UK Insurance Industr~r (1.ondon: The Technical Change Centre, No\ember 1983), p. 21. -Michael Denny and Melvyn Fuss, The Effects of Factor Prices and Technological Change on the Occupational Demand for I,abor: Evidence From Canadian Telecommunications, The Jourmd of Human Resources, XVII, 2, spring 1983, pp.-l 76.
PAGE 68
market effects. The study concluded that technological change out weighed wage/capital and wage/material ratios in affecting overall employment levels, and in the case of demand for telephone operators, whose jobs were the primary focus of technological change, far outweighed the effect of output growth. Small computers and word processors allow the standard office functions to be automated in small organizations, and incrementally, with relatively low capital investment at one time. Office automation need not involve construction of new facilities or extensive alteration of facilities. Where it replaces existing equipmenttypewriters, calculators, bookkeeping machines, and old telephone systemsthat stock does not represent large amounts of embedded capital. Most such equipment is more than 5 years old and has already been amortized. Office computers and equipment also enjoy a rapid tax write off. For all of these reasons, office automation may proceed more evenly, more widely, and more rapidly than other kinds of automation did in the past. The speed with which a technological change occurs, and its breadth, are both important in assessing the impact. A slower pace allows both individuals and the labor market to make whatever adjustments are possible. In this regard the potential office automation of small businesses is particularly important. Small firms, in many parts of the country, account for the preponderance of office jobs. A reduction in the number of office jobs available relative to the total work force would therefore be felt in all areas of the country as well as in all industrial sectors, although not with the same force in all sections and sectors. Part-Time and Temporary Employment The proportion of part-time and temporary workers has been increasing since the early 1950s. The number of voluntary part-time workers has remained between 13 and 14 percent since 1970 but the proportion of involuntary part-time workers has continued to increase (see table 2-4 and figure 2-5), indicating that the strongest factor in the growth Ch 2 Productivity and Employment l 59 Table 2-4.Changes in the Percent of Total Employment a for Involuntary and Voluntary Part-Time Workers as Percent of Total Employment, 1968-84 Part-time work Involuntary Year Total workers 1968 ........, ., 14 9 0 2.5 r, 1969, 155 2 6 1970 164 31 1971 ... ., 16.8 3 4 197 2 16.8 3 3 1973 .., ... 166 31 1974 . : 171 3 5 1975. ., 184 4,6 1976, . 180 4 2 1977 ..., 180 4 0 1978 .., 177 3 8 1979 .., ., 17.6 3 8 1980 ......, 184 4.5 1981 . 186 4 9 1982. ... 202 6 4 1983, ., 200 6.5 1984, ... . 189 5 7 force Voluntary workers 12.400 129 133 134 135 135 13,6 13,8 13,8 140 139 138 13,9 13.7 13.8 13,5 13.9 aThese cal rula! I rjns are for nonagr!cul u ~al workers agec 16 anc over SOURCES 1968.81 Labor orce stat!st IWS derived from the Cur,en( Popu/afjor~ Survey A Dafa BOOA Vo/ume f (Washington DC U S Depar(meot of Labor B u reau of Labor Stat I st I cs Septern ber 1982 Bu I let I n 2096) D 682 1982 Emploumenf and Earn/ngs VOI 30 N o 1 (Wash I nqtm DC U S Department of Labor Bureau of Ldhor Statlsttcs Janu~, y 1983) p 169 1983Employment and Earnings VOI 31 N,) 1 (Washlnq ton, DC U S Department of Labor, B u reau of Labor Stat I st cs Jan u ary 1984) p 194 1984 Ernp/o) rnt?n( an{f Earn I ngs VOI 32 N c, 1 (Washington DC U S Department of Labor Bureau of Labor Statls IICS January 19851 p 192 Figure 2.5Changes in the Percent of Total Employment a for Involuntary and Voluntary Part-Time Workers, 1968-84 20 I 15 -.* ..**** 10 l l l l l O* l l l l l 8 l l l l l l l U Voluntary R Involuntary t5 0 I I I I I I I I 1968 1970 1972 1974 1976 1978 1980 1982 1984 Years SOURCES 1968.81 Labor fore e stat tst ICS der!ved from the Current Popu/a/IcI,I Survey A Data Book Vo/urne 7 (Washington DC U S Department of Labor Bureau of Labor Stat I st ICS September 1982 Bu I let I n 2096) p 682 1982Ernp/oyrnenf and Earn/frgs VOI 30 No 1 (Washington DC U S Department of Laho! Bureau of Labor Statlstlcs January 19831 p 169 1983Employment and Earnings VOI 31 No 1 (Washing ton DC U S Department of Labor Bureau of Labor S(at{st!cs Janu ary 19841 p 194 1984Emp/oyment and Earn(ngs VOI 32 No 1 (Washington DC U S De~artmerrt of Labor Bureau of Labor Stat/s tics January 1985) p 192
PAGE 69
60 l Automation of Americas Offices is not workers choice of a more flexible lifestyle, but employers response to economic pressures. In some industries and some organizations, slack workloads lead employers to convert workers to part-time in preference to a layoff. Other employers, however, are adopting a policy of keeping a minimum-size work force, which can be temporarily augmented when necessary .48 There are reports that in other industrialized countries automation has greatly increased part-time work; for example in Japan, introduction of part-time workers and subcontraction has grown massively. Office automation and creation of a part-time work force are in some situations alternative or competing strategies for cost-cutting but they may also be complementary. Part-time workers (considered by BLS as an employee working less than 35 hours a week) are cheaper than a proportionately smaller number of full-time employees because they often are paid lower wages and do not qualify for benefits packages, regular yearly wage increases, or job security agreements based on seniority. There have been many anecdotal and press reports of companies reducing work hours to one or two hours fewer per week than would qualify workers as full-time employees, but few companies are willing to admit formally to this practice. The biggest advantage of part-time workers for employers however is that of load-leveling; that is, they can be used during parts of the day or week when the workload is heaviest. 50 T o the extent that office automation allows the work force to be reduced and workflow made more efficient, it may obviate some interest in moving toward a part-time work force. See, for example, a recent article in Business Week, PartTime Workers: Rising Numbers, Rising Discord, Apr. 1, 1985, p. 62, reporting explicit statements by several company spokespersons about reluctance to staff to full capacity. Katsus Nishiyama, Introduction and Spread of VDT Work and Their Occupational Health Problem in Japan, to be presented at the 5th UOEH International Symposium in Japan, Sept. 19, 1985. Business W~k, op. cit. Another strong factor has been the growth in demand for part-time employees by fast-food restaurants, shopping centers and shopping malls, and neighborhood banking locations, many of which are open long hours, at night, or on Sundays. But in other situations, office automation encourages the creation of a part-time work force. Where it is used to standardize and deskill work many employers have found it profitable to use part-time, low-paid workers. Some have reportedly moved to suburban locations to take advantage of the availability of housewives willing and eager to work part-time at low wages because there is another primary wage-earner, with a full benefits package, in the family. As discussed in chapter 7, office automation also makes it feasible to use homebased workers, on a part-time and piece-rate basis. In the long run, office automation may stimulate a stronger trend toward use of parttime or temporary workers by allowing employees to maintain a minimum work force that will need supplementing during hours or seasons of work overload; and by standardizing the basic skills needed by clerical workers and some kinds of professional and technical workers. In 1955, only 8 percent of American workers were part-time; 51 this rose fairly steadily to about 15 percent in the late 1960s and continued to rise to 20 percent by the 1980s. (See table 2-4.) Thus, about one-fifth of American workers are working part-time. Women are much more likely to work part time, often in order to combine paid employment with child care. About 29 percent of working women work part time, compared to 12 percent of working men. About 21 percent of teenagers aged 1619 and employed, are working part time. In 1983, in the FIRE industriesleaders in office automationonly 11 percent of employees were part-time. This sector ranked fourth among major industry sectors, after the wholesale and retail trade (32 percent), service industries (27 percent), and construction (14 percent). In the service industries, a large proportion of the part-timers were probably also office workers. In the office-oriented sectors of banking and insurance of other industrialized countries, however, part-time work is expected to increase, ] New Work Schedules for a Changing Society (Elmsford NY: Pergamon Press, 1981), p. 45.
PAGE 70
Ch. 2Productivity and Employment 61 According to the International Labor Organization, part-time employment in the banking and insurance industries is rising in its member countries, and in Sweden over 26 percent of banking and insurance employees are part-time. 52 ATMs have probably reduced the need for part-time tellers and clerks in the United States. But other forces are now at work. Four of the biggest eight accounting firms, and many financial service firms including Citibank, Travelers Insurance Company, and other major employers of clerical workers such as Control Data Corporation (CDC) are now emphasizing part-time employment. 53 Travelers Insurance Company has developed a job bank of retired professionals for temporary market research and product development, and plans to train them for use as programmers, parttime. CDC has a formal program of using parttimers, which has been in effect for 2 years. The goal is to have 15 percent of their work force (chiefly clerical and production workers) on part-time or temporary status and another 15 percent as independent contractors. This goal has been partly realized; by 1984 CDC was reported to have 4,500 part-time workers, or 10 percent of their work force. Closely related to part-time work is temporary work, which for employers is another strategy for workload leveling. Many clerical workers are temporaries, but there is a growing trend toward using temporary programmers, systems analysts, computer engineers, and data communications specialists. 54 Temporary workers can be called in on short notice when work is briefly or seasonally heavy, and can be dismissed almost instantly and without penalty. From the employers viewpoint temporaries are part-time workers for International Labor Organization, Advisory Committee on Salaried Employees and Professional Workers, The Effects of Technological and Structured Changes on the Employment and Working Conditions of Non-Manual Workers, Eighth Session, Geneva, 1981, pp. 50-51. Joann Lublin, Shorter Hours: More Managers Are Working Part Time; Some Like It But Others Have No Choice, Wall Street Journal, June 2, 1982, p. 50. John J. Davis, President of Worldwide Computer Services, Inc., Is There a High-Tech Pro in Your Future? Management Information Systems Week, May 22, 1985, p. 64. Pholo cred(t Kelly Serwes, lnc This simulator exactly duplicates several major word processing systems and is used for testing in an employment services firm. whom the organization has no responsibility for long-range job security. The worker who is individually hired on a temporary basis generally suffers the disadvantages of apart-time workeri.e., not qualifying for benefits and relatively little chance of promotion, and by definition has no job security. Many temporary office workers, however, are employed by firms within the new temporary employment service industry; that firm provides them with assignments to client firms. The worker may be available to the employment service firm fulltime or nearly full-time, or may wish to work only occasionally or sporadically. Some employment service firms are now providing their regular workers with prorated benefits similar to those that they might receive as permanent employees of a large firm. More generally, however, temporary workers do not have such benefits. The temporary service industry is growing rapidly, nearly twice as fast as GNP over the last 14 years; and faster than the computer equipment industry, to a payroll of $5.5 billion in 1984. 55 At first, it appeared that autoThe National Association of Temporary Services reported a payroll of $431 million in 1971 and $5.50 billion in 1984, an average of nearly 20 percent growth per year. (Figures sup(continued)
PAGE 71
mation would be a barrier, since temporary workers would be unfamiliar with equipment, which varies widely between offices, but the larger temporary agencies provide training in a broad spectrum of office automation equipment and applications, particularly word processing. Also, automation has standardized and de-skilled some tasks for which temporary workers can be used. Clerical occupations accounted for over 60 percent of temporary assignments made by the industry in 1980, although it accounted for only 49 percent of the revenue since clerical wages are generally lower than those in technical, medical, and industrial assignments made by the temporaries industry. Temporary computer and data communication specialists, and other professionals, are in growing demand. They can offer up-to-date knowledge of current systems, languages, and protocols, because of recent schooling and varied experience, and may choose to work as temporaries for fear of becoming trapped in a narrow specialty or job where their knowledge will gradually become obsolete. 5G A computer services official attributes the trend toward use of temporary professionals in computerrelated work to the triangle brought about by mounting costs in corporate and government-mandated fringe benefits, the inability of many companies to meet peak workloads with their permanent staff, and finally the growing number of professionals who desire to change their work patterns. Closely related to temporary employment are employee leasing and use of independent contractors. Employee leasing may be used by employers as a still longer term strategy for workload leveling (by month, year, or project-duration), but it is more generally used by very small firms or professional offices (doc(continued from p. 61) plied by the National Association of Temporary Services and also based on the Census Bureaus Count~~ Business Patterns, Vital Statistics of the Temporary Help Industry, Contemporary Times, vol. 2, issue 6, fall 1983.) I]a\i:, op. cit., says In practical terms, the! do not want to spend the next 5 ~rears of their careers learning how to appl~: Unix or C into an insurance conlpan-v microcomputer system. Da\is, op. cit. tors, lawyers, dentists) to shift the administrative costs and benefits costs associated with employees to a contracting firm, which can benefit by economies of scale. 58 The leased worker usually enjoys a full benefits package, although he or she is not guaranteed permanent employment (in practice, the job security may be about the same as in conventional forms of employment). Office automation appears to have given a large boost to the growth of independent contractors offering business services such as word processing, data entry, and computer programming. Independent contractors are selfemployed, with all the risks and benefits this entails; he or she assumes the costs associated with slack work periods and loss of workers benefits in return for autonomy. The work may be done in the employers facility, with the contractor/worker effectively indistinguishable to observers or coworkers from employees. The work may however be done in the contractor/workers home, using the communication capabilities of office automation. Some clerical and professional independent contractors are entrepreneurs, or small business men and women, seeking multiple clients either at one time or in sequence. They may or may not plan to expand their activities and take on employees of their own. Many contractors on the other hand work for only one firm and are in effect employees without the benefits otherwise associated with employment. The unresolved tax and legal issues associated with independent contractor status are discussed at greater length in chapter 7 in connection with home-based clerical workers, many of whom are former employees converted to the status of independent contractor. Part-time and temporary employment and independent contracting are likely to increase as automated offices move toward a lean work force with need for occasional supplementary business services, and as more workers are familiar with the equipment. There are strong %om-e service contracting companies make their profits from the interest on advance deposit of the monthly fees paid I]y the client to cover wages plus associated costs.
PAGE 72
Ch. 2Productivity and Employment l 6 3 benefits in it for workers as well as for employers. Many people prefer and actively seek part-time work. Students, mothers, and retired people often want to work less than a standard workweek; others want more time for families, education, or recreation. They choose to trade income for leisure time, and are willing to pay the additional costs in terms of loss of benefits such as health insurance, lack of job security, and diminished likelihood of promotion and advancement. The standard 40 hour workweek has not changed since the 1930s, and part-time work is the way some people create their own shorter workweek, Many temporaries choose this form of employment because they want or need the flexibility it gives them.. Some use it as a form of job-hunting, or trying out potential employers. However, some temporaries are unable to get assignments as regularly as they wish, and find the unpredictability y of their income a severe disadvantage, but have been unable to find permanent employment. At a minimum, part-time work is preferable to unemployment, Employers sometimes convert full-time employees to part-time status during a recession, in preference to laying them off and losing a valuable worker. 59 If part-time work is beneficial to many employers and is sought by many employees, under what conditions is it a public policy concern? First, if enough full-time jobs are eliminated-i.e., converted to part-time jobs, opportunities will be diminished for those who must have full-time work to make enough money to support themselves and their dependents. Second, in the United States, many social services and income protection mechanisms are provided not directly by tax payers but through employee benefits packagese.g., health insurance, life insurance, income duri n g illness or childbirth, pension plans, and to some extent training and higher education. These protections are much more costly, if they are available at all, on an individual basis. If conversion to part-time work means that a sizable proportion of the population no longer has these protections through employment, then the taxpayer is in the long run likely to bear more of the burden of the illness, old age, and death for these people, and the average level of health and well-being of the population is likely to decline. Society may be willing to bear this risk, if that is the price of allowing people to choose part-time work. If part-time work is not a choice, but the only alternative available to them, and especially if this limitation on choice is the result of employers decisions, then the public policy issue becomes one of whether this shifting of responsibility for basic protections from employer to employee is acceptable to the society at large. Historically, the choice of full-time or part-time work has been regarded as the individuals prerogative. We must then ask: is this still a free choice, and will it be so in the future? To what extent is involuntary part-time work increasing? The official part-time employment figures based on annual aver-ages do not tell the whole story. The number of people who work part time at some time during a year is often double the annual average. For example, in 1978, a recession year. the annual average was 21.4 million part-time workers, but a retrospective survey indicated that 40.9 million people, at some time during 1978, had only part-time work. () While the annual average showed 3.4 million of the part-timers as working part-time involuntarily (that is, because they could not find full-time work) the retrospective survey counted 10.1 million. The number of involun tary part-timers has been increasing, as shown in table 2-4, to more than a quarter of all parttimers (and about 5 percent of all employed)
PAGE 73
64 l Automation of Americas Offices at present. Recent increases in involuntary part-time employment and multiple-job holding for women suggest that there may not be enough full-time jobs for those that seek them. 62 Thus the number of people who want to work part-time may be smaller than assumed, and the change in attitudes and lifestyles cited above as a factor in the trend may not play as strong a role as is often assumed. Public interest in part-time and temporary work is therefore twofold: 1. 2. that the number of involuntary part-time workers not increase to undesirable levels, causing a deterioration in income levels, because full-time jobs have been converted to part-time jobs; and that the costs of worker protection not be shifted from employer to worker to a degree that ultimately causes them to be borne by the taxpayer. These unanswered questions again point to the need for more careful and systematic monitoring of trends in employment, since a significant increase in the number of long term involuntary part-time workers would challenge the adequacy of existing mechanisms for income security and other employee protection mechanisms. Analogies From Past Waves of Automation Throughout modern history, mechanization and automation of work have brought dire warnings of unemployment. 63 But employment has continued to expand. Mechanization and automation have contributed to, or driven, this expansion by reducing the costs of food and A~cordi~g to Professor Eileen Appelbaum in a talk prepared for presentation to the Panel on Technology and Womens Employment of the National Academy of Sciences, Washington, DC, Feb. 17, 1985. ] Mechanization is the use of machinery as a substitute for human or animaJ labor. Automation is a narrower term, meaning the use of machinery that makes decisions about the work without human intervention; that is, machinery with control systems that incorporate the principle of feedback to finetune or correct the machinerys operations. Computerization carries this internal decisionmaking a great deal further, with the use of information stored in memory and by the sensing of external conditions. material goods, stimulating the market for them, providing paychecks for workers to buy those goods, and creating capital to be invested in production of more and still cheaper goods, further stimulating the market. In general, workers displaced by mechanization have taken other jobs in the same industry as it expanded, or moved into new or expanding economic sectors. The argument from history is powerful; in general, technology creates rather than destroys jobs. That is why developing nations, with exploding populations, struggle to industrialize. There are however several important considerations to be noted. The great waves of mechanization and automation in the past were still part of the continuing industrial revolution. The United States was, in the 19th century, a developing nationan agricultural nation becoming an industrial nation, with an expanding national market based on plentiful resources, in which the consumers who bought goods were also the workers that produced them. From World War II through the 1950s, at least, U.S. technology enjoyed worldwide preeminence across the board. But this is now a mature economy, strong but with increasing competition for both domestic and world markets. Imports are a major factor in the economy. It is therefore not clear that the American economy will grow, in the future, at the vigorous rate of the past. In an economy that is growing more slowly, new jobs are created at a slower rate, and workers do not enjoy the mobility they have in a rapidly growing economy. Secondly, past waves of automation have been concentrated in one or a few industries, for example, at one time agriculture, at another period commodity manufacturing or industries that could use assembly line techniques. Automation proceeded unevenly across economic sectors, crafts and occupations, industries, organizations according to size, and geographical regions. Jobs were increasing in some industries and occupations, when jobs at approximately the same skill level were decreasing in other industries. Large companies automated well before small companies. Many
PAGE 74
Ch. 2Productjvjty and Employment 65 kinds of manufacturing automation, for example, have never been adapted to batch manufacturing or for small machine shops. This contrasts with office automation, which can be used in basic office functions across all industries, and especially with small computers and stand-alone word processors, which allow even small offices to automate. Although total employment has grown through and after historical waves of automation, each has left behind some structural unemployment. In many cases, older workers failed to make the adjustment and new jobs went to new workers with more recent training, new skills, or more flexibility. In other cases, new jobs and new industries more than compensated for lost jobs in number, but were located in other regions, leaving displaced workers behind. (The coal miners of Appalachia are a pertinent example.) Agricultural employment declined steadily from 27 percent of all employment in 1920 to 2.7 percent in 1980 (an absolute loss of 8.7 million jobs) as agriculture was mechanized. See table 2-5. However, employment was created in food processing (1.6 million jobs in 1982), in agricultural research, and in transport and sales of food. As population and the economy grew, blue-collar employment stabilized. There were only 3 percent more blue-collar jobs in 1980 than there were in 1950, while the work force grew by 69 percent in those decades. The 30.5 million blue-collar jobs in 1950 were over half of total U.S. employment, while the 31.5 million blue-collar jobs in 1980 were less than 32 percent of total employment. Now the number is decreasing. BLS reports that 2 million manufacturing jobs have been lost since 1979. 64 Had white-collar employment not been expanding rapidly, new workers could not have been absorbed into the economy. At the beginning of the 20th century, there were only 5.1 million white-collar jobs, accounting for under 18 percent of all employment. By 1950 these jobs had quadrupled to nearly 22 million, but still accounted for less than 37 percent of employment. In 30 years-less than the working lifetime of an office worker the number of white-collar jobs has more than doubled to nearly 52 million jobs, accounting for at least 55 percent of all American workers. Structural change in the economy has created jobs in some sectors while it displaced jobs in others. The result is a net increase in employment, and there is also a more equitable distribution of employment opportunities (e.g., better status jobs for more people, and more jobs open to women). Creation of jobs has by and large kept up with both population growth and growing participation in the labor force. In 1950,57 percent of the population was in the labor had grown to nearly force, but by 1980 this 64 percent. BLS originally reported to the Joint Economic Committee that 8 million jobs had been lost, but issued a correction after this was reported in the press. See BLS Corrects Figures on Factory Job Losses, Washington Post, June 18, 1985. Table 2-5.Shifts In Employment by Industry Sectors, 1900-80 1900 1910 1920 1930 1940 1950 1960 1970 1980 Total labor force (000). .29,030 37,291 42,206 48,686 51,742 58,992 67,990 79,802 104,058 White collar: Growth . . . 560/o 320/o 36?40 12%0 3470 260/o 39!0 420/o Labor force . . 18/0 21 /0 250/o 290/o 31 0/0 37 /0 400/0 47 /0 51 /0 Blue collara: Growth . . . 370/0 140/0 190/0 11%0 14 /0 9/0 190/0 21 /0 Labor force . . 45/0 480/o 480/o 490/0 520/o 520/o 490/0 490/0 450/0 Farm: Growth . . . 60/0 /0 /0 130/0 /o 0/0 /0 /0 Labor force . . 380/. 31 /0 270/. 21%0 17/0 12 /0 60/0 30/0 30/0 %!s ~flc)udes rrrarrua/ and serwce Workers SOURCES U S Department of Labor, Bureau of Labor Statlstlcs, H/stor/ca/ Abstracts, Co/onla/ Times to 1970, Series D, Nos 182.232, p 139, and U S Department of Commerce, Statfstica/ Abstracts of the Urr/ted States 7985, p 400
PAGE 75
66 l Automation of Americas Offices But it is not clear that the creation of new jobs has completely compensated for the long range labor-saving effects of mechanization and automation. High unemployment rates (7.5 to 7.1 percent) have persisted in spite of the highest employment ever achieved in this country. In 1984, after recovery from a recession, 3.5 million new jobs were created (the second highest growth rate in our history) but unemployment did not significantly decrease and stood at 7.2 percent at year end. 65 If those workers are counted who have dropped out of the labor market because of discouragement, or have accepted part-time work because they can not find full-time work, the unemployment rate would be several points higher than it officially is; for example, in early 1985, about 10.8 percent rather than 7.3 percent. The number of jobs in manufacturing declined by 1.6 million from 1979 to 1984, and in the goods-producing sector very few industries employ more workers now than before the recession began. 66 During a recession, some markets maybe lost to international competitors, and some organizations do not recover. Also, organizations tend to adjust slowly to labor-saving technology, preferring normal attrition to layoffs; but when layoffs are forced for other reasons, they are likely to take advantage of this to eliminate redundant jobs permanently. The normal level of unemployment has in fact been rising for at least half a century. Several kinds of unemployment are usually distinguished. One kind is frictional unemploymentthat which is normally attributed to the demise of individual firms and the mobility of workers moving between schools, jobs, and occupations. Another kind results from relatively discontinuous or sudden expansions of labor supply, for example, disbanding of a military force, a wave of immigration, or less precipitous but still unprecedentedthe increased participation of women in the labor force in recent years. Third, there is cyclic un U.S. Depa,tnwnt of I,abor, Bureau of I.abor Statistics The ~jn]plo~n~ent Situation (monthly) and l~mplqJrment and Earnings. Januar~ 1984: I,inda LeGrande Employment Status of the N at,ion: Data and Trends, Congressional Research Ser\ice. 1 \sLIe Brief 11182097, updated Ma~ 6, 1985. I.e(;rande, op. cit., p. 3. employment resulting from fluctuations in aggregate demand, which can be an acute and serious problem during recessions but declines when the economy recovers. The fourth kind of unemployment and the most serious in longterm considerations is structural unemployment, often defined as a mismatch between the supply of jobs and the supply of workers with the skills needed for those jobs, but in theory also possible when there are not enough jobs, at any skill level, to engage all would be workers. Until about 1970 it was generally assumed in this country that an acceptable level of total unemployment was about 3 to 4 percent. 67 But the rate has not been that low, even in periods of expansion, since 1969. As shown in table 2-6, unemployment rates have been rising for about two decades, not falling back even in boom years to previous lows. This long-term rise in unemployment has been attributed to many causes. One is demographic-the flow of young people and women into the labor market during the 1970s. Others are shifting industry patterns (e.g., involuntary job loss in the automobile industry as a result of foreign competition), changing life styles (willingness of people to take temporary or part-time jobs for the sake of leisure time, or greater mobility), and slackening of aggregate demand. But as shown in table 2-6, the increase can be seen through both the troughs In fact, unemployment rates frequcntl~ exceeded this figure in recession years throughout th<~ centur} and from 1931 to 1910 was higher than 14 percent, with mt)rf than 20 percent of thtj l~h~i. forc~ unernplo~wd in 1932. 1 ~~~ ~;. 19;\4, itnd 1 !);l~), During the t~ar ~em-s unemplo~rnent Jf :IS \ lndc,r ? perc(>r-. t, :1 nd }Ias heen on an upwnrd ~lope since then. Table 2-6. National Unemployment Rates Dt.lring Recession Troughs and Recovery Peaks, 1961-84 I+ecesslon troughs Recovery peaks 1969 -70... ..5 80/0 1973 -75., ., .8.3 1980, ... .75 1981 -82.., ..........10.6 1961-69, ., 3,6/0 1973, ... ... .4,8 1979. ., . .6.0 1981 ., . .7,4 1984. . 7,1 SOURCE Bureau of Labor S[attstlcs data compllecl by Michael Podgursky Sources of Cecular Increases In the Unem@oymenl Rate 1969-1982 Afonth/y Labor Rev/ew July 1984 p 20
PAGE 76
Ch. ZProductivity and Emp/oyrnent 67 and high points of business cycles. Economist Michael Podgursky notes that there has been a long-term rise in involuntary job loss and argues that: rising structural unemployment in traditional segments of the labor force may also have played a significant role the secular rise in the unemployment rate since 1969 seems to have been generated by more than just labor market adjustment problems associated with a rapidly growing labor force. 68 It is worth noting at this point that Podgurskys analysis suggests that structural unemployment so far may have affected primarily blue-collar jobs. This would argue that both automation and rising imports of capital goods have played a role. Both will also increasingly affect white-collar work in the future (see chapter 8 for a discussion of off-shore sourcing of data entry work). In summary, the number of jobs has continued to increase through waves of mechanization and automation in the past; the U.S. economy was growing strongly, and an assumed major driver in this growth was technological advancement. But there are disturbing signals that structural unemployment has also grown. The Future White-Collar Labor Supply The number of office jobs is likely to grow more slowly at best, and at worst to decline, with a possible precipitous decline in lower level clerical jobs such as data entry if certain technological developments proceed as anticipated. The effect on employment levels must be considered in terms of the supply of labor or the demand for jobs. hl ichae] ~)(d~wrsk: Sources of S(,(ular 1 ncr([lses in the I nenlpio}nlent lia Ce, j ~jog. I fj~~, ,IIOn/h)l I.;~lJor Retieu. ,Jul} 19(W, p. 2 }. During the coming decade, from 1985 to 1995, the population will grow by about 10 percent. But the work force will grow about 16 percent, from 113.5 million to well over 131 million; nearly 18 million more jobs will be needed. 69 There will be fewer young workers entering the work force each year; the number of people in the work force who are under 24 will in fact decline as will the number of workers 55 and over, while the number of prime age workers, age 25-54 is growing. These changes of course reflect wide variations in the birth rate in past decades; the average age of workers will increase. About 65 percent of the workers added to the work force will be women (by 1995, they will make up at least 47 percent of the work force). The number of working women between the ages of 35 and 44 is expected to more than double, and the number between 45 and 55 should increase by nearly 60 percent. Women in these age groups who are already working are heavily concentrated in clerical occupations. This is a demographic group that will be strongly affected by the outlook for office jobs over the next 15 years. The proportion of nonwhite workers will also be growing; now 12.5 percent, they will be 14.5 percent of the work force by 1995. The number of black women in the work force, for example, will increase by over 50 percent. Since minority women are disproportionately represented in lower level clerical jobs, this is another group that will be differentially affected by office automation. A further discussion of the effects on these groups is in chapter 12. This is the middle grow~h scenario used I}s H 1.S; SWJ f;mpfo,}ment IJro.lectjons for 1995, Bulletin 2 19, I?l :+rch 19HI.
PAGE 77
68 Automation of Americas Offices CONCLUSIONS The most likely outlook appears to be slowing growth in office employment over the next decade; some decline in office employment could begin by 2000. Slowing employment growth, or even decline, is most likely to occur in clerical occupations but may also affect lower and middle management positions. This outcome is not certain or inevitable. Strong growth in the U.S. economy and continuing growth in demand for information, and information-based products and services may outweigh the labor-saving achieved through office automation. Nor would slow growth in office employment necessarily result in lower overall employment levels. Growth in other occupations could more than compensate for a decrease in office jobs, especially if higher office productivity contributes significantly to the productivity of U.S. industry and its competitiveness in world markets. The possibility of slow growth or decline in office employment, which now occupies about 45 percent of all employed Americans, is nevertheless something which Congress should watch closely, in order to take preventative or corrective actions in a timely fashion. The further possibility of a significant increase in parttime and temporary work, at the expense of full-time employment, should also be watched carefully, lest it leave a growing proportion of American workers without essential benefits, income security, and other social protections. As has been noted throughout this chapter, however, the ability of Federal policy makers to monitor technological change and its effects on employment and the structure of the economy is weak. It is limited both by inadequate data and by lack of capability in technological and economic forecasting. The latter limitation in turn, reflects in part the state of development of these disciplines themselves; however, in the civilian agencies little resources are being allocated to improving these capabilities and recent budget cuts, may have further eroded government capability for foresight and planning, at least in the important area of information and communication technology development. POLICY CONSIDERATIONS: LABOR MARKET ADJUSTMENT OPTIONS The Need for Monitoring of Structural Economic Change Related to Information Technologies While the possible long-range effects of office automation can be foreseen, they are subject to many and complex uncertainties related to broader changes in the national economy and the global economy, as well as to natural social adjustments and accommodations and to specific policy interventions. Nevertheless, the potentiality is troublesome enough to merit both careful monitoring and systematic contingency planning by responsible agencies of the Government. That kind of serious monitoring and planning is not being adequately done. Executive agencies have few incentives to warn of possible long-range problems when such warnings, or the preparatory actions they imply, may call into question immediate administration policies or the assumptions around which they are framed. Congress may, therefore, wish to consider now how such monitoring and long-range planning may be set in motion. There are serious institutional barriers to such analysis within the executive branch of Government. The first necessity for analysis
PAGE 78
of emerging and potential employment problems related to structural change in the economy is the availability of time-series data organized in appropriate categories. There are at present troublesome deficiencies in the way in which labor data is collected and organized for use by government analysts. A second necessity is the continuing development of capability for monitoring and forecasting technological change. To analyze the employment implications of technological change, there must be a close link between technical, economic, and social science knowledge and analytical expertise. There is no institutional locus in the executive branch of the Government developing excellence in the technical monitoring and forecasting of information and communication technology and studying the economic, social, and political implications, despite the central role that information and communication technology now plays in the economy and in the Government itself. On the contrary, some of the relevant but partial and fragmented functions and capabilities that have been developed along those lines have recently been curtailed or weakened by budget cuts (e.g., the planning and forecasting elements within the Institute for Computer Science and Technology in the National Bureau of Standards). Congress should therefore consider means of mandating and implementing a mechanism or governmental unit within the executive branch with the capability for systematic monitoring, analysis, and reporting of changes in the structure of the economy related to fundamental changes in the technologies of communications, computers, and information management. Longer Range Policy Options If, as it appears possible, office automation will over the long run lead to inadequate growth in demand for office work or outright decline in the number of office jobs, or in the narrower but still large category of clerical jobs, what could be done about it? The policy options discussed below are long-range options, interventions to be considered if and when it appears Ch. 2Productivity and Employment l 6 9 that white-collar unemployment is becoming a serious problem. Discouraging the spread of office automation in the United States is clearly undesirable, because of the benefits it promises in terms of productivity and in terms of the quality of work life; and discouraging it is also virtually impossible under the U.S. economic and constitutional framework. Some of the marginal effects of office automation on employment could be controlled directly. For example, Congress may wish to consider options to discourage the off-shore sourcing of office work, or of the narrower category of data entry, should this increase to the point of significantly affecting clerical jobs in the United States. Conventional kinds of policy intervention would aim at improving labor market adjustmentthat is, helping displaced workers get new jobs. These mechanisms might include a broadening of the applicability of labor market adjustment support to white-collar workers, and an increase in the level of that support. In the matrix of 1982 dollars, the money spent by the Federal Government on general employment and training programs and for the Federal Employment Service per labor force participant has fallen from $46.35 in 1970 to $30.30 in 1982, a 35 percent decrease. This is about one-quarter of the expenditures in some other industrial nations, for example, Sweden. As pointed out by an expert in labor adjustment policy: 70 Current policy takes a passive orientation toward the labor market and services only the most disadvantaged workers. What is required is a more activist policy in which structural change is anticipated and a broad segment of the labor force is assisted in adjustment. WithMi~hael-P~d~rsk~, IJni\ersit~ of hl[lss:iclll]s[~tts, I,atx)r hlarket Policy and Structural Adjustment, a paper prepared for the Conference on U.S. Industrial Polic?r and International I)e\elopment, held b~ the Overseas 13e\elopnwnt Council, J$ash ington, 1 )C, hl ar, 4, 1983, l)odg-ursk~ made this argument in the context of displacement of manufacturing workers and o\erall structural changes in the econornf and was not specifically referring to white-collar displacemcnt.
PAGE 79
70 l Automation 0f America Offices out such a change existing employment and training programs will continue to play only a marginal role in assisting workers in the mainstream of the industrial labor force who face economic hardship as a result of ongoing structural changes in the economy. (Emphasis added. ) An alternative or complementary strategy is to focus Federal programs on those office workers apt to be most directly and strongly affected, and also relatively disadvantaged in terms of current employment status. This group includes: 1 ) those in specific clerical occupations generally at lower levels of the wage scale, most of whom are women; 2) minority workers; and/or 3) all women office workers, since even in managerial and professional occupations women as a group have less seniority than men and are concentrated at the lower levels of the occupational hierarchy most likely to be affected by automation. Existing job training programs and labor exchange or employment service systems that provide labor market information are primarily framed around blue-collar employment.; 71 programs available to displaced workers in the automobile and steel industries, for example, have given relatively little attention to office workers in those industries. There may be ways to improve the quality and availability of labor market information and counseling services for office workers, with an emphasis on forecasting changes in occupational demand. working hours by taking second (part-time) jobs, which would tend to make the strategy ineffective. In effect, the standard workweek may be shortened without policy intervention if the use of part-time workers increases. This has disadvantages from a public interest viewpoint, because as discussed above it would result in a deterioration of income security, and very likely a long-range increase in the costs of necessary social services and/or an increase in the share of that burden borne directly by taxpayers. To some extent that problem might be alleviated by laws requiring the prorating of all workers benefits packages, stronger controls over conversion of employees to independent contractor status (or more stringent definition and clarification of that status), and nationwide eligibility of involuntary part-time workers for prorated unemployment benefits. This would lead to a more rational allocation of labor resources by eliminating the advantages that accrue to employers who substitute parttime workers for full-time workers not t O level the workload but to save the cost of fringe benefits. Rigorous cost-benefit studies would be necessary, however, to assesss the desirability of such policy action~: they should include cost-effectiveness studies to determine the relative advantages to employer sponsored fringe benefits publicly provided social services.
PAGE 80
What Action Is Needed N O W ?
PAGE 81
Chapter 3 Training and Education for Office Automation
PAGE 82
Contents Page Technology and Training. . . . . . . . . . . . 76 The Stakeholders . . . . . . . . . . . . 76 Basic Skills Required for Office Automation. . . . . 78 Determining Training Requirements . . . . 79 Methods for Delivery of Training . . . . 79 How Is Training Obtained? . . . . . . . . . 81 The Adequacy of Office Automation Training . . 83 Education for Office Automation Technologies 85 Teachers and Computers . 87 Curriculum Development . . . . .. 87 Access coeducation. . . . . . . . 88 Policy Considerations. . . . . . 89 Policies in Foreign Countries . . 89 Existing Legislation . . . 89 State and local Practices 91 Policy Options . . g~ Tables 7ahle No. [~~g~ 3-1. Average Costs for Training Among Selected Office-Rela~ed Occupations 77 3-2. Sources of Training Needed for Obtaining Current Job Among Representative Selected Office-Related Occupations 84 3-3. Sources of Training for Skills lmpro~ement Among Re[*rcsentative Office-Related Occupations 85 Figur e Figut-e NO. P2ge 3-1, The IJse of Computers in Elementary and Secondary School Education 86
PAGE 83
Chapter 3 Training and Education for Office Automation schools, private schools, commercial schools, and employers. hluch of this training is recreational, such as crafts, photography, and travel classes. However, much of it is also job related: for example, computer courses. There are signs that a large part of the population does not, have access to training, which creates inequities in opportunities for available jobs, The lack of coordination on a national level of information on training and education for work and on job opportunities hampers t he ability of individuals to plan careers and t O adju st to changing job markets. The benefirs for the individual changing new skills are notonly higher income and greater imcome security, but also an increase in selfesteem and con fidence. which promotes the ability to learn additional new ski11s. Training and continuous education and opportunities for acquiring office automation skills
PAGE 84
76 l Automation of America Offices TECHNOLOGY New technologies and the accompanying training needs seem to operate in cycles. The training cycle for a technology has been described as follows: l new technology introducedemployers provide extensive training and upgrading of employees because of lack of available expertise in the work force; technology becomes widely adopted and equipment is standardized-specific skills become general skills; employers lose employees to other firms; employers cease to provide general training; training is shifted out of the workplace and into the schools, and firms focus on firm-specific skills; increased demand makes it feasible for public and private schools to standardize and formalize training; and the industry using the technology, or the technology itself, declines; demand for skills contractstraining focused on replacement needs of the firm and on retraining of displaced workers. Office technologies are in the early stages of this cycle. Some employers are continuing and even increasing their expenditures on training, but with the largest proportion spent on management skills training. Some are beginning to require skills in using automated equipment as a condition for employment, as they find a more plentiful supply of already trained workers. Public and private educational systems are now offering training in office automation as well as providing the genelal education needed by office workers. U.S. companies spend $40 billion per year on further education and training of workers. z Over 21 million people participated in adult education (part-time, nondegree studies) in 1981. 3 Fifty-seven percent Patricia Flynn, The Impact of Technological Change on Jobs and Workers (Waltham, MA: Bentley College, March 1985). -Manpower Comments (Washington, DC: Scientific Manpower Commission, July-August 1985), p. 7. Digest of Education Statistics, 1983-84 (Washington, DC: U.S. Department of Education, National Center for Education Statistics, December 1983), p. 157. AND TRAINING of these courses were in formal educational institutions such as universities, vocational and trade schools, community colleges, or elementary and high schools. Twenty-seven percent of the courses were provided by business, labor, and professional organizations or government agencies. Sixty-one percent of the courses taken were job related, that is, were taken to improve skills for a current job or to get a new job. 4 The Stakeholders According to the Bureau of Labor Statistics (BLS) the work force is expected to grow by 23 percent by 1995. BLS projected growth of 1.6 percent per year between 1982 and 1990. This would slow to 1.0 percent per year between 1990 and 1995. 5 This is based on a projected population increase of 12.5 percent by 1995. 6 About 75 percent of the U.S. labor force for 2000 is already in the labor market, creating a great need for adult learning opportunities. As the baby boom generation ages, the need for teaching new skills to an adult work force has been termed an adult learning crisis. The term crisis refers to the wide gap in future skill requirements for work in automated offices and the current capacity to train adults, and was identified for example in the Leontief-Duchin employment forecast described in chapter 2. Changes in elementary and secondary education will have little immediate impact on these adult learning needs in the next two decades, since the formal educational system may reach only 25 percent of that work force that are new entrants. Traning, October 1983, pp. 54-68. Randolph Brown, Demographics of the Current and Future American Work Force, Profit Sharing, vol. 32, November 1985, pp. 5-17. The BLS projected population for 1985 was 237.5 million, The population, according to the Bureau of the Census, actually reached 238 million in May of 1985 that indicates that the total for 1985 will be somewhat higher than was projected. Lewis J. Perelman. The Learning Enterprise: Adu)t Learning, Human Capital and Economic De\elopment, The Council of State Planning Agencies, 1984, p. xv.
PAGE 85
Because of continuing technological change, office workers will have to train and retrain over a lifetime. This will have a heavy impact on the resources of both providers of training and of those who are trained. In 1981, when the average annual earnings in private industry were just over $13,000, 42 percent of the participants in adult education had incomes of over $25,000, 42 percent had incomes of $10,000 to $25,000, and 12 percent had incomes of under $10,000 (the remainder were unreported). Average costs for training in some officerelated occupations are illustrated in table 31. Additional costs to the trainees are reduced leisure, family, and personal time. These costs can be a substantial barrier, especially for those with child care and other family responsibilities. Women, who are heavily employed in clerical work and increasingly striving to move up in organizational hierarchies, have a large stake in the changes in jobs and training. Women have constituted a large proportion of the increase in the labor force in the past 15 years. Blacks have also been moving into office jobs in the past three decades, increasing their participation rate in clerical jobs from 2.7 percent in 1950 to 10.2 percent in 1984. They, too, have a strong interest in the changes occurring. Black and other groups will enter the work force at a faster rate than whites, and will account for about one-fourth of the projected increase in the labor force to the year 1995 9 Hispanics will also make up a larger Digest of Education Statistics, 1983-84, op. cit., p. 157. Monthlev I.abor Retiew (Washington, DC: U.S. Department of Labor, Bureau of Labor Statistics, November 1983), p. 3. Table 3-1 .Average Costs for Training Among Selected Office-Related Occupations Costs (dollars) Hours Occupation Public Private Public Private Ch. 3Training and Education for Office Automation 77 part of the office work force because of a higher birthrate than whites and because of high immigration rates. The group most effected in the past by discrepancies between skills required for available jobs and their own education and skills, are young workers, and especially those from inner-city minorities. More than 40 percent of the unemployed in January 1983 were under 25 years old. 10 The rate was 50 percent for black teenagers. This situation may be somewhat alleviated by the decrease in number of young workers, but the problem of lack of needed skills will continue to limit the opportunities in office work for many young minority workers. The 35 to 50 year olds who will constitute the largest group in the work force until after 2000 have completed their basic formal education. Their additional training and education will be achieved through on-the-job training or through their own efforts outside the workplace. Older workers also have a stake in changing skill requirements. There is no evidence to prove that age is directly and linearly related to performance, or to learning, but there is less incentive for organizations to offer continuing education and training to older workers as they approach early retirement age. Attitudes on the part of managers that reflect their own perceptions of a workers capabilities play a large part in determining what training is offered. While there is no evidence that intelligence, learning ability, memory, or motivation decline with age until very late in life, 12 this perception can seriously affect the kind and amount of training and retraining that is offered to older workers. Accounting ... . $488 $2,893 1,238 1,019 Business administration 39 5 3,913 1,148 1,198 Secretary, . . . 541 2,903 998 1,043 Computer programmer . 551 3,473 1,276 704 Clerk . . 507 1,870 924 785 SOURCE U S Department of Ed;catlon National Cen!er for Educa!lon Stat[stlcs Dfgest of Educ~fIon Stat6t/cs, 79831984, Washington, DC table 139 . The Emplo.vment Situation: December 1982 (Washington, DC: U.S. Department of Labor, Bureau of Labor Statistics, Jan. 7, 1983). U.S. Senate, Special Committee on Aging, The Costs of Emplo.ving Older Workers ( Washington, DC: U.S. Government Printing Office, September 1984), p. 4. -Ibid., p. 59.
PAGE 86
Basic Skills Required Office Automation for Whether skills required to work in automated offices are at a higher or lower level than those they are replacing, and whether more or less training and education will be needed by future employees, has been the subject of much debate. The basic skill requirements for all office work can be obtained through the traditional educational system. These include reading, writing, spelling, and some math proficiency. Problem solving, abstract thinking, communications, and interpersonal skills are increasingly important. But studies have concluded that one-fifth of the Nations adults do not have adequate reading and writing skills to function competently in the labor force. Some office automation equipment will foster jobs that require less skill than do manual operations, but many jobs found in an automated office require a higher degree of discretion, initiative, understanding, and creativity. A few specific skills become redundant, but many workers must cover a wider span of work activity than before, often in a shorter time span. The degree of automation achieved may determine the skill levels required. In the factory, levels of automation vary from a power tool that is hand controlled to a robot that identifies and selects appropriate actions and corrects its own performance while operating. The skills required can increase as tasks are automated but only to a certain level. 14 When the automation reaches higher levels the required skill levels can decrease, as the worker is required only to monitor the machine and respond when the machine warns that something is wrong. See Norvell Northcutt, Adult Functional Cornpetenc~, Adult Performance Levels Project, Industrial and Business Training Bureau, University of Texas, Austin, 1975. James Bright, The Relationship of Increasing Automation and Skill Requirements, Employment Impact of Technological Change, Appen&x Volume II: Technology and the American Econom.v, National Commission on Technology, Automation, and Economic Progress (cd.), Washington, DC, 1966, pp. 11-209. Most offices are now at the lower to middle levels of automation; it is likely that the skills required are more complex at this time because of the many sets of rules that need to be learned to operate the new equipment, and because of the changes in the work process and in relationships between work groups. Among clerical workers surveyed by Kelly Services, Inc., 88 percent of the 613 respondents believed that their skills were increased by automation and that this would help them obtain salary increases, even though only 30 percent of them had achieved such increases since acquiring new skills. Some experts 15 challenge the popular belief that computer training should be basic in schools in order to prepare students for the workplace, on the grounds that the higher the technology, the lower the skill level required. A BLS analysis indicates that only a small percentage of new jobs in the future will require computer literacy beyond what can be learned on the job in a few hours or days. Others challenge the assumption that jobs requiring use of a computer are automatically transformed into knowledge work. Researchers seem to be in agreement that the number of jobs available in the future requiring in-depth knowledge of computers are not a large proportion of projected new jobs. However there is also general agreement that skills will change for many jobs (particularly office jobs) and that training and retraining, probably throughout the lifetime, will be required for many workers. As lower level clerical jobs are automated and eliminated, the remaining jobs will require higher level skills. To what extent employers are willing to provide this training on the job will depend on the availability of workers who have obtained the required skills elsewhere. From the workers point of view, achieving these skills will be critically important in obtaining employ r For example, Douglas Noble, Computer Literacy and ideology, Teachers College Record, vol. 85, No. 4, summer 1984, pp. 602-614. Henry Levin, Jobs: A Changing Workforce, A Changing Education? Change, vol. 16, october 1984, pp. 32-37.
PAGE 87
ment and in making job changes throughout to be reasonably competent. However, there their lives; 81 percent of adult Americans feel are man-y jobs in insurance, banking, and elsethat additional training will be required of where, in which the automated system is an them because of changes in the workplace. ); integral part of the work process. Weeks or even months of training and practice may be Determining Training Requirement s needed before an employee is fully competent. The major factors to consider in determining training requirements are the needs of the users (their current skill level and learning needs level), the nature of the technological applications and products in a specific office, and the characteristics of the job. Users of automated office technologies include: 1 ) those currently unfamiliar with computers; 2) those who currently use computers as tools to perform specific tasks but are unaware of how the computer works; and 3) those who program, perform systems analyses, and do other work that requires understanding of how computer systems work. OTA case studies indicate that an increasing number of computer users fall between divisions 2 and 3; they are not computer professionals and they usually employ software packages developed by others; but they may occasionally write small programs to improve the computers effectiveness as a tool. The tasks to which office automation can be applied include: 1) tasks that require only a well-defined, step-by-step procedure; 2) tasks that require a limited amount of problem solving; and 3) tasks that involve analyzing and manipulating data to achieve some goal. g For some tasks in the first category, training may be briefas little as a few hours. Even for tasks in the third category, learning to use a computer as an effective tool may take relatively little time when the worker already has other expertise. For example, an economist learning to use a statistical package may require only a few days training and practice, America at Mork: The E\olving Role of Proprietary tocational I+; duration, ITT Educational Ser\ices, Inc., Indianapolis, I h, 1982. Summary of a survej of a representatikre national cross-section of more than 1,000 adults. Paul liarmon, Training: Psychology Meets Technolo~, Computer Jlorld, ~lay 2, 1983, p. 9. Ibid., p. 12. The organizational structure and environment also has a bearing on training requirements and success. Skill needs change as an employee moves up in the hierarchy of the organization. Training managers to supervise workers in an automated office environment is a different process from training workers to use the equipment. The immediate working environment depends largely on the philosophy of the organizations management. Management may or may not, for example, consult employees about the implementation of office automation, the redesign of the workflow, and training methods. Methods for Delivery of Training Research has demonstrated the importance of hands-on experience in learning office automation skills; how well people learn a new skill depends heavily on how much engaged time they spend on the learning, n although experts differ as to whether this applies to less motivated trainees to the same degree as it does to the highly motivated. Hands-on and on-thejob training assure that the trainee is engaged during the learning session. Self-teaching (and mutual learning) appears to be the most common mode of training, followed by home study. But there are some problems with self-teaching. The lack of formal, guided instruction for all employees creates an unequal knowledge base, leaving some employees at a disadvantage. For example, many workers are not allowed training time on the job, and have responsibilities that take up their time off the job, and so are prevented from learning even when they want to do so, -Raymond Nickerson, Information Technology and Ps?,chology, Third Annual Houston S~mposium (New York: Praeger Publishers, 1982), p. 203.
PAGE 88
Generally, when office automation equipment is purchased, the vendor provides a limited amount of orientation training, often in a classroom setting or through computerbased training. The users must then experiment independently to determine what applications are best for their own specific tasks. Coworkers supply additional knowledge. Key workers are often used to train others in an office and assist them in acquiring office automation skills. These internal trainers may receive formal training from the vendor or may develop expertise on their own. Then they must interpret the skills and tutor other staff members. Other office workers benefit by learning from someone who knows both the business operations and the system. Beneficial as this short cut to formal training may seem, it may not be as productive as it first appears. The internal trainers are often volunteers whose formal job descriptions do not take this role into account and no allowance is made for it in their official work schedules. 21 They must balance the importance of their training activities against the possible loss of productivity in their own assigned duties. While the key workers often enjoy their teaching role, they may or may not be good teachers and may or may not treat coworkers fairly in sharing their time and attention among them. User groups are formed by users of a technology to share information and to assist each other. Such groups are usually formed by employees and are most often managed and maintained by the users themselves. User groups are also encouraged by equipment vendors. Outside help or intervention is rare. During meetings, users take on the role of teacher, translator, trainer, problem solver, and student. Centers for learning, testing, and exchanging information (often called user or technical information centers) have been established in Tora 13ikson, Don Mankin, and Cathleen Stasz, Individual and Organizational Impact of Computer-Mediated Work: A Case Study, The Rand Corp., OTA contract report, March 1985, p. 42. many organizations, as a place that employees may go to learn about automated technologies. These user information centers may offer a variety of services includinginstructional classes, computer-based training, hardware and software testing, rating guidelines for applications and prepurchases, and educational and informational publications. The training is most often self-initiated by the worker seeking basic skills or further applications knowledge. These forms of in-house, group learning are beneficial because people feel rewarded when they meet new challenges on their own. Computer-based training (CBT) for learning office automation skills is increasingly available. There are multiple choices of off-the-shelf equipment and systems, standard sets of equipment and software programs that can quickly and easily be adjusted to a variety of requirements. Some include optical disk and video text. As more manufacturers and vendors have entered the field, the cost of CBT hardware has decreased; it is no longer a prohibitive factor in most cases. In its 1984 industry survey, Training magazine reports that 46.4 percent of all responding organizations use computers for training. 22 Computer-based training is popular because it reduces the two most commonly cited problems of training-cost and time. One expert reports that the use of CBT reduced course length and that students trained on such techniques achieved the same or a better level of performance than was achieved by those trained in the longer conventional instruction courses. A review of the literature found no evidence that the use of computer-managed instruction (CMI) or computer-assisted instruction (CAI) caused students to do less well than control groups receiving other forms of instruction within the classroom. 23 Two matters cause concern when CBT is usedthe fidelity of the training system in simulating the work enviTraining, October 1984, p. 56. S* Mildred D. Jarvis, Computer Based Training: Lessons Learned, Proceedings of the Human Factors Society 28th Annual Meeting, 1984, pp. 515-519, also her reference to Orlansky and Strings report on the cost-effectiveness of CBT in military training, 1979.
PAGE 89
Ch. 3 Training and Education for Office Automation l 8 1 Phofo cred(t D~gfta/ fqu~~rnen( CorP Computer-based training ronment, and the quality of the written documentation used in the training. The closer the learning situation follows the work setting, as opposed to merely presenting drill and practice exercises, the greater the applied learning that is acquired. Poorly written documentation prevents trainees from advancing through the stages of learning or achieving the highest level of learning possible. Home study is a growing alternative to formal classroom-based training. CAI and CM I packages are offered, for example, by the National Radio Institute (NRI), which provides technical correspondence courses. Cost is the restricting factor for home study because it is expensive to convert courses to technologyoriented modes of delivery and because the students must often purchase their own hardware and software to study at home. However, for basic education, TV-presented courses are very cost effective. While sending employees to a training seminar or conference can cost approximately $40 per hour, and a university course can cost $7.50 per hour, a TV course costs only pennies per hour. How Is Training Obtained? A survey of Fortune 1500 firms indicated that companies, when automating offices, are America at ~ork: The Management Perspecti\7e on Training for Business, IIT Educational Senices, Indianapolis, IN, 1983, p. 44.
PAGE 90
most likely to hire new employees with the required skills (44 percent), switch employees into positions requiring no additional skills (40 percent), or reduce the number of employees (40 percent). Thirty-six percent switched employees into positions that required additional training. In addition to formal education institutions, many public and private sector organizations, including unions, are involved in providing training and/or education, but demand so far outstrips supply. Some experts have questioned whether the capacity exists to respond to all training and retraining needs. Many business organizations are increasing their budgets for training. A 1983 survey of 1,821 private and government employers showed that 1984 training budget increases were predicted by 47.4 percent of the companies surveyed, but 53.5 percent actually did increase those expenditures in 1984. Although it varies considerably by industry, the dollars spent on nontechnical training such as sales and supervisory skills, outweigh the dollars spent on technical training such as word processing, programming, and electronic testing. Also, according to this survey 26 managers were much more likely to receive training than lower level employees. Midlevel managers and first line supervisors received an average of 32.5 hours of training, executives received 28.3 hours, and professionals, 27.2 hours, while administrative and secretarial employees received approximately 11 hours. Most of the companies surveyed offered both in-house and outside training with executives most likely to receive outside training and lower level employees most likely to receive in-house training. Only 31 percent of these companies engage in retraining of employees, usually lower level employees. In a survey of selected clients in Chicago, Price Waterhouse 27 found that 42 percent of .-. Training Budgets : In the Pinkand the Green, Training, october 1984, pp. 16-31. Training Magazines Industry Report, !ll-aining, October 1984. Price Waterhouse Office Automation Survey, Chicago, 1],, 1984, these firms provide on-the-job training, most often with vendor prepared documentation. The figure went up to 64 percent for the smaller firms surveyed. This agrees with the BLS survey showing that 50 to 60 percent of workers gained qualifying skills on the job. Temporary agencies anxious to increase their supply of trained workers are offering word processing training to potential employees, often by means of computer-aided instruction and simulation. A standard for basic, intermediate, and advanced skills has been developed by one agency .28 This standard requires that an operator with basic skills be able toset up the system, keyboard, create documents, make minor corrections and proof, store and file text, recall/retrieve text, and print text. Advanced operators should also be able to execute special software packages, develop graphics, write special programs, and supervise other operators. The Kelly surveys 29 found in 1982 that 52 percent of the companies surveyed developed their own training programs and 51 percent .. .- -Manpower-The Temp Agency-Launches New Approach to WP Training & Placement, Inside 14rord Processing, vol. 4, No. 6, June 1983. The Kelly Report on People in the Electronic Office (Troy, MI: Kelly Services, Inc., 1982); The Kelly Report on People in the hlectronic Office 11: How Office 1$orkers Triew A utomation (Troy, MI: Kelly Services, Inc., 1983); and The Ke]l.v Report on Peop)e in the Electronic Offke /11 (Troy, M 1: Kell~ Services, Inc., 1984). & phofo Credl( ManPower /nc
PAGE 91
trained their own personnel. Fifty-one percent used a manufacturer-developed training program and 32 percent used the manufacturers trainers. A 1983 survey showed that 34 percent of the nonmanagerial employees surveyed received their training from a vendor representative, 28 percent from a supervisor, 26 percent from a manual (self-trained), and 11 percent from an outside consultant or a class off the premises. Professionals and managers in offices are not neglected in the market for office automation training. Courses focusing on the upper level employees need for training are becoming common. Many professionals and managers learn such skills on the job, through their own efforts, by means of a manual or by just fooling around with the machine. The availability of simplified software in specific professional fields is making it easier for the professional to be self-taught, and new ways to get work done are finding their way into management training curricula, Data on the sources of training for current jobs were developed by Carey and Eck at the Bureau of Labor Statistics. Table 3-2 shows that in 1983, 70 percent of computer systems analysts and scientists received training for their current jobs in a school, mostly 4-year colleges. Fifty-seven percent of secretaries, stenographers, and typists and 22 percent of records clerks received school training. Seventeen percent of all workers received qualifying training in 4-year (or more) college programs. Professionals working in offices, such as economists, statisticians, engineers, etc., usually obtain their qualifying training in a 4-year college. Business administrators and managers are also increasingly qualifying for their jobs through college education. Employers seem to be requiring higher qualifications for many jobs as a more highly educated work force becomes available. This reduces opportunities for less educated employees to be promoted to higher level positions. hlax Care} and Alan F;ck, HOW Morkers (jet Their Training. occupational outlook Quarter?. tol, 28. wint,er 1984, pp. 3-21. Ch 3 Training and Education for Office Automation l 8 3 Formal company training programs reached 27 percent of computer systems analysts and scientists, 30 percent of operations and systems researchers and analysts, 17 percent of general office supervisors, 25 percent of insurance adjusters, examiners, and investigators, but only 4 percent of secretaries. Informal onthe-job training was received by considerably larger proportions of the occupations shown in table 3-1. About 55 percent of a sample of all workers employed in January 1983 indicated that, according to their own perceptions, they needed specific training to qualify for their current jobs. One-third of all workers had undertaken skill improvement (see table 3-3 for some typical office-related occupations) since obtaining their current jobs. Although only about 5 percent of all workers obtain training from high school vocational programs, a large proportion of these are office workers. Thirty-five percent of secretaries have received vocational training as have a large proportion of computer systems programmers and computer systems operators. Thirty-five percent of the 5.9 million enrolled in public vocational education in 1981 -82 31 studied office trades, and 23 percent of adult education participants studied business-related courses. Business/office school enrollment increased 38.1 percent between 1975 and 1981, while vocational and technical schools and institutes suffered decreasing enrollments during that time. The Adequacy of Office Automation Training A recent study found that in the last 20 years, the post-secondary education field was able to accommodate quite well to changes in demand for its services. However, secondary Occupational Projections and Training Data (~~ashington. DC: U.S. Department of I,abor. Bureau of I.al)or Statistics, 19X4). table c-l Digest of Lducation Statistics, 198.3-84, op. cit. Sue Berr~n~an, The .4djustn]ent of louth and l.duc:]tional Institutions to Techn(jlogi(>till\ Generated (hangps in .Skilf Requirements {J$-ashington, I)C: National (omission for l~;mplo~ment Polic?, Ala? 19851, p, 66.
PAGE 92
84 Automation of Americ a Offices -Cw-in-=tv=?mnlmmcclm I=mcNeml . . . . . . . . . L m .. El : a) $ ; u : c rd
PAGE 93
C/r, 3 Training and Education for Office Automatlon l 8 5 Table 3-3.Sources of Training for Skills Improvement Among Representative Office-Related Occupations Sources of trainlng (percent of occupatlonal employment) vocational education in comprehensive schools was relatively unresponsive to changes in skill requirements. This problem may be exacerbated by declining enrollments and the need to delegate scarce resources to the academic studies, for that there is greater demand in these comprehensive schools. By contrast, private and public schools that focused on vocational education are more adaptable to changes in skill requirements in the labor market. Although the adequacy of any training or retraining depends on the office, the job, and the individuals needs, pability, it is possible h c 37 38 16 9 26 26 16 7 13 6 25 13 15 25 22 16 12 16 11 25 11 19 12 23 23 26 & 8 7 8 2 7 1 2 3 3 2 6 5 4 terms other factors influencing the quality of training. The quality is dependent on the quality of the instructional materials and design, the quality and availability of instructors, and the range of courses that are offered. There is currently no legislation to regulate, assess, or accredit the content or the quality of the courses offered in the private and commercial sectors. Those who invest in such training are at the mercy of the market. This lack of quality control can be expensive and can drain the resources of individuals and organicommitment, and cazations. to discuss in general EDUCATION FOR OFFICE AUTOMATION TECHNOLOGIES Computers have assumed such an important find themselves pressed for expanding the curfunction in the contemporary practice of busiriculum: ness, industry, science, and scholarship that almost no student can expect to remain isol Iated from these tools. This imposes on education the additional task of preparing students for jobs in which they will use the computer as a partner. 34 Schools and colleges l C-harles ~lossman, ,4ssociate Vice President, Academic Resource I]lanning, California State [Jni\ersit~. Fullerton, in joint hearings hefore the L]. S. Congr-ess, I louse, Information Technology in F;ducation, Apr. 2-3, 1980, p. 145 by students many of whom expec t schools to provide them with access to computing and instruction in computerrelated subjects, and by the job market-because employers expect applicants to have some basic computer literacy before they arrive on the job.
PAGE 94
86 l Automation of Americas Offices According to a recent national study, computers in high schools are used two-thirds of the time to teach children how to use a computer (computer literacy), and 18 percent of the time for drills and practice in various subject matter. Results of a survey of 1,082 elementary and secondary schools, conducted by the Center for the Social Organization of Schools, the Johns Hopkins University, are shown in figure 3-1. It appears that relatively little is being done to exploit the computers potential for teaching traditional subject matter in a more efficient, interesting, or effective manner. Also, time is not spent on the computer to teach and develop office automation skills for future jobs. Computer-assisted instruction (CAI)the computer as teacher has been introduced to improve the delivery and productivity of general education. CAI has been found to reduce by 10 to 30 percent the amount of time students need to master a subject. 35 CAI supplements the teachers efforts in the traditional classroom setting and provides specialized individual instruction. Education Turnkey Systems, Inc., Uses of Computers in Education, prepared for the National Commission for Employment Policy, April 1985, p. 42 Figure 3-1 .The Use of Computers in Elementary and Secondary School Education 70 r 6A 60 50 40 30 20 10 0 L Literacy 40 [ Drill 18 L 24 Game s Other Functions q Elementary = Secondary SOURCE Flqure derived from School Uses of Microcomputers ReDorts Fr~m a National Survey (working paper) Center for Social Organl zatlon of Schools The Johns Hopktns Unlverslty No 2, June 1983 The effectiveness of CA I is hindered by the lack of high-quality software. The Educational Products Information Exchange Institute, in cooperation with the Consumer Union, evaluated 600 pieces of educational software in 1984 and rated only 5 percent of what was examined, or 30 programs, as first-rate. 36 It takes about 200 person-hours to create 1 hour of conventional CAI. Work is being done at the Advanced Computer Tutoring Project, Carnegie-Mellon University, Pittsburgh, Pennsylvania, to reduce the time required to develop CAI. 37 Attempts are also being made to develop intelligent CAI, or ICAI, which allows for conceptual modeling and interactive instruction. Computer-managed instruction (CMI) is an automated technology used for the management of the school system and for managing the instructional flow in the classroom. It aids both the teacher and student by providing preand post-tests to evaluate advances in learning levels, offering diagnostics, helping students with assignments, and keeping student records. The teacher need not be present for all testing, assigning, and recordkeeping. Cognitive diagnostics for learners can also be provided by computer programs. The hardware available for educational uses and teaching office automation skills has focused on using microcomputers. Video disks in the future may play an important role because they allow for greater interaction in instruction by simulating actual situations. Although information on educational software, evaluations, and availability is published, the price of these analyses limits access to this information for many educators. Most school districts lack the means to identify better software, and in most schools only a handful of teachers have the training to make effective use of computers. 38 h Edward B. Fiske, Computers, In Most Schools, Have Brought No Revolution, /Vew York Times, Dec. 9, 1984, p. 80. John R. Anderson, C. Franklin Boyle, and Brian J. Reiser, Intelligent Tutoring Systems, Science, Apr. 26, 1985, p. 228. Fiske, op. cit., p. 1.
PAGE 95
Ch. 3Training and Education for Office Automation l 8 7 Teachers and Computers The proportion of college freshmen choosing education as a major has declined from 22 percent in 1966 to 4.7 percent in 1982, 39 because of the job disincentives common to the teaching profession (low status, low pay, low-advancement opportunity, and diminished psychic income). Teachers are being pursued by industry to fill training positions and they are finding these jobs more rewarding than teaching. The difficulties of competing for manpower and other resources in a rapidly growing hightechnology market are acutely felt by educators at the university level. Over the decade from 1970-71 to 1981-82, B.A. degrees in the computer sciences increased from 2,388 to 20,267, nearly 750 percent. Yet a large portion of faculty positions are unfilled and the number of Ph. D.s graduating each year has dropped substantially. The shortage of faculty in the field of engineering and computer science has been attributed to the fact that industry, by offering higher salaries and other incentives, has been able to draw academics and students away from universities. 40 The percentage of computer science faculty leaving for industry is twice that of any other field of engineering. Curriculum Developmen t Efforts have begun to update curricula and include more courses in the use of computers, but office automation and other technologyrelated subjects are caught in the lag between need, development, and implementation of new curricula. This is especially obvious at the secondary and post-secondary levels of education, where these skills are most often learned. The rapid pace of technological change has complicated the delivery of appropriate courses. In a report on office automation productivCenter for Strategic and international Studies, 7ec/mical Excellence in America: Incentives for Investment in Human Capital, Debra van Opstal (cd. ) (Washington, DC: Georgetown University, October 1984), p, 1. U.S. Department of Education, National Center for Educational Statistics, Condition of Education, 1984, table 2-12. ity, 4 Russell Aldrich of Apple Computer, who teaches at Golden Gate University in San Francisco, says that: Many times, a particular course was a good course, and a required course 5 to 10 years ago. But because technology is progressing rapidly, we are seeing so large a revolution in certain areas of office productivity, communications process, and management roles, that what was on the frontier 2 years ago is no longer relevant today. What is lacking is a mechanism for higher education institutions to quickly and accurately evaluate the information industry pulse so that it can design and plan curriculum that anticipates educatiomd needs and fills to them.42 (Emphasis added.) The Office Systems Research Association (OSRA), in a cooperative effort between educators and business sector representatives, is working to develop a model curriculum in office systems for universities and colleges. Office systems is defined as the business function related to the coordination and management of the information resources of an organization. Generally, this includes responsibilities for automated and manual office equipment, human factors, and office procedures. According to OSRA, managers in the office systems area are responsible for a business unit too complex to rely strictly on a computer background or management techniques applicable to the traditional office environment. The Office Systems Model Curriculum Project was begun in 1984 and is aimed at providing a framework, and possible standardization, of office systems curricula for all schools. OSRA plans to have a draft of their model ready in mid-1 985. 43 Frank Freudberg, Office Automation Producti\it~r: Lost En Route to the Promised Land (V1illow Gro\e, PA: Association of Information Systems Professionals, 1984), p. 14. -Freudberg, op. cit., p. 15. {This information is based on correspondence and conversations with Bridget OConnor, OSRA Vice President, Professional Studies, Business Education Program, New York Uni\ersity, March 1985.
PAGE 96
88 l Automation of America Offices Access to Education To the extent that computer literacy and computer expertise are needed for success in getting and keeping jobs, inequities in receiving computer experiences in school are especially serious for low-income, female, and rural students. They translate into social and economic inequities by giving some people more effective tools for working and living in an age of information technologies than others. acy associated with wealth of the school, community size, region, gender, and race. Student enrollment in computer programming is much lower in schools that qualify for Title I assistance (by having a large percentage of the parents with income below the poverty line) than in schools that do not qualify. After a survey of schools in 1983, Quality Education Data, Inc., reports that the 12,000 wealthiest schools are four times as likely to have microcomputers as are the 12,000 poorest schools. Rural and disadvantaged urban communities provide computer learning opportunities at a much lower rate than other communities 45 Among students at age 13, less than 17 percent of the rural/ghetto students, but 32 percent of those living in urban/rich areas, reported use of computers in schools. Eighteen percent of junior high school students in small towns report school computer use, compared with 26 percent in large cities. Each year the National Assessment of Educational Progress (NAEP) surveys 44 a stratified random sample of students aged 9, 13, and 17 in approximately 700 schools in the United States. Data from the 1982 survey show inequities in access to computer literStudents living in the South are much less likely to have used computers in schools than students living in other parts of the country. Those in the Western States are twice as likely as southern students to receive such experience. Young women in secondary schools are less likely than young men to spend time with computers and to enroll in computer classes. 46 Females are less likely to take computer programming classes than males; one study in 1983 showed that 8 percent of the females and 14 percent of the males have enrolled in programming courses for at least one semester. The Womens Action Alliance of New York City has devised several school-based strategies to overcome this problem, working with parents and teachers to increase girls use of computers. The results after a year of trial at a i These and the following figures are based on the work of Ronald E. Anderson, Wayne W. Welch, and I.inda J. Harris, Computer Inequities in Opportunities for Computer Literacov, University of Minnesota, based on work supported by the National Science Foundation under grant SED 8022125Ao I, 1983, p. 4. As shown in studies by Anderson, Welch, and Harris, op. cit., 1983.
PAGE 97
Ch 3Training and Educat/on for Office AutomatIon l 8 9 middle school in Wisconsin, showed a 42-percent increase in computer usage by girls. 4T While the NAEP data showed some racial inequity in computer exposure in 1978, more recent results reveal no significant difference between black and white students when income differences are equated. Apparent inequity between black and white students in use of computers in schools and for enrollment in computer programming courses is derivative of income and rural/urban differences. Inequities in computer access identified here point to the need for attention at all levels of the educational system: National, State, community, district, school, and the classroom. POLICY CONSIDERATIONS Policies in Foreign Countries In some countries, training and retraining programs are used as a basic instrument to deal with specific labor market problems. The successes and failures of foreign experiences are rich in lessons for shaping U.S. policy. The European and Japanese emphasis is on strong vocational education as a basis for work life and on periodic training to keep skills up to date. Their philosophy is that these programs serve not only low-wage workers, but a broad segment of the primary labor force. 48 However, the degree of government involvement differs among these countries. In France, the Law for Continuous Training, implemented in 1971, committed the government to provide training opportunities to adults and youth. In Sweden, where workers are considered to have a right to training, the government plays a strong role in adult training and retraining. A close coordination among government, employers, and unions helps ensure that government programs mesh well with national needs. The flexibility of the system in identifying new jobs and in retraining workers has reportedly added to the success in Sweden. Union involvement and training vouchers encourage wide participation in West Germany; the government and private industry both play hlichael l)odgursk~, I,ahor Jlarket lolic~ and Structural Adjustment, paper prepared for the conference on [1. S, Industrial lolic~ and International Development, ()~rerseas I)e\elopment (ouncil, \fashington, 1)(2, NI ar. 4, 19/+3, p. 17. major roles. In Japan almost all efforts are by private companies. Japan continually retrains selected workers in the lifetime employment system. Increasing labor-market transparency or skill transferability, providing job referral, counseling, testing, training, and relocation assistance to workers, are regarded as a matter of high public interest and the public employment service plays a major role in labormarket adjustment policy in Europe and Japan, Existing Legislation This section will present a summary of legislation and regulations related to education and training for automated office work and some options for congressional consideration. The potential for successful delivery of education for office automation skills and knowledge already exists in the traditional education system, most often addressed at the secondary and post-secondary level, where much of the Federal legislation is focused. Educational policies and legislation set the framework for addressing new and changing education needs. The higher education delivery system in the United States is a complex matrix of private and public institutions that function with varying degrees of independence and dependence on State agencies. The Federal Government direct role in institutional control has been limited to setting criteria for an institution to participate in Federal programs or receive Ibid. p. 26.
PAGE 98
90 l Automation of Americas Offices Federal contracts. Various congressional actions have been taken to draw attention to selected national problems, to provide the fiscal resources needed to address these problems, to support research activities that have national and international implications, and to complement and supplement the role of parents and State and local governments in supporting individuals and institutions. Although there is no existing Federal legislation specifically addressing office automation education, there are related laws that could provide a vehicle for delivery of office automation education. Direct Federal involvement in training and employment programs is considered to have begun with the Area Redevelopment Act (ARA) of 1961, although indirect Federal involvement in training through vocational education programs began with the Smith-Hughes Act of 1917. The Manpower Development and Training Act (MDTA), enacted in 1962, focused training efforts on workers displaced due to automation. The Johnson Administrations war on poverty brought about a wide range of work experience and training programs targeted on the poor, minorities, and youth. The Comprehensive Employment and Training Act (CETA), passed in 1973, absorbed many existing work and training programs and was designed to operate primarily at the local level. 50 CETA underwent amendments during its history that expanded its purpose and reach. The Trade Adjustment Assistance (TAA) program reauthorized in 1981, also authorized funds for training. The Job Training Partnership Act (JTPA) replaced all portions of CETA on September 30, 1982. This new program shifted the focus away from direct Federal involvement in training programs and established a business-government partnership in the provision of training for job skills. Training programs are operated by the States in combination with local-area governments and Private Industry Councils (PICS) and provide U.S. Congress, Library of Congress, Congressional Research Service, Job Training Programs: Reauthorization and Funding Issues, by Karen Spar, Issue Brief No. IB82005, Feb. 8, 1982. for the training of unemployed displaced workers in skills relevant to real employment opportunities in the area. These programs were not primarily aimed at office workers or at office automation training. Target groups for legislated training programs have traditionally included about 8 percent of the population. These groups include low-skilled adults (especially women and minorities), disadvantaged youth, and residents of economically depressed areas. The gaps in public policy left by JTPA and similar programs relate to: 1) the retraining needs of the large-middle tier of employed but at-risk workers; and 2) the basic skill needs of the 20 million or so functionally illiterate adults in the work force who will not be touched by reforms in elementary/secondary education, and whose learning handicaps prevent them from benefiting from job-specific training. The Vocational Education Act (Perkins Act), as revised in 1984, provides focus on educational needs at the secondary and post-secondary school level. The Perkins Act grants funds to States to make vocational education programs accessible to all persons, but follows the current trend of placing greater control and responsibility at the State and local levels. Targeted groups include: l handicapped and disadvantaged persons, l single parents and homemakers, l adults in need of training and retraining, persons in programs designed to eliminate sex bias and stereotyping in vocational education, and l persons incarcerated. The act is designed to improve the quality of vocational education programs in order to give the Nations work force the marketable skills needed to improve productivity and promote economic growth. Under this act, the acquisition of office automation skills may be included in State programs for general vocational opportunities. Part E of Title III, Industry-Education Partnership for Training in High-Technology Occupations, allows for .. Perelman, op cit., p. 27.
PAGE 99
Ch. 3 Training and Education for Office Automation l 91 grants to States to provide incentives for business and industry and the vocational education community to develop training programs for high-technology equipment, systems, and processes. These programs are intended to be closely tied to the local labor market and skill needs. The act states that to the maximum extent practicable, funds will be utilized in coordination with JTPA to avoid duplication of effort and to ensure maximum effective utilization of funds under both acts. While the goals of the two acts are similar, the manner of achievement is different. The focus of the Perkins Act is public vocational training and education, while the focus of JTPA is on industry-sponsored training. The two acts overlap where there is the greatest need for assistance. Funding for these programs is the key to their success. However, the level of Federal funding for employment and training decreased by 49.7 percent between 1981 and 1984. Funding for elementary, secondary, and vocational education decreased by 8.9 percent during that period. There were some increases for fiscal year 1985 17.5 percent for elementary, secondary, and vocational education; and 14 percent for training and employment; but decreases are again projected for fiscal years 1986 through 1990 in training and employment functions. 52 Federal aid to State and local governments for vocational and adult education decreased by less than one-half percent between 1981 and 1984, but is projected to increase by 23.7 percent in 1986. Federal aid to State and local governments for employment and training services was decreased by nearly 57 percent between 1981 and 1984, and is projected to increase by only about 13 percent by 1990, with some fluctuations. If this trend continues, State implementation of this act will not be effective. 53 The Employee Education Act amends the Internal Revenue Code to extend for 2 years ~:xecutive office of the President, Office of hlanagement and Budget, Historical Tables. Budget of the U..!. (;o~rernment, Fiscal Year 1986, tahle 3-3 (19) and (20), 1985. I bid., table 12-3 (431 and (52). the income tax exclusion for amounts received by an employee under a qualified employer provided assistance program. Unless extended by Congress, this act will expire at the end of 1985. The act limits the exclusion to $5,000 of educational assistance furnished to an individual during a calendar year. State and Local Practices Typically, State funding and support of education has been related to economic development. Several private sector industries and research and development firms have recently moved toward more direct involvement with higher education institutions by starting joint education ventures. While joint ventures are encouraged by some Federal laws, many businesses have traditionally contracted locally for training and education programs for their employees. States have assumed a greater role during the past few years, as they enact student assistance programs to supplement declining Federal support and shoulder the responsibility of Federal programs. As State policy makers have been confronted simultaneously with revenue declines and requests for additional funds for higher education as well as for other human resource investments, some have taken the unpopular course of raising taxes to provide additional support. Others have reduced real levels of support for higher education in general and for research universities in particular. The funding pattern has been unique to each State. Some States are making strong efforts to keep up with the new teacher and curriculum needs. For example, the State of Minnesota has made new office automation skills and computer knowledge a high priority for vocational education programs. Efforts are being made first to increase the competency of teachers through conferences and workshops on hightechnology equipment, office systems, telecommunications, and curricula design. An integrated office system has been installed for teachers hands-on experience. The system provides administrative assistance for teachers.
PAGE 100
92 l Automation of Americas Offices The States vocational program will also include more opportunities for the adult and part-time student communities. Another example of State activity is in Texas, where House Bill 246 will be implemented in all school districts in the 1985-86 school year. This law mandates computer literacy courses for middle school (seventh and eighth grade) students and computer science classes for all high school students, as requirements for graduation. The computer literacy course includes history, terminology, basic programming skills, and issues of computer use in the society. Computer-related course credits are not required at the elementary school level, but many school districts are providing some type of computer awareness at that level. Policy Options Federal actions that could encourage education and training opportunities related to office automation might include: l l l l l l l designate funds specifically for office automation education programs through new legislation or existing vocational education legislation that provides aid to schools; increase funds for programs through existing legislation (e.g., the Perkins Act); promote access for students in low-income families and in rural areas, and for women, through direct funding and establishment of service programs; increase Federal attention to maintaining the capacity and quality of research in the Nations universities and the flow of new talent into academia; provide career counseling, job guidance programs, job search assistance and training programs that do not penalize those receiving unemployment insurance; encourage ties with the private sector through cooperative educational efforts, curriculum development, and industrybased activities for office automation skills; establish support programs or tax credits for small businesses to ensure that the education of employees is adequate to maintain current levels of employment and productivity; provide training grants and tax credits to businesses that incur expenses for office automation training. Also provide incentives to establish training programs in office automation skills, including employment tax credits for employers who train and educate workers. Employers could be given the same kind of tax writeoffs for training that they get for plant modernization; establish a program of Individual Training Accounts (ITAs), similar to the Individual Retirement Accounts. The ITA may require employers and employees to make equal contributions to a bank account; this money would not be taxed until withdrawn. The account could be drawn by employees to pay for retraining when needed. This would direct resources toward retraining needs of mainstream workers now being neglected. A key objection to this proposal is that workers may not be able to choose and direct their own retraining in the best way, or appropriate programs may not be available; expand the tax deductibility of training for current and new occupations related to office automation; change unemployment regulations to permit receiving unemployment benefits during training. Encourage States to consider those in training programs as still available for work, and to reserve parts of their State funds to establish permanent funds for retraining; and encourage union involvement in negotiations for skill training opportunities. Include requirements that firms give notice of plant closings in advance and aid in the retraining of displaced workers. Congress may determine to take none of the above actions, making the conscious choice to preserve the status quo. Since there is no existing regulation of office automation training, the training available commercially is likely to grow, but with existing inequities in access and quality unchecked.
PAGE 101
Chapter 4 The Changing Natur e of Office Wor k
PAGE 102
Contents Case Studies . . . . . . . . . . . . . . . The Individual and the Work Process . . . . . ........ . . The Nature of Office Work . . . . . . . . . . . . The Work Process . . . . . . . . . . . . . . The Information Age Office . . . . . . . . . . . . Task Change and Computer-Mediated Work . . . . . . . . Job Changes With Automation . . . . . . . . . . . Management and Professional Jobs . . . . . . . . . . Clerical Jobs..... . . . . . . . . . . . . . . Job Ladders and Mobility . + . . . . . . . . .......+. Organizational Structure and relationships . . . . . . . . . Power and Access to Information.. . . . . . . . . . . End-User Computing Power . . . . . . . . . . . Communication . . . . . . . . . ........+ .....4... Dispersion of Work Activity . . . . . . . . . . . Implementation . . . . . . . . . . . . . . . Reason for Adoption . . . . . . . . . . . . . Key Actors . . . . . . . . . . . . . . . Adaptive Planning . . . . . . . . . . . . . Users in the Implementation Process . . . . . . . . . . Figures 95 96 96 97 101 103 104 105 107 109 111 111 114 114 116 117 118 118 119 119 Figure No. Page 4-1. Possible Variations of Work Process for Word Processing . . . . . 98 4-2. Model of an Integrated Customer Service System . . . . . . 102
PAGE 103
Chapter 4 The Changing Nature of Office Work Translating the capabilities of office automation technology into tangible operating benefits for an organization illustrates effects at many levelsindividual tasks, the work process, organizational structure and culture, and quality of working life. These effects are at the core of the most interesting and pressing issues related to office automation. Yet, they are difficult to address because it is not possible to give a simple or universal description of the effects of automation on workers, jobs, or organizations. This chapter looks primarily at the effects of office automation on the nature of office work, on job content, and on organizations. It examines the effects of office automation on the work process; on specific tasks, skills, and jobs; and on promotional opportunities. Next it discusses the possibility of structural or cultural change in organizations that adopt office automation. Finally, it reviews some of the major factors that contribute to successful implementation of office automation. One point to be emphasized throughout this chapter is that the technology itself is not the only factorand may not be the most important factorin bringing about these changes. Managerial strategies-decisions about the organizations goals and the role of people and technology in achieving themare of major importance in determining how the technology is used. In many cases, it can be seen that outcomes are more dependent on how the technology is used rather than on what specific equipment is employed. CASE STUDIES Many of the examples used in this chapter were taken from case studies of offices using automated equipment. In addition to the case studies in the literature, OTA commissioned a number of small case studies, discussed in more detail in appendix B. Case studies provide a rich source of detailed information about the process of implementing office automation and the possible changes in tasks, jobs, work processes, and organizations. They allow one to view the intended and unintended consequences at many levels, and in a variety of contexts. In this chapter, case studies are used to provide illustrative details about the effects of technology in a wide variety of organizations. These observations form the basis of some cautious generalizations about the effects of office automation on organizations. Caution is needed in making use of case material for a number of reasons. First, the methodology and level of detail of published case studies vary widely. Some are highly quantitative, others depend largely on qualitative, participant-observer, or anecdotal information. Second, the organizations studied are all unique. Authority structure, corporate culture, management philosophy, internal dynamics, financial health, and operating environment are different. To the extent that these factors of themselves cause certain outcomes, or moderate the effects of technological change, it is difficult to state authoritatively that the result observed at one study site can be expected at other locations. Third, the technology itself is defined differently from one organization to another and from one study to another. Office automation in one organization may be word processing capability used almost as a direct substitution for typewriters; in another it may be an extensive multiuser, multifunction operational system governing the production of the organizations primary product. Many possible systems and combinations of functions, includ95
PAGE 104
96 l Automation of Americas Offices ing text processing, messaging and communication, decision-support software, graphics, and numerous special purpose applications may have a separate effect on the individual, the job, or the organization. The choice of features, the way the work process is organized to use them, and the order and timing of system introduction, often make each case of office automation unique. Finally, the time factor must be considered. User reactions and use of technology in a honeymoon phase shortly after the equipment is up and running may be different from what happens at a later date when new capabilities or limitations are fully understood. Some changes in the work process may take place immediately after implementation, while others may not become evident for some time. Thus, the findings of a case study may depend on where the organization was in the system life cycle when the study was carried out. There are few longitudinal studies that follow the same set of organizations over time. Even within one study it can be difficult to compare before with after in an environment where new technology is introduced gradually over a period of months or years, and where systems are continually being upgraded or expanded. Despite these difficulties, case studies are often the only information available, and they capture the detail necessary for understanding dynamic change within the organizations. THE INDIVIDUAL AND THE WORK PROCESS The Nature of Office Work Office automation has transformed office work. Many offices are now semiautomated environments where people and computers interact. An office, in the sense of a working unit of people doing information processing work, is seldom a totally automated environment. A fully automated procedure is one that does not require human intervention in order to produce its final output. Automatic data processing falls into this category; all transformation of information is internal to the computer. But such fully automated procedures usually make up only part of the work. Office work usually involves a series of steps, some are fully automated, some are manual, and some require interaction between people and computers. For the purpose of this chapter, a task will be considered any clearly definable activity that forms a step in producing the final product of the office. Tasks may be performed by people or by machines. The work process is the social and technical organization of work. It is the way people and machines are organized to produce a resultthe way information or materials flow from one to another until the final product is completed. Jobs are organizationally defined positions that are usually associated with a bundle of tasks and a particular role with a defined set of responsibilities in the work process. The work unit or office is a group of people of any size, with responsibility for producing some identifiable final product. The work unit has a skill and task mix, that is, its members are at different levels, have different roles, and perform a range of functions to produce their product. Skills are the attributes and knowledge that workers need to perform useful activities. The impact of office automation on office work goes beyond changes in tasks. It can introduce new tasks, change the nature of the skills required, modify the work process, and ultimately can cause or be associated with changes in jobs and in organizational culture and structure. This is at least in part because computer-based technology bundles tasks in a way that is different from the way they would be done manually. When part of a process becomes automated, the tasks that remain to be done by people may be different kinds of tasks. A new process is necessary for integrating the work done by people and the work done by machines to accommodate those differ-
PAGE 105
ences. When the work process changes, the organizational structure, which defines the roles and responsibilities of the members, can also change. Take, for example, an office where a professional researcher writes reports and a secretary types them. Introducing a word processor can change the work process in a number of ways. First, the nature of some tasks will change, and the new tasks will require different skills. If the professional decides to use the word processor, then composing on a keyboard will require skills quite different from those of writing by hand. Some skills needed for old taskslike the secretarys ability to neatly paste up final copiesmay seldom be used. Further, a change may take place in the process that governs the interactions of author, secretary, and other people needed to produce the report. Partially completed drafts may no longer pass between the author and the secretary. The secretary may be left out of the revision procedure entirely, or at least the secretarys responsibilities may no longer include the task of typing the authors reports. There are also some points where the technology does not dictate a change, but allows an opportunity for choice. The word processor could be given to the secretary instead of to the author. The author might still hand write drafts, with the secretary keying in the document and making corrections on the word processor. In this case, the word processor is more like a simple substitute for the secretarys typewriter. The relationship of author and secretary, and the flow of work between them, might have been altered only a little, although the secretary will have to learn new skills. In another scenario, that relationship might be severed completely. The secretary might be reclassified as a word processing operator and moved to a newly created pool to key and correct drafts for many authors, as part of a change in the structure of the organization. The use of new technology also introduces new steps and new tasks. For example, use of an automated system introduces the need Ch, 4 The Changing Nature of OffIce Work l 97 to make backup copies, design and use an electronic filing system, and manage disk space or diskette storage. Decisions about whether these tasks will be handled by all users, by secretaries, by a system administrator, or by some combination will affect the work process. There is also the possibility that the availability of the technology will catalyze its own use. In the experience of many authors, the ease of editing a document on a word processor sometimes leads to additional rewriting, so that this step can sometimes seem to be repeated almost indefinitely. Thus, even with this simple example it is possible to see how changes at the level of the task, job, work process, and organization could take place as a result of office automation. The Work Process The bundle of tasks an individual performs describes the job, but the job is more than merely the set of tasks. The job is an organizationally defined position. In many organizations it is characterized by a job description that legitimizes the position. People identify with their jobs. The job defines a role within the work process and within the social structure; it is the point of articulation between the individual, the technology, and the organization. Thus it is not sufficient to talk about how office automation affects individual tasks. Jobs also change-the bundle of tasks, the role in the work process, the position in the organization, and the self perception. Division of labor is necessary when there is too much work for one person to handle. Once a division of labor is made, the work process governs the relationship of the various workers to the partially completed product. One way of looking at work process is as a continuum that runs from the most integrated to the most differentiated. Figure 4I shows five possible steps along this continuum for a word processing office. The most integrated is the one where authors do their own word processing and have complete responsibility for all steps in the process and control
PAGE 106
98 Automation ot America Offices Figure 4.1 Possible Variations of Work Process for Word Processing Integration of work process Group of authors, one operator ~ Team of authors, team of operators Many authors, centralized WP pool with task specialization Differentiation of work process SOURCE Adapted from S M Pomfret, C W Olphert, and K D Eason, Work Organ tzatlon Impllcatlons of Word Process ing, Proceed/rigs of the Ist /F/P Con/ererrce on f-/urnan-Corn~uter /n/eract/on (/N TERACT A VOI 2, 1984, pp. 357-363
PAGE 107
Ch. 4The Changing Nature of Office Work l 9 9 over the end product. The most differentiated is one where authors send work to a centralized word processing pool where operators have specialized taskssome key in text, some proofread, some do corrections, etc. Between these extremes might be one-on-one relationships between author and operator, or situations where one operator or a team of operators focus on all stages of work for a particular group of authors. The basic concept of work integration or differentiation can apply to a large extent regardless of the level of technology employed. The example above works equally well if the technology used is typewriters or quill pens. There is a popular perception that advancing technology is associated with increasing differentiation and fragmentation of work. The manufacturing assembly line, with workers performing repetitive tasks, is a product of the industrial revolution, not of earlier craft manufacturing. Many peoples only image of office automation is a picture of the centralized word processing or data-entry pool. Although there has been a connection between computerization and fragmentation of work in the past, other forces are also at work. Office automation technology offers the possibility of reintegrating the work process in some cases. Another way of looking at the work process and the way it changes with automation is in terms of the three generic types of work processes in offices identified by Giulianopreindustrial, industrial, and information age. ] The preindustrial form of organization is used by many small and medium sized offices, for example, a real estate brokerage or professional office. These work units may have a number of workers with different jobs, or roles, in the work process. However, each worker is to some extent a craftsman. Each works with some independence and performs a variety of tasks. There has been little effort to completely standardize or systemize the work process. A variety of individual work styles is tolerated. lin~>~nt l;. (li~llimt), The ~le~h~niz~ti~n of office Work, Scientific ,4merican, September 1982, pp. 149-164. There may be some fuzziness or overlap in responsibilities and workers may be sufficiently familiar with the work of others to switch tasks occasionally, or take over in an emergency. While the preindustrial style of organization works well in many contexts, it is often inefficient in handling large numbers of transactions. Thus, in the insurance industry, banking, or the billing departments of large firms, where high transaction volume is handled, industrial-style offices evolved. They are designed to organize people to serve the needs of a large, rigid production system. The industrial office is a production line. Workers are differentiated into functional groups, for example-typists, log-in clerks, validation clerks, and signature control clerks. Each group has its own supervisor, and is responsible for some step in the processing of a transaction. Documents related to customer transactions flow from one functional area to the next, from one out box to the next in box, receiving some incremental processing at each stop. The flow of work in an industrial office is consciously designed according to principles of scientific management first articulated for manufacturing by Fredrick W. Taylor in the early 1900s. With Taylorization, a complex production process is analyzed and divided into a series of simple tasks that can each be performed quickly and efficiently. workers are assigned to perform a single task, or a narrow range of tasks, in a routine and repetitive way. The industrial form of work organization is a deliberate attempt to increase efficiency by rationalizing the work process, and by reducing individual discretion and variation. Because industrial-style offices were ones in which the work process had been consciously analyzed and the individual tasks identified, isolated, and standardized, they were ideal candidates for the early introduction of computers. Certain tasks could be automated completely, making use of the computers ability to do large batches of calculations quickly. However, many manual tasks remained in preparing data for the computer or making use of computer output.
PAGE 108
100 l Automation of America Offices Some have argued that computerization allows managers to speed the process of Taylorizing and routinizing office work in order to reduce costs and increase management control. Certainly, as automation began to be used in offices, it appeared that Taylorization and computerization were mutually complementary and interdependent processes, as computers were first introduced in those areas where manual processes had already been industrialized. Indeed, computerization seems to have intensified the factory-like nature of industrialstyle offices. Pools of functionally similar workers existed before, but the trend toward specialization increased with computerization, especially in routine data-entry tasks: Routine keyboarding was separated more sharply from other clerical functions and was often spatially isolated As a result of the heightened fragmentation of work, processing personnel (both data processing and word processing) typically worked at machines all day. both the technical and social relations of work in these centralized word processing and data processing centers were factory-like. The work process was machine paced and often machine supervised; autonomy of the operator was minimal; competence measured by manual dexterity and speed. 4 Taylorization is not limited to clerical processing functions. It isolates predictable tasks. Professional or managerial jobs can be analyzed and the more routine functions stripped away. Baran and Teegarden note that underwriting, the main professional occupation in the insurance industry, has been increasingly rationalized. At one firm, this was done by splitting off the lower level functions and creating a new clerical position called underwriting technical assistant to perform them. Remaining professional positions were divided into specialty categories o 5 For example, see Harry 13raverman, Labor and Monopolv Capital: The Degradation of Work in the Twentieth Centur~ (New York: Monthly Review Press, 1974), pp. 293-358. Barbara Baran and Suzanne Teegarden, Womens Labor in the Office of the Future: Changes in the Occupational Structure of the Insurance Industry, Department of City and Regional Planning, University of California, Berkeley, 1983, p. 11. II bid. Ibid. Some offices are in factories. Note that some of these clerical workers are protecting their ears from excess noise. Photo credit M/chae/ Srmth Some offices where large numbers of transactions are handled resemble factor~es, The industrial style of office organization works well for highly routine processes. It has some disadvantages, in dealing with exceptions or nonroutine situations, and is usually not efficient for customized services. Some experts also note that it is susceptible to possible long-term growth of overhead costs. For example, in order to deal with error detection and correction it maybe necessary to add more steps, thus making the process longer and more unwieldy. Generally, no one employee has all the information needed to correct an error, so mistakes may persist or build on one another; papers might cycle through the system several times before they are corrected.
PAGE 109
Ch. 4 The Changing Nature 0f Office Work 101 It can become difficult to provide quick, accurate customer service when work is processed in a production line fashion. The process of issuing an international letter of credit at one bank required 23 steps performed by 14 workers over a 3-day period, and generated stubs and carbon copies to be stored in a number of locations. A customer with a question had a difficult time finding the one worker with the right bit of information to answer it, and if a customer wanted a transaction to be modified, the whole process had to start over again. Finally, for many workers, the endless repetition of a limited range of tasks is boring. Yet information-handling tasks, even boring ones, usually require a high degree of focused attention. The result can often be a high error rate in the work, and a high turnover rate among the employees. These problems will be discussed at greater length in the next chapter. The Information Age Office For some businesses, office automation offers the possibility of restructuring the work process in a way that reduces some of the problems of the industrial style of work organization. This new approach has been called post industrial or information age and is characterized by an electronic reintegration of the steps involved in producing a product. Electronic reintegration of work represents a new approach to rationalization of work. Instead of attempting to make the work process efficient by rationalizing each task and function separately, this approach seeks to rationalize a whole procedure, perhaps along product or market lines. This is possible because the technology allows: 1) the integration of information from many sources, and Z ) the distribution of information to many locations. A general example can be seen in the case of the bank mentioned above. Under the old system, each worker performed one or two steps in the processing of all letters of credit. Under the revised work process, each worker, with the aid of a computer workstation and a client database, performs all the necessary steps in processing letters of credit for a particular set of clients. The database contains all information related to the customers account. The customer service worker is the single point of contact between the customer and the bank for corrections or inquiries. An illustration of the capabilities of an integrated customer service system at American Express is shown in figure 4-2. customer service representatives have the ability to deal with a wide variety of activities related to their cases. Working from the on-line database they can answer telephone inquires, send written replies to mail inquires, make credit adjustments, stop automatic duns, or issue a special statement. In addition to the on-line database, the customer service workers have access to historical records stored on computer output microfilm. Electronic reintegration allows a worker a view of the whole operation and gives a variety of tasks to perform. Roth factors are indicative of a de-Taylorization of work, and probably contribute to greater job satisfaction, and in many cases, greater autonomy and responsibility. However, electronic reintegration does not necessarily lead to greater autonomy, responsibility, or discretion. Because the work is dependent on the use of the computer, it can be subject to machine monitoring and pacing. When a whole process cannot be handled by one kind of worker, electronic integration can still be used to accommodate a decentralized team approach to the division of labor. Some insurance firms have created small teams of raters, underwriters, and clerical support workers to work on specific product lines. Workers that were formerly separated by function are now integrated into a team that serves a specific market. Organizing the workers in teams rather than isolating them in functional groups allows them greater understanding of the whole production process and facilitates com munication.
PAGE 110
102 l Automation of America Offices Figure 4-2. Model of an Integrated Customer Service System Case 3.. 1. Open a case [ o 2. Post an action 3. Create an adjustment ~ 4. Close a case i I L 1 SOURCE Jay W Spechler D[rector Performance Englneerlng American Express Co personal communication 1985
PAGE 111
.. Ch. 4 The Changing Nature of OffIce Work 103 Baran and Teegarden suggest that in the insurance business, firms will tend to continue to use the industrial-style work process to produce and distribute standardized products. Much of the work will be highly routine, and the work force will consist of a shrinking number of low-skilled clericals and a few highly skilled professionals. With these products the goal is to reduce unit costs as far as possible and to minimize labor costs. For more specialized products, those that require customizing or responsiveness to individual client needs, there may be more use of the team approach along with integrated office automation systems. 7 Baran and Teegarden, op. cit., p. 21. TASK CHANGE AND COMPUTER-MEDIATED WORK One commonly noted characteristic of work with computers is that they make work more abstract and alter the workers relationship to the task. This is sometimes called a computer mediated relationship, in that the individual must now do the task through the medium of the computer system rather than through direct contact with the objects of the tasks Using any sort of tool can, of course, distance a worker from direct tactile contact with an object. But most manual or power tools still allow the worker to remain cIose enough to the object to get direct feedback about the condition of the object, the status of the process, or the need for adjustments. When a computer is used, feedback is indirect, in the form of symbols generated by the information system. Office work is already abstract. What is actually being processed is information. Yet from the point of view of the worker, this information becomes concrete because it is carried on physical objectssignature cards, account ledgers, invoices, letters, reports, and checks. Most manual or mechanically aided office tasks involve transforming the information by manipulating the objectcopying, typing, updating, or signing it. When computers are introduced, many of these objects disappear completely. The worker is left to deal with intangible information, which is transformed in invisible ways inside the computer. See, for example, Shoshanah Zuboff, New L$orlds of Computer-hlediated Work, Har\ard Business I?etiew, SeptemberOctober 1982, pp. 142-152. The rules by which information is prepared or transformed also change with computer mediation. Quantitative comparisons are easy for computers to handle, qualitative judgments are not. Thus, in the process of restructuring work to be done by computers, qualitative aspects of information are either quantified or lost. Correctness of data or procedures must be redefined to fit the formal logic of the computer, and formal correctness can become more important than the relevance of information content. g Human judgments are often replaced with computer-based decisions and human error detection or correction become difficult. The language of the human-computer interface is different from human language, as well. Although computers are becoming easier to use, they still require human operators to learn specific codes and procedural commands. Further, even when these commands are similar to plain English, the formalized process by which the computer works may be quite different from human thought patterns. The human operator must learn to think and to perform actions in a precise order that reflects the logical sequence built into the software. The ability to think in a way that parallels computer logic is not a trivial intellectual discipline, and represents a new way of working for many people. For example, retrieving information from a computerized database re... Further discussion is in Gunnar Aronsson, Changed Work Qualification Structure in Computer-hlediated Mork, National Board of occupational Safet~ and Health, Solna, Sweden, 1984.
PAGE 112
104 l Automation of America Offices quires different modes of thought than finding a folder in a filing cabinet. Physical search is aided by physical aids to memorythickness of the file, color, a coffee stain on the corner. These do not exist in a computer-based file. There is only a name, constructed according to rules inherent in the computer logic. A search through a database requires use of the right key words, in the proper relationship; otherwise the search may yield fault y results. An example of how tasks change when computers take over part of the process may be seen in banks. With computerization, the process of fabricating accounts, the principal activity of a bank, was made internal to the computer. These procedures had previously been done manually, or with the assistance of mechanical accounting machines. The role of human operators then changed from one of doing banking to one of surveillance of the computer as it does banking. Adler points out that: a series of tasks formerly considered the very essence of bank work have been eliminated, including accounting imputation and adjustment, classification of documents, multiple entries of data, manual data search, and supervision by signature. 10 He also notes that new tasks were introduced. For example, Accountants now diagnose and rectify residues and anomalies listed by the computer system. New types of errors and fraudappear. 11 Paul Adler. Rethinking the Skill Requirements of New Technologies, Working Paper 9-784-027, Graduate School of Business Administration, Ilarvw-d University, Boston, October 1983, p. 17, Ibid. Some observers note that the increased abstraction can affect peoples understanding of the work. Some managers in banking have noted, for example, that new employeesthose who have only worked on the automated systemsunderstand the system, but dont necessarily understand banking. They think like computer programmers instead of like bankers. As one manager said: Now you make an input and its gone. People become more technical and sophisticated, but they have an inferior understanding of the banking business. New people have no idea of the manual procedures so they never see or understand the process, People start creating programs that dont necessarily reflect the spirit of the operation.lz Clearly employees acquire new skills, they learn to interact with the computer, they may never learn certain old skills. To the extent that processes are automated and these skills are not needed any more, this may not be a problem. But, if the skills are still needed, either at that job level or at another, there may be costs to individuals or the organization for allowing them to be lost. To continue the example above, an understanding of basic banking practice probably is necessary for promotion above the lowest operational levels. To the extent that employees do not gain it by doing bank work, it may be necessary to acquire it through formal training. -Shosh&ih Zuboff, Problems of Symbolic Toil: How People Fare Mith Computer-Mediated Work, Dissent. winter 1982, pp. 51-61. JOB CHANGES WITH AUTOMATION Office automation can change jobs in many esses might be eliminated. This does not necways. The automated system may completely essarily mean that the people are eliminated take over certain tasks. To the extent that cerfrom the organization-they may be retrained tain processes become completely automated, and transferredbut particular positions no jobs that consist solely of those tasks or prolonger exist.
PAGE 113
On the other hand, office automation creates new tasks, either supporting the system or related to new products made possible by the system. These tasks may be incorporated into existing jobs or new jobs may be created, as in cases where a trainer or systems specialist position is created after introduction of office automation. Finally, very commonly, a change in the work process modifies jobs through the addition or deletion of tasks. This can also result in a change in the boundaries between jobs, transferring a task from one job to another. Any of these changes may or may not be accompanied by a formal change in job titles, descriptions, or compensation. The boundaries between clerical and professional and between clerical and managerial are showing interesting changes with office automation. Some researchers have referred to this redistribution of labor as the clericalization of professional work and the professionalization of clerical work. I S Most clear is the shift of keyboarding tasks when professionals and managers acquire terminals or personal computers. Individuals who would never have typed their own memos and reports now routinely draft them on their personal computers. Clerical workers have not stopped doing keyboard work completely, however. Secretaries may work on documents at a later stage to revise, format, or print them, or they may still key documents that professionals and managers choose not to key themselves. ~evertheless, many organizations have seen a decrease in the amount of time secretaries and other clerical workers spend in keyboarding or revising original material when professionals and managers begin using the system. By the same token, secretaries and some other clerical workers have taken over tasks formerly defined as professional. In case studies of two different banking organizations, for example, secretaries began using on-line See, for emmple, tJ~an (lr~enk]aum, C>dncl Pullman, and Sharon szymanski. F;ffects of office i4utomation cm the Pul)lic Sector Workforce: A Case Study, OTA contract No. 4337990.0, ~lpril 19/!5. Ch 4The Changing Nature of Office Work l 105 databases and statistical software available on their terminals to collect data and do analyses formerly done by economists. Clearly, new technology is not the only factor in clerical workers being able to take on new and interesting tasks; there must be management support for these job changes. The question of whether changes related to automation lead to an increase or a decrease in skill is a hotly debated one. Often the terms job de-skilling and enrichment are used to describe the changes. Enrichment usually implies an increase in task variety, autonomy, skills acquisition, and other factors that are considered to contribute to job satisfaction. De-skilling usually means the removal of these features, usually through simplifying the work and narrowing the number of tasks performed, and is usually associated with rationalizing or Taylorizing the work. Both de-skilling and enrichment could be accompanied by job enlargement that is, an increase in workload. It is clear that the impacts of office automation will vary from one job to another. The following is a brief summary of some generic changes that have been seen to affect different categories of workers. Management and Professional Jobs Increasingly, computers are helping professionals get more work done by providing more and better information, by providing tools to aid in formulating professional opinions, and by taking over some routine decisions. Traditionally, professionals and managers have high levels of autonomy in their work. Use of new technology can give managers and professionals better control over their time and even greater autonomy in their jobs. For example, electronic messaging may reduce telephone .. .C. Amick and J. Damron, Considerations in Defining Office Automation: A Case Study in the Eastern Africa Region of the Morld J3ank, Human Computer Interaction: Proceedings of the Fimt USA-Japan Conference cm IIuman Cmnputer Interaction. Gavrie] Sal\rend} (cd. ) (Amsterdam: ~jlse~rier Science Publishers, 1984), pp. 439-445. See also app. B, Computer hlediated \tork in Commercial Banking. B. Kaplan, et al., Job Demands and Workers Health (}$ashington, DC: U.S. Go\rernment Printing Office, 1975),
PAGE 114
106 l Automation of Americas Offices call-backs, or in some cases, the number of meetings. Managers can get up-to-date information on factory operations or business activity by accessing on-line databases within their companies. They may be able to manipulate these data, to do what if studies to determine the probable outcomes of their decisions before actually making them. Managers and professionals can avoid the need for outside data processing or other expert help by using spread-sheet software or information databases at their own terminals. They can produce polished reports on their own printers without a secretary or a graphics artist. 16 Along with these advantages, there can be some drawbacks as well. Greater autonomy, for example, can have its disadvantages. When managers and professionals demonstrate less need for a secretary to do their typing, they may lose the secretary entirely, even if they still want one for other dutiesor for a status symbol. In another example, a manager who is pleased to get up-to-date information about factory operations through the office automation system, may be dismayed to find that the same information is available to a senior executive, who is therefore capable of more closely monitoring or second-guessing decisions. In this case, the use of the technology might lead to a reduction in autonomy for the lower level manager. Knowledge engineers have been attempting to identify the skills, information, and expertise that go into managerial and professional judgments and to incorporate these into software. The results are a variety of computerbased models, expert systems, and decisionsupport systems. Some analysts have expressed fears that computer-based models are making professional and managerial work more routine. There are also fears that some people are going too far in allowing them to replace human judgment. Human judgment is not perfect, but its imperfection is generally recognized and expected. Many people have higher expectations of computers, and sometimes a misplaced faith in their accuracy. A computer-based model, no matter how good, is only a partial representation of complex relationships, and its internal assumptions, theoretical biases, and mathematical quirks can produce results that are not consistent with the real world. Overconfidence in the effectiveness of computer-based models, some believe, can lead to neglect of other forms of research that were, before computerization, the basis of expertise. Skill in these areas is still important as long as the model is incomplete. 17 The ability of computers to give the wrong answer to six significant figures is sometimes called the tyranny of illusory precision. Its danger to professionals and managers has probably been exacerbated by the growing use of computer graphics. Advertisements by computer companies rightly tout the greater perAlexia, Martin, Office Automation: Catalyst for Change, SRI, International, 1983, p. 9. Linda Sandier, Securities Risk: Wall Street Is Finding Its Trusty Computers Have Their Dark Side, Wall Street Journal, Dec. 4, 1984, p. 1.
PAGE 115
Ch. 4 The Changing Nature of Office Work 107 suasiveness of good graphics (the CEO adopts the proposal illustrated by manager As colorful pie charts, but he throws manager Bs laboriously typed tables into the trash). However, slickness of presentation and effectiveness of results are not necessarily related. Some professionals have expressed fears that too much use of computer-based information or decision-support software is limiting creative problem solving. Much of the value of a professional judgment is that it weighs many factors, many of which are unquantifiable, or even ambiguous or fuzzy. Someone who formulates a professional opinion gives weight to the nonquantifiable information in a gut feeling based on experience. Information available through a computer database may be more complete than information accumulated through other means, but it is also perfect -ambiguity has been removed. Some writers have suggested that ambiguous situations provide a free space for creative thinking that is fundamental to professional work. 18 The problem may be short term. Continued research and development along with practical use will further demonstrate both the potential and the limitations of computer-based tools in aiding decisionmaking. If professionals and managers who put too much reliance on computer systems make more mistakes than those who rely more on traditional skills, they will presumably prove less effective over time. Clerical Jobs It is in the area of clerical jobs that most studies of the effects of office automation have focused, and it is here that many of the impacts seem apparent. 19 Nevertheless, it is hard to generalize, because so many impacts depend Zuboff, op. cit. Exam les include: Mary C. Murphree, Rationalization F and Satis action in Clerical Work: A Case Stud~ of Wall Street I,egal Secretaries, Ph.11. dissertation, Department of Sociology, Columbia University, 1981: Evelyn N. Glenn and Rosalyn 1,, Feldberg. Proletarianizing Clerical L$ork: Technolo~r ond organizational Control in the Office, Case Studies in the Z.abor Process, Andrew Zimbalist [cd. ) {New York: hlonthly Re~iew Press, 1978), pp. 51 -72; and Robert A. Arndt, and 1,arr~ Chapman, Potential Office Iiazards and Controls, OTA Case Stud?, 1984. on how the technology is implemented in a particular organization, rather than on the technology itself. For many clerical workers, both quantitative and qualitative data show that workload has changed following the introduction of office automation. The work pressure scale from the NIOSH study shows that the introduction of a computer terminal has led workers to report greater work pressure. In most cases this is because management choice has been to redesign work according to the industrial or production-line model. The most common scenarios are those of the secretarial worker, the data-entry clerk, 20 or the directory assistance operator. 21 In each case, rationalization of the job fragments it into its component parts. A secretary, for example, who previously did filing, typing, answering phones, and a number of nonroutine tasks may end up only doing word processing. Work pressure may also increase when the work is less paced by the person and more by the computer terminal. In the extreme case, the directory assistance operator may be expected to take a call every 30 seconds, with calls continually forwarded to the operator by the computer. It is this combination of pacing and specialization that leads to increased workload. 22 Whether the workers autonomy and control are increased or decreased on the job can depend on how the automated system is designed. This can be seen, for example, in some of the automated collection systems being used -Michael J. Smith, B.G. F. Cohen, and 1,.\t. Stammerjohn, An Investigation of Health Complaints and Job Stress in Video Display Operations, I{uman Factors 24:4, 1981, pp. 387-400; Gunn Johansson and Gunnar Aronsson, Stress Reactions in Computerized Administrative Work Reports for the Department of Psychology, University of Stockholm, Supplement 50, 1980. Note the example cited in B.C. Amick and D.D. Celentano, Human Factors Epidemiology: An Integrated Approach to the Study of E{ealth Issues in the Off ice, Human Aspects in Offi& A uhxnation, B,G. F. Cohen (cd. ) (Amsterdam: I+; lse\ier Science Publishers, 1984), pp. 153-166. -R. Feldberg and E. Glenn, Technology and L%ork Deg-radation: Effects of Office Automation on Momen Clerical kforkers, Illachina J,.Y. Lkw: Feminist Perspectites on 7echnolog~. (New York: Pergamon Press, 1984).
PAGE 116
108 l Automation of America Offices both by private firms and government agencies. Often these are implemented in a way that reduces autonomy. Under a manual system, the individual worker had some discretion about which account to pursue, when to start a new file, or how often to call. Many automated systems make all those decisions. The system makes other decisions as the worker keys new information into the files during telephone calls with the delinquent accounts. The workers individual assessment of priorities or intuitive judgments of which people are most likely to respond to extra cajoling become less important. The automated system determines when, how often, and how long each call should be. Sometimes the worker has no understanding of the whole case, except through the notes provided by the system. Under a manual system, a worker ends each day with a big pile of completed work and a sense of accomplishment; a machinepaced system may continue to send new work at the same rate, no matter how much is accomplished. On the other hand, if designed differently, the same sort of system can maintain worker autonomy and control while providing the advantages of a computer-based information system. For example, work can be bundled so that one worker, or a team of workers, follows a certain set of accounts from beginning to end, and has the satisfaction of seeing the big picture in each case. If the system is modifiable by the user, so as to give some discretion about timing and duration of tasks or to allow use of individual expertise, it can offer the best features of both the manual and automated systems. Autonomy is not directly tied to the technology and is related to characteristics of the organization and the work process prior to automation. Changes in the work process can lead to a redistribution of tasks, which may either reduce or increase workload for clerical workers. Where professionals and managers have begun to do most of their own keyboard work, or where keyboarding is sent to a word processing pool, secretaries may find their typing load considerably reduced. How this change is handled depends on managerial and individual decisions. In one OTA case study site, a New York City office, secretaries organized themselves to solicit overflow typing from other departments. In the case of one Wall Street law firm, legal secretaries found they were losing their keyboard work to the word processing pool and their lawyering tasks to paralegals. As a result, their work is generally now limited to performing primarily those nonroutine tasks that involve social skills, time emergencies, and multiple contingencies, such as fielding phone calls, gatekeeping, coaxing rush jobs through the bureaucracy, etc. The firm has made a deliberate effort to reduce the number of secretaries and to eliminate the one-secretary-one-attorney relationship. Increasingly, secretaries work for more than one attorney or as members of small clusters in teams with other secretaries. 23 The allocation of tasks is a management decision. In the case of the law firm, business and economic factors moved management to allocate tasks in a new way. A centralized word processing pool appeared to minimize capital costs and ensure maximum use of equipment. Technology helped make the reallocation of tasks possible, but did not guarantee it. The other change in the legal secretaries work had more to do with accounting procedures than with technology; lawyering work done by paraprofessionals is billable to clients, but the same work done by legal secretaries is considered an overhead expense. Economic factors motivate both the acquisition of new technology and the assignment of work. For secretaries at one major bank the increase in workload relates not to fragmentation, but to the addition of new and interesting tasks. When personal computers were introduced into their work unit their keyboarding duties did not decline; the unit withdrew from the typing pool it had used previously, making secretaries responsible for more key-Mary C. Murphree, Brave New Oifice: The Changing World of the Legal %cretary, Jf.y Trouble,s Are Going To Have Trouble With Me: Everyday Trials and Triumphs of 14romen 14orkers, K. Sacks and D, Remy, (eds.) (New Brunswick, N,J: Rutgers University Press, Douglas Series, 1984).
PAGE 117
Ch 4The Changing Nature of Office Work l 109 boarding. At the same time, the secretaries discovered that the statistical packages and databases available on their terminals allowed them to do some types of work previously done by professionals. They found this work an enjoyable challenge and do not wish to give it up, even though it increases their workload. 24 Job Ladders and Mobility Changes in job content may affect career mobility within organizations (internal labor markets) as well as mobility within the economy at large (external labor markets). In many firms employees have the possibility of moving up job ladders within the organization acquiring new skills, assuming greater responsibility, and receiving higher compensation. The development of these opportunities and the cultivation of this internal labor market has been very important to some firms. Some indeed consider the experience of working up through the ranks, with its concomitant understanding of internal operations and demonstrated loyalty to the firm, to be more important than any amount of outside training. These firms may have a policy of filling certain positions only through promotions rather than from outside. Both company policy and the shape of the job ladders may change over time due to any number of factors, including technological change. What will be the impact of those changes directly related to office automation, or to the changes in work process or job definition permitted by technological change? On the one hand, the fragmentation of work into a multiplicity of narrowly defined jobs has traditionally aroused concern about the availability of good jobs. On the other hand, the prospect that office automation may lead to more standardized jobs raises concern about potential polarization of job opportunities into low skill, low wage and high skill, high wage jobs with breaks in the job ladder between them. The import of diminished mobility must be judged relative to the context, which includes f~.{. .lmick and J. I)ammn. op. cit., pp. 1:~$.1 45. both internal and external labor market conditions. In many industries, internal job ladders have been significant factors in the past. Capable and ambitious workers could learn the industry from the ground up, coming in as clerical workers and working their way up through lower level supervisory (or paraprofessional) jobs into management levels. In the insurance industry, for example, clerical workers might become raters and later underwriters. Some technologies provided a natural learning sequence so that workers learned on the job how to carry out more complex tasks and operations. Firms that anticipated future growth wisely saw to it that there were pools of workers at the various intermediate skill levels available for promotion so that expansion could be smooth and free of external labor market constraints. The insurance industry traditionally selected and trained managers from within the industry, usually from within the firm. In other industries too, such as retail trade, the internal job market was the most significant and employers provided much formal and informal training for lower level employees. Although the messenger-to-manager or clerk-to-CEO paths were trod to the end by very few, the possibility that a combination of diligence, striving, and luck could take one to the top was a powerful motivating factor. Some researchers have suggested a growing trend in recent decades to externalize both training and recruitment. In other words, employers are depending more on lateral recruitment of managers and skilled workers trained in colleges and business schools. 27 This also It has been pointed out b~T se~eral researchers that this job ladder worked well only for male clerical workers before the earl~ 1970s; women clerical workers seldom made the climb. Under pressure of equal opportunity initiatives this became a significant job ladder for women in the 1970s, but is now said to be truncated by the automation of rating and mu tint underwriting. -*Eileen .Appelbaum, School of Husin(lss tldnlinistrati(~n, Department of Economics, Temple Universit?r, personal communication, Feb. 5, 198,5. See for example, rhierr~ ,J, No~elle, Plmplo~n~ent an d Career opportunities for Women Minorities in a Changing I~conom~: The F: xperience of I.arge and Nlediurn Sized Firms, Conser~ at ion of I IU man Resources, (oIu mbia I Ini\ersi t?, ,J anu ar}. 19N3,
PAGE 118
110 l Automation of Americas Offices applies to lower level workers, because of the longer years of schooling for the general population and the growing availability of vocational schools and community colleges, employers can rely on the public sector for training once provided within the firm. For those who cannot or do not get that training and educationfor reasons related to socioeconomic conditions, culture, talents, or the driving necessity of earning a living early in lifea critical alternative path to better jobs is being eroded. Office automation could further weaken and truncate internal job ladders by completely automating some of the jobs that provided the intermediate rungs, and by encouraging employers to rationalize, simplify, and narrow tasks so that workers learn about only smaller and smaller fragments of the total work of the organization. This also makes the individual worker less and less valuable to the employer since he or she can be readily replaced by others who require only the briefest training, or have already learned the same simple procedures in other firms or other industries. Their skills are, in other words, fully standardized. This tendency from one perspective increases the mobility of workers, who can move relatively easily from one firm to another or from one industry to another with fully transferable skills. But it does not increase their upward mobility at the same time. If the supply of lower level jobs shrinks and the supply of higher level jobs increases, such workers may not benefit because the higher level jobs are likely to be filled laterally, with people who have had the benefit of higher educationor more recent education. Some researchers believe these trends are leading to a polarization of the labor force into low skill, low paid jobs and high skilled, high paid jobs, with few opportunities in between. Empirical evidence regarding any economywide trend toward job polarization within organizations gives conflicting signals. Polarization of jobs has occurred in such office-oriented industries as banking and insurance; examples are provided in previously cited research by Baran and Appelbaum. Conditions in those industries are not necessarily generalizable to general office circumstances, where initial staffing patterns are different and motivations regarding job design are also different. In many general office environments with lower levels of clerical employment (the extreme being the one-secretary office) office automation is not driven as strongly by the objective of immediate reductions in force. In many such offices, there is perhaps a stronger need for quality improvement, greater timeliness, and workload leveling. The result could be a greater tendency for job enhancement and possibly improved internal job ladders. Many people advance their careers by moving from one organization to another. Thus, the external labor market is also a source of opportunity. In professional, technical, and managerial occupations, advancement by moving between organizations, as well as by moving within them is relatively common. A study of occupational mobility noted that most job changes during the study period occurred within the same major occupational group, especially for both male and female professionals. 24 Indeed, mobility of these personnel seems to be the foundation for the employment agency and placement business. Although some people can advance readily via the external labor market, some groups find that more difficult. In many cases, a change in jobs does not reflect advancement. It was noted in the previously cited occupational mobility study that a relatively high percentage of women who had recently moved into manager or sales positions had previously been clerical workers. However, the largest group of women changing jobs were those who moved from one clerical occupation to another. 29 For lower level clerical workers, advancement through job changes is likely to require formal, external training, if on-the-job acquisition of skills needed for higher level positions is diminished. A reduction in intraorganizaKEllen Sehgal, occupational Mobilitj and Job Tenure in 1983, A40nthly Labor Retiew, October 1984, pp. 18-23. q Ibid.
PAGE 119
Ch. 4The Changing Nature of Office Work 111 tional career ladders is a shift in training burcation and training is uneven; job opportuniden from employer to employee. That burden ties may be increasingly tied to social class is aggravated inasmuch as access to good eduor economic means. ORGANIZATIONAL STRUCTURE AND RELATIONSHIPS Changes in the structure of organizations related to introduction of office automation can be influenced by economic, environmental, or cultural considerations, in addition to technology. Managerial strategies-upper level decisions about the organizations goals and the role of technology and of people in achieving themare probably the major considerations. Formal structural changes (i.e., something that would be noticed on the organization chart) might take place during implementation of office automation or shortly afterwards, perhaps reflecting a change in the work process. For instance, in some firms, the creation of administrative centers and the aggregation of all support personnel in word processing pools accompanied the decision to adopt office automation technology. On the other hand, change may not be reflected formally at all, but may only be seen as changes in communication patterns or in power relationships between groups or departments. Organizational effects often take place over a period of time, as part of an evolutionary process. As familiarity with technology and its capability grows, structural changes may be introduced that allow the organization to take better advantage of those capabilities. New product lines made possible by the technology may emerge, which in turn give rise to new work groups or departments. Power and Access to Informatio n Early research on the computerization of organizations suggested that introduction of computing appeared to change power relationships, and a central preoccupation of research in the 1970s, was whether computers led to growing centralization of power among upper level managers. The sources of organizational power include such things as control over resources, information, critical technical skills, or coping capability .30 As control over the computer offered senior management many of these things, some researchers predicted that computerization would increase the centralization of decisionmaking power. These studies were done for the most part in the era of mainframe computers. However, even in the age of personal computing and distributed processing, there may be reason to believe that integrated information systems can aid in the centralization of organizational power. Possible routinization of middle management work, through dependence on computer-based information systems and decisionsupport models, was discussed earlier. Some researchers have pointed out that any worker whose job becomes more routine becomes less powerful, less able to offer a unique contribution to the organization. While middle managers may retain responsibility y for making certain decisions, their range of choices could become more rigidly circumscribed by the assumptions inherent in their decision-support system. Their authority could be accompanied by increased supervision and control, as higher level management will have access to the same information and decision aids. 31 Another expert has suggested that paradoxically, [centralization] may be manifested by locating decisions at lower levels but controlling decision outcomes through the provision of performance records. 2 It has been suggested that the widespread use of integrated office autoJeffrey Pfeffer, Power in Organizations (Nlarshfield, MA: Pitman Publishing Inc., 1981), p. 274. M. Lynne Markus, S-vstems in Organizations: Bugs and Features (Boston, MA: Pitman Publishing Inc., 1984), pp. 52-53. &Daniel Robey, Computer Information Systems and Organizational Structure, Communications of the .4CAI, October 1981, pp. 619687.
PAGE 120
mation systems reduces the need for middle management and could ultimately lead to a flattening of some organizational pyramids .33 The concept of centralization is too simple to describe all the possible changes in power relationships that might result from the introduction of office automation. Organizations are not monolithic; different departments, work groups, or factions may be affected differentially. A set of case studies found that introduction of automated production scheduling systems into Danish plants altered the power relationships among management groups because of differential access to information. The plants had a matrix organization, with major decisions being made by interdepartmental teams of production planners, production managers, and plant managers. 34 The new computer-based production planning system was owned and mainly used by production planners. It seems to have concentrated up-to-date knowledge in their hands, thus giving them greater power over production decisions and higher status as compared to production managers and plant mangers. These shifts of influence were evidently unintended by the designers of the system. One expert 35 in interpreting the same study, points out that this concentration of power caused the matrix arrangement to lose some of its anticipated effectiveness, although it was still retained, largely for symbolic or ideological reasons. In one of the OTA case studies, Aircraft Instruments Plant, the introduction of the MRP II planning and scheduling system seems to have given greater power to production-support office workers while taking some discretion away from the supervisors on the factory R~~f T Vv~and, Integrated Communications and Wrork h~fficiency: Impacts on organizational Structure and Power, paper presented at the International Communication .Association ,4nnual Con\rention, Honolulu, Hawaii, hlay 22-27, 1985, p. 16. IN. Bjorn-Anderson and P. Pederson, Computer Systems as a Vehicle for Changes in the Management Structure, I nformation Systems Research C~roup, University of Copenhagen, 1$orking Paper 77-3, 1977. Cited in Rob Kling, Social Analyses of Computing: Theoretical Perspectives in Recent Empirical Research, CoI]]p~ltingSLlr\re\r 12 (March 1980), pp. 61-110. llobey, op. cit. floor. (See appendix B.) While access to up-todate information is partly responsible for this shift, even more important is the ability to act on information. Under the new system, only production-support staff have the authority to override the system when the automatic response would be inappropriate. Because of the complexity of the work and also because of some conceptual problems in the systems design, such conditions arise frequently. Previously, factory supervisors might have dealt with exception conditions on their own authority, but now they must request permission from production support. On the other hand, where shifts in relative power are inappropriate to the corporate culture of any organization, management has the option of reinforcing existing lines of authority and modifying the automated system to ensure that change does not take place. One example is that of a military organization that replaced its manual logistics system with a computerized one. Under the old system, senior officers received requests for equipment transfers and had junior officers to compile a detailed report on which to base a decision. The junior officers gathered information for these reports through written documents and through telephone contacts with a network of supply officers at other installations. With the new system, the junior officers worked at computer terminals with on-line access to information about the location and status of equipment, a decision-support system to calculate least-cost routing, and the means to implement transfers. Because the junior officers had timely and complete information, people began making requests for equipment directly to them rather than going through proper channels, and the junior officers sometimes could not resist the temptation to respond, even though they did not have authority to do so. Senior officers attempted to have the new system removed. In the end it was retained, but it was redesigned so that junior officers were unable to take action based on their decisions. 36 Markus, op. cit., p. 74.
PAGE 121
Ch 4 The Changing Nature of Office Work 113 Pho [o cred(f E?(// Ke//ey Commu n~cat(on Workers ()+ 4 me, IL a Advances in microelectronic technology have made telephone operators more productive
PAGE 122
114 l Automation of Arnencas Offices Many firms are using integrated databases, and employees at various levels need access to it in order to do their jobs. However, access to different parts of the database can be stratified so that existing organizational lines of authority are not threatened. End-User Computing Power Since one source of power is control over resources and since the computer itself (or what it can do) is a resource, there has been some dispersal or shifting of power in many organizations. Only a few years ago computers were a scarce resource; data processing professionals usually acted as gatekeepers between all other departments and the computer, and thus were in a position of power. Over the past 5 to 10 years, with the declining cost of electronic hardware, computer power has become much more widely dispersed throughout many organizations. High costs of computer equipment used to make centralization more attractive. Lower prices, increased capabilities, and the ability to link workstations in networks now make it possible for firms to disperse data processing and word processing capabilities. This does not necessarily mean that the word processing pools have disappeared (although they have in many firms) but that word processing capability is now available to private secretaries, receptionists, professionals, and others. Nor are centralized data processing departments likely to disappear, as many organizations continue to need the power of mainframe computers and the expertise of computer professionals. However, the ability to do a growing variety of data processing activities can now be brought directly to every professional and manager. In some organizations there maybe an overall strategy to keep centralized control over computer resources, to maintain compatibility between systems, to make sure that personal computer users can access databases, etc. This function may be performed by the data processing department or by some other office automation group. Where some other group is designated, turf difficulties with the data processing department have often arisen with the continuing convergence of word processing, data processing and communication. In one Fortune 500 company, conflict between the data processing department and ad hoc office automation group was finally resolved when the office automation group was assigned responsibility for managing all resources for data and word processing within the firm. The data processing department was limited to providing data processing services for external clients .37 In many cases, organizations have acquired office automation, especially personal computers, without a grand plan, in response to grass roots decisions at the individual or departmental level. Later attempts to standardize or to centralize control of the organizations information resources inevitably result in battles when different kinds of equipment or different philosophies of operation have already become entrenched. No system is perfect for everyones needs, and many groups may want to influence the development of office automation systems to make them most suitable to their own needs or most acceptable from their own bias. Communication Additional organizational implications of office automation are related to their effects on communication within organizations. An interesting sidelight on the case involving the military logistics office is the way in which the new system changed communication patterns. The senior officers were left out of the loop, when supply officers began contacting junior officers directly for equipment transfers. In addition, however, the flow of information reversed along the junior officers network of telephone contacts. The junior officers had created these networks to call out for information. However, once the on-line system came into place, the junior officers became a source of systemwide information. Their form Ginger Levin, Excellence in Information Resource Management, GHL Inc., Washington, DC, 1984.
PAGE 123
er informants made use of the same telephone network, with its established informal working relationships, past favors, friendships, etc., to call in for information or for emergency equipment transfers. Changes in communication patterns maybe at the base of many of the organizational changes associated with office automation, simply because communication is such a major part of white-collar workers jobs. Managers, especially, spend a large portion of their time communicating. It is estimated that from 46 to 77 percent of their day is spent in oral communication, including both telephone conversations and face-to-face meetings. 38 The introduction of new pathways of communication may bypass traditional gatekeepers and make profound changes in who talks to whom. Although some researchers have predicted that full access by many workers to a linked electronic system could lead to a complete reshuffling of power relationships, giving most power to those with access to the system, 39 there is little evidence that formal hierarchies are greatly affected by innovations like electronic mail and messaging. In fact, these systems are often constructed in such a way as to reinforce the existing hierarchy, e.g., automatically sending information copies to managers of all their subordinates messages, but not the other way around. There are some anecdotes of superiors using the system to intensify pressure on subordinates (I requested that information an hour ago! Dont you check your messages?). In addition, even in a completely wide open system, where everyone theoretically has access to everyone else, it is doubtful that everyone has something useful to say to everyone else. However, there are cases where use of an electronic mail system has lead to the undermining of an organizations traditional hier~largarethe 11. Olson, New Information Technology and Organizational Culture, .$11S Quarter-l> Special Issue 198,2, pp. 71-92. See for example, Carol T. Gaffney, The Impact of office Automation on Power in organizations, Proceedings of AFI PS Office Automation Conference, Philadelphia, February 1983, pp. 216-21 Ch. 4The Changing Nature of Office archy and authority structure. In one example, members of a disbanded project team used their firms electronic mail system to continue work on an abandoned project, eventually convincing management to revive it. At the same firm, the electronic mail system became a means of exchanging complaints and criticisms about management. 40 More recently there have been other reports of electronic messaging systems being used for flaming, or emotional outbursts related to organizational or other problems .41 Such subversive use of electronic mail probably depends a good deal both on corporate culture and on the configuration and capabilities of the system. Anecdotes from other firms suggest that their electronic mail is never used for complaints or any communication with implications for internal politics because electronic messages are more easily traced than paper ones. It is not yet clear how effectively electronic messaging will replace face-to-face or telephone communication. Early evaluations have not shown consistent patterns of replacement of telephone, face-to-face, or written messages with electronic messages across organizations. At some organizations, as at OTA case study site Company XYZ, electronic mail has been consciously rejected because personal communication is considered important to the corporate culture. In another OTA case study, Office of the Special Trade Representative, internal electronic mail is used extensively for circulation of documents for comment and it would be used externally if other agencies were equipped. Many managers do not use electronic mail extensively to replace face-to-face meetings or telephone calls. This may be attributable to two factors. First, use of the technology may interfere with the personal management style. Second, there may be dissonance between the electronic system and the type of information being transferred. In one study, over one-third of the messages exchanged in face-to-face or Tlarkus, p. 60, citing Ralph filmmett, TNET or Gripenet? Datamation, .Notember 1981, pp. 48-58. 1 Con\ ersations by Computer, ~~lectronic .Ser}ices Unlimited, october 1984.
PAGE 124
116 l Automation of Americas Offices telephone meetings contained soft information, that is, they conveyed opinion or conjecture. It is possible that messages of this sort cannot be effectively communicated through electronic mail. 42 Another researcher has noted that teleconferencing also is more successfully used for giving or receiving factual information, and less so for complex tasks like bargaining and persuasion. 43 These factors make it likely that these technologies are more likely to reinforce existing trends in the organization rather than to bring about radical changes. Major changes in the way people communicate at work may come, not from electronic mail, but from common access to databases and to computer operations that control the work process. If office automation offers employees new access to information, it can also change their informal patterns of communication. For example, where coordination between work groups is taken care of in informal hallway conferences among work group members, there is opportunity for social interaction as well as the exchange of work-related information. The introduction of an office automation system with a shared database may allow these groups to accomplish coordination through the system, and the people may no longer need to see one another for that purpose. At Aircraft Instruments Plant, an OTA case study site, coordination through the system, and the elimination of informal conferences and negotiations, were clearly goals of the new MRP II system. Although these goals have not yet been achieved, the new system has clearly changed patterns of communication within the plant and its offices. Informal conferences continue, but now a major purpose of negotiation is to get permission to override the system. (See appendix B for further information. ) This integrative quality of computer systems, which when successful allows coordination without physical proximity, leads directly to the subject of the next section. -Olson, op. cit. Thomas hlancie~rille, The Spatial I+~ffects of Information Technology, Some I~terature, Futures, February 1983, p, 67. Dispersion of Work Activity Perhaps the most powerful effect of the communication features of office systems is the ability to change the location and geographic distribution of work activity. Once words, data, or pictures have been converted to electronic form they can be sent to a device across the room, across town, or halfway around the world with almost equal ease. As a result, people do not have to be aggregated in one place in order to work in the same office. Traditionally, information-intensive services have located in large cities where they have the advantage of proximity to other organizations and people. To the extent that telecommunication can substitute for proximity, organizations will have many more choices in where to locate. Many see this as leading to a more geographically dispersed style of functioning because: the increasing spread (and cheapness) of telecommunication reduces the former external economies of physical proximity, we see the dispersal of corporate headquarters and major white-collar operators like the insurance industry from the decaying central cities to the suburbs. 44 Firms might minimize use of expensive big city office space not only by locating in the suburbs, but by moving to or establishing branches in small cities or towns, or even in foreign countries so long as a suitable work force and an adequate telecommunications system are available. The departments most likely to be located in remote locations are those that need little outside contact or that primarily engage in routine communications. 45 So long as coordination and integration of work can take place electronically, there is no longer much need for spatial proximity. In a study of automation in the insurance industry, it was pointed out that the centralizing and dispersing tendencies of automation are complementary. 46 Procedures for insurance Ibid. Ibid. Barbara 13aran, *Technological Innovation and Regulation: he Transformation of the Labor Process in the Insurance Industry, prepared for Technologyand Economic Transition Project, Office of Technology .Assessment, OTA contract No. 3433 -3610.0 Washington, DC, ,Januar~r 1985, p, 95.
PAGE 125
Ch 4 The Changing Nature of Office Work 117 rating, underwriting, and claims handling are increasingly being standardized and automated. A recent trend has been to centralize physically the databases on which these activities depend. Many insurance companies are now consolidating their regional offices into two or three highly automated regional centers. At the same time, these firms are disbanding some of their large centralized dataentry pools. Data-entry work is being dispersed to functional departments located in agencies or field offices, connected by on-line terminals to computers at the regional centers. Functional units within the company can be integrated electronically through the computer even if they are geographically distant. At the same time, because in all cases information is being increasingly consolidated in central master files, although access to it is proliferating and decentralizing spatially, decentralized production is fully compatible with centralized direction and control. 47 Spatial dispersion of office work could have major effects on the labor force of the future and could have differential impacts based on race and class. Large data processing pools often employ high proportions of minority women in low-skilled clerical jobs. These are the jobs that are increasingly being automated away. The more highly skilled clerical jobs that 1 hici. remain will increasingly be located outside of cities as organizations search for lower cost, nonunionized or more highly educated labor. In one insurance company relocation, labor force characteristics were explicitly important in the location decisionthe company sought to place its offices in communities where women were well educated, of German descent, with unemployed husbands as these were considered the most likely to put in a good days work. 4 There is probably a limit to how far spatial dispersion can go. Most organizations are unlikely to scatter branch offices across the country simply because office automation makes it possible. Certain activities may not lend themselves easily to dispersion. Because of the unsuitability of telecommunications for activities like negotiating, persuading, or exchanging soft information, qg it is likely that some critical mass of employees would have to remain near each other for face-to-face meetings. In addition, other cultural, logistical, and cost factors unrelated to the features of office automation may limit the number of separate installations. Baran and Ikegarden, op. cit., pp. :]0-31. olson notes that this ma~ he one reason wII} man} managers haye not started using electronic mail to replace faceto-face meetings. See Olson, op. cit. IMPLEMENTATION While advertisements are full of the promised benefits of office automation, there are a vast number of horror stories about unsuccessful attempts. The technological and organizational barriers to be overcome, the internal political battles that often must be fought, the planning and redesign of work to be done, are daunting. After hearing and seeing some of the things that can go wrong, those responsible for bringing office automation to their own organizations may consider an implementation to be successful if it is not completely botchedif productivity does not plummet or employee turnover does not rise to the ceiling. Paul Strassman, formerly Vice President of Xerox, noted that many organizations greatly underestimate the organizational costs (as opposed to technological costs) of automation. Organizational costs, for example, inefficiency while learning new procedures, time lost for training, time spent negotiating with peers about new work processes, can add up to several thousand dollars per employee in the first year if properly noted and accounted for. 50 Where organizational issues like job redesign and workflow restructuring have not been
PAGE 126
118 Automation of Americas Offices properly managed, organizational costs can more than offset productivity gains expected from new technology. A number of firms are demonstrably worse off with automation than they were without it, at least in the short run. However, these new tools are becoming increasingly available and increasingly necessary for doing business in the present era. It is possible to learn something from successful implementations to avoid the problems experienced in horror stories. The success of the introduction of office systems into an existing organization has been found to depend on a number of circumstances that surround the implementation. Although various researchers disagree on the relative strength of their influence, there seems to be some general agreement on their importance and direction. 51 These variables include the reason for adoption, the involvement of key actors, the use of adaptive planning procedures, level of user participation in planning and decisionmaking, and training and incentives for users, and training. Reason for Adoption In most discussions of successful installations of office systems, there was a clearly identified organizational objective. Organizational goals involving improved outputs have been associated with success more often than organizational goals defined only by cost reduction. Cases where a technological opportunity was seized, without demonstrated organizational need were those where there was least evidence of success. 52 These findings are in line with admonitions to tie system planning closely to the offices business function. 53 Where there is a clear understanding of the output to be produced and Tora Bikson, Barbara Gutek, and Don A. Mankin. Implementation of Information Technology in Office Settings: Review of Relevant Literature, Rand Corp., Santa Monica, CA, November 1981. ) Ibid. Michael Hammer and Michael Zisman, Design and Implementation of Office Information Systems, I.aboratory for Computer Science, Massachusetts Institute of Technology, OAM-005, May 1979. the role of automation in improving that output, there is a greater likelihood of selecting and designing a system which achieves that objective. Although cost reduction can be an important objective, a narrow focus on cost reduction alone may not have the desired results if the organizations goals and the work process leading to the achievement of those goals have not been thoroughly analyzed. In the case of Company XYZ, an OTA case study site, use of information technology was seen as a competitive weapon, a means of both cutting costs and increasing market share. However, cost cutting was placed in context of the type of work done at XYZ and the corporate philosophy. The goal was to replace old, rigid, batchoriented information systems and manual technology with flexible cutting-edge electronic tools, and concurrently to give users a renewed sense of power, insight and enthusiasm about their tasks, so as to improve organizational performance. 54 This focus on user needs led to the criterion that any system selected should be manipulable by users and should: augment the worker rather than automate the work. Being able to ask good questions, do insightful analyses, take the initiative, and make a decision are emphasized as distinctly human skills that computers can assist but not replace. While other firms will have different goals and a different philosophy about the role of their workers, analysis of organizational goals and the role of computers in achieving them is still important. Key Actors Another important element in successful implementation is the key actor. This can sometimes be a top management official who Tora Bikson, Don Manken, and Cathleen Statz, Individual and Organizational Impacts of Computer-Mediated Work: A Case Study, prepared for Office of Technology Assessment, 1985, p. 21.
PAGE 127
Ch 4The Changing Nature of Office Work 119 clearly supports innovation, but in many cases a technological entrepreneur somewhere below the top level of the firm maybe the prime motivating force in a successful introduction of new technology. In some studies it has been found that the role of the unit manager is essential in successful implementation. 55 Support and understanding of top management are also crucial. In the case of XYZ, the CEO was the inspiration behind the move to automation, although actual responsibility was delegated. Having management support guaranteed needed resources and also provided guidance in developing a system that matched corporate philosophy. On the other hand, lack of top management support seems to have aggravated some of the problems in the Aircraft Instruments Plant case study. For example, insufficient funds and time for training, both of which undercut the success of the implementation, were probably related to lack of senior management understanding of the need. Senior management did not ease production quotas when the new system was first introduced, even though it would be reasonable to expect some decline in productivity as people learned to use the system. Further, ongoing performance measures were not modified to fit the scheduling process inherent in the system. Thus, toward the end of each month managers all scrambled to bypass the system and manually schedule projects with high dollar value but low priority, in order to meet dollar quotas. This is likely to continue as long as management priorities and system priorities are in conflict. Adaptive Plannin g A planning process that is flexible and continuous, both before and during introduction, also seems to be associated with successful implementation. It is inevitable that both the organization and the system will change somewhat in the process of implementation. A plan set in concrete, which cannot be modified to accommodate changed user needs or new Ibid,; and ~ikson, ~utek, and h!ankin, op. cit. information, can seldom be successful. In some cases, with sufficient flexibility on the part of the planners and the users, even a drop it in their laps approach to implementation one where users have minimal involvement in planningcan work. A generic system can be modified to fit an organization needs during the early days of operational use. However, users may not be sufficiently flexible, and adaptability may not be the only requisite for success. In general, research in t he literature as well as how to articles in the computer and management trade press agree that involving users in the planning process is highly desirable. Users in the Implementation Process There are at least three areas where users must be taken into accountdesign and implementation, training, and incentives for users. Before the user can be adequately involved in the process, however, it is necessary to ask who is the user? Although the term user-driven is often used in designing office automation equipment, there are several conceptions of who the user is, and it is not always clear which user is being addressed or represented. Wynn points out at least three levels of user. 57 One is the organization or department that intends to develop an automated system. Another is the person or group of people within the organization with the authority to make decisions about the purchase of equipment. This is often the user that the vendor is trying hardest to please. The third is the end user, that is, the person who actually operates a computer or terminal. The needs, opinions, and level of knowledge about office practices of these different types of user are not necessarily identical; they may not even be in harmony in some cases. )Ron H. Epstein, An Approach to Inb-educing and Evaluating Automated Office Systems, Electronic Office: Management and Technology (Pennsauken, NJ: Auerbach Publishers. 1980). Eleanor 11. J$ynn, The User as a Representation Issue, Proceedings of the Hawaii International Conference on SJTstem Sciences, 1983.
PAGE 128
120 l Automation of America Offices User Involvement in Design and Implementation Successful implementations of office automation technology tend to include users of all types in the process. While there maybe some difficulties in involving future end users in system selection when they are not yet familiar with computers, there are benefits to be derived from the investment of time and effort necessary to bring them up to speed. No one knows better than the actual end user exactly what he or she does and how it might be done better. User needs may not be well understood by system designers or vendors, or even by managers in the same firm. Attempts to divine needs without asking can too often lead to the use of social stereotypes (clerical workers do not make decisions, managers will not type) instead of fact. Getting the user perspective early in the planning process can save costly retrofitting or other problems later. Methods for involving the users in the design process can include interaction with the entire user community by collecting design data through surveys and polls, or by having user representatives on the design task force or committee. Finally, involving users in the design process can help to overcome some of the resistance to use of new systems often classified as fear of computers, technophobia, or simply fear of change. While some have assumed that certain types of people naturally resist new technology (e.g., older workers, less educated workers, etc. ) Wynn notes that: The hidden assumption in these notions is that the reason people may not rush to be new users of technology is a) psychological rather than rational in nature; b) the fault of the user not of the equipment. Both of these assumptions, if acted upon, cause development organizations to do nothing at all to solve the problem, to go ahead and design as they see fit and see if people can be forced to use the resulting system, or to go to great lengths to cater to the users supposed psychological and cognitive incapacity, mostly by trying to advertise away the problem. 58 Eleanor H. Wynn, Linking User Responses to the Design Chain, AFIPS Office Automation Conference, San Francisco, Apr, 5-7, 1982. Several studies have shown that what managers or system designers sometimes perceive as clerical workers irrational fears of technology are actually very rational concerns. Wynn found in open-ended interviews at several firms that clerical workers were concerned, for example, that they might not be able to learn the system, that the equipment chosen was wrong for the job, that new measures of job performance were wrong for the job, that they might lose their jobs and be unemployed. She concluded that people can be seen to resist not change itself, but change for the worse. 59 Managers often do not perceive these specific concerns, but view them as generalized irrational fear. In one survey of Fortune 500 firms, over half (57 percent) of the managers attributed employees apprehension about new systems to general, unfocused fear, while only 8 percent attributed it to fear of computers and 4 percent to skepticism about managements ability to manage the change. On the other hand, only 20 percent of clerical workers attributed other workers apprehension to general fears. Most (30 percent) identified concern about computerspresumably fear of being unable to learn to use the new system. About 25 percent cited skepticism about management choices, and 22 percent cited job security as the major worry. 60 Training Training employees to use automated equipment is essential to making most effective use of it, and most researchers agree that the importance of training to a successful implementation cannot be overemphasized. As Strassman notes, Training, training, training: these are the top three priorities to changing work in the automated office. 61 A training plan is a necessary part of the implementation plan when an organization begins using office automation system. In addition, continuing efforts Ibid. )( Research and Forecasts, Inc., !lhe Kelly Report on People in the Electronic Office 11: E?ow Office Workers View Automation (Troy, MI: Kelly Services, Inc., 1983), p. 10. Paul A. Strassman, Information Pa~ofi} The Transformation of Uork in the Electronic Age (New York: The Free Press, 1985), p. 81.
PAGE 129
Ch. 4 The Changing Nature of Office Work 121 are necessary for training new employees and for upgrading the skills of current ones. An adequate training program may often constitute one of the major costs involved in introducing automated office equipment. Unfortunately, many firms grossly underestimate the need. In a number of surveys, workers and managers in automated offices have cited lack of adequate training as one of the major problems related to the implementation of office automation equipment. 62 Training is discussed in more detail in chapter 3. Incentives for Users The introduction of new technology can cause, or at least be associated with, changes in the way people do their jobs; in the way they interact with other people; and in the amount of power, authority, and self-determination, they have at work. To the extent -See ~~r ~xample, Iioney.well Technalysis, Office Automation and the 11orkplace, Ilone~rwell, Inc., No\renlher 1984; and Kelly Ser\ices, Inc., The KelIJ Iieport cm People in the I;lectronic Office, results of sur~e~w performed h~ Research and Forecasts, Inc.. three \olumes, 1984. that people perceive the changes associated with a new system as changes for the worse, disimplementation or damaging resistance can result. Many successful implementations have featured incentives to encourage users to accept the new system, acquire new skills, or accept changes in the social or organizational context. While benefits for the firm, such as increased productivity, may be important to some workers, more personally relevant rewards are also important, for example higher wages, opportunity for other training or advancement. However intangibles such as protecting the workers self-esteem are also extremely important. In this regard, involving workers in the change process can in itself be an incentive. [The employee] should see himself as the master of the machine, not its servant; not as a victim of the office design, but as a participant in it When invited as collaborators, many office workers will respond with enthusiasm to office system automation. 63 Hammer and Zisman, op. cit., pp. 34-3,5.
PAGE 130
Chapter 5 Office Automation and the Quality of Worklife
PAGE 131
Contents Page Section I: Stess and the Quality of Worklife . . . . . . . . ~ 126 Sources of Stress. . . . . . . . . . . . . . . 127 Strategies for Alleviating Adverse Stress Responses. . . . . . . 136 Section II: Office Automation and Health, Illness and Disease. . . . . 137 Visual System Outcomes . . . . . . . . . . . . 139 Musculoskeletal System Outcomes. . . . . . . . . . . 141 Reproductive System Outcomes . . . . . . . . . . . 144 Stress-Related Outcomes . . . . . . . . . . . . 148 Section III: Interventions . . . . . . . . . . . . 151 Office Design and Workstation Design . . . . . . . . . 151 Public Policy and Quality of Worklife . . . . . . . . . 155 Research Needs . . . . . . . . . . . . . . 162 Labor-Management Relations and Office Automation . . . . . . 163 Tables Table No. Page 5-l. Working Conditions Likely to Produce Stress Responses in Offices . . 127 5-2. Exposures in the Office Other Than the Video Display Terminal . . . 138 5-3. Epidemiological Criteria for Assessing Causality . . . . . . 139 5-4. individual and Work Environment Factors Associated With Visual System Outcomes . . . . . . . . . . . 140 5-5. Visual System Outcomes: Known Acute or Potentially Chronic. . . . 140 5-6. individual and Work Environment Factors Associated With Musculoskeletal Outcomes . . . . . . . . . . . 141 5-7. Musculoskeletal System Outcomes: Acute or Potentially Chronic . . . 142 5-8. Reported Cases of Reproductive System Outcomes in Computer-Mediated Workplaces by Work Site and Job . . . . . . . . . 145 5-9. U.S. Occupational Exposure Standards . . . . . . . . 147 5-10. international Radiation Protection Association Occupational Exposure Limits to Radiofrequency Electromagnetic Fields . . . . . . 147 5-n. Workers Compensation Claims for Employees Working at Video Display Terminals. . . . . . . . . . . . 156 5-12. Examples of Mental Stress-Mental Disability Claims for White-Collar Workers . . . . . . . . . . . . 157 5-13. Union Membership by Industry and Occupation . . . . . . 166 Figures Figure No. Page 5-1. Characteristics of the Office Setting That Can Contribute to an Individuals Quality of Worklife . . . . . . . . . . 126 5-2. The Proportion of VDTs Installed in New Office Space Compared to Existing Space, 1984-90 . . . . . . . . . . . 151 5-3. A Comparison of Workers Compensation Claims for Mental Stress and Other Occupational Diseases . . . . . . . . . . 157 5-4. Employed Wage and Salary Workers Covered by a Union or Employee Association Contract in 1983-84 . . . . . . . . . . 166
PAGE 132
Chapter 5 Office Automation and the Quality of Worklife This chapter is concerned with the individual in the office and the way in which office automation affects the individuals quality of worklife. A traditional definition of the quality of worklife is the degree to which members of a work organization are able to satisfy important personal needs through their experiences in the organization. The relationship of the worker to the office environment affects health, well-being, and productivity. Office automation changes the physical and psychosocial dimensions of the workplace and the work process. Of the many characteristics of the office, no one factor is the sole determinant of quality of worklife, especially not technology; but technology can change the social processes involved in producing an output or the way a person does a set of tasks. This chapter considers stress, office and workstation design, the human-computer interface and the way these relate to an individuals health, wellbeing, and productivity. The discussion is two-pronged; one emphasis is on how new office technologies affect the quality of work life; the second is on public policy. It focuses on factors that contribute to organizational effectiveness by improving work quality and quantity, reducing turnover and absenteeism, and ultimately indirect labor costs.
PAGE 133
126 l Automation of America Offices Figure 5.1 .Characteristics of the Office Setting That Can Contribute to an Individuals Quality of Worklife Fundamental elements Work process Organization I Technology I Physical envionment Huma-computer interaction Individual Building design Office design Workstation design The subjects discussed in this chapter are: station design and labor-management relations. A discussion of labor-management relations ends this chapter because many of the problems cited in chapters 4 and 5 can potentially be solved or alleviated through some form of worker involvement in the decisionmaking process, such as collective bargaining. SECTION I: STRESS AND THE QUALITY OF WORKLIFE Stress is a term often used in both conversation and scientific debate to describe troublesome experiences in daily life, ranging from the illness of ones child to an acrimonious session with the boss. Some stress is routine, the daily hassles of work. But, typically, events that precipitate stress responses alter or intensify daily routines and roles of worklife. Characteristics of office automation that lead to changes in the work role can thereby produce stress responses in the worker. Work-related stress is the psychological or biological response to conditions in the work environment. Stressors are the conditions that cause stress. 6 Stress can be either good or bad, but good stress is usually called challenge. . .eonard 1. Pearlin, Role Strains and Personal Stress, Psychosocitd Stress: Trends in Theory and Research (New York: Academic Press, 1984). A workers role is defined organizationally by the job heshe occupies. This model of stressor and stress response is a common model used and was originally developed by Hans Seyle and Walter Cannon in the 1920s and 1930s. For a good overview of this model of stress see, Hans Seyle, The Stress of Life (New York: McGraw-Hill Publishing Co., 1956). Office automation, by changing the conditions of work, can elicit psychological and biological responses; i.e., produce stress. Psychological stress responses to working conditions include challenge, boredom, anxiety, mental fatigue, depression, satisfaction or dissatisfaction, and feelings of security or insecurity. Biological responses can include chronic or periodic arousal, 7 muscle fatigue, headaches, and psychosomatic symptoms. This list is not exclusive. Any of these can occur among office workers. Office automation offers the organization a unique opportunity to create working conditions that challenge the individual worker and at the same time decrease the frequency of adverse stress responses. a --Chroni~ arousal refers to neural, hormonal, and immunological responses to external stimuli. In many instances the term neurohormonal arousal is used to refer to these distinct stress responses to indicate the controlling role played by the brain in initiating them. Traditional stress models emphasize that stress responses follow a U-shaped curve. Too little environmental stimulation leads to underload and stress responses such as boredom. Too
PAGE 134
A stress response sometimes comes from anticipation of an event as much as from the event itself. The anticipation of a computer going down, for example, can produce stress responses. New technologies often lead to new and conflicting expectations. The term technostress has been used to describe consequences of office automation: Technostress is a modern disease of adap tation caused by an inability t O cope with new computer technologies in a healthy manner. It manifests itself in two distinct but related ways: in the struggle to accept computer technology, and in the more specialized form of overidentification with the computer. ) The stress response is not technologically determined, although the word technostress conveys that impression. Many conditions that produce a stress response in the automated office are not the result of the new technology, and solutions to the problems can not often be achievecl through modifications in the technology. The consequences are contingent on how an organization implements new office technologies. Most office workers look forward to working on computer systems that automate manual tasks they would rather not do. 10 Table 5-1 shows a list of conditions that produce stress. These stressors are not unique to office work: they are also found, for example, in factory work. In fact, some people talk about the office becoming the factory of the future. Table 5-1 .Working Conditions Likely to Produce Stress Responses in Offices Working conditions in all offices: Increased workload coupled w!th. a) I I mlted job control, expanded job control Repetltlve task(s} In the job Machine pacing of work Lack of t I me for training to acqulre new skills Competing roles at work Lack of career o p po rt LI n I t Ies I n t hc organ I i7at 10 rl Working conditions in automated offices: Electronic mon Itoring as a form of su pervlslc)n and employee monltorl ng Electronic monitoring as a form of task feedback Higher expectations for speed of work coupled ~lth computer system response delays Off Ice work dependence on computer system an(~ system delays The computer med I at Ion of work and/or the pro blerms with the human computer d~alog Social Isolatlon with the primary I nteractlon the computer Increased soc I al contact and soc [al part IC i pat 10 n VI;) computer networks {~~)Ff,~ Off e $ 11 Ill I fq} ii.. I L ~ IT t These conditions have always been felt in offices; new technologies can intensify them. They can also ameliorate them. Typically, the adverse conditions occur together and especially affect lower level jobs. Most have been associated with low job satisfaction or with illness, Each can be alleviated with work redesign strategies, contributing to better employee health and attitudes toward work. Table 5-1 also shows a list of office conditions characteristic of office automation that produce stress that may permeate all levels of the organization. Sources of Stress Factors that bring about stress include increased workload, decreased control over work, repetitive work, machine pacing, inadequate training, ambiguity in role, consciousness of limited career opportunities, computer monitoring of work, unattainable expectations for speed, the abstraction inherent in computermediated work, and increased social isolation. Workload and Control Office automation can increase the number of tasks a person has to complete; increase the
PAGE 135
128 l Automation of America Offices complexity of the tasks; speed up the pace; lems, attempts are typically made to exert or increase the amount of information needed greater organizational control over what the to complete a task. ] Even the change to an worker does. This can be achieved more easily on-line system from a batch processing systhrough office automation. Greater organizatem can increase workload. 12 tional control tends to increase the problems These increases may or may not be transient. of repetitive or machine paced jobs; offering The ability to modify work strategies is a co-workers even less control and resulting in a ponent of the workers control over the job. ]3 vicious circle. The productivity of the worker If the worker has the ability to modify his/her is stabilized below optimal performance level. work strategies, the adverse stress responses Two patterns appear to be developing. The can be buffered and transient. When the worker key difference is the level of control. When conhas little control, the adverse stress responses trol is decreased, either through embedding tend to persist. Workload and job control in decisions in the system, electronic monitorconjunction affect attitudes about the job, the organization, and the workers health. Computer-based communication networks have been shown to lead to increased workload by making the transmittal of documents easier and thus more frequent. This leads to the professional having to sift through many memos to find the relevant one; the case of information overload. This also may be a transient effect that will disappear as the organization learns to utilize the system effectively. A primary motive for office automation is increased productivity, but many stress responses lead instead to increased absenteeism and turnover (especially under good economic conditions when jobs are plentiful). These responses have also been associated with decreases in performance and even with some instances of sabotage. 14 To alleviate these prob --Worklo~d is the level of demands the job places on the worker. Technological change can lead to decreases in the amount of time spent at a particular task. However, many studies have documented that workload more often increases with office automation: Jon Turner, Computers in Bank Cierical Functions: Implications for Producti\it~ and the Qualit~ of Life, Ph.D. dissertation, Columbia University, 1980; Michael J. Smith, et al., An In\estimation of Health Complaints and Job Stress in Video Display Operations, Human Factors, vol. 23, No. 4, 1981, pp. 387-400; Rob Kling, The Zmpacts of Computing on the Work of Jfanagers, Data Anal~sts, and Clerks, J$orking Paper, Public Policy Research Organization, University of California, Irvine, 1978. -Jon A. Turner, Computer illediated Wrork: The Interpla.v Between Technolog~r and Structured JobsClaims Representati~es in the Social Securit&v Administration, Working Paper, Department of Computer Applications and Information Systems, New York University, 1985. Control is the ability of the employee to arrange work in terms of speed, task priority, and ways of performing the task. Ilrod, op. cit.; SEW also Gavriel Salvendy and M.J. Smith (eds. ), hfachine Pacing and Occupational Stress (I,ondon: Taylor & Francis, I,td., 1981). ing, social isolation, or change in the job (e.g., the development of a single repetitive task), the interaction between increased workload and decreased control leads to adverse stress responses. Alternatively, when the increase in workload is associated with better opportunities to arrange the work, this can produce positive stress responses, or challenge, for employees. Repetitive Tasks Office automation can result in repetitive tasks, which lead to increased anxiety, boredom, and dissatisfaction with work. The introduction of a word processor or microcomputer can, however, also reduce the repetitiveness of tasksfor example, the secretary makes a few corrections rather than retyping a page. Taking breaks from the monotonous repetition provides psychological relief, but sometimes a need for breaks is overridden by the pressure of work. Machine Pacing When control over when to do a task is determined by the computer system, tasks are said to be machine paced. 16 When data entry or other tasks are machine paced, the office begins to mimic the factory assembly line. Machine pacing both increases the workload and decreases control. It can lead to anxiety, de.. ] F3rod, op. cit., p. 43. A task is said to be machine paced when the worker has no control over the initiation of the work c~cle or the duration of the work cycle. Typically machine pacing is associated with short-cycle repetitive tasks.
PAGE 136
Ch. 5Office Automation and the Quality of Worklife 129 pression, boredom, dissatisfaction, frequent health complaints and, decreased productivity with increases in error rates. 17 It also can allow higher pay because of an increased rate of production. Training The employee is sometimes expected to learn how to use new equipment with very little time allocated for training. Management expectations about productivity gains rise quickly, so the worker is faced with simultaneously maintaining the prior level of productivity, acquiring and mastering the new skills, and demonstrating productivity increases. He/she can be caught in a circle of expectations and increased workload demands, and may never learn to use the system effectively. Without the time to develop skills, the worker has limited opportunities for moves up the career ladder. The daily frustration over lack of time to learn may lead to psychosomatic complaints, as well as anxiety, depression, and dissatisfaction. Competing Roles Formal recognition of the responsibilities of managers, technicians, professionals, and clerical staff serves to differentiate their work roles, also recognizing interdependencies among members of the office group. 18 Role ambiguity results from the rebundling of tasks through office automation. The professional can do word processing and the secretary can do statistical analysis. Who does what is a key management decision. Without role differentiation it is unclear who is responsible for particular problems. Sal\wndj and Smith op. cit. .%x also: h]. N. (_k)rlett an d J. Richardson (C>dS. ), .$tress. \i-OI_k ~~S&, ~~d Productilritj ( New Im-k: tJohn ifile~r & Sons, 1981); and Robert Caplan, et al., (lot) Den?and.s and \lrorkers l{ealth. LJSDI 1 EJ$ ( N 10S11 ] 75-160 ( t$ashin~ton, I )(: {J. S. ( ;oternnwnt ]]rintin~ office, 197>1. Ij. (i, [>. (ohen, organizational F-actors Affecting Stress in the Clerical \lorker, human ,lspects in office .4 uton]ation, 13,(;. 1. Cohen (cd. ) ( /\nlsterdanl: l~lsetier Science Publishers, 19H I), pp. 33-42. Conflicting roles can develop when employees are expected to perform tasks that are not part of their job description or are not recognized or given social, monetary, or organizational rewards. With computer-based messaging systems, managers and professionals have to answer the phone, a task that the secretary may have done in the past. 19 One major source of role conflict for women is conflict between home and work roles. Either ambiguous or conflicting work roles lead to uncertainty in responsibility and performance and to anxiety, dissatisfaction, and insecurity. Lack of Career Opportunities Lack of career opportunities can create feelings of job insecurity and belief that one will be replaced by the computer. These feelings are exacerbated by the impression of limited job mobility. Traditionally a worker came into an organization, learned skills, and moved up the job ladder. Today companies more often hire people from outside with the skills needed. As career opportunities become less apparent the employee can become frustrated. Electronic Monitoring Electronic monitoring can take two forms supervision and feedback. This can be illustrated by the example of an information system that stores data on keystrokes and errors. One form of supervision has the computer system examine the workers performance and send the supervisor a message about her performance compared with prior records and organizational standards. The worker has no responsibility for when the monitoring occurs or how the information is used. 20 Another form, job feedback, lets the employee use the inforH. C. Amick and J. Damron, Considerations in I~efining office Automation: .4 Case Stud~ of the Eastern Africa Region of the Morld Bank, Proceedings of the First i) S.4-,Japan Conference on Human-Computer Interaction, Honolulu, ii:~uaii, CJavriel SaJ\rendy (cd. ) (Amsterdam: Else\ier Science Publishers, 1984), pp. 439-.! 45, This example also points out the problem of separating these factors from other stress factors. In this situation. the worker is not onls being monitored but also ma~: be machine paced and has no control oker the task or the Job.
PAGE 137
130 l Automation of America s Offices I I Photo credit Cornmunlcattons Workers of Arner(ca In the communications industry, workers can be continuously monitored. These charts show the monitoring of employees by the Average Work Time System. Every call is timed and the total volume of calls handled per day by all workers is then used to calculate an average number of calls for a given period of time. An AWT figure of 100 means that the operators performance was average mation to find out, for example, how many files were completed in a given period of time, and modify her performance accordingly. The first often produces adverse stress response and the latter positive challenge. With electronic monitoring for supervision, maximum consistency and reliability in performance are sought. One distinction between traditional work monitoring and electronic monitoring is that the former is intermittent and the latter continual. Continual monitoring places added pressure on the employee to produce at the pace and level established by the machine. This leads to increased anxiety, fatigue, psychosomatic complaints, and job dissatisfaction. 21 The employee, knowing he/ she will be expected to perform at a specified -ecause the terminal provides the worker with all the necessary information to complete any task, there is no need for the employee to Iea\e the workstation. This constrained posture for extended periods of time is a key element in \isual and musculoskeletal problems.
PAGE 138
Ch. 5Office Automation and the Quality of Worklife l 137 level throughout the day, feels chronic arousal and job dissatisfaction. 22 Electronic monitoring can also offer challenge and job satisfaction, if it is designed and used as a form of task feedback. Traditionally feedback on performance and product quality is usually notice that an error has been made. Or it is received infrequently (in quarterly or even yearly evaluations by the supervisor), and the worker begins to believe that quality is not a major concern. Quality checks through an on-line work monitoring system can improve the reliability of performance, and increase the responsibility for the product; this can lead to greater job satisfaction and greater commitment of the worker to the organization. 27 Typically, electronic monitoring has been carried out among support staff and supervisory staff, but as microcomputers and minicomputers are linked with larger mainframe systems, any person can be monitored. One example of productivity y monitoring of professionals is from a loan administrator using a Management Information System (MIS): Each employee is watched by the system, ,. To be tracked on the system is to be monitored by it on a daily basis with regard to every conceivable item of work which could be performed on the job. The MIS does perform some essential functions in helping to manage the enormous work load, hut often it was looked upon as an excuse for increasing the work load even more. Preparing the daily MIS input sheets at the end of the day is a dreaded duty, done hurriedly before leaving the office. The MIS theoretically tracks everything which the loan administrator performs during the course of the day. The entire job of the loan administrator is dissected into every minute function which might be performed. Each function is assigned a certain time standard for how long it should take to be performed. The system calculates each Eor an example {)f how prior knf)wlcd~c of an e~ent can produce a stress response we the work of hlm-ianne Frankenhauser, (oping J\ith Stress at t$rork. Znfernal iOn:Jl ,Jotlrna) of l?wft h Swvjces, ~ol. 11, No. 4, 1 W 1. pp. 491 -h 10. Richard l+;. Ralton, From (ontrol to Commitment in t h( Workplace, Iiar\-~ird llusinets.~ Retiew. Jo]. 35, No. 23 1985, pp. 76-~ i. workers daily productivity as a percentage figure based on what the employee reported as having done that day. z If a professional, manager, or secretary is making entries into a calendar, then the persons schedule can be examined through access to the files. In one large American company, the Chief Executive Officer plans to use the computer system for just this purpose. Employees in the networked system can also get more immediate feedback about a project to which they have contributed. This provides them with a broader picture of what the organization is doing, and so leads to increased satisfaction and commitment. Higher Expectations of Speed A subtle effect of office automation is the generation of higher expectations of speed. This can lead to a distorted sense of time and distorted expectations about behavior. One characteristic of the computer system is the ability to respond quickly to commands. The manager and the clerical worker alike can develop a new time reference. Days, hours, and minutes take on new meaning as time is compressed and accelerated. Jobs that took days before computerization are expected to be done in hours. Software that was once appreciated for its speed, such as VisiCalc, is suddenly viewed as clumsy and S1 0 W 26 The worker begins to notice response delays in the microcomputer or larger computer system. As he/she becomes more proficient, he/she develops a high expectation of speedy responses from the computer at known intervals. When delays occur, the pattern of work is interrupted as the worker is forced to waste time waiting for the system to respond. This leads to anxiety and dissatisfaction. When the delays persist above a certain time there is an increase in the error rate for keystrokes. Alan Mrestin, PnItic> issue.~ in the .filonitorin~~ of F.mplo-~ee 11(xk m 1lITs in the (lffice ~;n~ironnwnt, contractor r(port prepared for the office of Tmhnolo~q Assessment, 19H5, Nlark I)[~tt~, *G]? JIelch l)owering Firm Into (;lohal (?ornpetitor: Part [, The T!-a.shington lo,st, .Sept. Y;3, 19tiI, Jlrod, op. cit.
PAGE 139
132 l Automation of America offices This new time reference results in higher expectations of oneself and others. A worker may seek to discover ways of circumventing organizational procedures to get work completed more quickly. Going through channels or doing excessive paperwork becomes a barrier. This also can lead to increased anxiety and tension and feelings of work pressure, or to a desire for social isolation because one would rather work with the computer than with a persononly the terminal can respond quickly enough. This tension requires some form of stress management. Like the individual, the office can become dependent on the computer system and vulnerable to system delays. This is perhaps largely a transient problem associated with the implementation of a new system, although changes in both software and hardware will continue, and the work process will become more and more dependent on the computer. When the system breaks down the worker looses control over the work and it piles up. Messages cannot be sent to coworkers in other buildings, memos cannot be printed or distributed. A sense of anticipation develops in the worker when will the system be back up and running? One study in a Swedish insurance company found that when the computer system went down workers responded with increased arousal, higher blood pressure, fatigue, and feelings of being rushed (work pressure). Computer Mediation of Work The computer mediation of work is likely to produce stress responses that vary from boredom and job dissatisfaction to challenge and increased productivity. Three characteristics of computer-mediated work that are likely to produce stress responses are: 1) human judgments being built into the system, Gunn Johansson, Computer Technology: Stress and Health Relevant Transformations of Psychosocial Work Environments, Proceedings of the First USA-Japan Conference on Human-Computer Interaction, Honolulu, Hawmi, Gavriel Salvendy (cd.) (Amsterdam: Elsevier Science Publishers, 1984), pp. 347-354. In this study, VDT operators performed data-entry tasks and were compared to non-VDT operators performing similar tasks. 2) the abstraction computer dialog. of work, and 3) the humanJudgment is fundamental to work, yet a principal goal of automated work can be to reduce judgment and therefore error. Alternatively, automation can be used to create more opportunities for exercising human judgment. The key to the former approach is the substitution of computer algorithms or decision rules for judgment. This necessitates a formalization of skills traditionally obtained through experience. Problems arise when the worker is not provided with new opportunities to exercise judgment. For example, automated systems may assign priorities to tasks or refer accounts to employees as they complete previous assignments, decisions previously made by supervisors. This is usually done when automating tasks that in any case are very standardized. The computer mediation of work extends to professionals and management. Attempts have been made to define decisionmaking so that a level of reliability and predictability in judgments can be maintained. The shifting of partial control from the individual to the computer system also is a defense against data contamination that could result from unlimited access to an on-line system. 28 Risk-taking behavior and creativity, especially of professionals and managers, can be constrained when decisions are embedded in the computer system. The meaning and challenge of work disappears and boredom and job dissatisfaction ensue. A constant state of chronic arousal places a biological burden on the worker. The airline pilot or the nurse who monitors a control panel instead of the plane or the patient, can become bored and dissatisfied, feeling that their training is useless if the computer is going to make the decision. g Although uncertainty may lead to errors, freedom of choice leads to creative and inspired decisions, which %hoshanah Zuboff, New Worlds of Computer-Mediated Work, Harvard Business Review, vol. 60, 1982, pp. 142-152, (hereafter referred to as New Worlds, 1982). Shoshanah Zuboff, Problems of Symbolic Toil: How People Fare With Computer-Mediated Work, Dissent, winter 1982, PP. 51-61 (hereafter referred to as Symbolic Toil, 1982).
PAGE 140
sometimes go against the grain of organizational logic, and it is this logic that is typically programmed into the computer. so Nowhere is the use of computers to approximate human judgments potentially so dramatic as in the area of expert systems. 31 The goal of developing an expert system is to place in the computer the knowledge and decisionmaking processes of the best workers at a particular job. This can enhance a job by providing large amounts of information in a manageable form, allowing a person to ask questions that were difficult to ask before. Expert systems can evaluate decisions and suggest alternative solutions to problems. The system can serve as a learning tool for new workers, who by asking questions and giving answers can simulate experience that may otherwise have taken years to encounter. Expert systems for use in offices are still in an early phase, but one area already impacted is insurance underwriting. (See box A for an example of the human-computer dia.. llistori~~ll?, researchers ha~e argued about whether tht bureaucratic structure of the organization constrains human beha~ior by increasing centralized control and explicitlsr defining the rules by which the tasks should be performed and the job elaluated. These trends can reduce the effectiveness of the organization as well as the emplojee. The implementation of office automation can allow man~ of the rules to he embedded in the s~stenl. For discussions of this idea see, ill ax 14eber, The Theor\ c)f Social and l
PAGE 141
734 l Automation of America Offices Box A.Expert System Used in Insurance Underwriting: An Example of the Decisionmaking Process A publishing company in California has decided to insure their data processing facility. They hire insurance agent Smith to find the best possible policy for them. Mr. Smith approaches Ms. Jones of Insurance Company West to develop a policy for the data processing facility housed in a 25-year old brick building. Ms. Jones goes to the computer terminal and starts a file called Publishing Company Account. She then accesses the expert system to get assistance in defining the parameters of the policy. First, she must enter information about Mr. Smith, the agent. The computer system has a large database of agents the company works with, describing the type of work they do, their reliability and the quality of the information they provide. (This database is updated quarterly by the company to enter information on new agents and update the information on old agents.) The system evaluates Mr. Smith. Uncertainty coefficients are then estimated by the computer system based on information about Mr. Smith. The expert system uses this coefficient to determine the most appropriate questions (prompts) to Ms. Jones. Any equations used in final recommendations by the system take into account the agents uncertainty coefficient. Then Ms. Jones enters the basic information about the facility. The computer responds that in California, 20 years ago, reinforcing bars were not mandatory in brick buildings. Therefore, earthquake damage must be excluded from the policy. The system may then prompt Ms. Jones for information about smoking-is it permitted in the data processing facility, and in what rooms? The system prompts Ms. Jones about whether flammable solvents are stored in the room. A value may be assigned to designate the importance of this information. The system then combines this information with other information. It prompts Ms. Jones when necessary, and comes to a decision about the policy just as an expert underwriter would. SOURCE: Baeed on an interview with Steve Weyl of Syntelligence, to perform the same task over and over within the rules defined by the computer system. In the extreme case, the worker monitors information as it comes to the screen, but all judgments are made in the system or by some other person at another location. For workers whose job involved interaction with people, this change can mean a loss of the sense of the meaning of the work. Computer-mediated work is characterized by a human-computer dialog. The quality of the dialog is important to the quality of working life. The user gets from the software information about the current state of the system, the work, how to proceed or return to an earlier part of the process, the future stages of the production process, and the consequences of further command sequences. If the software does not provide the needed information, there is decreased control over the work, bringing Inc., Santa Monica, CA, 198S. anxiety, and job dissatisfaction. 36 The worker may use the system less and less, or not attempt to fully use the system capabilities; a circle develops leading to further anxiety and less productive use of the new office technologies. A satisfactory human-computer dialog can counter the sense of abstraction, giving the operator a greater sense of control. Thus, transitory problems with office automation may be alleviated with appropriate attention to the human-computer dialog. When the dialog uses concepts and phrases that match the operators task vocabulary, 37 this encourages the use of the system and increases the likelihood of greater productivity y. Jon Turner and Robert A. Ka-asek, Stjftware ~rgc)nOrnks: Effects of Computer Application Design Parameters on operator Tasks Performance and Health,llrg~~~~lj~s, ~o]. ZT, No. 6, 19S4, PP. 663-6!30. Turner and Karasek, op. cit.
PAGE 142
Ch 5Off/cc Autormtion and the Quality of Worklife l 13 5 However, as users become more advanced they begin to prefer more formal dialog. 38 Men u driven systems guide the novice user; as the user becomes more knowledgeable, this may become tedious; the user feels held back by the system. This can also produce anxiety, boredom. and job dissatisfaction. More transitionality in the way that people and computers interact or interface will alleviate this source of stress. 39 In a dialog based on transitionality, the user can designate changes in the human-computer dialog as he/she becomes more proficient. Social Participation or Isolation The office is a social environment, and as the technology changes, the social relationships of work may also change. This is important because the social group can be a buffer between the worker and other characteristics of the work environment .4 Communication and social participation in the work process is central to the productivity of the work unit and the maintenance of the organizational culture. This social network can be particularly important in a dynamic work environment. At times of heavy workload, when peak worker performance is required, the resulting adverse stress responses affect the employees capability to perform. A social environment capable of ameliorating these conditions for the worker is often not considered in evaluating the affects of office automation on the quality of working life. Yet, new office technologies can either create more opportunities for communication and supportive interaction or close off channels. There is a tendency for workers to spend more and more of the day interacting with a computer terminal. In Japan, researchers have recognized this potential: There will be more need for lounges and discussion rooms and the like, to break the routine of stress from the machines. As tasks become more mechanical and isolating, more group activities and worker clubs and incentive systems need to be developed t o keep up team spirit and morale. The worker may spend so much time at the terminal because he/she is electronically monitored and paced or because there is no need or opportunity to interact with others, resulting in social isolation, which has been shown to be associated with depression, anxiety, job dissatisfaction, muscular fatigue, and psychosomatic symptoms. 42 When the computer becomes the primary source of interaction in the office, coordination of the work process is transferred from the social to the technological milieu. .4 technological form of communication develops and both work and communication tend to become more formalized. Whether this will adversely impact the quality of worklife is unknown at the present. One of the primarly reasons for social interaction is to communicate value judgments via both verbal and nonverbal means. The content of the communication may be important although not directly pertinent to the work process. For example, the employee may be performing poorly because of circumstances at home, or an illness; when evaluation is based primarily on computer monitoring, supervisors may not recognize these circumstances 43 Alternatively, increased social contact can be provided via the computer system or via user groups, because computer-based communication can remove both time and distance
PAGE 143
136 Automation of America Offices constraints on communication. 44 This can lead to greater employee involvement and commitment to an organization along with increased job satisfaction. Strategies for Alleviating Adverse Stress Responses The strategies for alleviating adverse stress responses simultaneously encourage challenge, or positive stress responses. Three different approaches concentrate on different parts of the model that was shown in figure 5-1 and have different goals. Stress management programs focus on the individual, job redesign programs focus on the job, and sociotechnical programs focus on the work process. No single approach is always adequate. A participatory style of management in the implementation of new office technologies will help to identify the methods appropriate for a successful implementation. Stress Management Stress Management refers to stress reduction programs tailored to the needs of the individual, 45 providing training to deal with conditions either at work or home. 4G These programs can include meditation, progressive muscle relaxation and biofeedback, and cognitive/behavioral skills. The advantages of these approaches are that they do not require any changes in the work process, organization, or the physical environment. They assume no fault for the stress responses on the part of either management or the worker, but place on the worker the responsibility of coping with adverse stress responses. In one Swedish study, it was found that 50 percent of the system messages would not have taken place without the system. Cited in Alexia Martin, Office Automation: Catalyst for ChangeManaging the Transition to the Automated Office, SRI Research Report No. 694 (Stanford, CA: SRI International, 1984). I,awrence R. Murphy, A Comparison of Worksite Relaxation Methods, Human Aspects in Office Automation, 13. G.F, Cohen (cd. ) (Amsterdam: Elsevier Science Publishers, 1984), pp. 257-265. Recently, new software has entered the market that assists the worker in managing stress. One such program is Relax, which uses biofeedback techniques along with muscle relaxation methods for managing stress. See Joseph Hager, The Stress Manager, ,, ~>cjr ~ror]d, September 1984, PP. 22~-232 Data have not shown conclusively that these programs are cost effective. Their disadvantage is that working conditions are not modified to remove the source of the adverse stress response. In fact, research has shown that to realize the benefits of stress management typically requires extensive and costly training of the worker, not the typical market offering of a two-day seminar. There is no credentialing of the trainers and there are vast differences in the quality of the training. Therefore, although stress management programs are popular, the employer is often faced with having to make a decision about them without sufficient information about the quality of the training and the likely results. Job Redesign Job redesign or task redesign are both ways to modify the working conditions to change stress responses. Office automation presents an opportunist y to develop job or task redesign strategies. The central thesis of job design is that through changes in characteristics of a job such as feedback, autonomy, task identity, and skill variety, the internal motivation and thus the performance of a worker is enhanced. Task redesign focuses specifically on task(s) rather than on a job. The scientific management philosophy is a method of task redesign, which in office automation primarily involves changes at the human-computer interface and the allocation of tasks between the worker and the computer system. Unlike stress management, job redesign changes the conditions of work. When jobs are designed with little meaning and little responsibility, and the worker has no information about how well he/she is performing, there can be motivational problems. 47 For example, the overrationalization of clerical work results in repetitive jobs, lacking responsibility and offering little scope for personal development. 4g Richard J. Hackman and Greg Oldham, ltrork liedesign (Reading, MA Addison-Wesley Publishing. 1981). These types of jobs may lead to high labor turno~rer, absenteeism, and recruitment problems. This leads to inefficiencies through lost time, recruiting and training, and costs asso(continud)
PAGE 144
Ch. 5Office Automation and the Quality of Work/i/e 137 Many job redesign programs seek to expand these jobs. 49 Job redesign, unlike stress management, requires that management recognize a problem, and change working conditions. This could involve relatively high initial costs. Another disadvantage is that job redesign efforts tend to focus on a single class of jobs rather than on the work process; and on internal motivators such as self-esteem rather than external motivators such as career ladders and pay. 50) There may be few opportunities for successful job redesign if an organization has only a limited set of tasks to be done. Sociotechnical Systems Sociotechnical systems developed as a counter strategy to the traditionally structured organization. Instead of each worker doing a single task or set of tasks with no clear connecciated with o~errnanning to deal with absenteeism. For an example of how to measure the impact of beha~ior on organizations see, Jla}ne 1:. {ascio, Costing liuman Resources: 7he f+inancial Impact of lleha~ior in organizations (Boston, N1:4: Kent I)ublishing C()., 19H2}. Iror an example of a job redesign program in white-collar work see, Rodgyr \tr, (; ri ffeth, \loderation of the hlffects of Job F;nrichment b~ I)articipation: .4 I,ongitudina] Field F~xpediment, organizational Beha\ior and human l)ecision lro(PSS, \ol, 35, 1985, pp. 73-9;1, I n this experiment, desk receptionist jobs were increased in terms of task identity, skill ~ariet~, autonom~. and j oh feedback, resulting in increased j oh sat isfaction and self-esteem. In re~iews of job-redesign experiments conducted o~er the past 20 years, it has been found that a key predictor of n~aintaining the changes in producti~it}, quality of work, abwnteeism, and turno~er was an organizational recognition of t he new tasks and skills a person was using in his her work. HasicallJ, the workers wanted changes in the reward s~stem to reflect the new job. I f these changes did not de~elop, then man~ of the problem~ the jot>redesign strateg~ was intended to resol~re returned. tJohn P;, Kell~, Scientific ,Ifanagement, ,Job Redesign and It ork Ierformancr { Yew }ork: Academic I)ress, 19821. tion to an end product, workers collaborate to produce a product. Many offices already have small working groups such as the secretary and the manager. However, office automation sometimes changes this, and may dissolve familiar interdependencies. In sociotechnical work systems, members can share decisionmaking about who is to do what tasks and how they will be executed, with each person capable of doing a variet y of tasks. The pay/reward system can be tied to the skills in order to encourage skills acquisition and career development. There are very few examples of such semiautonomous work groups in white-collar organizations; and those few typically involve lower level staff .5] The advantage of this approach to stress reduction is that it requires changes not just in the job but in the work process. The principle disadvantage is that changes required at the organizational level and training of workers in group processes such as decisionmaking may be costly. But in addition to reduction of stress there maybe additional benefits such as changes in pay structure to reflect the new skills demanded by computer-mediated work, with the incentive to continue learning, increased sense of job security, and reconfigured job ladders. Continuous electronic monitoring may no longer be needed if the group makes its own decisions about the speed and pace of work, monitoring each other.5z Recently, there has been a growth in the use of sociotechnical system programs in all le~e]s of the organization. For examples, see Cal\in Pa~ra. Itlanaging .Veu Office 7echnolo~~?: ,4n organizational Strate~ { New }ork: Free Press, 198:31. -There also can be problems with autonomous groups such as exclusiiity and excessi~.e pressure to conform. SECTION II: OFFICE AUTOMATION AND HEALTH, ILLNESS AND DISEASE The video display terminal (VDT) is the dewas depicted in figure 5-1, the VDT is only vice most often discussed in relating illness one part of the work environment. Much of and disease patterns to office automation. It this section will focus on the VDT, since most has become a symbol or a scapegoat. But as of the research has done so; but the VDT must
PAGE 145
738 l Automation of America Offices be considered in the context of workstation design, office design, and the task and job. Occupational health research has traditionally focused on blue-collar work; office work has been considered the least hazardous of occupations. The growth of the white-collar labor force and the ubiquitous applications of the new office technologies demand that this assumption be reexamined. By the year 2000, it is estimated that most white-collar workers will use a terminal. Therefore, office automation requires the attention of public health officials because of the magnitude of the work force at potential risk for any potential illness and disease outcomes. 54 The list of potential exposures in the office is large and diverse. Table 5-2 showed the list of exposures for workers in all workplaces, many difficult to measure because they occur 1 Iealth has been defined b~ the World Health Organization as the state of complete social, psscho]ogical, and ph?wical well-being and not the absence of disease. Illness is a subjective state where the indi~ridual is aware of dysfunction. Disease is generall~ a pathological process that creates and ends in an altered physiological or ps~chological state. Risk is the probabilit~ of a disease or an illness free individual developing a speclflc illness or disease over a specified period of time. It is conditional on that individual not developing an other illness or disease during that period of time. The VDT IS sometimes placed alongside other technologies (such as the microfiche reader). Some of the symptoms workers report can be traced to the different lighting requirements of different technologies In the workplace Table 5-2. Exposures in the Office Other Than the Video Display Terminal Indoor air pollutants: Micro-organisms Cigarette smoke Dust and particulate Chemical emissions from machines: Ozone from photocopiers Methyl alcohol from mimeographs Ammonia from blueprint machines Chemicals in paper/office supplles: Formaldehyde from carbonless copy paper Noise from office machines, ventilation systems, and conversation Static electricity Low humidity Inappropriate lighting design: Excess lighting Insufficient Iightlng Glare Overcrowding and lack of privacy Poor ventilation Awkward and static work postures NOTE This IISI does not Include any factors relate J 10 either the lask job or organ !zatlon wh!ch may be associated with stress responses These arP (j I scussed I n another sect Ion SOURCE Jeanne M Stellmarl Of f!ce Automation A Publ(c Health Pe~>pectlve on the Potential Acute and Chrontc Effect: paper presenle[l at the Of flee of Technology Assessment Symposium or) f he Impacts uf Off !ce Autornat!on and Computef Mediated Work on the oudht y of V~ork I Iff! Dec 1012 1984 at low levels. ss New office technologies must be considered in conjunction with these exposures to understand their full impact. A major problem is the lack of long-term statistics on the illness and disease of office workers. The most comprehensive and thoroughly analyzed national statistics are 25 years old, from the 1960 Health Examination survey. An examination of the medical literature indicates that little valid data exists even on absenteeism and sick leave rates beyond specific industry studies. This makes it impossible to compare historical trends of the morbidity and mortality of office workers with present rates. 57 Jeanne \l, Stellman, office Automation: A Public Iiealth l)erspecti~e on the Potential Acute and Chronic Effect s, paper presented at the office of Technolo~ Assessment S~mposium on the Impacts of office Automation and ComputerMediated Work on the Quality of Work]ife, Dec. 10-12, 1984. National Center for health Statistics, Selected l{ealth (characteristics b~ Occupation: United StatesJul~. 1961 -June 1963, Vital and Health Statistics, Series 10, No. 21 (14ashington, DC: U.S. Goternment Printing office, 1965). The problem with using industry studies to establish national trends is exemplified by work conducted in the late 1960s on the relationship between social class and coronar~ heart disease. 1 ndustr? studies showed a positi~e relationship between (continued)
PAGE 146
The four major categories of outcomes to be discussed are visual system outcomes, musculoskeletal system outcomes, stress-related outcomes and reproductive outcomes. 58 Causal relationships between some characteristic of the office environment and the outcome are assessed using the standard epidemiological criteria for determining causality, listed in table 5-3. The risk of these outcomes can be reduced by job and task redesign, office design, and workstation design. The question remains open as to whether new office technologies lead to an increased risk of illness and disease. Visual System Outcomes Fifteen years of research has shown that VDT workers report a high prevalence of visual social class and coronary heart disease, The higher ~he social class the higher the incidenw of coronary heart disease (Cl 11)). I n communit~ and national studies, the relationship found was ne~ati}w the lower the social class the higher the incidence of C 11 D. one reason for this inconsistenc~ is selection hias, J$hite-collar workers are likel~ to be kep~ longer bJ the organization, and maJ be more likel~ to return after serious illness. Thus blue-collar workers aro no longer in the industry to he counted, but they, are still found in the comrnunit~., and arc counted in community studies. This phenomenon is the 4 health~ worker effect that over time leads to underestimates of the pre~ralence of an outcome in industrs. A nether health outcome whose appearance is associated with office work is dermatitis, There ha~.e heen too few cases to reliahl~. identify a singl( cause. Some researchers ha~e h~pothesizeci that this is due t~) an elect ros~atic field that huilds up between the terminal and the operator, The unanswered question is wh~ the dermatitis onl~ appears on the face. I t maJ be that when one cleans off the screen and then touches ones face the dust and fib[lrs collected on the screen are transmitted to the face. I n a recent stud~., it was found that setreral dermatological problems occurred in a greater frequenc~. among 1 1) operators The researchers were unable to determine whether this was due to phjsical factors such as the electrostatic field, ps~chological factors, or pure chance. Thej concluded that Working at the t 1) ma~r not create but exacw-hate dermatological prot)lems. Carola I.indcn and (Jan P]. Wahlher-g, JIor-k at \ id(v) I)isp]a.b. Terminals. .~\ n Plpidetr]io]ogic;i] ] lealth [n~e.sti~ra t ion of of fi[v IS \ ing workplace exposu rt~s SLYI the ( )TA report, lr[>~ (n(inAr lflnt.s.s ;]nd Injurj in the llorkpi;]~v, ().+1-1 1-256 (Jfrash ington. 1)(: ( S. (io~crnm(nt I%in( ing office, 19851. Ch. 5Office Automation and the Quality of Worklife 139 Table 5-3.Epidemiological Criteria for Assessing Causality 1, Strong association of factor to outcome 2. A dose-response relations i p between the factor and the outcome 3. Clear temporal relationship between the factor and outcome 4. A consistency of findings across studies 5, A biologically plausible explanation for the observed association NOTE These crlterla do not al I have to he satlsfled at one t (me and In one st ud~ to suggest a rel at Ion sh I p or Iac k of a rel at I on sh I p between the ne A off Ire technology application and a worker s health For example (f rICI biologically plauslble ex planatlon car be offered for the obser, ed assoc! at I on that may be due to I nadequate kn owl edge or measu rer~l vr~ t techn I q ues Thls makes It very dI f+ c u I t to refute a rel at I onsh I D bet w ee r, a factor and an ou I come based 01 a I ac k of blolog~cal or blom PC har] I r al evidence SOURCE A M Llllenfeld and D E Llllen$eld Founcfaf~ons of Ep/demlo/oqk ( Ne A York Oxford Unlverslty Press, 198(II strain, ranging in various studies between 47 and 91 percent of all operators. Comparison to other workers suggests that VDT work is associated with an increase in visual symptoms across most occupations, but especially in those with heavy visual demands. Table 54 shows many of the factors that have been associated with asthenopia (visual fatigue) and perception problems in automated offices. Visual symptoms are associated with specific characteristics of the individual, the physical environment and the work process. But based on the current evidence, causal relationships cannot be established between office automation and visual pathologies. Asthenopia and other acute 60) visual system outcomes reported by VDT workers are listed in table 5-5. A major problem is the lack of correlation between reported visual symptoms and clinically observable signs. Furthermore, there are no national data with which the prevalence rates can be compared. In those studies where visual functions (eye movement, visual acuity, pupil size, blinking, accommodation and vergence) have been objectively measured, no significant differences have been observed between non-VDT and VDT work, or over periods of time at VDT work. Yet some studies have shown that the amount of time spent at i+\c.utt outcomes arc transi(nt and closel? associated in time with th~, (~xposure, (hronic outcomes are persisttint and ~ndu ring an(i distant 1} or i nd i r[ct lJ associated vit h the exposure.
PAGE 147
140 Automation of Americas Offices Table 5.4.individual and Work Environment Factors Associated With Visual System Outcomes Individual: Age Uncorrected refractive errors Work environment: Glare on the screen Flicker and jitter Resolution of characters High contrasts between source document and screen Tasks requiring long continuous looks at the screen without a break Air humidity High visually demanding task, like data entry and data acquisition Low control jobs where the worker is unable to take a break to relieve the load on the visual system NOT = A~though the professional who wo~ks-at the VDT IS not as llke~y to report as many vtsual problems as people doing data entry acqulsltlon or dialog tasks It IS important to recognize they tend to report more symptoms than workers not us I ng [he V DT For a comprehensive revl ew of the I I terat u re see Ulf Bergqvlst, Video Display Terminals and Health A Technical and Medical Appraisal of the State of the Art Scancffnavfarr Jouma/ of Wor k Errv/ronrnenf and Hea/th VOI 10 SupP 2 1984 SOURCE Off Ice of Technology Assessment Table 5-5.Visual System Outcomes: Known Acute or Potentially Chronic a Known acute outcomes: Asthenopia: b Burning eyes Itching eyes Irritated eyes Perception problems: Double vision Blurred vision No existing data shows that working at a VDT can lead to the development of cataracts, 61 but neither has the VDT been thoroughly absolved. While in some industries people have worked at VDTS for years, no study has examined the prevalence of cataracts in such groups of workers. No long-term study has yet shown either clinically observable changes in the eye functions or pathological changes in the eye. There is evidence that work at a VDT leads to more frequent changes in eye glasses. 62 This has been attributed both to the aging of the workers, and to a greater tendency for VDT workers, as compared to others, to identify minor visual problems. A possible alternative explanation is the development of acquired myopia (nearsightedness), which is characteristic of close visual work such as reading. If working at the VDT is found to increase the chance for acquired myopia, the visual health of a large sector of the population will be affected. As the debate about the effects of VDTs on the visual system of workers continues, there is a growing need for a well-designed longitudinal study to determine whether the acute conditions persist. Since little is known about potential biological or biomechanical mechanisms for the initiation of pathological changes in the visual system attributable to new ofPotential chronic outcomes: fice technologies, research should also focus Acquired myopia here. a There are no c-h ron IC outcomes that have been observed I n any c I In Ical I nvest I gallons of VDT operators bAStheflOpla IS a term used to describe vtsual fatigue or weakness N O clear defl Eye exams are recommended by the Nan It Ion of visual fatl gue IS employed In the research and it I s sometl mes referred tional Institute for Occupational Safety and to as visual strain SOURCE Off Ice of Technology Assessment Health (N IOSH) before beginning VDT work and periodically thereafter. G3 In many ina terminal is directly proportional to the number of workers reporting symptoms and to the extent of their complaints. That complaints cannot be tied to organic change does not, in other words, mean that they are not real. There is also evidence that visual problems can influence the development of musculoskeletal symptoms. The acute outcomes in table 5-5 can be reduced with changes in the workplace, such as appropriate lighting levels, computer terminal design, workstation and job redesign, and prescriptions to correct refractive errors. National Research Council, Video Displays, Work and Vision: Panel on the Impact of Video Viewing on Vision of Workers (Washington, DC: National Academy Press, 1983). However, in a recent workers compensation case, the worker received an award for incipient cataracts. The workers compensation board found there was a recognizable link between the employees occupation and disability-see, Legal Rights for VDT Users (Cleveland, OH: Working Women Educational Fund, 1985). Wlf Bergqvist, Video Display Terminals and Health: A Technical and Medical Appraisal of the State of the Art, Scandinavian Journal of Work, Environment, and Health, vol. 10, SUpp. 2, 1984. J. Donald Millar, Assistant Surgeon General and Director of N IOSH, Statement before the Subcommittee on Health and Safety, Committee on Education and labor, U.S. House of Representatives, May 15, 1984.
PAGE 148
stances, visual complaints are in part due to uncorrected refractive errors. Eye examinations would identify and allow correction of this problem. There is no evidence as to how often eye exams should be given for good visual health. They are particularly important for people over 40 years of age, when the visual system begins to change with the onset of presbyopia (farsightedness). Musculoskeletal System Outcomes Musculoskeletal disorders rank very high among health problems in the frequency with which they limit activity. 64 Over 10 million people report limitation of activity due to musculoskeletal disorders as compared for example with about 7 million reporting limitation of activity because of circulatory problems. Musculoskeletal disorders are the leading cause of disability of people in their working years. From 1973 to 1983, industry data sources report low back pain as the primary cause in 20 to 25 percent of the yearly workers compensation cases. 65 As the working population ages, the prevalence of musculoskeletal disorders and the potential costs to individuals and society increase. No recent data are available to provide more accurate estimates of the extent of musculoskeletal disorders in the aging U.S. population, and there is no breakdown of the older data by occupation. Therefore it is difficult to estimate the impact office automation has had on the prevalence of musculoskeletal disorders. A national data collection system is needed to determine the contribution of the workplace to the prevalence of musculoskeletal system outcomes. NIOSH has identified musculoskeletal disorders of an occupational origin as one of the top ten health problems affecting workers today. Over the past 15 years, they have been ,Jennifer Ii. Kelsey, Epidemiology? of ,}lusculoskeletal I)isorders, Llonographs in ~;piden~iolog~ and Hiostatist i~.s, YTO1. l] ~ New York: oxford Uni\ersit~ Press, 1982). Cited in Charles F. Spakell and ~$illiam F. N1cKeon, Pre\enting I,ow f] ack f)ain in I ndustr~, Bu.wne.s.s and }Iealth. Jo], 2, No. 6, 19x5, pp. 16-19. This is supported h~ unpublished anal~ses of H 1..S data. one of the major categories of self-reported complaints by VDT workers, a high percentage of whom report pains, stiffness, cramps, and numbness in the back, neck, shoulder, arms, and hands. Comparison with other office workers reveals an association between VDT work and increase in self-reported symptoms in specific types of office work. 66 Table 5-6 shows factors that have been found to be associated with musculoskeletal system outcomes among VDT operators. Few studies have weighed the relative contribution of each factor 67 Characteristics of the individual, workstation design, and the work process are associated with the reported symptoms. There is a suggested association between length of time at the VDT and an increased levels of reported musculoskeletal symptoms, 68 but the current scientific evidence is inadequate to demonstrate a causal relationship between any The data usLId to make the ohsertat ion are inadequat~~ t{) make an}. strong asscr-tion across all j ohs where the worker uses the computer terminal. ITor example, stx S, 1,. Sauter, et al., tJoh and ] ]ealth I nlplicaLions of \I)rI (Ise: Initial ltesults of the \$isconsin-N 10S1 f Stud~, ((jrllr~][lni(:iticjns of t he.4 sscxiat ion of Comput kg .ll:/chinar~-, I.ol. 26, pp. 2H 1-2!)4, 19H~l. Nl ichael .J. Smith, I;rgonomi( and Str~ss Aspects of ( )f fice .,\ut omation, paper presented at t }1( of ficv of echnolo~? SJm~posiunl on the 1 mpacts of office )\uton~ation and (onlputerNlediated Work on the Qualit} of Worklife. I)ec. 1 ()12, 1984. Table 5.6. Individual and Work Environment Factors Associated With Musculoskeletal Outcomes Individual: Age Existing musculoskeletal disorders Visual system characteristics Near-s ightedness Presbyopla Bi or trl-focal eyewear use during VDT operation Work environment: a VDT/workstation-user interface Keyboard and hand Body and chair Eyes and screen Repetitiveness of task(s) Length of time spent at task(s) Extent of physical constraints aT hese factors .+ I I I be d I sc u ssed r more det aI I I n he sect I on on A orh ~ t df [j n desl q I anrj t h e q u al I t ~ of $, orkl I fe SOURCE Of f!ce of Technology Assessment
PAGE 149
characteristic of the automated office and any musculoskeletal system outcome. 69 Acute musculoskeletal system outcomes are listed in table 5-7. Back, neck, and shoulder problems are reported more frequently (50 to 80 percent of VDT operators) than arm, wrist, and hand problems (20 to 40 percent of VDT operators). This ranking of self-reported symptoms by site corresponds to a ranking based on a national sample of U.S. adults. 70 Furthermore, in the national sample it was found that Although no experiments have been conducted showing a causal relationship between new office applications and musculoskeletal outcomes, there are biologically plausible explanations; for example, the experience of pain and discomfort by the worker who sits at the terminal for extended periods of time in constrained postures. Static loading of muscles results from the efforts of the human body to maintain a fixed posture for extended periods of time. This keeps the muscle in a constant state of contraction instead of a dynamic state of contraction and relaxation. Thus, waste products build up in the muscle and blood is not circulating in the muscle to provide the needed nutrients. This results in cramping and pains. Tichauer has shown that the peritendinitus that occurs in the lower arm of typists is caused by excessive static workload. (E.R. Tichauer, Biomechanics Sustains Occupational Safety and Health, Industrial Engineering, vol. 27, 1976), pp. 46-56. I,inda S. Cunningham and Jennifer L, Kelsey, Epidemiology of Musculoskeletal Impairments and Associated Disability, American Journal of Public l{ealth, vol. 74, No. 6, 1984, pp. 574-579. Table 5.7.Musculoskeletal System Outcomes: Acute or Potentially Chronic Known acute outcomes: Pain in shoulder Cramps in arms and legs Pain in back: Upper Lower Cramps in shoulder and neck muscles (stiff neck) Soreness, tingling or numbness in wrist and fingers Potential chronic outcomes:a Cervicobrachial syndrome b Prolapsed lumbar intervertebral disc Cumulative trauma disorders aThe Cbrofl!c outcomes have not been observed In my Population Of Office workers They are potential chronic outcomes based on self. reported acute musculoskeletal outcomes and biological and blomechanical plausibility bcewlcobrachlal syndrome IS a term used to define a cluster of signs In the shoul der and neck region It IS a functional and organic disorder occupationally produced on the bas!s of muscular and mental fatigue resultlng from stat!c and/or repetitive exerilon of the arm and hand muscles Chronic pain can result from the compression of nerves In the neck radlatlng down the arm ccumulatlve trauma dtsorders are a class of musculoskeletal outcomes, that are potentially Important Those In the literature associated with office workers are tenosynovltls and carpal tunnel syndrome Carpal tunnel syndrome is a disorder of the wrist In which the median nerve IS compressed against the transverse carpal Ilgament wlthln the carpal tunnel SOURCE Off Ice of Technology Assessment the level of self-reported symptoms corresponded to physician-observed abnormalities. There is, however, no evidence from studies of office workers using VDTs display that self-reported outcomes predict clinically recognizable changes in the musculoskeletal system. 7l There is a need for research to clearly identify the relationship between reports of musculoskeletal outcomes by the worker and medically verifiable outcomes. Based on the current scientific evidence, it is impossible to determine whether the acute outcomes listed in table 5-7 lead to any pathological changes in the musculoskeletal system or to any disability for workers in automated offices. The chronic outcomes reflect potential long-term consequences based on evidence from other occupations and biological and biomechanical evidence. Potential long-term consequences of new applications of office technologies for the musculoskeletal system must be considered. As Kelsey notes: [Impairments of the back and spine and arthritis and rheumatism account for the greatest amount of time lost from work. In 1976, it was reported that a reduction of one day per year in the average annual absenteeism rate among the labor force of the United States would increase the gross national product (GNP) by $10,000 million. Accordingly, the GNP would have increased by $19,000 million had musculoskeletal disorders been prevented. 7z If absenteeism rates can be affected by design of new office technologies, this has far reaching implications for health care costs. Therefore, research should be conducted to determine the relative contribution of office automation to the incidence of the chronic outcomes listed in table 5-7. Studies that have attempted to correlate self-reports with medical exams have not demonstrated a clear relationship. For an example see the work of W. Hunting, T. Laubli, and E. Grandjean, Postural and Visual Loads at VDT Workplace, I, Constrained Postures, Ergonomics, vol. 24, No. 12, 1981, pp. 917-931. Kelsey, 1981, op. cit., p. 7.
PAGE 150
.. Ch. 5Office Automation and the Quality of Worklife l 143 Photo credit Communlcat(ofls Wor/ters of America Workers adjust to their working conditions. Sometimes this can lead to musculoskeletal strain, The first two, cervicobrachial syndrome and prolapsed lumbar intervertebral discs, can potentially develop among either professionals or clerical workers. 73 They can be related to increased time spent sedentary and in constrained postures, which could be blamed on the job or the design of the office or workstation, but most likely is related to all three. Cumulative trauma disorders are more likely to develop among workers doing highly repetitive keying operations at the VDT for extended periods of time, mostly data-entry clerks who key as much as 12,000 keystrokes per hour. A prolapsed intervertebral disc occurs when the nucleus pulpous of the disc protrudes outside the annulus fibrosis. The disc tends to protrude onto a nerve root causing pain. In the lumbar region this pain radiates down the leg. Kelsey estimates the incidence of prolapsed lumbar intervertebral disc at around 0.1 to 0.5 percent per year (Kelsey, op. cit., 1981, p. 151) in the population of ages between 24 and 64 years. These are the wear and tear disorders. Many of the early warning signs (tingling, numbness, and pain) of these chronic conditions can be alleviated with changes in the working conditions. 74 There is a relationship between visual system outcomes and musculoskeletal system outcomes. For example, a person who is nearsighted most likely will get a pair of reading glasses that typically allow the person to read clearly at about 33 cm. However for work at the VDT, a person is on average only as close as 50 cm from the screen. The worker then leans forward, and the posture results in musMuch of the research on the prevention of cumulative trauma disorders is based on redesign of the machines in assembly line operations. See %eventing Illness and Injury in the Morkplace (op. cit. ) for several examples of the prevention projects.
PAGE 151
144 l Automation of Americas Offices culoskeletal strain. As the pain gets worse is between 10 and 20 percent. 78 No site has through the day, the worker adjusts by movenough data to show conclusively any assoing her body back in the chair, increasing the ciation between office automation and any advisual distance, which can lead to visual strain. verse reproductive outcome. 79 NIOSH recommends that a worker take periA Finnish study looked at the relationship odic rest breaks every 2 hours when doing work between occupation and spontaneous aborat the VDT to prevent acute asthenopia and tions, and found that industrial and construcmusculoskeletal fatigue. 75 This can be enforced tion workers had one of the highest ratios through organizational rules, through a job (13.01), and clerical and managerial workers redesign strategy giving the worker a variety one of the lowest estimated ratios for sponof tasks to break up the VDT work, or by altaneous abortions (9.91); 80 the lowest ratio was lowing the worker the discretion to take breaks in administrative work (8.16). when she feels the need. The last is especially advantageous in settings where the work is unpredictable or there is a high worker-to-VDT ratio with competition for terminals, but may demand more flexibility than an organization would normally provide. Sometimes people will not take breaks unless they are required. Reproductive System Outcomes Many occupational hazards affect the reproductive system. 76 The concerns about the reproductive hazards of working in an automated office focus primarily on the VDT. Most of the early studies considered only female operators, but both genders should be considered in discussing reproductive system outcomes. The concern over working at the VDT developed because of reported clusters of spontaneous abortions and birth defects among VDT operators. Table 5-8 shows a list of the work sites where these clusters have been reported. No adequate explanation has been offered for these observed clusters, 77 which include a wide variety of pregnancy outcomes with the majority being spontaneous abortions. The estimated prevalence rate for spontaneous abortions in the general population r-M-ifiar~_op. cit.; this recommendation has been challenged in the scientific community. bFor a complete listing and a discussion of the potential control technologies see Reproductive Hazards in the Workplace, Office of Technology Assessment, Washington, DC, 1985. -The argument is made that these clusters are chance occurrences. The rate of spontaneous abortions is not at odds with the normal rate that occurred during the time period of the reported clusters. For a lengthy statistical discussion on this see Bergqvist, op. cit. Two studies in Sweden and Japan examined the relationship between VDT work and reproductive system outcomes. Little information is available on the Japanese study .8 The Swedish study infers VDT exposure from occupations; the measure of work at a VDT is indirect. 82 Nevertheless, this retrospective 7KIn an analysis of the ability of women to recall whether they had a spontaneous abortion or not; one in four cases of spontaneous abortion was not reported. Gestational age of the time of abortion was the major determinant of recall. Allen J. Wilcox and Louise F. Horney, Accuracy of Spontaneous Abortion Recall, American Jourmd of Epidemiology, vol. 120, No. 5, 1984, pp. 727-733. There are several methodological prclblems with examining reproductive hazards in the workplace. First is the problem of selection bias. It is known that a predictor of an adverse pregnancy outcome is prior adverse outcomes. Women who have a full-term pregnancy tend to leave the work force, and therefore women with prior adverse pregnancy outcomes will be over represented (Gosta Axelsson Selection Bias in Studies of Spontaneous Abortion Among Occupational Groups, Journal of Occupational Medicine, vol. 26, No. 7, 1984, pp. 525-528). The ratio is calculated as the number of spontaneous abortions times 100 divided by the number of births. K. Hemminki, M. L. Niemi, I. Saloniemi, H. Vainioh, and E. Hemminki, Spontaneous Abortion by Occupation and Social Class in Finland, International Journal of Epidemiology, vol. 9, No. 2, pp. 149153, 1980. Conducted by the General Council of Trade Unions, it has only been reported in Japanese Miscarriages Blamed on Computer Terminals, New Scientist, vol. 106, No. 1457, 1985, p. 7. Among 4,500 women, 250 became pregnant or gave birth after work at the VDT and 91 had problems with pregnancy. There was a reported association between hours per day, work at the VDT, and problems with pregnancy or labor: for more than 6 hours VDT work per day percent had problems; for 3-4 hours VDT work per day percent had problems; for less than 1 hour VDT work per day percent had problems. It is impossible to draw any conclusions from this data since no information on the research design is known. JBengt Kallen, An Epidemiologic Study of Work With Datatermnals and Pregnancy (En Epidemologsk Studie over Arbete med Dataskarm och Graviditet), translated by Anita Dvorak, research report from the University of Lund, Sweden, 1985. Data reported herein is taken from a second report, Ad (continued on p. 146)
PAGE 152
Ch. 5Office Automation and the Quality of Worklife 145 Table 5-8.Reported Cases of Reproductive System Outcomes in Computer-Mediated Workplaces by Work Site and Job Company city (date) Sears-Roebuck/ Dallas, TX (5/79-6/80) Toronto Star/ Toronto, Ontario (5/79-5/80) Air Canada/ Montreal, Quebec (2/79-4/81) Defense Department/ Marietta, GA (10/79-10/80) Defense Department/ Marietta, GA (10/80-10/81) Pacific Northwest Bell Telephone/ Reston, WA (7/80-12/81) Solicitor General Offices/Ottowa, Ontario (4/79-4/82) Surrey Memorial HospltallVancouver, British Columbia (1978-82) Non~VDT i miscarriages 1 premature birth 8 full-term 3 full-term Outcomes VDT 2 miscarages 4 full -term b Job Financial records processing clerk a Classified ads processing clerks 4 birth defects 7 miscarriages 6 full-term NIA C Part-t i me ticket agents NIA 3 birth defects 7 miscarriages 5 full-term 14 full-term NIA NIA NIA 2 birth defects 1 still born N 1A N 1A 4 miscarriages 1 premature 2 respiratory diseases 1 full-term NIA NIA 3 miscarriages 1 birth defect 1 premature 1 bronchitislful l-term 1 full-term 10 miscarriages 9 full-term Accounting department clerks N 1A Toronto, Old City Hall Ministry of Attorney General (1980-81) United Airlines/ San Francisco, CA (1984) Southern Bell Telephone/Atlanta. GA (3/81-9/83) General Telephone/ Alma, Ml (1983-84) Clerks Secretaries 23 miscarriages 61 full -term d NIA Airline reservation clerk Telephone operators 26 miscarriages 159 full-term NIA 7 miscarriages 15 full-term Telephone operators 12 miscarriages 2 still births 3 premature 15 full-term Library/ Aarhus, Denmark Department of Public Employees/Runcorn, England (1974-82) NIA 8 miscarriages 2 full-term NIA NIA 8 miscarriages 4 still births Full-time VDT operators 12 malformations a Compares on Iy fu I I.t[ me VDT operators with non-VDT operators who may have worked as much as 6 hours per month at the terminal bFul{ term refers t. the healthy dellvery of a newborn after three trimesters These are commonly termed Ilve births CN 1A means no lnformatton reported dTh, s excludes 23 Induced abortions and 9 full-term del!venes of wives of employed men SOURCE Off Ice of Technology Assessment, compiled from several repofts and news sources
PAGE 153
146 l Automation of America Offices study found that women who worked at the VDT were 50 percent more likely to have miscarriages (crude risk ratio was 1.5 with 95 percent confidence limits of 1.1-1.9) than other women of similar age and occupational group. With adjustment for smoking and stress, confidence in the risk ratio being greater than one decreased. 83 For all types of adverse reproductive outcomes, there was a significantly greater risk for those who spent over 40 hours per week at the VDT. This suggests that further investigations are needed that directly measure VDT work and follow a group of men and nonpregnant women to see whether there is an excess of adverse reproductive system outcomes. No epidemiological study has provided reliable data comparing VDT to non-VDT operators. Two explanations other than chance have been offered for the clustersstress, and radiation emissions. There is no epidemiological evidence linking stress in office work (either physical or mental) to spontaneous abortions or still births. But while no evidence exists, no research has thoroughly explored the issue. The greatest interest has been in the impact of radiation 84 and whether radiation from the VDT is the single source exposure. There is no evidence that ionizing radiation above established safe thresholds is being emitted from currently sold terminals. 85 Ionizing radiation from the VDT is not a likely explanation, according to NIOSH, the Federal Drug Administration, and the Government of Canada. 86 The possibility of nonionizing radiation must be considered; but future developments in flat panel display technologies may eventually replace cathode ray tubes as the dominant human-computer interface. Recent concern has focused on very low frequency radiation, particularly the magnetic pulse generated from the coil of the cathode ray tube. This concern developed because studies in Spain related a magnetic pulse to morphological abnormalities in chicken embryos. 87 The evidence has not been confirmed in any other investigation trying to reproduce the patterns of electromagnetic radiation emitted from a VDT. 88 Currently, there are international efforts to study the effects of pulsed electromagnetic exposure on chick embryo development, but measurement of this and other forms of nonionizing radiation in the field is difficult. Further research should be concerned with the development of reliable instruments for the monitoring of radiation being emitted from VDTS. Based on current evidence, no long-term risk is associated with very low frequency radiation emitted from the visual display terminal, but continuing evaluation is warranted in light of Hans Malker, Critical Comments on the Report An Epidemiologica/Study of Work With Video Screens and Pnsg-nancy Outcomes, Memorandum, Swedish National Institute of Occupational Safety and Health, Feb. 2, 1985. This risk ratio dropped only 10 percent when the effects of smoking and stress were entered into a logistic regression analysis (adjusted risk ratio 1.4 with 95-percent confidence limits of 1.0-1.8). No data was reported on the effects of chemical exposure or heavy lifting on the relationship of work at the VDT and miscarriages. If the relationships persist after adjusting for these potential confounders along with other predictors of miscwriages (e.g., prior spontaneous abortions ~d parity) then the researchers can make a stronger inference. This will always be limited by the absence of a direct measure of VDT use. Three major types of radiation are associated with biological effects. They differ by the energies involved. Ionizing radiation has the most energy and is known to cause cancer and birth defects. Microwaves and radiofrequency nonionizing radiation have been associated with fertility problems and blood abnormalities. Extremely low frequency and electromagnetic nonionizing radiation is the least energetic of the three categories and has not been associated with any adverse biological effects. r William E. Murray, Video Display Terminals: Radiation Issues, IEEE Computer Graphics and Applications, April 1984. Millar, op. cit.; FDA, Drug Bulletin, vol. 14, No. 1, April 1984; Environmental Health Directorate, Investigation of Radiation Emswons From Video Display Termnals, Department of National Health and Welfare, Ottawa, Canada, 1983. This evidence comes from two reports: J.M.R. Delgado, J. Leal, J.L. Monteagudo, and M.G. Garcia, Embryological Changes Induced by Weak, Extremely Low Frequency Electromagnetic Fields, Journal of Anatomy, vol. 134, 1982, pp. 533-551; and A. Ubeda, J. Leal, M.A. Jimenez, and J.M. R. Delgado, Pulse Shape of Magnetic Fields Influences Chick Embryogenesis, Journal of Anatomy, vol. 137, 1983, pp. 513-536. *Arthur W. Guy, Health Hazards Assessment of Radio Frequency Electromagnetic Fields Emtted By Video Display Terminals, report prepared for IBM Office of the Director of Health and Safety, 1984; Kjell Hansson Mild, Personal Communications, Swedish National Board of occupational Safety and Health, 1985; S. Maffeo, M.W. Miller, and E. L. Carstensen, Lack of Effect of Weak Low Frequency Electr~Magnetic Fields on Chick Embryogenesis, Journal of Anatomy, vol. 139, No. 4, 1984, pp. 613-168.
PAGE 154
Ch. 5Office Automation and the Quality of Worklife 147 Table 5-9.U.S. Radiation type X-ray ... Ultraviolet (near) ., Visible Radiofrequency Occupational Exposure Standards Occupational exposure standard Source 2,5 mR/hr OSHA 1,000 microW/cm a ACG I H 2.920 fL ACGIH h electromagnetic fields, frequency range (MHz) 0013 100 mWlcm b ACGIA b 10-100,00 0 10 mWlcm b OSHA aCo(j~ of Federal Requ Idt Ions TI tle 29 Chapter XVII Part 191096 Ion!zt ng Rad dt I o n Occ u pat tonal Safety and Health Ad m I n $1 rat (on Washl ngton DC 1980 hTh rpsho(cj Llrn!t for Cherntcal Substances and phy SICEil Agents lr~ tne @JOrk El ,v!, onment Wlt h I nt ended Changes for 1983 and 1984 American Conference or Go~ernment Indusl1-10. 10/Fa >10-400 1 >400-2,000 F a /400 >2,000-300,000 . 5 aF refers to the frequency tn Mfiz thus 10 calculate the speclflc I Imtt wlthln the range the frequency must be factored In accordinq to the expression SOURCE International Non Ion!zlnq Radlatlon Commlltee of the International Radlat[on Protecllon Association Interim Guldelhnes on Llmlts of Ex pos,ure to Radlofrequency Elect romagnetlc Fields (n the Frequenc Y From 100 KHZ to 300 GHz Hea/th Phys{cs VOI 46 No 4 1984 pp 975984 the large number of people exposed, to see if modified or new standards need to be developed. 89 Standards shown in table 5-9 reflect the available evidence on the potential effects of chronic exposure at low levels. The American Conference of Government Industrial Hygienists (ACGIH) exposure limit for electromagnetic fields in the frequency range of 0.01-3 A recent report from the U.S. Environmental Protection Agency concluded there were no consistent biological effects at the molecular and subcellular level in laboratory experiments. The report suggested continued revision of general conclusions because of limited knowledge about chronic low-level exposures and the existence of frequency-specific effects and power-density windows. (U.S. Environmental Protection Agency, Biolo@cal Effects of Radiofrequency Radiation, Joe A. Elder and Daniel F. Cahil (eds. ), EPA-60018 -83-026F (Research Triangle Park, NC: Health F~ffects Research Laboratory, 1984). ) MHz is 100 mW/cm 2 This threshold level is 10 times higher than the VDT emissions in the 0.01-100 MHz range measured in a study by the Bureau of Radiological Health of the Environmental Protection Agency .90 The International Radiation Protection Association (IRPA) standards are 10 times lower for a similar frequency range than standards set by ACGIH (0.01-3 for ACGIH and 0.01-1 for IRPA). Because there are differences in the limits of exposure, the standards should be constantly reexamined in the light of further research. NIOSH is currently conducting a retrospective study to ascertain whether working at a VDT is associated with spontaneous abortions. Similar studies are being conducted by Dr. Irving Selikoff at Mt. Sinai Hospital (in conjunction with 9 to 5, the Working Womens Association), by Dr. Kelley Brix at the University of Michigan and in Sweden, Finland, Canada, and Denmark. Again, there is no reliable national registry for monitoring the prevalence of reproductive system outcomes as they relate to occupations. The establishment of such a system could help prevent the public fear that has arisen about VDT work. Metallic shielding has been discussed as one way to avoid any possibility of radiation being absorbed by the operator, especially shielding the flyback transformer. This shielding is relatively inexpensive and is usually provided in newer models to satisfy FCC requirements for reducing electromagnetic interference. ] The government of Sweden has recommended that any pregnant employee have the right to be moved, at full pay, to a job with no exposure until the baby is born. The International Labor Organization recommends that a worker who is considering becoming pregnant be allowed to transfer to another job. 92 However, the risk of reproductive system damage, if it exists, may be just as great for a man as for a woman. Most recently, the Service Employees International Union (SEIU) has negotiated Cited in fieproducti~e Hazards in the 11orkplace, op. cit. Guy, op. cit. -International I.abor Organization, Guidelines for \DTs, (~ene~a, Switzerland, 1985.
PAGE 155
148 l Automation of America Offices in collective bargaining with Equitable Life Insurance the rights of employees to leave VDT jobs where the believed risk for an adverse pregnancy outcome is high. These measures would at a minimum alleviate fears. Stress-Related Outcomes Work-related stress is not new to offices, but recent epidemiological evidence has called attention to the potential long-term consequences of stressful office working conditions. One study of the health records of 22,000 workers in 130 occupations found that clerical workers had the second highest rate of stress-related diseases. Analysis of the Framingham Heart Study data showed that women clerical workers developed coronary heart disease (CHD) at about twice the rate of other women workers and women at home. 94 The workplace factors that predicted the development of CHD among clerical workers were a nonsupportive boss and low-job mobility. Any conclusions about the contribution of the new office technologies to the incidence of disease must thus consider that women clerical workers were already at a greater risk for the development of certain stress-related diseases. The new technology may or may not intensify those characteristics of the work environment that are associated with the elevated risk, increasing ones chances of developing disease; they could also improve working conditions, decreasing the likelihood of disease. There are no reliable estimates of the costs to the organization or to society of stress-related illnesses, absenteeism, tardiness, employee turnover, decreased quality and quantity of output, unscheduled machine downtime (due to employee tampering) and workers compensation awards. These outcomes can lead to increased medical care costs and decreased organizational effectiveness. 4{ Mich~ei J. Smith, Michael J. Colligan, and J.S. Hurrell, A Review of NIOSH Psychological Stress Research, (lccupatiomd Stress: Proceedings of a Conference on Occup,~tional Stress (Washington, DC: U.S. Government Printing Office, 1977). ( Suzanne G. Haynes and Manning Feinleib, Women, Work and Coronary E{eart Disease: Prospective Findings From the Framingham Heart Study, q American Journal of Pub]IcS }Iealth, Iol. 70, 1980, pp. 133-141. Stress-related outcomes can be divided into those that are acute and transient 95 and those that are persistent and enduring. Anxiety and depression have been shown to be related to myocardial infarction, angina pectoris, and coronary death. 96 Therefore, correction of working conditions that produce these stress responses can help to prevent stress-related diseases. 97 As important as are acute conditions, the potential for chronic diseases is the key public health issue. The population potentially at risk in offices is large. If conditions that can lead to chronic illness and disease can be identified, management can use this information to guide implementations. The biological plausibility of changed working conditions leading to a pathological change in the body has been the subject of debate for decades. Chronic arousal is one currently acceptable biological pathway (see box B for an example of a plausible link to disease). Chronic arousal is a biological adaptation to the environment. When one continues to respond daily to the same stimuli, arousal can become part of the normal biological adaptation to, for example, work a method of activating the bodys resources to meet the demands. This can lead to a form of healthy maladaptation, in which the worker completes tasks, but at some biological costs. The eventual cost can be illness due to lowered immunological resistance, or chronic illness that lowers the ability to perform, such as cardiovascular disease or peptic ulcers. The danger is hidden in that the chronic heightened state of arousal does Stress can contribute to the development of acute visual and musculoskeletal system outcomes. For example, muscle fatigue can be the direct result of a stressful working condition, which can stimulate brain stem activity, which in turn can cause muscle fatigue. C. David Jenkins, Psychosocial and Behavioral Factors, Prevention of Coronar~ Heart Disease: Practical Management of the Risk Factors. Norman M. Kaplan and Jeremiah Stamler (eds. ) (Philadelphia: W. B. Saunders, 1983), pp. 98-112. This type of prevention strategy may be found to be the most cost-effective, since it has the potential to reduce the longterm development of disease. Also, if the working conditions that produce stress responses in workers are modified, this may alter certain behavioral outcomes including smoking, drinking, and overeatingall considered socially acceptable ways to cope with the stresses of life, but primary determinants of premature morbidity and mortality.
PAGE 156
Box B.Chronic Arousal: The Link Between Disease, The Case of Ch. 5Office Automation and the Qualify of Worklife 149 Working Conditions and the Natural History of Cardiovascular Disease* Early theories of stress and disease held that humans instinctively respond to challenging conditions from their environment through activation of pituitary and adrenal gland functions. Walter Cannon in the 1930s showed the importance of the sympathetic adrenal system in the flight-or-fight syndrome. He showed there were limits to how well the system could adapt to stimuli; afterwards bodily harm could result. OnIy in the last 20 years have scientists begun to unravel the complex interrelationship between the brain, immune system, and endocrine system. There are competing models of this complex system, but several conclusions can be stated. Sympathetic adrenomedullary (SAM) stimulation is accompanied by evaluated blood pressure and heart rate, heightened myocardial oxygen requirements, increased levels of circulating epinephrine (E) and norepinephrine (NE), elevated plasma concentrations of free fatty acids (FFA), and increased plasma renin activity. These can predispose the individual to cardiovascular disease (CVD). The release of the neuroregulator, catecholamine, as a consequence of SAM activity is central to the most prominent theories of stress and CVD. In its efforts to mobilize reserve energy resources, catecholamin es hydrolyze triglyceride stores into FFA and glycerol. FFAs are either utilized in the production of energy or taken up by the liver and adipose tissue then resynthesized into triglycerides that are secreted as a component of very low-density lipoproteins (VLDL). VLDLS have been shown to be significant components of the atherosclerotic process. When catecholamin es are chronically secreted in great excess of the bodys energy requirements to meet the demands of the work environment, the abundance of FFAs can ultimately result in damage to vascular walls caused by circulating VLDLS. Damage to vascular endothelium can also occur as a result of increased arterial blood pressure or turbula.nce caused by SAM hemodynamic responses. Damaged vascular endothelium is more susceptible to atherosclerotic deposits than healthy vessel walls. These processes suggest several possible pathways between the challenges created by working conditions and the development of cardiovascular disease. ahig Sxtion is bas~ on a S1.lrnrnary prepared by Dr. Andrea LaCroix. For a full description see Andrea LaCroix, Occupation Ex~sure to High Demand/Low Control Work and Coronary Heart Disease Incidence in the Frarmngham Cohort, doctoral dissertation, University of North Carolina, Chapel Hill, 1984. A more complete description of the various biological procesaes can be found in The Institute of Medicine, Research on Stress and Human Health (Washington, DC: National Academy Fress, 1981). not necessarily correlate with the workers attitudes about work. High motivation and satisfaction does not obviate the possibility that a person could be at risk. Three working conditions have been associated with chronic arousal: l social isolation or lack of social support; lack of control over the timing, speed, and variety of tasks; and heavy workload (especially repetitive and machine-paced tasks). Each of these conditions has also been postulated as likely to arise from office automation. No evidence has shown that all three must occur to elicit deleterious biological responses; however, in some jobs they all occur. Epidemiological evidence has shown the most consistent links between heart conditions and workload and control. In a 6-year prospective analysis, men with jobs characterized by a heavy workload and limited job control were found to have 1.4 times the normal risk of CVD morbidity. 98 In a case-control study of myocardial infarction and occupational exposures, it was found that hectic work and low control over work tempo and skill variety were associated with myocardial infarction in men under 55. 99 In a study of Swedish workers who had changed jobs, those whose new job had greater control had fewer coronary symptoms The CHD morbidity measure was a self-report indicator of chest pain, dyspnea, hypertension, and heart weakness. R. Karasek, et al., Job Decision I.atitude, ,Job Demands, and Cardio}ascular Disease: A Prospective Stud} of Swedish N!en, American Journal of Public l{ealth, \ol. 71, No. 7, 1981, pp. 694-70,5. II. A] fredsson, et a],, M~.ocardial Infarction and I)s~chosocial Jtork l~nyironment: An Analysis of the Male Swedish Mrorking Force, Social Science and Aledicine. \ol. 16, 1 W2, pp. 463-467.
PAGE 157
150 l Automation of Americas Offices than workers whose new job had less control. l00 In the only American prospective study, women who described their job as having a heavy workload with limited job control had a threefold greater risk of developing coronary heart disease (CHD) as women reporting a heavy workload but having control over their work. Men did not exhibit the same relationship. At greatest risk were clerical women, who had a 420 percent greater chance of developing CHD (relative risk 5.2 with 95 percent confidence interval of 1.80-15.08). These associations persisted after controlling for the traditional risk factors for CHD. 101 Based on the available evidence, lower level staff appear to be at greater risk for the development of stress-related diseases. This however need not be true with office automation because the changes in the working conditions can affect all levels of staff. Alternatively, office automation can be implemented so as to add control and encourage more social interaction. Currently, the evidence for a relationship between stress-related diseases and VDT work is still sparse. A principle problem in drawing conclusions is that the long latency of chronic diseases prevents early recognition. However, in the natural history of disease certain early indicators could be expected such as psychosomatic symptoms (sleeping problems, dizziness, nausea, and stomachache), high blood pressure or angina pectoris. 102 Only preliminary evidence exists for the likelihood of deleterious long. term health effects on office workers, or specifically on VDT workers. Most studies show that support staff who work at a VDT report more psychosomatic symptoms than either other support staff or professionals who work at VDTS. These studies, in general, also show that women report more psychosomatic symptoms, as do people in jobs characterized by R. Kar~sek and B. Garden, Managing Job Stress, Work. ing Paper, Columbia University, Department of Industrial Engi. neering and Operations Research, 1984, ( LaCroix, op. cit. These can also be explained by other factors such as sit uations outside of work or biological predisposition. Research efforts should in the future attempt to differentiate between the risks associated with work and those associated with lei. sure. Also, physical and psychosocial working conditions can produce similar adverse stress responses and should be differen tiated in future research. little control over pace, a heavy workload, and lack of social support. This is consistent with the general literature on stress-related disease. No study can fully answer the questionwhat contribution to the risk of disease can be attributed to office automation? But two recent studies suggest that there is a potential for these psychosomatic symptoms to develop into chronic conditions. In a cross-sectional study of clerical workers in the communications industry, those who worked at the VDT were at about a two-fold greater risk for the development of angina pectoris. 103 This is the first study to demonstrate a relationship between automated office work and a valid precursor of CVD. A second study of Finnish workers found that workers in automated offices were at a 106 percent higher risk (relative risk 2.06, 95 percent confidence interval 1.43-2.69) for one or more chronic illnesses. 104 These projects point to the need to examine the potential effects of office automation to determine under what conditions a worker is likely to be at a greater risk for the development of a chronic disease. The analysis of angina was restricted to a sample of 650 women. There were no significant differences between respondents and a 10-percent sample of nonrespondents. For a full description of the results of the study along with the limitations of the study see, Suzanne Haynes and Andrea LaCroix (University of North Carolina at Chapel Hill), A Cross-Sectional Study of the Health of VDT Operators in the Telephone Industry, unpublished manuscript, 1985. The measure of angina used was the Rose Angina Questionnaire. Several studies have shown that this self-report measure is as valid and reliable a predictor of angina pectoris as a physical exam by a physician. (Lawrence M. Friedman, et al., Assessment of Angina Pectoris After Myocardial Infarction: Comparison of Rose Questionnaire With Physician Judgment in the BetaBlocker Heart Attack Trial, American Journal of Epidemiology, vol. 121, NO. 4, 1985, pp. 555-563. ) The ratio reported is after adjustment for age, sex, baseline health status, health risk factors and baseline systolic blood pressure. The risk did not change when ergonomic strain, job strain and social support from the supervisor were taken into account. There was a 21-percent increase in the risk when the control over the pace of work was taken into consideration. Again, the unanswered question is whether the job was characterized by low control prior to automation or whether working conditions were exacerbated by the automation of the work process? The chronic illness measure was composed primarily of heart disease (70 percent) with cases of cancer, chronic musculoskeletal disorders, and gastrointestinal diseases. These results are preliminary and further analysis is currently being carried out. For a description of the study see Mary Haan, Health Effects of Automated Office Work, paper presented at the Office of Technology Assessment Symposium on The Impacts of Office Automation and Computer-Mediated Work on the Quality of Worklife, December 10-12, 1984.
PAGE 158
.. Since the evidence relating working conditions in automated offices to stress-related diseases is still limited, it may be thought premature to begin job and organizational redesign strategies. The major dilemma facing managers is whether it is more cost effective to wait and make job and organizational changes later, or to use the new technologies to usher Ch. 5Office Automation and the Quality of Worklife l 151 in job and organizational changes. The major dilemma facing public health officials is that by the time the population of VDT users has worked long enough to manifest chronic outcomes there will be few people to use as controls, since most office workers will be working at the VDT. SECTION 111: INTERVENTIONS Office Design and Workstation Design More and more organizations consider the design of the office, building, and workstation as integral parts of the new technology. These characteristics of office environments can contribute to quality of worklife both directly, and indirectly by ameliorating stressful working conditions, Office Design The introduction of new office technologies into a building can affect the way the building performs the distribution of office space, heating, ventilation and air conditioning, power and wiring, and acoustics. 105 Running of cables through offices can be the cause of falls. Office design can balance the impersonality of some computerized office tasks. However, new buildings will only accommodate about 5 percent of the computer terminals installed between now and 1990 (see figure 5-2). The single greatest area of impact will be the retrofitting of older buildings to accommodate new technologies. Retrofitting can be as simple as putting in a new lighting system or as com 1t has been estimated that only 2 percent of the costs of conducting business in an office building over 40 years are devoted to building design. Six percent goes to operation and maintenance and 9.2 percent to labor. Therefore, improvements in the quality of worklife through building design offers a potential cost-effecti~e ie~er for organizations. This estimate is based on 1971 dollars, but the differences in scale still exist today, see Francis T. Ventre, Documentation and Assessment of the GSA PBS Building Svystems Program: Final Report and Recommendations, NBSIR 83-2777, General Services Administration, Washington, DC, 1983. This general ratio of costs has been reproduced in another 7 year study of office design and its impact on the quality of work (Michael Brill, presentation at Conexion Atlanta, GA, Nov. 6-9, 1985). Figure 5-2.The Proportion of VDTS Installed in New Office Space Compared to Existing Space, 1984-90 40M 30M 20M 10M 1984 1986 1990 Year = New building s Most video display units will be installed in existing office space. Even if every person in new office construction in the United States had a VDU, only about 5 percent of the VDUS would be in new buildings. Therefore, accommodating them is mainly a retrofit job. (Assuming approximately 200 gross square feet for each workstation with a VDU, then if the height of each bar represents the total space occupied by workstations with VDUS, the shaded part of each bar represents the cumulative total of VDUS in new office space.) SOURCE Michael Bell, Harbinger Group Inc 1985 plex as complete rebuilding of the internal structure of the building. When new office technologies are haphazardly introduced without considering the extra demands placed on the building, new health and performance problems can develop or old ones can be amplified. Office Space Office automation allows the redesign of workspace to accommodate private and shared
PAGE 159
152 l Automation of America Offices work. Allocation of space is always a critical factor in offices; it signifies status and organizational commitment to the worker. It provides some control over interruptions and privacy and encourages or discourages social interaction, factors associated with job satis faction and performance. A personal computer and attachments take about 12.5 square feet of office space, which increases individual workspace needs. 106 With the cost of office real estate rising, new technology is sometimes being crammed into inappropriate spaces that can lead to shifts in status, and awkward work postures associated with visual and musculoskeletal strain. Noise levels can be distracting, especially for workers in an open plan office, 107 and devices such as acoustical printer covers may be needed to reduce noise levels. Changes in office layout may be necessary to reduce the isolating effects of computer work. Air quality is a major problem in modern offices. Heating and cooling problems can be exacerbated by office equipment that generates heat, and by changes in lighting that are made to accommodate VDT users. 108 Hot spots caused by concentrations of office equipment can change the way the building distributes hIn one study, Office Research Into Buildings and Information Technology (OR13 IT) it was found that new information technology only increased the secretaries space requirements by 50 percent (Duffy, Eley, Giffone, and WorthingtonArchitects, ORBIT Report, 1983). It has generally been argued that open space offices are most beneficial to workers doing repetitive and boring tasks; they provide needed social stimulation. Alternatively, people doing creative work require private space. However, this has recently been challenged by researchers who have found that even workers with boring and repetitive jobs prefer private spaces to open spaces and giving it to them increased their job performance (E. Sundstrom, et al., Privacy at Work: Architectural Correlates of Job Satisfaction and Job Performance, Academy & Management Journal, vol. 23, 1980, pp. 101-1 17). wit is generally estimated that a workstation and operator can generate 18.0 Btu/square foot of heat throughout a building. For discussion of these problems see American Society of Heating, Refrigerating, and Air-Conditioning Engineers, ASHRAE Standard 62-73. Standards for Natural and Mechanical Ventilation (New York, 1973); ASHRAE Standard 55-1981. Thermal Environmental Con&tions for Human Occupancy (New York, 1981); ASHRA Standard 90-75. Energy Conservation in New Builchng Des@ (New York, 1977); and ANS1/ASHRAE Standard 62-1981. Ventilation for Acceptable Indoor Air Qualit.v (New York, 1981); also Ventre, op. cit. air, heat, and cooling, and make indoor air pollution and thermal discomfort worse. It is recommended that in refitting a building for extensive office automation, the building be remodeled as a set of microzones composed of comparable offices with localized control of the heating, cooling, and ventilation system. Workstation Design The workstation includes the table or desk holding the terminal, the chair, and the equipment that an operator users. In many offices these have not changed, and the microcomputer merely replaces, or even sits next to, the typewriter. Yet several characteristics of the VD T 109 differ from conventional office technologies and impose different physical demands on the worker: l l l l The terminal is self-luminous. There is a transient display on the screen, whereas the display in printed text is constant. VDTs usually have highly specular curved glass surfaces that reflect light. Information is presented in a vertical plane as compared to the horizontal plane of paper or the angled plane of the typewriter. In conventional offices the illuminance has been kept high for reading, with other light sources, such as windows, contributing to the illumination. For VDT work, much lower illumination is best. 110 But the worker often looks back and forth from the VDT to printed text and other equipment, so a compromise is required; task lighting (lamps) can be used for additional illumination where needed. Tradi The discussion of VDTS usually is about the monochromatic raster scan CRT, which is like a television screen with some additional electronics. Other kinds of VDTS that may be much more common in the future are the plasma panel, liquid crystal, and electroluminescent displays. See App. A for descriptions of the technology. (The American National Standards Institute recommends a general lighting level of 750-1600 I UX for traditional desktop reading; other groups recommend no more than 200 I UX for rooms where VDTS are used. American National Standard Practice for Office Lighting, ANSl A132.1 (1973); AT&T Bell I.aboratories, Video Display Termnals; Preliminary Guidelines for selection, Installation, and Use (1983); NIOSH, Potential Health Hazards of Video Display Termnals (1981).
PAGE 160
Ch. 5Office Automation and the Quality of Worklife 153 tional office and office furniture design assumes a depressed line of sight, but the vertical position of the VDT brings windows, ceiling lights, etc., into the line of sight, causing glare. The reflective surface of the screen can also contribute to eye strain. On a VDT the display is transient and there is a characteristic flicker (assuming that the VDT is a cathode ray tube) which is often imperceptible to the worker, but can probably lead to visual fatigue and asthenopia although it has not been linked to visual performance. 111 The readability of the screen is also related to the structure of the dot matrix used to create characters, the contrast between the characters and the background, and the viewing distance of the person from the screen. Increases in the size of the matrix have been shown to increase reading performance, 112 but this is not within the control of the user or, generally, the purchaser. The contrast between characters and screen can be controlled by the user, and should be placed for ease of use. The fonts used in VDTS are chosen by engineers for convenience in design, and are less easy to read than print fonts evolved through years of practical experience and user preference. 113 They assume a set viewing distance between viewer and screen, but some people lean back in their chairs and put the keyboard on their lap or in other unlikely places. This is especially likely when the worker has such a small table or desk that cannot hold documents as well as the keyboard and screen. Many office furniture designers have evidently not real If the refresh rate is at least 65 flashes per second, there is no perceptible flicker; VDT refresh rates are about 30-70 Hz. Television set flicker has occasionally been associated with photosensitive epileptogenic seizures, but the refresh rates known to induce such seizures are low, from 8-14 Hz. It is thus highly unlikely that such seizures would be induced by VDT work. A 14-percent increase in reaction time and a 10-percent decrease in errors was shown in a study by Haubner, et al., Visual Dispiay UnitsCharacteristics of Performance, Commission on International Lighting, 20th sess., 1983, as reported in Bergqvist, op. cit. National Academy of Sciences, op. cit. For further discussion see l+;. Grandjean, W. Hunting, and M. Pidermann, 4VDT Workstation Design: Preferred Settings and Their Effects, Human Factors, ~ol. 25, No. 2, 1983, pp. 161-175, Photo credl( O~(fca/ Coat/rig Labora(or/es /fJc One way to control glare and enhance readability is to use a filter ized that office workers frequently use both paper and a terminal in the same task. The visual and musculoskeletal strain that can result from the wrong viewing distance is best managed by providing fully adjustable equipment so that each user can adapt the workstation design to fit his/her own needs. A tiltable screen and adjustable desktop can solve many problems. The keyboard is likely to remain the dominant input technology for most office workers; 115 keyboard work has been associated with carpal tunnel syndrome and cervicobrachial syndrome. The physical characteristics of the The usual QL$ERTY keyboard dates back to the 1870s, and was recognized as an international standard in 1966 although it has been challenged by the Dvorak ke~board, with a different arrangement of the keys. Some studies indicate the Dvorak would improve typing speed by 25 percent and shorten the time needed to learn to type. It was designed (after World War II) to avoid excessive wrist movement b~ locating the most commonly struck keys in the center row, possibl~ reducing musculoskeletal strain, There is currentl~. a resurgence of interest in the Dvorak, but both employers and employees are seemingly unwilling to undertake the retraining that would be necessary to make the change. See ,Jan Noyes, The QWERTY Keyboard: A Review, International Journal of Alan-illachine Studies, vol. 18, 1983, pp. 265-281.
PAGE 161
154 l Autornation of Americas Offices keyboard associated with precursors of these syndromes, such as its height, can be altered. Some keyboards are attached to the terminal and resemble the typing keyboard in height from the table (generally over 40 mm); others are detached and built lower, which has many advantages since a determinant of musculoskeletal symptoms in forearms and wrists is the ability to intermittently rest the arms or wrists on the table. 116 A detachable keyboard can also be arranged for comfort and viewing distance. Chairs are critical to the comfort of any office worker; and with office automation people tend to spend longer blocks of time in one place. The chair should allow one to change the seated posture without adopting an awkward position; 117 and it should be easily adjustable. To prevent lower back problems, office chairs should have an adjustable lumbar back support and the back rest height should extend beyond the lumbar region to the thoracic region and be adjustable for height, angle, and pressure. The seat should also be adjustable to allow people of different sizes to get about a 90-degree angle between upper body and thighs. Table height is also important. If used alternately by several workers it should be easily adjustable to accommodate different heights. Desks should not restrain knee and thigh movement, and should have enough space for working documents, screens, and keyboards to be moved around. One of the most important elements of office design strategy is to make sure that office workers understand why adjusting lighting, furniture, and temperature controls is Operators may not prefer suggested optimal keyboard heights and their preference may not affect performance. Teresa Burke, Effects of Keyboard Height on Typist Performance and Preference, Proceedings of the Human Factors Society, 28th Annual Meeting, 1984, pp. 272-276. See also Grandjean, op. cit. Most keyboard operators maintain an erect posture for only 2 to 3 minutes at a time. Standards and recommendations are usually based on this position, even though studies shc~w that people prefer leaning slightly forward or slightly backward. When tables andior chairs are not adjustable, the user sometimes must assume awkward postures. This woman may eventually develop neck cramps from continually looking down Photo credlf M/cbae/ Sm/th Sometimes the worker finds her own way to sit comfortably important, and how to do it. 118 A Finnish study found that if data-entry workers were trained to recognize ergonomic problems and prevention strategies, the level of clinically recognized This has in fact been mandated in the OSHA Act (Sec. 21c), which says that the Department of Health and Human Services must provide for the establishment and supervision of programs for the education and training of employees and employers in the recognition, avoidance, and prevention of unsafe or unhealthful working conditions,
PAGE 162
Ch. 5Office Automation and the Quality of Worklife l 155 -.neck, shoulder, and elbow problems was sigIn developing standards or guidelines, current nificantly reduced over a 6 month period. 119 tax codes should be examined to ascertain that they will not act as a barrier to intended Public Policy and Quality of Worklife change. Public policy interventions must be considered in light of both the scientific evidence and the political and economic climate. Many organizations look to new technologies to provide them with the competitive edge, and some argue that health and safety regulations inhibit the full use of new technologies and further innovations. A report issued by the Presidents Commission on Industrial Competitiveness recommended that no new health and safety regulations be established; society should not seek to eliminate risk, and risk below minimum levels should be left unregulated, or to industry self-regulation. 120 Executive Order 12291 (1981) provided that except where expressly required by law, the costs of a regulation should not exceed its benefits. This must be considered in discussion of the potential effects of office automation. Cooperative actions by employers and employees, when they are sought in good faith and are successful, are in theory a more cost-effective strategy than regulation because they can be self-enforcing. Investment tax credits have encouraged the adoption of capital-intensive technology to improve the efficiency and productivity of the work process, yet often working conditions that emerge from this technology adoption produce adverse stress responses. Tax policies can also directly affect the way physical characteristics of the office are designed. For example, some experts suggest that technicalities in the way depreciation is allowed have encouraged open plan offices so that partitions separating office spaces can be depreciated. R. Kukkonen, et al., Prevention of Fatigue Amongst Data k;ntry operators, unpublished manuscript. Using a quasiexperimental design, the researchers were able to demonstrate significant decreases in neck and shoulder symptoms, back symptoms, and clinically defined neck shoulder tenderness and hardening. Report of the Presidents Commission on Industrial Competitiveness, as cited in The Occupational Health and Safet~ ,Vews]etter, ~ol. 15, No. 5, p. 3, 1985. hlichael Beer and James W. Driscoll, Strategies for Change, Improting Life at I! ork, op. cit. 1977, pp. 364-453. Workers Compensation Under common law, employers have a duty to exercise reasonable care in protecting their employees from harm in dangerous situations. Remedies under the common law of torts have been proposed for many health and safety problems. But in the United States, the laws creating workers compensation generally make this system the exclusive remedy for employee disability, 123 although in many instances this exclusionary principle has been circumvented by third-party liability suits in which the employee sues the manufacturer of equipment or material (e.g., the thousands of suits against Johns Manville as the producer of asbestos materials). Workers compensation does not cover spontaneous abortion, miscarriage or birth defects, since this class of injury does not impair the -These typically fall into three categoriesnegligence, product liability, and strict liability. In negligence actions, the employee must demonstrate that the employer was negligent ir. the maintenance or inspection of equipment or did not adequately warn the employee about risks inherent in its use. These actions must be shown to cause proximate harm to the employee and not to be a part of normal dangers inherent in the job. Since it is very difficult to demonstrate a causal link between VDT use and any adverse health outcome, this area of legal action has limited utility. The same analysis can be applied to product liability and strict liability?. William Presser, Handbook of the Law of Torts, 4th ed. (St. Paul: Wrest Publishing, 1971). Workers compensation is a broad disability>. insurance program in which awards based on loss of earning power are paid to workers whenever they are injured on the job, Under workers compensation the employee must establish that the condition arose from employment. (T?picall~ the terms proximate cause, producing cause and contributing cause are used to refer to the relationship between the working conditions and the injury or disease. ) lfhere a claim is upheld, a schedule of payments is set up for the worker based on some formula that determines the extent that the worker is disabled or suffers a decrease in wage-earning capacity. The clear intent of the statute is to compensate all work-related injuries. (This section draws on the work of the National Council on Compensation Insurance report, Emotional Stress in the11orkplace,NTeu. I.egal Rights in the Eighties, New York, 1984. ) When an injury is found to be covered by a workers compensation act, it is usually held that the statutory compensation is the exclusive remedy and any recovery at common law (tort law) is barred. (Presser, op. cit.)
PAGE 163
156 l Automation of Americas Offices workers earning power. 124 It has allowed for compensation for stress-related diseases and disabilities along with visual and musculoskeletal system injuries. Table 5-11 contains three examples of workers compensation claims for injury in automated workplaces. The worker must show a causal link between work in an automated office and the injury. Currently, for the VDT operator there is very little direct causal evidence for visual and musculoskeletal injuries and for stress-related injuries. Where an organic cause can be found to explain the injury the claim is easier to interpret. Thus, demonstrating that some characteristic of the automated office environment led to a biological change in the worker is the clearest vehicle for establishing a compensation case. These conditions while impairing the workers ability to work, may not be medically detectable, so the primary instrument for identifying illness and injury to the office worker is through the reporting of symptoms. Thus, it is difficult to demonstrate proximate cau sality. If the condition predated employment or was aggravated by conditions outside the job the claim may be disallowed or the compensation reduced. For the VDT operator who goes home and watches television (also a cathode ray tube~ or has personal problems that produce adverse stress responses, it is difficult to argue that the work was the source of the injury, although John Parry, Jeanne Dooley, David Rapoport, and Johr Taylor, Are VDTS Hazardous to Your I.egal Health? Men tal and Physical Disability Law Reporter, vol. 8, No. 4, 1984 pp. 342-360. For a more complete discussion of the workers compensation system and its evolution see Reproductive Haz ards in the ltorkplace, op cit.; or Peter Barth and H. Allan Hunt, Vorkers Compensation and Work-Related Illnesses ana Diseases (Cambridge: MIT Press, 1980). workers who spend over 7 hours a day at a terminal may make such an argument effectively. In addition, many chronic debilitating visual and musculoskeletal injuries do not occur suddenly and unexpectedly, so that claims are likely to be awarded only in States that accept gradual developments of an injury over time. The workers compensation system was meant to also serve as a preventive measure; forcing employers to pay for injuries and diseases arising from employment should encourage them to develop preventive strategies. Emerging office problems may develop into compensatable disabilities, but the current work force would not benefit from the incentives for prevention created by the rise in compensation claims. Yet todays employee cannot sue the employer in most States through the common law of torts since visual, musculoskeletal, and stress-related conditions are compensable. Mental Disability. There has recently been an increase in workers compensation claims related to mental stress. The California Workers Compensation Institute found that stress claims doubled from 1980 to 1982, while claims for other disabling work injuries decreased during the same period. 125 Workers compensation for mental disability arising from employment is greatest among younger workers, as shown in figure 5-3. Mental stress is significant by itself, and may also predict future development of stress-related diseases. If it does, any increase in claims because of office automation could be costly to society. Table 5-12 provides several examples of successful workers comCalifornia Workers Compensation Institute, Bulletin, Apr. 20, 1983. Table 5-11. Workers Compensation Claims for Employees Working at Video Display Terminals Job Probler n Word processing Stress of 6 to 8 hours VDT work per day Insurance claims processor Repetitive work al VDT lead to wrist pain Word processing Severe headaches Sensitivity to light Visual pain aset!led prlo~ to wkwwflt bClalm awarded SOURCE Legal Rlgbfs for VDT Users (Cleveland OH Working Womels Educational Fund, 1985) Injury Nervous brea~down a Carpal tunnel syndrome b Accommodative spasm a -.
PAGE 164
Ch. 5Office Automation and the Quality of Worklife l 157 Figure 5-3.A Comparison of Workers Compensation Claims for Mental Stress and Other Occupational Diseases 30 25 20 15 10 5 0 1-29 30-39 40-49 Age 50-59 60-99 Mental stress Other occupational diseases I I I I SOURCE National Council on Cornpensatlon Insurance Ernot~onal Sfre.ss (n the LVorkp/ace New Lega/ Rights ~n (he .E~ghltes (New York 1985) Table 5-12. Examples of Mental Stress Mental Disability Claims for White-Collar Workers Alleged Allege d Occupation of clalmant work-related mental stress mental disability Advertising manager Overworked, supervisor requestedea-rly retirement Anxiety, depression Clalms director Job pressures Psychiatric illness Data analyst/control clerk Inability to perform job duties Psychological disability Insurance underwriter Increase in job duties Mental breakdown Secretary Increase in job duties Depressive neurosis SOURCE Ernot/ona/ SIress ~n the WorkFace New Lega/ R/ghts /n fhe E/ght/es (New York Nat!onal Council on Compensahon lns~rance 1984) pensation claims for mental disability, showing the variability in both the alleged cause and the disability. The workers compensation system can be considered as an early warning system for the effects of office automation on the mental health of the worker. Claims for mental stress and mental disability decrease with increasing age and are greater among women than men. 126 This could be the result of a cohort effect; younger workers are probably more likely to demand rights in the These results are based on a random sampling of claims from 1980 to 1982 in 13 States representing geographic variability. The study is described in National Council on Compensation Insurance, op. cit. workplace than are older workers. It also corresponds to the trend of increasing automation of the workplace that affects younger workers. Since a majority of the claims have been filed by women (54.2 percent), it may be logical to infer a relationship with the recent technological changes; the workers first affected by office automation in the 1960s and 1970s were predominantly women. 127 -It has been suggested that the highly publicized mental disability claims associated with stress may lead to a greater reporting and filing of similar claims. It may be that the legal recognition of mental disabilit~ claims has led to an increase in the number of claims. This is important in mental stress claims because of the universality of stress across all occupations.
PAGE 165
158 l Automation Of America's Offices Indemnity costs for gradual mental stress claims were less than 60 percent as high as the costs for other occupational diseases per average claim in 1980, but surpassed these claims in 1982. Also, the average medical costs incurred for gradual mental stress claims passed other occupational diseases in 1981. 128 Mental disability awards could become an economic burden for industry. If workers attribute their adverse stress responses to the VDT and workers compensation claims become a common avenue for compensation, then the system may become overburdened, with the employer becoming a universal insurer. The workers compensation system is at a crossroads in determining whether a worker who is unable to work because of mental disability is to be treated differently from a worker unable to work due to a physical injury arising out of employment. There have been relatively few compensation claims specifically related to work at a VDT (about 30), most for physical injury bul: some for mental injury. In one case a VDT operator claimed that the VDT contributed to an already high-pressure atmosphere, leading to three stress reactions during a 7-month period. Her claim was upheld by the New Jersey workers compensation board. There seem to be three compensable categories of mental stress. The stress response as a usual condition of the work environment is perhaps the most liberal and far-reaching of the categories. The recognition that gradual accumulation of mental stress is a cause of the mental disability is important since many stress responses in offices result from the incremental or chronic conditions of work rather than from any single acute event. Second, the stress responses can result from a continuing set of unusual conditions in the work environment, for example, an increase in job duties because of new office technologies, Third, the stress reaction can be due to a suds den and unusual event in the work environment, for example, the witnessing of a cowork J Nat~onal Council on Compensation Insurance, op. cit., p 6. ers heart attack. The only condition under which this could be associated with office automation is if the technological change itself were construed as a traumatic event. Perhaps an acute fear of being exposed to a terminal that purportedly emits hazardous radiation could lead a person to an episode of mental disability. 129 Fifteen States have no relevant statutory limitations. Seven State courts have concluded that mental disability is to be treated no differently than physical disability if working conditions are the cause of the disability; nine have ruled mental disability is not compensatable. Most of the States that have taken the position that mental disability claims are not compensable have strictly interpreted the statutes defining injury to be of a physical nature. Eight States have adopted the idea of an acute event in employment as proximate cause. They allow mental disability claims when there has been some traumatic event that can be identified as the proximate cause of the disability. Eleven States have upheld the right of the worker to claim a mental disability if the stress exceeds the stress of everyday work. In these States, there have been cautions about limiting the scope of the compensation system to avoid the development of a social health insurance program. The Federal Government acted first (1908, 1916) to establish workers compensation for Federal railroad and then for other employees. 130 Since then, the workers compensation system has evolved at the State level relatively without Federal interventions, although there have been proposals from time to time to create a Federal system. One role the Federal Government might play is to serve as an information clearinghouse for issues of office automaZ9This is analogous to the idea of a pollutant on the job, where the effects of the pollutant are unknown. The Montana Supreme Court has allowed such a compensation to be awarded to a worker who developed psychological problems on exposure to pollutants at work (Alcklahon v. The Anaconda Company, No. 81-34 (Montana Supreme Court, Mar. 29, 1984), 8 MPDLR 291). M.B. Kent, A History of Occupational Safety and Health in the United States, 1983, as cited in Preventing Illness and Injury in the Workplace, op. cit.
PAGE 166
Ch. 5Office Automation and the Quality of Worklife 159 tion and stress responses, to provide States with the most recent scientific evidence for making decisions. Clearly, automation is the dominant change in office work and many mental disability claims in the next two decades are likely to revolve around the effects of the new office technologies. Standards or Guidelines Since ergonomic factors in office equipment are important to the health of the American worker and the productivity of American business, there have been proposals that ergonomic standards or guidelines be developed, either by government or other institutions. A recent discussion of regulatory policy pointed out that health and safety regulations should focus on threshold limit values on exposure as oppposed to means for reducing exposure such as design standards. 132 Standards can only be developed with the appropriate measurement techniques. 133 Working conditions that produce adverse stress responses can be measured in many different ways and scientists disagree on which ways are best. Therefore, the development of standards for working conditions that produce psychological and biological stress responses in individuals may be an option to consider only when there are valid and objective measures of the working conditions. 134 Any standards development process would have to consider issues raised in the recent benzene standard case decided by the U.S. Supreme Court and the noise standard case de .- In 1981, the Science and Technology Committee of the House of Representatives held hearings, The Human Factor in Productivity. These hearings are the most recent attempt to gather information about the role of human factors in industrial competitiveness (see, The Human Factor in Innovation and Productivity, U.S. House of Representatives, Committee on Science and Technology, Subcommittee on Science, Technology, and Research, 97th Cong., 1st sess., 1981). L Douglas Ii. Ginsburg, Administrative Efforts to Enhance the Opportunities for SelfRegulation, I,abor Law Journal, vol. 3,5, No. 12, 1984, pp. 731-735. 1Becker, op. cit. Edward E. I.awler, Should the Quality of Work Life be I.egislated? 7he Personnel Administrator, vol. 21, No. 4, 1976, pp. 17-21. See also, Edwin A. I.ocke, The Case Against I.egislating the Quality of Y$ork I.ife, The Personnel ,4drninistrator, vol. 21, No. 4, 1976, pp. 19-21. cided by the U.S. Circuit Court of Appeals in Richmond, Virginia, which exemplify the impact judicial decisions may have on occupational health and safety policy. In the noise case the judges invalidated a standard on the grounds that it might on the basis of medical examiniations require employees to take actions concerning hearing loss caused by nonoccupational factors. 135 While progressive hearing loss can be identified through medical exams, the exams can not distinguish between loss resulting from occupational and nonoccupational sources of noise. The ruling has been reversed by the same court. If upheld, it would have probably severely restricted the development of Federal standards for automated offices where most exposures are chronic and confounded by nonoccupational factors. The Supreme Court in 1980 upheld an earlier lower court decision striking down the 1978 OSHA benzene standard. They found (5 to 4) that OSHA had not established a threshold limit for benzene posing a significant risk to the worker. The Secretary of Labor must demonstrate that a workplace threatens the worker with a significant risk of harm. 136 Althoug h providing limited guidance as to what significant risk means or how it should be calculated, the Supreme Court decision has demanded a more rigorous scientific treatment of occupational safety and health standards. 37 The setting of standards designed to improve the health and well-being of the office workers would have to take into account this decision. However, standards can be developed and adopted voluntarily .There are three general types of standards; engineering, informational, and administrative. Engineering standards (which include design standards) establish alternative means of interfacing with the technology or alternative means of building the technology to ensure the health and safety of Reported in the Occupational Safet,\ and Health I.etter, vol. 15, No. 1, 1985, p. 2. ) Industrial Union Department, AFL-C1O v. American Petroleum Institute, 448 U.S. 607 (Supreme Court, July 2, 1980). For a full discussion of the implications of this decision for OSHA standards setting see, Pre~enting Illness and Injur.~ in the Workplace. op. cit.
PAGE 167
160 Automation of Americas Offices the worker. Informational standards include product labeling, instructions for proper use, and other means of transferring information from the employer or manufacturer to the worker. Administrative standards establish organizational policies such as rest breaks. A key policy issue is how to ensure that the standards development process is based on sound scientific evidence. 138 Design Standards.The possibility of office and workstation-design standards has received much attention, spurred on by bills introduced into at least 18 State legislatures, and by the development of standards in other countries. The Governor of New Mexico has issued an executive order (No. 85-1 1) mandating ergonomic guidelines for State employees. Such proposals and actions are based on the assumption that there are definable and measurable characteristics of both the office and the workstation that, if modified, will improve health and worker performance. A major issue is whether standards should be voluntary or required-self-regulation or governmental regulation. It is argued that self-regulation is more flexible, and standards can be tailored to meet the specific office automation application. However, a second question is whether many organizations would voluntarily pay the front-end, or retrofitting cost that might be entailed. 139 Another issue is that of fairness; standards development committees composed of special interest groups are often suspect. Especially in promulgating mandated standards, balanced groups of consumers, producers and other affected parties may be preferable for the development and enforcement of standards. 140 A major impetus for currently proposed State bills is a concern over the inadequacy of voluntary standards for these reasons. One option at the national level would be to establish a national commission to oversee standards development or to empower the National Bureau of Standards with such duties. 141 A third problem is that of the appropriate level of specificity for design standards. The level of specificity in State bills varies widely, indicative of the uncertainty as to the most effective way to modify the work environment to accommodate new office technologies. 142 International standards reflect differing approaches, different populations, and different goals. The very specific German standards are meant to standardize the building and production of office equipment, while Swedish standards are more like guidelines to be used in negotiation between the unions and management. Japanese guidelines are not binding and are very general. Other major issues in standards development are how to define enforcement mechanisms for government standards and how to ensure that the development of standards spurs rather than impedes the development of more protective technology in the future. Standards can be established at either the Federal, State, or organizational level by the use of procurement schedules. 143 California, New Mexico, Massachusetts, and Wisconsin have all developed procurement guidelines for the purchase of workstation equipment in the public sector. An examination of the currently proposed office standards for State bills demonstrates problems in the standards development process; general lighting levels vary from 500 IUX or less to greater than 700 IUX. International standards also do not agree. There is no general national or international consensus on the best way to establish standards for office automation. The lack of consensus may reflect the generally agreed need to maximize flexibility in current design to accommodate a variety of future needs. This need for flexibility is clearly reflected in the mechanisms for controlling glare in proposed State bills, which range from installing indirect lighting; to antiglare.filters; to proper placement of the terminals with respect to windows; to task lighting; to screen hoods. q Francis Ventre, Transforming Environmental Research Into Regulatory Policy, Responding to Social Change, Basil Honikman (cd.) (New York: Halstead Press, 1984). Ginsburg, op. cit. Currently, the National Bureau of Standards coordinates within the Federal Government and assists the private sector in the procedures and policy for the development and application of standards; however, they have no enforcement powers. Currently, the Human Factors Society in conjunction with the American National Standards Institute is developing a set of standards for the physical and perceptual ergonomics of visual display terminal workstations. These standards contend that ergonomics is highly applications dependent and looks only at word processing, data-entry, and data-inquiry tasks. Ostberg, et al., Ergonomic Procurement Guidelines for Visual Display Units as a Tool for Progressive Change, The Procdngs of the Eleventh International Symposium on Human Factors in elecornmumcations, Sept. 9-13, 1985, Cesson-Sevigne, France.
PAGE 168
Ch. 5Office Automation and the Quality of Worklife 161 -. Information Standards. These might provide the employee and employer with information about, for example, the recommended lighting levels for using a particular display terminal. Rhode Island has just passed a law requiring the State Department of Labor to develop a brochure on VDT use and to hold training sessions throughout the State. Health and Safety Standards.They are administrative standards. At the Federal level, an initial policy consideration would be to decide whether they should be developed within a single agency or by several agencies. The former would eliminate replication, while the latter would ensure that all areas are covered. Currently, OSHA, the Consumer Products Safety Commission (CPSC), and the Department of Health and Human Services (DHHS) can develop standards that might be adapted to address new office technologies. For example, the OSHA Health Hazard Communication standard is intended to disseminate information about chemical health hazards and only applies to the manufacturing sector. For OSHA to disseminate information about health and safety issues in office automation under the Health Hazard Communication Standard, that authority might have to be broadened. 144 Alternatively, the recently proposed High Risk Occupational Disease Notification and Prevention Act of 1985 (H.R. 1309), would establish within NIOSH the authority to disseminate information to persons at an elevated risk for the development of disease. Only workers at an elevated risk will receive counseling and medical monitoring. The disadvantage to this approach is that it requires prior knowledge about the development of the disease and the risk to the worker. A broader approach is to provide information to all employers and employees so that the factors associated with health and safety can be discussed before the equipment is installed. This would require input from many sectors of the Federal Government. A recent court decision suggested that OSHA expand the scope of the standard to include nonmanufacturing industries, ~l,s, Court of ,,~ppe~s, ~jrd Circuit, Case No. 83-3554. Eye examinations, rest breaks, semiannual equipment inspections, and transfers for pregnant VDT operators could be handled through administrative standards. Several countries have standards or guidelines regarding visual exams. 145 Legislation calling for rest breaks has been proposed in several States, and several nations have guidelines or standards with respect to rest breaks. 146 Although Swedens practice of 15 minute rest breaks after every hour or 2 hours (depending on the work) is often cited, this is not a national standard but results from agreements between labor organizations and management. This will change with the issuance of the Ordinance Concerning Work With Computer Displays. 147 This ordinance requires that employees who work more one-half hour consecutively on the VDT on a daily basis, must receive eye exams. In Britain, the Health and Safety Executive recognizes that the most satisfactory length of pause can only be determined by consideration of the individual operators job ... 148 and recommends that natural breaks be built into jobs, with a mix of VDT and non-VDT work; goals are provided to be used in labormanagement bargaining. Guidelines.The examples above point to the problems of standards development, which is both a scientific and apolitical process. Communications between the scientific and political communities are sometimes lagging and mechanisms for encouraging more dialog are needed. Japanese pro~isional guidelines for TIII work calI for a visual exam as part of the preemplol-ment medical exam, with regular eye tests thereafter. J1est German regulations sa} that VDT workers should have eye exams mer~. 5 years until age 45 and every 3 years thereafter. In France, two regulations stipulate eye exams and special medical surxwillance for JDI users. But Great Britain does not recognize the need for ~isual standards for VDT usage which are an~. different from other clerical work. TWO Swedish standards call for periodic intermissions without specifying length or frequency?: ,J apan pro~isiona] guidelines call for at least a 10-15 minute hreak after each hour of VDT work. Ordinance Concerning 11ork M?th Computtr Displa.~s, translated b~ Dr. Olo\ Ostberg. This ordinance has not gone into effect yet. Great Britain, Iiealth and Safety Executive, Iisual Displa~ [)nit.s (I,ondon: Her Majest?,s Stationar?T Office, 1983), p. 1 :?.
PAGE 169
162 l Automation of America Offices An alternative to mandated standards is the development of guidelines articulating goals. Both Sweden and Japan have established goals for workstation design. In Japan, these are wholly voluntary; in Sweden the guidelines are presented as the governments position and are used in collective bargaining. Guidelines recognize the variability yin the way office technologies can be implemented. They can be established within the broader context of an institutionalized change process. In some countries this is environmental legislation; 149 for example, the Work Environment Act of Norway (1977) says, An effort shall be made to avoid monotonous repetitive jobs and jobs which are determined by a machine to such an extent that the worker is prevented from altering his rate of working. Such acts give those involved in the change process a national policy against which to evaluate their efforts. Such goals and guidelines are only as effective as the commitment on the part of employers and employees to cooperatively achieve them, and on the part of government to monitor and actively encourage their progress. Research Needs Congressional hearings have addressed the issue of possible hazards of VDT work, 150 and much has been written on the subject, but workers still wonder if new technology is dangerous. A possible public policy response is to encourage research on various illness and disease outcomes associated with the computer mediation of work, both to answer the questions in the minds of the public and to demonstrate .. ... L Federal Republic of Germany: The Works Constitution Act of 1972; The Netherlands: Working Environment Act of 1980; Norway: Work Environment Act of 1977 (also known as the Act Respecting Workers Protection and the Working Environment); Sweden: Working Environment Act of 1974; Denmark: Act Respecting the Working Environment; German Democratic Republic: Labour Code as Amended June 1977, Thes{~ acts deal with all working conditions, including office automation. U.S. Congress, Committee on Science and Technology, Subcommittee on Investigations and Oversight, Potential Health Effects of VDT Termnals and Ra&ofrequency Heaters and Sealers, 97th Cong., 2d. sess., May 12-13, 1981. U.S. Congress, Committee on Education and Labor, Subcommittee on Health and Safety, OSHA Oversight: Video Display Termnals in the Workplace, 98th Cong., 2d sess., 1984. the feasibility y of using new office technologies as both a public health and a productivity tool. The OSHA Act (section 20) mandates continued research and demonstration projects exploring emerging problems including psychological factors (such as stress) created by new technologies. Office automation is one technological change with such a broad exposure that Congress may wish to insist that the Department of Health and Human Services put high priority on investigating the health and safety issues, demonstrating practices to reduce any potential effects, and disseminating such information. There is a need for at least three directions in the research. The first is to follow office workers over time to develop risk estimates of the relative contribution of office automation to the natural history of a disease, focusing on those populations at greatest risk today (clerical and technical workers), and also considering the changes in the work of managers and professionals. This field research should be complemented by lab research to test relationships between working conditions and biological processes. The second is to define the appropriate measures of the physical conditions of work. Field evidence that compares objective measures of the physical environment to subjective measures of the quality of worklife is limited. The tools for measurement must be sharpened and refined to assess the many subtle causes and outcomes characteristic of office work. If it could be demonstrated that subjective measures are valid indicators of the physical environment and discriminate the various health, performance and job attitudes, less costly evaluations of the work environment would be possible. Third, intervention and field evaluation studies are needed to verify claims that changes in ergonomic or other working conditions can improve worker productivity and reduce health problems. These efforts should focus on high risk groups of office workers or identify high risk working conditions and develop modification strategies.
PAGE 170
Ch. 5Office Automation and the Quality of Worklife l 163 The Health Promotion and Disease Prevention Amendments of 1984, which establish research centers for disease prevention and health promotion across the country, administered by the Centers for Disease Control. These centers could conduct the research projects mentioned above. Finally, it is important to know the true extent of the problem. How prevalent are poor working conditions? Section 20(a)(7) of the OSHA Act stipulates that the Department of Health and Human Services conduct industrywide studies of the effects of chronic low-level exposures on the health of workers. This need for reliable data also points to a need for a centralized data collection system for monitoring the emerging problems. There is now no single source of this data. The current industry and occupational surveillance systems are disparate and difficult to link together. In testimony before Congress the Assistant Surgeon General noted: The activities do not provide a comprehensive epidemiologic surveillance of occupational diseases and injuries in the United States Unless our efforts are targeted toward comprehensive data collection and synthesis, the confusion will only grow worse. 151 The current data collection system at the Bureau of Labor Statistics (BLS) is not able to identify the extent of office automation in white-collar occupations. There is a need to develop not only reliable indicators of national trends in morbidity, but also broad national descriptions of those working conditions likely to place the worker at highest risk. The Job Training Partnership Act of 1982, Section 462(b) states: the Secretary shall maintain descriptions of job duties, training and education requirements, working conditions, and characteristics of occupations. There is a need to further refine the definitions used in the Dictionary of Occupational Titles to reflect the current changes in the workplace. An early r ] Assistant Surgeon General J. Donald Mdlar, cited in Occupational Iflness Data Collection: Fragmented, Unreliable, and Sevent-y Years Behind Communicable Disease Surveillance, U.S. Congress, Committee on Government Operations, House Report 98-1144 (Washington, DC: U.S. Government Printing Office, 1984). warning system could be developed by linking these data sources. 152 This would allow researchers and policy makers to estimate the relative severity of problems among different classes of workers. Legislation recently introduced into the House of Representatives would establish an early warning system to identify workers at an increased risk for occupational disease (H.R. 1309). This bill could be amended to include language establishing a national morbidity surveillance system for the development of chronic diseases related to technological changes. Labor-Management Relations and Office Automation As has been pointed out, one strategy for improving quality of worklife is through standards development, and a second is the setting of goals or objectives, incorporated in public or private sector guidelines. Another way to work toward goal-directed change is through collective bargaining. In Europe, there is much emphasis on involving worker representatives in office automation planning. In some countries there is a national or local work council to which industry supplies information about office automation plans. This creates the framework for monitoring how well organizations are moving toward established goals, minimizing the need for government enforcement. 153 This process of involving workers in technology planning is part of a process called codetermination, a phrase that has gained some popularity in this country. While the legal right to be kept informed about technological change has not been ex For the importance of linking databases to inform policymakers, and current impediments see, Legal and A dministra tive Impediments to the Conduct of Epidemiologic Research, Task Force on En~ironmental Cancer and Heart and I,ung Disease, 1984. 1 ~ Federal Republic of Germany: Works Constitution Act of 1972; The Netherlands: Works Council Act of 1979; Norwal: Work Environment ,4ct of 1977; United Kingdom: Emplo~ment Protection Act of 1975; France: Act No. 82-915 of Oct. 28, 1982 and Act No. 82-689 of Aug. 4, 1982; Sweden: Act Representing Co-Determination of Work, 1976. For further detail see Automation, Ilork Organization, and Occupational Stress (Geneva: International I.abour organization, ]984).
PAGE 171
164 l Automation of America Offices tended to American workers, the National Labor-Management Relations Act guarantees the right to bargain collectively with management. Thus rather than intervening directly in office automation related quality of worklifc problems, the Federal government can choose to leave this to industry self-regulation, judicial decisions, and labor-management relations, with National Labor Relations Board interventions when they are called for. Collective bargaining could function as a forum for resolving quality of worklife issues related to office automation. 154 The National Labor Relations Act (NLRA) established collective bargaining to reduce the need for government regulation by providing bargaining power to all parties. But the ability to use collective bargaining as an avenue for discussion of stress, health, and ergonomic problems surrounding office automation is limited by the low level of participation of office workers in collective bargaining units. Fair representation in this process also depends on judicial decisions and National La-n There are other mechanisms through which workers can participate in org anizational decisionrnaking about technologiczd change, but they are specific to an organization or industry. There has been, for example, great interest in the Japanese system of quality control circles. One observer notes: Predecision joint consultation to solve the problems of manpower and employment due to drastic technological changes developed around 1960, and built up to become a basic part of the later Japanese industrial relations. This practice often takes the place of collective bargaining in Japanese industry. Akihiro Ishikawa, Microelectronics and Japanese Industrial Relations, Microprocessors, Manpower, and Society, Malcom Warner (cd. ) (New York: St. Martins Press, 1984). In Germany, the development of Work Councils at the local shop-floor level came after World War II. These councils are separate from unions in Germany and make many decisions at the local shop-floor level. These alternatives to collective bargaining never developed in the United States to any signif: cant degree. 1NLRA gives employees the right to organize and join a union. At least 30 percent of the employees of an organization must petition NLRB to have an election, and to win certificz,tion as a bargaining representative the union must win the votes of a majority of the employees. Once the union is certified, the employer is obligated to enter into collective bargaining over wages, hours, and other terms and conditions of employment to develop a contractual agreement, which typically will last 1 to 3 years. NLRB does not oversee specific contract terms, but can be asked to intervene if either party does not act in good faith. Any worker has the right to file an unfair labor practice char= if she believes constraint, coercion, or discrim; nation has been used. bor Relations Board (NLRB) decisions about the scope of the law and whether a series of acts and amendments designed to deal primarily with blue-collar occupations applies equally well to white-collar occupations. A complicating factor is that in white-collar work, distinctions between management and labor tend to be blurred. In recent cases over whether a group of workers has the right to organize, the courts have been confronted with problems of determining whether professionals are workers or part of management, and who makes decisions for the employ er. 156 As office automation leads to task bundling, the distinctions between management and workers will blur further. A recent Supreme Court decision reemphasized earlier questioning of the intent of Congress in several sections of NLRA : 157 was the intent to prohibit certain types of collective bargaining over technological change? Current laws may require clarification in this regard. While some labor contracts contain clauses concerning automation, negotiation about the decision to automate is not construed as mandatory by either NLRB or the courts. However, it is also not clear that management has an absolute right to automate. The employer does have the right to determine equipment needs and the size of the work force, and can change the nature and scope of the business, but only if it does not affect contract terms and conditions of employment. Technology bargaining has become a focus of the collective bargaining process and the extent of employee rights and employer prerogatives is decided on a case-by-case basis by NLRB and the courts. Marina Angel, Professionals and Unionization, Minnesota Law Review, vol. 166, 1982, pp. 383-457. In NLRA v. International Longshoremen Association, AFL-CIO, et al. (Case No. 84-861, June 27, 1985), the court said The only question thus to be decided is whether Congress meant, in enacting Sec. 8(b)(4)(B) and 8(e) to prevent this kind of labor-management arrangement designed to forestall possible adverse effects upon workers arising from changing technology.
PAGE 172
Ch. 5Office Automation and the Quality of Worklife l 165 The strategy of using collective bargaining for addressing quality of worklife issues related to office automation is limited both by the relatively small number of office workers covered by union agreements and by the nature of the process; both topics are addressed below. White Collar Unions and Worker Representation According to BLS statistics, only 19 percent of all U.S. workers belong to labor unions and among office workers the level of unionization is much lower. About 14 percent of clerical workers belong to unions and about 17 percent are in bargaining units covered by union contracts. 158 Professional specialty workers have a higher level of unionization, 23 percent, but this reflects the large number of teachers belonging to unions or professional associations. Only 6 percent of executive, administrative, and managerial workers belong to unions. }59 Unionization of office workers tends to follow industry lines; in industries where manufacturing workers are highly organized (see table 5-12) they are more likely to belong to unions than in other industries, such as the insurance industry, where office automation has been heavily adopted. The figures show levels of union membership for broad industry categories; the first two rows include most office workers. 160 It has traditionally been difficult for unions to organize office workers, primarily because of their close relationship with management and identification with the middle class. In every occupational category, the number of workers covered by a union or Bargaining units cotrer all workers whether or not the~ belong to the union. Paul 0, Flaim, l~ew Data on Union Members and Their I+; arnings, to be published in ?tfonthi.~ Labor l?e~.iew. Office workers are represented by a large number of unions including the Communications Workers of America, Office and Professional Employees (OPE I U], the Newspaper Guild, the United Food and Commercial Workers, the Service I+~mployees Internat~onal Union (SEIU), American Federation of (jovermnent I+; mployees (A FGI?), American Federation of State, County, and hlunicipal Employees (A FSCME~), and others. District 925 is a new nationwide office workers union formed by SE IU and Nine to Five: the National Association of Working Women, in 1981. employee association contract decreased slightly from 1983 to 1984 (see figure 5-4). The influx of women into clerical jobs may have retarded unionization, because they were a ready supply of low-paid workers, and have historically been slower to join unions. It has been suggested however that the changes in the office workplace brought on by office automation will increase the level of unionization among clerical workers, in those places where office work is becoming routinized and more like factory work. Alternatively, office automation can improve jobs; if unions can present themselves as mechanisms for ensuring these improvements, they may have greater appeal to office workers. A policy question to be considered then is whether the Federal Government should take steps to encourage technology-related bargaining in labor-management negotiations, and also encourage other mechanisms to improve the opportunity for office workers to be represented in discussions about office automation. Technology Bargaining Traditional subjects for collective bargaining have included wages, fringe benefits, and hours. Increasingly unions are also dealing with questions of health and safety, electronic monitoring, and job security. Technological change is also a major issue, but technology is a difficult subject for collective bargaining. One recent survey found technological change provisions present in fewer than 20 percent of current agreements. 162 Labor organizations can seek to deal with technological change issues by trying to influence: 1) the introduction of new technology per se; 2) the changing nature of the jobs; 3) changes in skills requirements or status; or 4) work force reductions. The choice of what technology to use (for manufacturing or for office work) has traditionally belonged to management, and many Roberta Goldberg, Organizing &omen Office Workers: Dissatisfaction, Consciousness, and Action (New York: Praeger Publishers, 1983). -Cited in Kmin Murphy, Technolo~cal (hange Clauses in Coilectitre Bargaining Agreements, Department for Professional Employees, AFI.-CI(), Publication #81-2, August 1982, p. 5.
PAGE 173
766 Automation of Americas Offices Table 5-13.Union Membership by Industry and Occupation Industry Transportation, Financial ~ communication, insurance, Occupation Total a Mining Construction Manufacturing utlities Trade real estate Service Government All occupations .15 6 17.9 24.3 26.5 396 82 2,7 72 359 Management/professional specialist 52 1.7 86 40 11 1 18 17 67 381 Technicians/ sales/administrative 80 3.3 2,9 10.4 34.8 6.6 22 5.1 302 Service 87 (b) (b) 352 490 44 12,4 86 394 Operator/fabricator/ laborer 33.6 32.9 25,7 392 471 204 (b) 11 3 362 aTb~ totals j~ld~ agriculture, fOreSrw, and fishery occupations which are not listed separately. Thts number includes Private sector union membership, excluding government membership bThe data do not meet publication standards SOURCE Larry T Adams, Changing Employment Patterns of Organized Workers, Month/y Labor Rewew, February 1985 pp 25-31, uses data from Current Population Survey, Bureau of Labor Statistics Figure 5-4.Employed Wage and Salary Workers Covered by a Union or Employee Association Contract in 1983.84 Executive, administrative, 1 & manager I Professional Technicians Sales Administrative & clerical support 1983 1984 0 5 10 15 20 25 30 35 Percent of work force 1 SOURCE. Paul O Flatm, New Data on Unton Members and Their Earnings, 10 be publlshed In the Month/y Labor Rewew labor agreements include a management rights clause that gives management sole authority over processes and types of machinery and equipment to be used, types of products to be manufactured, quality of material and workmanship required, selling prices of products 163 .. h Murphy, op. cit., p. 5. NLRA does not specifically identify technology as a subject for the bargaining process. There has been no definitive decision on whether or not management has a duty to bargain about implementation of new technology. In the past, the NLRB has: shown a willingness to extend the duty to bargain over the introduction of technological change to the decision making stage. The
PAGE 174
Ch. 5Office Automation and the Quality of Worklife 167 courts have tended to restrict the duty, requiring only that the employer bargain with labor over the effects of its unilateral decisions 164 In recent years, both the NKRB and the Supreme Court have tended to favor management prerogative to effect unilateral change. Yet, a recent Supreme Court decision has partially countered this trend. In a decision regarding the longshoremen, the Supreme Court ruled that: Elimination of work in the sense that it is made unnecessary by innovation is not of itself a reason to condemn work preservation agreements ; to the contrary, such elimination provides the very premise for such agreements. 165 Modified forms of the management rights clause have given the union or the employees right to advance notice of technological change and in some cases the right to consultation in the change. While there is no single blueprint, advance notice and joint consultation seem to be the hallmarks of building a cooperative dialog between labor and management; they reduce opposition and resistance to the change. 166 The Communications Workers of America (CWA), in its agreements with AT&T and the Bell operating companies, has negotiated the right to advanced notice. 167 Consultation might in rare cases mean participation in the planning and implementation of technological change, but more often it means the right to negotiate about the adjustment mechanisms that will accompany the change, for example, layoffs or retraining. Nicholas A. Ashford, The Impact of Office Automation cm the Qualit>r of Worklife: Policy I replications, paper presented at the Office of Technolo~ Assessment S~mposium on the Impacts of Office Automation and Computerhlediated Mrork on the Quality of l~rorklife, Dec. 10-12, 1984, p. 30, ~Vationai Labor Relations Board v. National Longshoremen Assn., op. cit. Ste~en Deutsch, Technological Change and Laborhlanagement Relations, draft report to the Bureau of LaborManagement Relations and Cooperati~e Prograrm, Department of I,abor, 19/i5; used with permission. Dwight 13. Dayis, \\orkplace High Tech Spurs Retraining El fforts, Iiigh Technology, November 1984, pp. 60-62. In Europe, advance notice and the provision of information are usually the first steps in collective bargaining agreements. 168 Advance notice typically means that management must give the workers and their representatives enough time to consult, negotiate, and prepare for the changes. Implementation of the new technology can be made contingent on a collective bargaining agreement. Some agreements specify that the employee be involved in any job redesign. 169 Close cooperation of unions and management in planning automation systems is rare. The NLRA precludes any person acting in the interest of the employer from being a member of the union. Historically, this has meant that supervisors were excluded. Therefore, if decisions of the planning of technological change are construed to be employer or managerial decisions, union members may be precluded from actively participating in some stages of the decisionmaking. One policy question for Congress to consider is whether the NLRA needs revision so that the traditional boundaries and relationship between labor and management does not prevent a cooperative planning process. Cooperative planning is not easy to achieve. Union leaders may be uninformed about technological choices, long-term company, plans or industry conditions. Management decisions may be made at the national headquarters while labor negotiations may take place on a regional basis. Even in Norway, where whitecollar unionization is high and where law requires management to inform unions about h Examples of the collective bargaining agreements are provided in i4utomation, }iork Organization and Stress, op. cit., 1984. h For example: The employees shall, individuall}~ or in groups, be given proper information about condi~ions at the workplace that affect their own job The employees shall be given an opportunit~ to take part in designing their ow n job situations as well as in the work of change and development that affects their jobs. Agreement on Efficienc~ and Participation SAF-I.OPTK. Swedish I?mpio.\er Confederation (Stockholm: Andren & Helm, 1982). f Recently, in the case Yeshiva University v. The NLRB, the Supreme Court ruled that faculty members were managerial employees and therefore not entitled to organize under the NLRA, see NLRB v. Yeshiva Univ., 444 U.S. 672 (1980).
PAGE 175
168 l Automation of America Offices industry plans and conditions, actual participation of unions in technological planning is rare. The process of alternating cooperation, conflict, and negotiation requires a large commitment of time and resources from the union. 171 Several recent labor-management agreements have centered around quality of worklife issues related to VDT use. In late 1984, claims processors at Equitable Life Assurance Company became members of District 925 of SEIU after a 3-year long effort. Points covered in the contract include requirements for detachable keyboards, adjustable chairs, an additional rest break from VDT work, and (in some cases) transfers to non-VDT work for pregnant workers. 172 One strongly worded agreement seeks to prevent the routinization of work through technological change. The contract between District 925 and a legal services organization states that: ., [t]he Employer recognizes that as a general matter the routinization of the secretarial profession through the introduction of new technological changes, such as mag cards, is undesirable, and the Employer has no present intention of doing S0.173 The problem of electronic monitoring has drawn the attention of at least 30 unions, many of which represent office workers. Several, including the International Federation of Clerical and Technical Employees and the AFLCIO, have adopted the position that no VDT monitoring clauses should be included in contracts, but thus far, no contracts have prohibited monitoring. CWA has won contract language providing that electronic monitoring will be used for training and not for discipline. The SE I U-Equitable contract provides that employees will be given full information about the monitoring system, access to their U. Briefs, C. Ciborra, and L. Schneider (eds. ), Svstem~: Design For, Writh and B-v the [Jsers (Amsterdam: Nor-th Ho] land Publishers, 1983). Stephanie K. Walter, A VDT Victory at Equitable Life Dents the Anti-Union Armor, Management Technology, Jan ll~y 1985, pp. 6-8. Ibid. own records, and the right to file a grievance if they believe their record is inaccurate.1 74 Some union contracts have focused on new job classifications and higher pay for new skills in office automation. AFSCME contracts with the cities of Los Angeles anci New York require wage increases when word processing systems are introduced. The contract between Equitable and District 925 also modifies the piece rate pay system previously used for VDT workers. 175 The CWA contract with AT&T and the Bell operating companies established quality of worklife committees to deal with the retraining issue, and workers are notified in advance when a job will end so they can be retrained. An agreement between OPEIUand the New York Stock Exchange binds the employer to train displaced employees for an available job resulting from such technological change or for other jobs which the producer has available. . 177 In some cases, contracts have specified that some form of income maintenance will be available to employees who must be moved to lower paying jobs as a result of automation. 178 Few unions have been able to negotiate a contract that guarantees no layoffs as a result of technological change, but some contracts have set forth who can be laid off and what severance pay, relocation benefits, or rehiring preference will be given to laid-off workers. However, the recent Supreme Court decision in favor of primary work preservation for longshoremen does uphold the right of the employees to keep secure their work: When the objective of an agreement and its enforcement is so clearly one of work preservation as is the one involved here, the lawfulness of the agreement under Sees. 8(b)(4)(B) and 8(e) is secure, absent some other evidence of secondary purpose. 179 Alan Westin, Privac-v Issues in the ,tlonitoring of Emplo-vee Work on VDTS in the Office Environment: Practices, Interests, and Po)ic.v Choims, prepared for Office of Technology Assessment, December 1984, pp. 119-125. -r Walter, op. cit. Davis, op. cit. ] Murphy, op. cit., p. 14. Ibid. Supreme Court, op. cit., 1985.
PAGE 176
Chapter 6 Confidentiality and Security Issues With Office Automation
PAGE 177
Contents Privacy, Confidentiality, and SecurityDefinitions. . . . . . . . Office and Data Protection . . . . . . . . . . . . The Period 1960-78 . . . . . . . . . . . . . . The Political and Legislative Climate . . . . . . . . . . The Third Phase of Office Automation . . . . . . . . . . The Handling of Client Data . . . . . . . . . . . Security and Confidentiality Issues . . . . . . . . . . Privacy Issues in Work Monitoring. . . . . . . . . . . Security and Confidentiality in Automated Public Offices . . . . . . Government Guidelines . . . . . . . . . . . . The Special Concern About Employee Privacy . . . . . . . . Accidental Losses . . . . . . . . . . . . . . Comparisons With the Private Sector . . . . . . . . . Policy Considerations . . . . . . . . . . . . . 171 172 174 175 178 178 179 181 182 183 183 184 185 185
PAGE 178
Chapter 6 Confidentiality and Security Issues With Office Automation Protection of privacy or confidentiality in recordkeeping (and security measures to accomplish such protection) is a concern that has continued from the constitutional era of quill, pen, and copybook through the invention of the telegraph, telephone, typewriter, microphone, duplicating machine, large-scale manual filing systems, teletype, electric accounting machinery (EAM), and first-, second-, and This chapter as a whole draws on a report to OTA, Privacy and Security Issues in the Use of Persona) Information About Clients and Customers on Micro and Persona) Computers Used in office Automation, prepared by The Educational Fund for Individual Rights, Alan Westin and I.ance Hoffman, principal investigators, 1985. third-generation computers. How much data about individuals and groups is needed? Who will use the data? How does one determine what information about oneself should or must be made available to others? What uses will be made of the information? 2 The third phase of office automationsmall computers linked or networkedfurther raises these questions, as the ubiquitous placement of computing devices in offices gives more people more opportunities to access records. Alan Westin, New Issues of Computer Privacy in the Eighties, Proceedings of ZFIPS 11or]d Congress, Paris, France, 1983. PRIVACY, CONFIDENTIALITY, AND SECURITYDEFINITIONS Since analysts from various fields and disciplines write about privacy and security issues, there are differences in the use of terms and concepts by these practitioners. In the broadest sense, privacy is a set of values dealing with individuality, autonomy, personal space, and personal information. 3 Privacy deals with the rights of an individual to limit others access to information about oneself, and the social or legal rules by which such claims are accepted or rejected in particular contexts. Viewed as a desirable attribute of the data and the way it is handled, this is better termed confidentiality the protection of privacy. Security deals with a data-collectors capacity to safeguard the existence and integrity of the data it has collected and to provide the proper degree of confidentiality as set by organizational or legal policy. For a general discussion of privacy and how it is defined and used in policy see, Priscilla M, Regan, Personal Information Policies in the United States and Britain: The Dilemma of Implementation Considerations, Journal of Public Poiiq, vol. 4, No. 1, 1984, pp. 19-38. Confidentiality and security are related but not synonymous. Confidentiality addresses the use of data about individuals. Security is concerned with the accidental or intentional theft, modification, or destruction of data. Breaches of security may compromise privacy; for example, the theft of a mailing list stored on magnetic tape is a result of poor security and may compromise the privacy of individuals on that list. The breaches of security may also be unrelated to privacy or confidentiality. 4 Respect for privacy in office automation involves three components: 1. Data collectionwhat personal information is relevant, necessary, and socially acceptable for an organization to collect to carry out its missions? 2. Protectionwhen should an organization record and preserve identified personal data, who should have access to it within A representative discussion by EDP experts of the relation between privacy and security considerations appears in Alexander Gaydasch, Jr., Principles of EDP ilfanagement (Reston, VA: Reston Publishing Co., 1982). 171
PAGE 179
172 l Automation of Americas Offices the collecting organization, and under what circumstances can it be released outside the organization to third parties? 3. Notice and accesswhen can the subject of data collection know that an identified record has been created about him or her, have the right to examine the record, and be able to challenge the accuracy, completeness, or proper use being made of the record? Survey evidence suggests that American:~ concerns about privacy are rising. A Louis Harris Survey indicated that during the period of 1970-78 American concern about the hvi~sion of privacy rose from about 33 to 64 percent. 5 This public concern was a factor in bringing about Federal privacy legislation. There was also a shift in employee attitudes about the prerogatives and responsibilities of employers with regard to employee data, adding further impetus to Federal policy activity. By > The Dimensions of Privacvv: A National Opinion Research Survey of Attitudes Toward Privac~r, conducted for Sentry I 1surance by Louis Harris & Associates and Alan Westin, 1973. 1983, another survey by Louis Harris 6 indicated that the proportion of the public concerned with privacy had increased further, from 64 to 77 percent. This chapter is concerned with both confidentiality and security issues raised as more and more organizations introduce new office technologies, in both the private and public sectors. While privacy or confidentiality and security are interrelated, this chapter first discusses confidentiality and privacy issues, then security issues in the protection of personal and client data. 7 Louis Harris & Associates, Inc., The lioad After 1984: The Impact of Technology on Society-A Nationwide Survey of the Public and Its Leaders on the New Technology and Its Consequences for American Life. Harris study No. 832033, 1984. Confidentiality and security of data involved in off-shore sourcing of data-entry work are discussed in ch. 8. Issues of software security are being addressed in another OTA report, intellectual Propertey Rights in an Age of Electronics and Information (winter 1985). Issues related to the security of Federal information systems are being covered in the OTA report, implications of Federal Government Information 7echnolog.y (winter 1985). Privacy issues in electronic surveillance are covered in Electronic Surveillance and Civil I.iberties, OTA-CIT293 (Washington, DC: U.S. Government Printing Office, October 1985). OFFICE AND DATA PROTECTION Much of the rising concern about confidentiality and security has been occasioned boy the advent of computers and large data banks, and these concerns have repeatedly been subjects of congressional action, resulting in new laws. 8 End-user computing, in which many peo ple may access and use organizational databases, raises new questions about how claims to privacy can be respected and the confider~tiality of information be assured. In a recent survey of privacy and security professionals, almost half, 47 percent, believed that as a result of the third phase of office automation In the mid-1960s, a Senate investigation was held to exanline the kinds and amounts of personal information collected by the Federal Government. This investigation is considered by many to be the beginnings of the national expression c~f concern about the collection and use of computerized personal records systems. (U.S. Congress, Senate Committee on the Judiciary, Subcommittee on Administrative Practice and Procedures, Government Dossier) (Washington, DC: U.S. Goverrlment Printing Office, 1967). there is a trend for broader or more detailed personal information to be collected. 9 Concurrently, this proliferation of computers is leading to decreasing compliance by managers with government-privacy regulations (45 percent). Networking, computer-based messaging systems and electronic mail will exacerbate these problems still further as personnel data can be circulated among more people. Computer crime legislation passed in the 98th Congress established penalties for indiWestin and Hoffman, op. cit., 1985. This survey was not a representative sample of all types of organizations affected by office automation. Rather, the participants were chosen based on their reputation in computer and business circles as active, advanced, and unusually skillfull users of office-systems technology and as leaders in dealing with the employee-relations and organizational change aspects of new office technologies. However, there is a paucity of quantitative data on the extent of the privacy and security problems. Thus, the information reported from this survey should be treated as indicators of problem areas and not as indicators of the extent of the problem,
PAGE 180
Ch. 6Confidentiality and Security Issues With Office Automation l 173 viduals convicted of theft, fraud and abuse associated with Federal computer systems. The first indictment and conviction under the new Computer Crime Act involved a person convicted of misrepresenting himself as a valid user of the computer system, establishing a super user code allowing future access to the computer he worked on and other systems to which it was linked. As organizations rely more and more on computer-based information to conduct their business and to keep track of their personnel, the privacy rights of clients and employees must be considered. Coupled with the fact that the purchase and installation of microcomputer systems is often haphazard and uncontrolled within organizations, an appropriate degree of confidentiality is more difficult to assure when access to data is widely distributed. Personal information is required to effectively run a business, and with telecommunications linkages between computers and between organizations information can be collected in greater quantity and shared more easily. Client information can be processed at the clients workplace and sent via phone lines to the organizations computer system for processing. Office automation also raises questions about traditional security measures in central electronic data processing (EDP) environments. The purchase and use of computers in office automation is not controlled by a single department. Once microcomputers are linked with larger computer systems, any individual on the system has the potential for unlimited access and distribution of information. Before the recent wave of office automation, the principle users of a computer system were professionals, and the two chief dangers were theft of funds or data by employees, and that of novices external to the organization breaking into the system. The dispersal of computers throughout an organization has extended accessibility to noncomputer-professionals. The computer professionals sociocultural system had within it values, beliefs, and concerns about privacy and security of information. Counterfeit Access De\ice and Computer Fraud and Abuse Act of 1 9X4, Public I.aw -473. Today, people who never participated in the sociocultural system of EDP environments, who have not been informed about the fine points of the law and ethics of protecting confidentiality or assuring data security, have easy access to an organizations databases. Violations can result from carelessness as well as through intent. This chapter differentiates current office automation issues from the confidentiality and security problems that have been present in many organizations for two decades as part of an EDP system. This distinction between centralized and end-user automation is probably transient and likely to disappear as offices tie their small computers into the larger EDP systems highly integrated information systems. The transient nature of current systems is important for policy considerations later, and the distinction is useful in discussing new problems that arise with office automation. Three developments in office automation are central to confidentiality and security issues, not only because they deal with the technological changes mentioned above, but with how the l l l technology is procured and installed: the arrival of stand-alone word processors and text editors in the late 1970s, perceived as higher-order typing instruments and generally procured and controlled by office administration staffs or user departments; the move to widespread professional microcomputers, now available at prices and with features that allowed organizations to procure them for stand-alone work, and not only when linkage to on-line mainframes and minicomputers was involved. Generally, these machines were ordered by user departments; though often with EDP-department guidance; and the explosion of personal computers (PCs) in the early 1980s, for professional, sales, and executive use as stand-alone computing devices, and increasing linkage of such PCs to minis and mainframes. PCs in most organizations were ordered by individuals or by user departments, with limited guidance or control by EDP departments or central administrative service functions.
PAGE 181
174 l Automation of Americas Offices Each of these developments fostered new ways of carrying out office work. By 1984-85, each of these three original types of stand-alone terminal/software machinery were being linked to minicomputers and mainframes. This building of communication between computers makes possible access to central or distributed databases. In the survey of computer security professionals mentioned above, 64 percent believed that the spread of electronic mail and messaging systems poses a real problem in maintaining confidentiality. The ability to download (take data from) or upload (add data to) mainframe databases, using small computers, makes it possible for any employee with access to the building or computer to generate a file of personal information. Seventy-six percent of the survey respondents felt the automated office environment allowed more people physical access to workstations and disks as compared to earlier waves of automation. Finally, the interconnection of microcomputers allows the exchange of data within units or across units of the organizations, independent of mainframe system controls. The following quote demonstrates the capabilities of the convergence of the various office technologies: In IBM we now have a world-wide network of more than 1,500 mainframes and 200,000 terminals. Any user at any terminal can send a message or a file to any other user. A user at any of those more than 200,000 terminals can connect to any application in any of those 1,500 systems. Programs and even entire applications spread spontaneously through the network, usually without management direction or intent and often without management understanding or knowledge. Employees can access the network from their homes. Some vendors and contractors use it [DP] management did not plan it. They bought it and built it but they were just as surprised as anyone else when they saw what God had wrought. 11 W.H. Murray, Security Programs, Functions, and Concepts for the New Computer Economics, Proceedings of the 1 lth Annual Computer Security Conference (Northboro, MA: Computer Security Institute, 1984). THE PERIOD 1960-78 EDP systems made it possible for client data to be: l l l l l collected and recorded more easily; analyzed, compared, and collated more fully; distributed to or made accessible to more people within the organization; amalgamated with data on individuals obtained from other organizations having computerized record systems; and disseminated more widely, both in intraorganizational networks and in response to specific demands from other organizations. These capabilities made it possible to provide more customized and personal services by business and government. They also raised the possibility of greater aggregation of data-computerized databases can break down the vital compartments and boundaries that helped keep client information confidential. This is what the privacy-and-data banks debate of the 1970s was about. 12 There were also several aspects of internal organizational control in the 1960-78 era, before end-user computing, that provided the framework for carrying out new privacy rules that were developed: l in the EDP era, small staffs of EDP professionals ran the data centers and controlled access to automated files; and centralized security procedures of access and audit were available to check (if necessary) that terminal users followed the legal and organizational rules. For an early discussion of the debate see, U.S. Senate, Committee on the Judiciary, Subcommittee on Constitutional Rights, Federal Data Banks and Constitutional Rights, 1974.
PAGE 182
Ch. 6Confidentiality and Security Issues With Office Automation 175 Computer-system security measures originally evolved from manual techniques for physical security carried over from the electronic accounting machines of the 1950s. Usually physical security (locking up sensitive card decks and tapes) was the prime technique. As first and second generation batch processing computer systems were introduced, followed by remote terminals and third generation systems, computer security techniques became more technological and sophisticated. Passwords were used to restrict access to information based on a users or terminals privileges or on the type of function being performed. A few systems, especially those involved in national defense, used cryptography or enciphering. Some systems provided other mechanisms to assist proper authorization, most notably the hardware rings provided by the GE/Honeywell machines of the late 1960s and early 1970s. During this period, it was hard enough to get a program to run at all on a given hardware system, and security took a back seat to other design criteria such as correctness, speed, cost, and utility. As third generation computer systems became more common in industry and government, the use of data within organizations changed and networks were built up. Rather than use one computer center, one or more systems could be accessed via user terminals. Indeed, commercial time-sharing networks grew to service the needs of organizations without the resources or inclination to start their own networks. Minicomputers (minis) appeared, but in contrast to todays microcomputer systems, minis were generally still purchased and operated by data processing specialists within the organization. At the same time, there was growing public awareness of potential loss of privacy and problems of fairness or due process. Thus, the security of computer systems took on added importance and led to additional safeguards such as logs, journaling, of important operations and more specific policies with respect to computer security, as a means to guarantee confidentiality and accountability. However, the most sensitive narrative information was still not computerized at this time; it remained on paper in manual files. Thus, most problems involving personal data in the 1970s still were associated with paper records. In the early 1980s, medical, banking, credit and employment information was placed in large databases, beginning the current rapid move to store personal, employee, and client data on-line. The advent of commercial products to enhance computer security, such as the ACF2 access control package for IBM mainframes and various vendors versions of the National Bureau of Standards Data Encryption Standard had not yet been introduced. Leading manufacturers were starting to take seriously the task of educating users about these issues. An advance guard of computer practitioners were becoming knowledgeable about computer security issues. Second-generation technological security issues (e.g., the idea of kernels in software engineering) 13 were just starting to be investigated. A security kernel is a small nuclear piece of the operating system that controls access to other parts of the computer systemeither information or data. This nucleus itself must be tamperproof so that its programs maj not be modified, allowing system operators to verify whether it has implemented the systems security policy through the programs. See G.J. Popek and C.S. Kline, Issues in Kernel Design, Ad~-ances in Computer Seczmit.}r, Rein Turn (cd. ) (Dedham, MA: Artech 130use, Inc., 1981), pp. 139-144; and R.C. Summers, An O\rerview of Computer Security, Il?ltf SJwtems Journal, vol. 23, No. 4, pp. 309-325. This de~elopment has been primarily in the Department of Defense; there are no commercially a\ailable security kernels. THE POLITICAL AND LEGISLATIVE CLIMATE The response of American society to both ied, 14 By the late 1970s, a detailed latticework the social change aspects of privacy protection and the computer-based handling of perSee for example, J$. H. YIare, Information Systems Securit~r and Pri\acy, (mnmunications of the ,4(, ~ol. 27, No. sonal data by organizations has been well stud4, 1984, pp. 31.5-321,
PAGE 183
176 l Automation of America Offices of laws, regulations, organizational policies, and social expectations regarding privacy protection in EDP systems had been put into place. Codes of fair information practices were embodied in law or organizational standards to govern the collection, use, and release of personal data on clients and customers, and to make this process visible and accountable to both data subjects and the public. Laws and regulations were promulgated to deal with privacy in particular fields of organizational recordkeeping-Federal agencies, banking, insurance, health care, education, credit-reporting, employment, law-enforcement, etc. The earliest statute was the Fair Credit Reporting Act of 1970 (15 U.S.C. 1681). This act requires all credit investigating and report ing agencies such as banks and retail charge card firms to make the records they collect available to the subject. Furthermore, it provides procedures allowing the subject to correct the information. Finally, it only allows disclosure to authorized customers. Since thi:s statute, many more have been enacted providing protection policies for information privacy. The Crime Control Act of 1973 requires that State criminal justice record keeping systems, developed with Federal funds, ensure the privacy and security of the information collected. The Privacy Act of 1974 (5 U.S.C. 552a) restricts the collection, use, and disclosure of individually identifiable information by the Federal Government. It gives the individual rights to access the information and to correct it. The Tax Reform Act of 1976 (26 U.S.C. 6103) protects the confidentiality of personal tax information. It restricts disclosure of tax information for nontax purposes. In furtherance of such laws, and in manly areas where no legislation had yet been enacted, 1Exampl=s would be the Federal Privacy Act of 1974 ( 5 U.S.C. 552a); the Federal Freedom of Information Act Amenc.ments of 1974, and similar jurisdiction-wide statutes in 12 States. At the statutory level, this is exemplified by the Fan--ily and Educational Privacy Amendments of 1974 (20 U. S. C. 12239), and by State medical, banking, insurance, and employeeprivacy legislation. many private and public organizations developed privacy codes or fair information practices standards to govern their own handling of client or employee personal data. The motives for such action were a blend of concern to meet legitimate privacy concerns of clients and employees; the desire to avoid the necessity of detailed legal regulation; and the judgment that fair-information-practices rules had a generally positive effect on accuracy, completeness, and timeliness in the management of automated data systems, and could be initiated without heavy costs in money or efficiency. A few corporations such as IBM, Bank of America, and Control Data Corp. had also developed employee privacy codes in the early to middle 1970s, but most of the 10,000 large private employers in the United States had not. By the time that end-user computing began to add new dimensions to the handling of personal data in office work in the early 1980s, there were still debates among privacy advocates as to whether the data collection and the confidentiality aspects of privacy had been ade quately dealt with, in what was coming to be called the first-generation of privacy protection measures. In spite of the emphasis on notice, challenge, and due process rights of data subjects, and workable procedures for strengthening confidentiality rules, critics argued that there had been too few limitations on what was appropriate information to collect about peoples transactions and activities in many sectors of business and government life. They also argued that merging data from different organizations about the same individual (e.g., in computer matching programs) threatened to shatter basic confidentiality standards. There was also debate among informed observers and the media about whether the necessary machinery with which to enforce privacy rights had been created, and about other questionsthe implementation of the Federal Privacy Act of 1974, the desirability of a continuing Privacy Commission or Privacy Ombudsman (as Canada and many European countries have), and U.S. Supreme Court rulings
PAGE 184
Ch. 6Confidentiality and Security Issues With Office Automation 177 in matters of data collection by business and government organizations. 16 The report of the U.S. Privacy Protection Study Commission in July 1977 urged new laws and voluntary codes in the private sector. What followed was a new series of statutes that affected information privacy. The Right to Financial Privacy Act of 1978 (12 U.S.C. 3401) limits the access of Federal agencies to information about the customers of financial institutions, by describing the procedures necessary to obtain that information. It also provides bank customers other assurances of privacy about some aspects of the bank records. The Privacy Protection Act of 1980 (42 U.S.C. 2000aa) prohibits government agents from conducting unannounced searches of press office and file records if no person in the office is suspected of having committed a crime. The Electronic Funds Transfer Act of 1980 states that any institution providing EFT or other services must notify customers about third party access to customer accounts. The Debt Collection Act of 1982 (Public Law 97-365) establishes due process procedures that Federal agencies must go through to release any information about bad debts to credit bureaus. The Cable Communications Policy Act of 1984 requires any cable service to inform the subscriber of any personally identifiable information collected. They must inform the user about how the information is to be used and how long the information is maintained on record. If the information is disclosed, the individual must be informed and restrictions are placed on how and to whom information is disclosed. Only after the passage of new privacy legislation in 1978-82 protecting bank depositors, There were debates over whether new systems or applications were covered by first generation privacy measures; for example, electronic funds transfer (EFT) systems, two-way cable systems, and Postal Service electronic-mail projects. There was also concern that plans for large sophisticated computer systems as replacements for older systems in the Federal Government might go forward without sufficient attention to privacy risks, for example, the proposed FBI Triple 1 system for processing criminal history records, and the planned IRS, Social Security, Secret Service, and Veterans Administration systems. and State laws on insurance, employment, and medical privacy did broad private sector institutionalization of privacy rules take place. The establishment of organizational procedures and regulatory or judicial enforcement of privacy was just becoming the norm when end-user computing began to spread in the early 1980s. In the mid-1980s as office automation is spreading, an extensive set of confidentiality protections for manual and EDP systems has been put into place; but debate continues as to whether these protections are adequate in scope and are being vigorously administered. The laws above basically deal with rights to confidentiality. Until recently, no Federal law dealt specifically with sanctions against the use of computers by individuals to commit a crime, or with trespassing by reading private computer files.] In October 1984, Congress passed the Counterfeit Access Device and Computer Fraud and Abuse Act of 1984, which makes it a felony to access confidential or restricted information related to national security without authorization, and makes it a misdemeanor for unauthorized persons to access the data banks of financial institutions, or to use, modify, destroy, or disclose information in a government computer. There are of course many criminal laws that can be used in prosecuting computer-related crime, but there are problems in applying many of the laws defining theft to cases where only virtual property (nonphysical property) is concerned. Thirty-three States now have computer crime laws, but some do not cover hackers who penetrate systems for fun rather than profit. The Privacy Act of 1974 (Public Law 93-579) and the Crime Control Act of 1973 (Public Law 93-83, sec. 524(b)) are, however, designed to prevent misuse of Federal records of all kinds in ways that would violate the privacy of citizens. List provided by the National Center for Computer Crime Data (Los Angeles) includes Alaska, Arizona, California, Colorado, Connecticut, Delaware, Florida, Georgia, Hawaii, Idaho, Illinois, Iowa, Maryland, Massachusetts, Michigan, Minnesota. Missouri, Montana. Ne\ada, New Mexico, North Carolina, North Dakota, Ohio, Oklahoma, Pennsylvania, Rhode Island, South Dakota, Tennessee, Utah, Virginia, Washington, Wisconsin, and Wyoming.
PAGE 185
178 Automation of Americas Offices THE THIRD PHASE OF OFFICE AUTOMATION The Handling of Client Data Confidentiality in an organization with new decentralized or networked office technologies requires consideration of a number of new factors: Low-security or no-security physical environments in offices. Except for military, diplomatic, and a few private-sector settings that operate in highly sensitive data modes, most organizations have put microcomputers onto desks in working areas that are open to passing fellow workers, service personnel, invited visitors, or even the general public. Storage of diskettes, location of printers, and other peripheral equipment is also usually in unsecured environments. More finished and relined information in office automation systems. In most organizations, word processors contain finished correspondence, memoranda, documents, charts, and reports. These stored materials may contain sensitive personal information on clients, proprietary data of the organization, information confidential to particular employees or units, and data about terms of contracts and other legal responsibilities. While much raw data and some reports of this kind are in mainframe storage, the information in word processors or professional/executive microcomputers is generally more complete, revealing, and easily found, because it includes explanatory text. More readily accessible information in office computer systems. An interloper, whether an employee or someone from outside, who seeks to extract sensitive data from the mainframe database generally has to know specialized mainframe software codes, as well as how to deal with special security protections that most organizations have built into mainframe databases. With microcomputers, unskilled interlopers are more able to call up data on screens, print out documents, or copy diskettes than they would be in EDP systems. The mobility of microcomputers and their data-storage media. Dumb terminals connected to mainframes are known to the system, and their uses are usually logged or monitored. Microcomputers can be moved around within organizational offices without central control or notice, and portable versions can be taken home or on trips. Information is archived on floppy discs rather than magnetic tapes. Less sophisticated office automation users. While 10 percent of organizational personnel may have been using terminals in the middle 1970s, either in EDP units or as EDP-trained operators in users departments, at least 25 to 35 percent of personnel in large organizations are probably using microcomputers today, and eventually this may approach 100 percent. Typically, new users have not been educated about confidentiality issues or security protections. Uncontrolled channels of data communications. Electronic message systems, bulletin boards, and microcomputer networks encourage users to send messages to everyone using these systems or to create their own distribution lists. As a result, messages with confidential data can move around either anonymously, or without control for their confidentiality. Fraudulent memos can be circulated for political purposes without an easily traceable origin. Reproducing machines, and before them, mimoegraph machines had the same effect, of course; computers merely allow this communication to be done in the privacy and safety of ones own office or home. 19 Wide ability to add information to or copy or extract information from corporate databases. Controls over alteration of data or unauthorized access become critical. 20 Microcomputers can be used to attack the security of mainframe and mini data, and obtain confidential client data improperly. This can in fact be done from any terminal, including those developed for mainframe use, but more understandable This situation may seem trite, but a recent State supreme court ruling stated that an employees privacy had been violated by an interoffice memorandum. S,JC Outlines Rules on E~mployer Rule in Workers Privacy, Boston Globe, July 7, 1984; as cited in Philip Adler, et al., Employee Privacy: Legal and Research Developments and Implications for Personnel Administration, Sloan Management Review, winter 1985, pp. 13-22), In this case upheld by the U.S. Court of Appeals, First Circuit on Aug. 6, 1984; the memorandum was on paper. This does not address. the problem of unauthorized users external to the organization accessing the databases. One commonly suggested procedure is the use of tiered passwords for accessing the computer and then the particular volume or file the information is stored in.
PAGE 186
Ch. 6Confidentiality and Security Issues With Office Automation 179 software instructions are expanding the number of persons able to probe password security, attempt unauthorized entry, or alter confidential data. Little or no hardware or software security protections. Because of the ways that word processors and microcomputers entered office work, there was neither a perceived necessity or an assumed market demand for built-in terminal hardware protections such as locks on the machines. Similarly, in buying or designing software for office systems, organizations have generally not specified audit trails, cryptographic protections, and other measures that could have been installedthough at added costs. Unpredictable individual and group behavior. There is no way to predict with any assurance how office employees and managers will use these powerful new capacities. There will be some conduct by insiders and outsiders arising out of the new opportunities in microcomputer use that pose risks, and this has to be the perspective from which managers assess risks to confidentiality and security. Security and Confidentiality Issues The potential vulnerabilities to client data emerging in office automation can involve either new situations created by microcomputer use, or can be extensions of familiar risks in manual and EDP recordkeeping. The types of potential end-user conduct that could violate existing privacy standards are discussed below. Data Collection Microcomputers permit the creation of any files or databases that the end user wishes to maintain. Either without awareness of or in deliberate disregard of privacy, end users may I+; xamining such potential twlnerabilities in office automation follows the studies of EDP s~stems impact done by the National Academy of Sciences in 1969-72; the HEW Ad\risory Committee in 1972-73: and the U.S. Pri\acy Protection Study Commission in 1975-77. Hopefully, the examination of potential office automation ~wlnerabilities can benefit from the experiences with risk-assessment and reality-testing that were de~eloped in 1969-77 EDP studies, as well as the practical experiences gained o~er the past decade in administering the standards promulgated (for client data) as the *first generation of pri~ac? and securit?r rules. put personal client information into their files that violate laws, regulations, organizations rules, or ethical guidelines. If there is no auditing or physical inspection of end-user files, then management may be unaware of such conduct, and suffer legal sanctions or have the companies reputation compromised. Confidentiality End users may take client information from central files and merge it or match it with other client data in ways that violate privacy standards. Through electronic mail and message systems, end users may send confidential client information inadvertently to those not entitled to have it, especially if automatic distribution lists are used. Subject Notice and Access Where end users create files improperly or record confidential data they should not have accessed, failure to inform clients that such files have been created and might be used in making decisions about clients could be a violation of several basic privacy laws (the Federal Privacy Act, State fair information practices acts, Fair Credit Reporting Act, State insurance privacy acts, etc.), as well as the organizational rules of many banks, insurance firms, brokerage houses, medical facilities, educational institutions, and other private organizations. Failure to provide opportunities for individual clients or potential clients to examine and challenge these files would be a violation of such laws or rules. It is critical to distinguish the privacy issues generated by the transition from manual record systems to EDP systems, from those of the current progression from centralized to decentralized computing. The arrival of EDP brought about revolutionary increases in two areas: 1) data-collection capacities (reducing costs and time constraints, magnifying dataanalysis capabilities, etc.); and 2) data-sharing capacities (circulating personal data within and between organizations). The response in terms of privacy laws and organizational codes was: 1) to increase the visibility of organizational activities affecting sensitive personal data
PAGE 187
180 Automation of America Offices (changing these from private kitchen work to public notices and descriptions, with specific rights of data-subject inspection challenge); 2) to stipulate broad relevancy and social acceptability standards for data collection; and 3) to provide rules and procedures for confidentiality or data sharing. The use of microcomputers in the office has the l l l foIlowing characteristics: They do not significantly increase the scope of data-collection capacities over existing EDP systems. However, because of their ease of use or cost, data not entered into the mainframe system can now be entered, potentially increasing the quantity of data collected. This requires managers to provide greater education and oversight of users to see that they know the limitations on data collection set by law or code, and do not violate these in personalistic data recording. They do increase the risks of improper circulation of personal client (or personnel) data within the organization and to outside organizations, threatening confidentiality, because they provide many more employees with a tool for accessing the records or information. They do increase due process problems. To the extent that stand-alone-diskettes and off-line storage of personal client data would not be known to data subjects, the use of that data for decisionmaking could not be challenged under privacy rules for subject access. Security How to protect records from accidental or deliberate destruction, loss, or theft is a security question. There are also differences between EDP and decentralized office automation. Surveys of private organizations show general agreement that by the end of 1984 there were significant security issues in use of microcomputers, not just for sensitive client and personnel data but also for general proprietary business information, financial data, national security information, and legally sensitive information. These problems have not been taken up yet by most top managements, and thus the policy directives and budget authorizations necessary to address this problem adequately have not existed in most private-sector organizations. The major problems developing with decentralized office automation are: Lack of clear identification of sensitive information. The basic requirement for sound security is to identify information that is sensitive and needs special protection. Any effort to protect all personal information in ordinary business or government organizations would be too costly and would virtually paralyze organizational programs. z Failure to provide adequate physical security for machines and storage media. Many microcomputers are not kept in locked rooms and diskettes are often not kept in locked cabinets or desks. Easy physical access to such microcomputer equipment poses real security risks; for exampIe, diskettes can be copied on another machine on or off the premises. Failure to have key locks on terminals. Most microcomputers do not have key locks that control on-off functions, enabling third parties to activate them. Weaknesses in password systems governing access to central databases. Microcomputers connected to mainframes in many organizations suffer from the same security problems as dumb terminals; users are casual about writing down or telling their passwords to others. But microcomputer users are probably less disciplined in handling password regulations than EDP-trained personnel. No logs or journals. Though central databases of confidential client data are often provided with audit trails, transaction logs, or journaling capabilities, these techniques are often not used when groups of microcomputers are connected to minicomputers or, through them, to mainframe files. Not recording efforts to penetrate security. Unlike mainframe systems, microcomputerbased office systems generally do not record efforts to enter restricted files without proper identification or passwords, or warn security officers that such efforts have taken place or are under way. -Many European countries do attempt to protect a large amount of the personal data collected.
PAGE 188
Ch. 6Confidentiallity and Security Issues With Office Automation 181 Absence of either security education for users or auditing of their practices. Even if sensitive information was identified, security measures were adopted, and records were kept of efforts to access databases improperly from microcomputers, most experts would agree that security of sensitive client data also requires that users be educated about security policies and procedures. Second, it requires that inspections or auditing be carried out to learn whether security policies are being followed. When it comes to security measures needed to safeguard sensitive client data in microcomputers and end-user office automation, current evidence shows that: 1) the techniques for installing security protections are well known; 2) the process of risk-assessment and vulnerability analysis to determine cost-beneficial policies is well known; but 3) these techniques and processes have just begun to be undertaken in the private sector, and are only somewhat further advanced in the Federal establishment. The first task in dealing with security is not to identify wholly new security approaches or methodologies, but to stimulate organizations to provide directives and resources with which to modify and apply known techniques and processes. There are some situations in which new forms of office automation may require special measures. For example, if combined voice/ data terminals are widely used, manufacturers may have to provide means of preventing the undetected turning on of the microphone capabilities of terminals, to ensure that neither officials within the organization nor outside hackers, or more serious intruders, used these terminals for organizational espionage. Local area networks present particular problems. Three different types of network configurations exist and present different security problems. In a star network, several terminals are connected to a central controlling device; the central computing power can be used to control data and software, maintaining a high degree of security. In a ring configuration (or a loop) the workstations are arranged in a circular network. Each station is linked by a repeater mechanism that monitors all passing information to see if any are addressed to that workstation. This configuration has gained more acceptance in Europe than in the United States. Because information travels around the ring, any workstation has the capability of accessing it. The tree or bus network, the most easily expandable of the three, does not require a central controller and, security is therefore contingent on the security capabilities of individual workstations. PRIVACY ISSUES IN WORK MONITORING The monitoring of office work by computers was discussed in chapter 5; but it is sometimes discussed as a privacy issue rather than as a quality of worklife issue. The privacy issue raised by computer-mediated work monitoring is whether the collection of operator performance data through machine capacities and its use to evaluate employees constitutes an intrusive form of employer surveillance that violates reasonable expectations of personal privacv by the employee. Whatever the pros and cons of computerized-work monitoring, it is probably not best posed in terms of privacy. The work is done on the employers premises; the activities are usually group settings open to view rather than individual activities taking place in closed or private rooms; supervision is a normal condition of the employer-employee relationship; and collecting quantitative data as to employee output has long been used in evaluating and compensating work performance in factories and offices.
PAGE 189
182 automation ofAmerica Offices Unlike the use of TV-monitors to watch asployees know that monitoring systems are sembly lines, or of hidden microphones to overused, employees have access to their individhear workers in cafeterias or restrooms, the ual records; and a procedure is provided for collection of operator-production statistics contesting the accuracy or fairness of applygenerated by system software does not repreing records for evaluative purposes. sent an intrusive act per se, provided that emSECURITY AND CONFIDENTIALITY IN AUTOMATED PUBLIC OFFICES The protection of data is a major concern in public-sector offices, especially in regard to matters of national security, diplomacy and foreign relations; the Federal role in monetary transactions; government funds transfers; and foreign trade. Most of the sensitive information about such subjects, however, is usually in large computer and communication systems, and is usually protected by encryption and a variety of other mechanisms. However, shared databases and the downloading of data to microcomputers for end-user computing is raising new concerns, especially since data that are not per se identified as sensitive can, when aggregated, reveal information that is highly sensitive. The Federal Government collects large volumes of detailed personal data about citizens and in particular about Federal employees. As early as 1967, a study of computerized Federal records revealed that the files contained more than 3 billion records and that over half of them could be accessed via computer terminals. 23 Several years later in 1974, another congressional committee found that 86 percent of the 858 known government databanks were computerized. 24 Successive waves of office automation have continued to provide greater access to these expanding data files. This information can be integrated through computer matching and other techniques and through the exchange of data between agencies and with State governments, in ways that cause deep concern about confidentiality. For example, when one applies for Food Stamps, Go\,ernn]ent Dossier, op. cit. Federal Data Banks and Constitutional Rights, op. cit the Department of Health and Human Services, which administers the program, matches ones name with those in the Internal Revenue Service (IRS) earnings file (which is in fact maintained and used by the Social Security Administration (SSA) to verify eligibility. SSA checks the data in the IRS earnings file to verify income reported by recipients of SSA retirement benefits. 25 In general, information-handling associated with end-user office automation is not classified information, but it is often sensitive information, especially when aggregated in certain ways. 26 Routine records, forms, and correspondence often contain information that would allow individuals to put together and profit from advance knowledge of government actions. For example, plans for siting highways and government facilities, impending regulations, actions that affect interest rates, sales of minerals and timber rights, etc., are very tempting. Personal data about Federal employees can be used to discourage whistleblowing. Less often voiced, but still important, is the concern about unauthorized access to ~ According to briefings and interviews provided for OTA by SSA. -This has chiefly been of concern to the Department of Defense; for example, the problem was exhaustively discussed in a planning conference held for the U.S. Army Information Systems Command by SRI in Tucson, AZ, in December 1984. Hypothetical examples given in an informal talk concerned the possibility of aggregating routine travel orders or schedules for key individuals to reveal the (undisclosed, sensitive) location and timing of small meetings; or the possibility of aggregating data on materials delivery to a site to reveal information about development of weapons systems, However, some civil liberties specialists have long been concerned about the ability to aggregate information from many sources about one indi\idual to produce a profile of his or her marital, financial, social, business, professional, and political activities.
PAGE 190
C/J 6Confidentiality and Security Issues with Office Automation l 183 information that is simply embarrassing or politically fatal to officials or to their nominees for official positions. In congressional members offices, there are grounds for concern about protection of information in computers, which is chiefly protected by IDs and passwords. (The privacy and security issues related to large Federal computer systems and databases will be covered in a forthcoming OTA report. 27 Even when nonclassified Federal data is normally stored in large databases or processed in central EDP centers, it is increasingly hable to be at some point accessed, handled, analyzed, or even generated using personal computers or terminals scattered throughout agency headquarters and field offices. Most of the concern about government security has focused on large information systems. There has been much less attention to the protection of data in day-to-day agency operations in which decentralized office automation is used. Violations of security and confidentiality need not be intentional or malicious; they are often the result of ignorance or carelessness. Office workers have not yet been acculturated to think routinely about computer security; there are stories of government workers routinely locking away sensitive papers but leaving the disks from which the copy was made lying next to their personal computer or word processor. Government computers are thus subject to the same risks that plague banks, corporate payroll and financial management operations, and other private-sector computer systems, plus some risks that are particularly political in nature. As extreme examples, records could be destroyed to cripple a government program or project; or calendars and trip schedules could be used to plan the assassination of a political leader or a foreign dignitary. Implications of FederaI Go\crnment Information 7echnoIo~;~ for Ci\.il I.ihertie.s and Congressional O\ ersight, op. cit. According to the Itrashington Post. an air traffic controller, angry about the Soviet in~asion of Afghanistan, deliberatel~ endangered an Aemflot jet carr~ing So\iet ~\mbassador Dobr~nin, b~ manipulating a signal so that a computer read the jet as a small craft and did not properly monitor and control its landing at a bus~ airport. (hlar~r Thornton, Age of E~lectronic Con\enience Spawning In\entl\e Thie\res, The 11ashin@on })o.~t, Yol. 107, hla~r 20, 1 984). Government Guidelines Security and confidentiality for nonclassified government information is covered by the Privacy Act of 1974 and: OMB Circular A-71 Transmittal Memo No. 1, July 17, 1978, which provides general guidance to agencies, on administrative, technical, and physical measures to increase security; and OMB Circular A-123, which sets standards for internal controls implementing the Federal Managers Financial Integrity Act and directs each agency to review and update its security provisions. The General Services Agency (GSA) and the National Bureau of Standards (NBS) in 1983 began development of guidelines for security in end-use computing and small office systems. The Office of Personnel Management (O PM) recommends personnel security policies for computer-related jobs. However, the General Accounting Office (GAO) has repeatedly criticized the lack of compliance with guidelines established in the circulars by Federal agencies. 29 In 1982, GAO said that: increasing Federal investments in automated systems have resulted in growing vulnerability to fraudulent, wasteful, abusive, and illegal practices because greater concentrations of information are accessible from remote terminals. The Office of Management and Budget (OMB) is now updating A-71, A-123, and other circulars related to computer security. It is expected that they will be combined into one broad policy statement. 30 The Special Concern About Employee Privacy An aspect of Federal computer security that deservesbut has not receivedspecial atten-. .-. -U.S. General Accounting Office, Centra) Agwncies Con]pliance 11>th OJfE Circu]ar ,4-71, Transmittal Nlemorandum No. 1, I, CD-80-561, Apr. 30. 1980: and Federal Information S?stems Remain High]jIulnerah]e to Fraudulent, 11asteful, .4 busi\e, and Illegal Practices q NIASAD-82-18, Apr. 21, 1982, Susan \l. Nlenke, Sur\re~ of Agencies Finds Nlan} Are Implementing Standards, Special Section: Securit?, (;olernment Computer ,\e\~s, Yotrember 1984.
PAGE 191
tion is the question of the ability to protect, the privacy of Federal employees. Agencies collect and keep much personal data about em. ployees and even applicants for Federal jobs, Much more data is likely to be collected than would be sought by corporation personnel offices. These files may be, and usually are. widely dispersed and only superficially protected. As files are computerized, they can, more easily be aggregated and accessed, or tampered with. If files are stored in personal computers or can be accessed or downloaded, more and more people will have access to them. and may use them for satisfying curiosity, for mischief, for exerting political pressure, or for sheer malevolence. This raises significant questions about protection of privacy in Federal computer systems and databases. Federal employees are the subject of much computer matching, which was described earlier; that is, the comparison of two or more computerized lists of individuals. For example, Federal employee lists were checked against lists of people earlier defaulting on student loans. A very early example of computer matching occurred during the Carter Administration; this was Project Match, through which names of Federal employees were matched with names on the roll of the Aid to Families with Dependent Children, in an effort to discover ineligible recipients. 32 Accidental Losses Telecommunications and computers are vulnerable to crime, mischief, and terrorism, but they are also vulnerable to unintended disruption, a point that is sometimes overlooked. This includes error or accident, and the simple breakdown or failure of equipment; it also includes destruction by fire, flood, earthquake, and other natural and technological disasters. Data can also be lost permanently, or made temporarily inaccessible, by electrical outages. As government activities become more and See Alan Westin, Personnel Practices in the U.S. Civil Service, Information Age, July 1982, pp. 149-169. According to information supplied by Federal officials to OTA. more dependent on microelectronics there is the strong possibility-and given enough time the virtual inevitabilityof electrical failures that affect critical activities or whole regions of the country, bringing all computer-mediated activity temporarily to a halt. Emergency response systems may be unable to respond adequately because some step or link depends on office automation that is down because of the same emergency. Routine recovery procedures may be hampered by loss of data in the same incident. GAO called attention to this danger in a strong report in 1981, identifying the need for further emergency planning by the Department of Energy and the Federal Emergency Management Administration. 33 Some government operations are highly time sensitive (e.g., transfer of funds, transmittal of orders to the armed services, air traffic control, response to natural disasters, law enforcement actions). Others are dependent on ready access to individual and case records (tax collection, processing of welfare payments, payrolls, etc.). As the government becomes more and more dependent on office automation for orderly performance of necessary activities, the result of any disruption-whether irreparable or of brief duration and narrow scope becomes more severe. Federal agencies have not adequately prepared for natural or technological disasters that might wipe out electronic information, according to GA0. 34 Many still have no contingency plans or rely on letters of agreement from other agencies to supply equipment in emergencies. This course assumes that the other agency will: 1) be able and willing to live up to its agreement even at the cost of prejudicing some of its own activities, and 2) will not -. ... U.S. General Accounting Office, Federal Electric Emergency Preparedness IS Inadequate, EMI)-81-50, 1981. U.S. General Accounting Office, Most Federal Agencies Have Done Little Planning for ADPDisasters, AFMD-81-16, 1980; and Federal Electrical Emergency Preparedness Is Inadequate, EMD-81-50, 1981. A contingency plan should have most or all of the following features: backup files at external storage sites, standby arrangements for renting processing time or services, a recovery operation center, a multilateral aid agreement involving five or more agencies, or a plan for reverting to manual (nonelectronic) operation if necessary.
PAGE 192
Ch. 6Confidentiality and Security Issues With Office Automation l 185 have been disrupted at the same time by the same events. Nonelectronic equipment is often not available, and reversion to manual systems not possible; data may exist only in electronic, machine-readable form. The cost of downtime or time lost because the computer is malfunctioning, is seldom calculated in justifications or cost-benefit analyses of office automation. In most cases, the downtime is merely an annoyance and a temporary problem. In some cases, it significantly aggravates the workload peaks and degrades the quality and timeliness of important services. In time-sensitive situations, such as response to local emergencies, it can mean catastrophe. Comparisons With the Private Sector In spite of these problems and concerns, the Federal Government is probably well ahead of the corporate sector in attention to the need for safeguarding privacy and security in the use of microelectronic office automation. 35 This reflects the implementation of provisions of r This is the conclusion from an OTA contracted study, F?-itac~ and .9ecurit.v Issues in Office A utmnation, Alan F. \testin and I,ance rograms and I)ri\ate Industr>: (omputer .4 buses, (~onlmitt.ee Print, 94th Cong,, 2d sess., 1976; and Computer Securit} in Federaf Programs, Committee l)rint, 9,5th (ong., 1st sess., 1977. last three Congresses related to computer security in general and in particular to the security of computerized government information. 37 In April 1984, the House Committee on Science and Technology issued a special report on Computer and Communications SeFederal Computer Systems Act of 1977, S. 1766 and Ii. R. H421: S. 40 and Ii. R, 6196, 1979; Federal (omputer S?rstems Protection Act of 1981, 11. R. 3.
PAGE 193
186 Automation of America Offices curity and Privacy 38 that recommended that Congress charter a national commission to study the issues and outline a framework for policy. While recognizing the threat from hackers and other outside intruders, the report said that the greatest threat is from personnel who are authorized users of the computer resources that they attack. The report was highly critical of recent policy and programs related to computer security. OTA concurs that violations of data confidentiality are most likely to occur within the organization by authorized users. External threats, whether from hackers, ordinary thieves, political opportunists, disgruntled former employees, terrorists, or others, nevertheless deserve greater attention. But these threats chiefly concern large systems. With small computers violations by insiders are most likely, and there are fewer safeguards against them. Both large systems and small computers are at risk from accidental disruption because of natural events, human error, and technological failure. Because decentralized end-user computing is raising new uncertainties about how well data is protected, Congress may wish to reexamine the structure of privacy laws and thek application to these technologies, perhaps through a special commission or task force. Stringent actions have been suggested. One possibility is to establish a new authority to implement, oversee, or enforce current and future statutes. This organization would have no responsibility for collection and distribution of data. There are precedents for this: Sweden for example has a Data Commission to oversee all records and their linkage. Another U.S. Congress, House Committee on Science and Technology, Computer and Communications Securit~r and Pri\ac.y, report prepared by the Subcommittee on Transportation, A\iiition and Materials, Committee Print, 98th Cong., 2d sess., April 1984. possible strategy is to have an ombudsman within each data-collecting organization. This is the approach used, for example, in Germany. However, current laws may be adequate to address the problems of decentralized automation. Whether organizations treat office automation and centralized EDP as one or separate and distinct components of office activity, the critical element is to adapt existing rules to apply to predictable oversights, carelessness, or misuse by some persons in the large end-user population, and to assign organizational responsibility and continuing oversight duties to effective units at various levels of the organization. If organizations are not willing or able to create and enforce policies to ensure that existing safeguards and guarantees are applied in end-use computing to protect their clients and their employees, Congress may wish to clarif y and strengthen through legislation the liabilities that such organizations incur by their failure. Current laws chiefly strengthen the right and the ability of an individual to control information about himself or herself. Thus, the individual is the final enforcer of principles of fair information use. Given the ubiquitous nature of the new information technologies and their linkages and systems integration, more specific data collection and data protection policies may become necessary as opposed to traditional policy approaches strengthening individual rights. For Federal agencies, strong oversight attention is merited to make sure that reasonable security provisions are enforced. Special attention should be given to the questions of: 1) the ability of essential day-to-day government operations to continue when computers cannot operate, 2) the need for protecting employee data, and 3) the adequacy of procedures for protecting correspondence and records with congressional offices.
PAGE 194
Chapter 7 Home-Based Automate d Office Work
PAGE 195
Contents Page The Controversy Over Home-Based Office Work. . . . . . . . 1$9 Historical Roots. . . . . . . . . . . . . . . 191 The Status of Home-Based Office Work . . . . . . . . . 192 How Many Home-Based Office Workers .Are There? . . . . . . 192 How Many May There Be in the Future? . . . . . . . . . 193 Mechanisms for Home-Based Work ..., What Work Can Be Done at Home? Parties at Interest . . . . . Employers . . . . . . Productivity . . . . . . The Workers: Why Are They at Home?. The Benefits and Costs for Home-Based Family Considerations . . . Effects on Society . . . . . . . . . . . . . . 194 . . . . . . . . . 196 . . . . . . . . . 197 . . . . . . . . . 197 ...,, . . . . . . . . 198 199 workers? : : : : : : : : : : : : : : : : : : : : : : : : : : : : 200 . . . . . . . . . 203 . . . . . . . . . 204 Legal and Regulatory Barriers to Home-Based Work . . . . . . . 205 Public Policy Issues . . . . . . . . . . . . . . 206 Encouragement or Prohibition of Home-Based Office Work...., . . . . 206 Regulation of Home-Based Office Work..... . . . . . . . . 208
PAGE 196
Chapter 7 Home-Based Automated Office Work l New information technologies allow the dispersion of office work over time and space. As many functions of an office become internal to computers, and as computers are increasingly linked by communication networks, it becomes less necessary for office coworkers to be located in the same room, or in the same building. When a workers primary interaction is with the system rather than with other people, she can do her work wherever she can get access to the systemin the office, while traveling, or even at home. Some futurists have predicted that the availability of low-cost computing power and telecommunications will increase the number of Americans working at home. The dawn of the Parts of this chapter draw on a contractor report prepared for OTA by Kathleen E. Christensen, New York University, Impacts of Hmne-Based H-m-k on }1mnen and Their Families, tJanuar~ 1985. information age will find millions of people telecommuting from their electronic cottages, that is, using computers and telecommunications to do office work in their homes. z At the present, there are only a few thousand Americans for whom working at home is a fulltime substitute for working in the office, but the number is growing and many more might be so employed by the mid-1990s. -The word telecommuting was probably coined by Jack Nines, now at the University of Southern California, in the 19s and reflected the interest at that time in working at home as a means of conserving automobile fuel and reducing urban air pollution. The term is not full~. appropriate for those who are hired (or contracted) specificzdl~ for home-based work and might not otherwise commute to an office to do the same work. In any case, o telecommuting is an awkward back construction since it means commuting from afar rather than working at a distance. Other terms often used are homework, remote work, telework, or flexiplace. Blue Cross calls its home-based workers, cottage keyers, or telenauts. The home as a computerized workplace is sometimes called an electronic cottage or an electronic sweatshop. THE CONTROVERSY OVER HOME-BASED OFFICE WORK The term home-based office work describes three very different phenomena. The first is the practice of working occasionally or sporadically at home instead of at the office, when it suits the purposes of the worker. The primary work site is still the office, and the worker continues to be a fully participating member of the office staff while enjoying greater control over when and where, and under what conditions the work is done. The occasional home-based worker is most often a professional. This kind of home-based work is not controversial. In other cases the residence is the primary work site. Some of these full-time home workers are entrepreneurspeople who have founded small businesses with headquarters in their homes. They may intend to move the business to separate quarters as soon as it can generate enough income to cover the overhead, or they may choose to keep their business at home no matter how successful it becomes. In any case, these owner-managers seek work from a number of clients and set their prices based on their perception of the value of their work and their competitive environment. These home-based businesses provide a variety of professional and clerical services from word processing to accounting to computer programming. This kind of home-based work is also, for the most part, noncontroversial. s Finally, the residence can also be the principal work site for workers who are employed by a single organization, but seldom or never work in the central office. As employees, their Christensen found that women who developed word processing businesses in their home, seeking multiple clients rather than tied to one corporate client, thought of themselves as, and functioned as, professionals and businesswomen, in spite of the fact that the work they were doing would be called clerical work if done in an office, and the women had in fact previously been clerical workers. Christensen, op. cit. 189
PAGE 197
190 l Automation of America Offices wages are usually set by the employer. Most home workers in this category are clerical workers, usually performing data entry or word processing; in some cases however, the home may be the primary site for professional work such as computer programming. Most of the workers in this last category are women. They may decide to work at home because they must combine work with other responsibilities such as care of young children or elderly relatives. Some other full-time home workers have physical disabilities or live far from commercial centers. Those who are enthusiasts for home-based work usually discuss it in terms of the first two images, the privileged worker and the entrepreneur. Those who oppose it are likely to) speak mostly of the last, the woman struggling to juggle two or more full-time responsibilities. There are three important variables here, which relate to controversial positions on home-based office work. One is the extent to which the home replaces a separate office as the primary work site. Another is the degree to which the home-based worker functions as a separate unit providing services to the organization, rather than as an integral and participating member of the organization. The third variable is the degree of choice exercised by the worker, either in choosing a job that is home-based rather than office-based, or in allocating his or her time between two locations. The controversy about home-based work can only be understood in the context of other social issues, including the long-range outlook for employment, the feminization of poverty, labor-management relations, protection for workers, the adequacy of child care sytems, and opportunities for the handicapped and the elderly. If we had full employment and ample social services, home-based work would probably not be controversial at all. The concern hinges on the question of whether, now or in the future, some workers are forced to work at home under undesirable conditions because the lack of certain social services deprives them of other options. Home-based workers themselves are not divided over the issue; they would be unwilling to give it up. Many have demanded the privilege or worked hard to persuade employers to grant it. Some have accepted the risks that go with free-lance employment in order to work at home. Some have made the basic decision to stay at home while their children are young, and regard the opportunity to do paid work at home as a pleasant bonus, the icing on their cake. Others would much prefer to work away from home but have not been able to find any satisfactory alternative. Their choice is to work at home, or to settle for child care arrangements that they regard as unacceptable, or not work at all and become dependent on others. Congressman Newt Gingrich of Georgia introduced a bill in the 97th and 98th Congresses (H.R. 2531, The Family Opportunity Act) that would have allowed a tax credit of 50 percent of the cost of computers bought primarily for nonrecreational use in the home. One purpose of the bill, which has not been reintroduced in this session, was to encourage computer-mediated work at home. 4 In contrast, the AFL-CIO has called for a ban on computer-mediated home-based work. The labor organization says that it is not possible to enforce health, wage and hour, and other worker protection measures in homes without invasion of workers privacy. The unions are concerned that many home workers are paid under piece-rate systems and are deprived of benefits packages. They also claim that overhead costs are shifted from employer to employee, and that the threat of sending H. R. 2531 would amend the Internal Revenue code to allow an income tax credit for 50 percent of the expense of computers designed primarily for educational, professional, or other nonrecreational use in the home. It would limit the amount of such credit for a taxable year to $100 multiplied by the number of qualified members of the taxpayers family. The bill was referred to the House Committee on Ways and Means and was not reported out. Computers used for work in the home are tax deductible; the bill called for a tax credit.
PAGE 198
work out to be done at home could be used to discourage office workers from demanding rights and benefits or from joining unions. 5 The Service Employees International Union (SEIU) also supports the call for a ban by refusing to enter into collective bargaining with employers who use home workers, but is not completely opposed to home-based work if ways can be found to regulate it. The implementation and enforcement of labor standards in home-based work is probably the central policy issue in this discussion. Those who call for a ban on home-based clerical work argue that such enforcement would be excessively costly or impossible, and that without it women (especially mothers), the elderly, new immigrants, and other disadvantaged minorities may be exploited. Those who favor home-based work argue that protective regulation should not become a reason to deprive workers of benefits that are often sought after, such as the privilege of choosing where they will do their work. Historical Roots Americans have a long tradition of working in the home to earn family income. Cottage industry did not entirely disappear with the development of factories. It is thought that about 10 million to 11 million Americans earn part of their income by working at home at a wide variety of craft, production, and service occupations. G This includes some farmers (although their number has been steadily decreasing) and many others who are only earning a little supplemental money at home, are part of the underground economy, or for other reasons are not counted in the labor force. The 1980 census counted about 1.2 million people in the labor force whose primary place of work was in their residence. 7 children. They are chiefly unskilled or semiskillt,d (and I recruited largely in tenement neighborhoods, often from recent immigrants or other groups with little or no indus[ rid experience. Emily C. Brown, Zndustria) Home Ilrork {Mashington, DC: U.S. Department of I,abor, Jfomens Bureau, 1930). Regulations, pt. 530: Employment of Home J4-orkers in Certain Industries. Title 29, pt. 530 of the Code of Federal Regulations. U.S. Department of I,abor, Emplo~ment Standards Administration, J$age and Hour Di\ision. White Iiouse Publication 1026. Revised h!arch 1980. See also, S. 2145, a bill to amend the Fair I.abor Standards Act of 1938 to facilitate industrial homework. and H. R. 6103, a bill to amend the Fair Labor Standards Act of 1938 to pro~ide that an emplo}er who \iolates sees, 6 or 7 of that act shall be liable to the employee in~ol~ed These bills represent opposing sides of the issue of relaxing homework restrictions. A bill to amend the Fair I.abor Standards Act of 1938, S. 665, was introduced by Senator Orrin G. Hatch of Utah in Nlarch 1985 ~ith eight cosponsors.
PAGE 199
192 Automation of Americas Offices workers. These are usually women housebound with small children, or retired people. How much work is done in this way is unknown, much of it perhaps being part of the underground economy (that is, not reported to IRS), but it has provoked little comment and aroused little or no controversy. The tasks that can be done in this way are limited and poorly paid, and the number of workers willing to do such work is also limited. Homebased work has not been a viable option for most employers or employees, nor a threat to the mainstream clerical work force. Two other categories of home-based office work have always been relatively common. The first includes the clerical and management aspects of home-based small businesses that most often involve crafts, personal services, and professional practices. In most communities, there are many such home-based commercial activities, even where they are officially forbidden by zoning laws and other local legislation. Another kind of home-based work is overflow, work brought home to be done outside of regular office hours. An AT&T survey in 1982 concluded that in 16 percent of households there is at least one person who frequently brings work homepossibly as many as 30 percent of all employed people. Few people doing office work at home, in any of these categories, have until recentl:y been using sophisticated technology. The AT&T survey found that most people bringRobert E. Kraut (13ell Communications Research), Telework: Cautious Pessimism, a presentation at the National Research Councils National Executive Forum: Office Workst[itions in the Home, Washington, DC, Nov. 10, 1983, p. 3 (manuscript, no date). ing work home with them used only the telephone, pencil and paper, and perhaps a calculator. These, plus typewriters, are more than likely the most frequently used equipment for cottage industry offices as well. Computers and telecommunication increase the viability of home-based work and make it possible for a significant portion of all office work to be done at home. The information that is to be processed or generated, the instructions for handling it, supervision and monitoring of the work, interaction between coworkers, and distribution of the output can now or at some future stage of technological development, all be done at a distance. In addition, American households are becoming equipped with computers that could be used for paid employment. By 1990, at least onethird of households may have a PC and some projections are much higher. z Much of the interest in home-based work using computers arose, however, years before the technology was ready to allow it. In the early 1970s, the need for conservation of gasoline and the problem of growing air pollution and congestion evoked much talk about the potential benefits of decentralization of work. If or when these problems again become high national priorities, they will surely act as a powerful stimulus for interest in home-based office work. About 15.8 percent of American households have a computer in 1985, according to an estimate supplied to OTA by Future Computing, Inc. (a division of McGraw Hill). This is based on a total installed base of 15.4 million in 1985, expected to rise to 38.8 million in 1990. They estimate that by 1990, 32.9 percent of households will contain a computer. THE STATUS OF HOME-BASED OFFICE WORK How Many Home-Based Office of Labor Statistics projection. ]3 But many of Workers Are There? these estimates include those who only do casEstimates of the number of home-based of i Homebodies, Forbes, September-October 1984, p. 10. Joanne Tangorra, in an article, Telecommuting, Working fice workers using electronic equipment at Woman 7:1 1:52-54; quoted in Telecommuting, Advanced Syspresent range from 10,000 to 30,000. The most terns Laboratory, Wang Laboratories, Inc., educational brief, frequently used estimate is 15,000, the Bureau 1984, p. 6, says that the number of full-time home-based, computer-based workers is only 1,000.
PAGE 200
ual or overflow work. 14 Probably there are only between 3,000 and 5,000 people who are doing office work, using microelectronic equipment, in their homes for outside employers or clients, and most of these began only in the last 4 years. While there was much discussion of telecommuting during the oil crisis there was little or nothing in the literature discussing real experience or examples. l5 A survey in 1970 identified few home-based workers using computers. But some corporations had already begun formal home-based work programs in the United States and in other countries. There are now at least 40 such programs in the United States. That could amount to only a few hundred people. But there may be at least 3,000 home-based workers in zOO companies that less formally allow employees to choose to work at home. 17 This does not include many people who work at home as contractors or free-lance workers. How Many May There Be in the Future? Some enthusiasts, such as Alvin Tofiler say that there could be as many as 15 million homebased white-collar workers by 1990. This is loosely based on the estimate that 15 million of today jobs could be moved to remote locations. 19 Jack M. Nines, University of Southern California, An O\erview of Office Workstations in the Home, presented at the National Executive Forum: Office l?orkstations in the Home, National Academy of Sciences, J$ashington, DC, Nov. 10, 1983. Ibid., pp. 1-2. Joanne H. Pratt, Home Teleworking: A Study of Its Pioneers, Technological Forecasting and Social Change 25, 1-4, 1984, p. 1. -Patrick Honan, Telecommuting, Will It J$ork For You? Computer Decisions, June 15, 1984, p. 89. Quoted in 13reathnach, op. cit., and elsewhere. An estimate by economist E;lisabeth Allison of Data Resources, Inc., quoted by Judith Gregory, op. cit., p. 3, and Business 1$eek, If Home Is k$here the Norker Is, May 3, 1982, p, 66, Ch. 7Home-Based Automated Office Work l 19 3 Others project 5 million to 18 million homebased workers (up to 18 percent of the whitecollar work force) using computers by 1990. 20 One estimate is that 50 percent of all whitecollar workers, or 26 million people, could do their work at home using computers. 21 There are similar estimates in other industrialized nations. For example, one organization has forecast that by 1990, 40 percent of people in Britain could be working at home. 22 All of these highly unlikely estimates (some of them even presented as projections, or at least possibilities for the future) assume, explicitly or implicitly, that some driving trends will continue or that some existing constraints will be eased. Factors most often mentioned as encouraging the growth of home-based work are lower equipment costs, the growth of electronic mail systems, improved technology for linking the home and central office, rising energy costs, stable or decreasing communication costs, and renewed inflation (which tends to increase the demand for part-time workers or an externalized work force). Some long-range social and demographic trends seem to make it likely that more people will be willing or eager to work in their homes in the future, or will see it as their only way to earn a necessary income. Married women increasingly want an independent income even when they choose the traditional role of housewife; in the absence of alternative provisions for child care they may be forced to combine work with family duties. The same choice or lack of choice may face people who must care for elderly relatives. Large numbers of See for example, Lad Kuzela, Office Old-Fashioned? lrJdustry Week, Oct. 19, 1984, p. 71; and Sally Jacobs, Working at Home Electronically, New England Business, May 21, 1984, p. 15, for summaries of recent forecasts and projections. In the Kuzela article, Dr. Jack Nines of the University of Southern California is quoted as saying that under different sets of conditions the number of telecommuters in California, by the year 2000, could range from 300,000 to 8 million. R.C. Harkness, Technology Assessment of Telecommunications-Transportation Interactions, Stanford Research I nstitute, Menlo Park, CA, 1977. -Business Equipment Trends 1983/1 984, compiled by Kern/Ferry International for Beta Exhibitions I,td., quoted in B. C. 13urrows, Information Technology-Its Impacts on Property Development, I.ong Range Planning, vol. 17, No, 4, August 1984.
PAGE 201
194 l Automation 0f America offices retired workers want or need a way to supplement retirement income on a part-time basis. Handicapped people seek the self-reliance that comes from an earned income. Many people simply like the idea of more flexibility in the use of their time, or wish to live in rural areas but are reluctant to spend large parts of the day commuting. However, many observers are skeptical about the likelihood that home-based work will expand greatly .23 The conventional office has proven to be a remarkably useful and stable institutional structure. 24 It has four valuable characteristics that cannot be matched by home-based work: the presence and cooperation of coworkers, its role in socializing the worker to the corporation and its unique culture, the prevalence of informal communication networks, and mechanisms for structuring the allocation of time. The office is also a major focus of social and recreational activities for many workers. The central office provides economies of scale in capital equipment acquisition, communication and cooperation of coworkers, access to central files and reference material, and the supporting superstructure of superior/subordinate relationships. 25 Some of these benefits are lost or attenuated with dispersion of the work to other locations. If home-based work threatens to disrupt established corporate culture, employers are likely to choose instead to increase productivity by further automation within the office. 26 .. -. -Margrethe Olson has surveyed corporations experimenting with work-at-home programs in 1983-84 and concludes that . while there is continuing interest in the prospect, there are no significant trends toward shifting large numbers into their homes either part or full time. (Margrethe H. Olson, New York University, Working at Home and Telematics: Myths and Realities, a presentation at the Office Automation Conference, Convention Center, Los Angeles, Feb. 20-22, 1984.) -Robert E. Kraut, Bell Communications Research, Telework: Cautious Pessimism, a presentation at the National Research Councils National Executive Forum: Office Workstations in the Home, Washington, DC, Nov. 10, 1983, pp. 17-18. -Steven S. Kawakami, Institute of Labor and Industrial Relations, University of Illinois, Electronic Homework: Problems and Prospects From a Human Resources Perspective, September 1983, p. 14, Kraut, op. cit., p. 19. There are other factors that may well retard the spread of home-based office work. Managers and supervisors sometimes oppose it, because it calls for entirely new techniques of supervision, instruction, and quality control. Home-based workers frequently report that they are resented by coworkers who do not have that privilege. 27 Many work-at-home arrangements depend heavily on telephone lines; if there is a significant increase in local telephone rates, home-based work may be less attractive to both employers and employees. s (In some cases, telephone costs are borne by the home worker.) Finally, it is generally assumed that further development of information and communication technologies will tend to encourage homebased work. It is also possible however that technological development may make some of it superfluous. If the use of optical scanning devices eliminates the need for much of todays mass data entry, a large portion of the work now done at home may be eliminated. Mechanisms for Home-Based Work Work at home can be supported either by an organization or by independent, free-lance activity. The worker, in other words, may be an employee, or may be self-employed. A corporation may have a formal program under which selected employees are offered the option, or employees may request home-based work (often under the condition that a supervisor also has the option of refusing to agree to it). Other corporations informally allow individual workers to negotiate the privilege of working at home, either full time or more often part time. Most employed home-based workers have previously worked for the same employer onsite. But some corporations have set up programs to hire workers not otherwise available to them or able to work, e.g., handicapped workers, mothers with small children, or suburban housewives. --Pratt, op. cit., p. 7; Jacobs, op. cit., p. 19. Honan, op. cit., p. 96.
PAGE 202
Ch. 7Home-Based Automated Office Work 195 Independent or self-employed home-based workers may work under contract for one or more organizations, or solicit piece work on an ad hoc basis. Often corporations shift employees to self-employed contractor status when they become home-based workers. However, if these workers contract with only one organization, they may in fact be employees in the eyes of the law, depending on how pay rates are set, and other details of the arrangement 29 Professionals are often operating as consultants, one-person firms, or part of very small businesses. Those who are employees, however, typically retain salary and benefits, whereas clerA recent IRS ruling indicates that they are employees, at least for some purposes. See p. 38 below. There have also been cases under the National I.abor Relations Board : rider which newspaper delivery truck driters, for example, were held to he employees rather than independent contractors as a company, had claimed, according to Dennis Chamot, Associate IJirector of the A FI.-C 10 Department for Professional l~mplo~ees (personal communication, ,Jan. 8, 1985). A portable terminal can be connected via telephone lines to business computer systems ical home-based workers are usually shifted to part-time status or independent contractor status, and do not retain employee benefits. Most clerical workers who are first hired for home-based work are not given employee benefits. 30 Whether workers are full or part time, they may not do their work in the traditional office hours of 9 to 5. The ability to control ones own work hours is often cited as a major benefit of working at home. 31 But the choice of work hours in practice is usually constrained. The workers sometimes find that they must work during the business hours of employers or clients in order to ask questions or receive instructions. More often, they must fit their work around family responsibilities, working during school hours, when another adult is at home to care for children, or when the children are asleep. It is common for them to work at night, often after a full day of child care, cooking, cleaning, and shopping. 32 Some workers use dumb terminals connected to an employers mainframe computer, usually by a ordinary telephone line and modem. An additional telephone line is often installed T!argrethe Olson studied work-at-home programs in 14 corporations in 1983. I n eight programs the workers were permanent full-time employees (salaries plus benefits). All of t,hese programs except one involved professional workers. The other six programs were for clerical workers. I n four programs, the~ were permanent part-time workers earning hourl? wages. one of these companies paid the workers no benefits, three prorated benefits. One program paid the part-time worker-s bj output (piece rate) with no benefits. The sixth program used contract workers. paid hourly rates, but with no separate benefits package. Margrethe H. Olson, (her~iew of 14ork-at-Home Trends in the United States, New York University, Graduate School of Business Administration, Center for Research on 1 nfornlation Systems (New York: New York Uni\rersit~, August 1983). p. 9. The literature suggests that the disparity in arrangements for professional compared to clerical Workers shown in olsons data is probably typical of corporate pro~rams. Olson found that about half of the workers in her stud~r did, howeter, approximate a 9-to-5 schedule. The others tended to work early mornings or late evenings, but on a fairll regular schedule. These were most often professionals, working at home under informal arrangements rather than formal programs. Those in formal programs. i.e., clerical workers, who worked odd hours usually did so because of the need to work around fanlill constraints. Olson, op. cit., p. 22. Kathleen l+;. Christensen, Impacts of IIome-Based \tork on J$omen and Their Families, contractor report for OTA, Januar~ 1985.
PAGE 203
in the home for this purpose. Sometimes a dedi cated leased line is necessary for security. Other workers use personal computers or standalone word processors, dictating machines and printers. 33 There have been problems with equipment in some programs, but by most accounts technological problems have been minimal. 34 The equipment may be owned and installed by the employer, and either lent or rented to the worker. The worker may own or lease her own equipment .35 Blue Cross clerical workers in North Carolina, for example, pay their employer $2,400 yearly rental for use of their equipment. 36 On the other hand, an employer in the Netherlands pays home workers an extra 2,000 guilders per month to cover the costs of operating a terminal at home. 37 What Work Can Be Done at Home? One of the most likely kinds of office work to be done at home is professional work that involves only one persons creative activity (programming, writing reports) or contacts usually carried out by telephone (sales, brokering). A second kind is clerical activities that are unitized, repetitive, and routine. 38 Data enIn Pratts recent study of 59 home-based office workers, 57 percent of the equipment was employer-owned terminals communicating with mainframe computers in the office; 31 percent was stand-alone computers; and 12 percent was standalone word processors usually not equipped for communication. Pratt, op. cit., p. 5. However, according to Elizabeth Carlson, second vice president of personnel, as quoted by Honan, op. cit., p. 96, Continental Illinois curtailed its program in 1980 and again in 1981, waiting for vendors to come out with usable equipment. In most of the formal corporate programs that Olson studied, the employer paid for installation of equipment and monthly telephone charges; in one program employees rented equipment from the company. Olson, op. cit., August 1983, p. 10. If Home Is Where the Worker 1s, Business Week, May 3, 1982, p. 66. Richard J. Long, The Application of Microelectronics to the Office: Organizational and Human Implications, Nigel Percy (cd.), The Arrnagement Implications of New Information Technology (London: Croom Helm, 1984), p. 106. In one study of corporate pilot programs, it was found that they had targeted either clerical workers, with the objective of cutting overhead costs, or professional workers, in the interest of retaining valued workers who demanded the privilege of working at home. These were distinct objectives and quite different kinds of pilot programs; no corporations had both types. Olson, op. cit., 1983, p. 7. try and word processing can be monitored and measured electronically, computer checked for errors, and paid as piecework. Raw data and finished work can be physically transported between office and home in batches, or can be sent by telephone. Whether professional or clerical, this work usually does not require much face-to-face supervision or collaborative effort between coworkers in real time. Supervisors are least likely to be able to work at home and there are now only a few firms where they do so. This may change in the future with the spread of electronic mail, computer conferencing, PBX, and other electronic tools that make possible cooperative documenthandling and reduce the need for frequent interactions with coworkers. 39 However, the extent to which supervisors will be inclined and willing to replace face-to-face interaction and supervision with electronic communication remains to be seen. Financial and computer service corporations have been most likely to experiment with workat-home programs. Metropolitan Life, Control Data Corporation, National Bank of Chicago, Continental Illinois, Southern New England Telephone, Seybold, and Aetna Life and Casualty Company have started such programs, although some of these have been terminated or suspended. 40 Small new companies have also experimented with the scheme. Few or none of the programs have been found in old, traditional industries. There is strong interest in work at home in other countries, but the movement has been strongest in the United States. In England, early enthusiastic projections of electronic commuting were not met, largely because of the high costs of telephone lines. 41 In France, there have been a number of experiments with working long distance from small neighborhood centers rather than homes. The national telecommunications authority, which sponsored these programs, made an explicit policy Olson, op. cit., August 1983, p. 7. 4Honan, op. cit.; Jacobs, op. cit.; Wang Laboratories, op. cit. Ursula Huws, The New Home Workers (London: I.ow Pay Unit, pamphlet No. 28, 1984), pp. 14-15.
PAGE 204
decision that home-based work would be too isolating. These experiments have generally not been successful and most of the workers dropped out. However some pilot programs for handicapped workers are now underway. Sweden has also experimented with neighborhood work centers, and there is reported to be much interest, but little experience, with working at home. 43 44 One multinational software company based in Britain, with subsidiary offices in Denmark, Ch. 7Home-Based Automated Office Work 197 Holland, and the United States, is deliberately structured so that nearly all of its workers are based at home. These are computer professionals, and in Europe, 96 percent of them are women with small children, and 95 percent work entirely at home. In the United States, 70 percent of the companys employees are women, and 50 percent are fully home-based. The workers agree at the outset, they will steadily increase both their working hours and commitment to the company and career, as their children mature. 45 lt)id., p. 19. Ibid. lklonica I+jlling, Swedish Center for Working Life, Stockholm, Remote J1ork Telecommuting, presentation at IF IPConference on }fomen, 11ork, and Computerization, Ri\a del Sole, ltal~, Sept. 17-21, 1984. Huws, op. cit.: and Nlarsha Johnston Fisher, Firm Turns Telecommuting Into a Realit~, AlIS, No\. 28, 1984. PARTIES AT INTEREST Employers Employers who offer home-based work have at least one of four possible objectives: response to employee demands, social responsibility, access to an untapped labor pool, or cost cutting. The last is apt to be the critical factor for the future spread of home-based work. Some corporations have begun home-based work programs, or allowed the option in individual cases, to hold on to particularly valued workers. Some allow workers on retirement to continue part-time work from their homes. Others such as Control Data Corporation and Metropolitan Life began programs in order to offer employment to handicapped workers. 46 A few corporations are reported to be turning to home-based work as a way of recruiting otherwise unavailable but highly qualified workers, such as mothers of small children, or suburban housewives. 47 Critics say that they are trying to avoid hiring poorly educated urban minority workers. Ilonan, op. cit., pp. 88-91. Pratt, op. cit. The strongest motivation for offering homebased work in the future is, however, likely to be the possibility of reducing costs. The employer saves money in terms of: l floor space and associated overhead costs, l equipment costs (in some situations), direct labor costs, and workers benefits. The office needs proportionately smaller facilities, with all of the reduced operating costs that entails, if many of the workers are at home. There workers pay for their own floor space, heating, cooling, and amenities; in some cases, they own and maintain the computer and other equipment. In effect, these costs are shifted to the workers. If the workers use terminals to communicate with the employers central computer, the employer benefits by the more intensive use of the computer because workers often use it, and can even be required to use it, outside of peak hours. For example, in a pilot program run by the Army Materiel Development and Readiness Command, there was a 64 percent increase in computer usage without additional cost. Three homebased employees were shifted to second and third shifts. The percentage of work time that the remote employees spent
PAGE 205
198 l Automation of America Offices on-line also increased by an average of 93 percent. 48 Home-based clerical employees are often paid less than their peers in the office, although this is not always the case. 49 If the workers are paid piece-rates, the employer is not paying for set-up time, time spent in collecting or delivering work, coffee breaks, bathroom breaks, discussions with supervisors, chats with coworkers, or any of the other unproductive time that office work includes. The organization can define the work by task and pay only when it is satisfactorily completed. The employers biggest cost saving may come from flexibility-workers can be used when needed and not used when work is slow, without the difficulties of firing and rehiring or the expense of unemployment compensation. This-is a great advantage for a company with a highly uneven workload. It means, however, that the worker does not have a steady income .50 Many homebased workers received no benefitshealth insurance, sick leave, vacations, pensions, etc. Of 14 corporate programs studied in 1983, for example, the 6 designed for clerical workers shifted their status from fulltime to permanent part-time workers, in which category workers-did not receive benefits. 5 Those classified as independent contractors are generally not eligible for unemployment insurance benefits when not working. Home based employees and independent contractors Mary T. McDavid, The ALMSA Work-at-IIome Proto type, presented at the National Executive Forum: Office Work. stations in the Home, .National Academy of Sciences, Washington, DC, Nov. 10, 1983. q Continental Illinois, for example, paid at-home workers the same salaries paid to in-office workers, expecting that costreduction would result from reduced need for office space. Honan, op. cit., p. 96. Business Week reported in 1982, for example, that Aetna plans to pay telecommuters by the project and to use them only for peak work, leaving them without a regular salary. They will also be ineligible for health and pension benefits. Blue Cross home-based data enterers are excluded from benefits, which for office-based workers amount to $2,000 to $3,000. The company says that those at home earn up to $3,oOO a year more, but they also pay $2,400 rent for their terminals. If Home Is Where the Worker Is, Business Week, May 2, 1982, p. 66. Olson, op. cit., August 1983, p. 11. studied by OTA for this project did not receive any employee benefits. Entrepreneurs had purchased their own health insurance and retirement plans. 52 It has also been suggested that unionbusting is, or will be, a motivation for some employers. At present, few office workers belong to unions, and home-based workers are likely to be even more difficult to organize than other office workers. The low pay for homebased clerical workers, especially piece-rates, could act over the long run to depress the pay of the main clerical work force. Critics argue that employers might use the threat of expanded homework programs to undermine attempts by office workers to organize. In spite of seemingly significant benefits to the employer, many companies resist the concept of work at home. It involves a significant change in traditional techniques of management and supervision, and managers fear a loss of control over quality, quantity, and pacing of the product. The workers themselves sometimes complain of poor instruction and a lack of feedback from managers. 53 Productivity Almost without exception, studies show that home-based workers are more productive than those in the office. Estimates of the increased productivity of home-based workers range from Control Data Corporations estimate of 15 to 25 percent to other estimates of up to 60 percent 54 and even 80 percent 55 for some work--_ Christensen, op. cit. Most of the home-based workers studied in depth were, except for those who were active entrepreneurs who owned their own business, regarded by their companies as independent contractors. Since they each worked only for one company, in most cases had been regular employees until they took maternity leave, and had been approached on the question of working at home in each case by the company rather than themselves soliciting home-based work, they probably should still be regarded as employees. Pratt, op. cit., p. 6. R.A. Manning, Control Data Corp., Alternative Work Site Programs, a presentation to the National Research Councils National Executive Forum: Office Work Stations in the Home, Washington, DC, November 1983. In a study of about 1,000 home-based workers by Electronic Services Unlimited, reported in Jacobs, op. cit., p. 15.
PAGE 206
Ch. 7Home-Based Automated Office Work l 199 ers. 56 In part, this is an artifact of measurement. With home-based workers, only actual working minutes are counted. In the office, there is a considerable amount of wasted time that is so fragmented that it is not noticed. However, many experts believe that much of this wasted time is in fact spent in informal help and support to other workers, or in the exercise of invisible skills beneficial to the company .57 Home-based workers who were interviewed for OTA said that they did not bill their employers for extra work they did in correcting mistakes. Some did not request payment for, or even report, hours that they put in on a project over and above the time that had been estimated as required, feeling that they should have been able to work fastereven though these estimates of required time had been made by the employer rather than by themselves. This kind of self-exploitation maybe common among those for whom home-based work is new and regarded as a rare privilege, and should be taken into account in assessing the reports of high productivity .5* Workers themselves tend to attribute the increased productivity to their control over their schedule, their ability to work when they most feel like working. In part, the additional productivity probably results from the fact that home-based workers are carefully selected, highly motivated, and working in an environment chosen and designed by themselves. Some observers argue that all of these effects are transitory and will disappear when hlc~at~d, op. cit. Nines concludes that gains of 20 to 50 percent are common. (,Nilles, op. cit., p, 3.) In one group, 6 percent of the workers perceived their own productivity, as well as the qualit,y of their work, to have significantly increased. (Pratt, op. cit., p. 5.) In another study, workers reported their perceived increase in producti~ity to be 35 percent on the average, with the self-reported range from 5 to 100 percent. In;risible skills are important aspec~s of the job that are not specified in job descriptions but make the worker traluable to an organization: for example, a secretar~ chatting hospitably with important \risitors, or a word processor showing coworkers features of the equipment that they have not yet disco~ered, Christensen notes that this response was typical for the emplojed home workers, but that it was not typical for entrepreneurs who had clear ideas about the \alue of their time and priced services to be competiti>e with similar businesses, the novelty of working at home is gone, and when less highly motivated or selected workers are involved. The cross-fertilization and mutual support that occurs in an office working group are valuable. Workers learn from each other on the job. When there is no chance to do this, productivity y may suffer in the long run. The Workers: Why Are They at Home? Because the number of home-based workers is uncertain, their characteristics-age, gender, occupational status, etc.cannot be described quantitatively with confidence. Certain groups are thought to be represented far out of proportion to their number in the total work forcewomen, single parents, the handicapped, and retirees. 59 Professional home-based workers include both men and women; the clerical workers are overwhelmingly women, and the evidence indicates that most have one or more children under 6 years old. 60 This is true in other countries as well as in the United States. The typical home-based white-collar worker, then, is probably not a male professional, but a young mother doing clerical work. 61 Most of the home-based clerical workers are women because: 1) most clerical workers are women, and 2) women are most likely to be responsible for care of children, the elderly, An analysis based on the 1980 census suggests that of people regularly working at home (but not necessaril~ doing office work or using computers), 57 percent were women compared to 43 percent of those working outside the home. (Robert E, Kraut, in on-going analysis, communication of Dec. 20. 1984.) Among those who volunteered to be interviewed b) 0rA contractor Kathleen Christensen there were two groups, roughl~. equal in size. Those who were in effect employees (although called independent contractors they worked only for one organization, usually their employer before they took maternit~ leave) were all women in their thirties, married or with partners, and with small children. Those who were entrepreneurs. with their own companies and seeking multiple clients, were typicall~ older, single, with no small children. An as-~ etunpublished survey of se~eral hundred home-based workers confirms this finding. Nlost home-based women workers with children under 17 are clerical workers, according to a major sur}ey: Kathleen h;. Christensen, National Survey on 14ornen and Home-Based 1$ork FamilJF Circle, Dec. 15, 1984 (publication of results in progress).
PAGE 207
200 Automation of Americas Offices or the ill. 62 Women are probably also most likely to need additional income after retirement, because they live longer than men, and are less likely to have adequate pensions. Most if not all home-based workers are now working at home by choice, in the sense that other jobs are available and no employers are known to require working at home as a condition of employment. In most cases, the demand for home-based work has so far come from the workers, not from employers, who are often reluctant or at least hesitant to provide this option. But this may already be changing, as organizations become aware of the cost-saving possibilities. 63 Choice in any case has a wide range of meanings. In some cases, it means that other options are not available, or are less attractive to a worker for reasons over which she has little controlstaying at home is necessary because of other responsibilities, because of social inhibitions (the traditional housewifes role), or because of physical disability. Often, however, the person has first decided that he she wants to stay at home and care for a family; the opportunity to do paid work at home is a secondary choice that is not a critical factor in the first decision. Other people simply prefer a lifestyle that does not include going to an office, at least not at any preset times. These people who have actively sought home-based work as part of a new lifestyle are more likely to be the professionals (including males) simply because professionals usually can exert more control over the conditions under which they earn a living. Corporations frequently report that they have formally or informally provided the option in order to retain valued employees. While women constitute 80.5 percent of all clerical workers, they are even more dominant in clerical occupations most likely to include home-based work. For example, over 96 percent of typists are women, but less than 23 percent of shipping clerks. Kraut, op. cit., p. 5. The group of 13 independent contractors who worked only for one client included nine former employees of an insurance company, who had been offered work to be done at home when they applied for maternity leave. Four others were housewives who had each been asked to do work at home by friends or husbands who were the proprietors of small companies. Christensen, op. cit. Many of these professionals believe however that home-based work, especially if done full time or nearly full time, seriously prejudices their chance of promotion and advancement. They have consciously traded-off advancement for a preferred lifestyle. 64 Women managers especially tend to believe that this choice is particularly prejudicial to their careers; some have done it in preference to dropping out entirely when they became mothers, and in order to maintain their expertise and credentials until they could return to work. 65 Those who look for home-based work reluctantly, because they must beat home, are more likely to be clerical workers. 66 They are less likely than professionals to be able to afford professional child care or specialized transportation for the handicapped. But even for these people, the opportunity to work at home may be cherished, in contrast to not working, or working under less desirable conditions. It is therefore not surprising that overwhelmingly home-based workers are pleased with their situation. This must be kept in mind when evaluating the implications of home-based office work. Controversy over home-based work does not arise from the dissatisfaction of the current home-based workers themselves, but from the possibility of future exploitation. The Benefits and Costs for Home-Based Workers? How workers assess the relative benefits and costs to them of working at home undoubtedly depends on the degree to which they exercised free choice from a wide range of employment options. All studies, and all journalistic accounts of home-based workers, indicate that most value the opportunity to control their own work hours. Many report that the quality of their leisure time is also improved because they can select and schedule recreational activities that Pratt, op. cit., p. 7. See also Breathnach, Honan, Jacobs, Olson, op. cit. Pratt, op. cit., p. 11. Also substantiated in Christensen, op. cit. For example, see Olson, op. cit., August 1983, p. 17.
PAGE 208
. are not otherwise available. Many of them find the home a congenial and comfortable working environment; this is not always the case for those who are combining work with child care in crowded living quarters. Many value highly, the additional time spent with their families, or the latitude to spend more daytime hours with family. The opportunity to care for children at home rather than have them cared for by others is the primary benefit for many people. It is not always true, however, that working at home allows the worker to control when he or she works. The home-based worker indeed often has a double constraint, having to fit work around family needs and the times when the company computer is down, such as on Sundays. Thus, they may have as little control as office workers over when they work, and they may even lose some of what is usually considered family time. Cost savings can be significant for some who give up work in an office. Workers have reported savings up to $200 per month for meals, $100 for gasoline, parking, and insurance; $100 for clothes and cleaning bills. Large costs of childrens day care may be avoided. 67 Those who are entrepreneurs avoid the cost of outside office space and take tax deductions for the use of their home. Eliminating the commute to work is for many a primary benefit. The average American worker now travels over 9 miles in each direction, often a 1to -hour commute. 68 The time and the stress spent in commuting is often resented even more than the cost of gasoline, insurance, car maintenance, and parking. These cost savings of working at home do not tell the whole story. There are also added costs, both direct and indirect. Some workers must lease or buy terminals. Most pay the telephone bills and some must pay for special dedicated telephone lines. Workers generally provide the furniture that they use in connection with the equipment. They pay higher electri Pratt, op. cit., p. 5. hlotor }ehicles hlanufacturers Association, Facts and Figures, Detroit, hl I, 1981. Ch. 7Home-Based Automated Office Work l 20 1 city bills because of the equipment, and some report additional heating and cooling bills. 69 Clerical home workers almost always earn less than their peers in the office, even within the same firm. Sometimes the pay rate is technically the same, but those at home are shifted to part time or piece rates without fringe benefits. They usually work fewer hours than those in offices, either because work is not regularly available, because of their home-related duties, or by choice. Part-time and piece-rate home-based workers usually report that their workload is very uneven. (For employers, the flexibility of labor supply is a major benefit.) The lack of a regular income can be as much of a problem as is low pay; those who are dependent on income from home-based work may live balanced on the edge of financial disaster. Health insurance coverage becomes a major problem for those converted to the status of independent contractor or part-time worker. The majority of home-based workers may be covered as dependents by the health insurance of spouses, but the separation from employee status leaves others to pay the high costs of individual coverage or be without protection .72 Pratt, op. cit.; and Honan, op. cit., p. 97. See for example, Olson, op. cit., August 1983. Other case studies of home-based clerical work have consistentl.v found that home-based workers make less than peers. However, some preliminary analysis by Robert Kraut at Bell Communications Research, shows higher hourl~ wages for home-based workers in a few clerical categories as compared to those not working at home. Reduced income is typical of the larger category of all work at home as compared to work away from the home. According to the 1980 census, 10 percent of home-based workers are below the poverty line, compared to 6 percent of other workers. Male home-based workers have an average hourly wage of $6.77 compared to $8.20 for nonhome workers; women home-based workers have an hourly wage of $3.06 compared to $4.80 outside the home. In addition, home-based women workers a\erage 30.6 hours work per week compared to 40.2 for those working away from home, 35 for male home-based workers, and 41.6 for other male workers. There may be several variables invol~ed in this phenomena. Pratt (op. cit. ) also reported that many retired people found home-based work not worth their effort because of the low pa~, even though many enjoyed ha~ing some work to occup~. their time. Pratt, op. cit., p. 10. -Among those interviewed b> Christensen, independent contractors working for one client were cotrered under a hus-
PAGE 209
202 Automation of America Offices While home-based workers have more control over their working environment than they would in most offices, in the sense that the decisions are their own, in fact, they often work under very undesirable conditions. It may not be possible to set aside space dedicated only to work. They may not know how to select, or perhaps cannot afford to buy, chairs and desks that accommodate the equipment and reduce muscular strain for the workers. 73 Poor lighting may cause eyestrain and headaches. It is possible that office automation equipment also introduces unrecognized hazards into the home. 74 One intangible cost to the worker may be the social isolation of working at home. Saloman and Saloman point to the importance of social interaction at work, and studies that have shown that small, cohesive work groups are for most workers the most effective and satisfying work unit. For these workers, separation from the office work group may in the long run result in dissatisfaction and low motivation. For managers and professionals, the social role of the workplace may be less important; they rank achievement, advancement, and the work itself as the most important factors in work satisfaction. The Salomans hypothesize that for some of these workers, the shift to working at home may cause the individuals motivators to shift. He or she may become less concerned with achievement, status, and recognition, and more concerned with bands health insurance. Those who were entrepreneurs with, or seeking, multiple clients were mostly not married; they said that the high cost of health insurance (in one case, $1,400 per year) was one of their greatest concerns, and one or two had gone uncovered for long periods of time. Christensen, op. cit. K1e]ecornmuting Review, Dec. 31, 1984, cited a survey of computing furniture needs conducted by the Business Products Consulting Group, which found that only 25 percent of personal computer users, whether in the home or in the office, had specialized computer furniture, and many of them complained of flaws in design that caused discomfort. Overall 7 out of 10 PC users had complaints about the furniture used with the computers. The situation is likely to be worse in homes than in offices, since home furniture is often jerry-rigged to hold office equipment. If fire breaks out, there is a serious hazard to residents and to fire fighters from the highly toxic gases produced when the plastic casing of computers burns. Any special cabling that might be installed in homes could also involve fire hazards, according to discussions with insurance risk assessors. affiliation and emotional contentment within the family and community, which will become more important to the worker than his or her organization. If these needs for belongingnes s cannot be met within the immediate environment, then frustration and dissatisfaction will result. 75 Home-based professionals whose salaries do not depend on the exact number of hours worked have remarked that minor illnesses, for which they would have stayed home from the office, now make them feel guilty-there is no ritual process of calling in to validate their reason for not working. Distractions and interruptions by family members, normal household noise, salesmen and solicitors, friends who do not respect working hours at home and repeatedly call or visit, are sources of annoyance and stress for homebased workers. 76 Some people have reported giving up work at home because they could not resist the temptation to eat or drink too much. Men have reported that wives resented having them under foot all of the time. 77 On the other side, some researchers have reported that both the loners and the gregarious types say that they can overcome these problems with time. Handicapped people report that they feel less isolated with work to do, even if their human contacts do not increase, and young mothers according to one survey said they were no more isolated than they would be if they were caring for children without paid office work to do. It appears to be the young, never married workers who are most likely to give up home-based work after they have tried it because they miss the social aspects of the office. 78 --. .Ilan Saloman and Meera Saloman, Telecommunicating: The Employees Perspective, Technological Forecasting and Social Change 25, 15-28, 1984. The Salomans here are drawing on the work of N.D, Dunnette and others in Work and Nonwork in the Year 2000, Dunnette (cd. ) (Californiiz Brooks/Cole, 1973). Saloman and Saloman, op. cit., pp. 23-24. Nelson B. Phelps, Mountain Bell Telephone Company, a Case Study, a presentation to the National Research Council National Executive Forum: Office Work Stations in the Home, Washington, DC, September 1983. Pratt, op. cit.
PAGE 210
Ch. 7Home-Based Automated Office Work l 203 Again, those who found home-based work unpleasant or stressful, have so far either had the option of returning to the office or have found this alternative even less acceptable. Thus, studies of home-based workers almost always find the workers enthusiastic about this lifestyle. Family Considerations Advocates of home-based work stress the benefits to family life of returning the worker to the home. Little research has been done to determine the real effects on family life. Such indicators as there are point to generally positive effects, but with some caveats .79 Mothers perceive several specific advantages for their children, in addition to the primary advantage of having a parent care for them, instead of strangers or a commercial facility. They believe it is good for their children to realize that mothers can do more than cook and take care of them. They want their children to see women performing a broader social role than that of housewife, however they themselves value that role. Some say that their children become more independent because they are not the only focus of mothers attention. They say also that their children become familiar with computers and what they can do. On the other hand, some mothers report that their children get less time and attention, and that the mother gets impatient when she is under pressure to get work done. For women struggling to earn an income and care for children at the same time, home-based work may be a golden opportunity, but it is not an unalloyed blessing. It involves significant stress, both physical and mental, and may create emotional strains within the family as well. so One expert says, It appears that work Ma~erial in this section, unless otherwise noted, is drawn from Kathleen Christensens study and chiefl~ from the responses of 14 mothers working at home, so that it is merely illustratiie. Iler findings as discussed in this section are, however-, in man~~ cases confirmed by or congruent with the somewhat scant~ and widel~ dispersed observations of other researchers about the family situation of home-based workers. Gregory, op. cit. discussing Cynthia Costello, On the Front: Class, (jender, and Conflict in the Insurance Workplace, Ph. D, dissertation, University of J4risconsin, Dept. of Sociolat home cannot be called a good solution to child care. 81 Mothers working at home typically try to work when other adults are at home to care for the children or when the children are at school or asleep; they are not so much combining work and child care as interweaving them. It is hard to find 40 such hours a week even when work is available and the income is wanted, and typically they work split shifts, often late at night. Those who must earn as much as possible because they and their children are dependent on their income, may therefore have little or no time for rest or recreation, like the cottage sweatshop workers of the turn of the century. Those who regard this income as discretionary often have babysitters while they are working. Since taking care of the children themselves is important to these women, the fact that they turn to baby sitters indicates two things: that doing paid work and taking care of children at the same time is difficult and stressful; and that it is important to them to have some other work, in addition to caring for a family, for reasons other than the marginal income. The second point is repeatedly confirmed by home-based working mothers, who say that they need something to occupy their minds or that paid work gives them pride ogy, 1984. In one study, 50 percent of mothers working at home reported that the-y found it necessary to have a paid bab~sit ter for part of the time in spite of their desire to combine work and child care. Also Margrethe olson, Remote Office l! ork: Implications for lndi~iduals and (~rganizations, CAIS No. 25. GBA No. 81-86 (CR), New York Uni\rersit~, (lraduate School of Business Administration (New York: New York Uni\ersit~), Olson, op. cit., August 1983, p. 27. -The 14 mothers inter~iewed by Kathleen Christensen, nearly all said the hours the~ could work were those in which the children were asleep, in school, being watched by the husband, or playing bv themselves and or in a good mood. These women are in ~radit-ional t we-parent households where the husband is the primary breadwinner. Six of the 14 mothers interviewed by Kathleen Christensen, all of whom were from traditional two parent households and said that the~. stayed at home to care for their children, relied on some form of paid child care while they worked. The mothers, whether or not they had child care, usually ga~e as their reason for working to ha~e something to do, so that m~ mind can keep working, or to show a ~hat I can do something besides cleaning the house and being a house~ife.
PAGE 211
204 l Automation of Americas Offices and self-respect (and respect from others) that is lacking otherwise. 84 In the few cases where effects on the marriage, or other personal relationships, of homebased women workers have been studied, most husbands were said by their working wives to be supportive and helpfulsometimes because they welcomed the supplemental family income, sometimes because they recognized its psychological importance to the worker, sometimes for both reasons. A few husbands were said to have mixed feelings, perceiving that work sometimes interfered with housework; none were reported to be actively opposed. Most of the women said, however, that their husbands helped no more with housework than they had before. The women had been, and still were, predominantly responsible for the household work, and a few spontaneously expressed dissatisfaction or resentment of this. Most were unable to draw boundaries between household and work responsibilities, and move back and forth between them during the day. Even those home workers without children to care for often report significant stress resulting from the lack of separation between work and family responsibilities. Saloman and Saloman, in a paper on home-based work from the perspective of the sociology of work, point out that: Work and family life today are not only physically separate entities, but each also gives rise to distinctly different role behavior that may have little in common. It is not only that the two different environments require different behavior, but that they also offer the individual a chance to express different aspects of his or her personality .85 Sociologists define a role as a set of structural demands being placed on the individual Ibid., citing D.T. Hall, A Model of Coping With Role Conflict: The Role Behavior of College Educated Women, Administrative Science Qmrterl.v 17 (4), 1972. See also J. Pleck, The Work-Family Role System, Social Problems 24 (48), 1977. in a given social position. 86 Role conflicts can result when the demands and expectations imposed by multiple roles operate at the same time. Working at home, the Salomans argue, can introduce identity conflict as it eliminates the sequential operation of the different roles related to home and work. For men, it has normally been acceptable for work (to) interfere with family life; at home, the reverse interference may happen more often. For women, who may have more social inhibitions to overcome in establishing a career role, the conflict may be particularly severe when the roles of mother, wife, housekeeper, employee, and career aspirant overlap in time and space. These sociologists say that the trip between work and home is often a useful separation between two arenas of social interaction. Thus, eliminating the commute may be a cost as well as a benefit. Effects on Society Workers and employers are part of one society and the interests and concerns of each are part of the public interest. But society at large may have a general interest separate from the specific concerns of either party. In this case, some of the costs that are ordinarily borne by employers or shared by employers and employees are shifted to workers for example, in the case of independent contractors, social security taxes, health insurance, pensions, and the costs associated with periods when the workload falls below normal levels for long periods of time. For people who are not employed, those costs ultimately may be borne by the taxpayers. Thus, a strong trend toward farming out work to independent contractors who were previously, or would otherwise be, employees, also implies a shift of life-cyle costs from the employer to the general public. This was a theme that recurred repeatedly in the interviews conducted by Christensen. Ilan Saloman and Meira Saloman, op. cit., pp. 15-18, 21.
PAGE 212
Ch, 7Home-Based Automated Office Work 205 LEGAL AND REGULATORY BARRIERS TO HOME-BASED WORK Zoning laws in some communities prohibit any paid employment in residential areas. 87 The Internal Revenue Service (IRS) has in recent years progressively restricted the income tax deductions that may be taken for home offices. It is not clear under what conditions homebased workers can claim deductions, especially if the workspace that they use is not entirely dedicated to work (many home-based workers use a corner of the dining room, bedroom, or living room). But when there is dedicated space, a recent IRS ruling says that the home office is to be treated as business property rather than residential property, and thus part of the proceeds of selling a house that includes a home office are not eligible for the standard exclusion from tax on the proceeds allowed to people who immediately buy another house, or who are over 55 when they sell the house. 8s There have until recently been unresolved questions about the status of home-based contractors who contract with only one organization, especially, if they have previously been employees of that organization. The National Labor Relations Board has ruled, in similar situations, that such independent contractors are in fact employees. The tax code says that anyone who performs services for any person as a home worker, performing work according to specifications on materials furnished by the other person and required to return the finished product to that person, is an employee. This holds unless that person has a substantial investment in facilities used for the work, in which case, he or she may be an independent contractor. 89 The Internal Revenue Service has ruled specifically that computers with word processing capability are not substantial investments in that sense, and that home-Tammara H. Wolfgang, Morking at Home: The (jrowth of Cottage Industry, The Futurist, ,June 1984, p, 31. e]ecommuting Re$iew, Dec. 31, 1984, p. 7. Sec. 312 l(d) (3) (C) of the Internal Re\.enue Code. based workers providing transcription services for one person or organization are employees at least for certain tax purposes. go The effects of these rulings on home-based clerical work are not yet fully apparent. It will open the door for home-based clerical workers, if they choose, to argue that they are employees, even if regarded by their client as contractors, and thus, entitled to certain worker benefits and safeguards. Government work-at-home pilot programs have been frustrated or terminated because of regulatory requirements. The pilot program of the Army Material Development and Readiness Command wanted to use either direct lines or telephones with a modem to connect home-based workers to the central computer. The Army Communications Command determined that for the government to provide telephone service to a home violated a Federal statute; 91 but employees could not use their own telephones because another statute placed some limitations on government acceptance of voluntary services from individuals. 93 The problem was solved by installing direct lines without voice capability; a legal opinion sanctioned occasional use of the employees home phone to communicate with a supervisor. This introduced a second problem: how to protect the government if personal property was damaged as a result of installing the communication lines or the use of governmentowned equipment. This was solved by requiring employees to sign a hold harmless agreement with the government as a condition for participation. -Internal Re\enue Ser\rice, Technical .4d\isor? hl[~morandum 8451004, Aug. 1, 1984. Index nos. 312 1.04 -00, 3306,0500, 3401.04-00. Stat. 32 U.S.C. 1348. Rev. Stat. 3679, 31 U.S.C. 1342. 4McI)avid, op. cit. That is, an agreement not to hold the go\wrnment liable for accidental damages,
PAGE 213
206 l Automation of Americas Offices A third problem was the question of responsibility y for injury to the employee while working at home. Workers Compensation (U. S. C., title 5, sec. 8102) covers Federal employees on duty, but the question was how to prove that they were on duty when the accident occurred. The solution to this problem was a written work agreement stating hours to be worked each day; participants had to formally request changes to their designated work schedule and get a supervisors approval. This of course eliminated one of the major benefits that workers typically see in working at home. Finally, this program ended when government auditors ruled that there was a risk of fraud or abuse in spite of electronic monitoring of work done at home. Some of these problems also appear in the private sector, for example, the problems with the installation of communication lines and questions related to the protection of home workers from work-related injury or illness. Installing an additional telephone line in a rented residence, for example, sometimes requires the permission of the owner. One study of home-based corporate employees found that half of them had no accident insurance, and most assumed that they were covered by Workers Compensation. 95 It appears to be true that home workers are covered by Workers Compensation if injured while working at home, and by the employers insurance plan if injured at the same location while not working-if the worker is covered by employee benefits. 96 In either case, how the worker is to demonstrate whether or not he or she was actually working when injured, is so far unanswered. . Pratt, op. cit., p. 8. Honan, op. cit., p. 91. The more general question as to how Occupational Safety and Health rules apply to homebased workers is also unanswered. Laws pertaining to the use of VDTS in the office have been proposed in 13 States but it is not known how these will apply to home-based employees. A British document illustrates that these questions arise in other countries. Draft regulations proposed by the U.K. Health and Safety Commission in 1979 says that: Home workers, properly speaking, work for the person who puts out work to them in the sense that they contribute to products which he markets. For this reason, those who put out work to home workers bear the prime responsibility for ensuring that, so far as is reasonably practical, no risks to health and safety arise. 98 In the next paragraph however, clerical workers are expressly excluded: The Commission, particularly in the absence of evidence of risk to home workers from these processes propose to exclude from these regulations all office type work undertaken in domestic premises. The document continues: It is important to reemphasize that any person requiring a home worker to utilize potentially hazardous processes in connection with clerical work is nonetheless bound by the requirement to ensure that risk is controlled. No further reasoning supporting the exclusion is given, nor are potentially hazardous processes in connection with clerical work specified or defined. Gregory, op. cit., p. 2. Health and Safety Commission, Home Regulations, consultative document (issued sion in compliance with its duty to consult of the Health and Safety at Work, etc., Act, Workers: Draft by the Commisunder sec. 50(3) 1974, 1979, p. 1. PUBLIC POLICY ISSUES Encouragement or Prohibition of Home-Based Office Work actively discourage or forbid it, or 3) take neither action. The primary public policy issue in homeParties at interest include: employers, homebased office work is whether the Federal Gov.. based workers, their families, other workers ernment should: 1) actively encourage it, 2) (especially clerical workers and working women
PAGE 214
with children), and unions. The issue also tends to engage people with an ideological position regarding home and family values, equity for women, and alternative life styles. Society in general has an interest, in that some potential costs may be shifted to the taxpayer when home-based workers are converted to the status of independent contractors. Congress, and State governments, could take a number of steps to further encourage home-based work, although these are probably necessary. These include: revising IRS rulings on independent contractors and relieving employers of some tax liabilities for them, for example, rewriting social security tax provisions; resolving the problems of applying workers compensation to work performed in the home; persuading States to grant exemption from zoning and building codes for computer-mediated employment in residential buildings; providing significant tax incentives for equipment purchased for home-based work; i.e., a tax credit rather than deductions; and clarifying and expanding tax deductions allowed on home office;, especially providing large deductions when all or a significant portion of family income is earned in the home (this would benefit some home-owning workers). If public policy is to discourage home-based office workor more narrowly, home-based clerical workthe clearest option is to prohibit it, as was done for some other occupations in the 1930s. However, this would require very careful definition to limit the prohibition to those kinds of office work that are subject to exploitation. The prohibition would very likely be seen, even by some of those it was designed to protect, as discrimination against women (who would be mostly affected) or against certain occupational and income groups. There are a few ways to discourage the spread of home-based work that are largely the converse of options for encouraging it: l l l Ch. 7Home-Based Automated Office Work l 207 strong enforcement of the rules requiring organizations to treat independent contractors in some situations as employees, with full benefits; placing a special tax on income derived from services provided from homes, for both the employer and the employee; and high telephone rates for digitalized information transmitted between office and remote workers, which would run counter to present policies of deregulation. Other government actions are possible that would both: 1) tend to discourage the spread of home-based clerical work, and 2) enhance the element of free choice for home-based work, thereby reducing the possibility that it will be used exploitatively. These actions are: l developing subsidized, high-quality day care centers for children of working parents; l providing larger tax deductions for expenses related to child care (or the care of elderly or infirm dependents); and l requiring or providing further significant incentives for modifying transportation systems and office environments to facilitate employment of the handicapped, and providing positive inducements for employment of the handicapped. Congress may choose not to take actions either to encourage or discourage home-based work. There are now few important legal or regulatory barriers to its growth; therefore the option of no-action will allow its spread. Eventually, home-based work is likely to grow since it offers benefits to both employers and many individual employees, while the costs that it imposes on some workers are generally considered acceptable in the absence of more desirable alternatives. Controversy about home-based work is likely to become a major policy issue only if and when one of three conditions obtain: the number of home-based clerical workers becomes a significant fraction of all clerical workers, so that this becomes a factor in the competitive position of office workers in the job market and in negotiations with employers;
PAGE 215
208 l Automation of Americas Offices l l cases of serious exploitation of homebased workers come to public notice through the media; and/or unions are highly successful in their attempt to organize white-collar workers Regulation of Home-Based Office Work If, however, the Federal Government neither actively promotes nor prohibits home based office work, then issues arise regarding its regulation to provide protections that are assured to other workers. In summary, these include wage and hour guarantees, assurance of safe and healthy working conditions, the right to negotiate collectively with employers, guarantee of equal opportunity, equitable pay, and equitable access to insurance, pensions, and other entitlements. At present, managers and professionals are generally exempt from the Fair Labor Standards Act of 1938 (which covers such conditions of work as wages and hours) and managers, but not professionals, are exempt from the National Labor Relations Act (right to collective bargaining), whether they work in an office or at home. They are assumed to protect themselves individually by negotiation with employers, although some belong to unions or professional organizations that bargain collectively. They are however covered by the Equal Opportunity Act and other recent work-related legislation. The chief concern in regulation of working at home therefore focuses on clerical workers. Predominantly female, nonunionized, and often bearing heavy responsibilities as mothers (increasingly, as single parents), they are particularly vulnerable to exploitation. Another growing concern is the vulnerability of new immigrants, of disadvantaged minorities, and of elderly workers to possible exploitation. It is quite possible that States or local governments may impose regulations on homebased work, especially with regard to health and safety issues. Questions with regard to regulation are: clarification of the application of existing regulations to home-based work (e.g., Workmans Compensation); what additional protection is needed for home-based workers? and what means can be devised for effective implementation and enforcement of regulations related to home-based work? With regard to the last question, there are concerns that any attempts to implement and enforce regulation of home-based work may destroy the benefits for which it is valued (i.e., autonomy over work hours), or may lead to unacceptable violations of the privacy of workers and their families. Thishow existing labor standards can be implemented and enforced for home-based workersis in fact, the critical policy issue most likely to confront the Congress in this area in the immediate future. It is the point on which opponents usually base their argument for an outright ban, since they maintain that real enforcement will be extremely costly, and in practice impossible. This is the problem that led to the ban on home-based work in some industries in the 1930s. However there has been little real examination of the possibilities and difficulties of enforcement today. It can be argued that the same difficulties would arise in enforcing a prohibition. It can also be argued that either a ban or regulation would be easier to enforce today than in the 1930s. Reporting requirements laid on businesses have proliferated, the rights of workers and the benefits they stand to gain by demanding those rights are larger, and people doing computer-mediated work are likely to be far better educated and more sophisticated in understanding to what protection they are entitled. The same technologies that make computer-mediated work at home possible, might be used to make it difficult to hide.
PAGE 216
Chapter 8 Off-Shore Office Work
PAGE 217
Contents Page The Present and Future Status of Off-Shore Offices . . . . . . . How Off-Shore Offices Operate . . . . . . . . . . . Linkage Arrangements . . . . . . . . . . . . . Types of Data Processing . . . . . . . . . . . . Examples of Off-Shore Offices . . . . . . . . . . . Factors in Moving Data Entry Off-Shore . . . . . . . . . Differing Views of Off-Shore Office Work . . . . . . . . . U.S. Off-Shore Companies . . . . . . . . . . . . U.S. Labor Organizations. . . . . . . . . . . . . Foreign Government Officials . . . . . . . . . . . Economic Development Organizations . . . . . . . . . Public Policy Issues . . . . . . . . . . . . . . Regulating or Prohibiting Off-Shore Offices . . . . . . . . Regulation of Data . . . . . . . . . . . . . Encourage or Do Nothing About Off-Shore Office Work . . . . . 211 214 214 215 215 216 223 223 224 225 226 227 227 227 230
PAGE 218
Chapter 8 Off-Shore Office Work 1 While the cost of electronic office equipment declines, the cost of employing Americans to perform routine office tasks is rising both per employee and in the aggregate. Some U.S. firms have established off-shore offices in order to take advantage of lower cost labor. Some of these offices involve manual clerical work such as coupon sorting. However, advances in information and communication technologies make it increasingly attractive to move data-entry operations off-shore. Data entry means converting information from hard or paper form into digital form so it can be stored electronically. This entry phase of data processing is commonly referred to as keypunching keyboarding, or keying. High domestic labor costs could lead to an international division of labor in data processing. Data-entry clerks in Caribbean countries, for example, typically earn weekly wages that range from about $15 (U.S. equivalent) to approximately $60, while their counterparts Ihis chapter draws hea~il~. from an OTA contractor report: (Christopher P. Astriab, .4n .+ls.sessment ofoff-.$hore Office 11ork, prepared for the U.S. Congress. Office of Iechnolog~ .4ssessment, contract No, 583-0630, Feb. 28, 198.5; and from an additional report from Consultant Anne Posthuma. in the United States may earn at least six times the latter figure per week. Even with transportation, communications, and other costs accounted for, total expenses for keypunching data off-shore may still not equal the cost of carrying out the process domestically. z Several data-entry facilities are presently operating in the Caribbean region, and their number is expected to increase. Labor costs are low and the region is easily accessible and close to the United States. Transportation and communication networks are reasonably well developed, literacy rates are high, and there are favorable tax provisions for foreign investors. The Caribbean is only one region, however, in which off-shore offices could proliferate. These operations can be simple to implement. The tools and raw materials are light, relatively inexpensive, and easy to transport. If full advantage is taken of communication technologies, distance and time become almost incidental factors when choosing sites at which to locate off-shore offices. -Fig-ures deri~ed from Caribbean Central American ~\ction data (Washington, DC, 1982). THE PRESENT AND FUTURE STATUS OF OFF-SHORE OFFICES While off-shore offices have existed for the data processing operations in the Caribbean. better part of two decades, the arrival of elecBarbados hosts seven such firms, while Jatronic communication and information techmaica hosts at least three, and St. Christophernologies has set the stage for new and possiNevis and Haiti each host one. In these four bly rapid growth in the near future. The countries, approximately 2,300 workers are number of companies entering data at off-shore directly employed in off-shore offices. The sites are still few, and the impact of the largest such office in the region is a couponphenomenon on foreign and domestic employsorting operation that employs approximatement and economic situations is at present inly 1,200 people in Haiti. significant. At least three off-shore offices in St. VinOff-shore data processing facilities are precent, Haiti, and Grenada closed down in redominantly located in the Caribbean region. cent years. Managerial and transportation Currently, there are at least 12 U.S. firms with problems were the causes; the ability of the 211
PAGE 219
212 Automation of Americas Offices employees to carry out the operations were apparently not in question. 3 In the case of Grenada, which hosted a firm for some 10 years, recent unsettled conditions there were cited as the cause. 4 Other countries including India, Singapore, the Peoples Republic of China, and Ireland also host at least one data-entry firm each. Approximately 27 coupon-sorting operations are located in Mexico, one of the first countries to host off-shore offices. The total number of people employed by these firms is not known. There are generally two types of firms involved in establishing off-shore offices. The first perform only their own clerical work offshore in order to reduce labor costs. The second group are vendors who have established off-shore offices to provide data processing orword processing services to clients in the United States. All signs indicate that off-shore data entry could undergo rapid growth over the next 10 to 15 years. The vendors who provide these services are optimistic about the future of the industry: In 3 to 5 years we expect to have 300 to 500 people (keying) in different locations (in Jamaica). This is a large market, perhaps $50 billion by 1990. 5 The growth industry in Barbados has to be information services and data processing As Barbarians acquire data processing skill, Barbados, with its investment in telecommunications, can develop software and other applications. The industry in Barbados has tremendous opportunity to grow and evolve, especially given. the governments commitment to it. G For now, any softening of data-entry markets is being offset by explosive growth of data-base services. Such growing needs will Donald Marsden, Coopers, and Lybrand, St. Kitts-Nevis. Personal interview, Miami, Dec. 7, 1984. Ibid. Mary Ramond, Jamaica On the Move, Business Week special advertising section, Sept. 17, 1985. (Gary Bechtel, president of Telemar Data Systems, Fairfield, NJ, is quoted.) Mary Ramond, Barbados 1984, Business Week, special advertising section, Apr. 16, 1984. give (data-entry firms) a big market to shoot at 7 In view of dramatic growth in data processing and expanding needs for digital data, these observations may be justified. There is considerable interest in establishing new data processing facilities abroad and some companies already in the business have cited expansion plans. Several companies are now studying the possibility of setting up keying operations in the Caribbean. Barbados, St. Lucia, St. KittsNevis, Dominica, Jamaica, Trinidad-Tobago, and Grenada are all possible sites for further growth. 8 Firms already in the region maybe expected to expand, but the scale of this expansion is uncertain. A firm in Jamaica that presently employs approximately 60 plans to expand to about 500 employees over the next several years. This growth-approximately 700 percentis probably exceptional, however. The whole industry now employs about 500 people in Jamaica. The Jamaican Government encourages this industry, and ,Jamaica Investment Promotion, Ltd.the government development agencyhas reportedly secured contracts for another company that would require it to expand significantly. While the industry has existed m the Caribbean for some time, investors are showing renewed interest since the passage of the Caribbean Basin Initiative Act by the U.S. Congress. 10 While no specific provisions in the act deal with data processing, the act does express the commitment of the U.S. Government to assist in the development of the region. Early signs of success in using satellites to transThe Instant Off-Shore Office, Business Week, Mar. 15, 1982, p. 136. Anecdotal information from various interviews, U.S. Department of Commerce, Caribbean Basin Business Information Center, and others. Audley Shaw, Director, North America, Jamaica National Investment Promotion, Ltd., New York, personal interview, .New York, Jan. 14, 1985. Gary Bechtel, President, Telemar Data Systems, Fairfield, NJ, personal interview, and presentation given at the Miami Conference on the Caribbean Off-Shore Key Punch Operations square table discussion, Miami, Dec. 7, 1984.
PAGE 220
mit data from the region are also attracting attention. Based on observations of people knowledgeable about the industry, it is probably safe to assume that at least l,000 to 2,000 additional data-entry jobs could be created in the Caribbean region over the next decade. In Mexico, the growth potential of data processing is high. An estimated annual growth rate of 10 percent has been cited, but competition from the Caribbean region is anticipated. 11 In other areas of the world, growth is dependent on: 1) companies willingness to establish operations at great distances from the United States, 2) the type of data they are dealing with (i. e., the lead time available for dataconversion work), and 3) the mode of operation being used (i.e., air transportation or electronic transmission). At the present time, the practice of flying hard copy to off-shore sites and returning the data on magnetic media is the predominant mode of operation. Only a few firms are using two-way satellite links to any extent, and the oldest has been operating for about 3 years. Unless satellites are used, it seems unlikely that time-sensitive work would be exported to areas such as the Far East or India. Air transport to these areas can be relatively slow and complicated, particularly if time is of the essence. Wage rates in these areas, however, can be much lower than in the Caribbean. In the Peoples Republic of China, a figure of $2.00 per week for clerical workers was cited. 12 In India, the labor rate for keyers may be as little as one-tenth to one-fifteenth of the U.S. rate, which more than compensates for transport costs, The president of one firm operating there estimates that it would cost approximately $65 to key 10,000 characters in the United States, while he can get the same job done in India for about $7 to $10, and the quality is higher. .. lMollie Shields, Commercial Officer, U.S. Embassy, Mexico City, Mexico. Telexed response to inquiry. Posthuma, op. cit., p. 5. Nick E}age, General Information Ser\ices, Philadelphia. Telephone inter~iew, July 3, 1984. Ibid. Ch. 8Off-Shore Office Work 213 At least one U.S. firm keys data in Singapore, but this site may have been selected because the data is then relayed to the firms office in Australia. Presumably, transmission costs also were an important factor. However, a World Bank report says that two-thirds of Singapores total output is services and over 35 percent of these services were devoted to the production and distribution of information. The primary information sector in Singapore contributes over 24 percent of gross domestic products. 15 In the Far East, therefore, Singapore would be a promising site for expansion of data processing work. As to the future of off-shore keying, there may be inherent limitations on growth. International telecommunications to link off-shore offices with the United States may not be available everywhere or may not have the tremendous growth in capacity that some expect. Not all raw information (i.e., hard copy) can be readily transmitted abroad for keying. Facsimile clarity is still a significant problem, particularly if operators must process large volumes of data. The use of facsimile equipment does not appear to be cost effective as yet. Facsimile transmitters may take as long as 6 minutes per page, . far too long to let overseas operators compete for time-sensitive jobs, according to one firm. 16 Advances in facsimile technology will, presumably, solve these problems. But the very fact that such advances are being made may drastically reduce the need to hire operators for many kinds of data-entry work. Perhaps the most important technological consideration when assessing the future growth of the industry regards optical character readers (OCRs). As advances in the field permit the widespread use of scanners, and the use of machine-readable documents increases, the need for human intervention in data entry will be reduced considerably. World Bank, ilorld De\elopment Report, 1982, New York, 1982. The Instant Off-Shore Off ice, p. 136. 52-644 0 85 8
PAGE 221
214 Automation of Americas Offices Most of the experts consulted on behalf of this study agreed that the arrival of cost efficient, highly capable optical scanners would mean the end of off-shore data-entry work. For example, an executive of one company believes that his firm will stay as long as the costeffectiveness of satellite transmission is not outpaced by advanced technologies that could eliminate the need for a majority of data-entry jobs altogether: To say that well be doing this for the rest of our lives is not accurate; but to say that well pull out in the near future is also not accurate. 17 Others in the industry agreed. It is thus reasonable to expect that off-shore keying may have an effective lifetime of only 15 to 20 years, but during that period it could grow rapidly and have a significant effect on U.S. clerical employment. Posthuma, op. cit., p. 18. (James Marston, vice president of data processing for American Airlines, is quoted. ) HOW OFF-SHORE OFFICES OPERATE Linkage Arrangements There are three principal methods of linking off-shore data-entry sites with data sources and end users in the United States. All three use some combination of air shipment or electronic transmission to move data back and forth between sites. The least sophisticated approach is to ship information by air in both directions. Typically, paper documents, microfiche, magnetic tapes, cards or discs, or audio recordings are collected at one or more U.S. sites for packaging. In the case of paper documents some preliminary hand sorting may occur at this time. A major U.S. airline, for example, sorts ticket stubs according to station, flight number, and passenger class before shipping them abroad, Documents are then packaged and shipped to the off-shore site via regular air freight or overnight courier services. On arrival at the host country airport, the packages are cleared through customs and delivered to the processing site by courier services or employees of the user firm. At the processing facility each keyer has a video display terminal that is typically linked to an onsite computer central proc essing unit. The information on the documents is keyed into the computer, the data is recorded on magnetic tapes, disks or cards, and shipped back to the United States, where the data can be stored, printed on paper, or fed into a computer program for analysis. The second method of linking sites involves air shipping documents from the domestic point of origin to the processing site and then electronically transmitting the digitized data back to its source or to other end users. Typically, the data processing sites are linked via terrestrial telephone lines to satellite Earth stations. The use of this method is, of course, limited by the availability and proximity of earth stations, but permits shorter turnaround times. Where satellite communication facilities are not available it maybe possible to use submarine cable links to the continental United States. Theoretically, the reverse of this process (i.e., transmitting data to the off-shore site and shipping hard documents back to the United States) is also possible, but OTA found no instances of such use. Finally, electronic links may be used to transmit data in both directions. Documents are facsimilied in the United States and transmitted to the off-shore site, keypunched, and then retransmitted back to end users. Information that has already been digitized, but requires further processing or revision may also be sent abroad by simply playing back data on magnetic media and transmitting it from the United States to the off-shore site. Two-way electronic links offer the shortest possible turnaround times, and may, in fact, be the only practical method of processing data off-shore when the ability to deliver it to end users within a very short time is essential. A
PAGE 222
major U.S. printing house that must process and deliver financial data to end users in as little as 15 hours is presently exploring the feasibility of establishing data processing facilities off-shore. Types of Data Processing The kinds of data processed off-shore and the types of processing they require are many and varied. Documents sent abroad for processing include manuscripts; legal documents; insurance or medical records; statistical data; financial statements; coupons for food and other products; order or subscription forms; business documents such as ledgers, payrolls and the like; ticket stubs; mailing lists; contest or sweepstakes entry forms; and audio recordings. In most cases, the documents are produced in large volumes on a relatively continuous basis, and their conversion to electronic form is essential to the end users. Processing in many instances is limited to converting textual or tabular information on hard copy into digital form so it may be more readily reproduced, stored, manipulated, or analyzed. One company, for example, keys in mailing lists and contestant entry forms, while another concentrates on keying textual and financial data for a Fortune 500 client base. Processing may also involve reformatting documents or revising and correcting texts that have already been proofread or edited and digitized in the United States. The printing firm noted previously may transmit to an offshore site rough copy that had already been keyed. The off-shore facility would key in the necessary revisions to make copy ready for typesetting. 18 Another firm reformats financial data at its off-shore facility into standard forms for use in the United States. g Examples of Off-Shore Offices The following capsule descriptions exemplify the types of off-shore offices and the methods by which they operate. Keith Adams, R.R. Donnelly & Sons Co., Chicago. Personal interview, Miami, Dec. 5, 1984, gBechtel, op. cit. from airline ticket stubs for a major U.S. airline, l l the keyed information is returned to the United States via satellite A California-based firm air-freights batches of paper copy to its keying facility in Singapore. From there, the keyed data is sent to its branch office in Australia via satellite. This allows the company to update the files of its Australian clients on a daily basis. A Texas businessman has an arrangement with students at Tsinghua University in the Peoples Republic of China to key in numerical and narrative information from surveys. Magnetic tapes containing the information are then flown to the United States for analysis. In Limerick, Ireland, a major U.S. market research firm keys magazine subscription information. The same firm has branches in Port-au-Prince, Haiti, and Chihuahua, Mexico, where product coupons are sorted and rebate lists are developed. A major U.S. international airline collects ticket stubs from its operating bases throughout the United States and flies them to its keying facility in Barbados. There, information on passenger class, point of embarkation, and destinations, etc., is keyed by a staff of almost, 300. The revenue database is sent over private leased lines via satellite to its data processing office in the United States on a
PAGE 223
216 l Automation of Americas Offices l l daily basis. The company transmits about 2.25 million characters to the United States every month. Information on data entry personnel efficiency is transmitted back to Barbados for analysis by managerial staff. The airline is now soliciting contract work in order to make its keying operation into a profit center. A New Jersey company has entered into a joint venture with a Jamaican firm to key financial data for approximately two dozen Fortune 500 clients. The data-about 50 percent textual and 50 percent tabular-is used for financial reports. The firm cited a document turnaround time of 1 to 2 weeks. It presently flies hard copy to the facility for keying, but anticipates using facsimile equipment in the United States to transmit information to Jamaica, and then return it to the United States via satellite. Approximately 60 people are employed in the firm. It expects to double its size sometime in 1985. A firm in Montego Bay, Jamaica keys in mailing lists for such companies as The Great American Sweepstakes and Publishers Clearinghouse. The responses to mass mailings are flown to Jamaica and mailing lists on magnetic media are flown back to the United States. A small, Philadelphia-based firm operates a joint venture data-entry service that employs about 20 people in Madras, India. The company concentrates on keying in past records, and disk-to-disk conversion of data for clients shifting from one kind of computer system to another. It had originally used a Kurzweil optical scanner that translates printed characters into digital form for storage on magnetic media. It found that use of the machine was not cost effective, and also encountered technical problems regarding the capability of the machine to read accurately. It gave the machine to a university and now uses data-entry workers exclusively to key in information. Factors in Moving Data Entry Off-Shore Several general factors encourage American companies to establish off-shore offices: l l l l l l l The growth in information processing and increased need to convert information from hard copy to digital form; rising domestic labor costs and the availability of lower cost labor off-shore; the availability of good communication infrastructure; the existence of suitable office space, transportation facilities, electrical power, and other support infrastructure; the availability of measures that can be taken to ensure the security of information in transit; the general absence of regulatory impediments to the exportation or to the flow of data across international borders; and the existence of politically stable and economically attractive environments in other countries. Information Market The director of one foreign government industrial development corporation says: In todays world, information resources have become abundant. and more efficient as a result of the progress stimulated by the new electronic technology. It seems fairly clear, to me at least, that as the U.S. economy continues its transformation from industrial activities toward information creation and distribution, that the demands for information services will continue. 20 The need for rapid access to information helped to stimulate the creation of electronic data storage and retrieval systems. Their existence, in turn, has accelerated the speed at which information must be generated, and in[Fred Gollop, Chairman, Barbados Industrial Development Corp., Bridgetown, Barbados. Gollop moderated the Off-shore Key Punch Operations square table discussion at the Miami Conference on the Caribbean, Miami, Dec. 7, 1984.
PAGE 224
Ch. 8Off-Shore Office Work 217 creased the volume of data needed to enter into computerized information systems, The president of a U.S. firm, presently engaged in off-shore keying, notes that . the history of the world, which is in writing, is about to be put into electronic databases throughout the world. 21 The capability and need to generate information in large volumes and on a relatively continuous basis is essential for the move to off-shore data-entry operations. It is the strength of this growing information market that makes owners of off-shore keying operations enthusiastic about their future. In the absence of an expanding information market, the viability of off-shore operations would be doubtful. Attracting workers would be difficult if work were sporadic, and cost efficiency could decline rapidly, thus defeating the purpose of establishing such operations in the first place. Domestic and Foreign Labor Costs In the recent past, there has been a trend toward an international division of labor, in which labor intensive manufacturing work has moved to countries where wages are significantly lower than in the United States. This has allowed many American manufacturing firms to maintain their competitiveness in international markets. The exportation of dataentry workone of the most labor intensive phases of data processingis now occurring for the same reasons. The hourly wage rate for keyers in other countries may range from one-fourth to as little as one-fifteenth of rates paid to U.S. keyers. 22 One U.S. firm reports that it was paying domestic keyers $9.50 per hour, whereas average hourly rates at its off-shore facility in Barbados are approximately $2.10.23 This is a savings of 78 percent in labor costs alone. While companies use different methods of estimating their savings by moving off-shore, those questioned indicated that it is costing them about 75 percent less off-shore than it would in the United States to obtain a comparable level of labor output. The availability of trained workers is a major consideration in choosing an off-shore site. Because the bulk of the work done at dataentry sites involves typing (typically, 60 to 90 percent of the employees of data-entry operations may be keyers), a pool of trained typists must be assembled. Companies located in Jamaica and Barbados report no apparent shortage of trained clerical workers; they invariably had more job applicants than positions. 24 Young women comprise about 99 percent of the workers. One firm reported that the average age of its keyers is 26. 25 Most of them have at least a secondary school education, while many have completed at least some post-secondary training. A keying rate of 10,000 keystrokes per hour is cited as an industry standard. 26 Higher rates may be achieved if conversion is limited to keying tabular data on a numerical ten-key pad, while rates may average somewhat lower if text data is being keyed. All companies surveyed indicated that a high literacy rate was a primary consideration. Literacy in English is strongly preferred to avoid the complications of crossing a language barrier, although language barriers may not be critical if keying is limited to numerical data. However, other firms have experimented with using non-English speaking workers for keying text, and report excellent results. The workers do not need to understand words to key the right letters. Workers can learn to recognize up to 100 characters without any prior knowledge of them, and to respond by hitting the correct key. Because thousands of ideograms are used to write many Oriental languages, workers literate in these languages may be particularly adept at character recognition. Bechtel, op. cit. -Page, op. cit. -Jordan, op. cit. -Bechtel, op. cit.; and Jordan, op. cit. c Jordan, op. cit. Bechtel, op. cit.
PAGE 225
218 Automation of Americas Offices The ability to achieve high accuracy rates seems to be a direct result of employing low cost labor. In order to assure high standards of quality control, two or three keyers may be given the same information to key, on the theory that no two people will make the same error. After copy is double- or triplestroked, parity checks may be conducted to spot errors. The original copy is then checked in order to correct those errors .27 It has been reported that by using this technique firms have been able to obtain very high accuracy rates even using non-English speaking keyers. It would obviously be expensive to achieve quality control in this manner in the United States. Using the hourly wage comparison cited earlier as an example, copy triple-stroked for one hour in the United States would cost $28.50, while the same task off-shore would cost $6.30. This estimate is for Barbados, where wage rates are higher than those for most developing countries. Some firms have noted that it is more difficult to find skilled managers locally than it is to find clerk-typists. However, this was usually overcome through programs to promote and train supervisory personnel, sometimes by sending them to the U.S. office for training. 28 The size of available off-shore labor pools may be an important consideration for large operations, especially in countries with small populations, such as many of those in the Caribbean. U.S. off-shore offices in the Carib~bean presently employ from 10 to as many as 1,200 personnel each. At least one firm chose to locate in Jamaica because keying operations companies already were absorbing most of the available labor pool in Barbados, and bidding for such labor might get a little steep particularly for a start-up company. 29 In sum, countries most apt to attract U.S. firms to off-shore offices are those with lowwage rates and a highly literate, skilled labor force of a size sufficient to support proposed -. -. Adams, op. cit. Becker, op. cit.; and Jordan, op. cit. Bechtel, op. cit. operations. The degree to which these characteristics are present promotes establishment of off-shore data conversion. If English literacy is not critical, then virtually any developing country willing to provide the necessary infrastructure and having a large pool of unemployed people could host an off-shore data-entry operation. Communication Technology Advances in telecommunication, in particular, the advent of satellite communication, have encouraged the operation of off-shore offices. Unlike manufacturing industries, which face high costs and pay a time penalty for transporting raw materials and products, the information industry may ship its raw materials and products thousands of miles in seconds when in electronic form. International communications usually take place by satellite or by undersea cable. The use of satellites appears to predominate, but the use of international submarine cables cannot be discounted, and the desire for greater security and reliability of data transmissions may prompt some firms to maintain access to both. A number of technologies are used to provide access to the international communication networks. Some firms have satellite earth stations on their own premises. Most, however gain access to common-use earth stations or to undersea cable terminals as subscribers to the local telephone system or by leasing dedicated lines. Firms transmitting large quantities of data generally find dedicated lines more economical since they can be designed to accommodate high-data transmission speeds. Thus, one criterion for locating off-shore data conversion facilities is the existence of a communication infrastructure with sufficient capacity to provide dedicated lines to large users. Microwave transceivers are widely used for private line communications in the Caribbean where an extensive microwave system is already in place. It has been suggested that cellular telephony could also offer the capability of providing data links between off-shore
PAGE 226
. Ch. 8Off-Shore Office Work l 219 offices and international communication terminals. Cellular telephony allows voice or data communications without the use of land lines. It could play a significant role in bringing the information age to Third World countries, allowing them to not only catch up, but perhaps also even surpass the phone networks that presently exist in the developed countries. According to a representative of a U.S. cellular telephone company: Cellular will provide nationwide data network capability and needed mobile and fixed services to all segments of an emerging nation. It may well assume the role of the wireline system replacement in smaller cities and in isolated rural areas. Cellular could be utilized to remove isolation, encourage dispersal of industry and increase the efficiency of projects located far from the urban concentrations of industry. ] Costa Rica is expected to begin installation of a cellular system in early 1985, which will probably make it the first Third World country to use cellular technology as an integral part of its telecommunication systems. Installation of cellular systems in several countries could increase the options of firms looking for off-shore office sites. Of importance to operators of many off-shore offices are the telecommunication regulations in the host country. In many developing countries, communication networks are owned and operated by the government in order to generate public revenue. In some cases, all phone equipment and services must be purchased or leased from the government, thus limiting users options with regard to the availability, type, and capabilities of equipment. A countrys reputation for providing hookup and maintenance service also deserves consideration. Clearly, it would not be desirable to establish an off-shore office where bureaucratic problems or lack of well-trained technical personnel caused significant delays in Charles T. 1 lagel, .4pplication of Cellular echnolog-? for L~e\wloping Countries, presented at the Miami (conference on the Caribhean Communications I nfrastructures squaretahle discussion, N!iami, December 1984. I hid. service, or impeded the timely and proper maintenance of the phone network. Facilities, Transportation, and Electrical Power By and large it does not appear that U.S. firms have had serious difficulties in locating facilities in which to set up shop, especially in the Caribbean. Many governments have developed industrial estates that rent or lease factory space at very low rates to foreign investors. In Jamaica, for example, the average rent for such space is a mere $.42 per square foot per month, or about $5 (U. S.) per square foot per year (1982 figures). The leasing of space at off-shore sites is considerably cheaper than in the United States, and in some cases so much cheaper that leasing costs may be a minor consideration in the course of site selection. Space may be leased or purchased outright from the private sector, depending on foreign ownership restrictions imposed by government. Or, it may be possible to expand existing locally owned data processing facilities, if, for example, a U.S. firm enters into a joint venture with an existing host country firm. The quality of available facilities in developing countries may vary considerably. Some -Caribbean Central American Action, op. cit. Photo credlf M/v/s(r) of Trade /~?cfustrj and Deve/oprnent S( Chmfoober and Neb/s Off-shore data processing offices occupy a wide range of fact I i ties from modern office buiIdings to converted factory shells. This one is In St. Christopher and Nevis
PAGE 227
220 Automation of Americas Offices office buildings may be on equal footing with those commonly found in the United States, while others may be simple corrugated metal or concrete block shells adaptable to many uses. One U.S. firm adapted such a shell for its data conversion operation in St. KittsNevis. Another spent approximately $1 million (U. S.) to upgrade an existing office building in Barbados. Its facility is regarded by the company, the Barbados Government, and the companys competitors as a showcase operation equal to the best facilities in the United States. 33 The availability of international air transportation may be essential if electronic linkage arrangements are not being used in both directions. Both regular air freight and overnight courier services are used. One reason for the proliferation of off-shore offices in the Caribbean is the availability of frequent, direct flights of relatively short duration to many of the islands from major cities in the United States. Time in transit from New York to Barbados, for example, is about 4.5 hours. This is particularly attractive if time-sensitive documents are being dealt with. It also makes it easier to manage off-shore facilities, since U.S. personnel may get to the sites relatively quickly if necessary. Companies generally indicate that availability and reliability of electrical power are not major concerns in their site selection. Some firms use on-site back-up electrical generators because power grids in some countries experience sporadic brownouts (i.e., voltage drops below specified levels) or blackouts. The cost of electrical power is not a major concern; in Jamaica for example electrical rates are comparable to those on the eastern seaboard of the United States. 34 In general, facilities are less costly off-shore, although infrastructures in many countries lack the reliability and redundancy that is taken for granted in the United States. The growth of foreign investment in less developed countries has, in many cases, spurred government efforts to enhance their infrastructures with a view toward attracting still more foreign capital, technology, and expertise. Security of Data Shipments and Transmissions With large amounts of data being transferred across international boundaries, security of data shipments and transmissions is a concern, but has so far been a minor problem in off-shore sourcing. None of the firms interviewed for OTA identified any special measures that they take to ensure the security of shipments. Yet the loss or interception of documents containing sensitive information could conceivably cause serious damage, particularly if, for example, information on stock offerings or other financial information were made known prematurely, or if private, intracorporate communications are intercepted. Interception of such shipments is not without its difficulties. One would need to know in advance about the nature and timing of shipments. If particularly sensitive documents are being shipped, air courier services may be used to ensure their safety. One courier firm in the Caribbean region indicated that its shipments are accompanied and monitored by their personnel on a continuous basis. 35 The companies shipping the information, as well as their clients, appear to be satisfied with its security. Intercepting electronic data transmissions is another matter; it requires technical knowledge, and may not necessarily yield usable, understandable data because encryption can be used. Even if unscrambled transmissions were intercepted, strings of characters may not be useful unless one knows what to do with them. 36 In some industries such as airlines, much information is already shared, so the value of information transmitted may be negligible to another airline. Nevertheless, wires can be tapped, and satellite transmissions can be picked up by parties other than the intended ones. Wiretapping ;The firm referred to has operating bases in Miami, New York, and other cities in the United States and the Caribbean. Jordan, op. cit. iGollop, op. cit. Ibid.
PAGE 228
Ch. 8Off-Shore Office Work 221 can be detected fairly easily because the amplitude of an electronic signal is attenuated when a third-party taps the line. 37 Identifying unintended recipients of satellite transmissions can be far more difficult. In either case, however, it would seem that a considerable amount of technical ingenuity would be required to not only intercept transmissions, but make them usable as well. Rules and Regulations At the present time, regulations of any sort that bear directly on the establishment or operation of off-shore data-entry services appear to be so few and/or minor that firms involved have expressed relatively little concern over them. Laws that generally affect off-shore keying operations are those concerning telecommunications. Such laws often mandate the use of government-owned Post, Telephone, and Telegraph (PTT) networks and equipment, since the PTTs of developing nations typically generate revenues needed for other government functions. Firms report that these regulations have been more of an annoyance than a major obstacle to off-shore sourcing. Foreign customs regulations also do not appear to have been a problem for data-entry operations. Many firms investing in developing countries are granted concessions including the privilege of importing raw materials and equipment duty free. The only notable complaints have to do with the difficulties in getting documents and other items cleared quickly through customs, especially on weekends, which can be a problem for firms keying time-sensitive documents. However, serious customs problems would probably spur authorities to make remedial actions to avoid losing sizable foreign investments. U.S. telecommunications and customs regulations at present raise no obstacles to offshore keying. Customs laws govern the importation of documents and magnetic recordings, but tariffs are extremely low or not levHough, op. cit. ied at all. 38 U.S. companies keying their own business data (e.g., payroll, inventory, archives, etc.) may import these records duty free. These are classified under schedule 8, pt. 7 of the Tariff Schedules of the United States Annotated (TSUSA) (1981) as item 870.10 records, diagrams and other data with regard to any business operation conducted outside the United States, whether on paper, cards, photographs, blueprints, tapes or other media. 39 Ordinarily, recorded magnetic tapes, cards, or discs are classified under TSUSA schedule 7, part 2, subpart G as item number 724.40 recordings on magnetic tape or any medium other than wire. 40 A duty of 9@ per square foot of recording surface is specified. Thus, a duty of only $9 would be payable for a recording on a standard 2,400 foot reel of .5" inch computer tape (which contains 100 square feet of recording surface). Even this duty may not be imposed if the recordings originated in a country that is beneficiary to the Generalized System of Preferences. Such countries may export most or all of their goods to the United States duty free. In any event, recorded media from data-entry operations are, by and large, regarded as returned U.S. goods and not subject to duties. 41 While the physical characteristics of the media itself are not advanced or improved at keying operations, it can be argued that such media has an intangible goodinformation added to it, and is, accordingly, increased in value. There has been sharp criticism in this regard, particularly by U.S. labor representatives. As one union spokesman says, You may John R. Gray, Chief, Classification and Value Di\ision, U.S. Customs Bureau, Miami, personal inter~iew, Nliami, Dec. 6, 1984. Tariff Schedules of the United States Annotated 1980, Washington, 1980, p. 774. Ibid. ] The Instant Off-Shore Office, p. 136. Specifically, they are listed under TSUSA schedule B, pt. 1 as item number 800.00products of the United States when returned after ha~ing been exported, without having been advanced in \ralue or improted in condition by any process of manufacture or other means while abroad.
PAGE 229
222 l Automation of Americas Offices be importing a $10 tape with $50,000 worth of information on it. 2 Such claims do not appear to be exaggerations in view of the prodigious volume of data that can be recorded on a given amount of magnetic media. If recorded at maximum density (data density is dependent on the type of recording equipment used), a .5" inch wide ninetrack computer tape can hold 6,250 bytes or characters of data per linear inch. A 2,400 foot reel of tape holds up to 94 megabytes (i.e., 94 million characters). To put this in perspective, a commercial directory listing the names, addresses, telephone numbers, chief executives, Standard Industrial Classification numbers, sales levels, and number of employees of the top 51,000 U.S. corporations comprises only about 30 million characters. 43 This equals approximately 1,500 pages of high-density text in very small print. The president of a domestic data-input company has argued that duties should be payable at least on the labor value added to processed tapes and disks. The directory mentioned above, for instance, probably took at least 3,000 person-hours to key at 10,000 keystrokes per hour. Using the $2.30 per hour wage for Barbados, it would cost roughly $6,900 to enter all of the data in the directory. His concern stems from the fact that his firm has lost millions of dollars in contracts to off-shore firms. 44 Certain countries have erected tariff and nontariff barriers to regulate the content and operation of data banks, and designate the methods and routes by which data can be transferred across international borders. Where these regulations exist, they usually stem from governments concerns over the privacy of their citizens, national security, perceived loss of cultural independence, and concern that unrestricted data transfers could lead to loss Dennis Chamott, Assistant Director, Professional Workers Division, AFL-CIO, Washington, DC, personal interview. Washington, DC, Dec. 17, 1984. The listing referred to is Wards Directory of 51,000 Largest U.S. Corporations, vol. 1 (Petaluma, CA: Baldwin H. Ward Publications, 1984). The directory consists of ,504 pages, 11 x 914 over 30 million computer data bytes. The Instant Off-Shore Office, op. cit., p. 136. of control over economic development and growth. 45 Off-shore data-entry operations have not been significantly affected by such laws because they have generally been established in the countries that have not enacted such laws. Many countries have not yet addressed the issues surrounding data and its transfer, and these countries tend to be less developed, have wage rates that are generally lower, and want the jobs, training, and capital that data-entry operations can provide. They view data entry as a basis for the development and advancement of computer and computer-related industries and capabilities. Thus, data-entry firms presently operate in a very lenient regulatory environment. Developing host countries are not inclined to kill the proverbial goose that laid the golden egg by legislating against the data-entry industry. The benefits they obtain from it far outweigh any gains to be had by levying discriminatory taxes or imposing other restrictions. Undoubtedly, any moves in this direction would quickly cause firms to relocate to countries with more favorable business environments of which there are many. U.S. laws impose virtually no burden on off-shore sourcing. Vigorous debate over regulatory matters can be expected to continue, however, as labor unions and domestic firms pressure authorities to extend tariff treatment to international trade in data and data processing services. Political Stability and Investment Climate in Foreign Countries The reason most frequently cited by U.S. firms in all sectors for not investing abroad is perhaps that they must deal with too many unknowns. Political stability of a foreign country and its policies regarding foreign, private investment are foremost concerns of U.S. businesses thinking about establishing off-shore enterprises. 45 James R. Basche, Jr., Regulating International Data Transmission: The Impact on Managng International Business, Washington, DC, 1984, p. 16.
PAGE 230
. -.. .. -. Despite these problems, U.S. firms have found that the Third World presents an abundance of business opportunities for those willing to assume the attendant risks. Many developing countries recognize the advantages of permitting foreign investment to enhance private sector development. Not only does it expand their capital base, but it also provides new opportunities for host-country owned businesses, brings in foreign technology and expertise, and provides training for their people, which in many cases is their most abundant resource. The host countrys investment policies are also instrumental in attracting off-shore business. Specific investment incentives differ depending on the country, but various forms of tax relief and import duty exemptions are commonly used. The investment incentives provided by many developing countries are helpful to investors, and their existence is generally a good indicator of a countrys need for such investment and recognition of its importance. Such incentives may, at the very least, provide an appearance of political stability and continuity, and an indication that foreign governments are aware of the need to create favorable business environments for investment by industrialized nations. Ch. 8Off-Shore Office Work l 223 The U.S. Government and the governments of other countries have encouraged investment in developing nations by helping firms to locate business opportunities, guarding them against the effects of adverse political developments, and providing attractive business incentives. At least one American company notes that both United States and foreign government assistance was instrumental in the establishment of its data-entry operation in Jamaica. Of significant importance to the firm was assistance provided by the Overseas Private Investment Corporation (OPIC). OPIC is a selfsustaining, profit-making Federal agency (a component of the U.S. International Development Cooperation Agency, which also includes the Agency for International Development) whose mandate is to mobilize and facilitate the participation of United States private capital and skills in the economic and social development of less developed, friendly countries and areas. To do this, the agency provides political risk insurance, direct loans and loan guarantees, and other service, All (of which) are designed to reduce the perceived stumbling blocks and risks associated with overseas investment. 46
PAGE 231
224 Automation of Americas Offices er accuracy rates for converted data can also be obtained abroad because low-labor cost permits redundant keying. Another argument raised in favor of offshore data processing is that in some cases the work will not be done otherwise. One company official pointed out that his firm is doing massive archival work for several major customers. The customers had considered doing it in-house, but finding the work too costly, they postponed it until they could find a company like his to do it at an economical rate. 51 U.S. Labor Organizations Representatives of U.S. labor unions argue strongly against the movement of data-entry operations off-shore. While the AFL-CIO has not yet adopted an official position, their general opposition to moving any type of work off-shore is based largely on the following points: l l l The creation of off-shore work represents a direct displacement of U.S. workers which is detrimental to our domestic employment situation, and does not always have a permanent, positive effect on the development of needy countries. Labor competition for jobs based on wage rate differentials is not fair or equitable, and is inherently exploitative of labor in less developed nations. Moving operations of any kind off-shore reduces effective control of our economy, and threatens our economic security. U.S. labor organizations argue that for every job created off-shore, one is lost in this country. Exporting clerical work to low-wage countries is the latest phase in a trend that has been going on for a long time in U.S. manufacturing industries. While the number of people employed in off-shore offices has had, as yet, no discernible impact on U.S. labor, this should not preclude our taking steps to stop this trend. Bechtel, op. cit. Further, it is argued, creating such work offshore does not always have a permanent developmental impact on poorer countries, because companies always seek areas where wage rates are lower. As a country develops and its wage rates rise, companies located there may move to still less developed and less costly areas in which to operate. 52 Union representatives argue that because our economy is becoming more heavily dependent on information and white-collar work, as opposed to manufacturing or blue-collar work, and as white-collar work becomes more automated, we could be taking away the last large group of jobs that the economy is supposedly going to have available. 53 Others argue that locating more and more jobs off-shore may jeopardize our economic security because: 1) investment in the economy is lost, and 2) off-shore operations exist in environments that cannot be controlled, thereby making the investments vulnerable to adverse political developments. The continued relocation of work out of this country is economically unsound because we are foregoing opportunities to create domestic jobs and new business enterprises. 54 Also, our dependence on foreign labor further diminishes the ability to function economically and independently, and, therefore, weakens the United States. There have been repeated nationalizations of U.S. investments abroad in recent years. Off-shore office work is particularly worthy of special consideration: What (were) doing now with shipping jobs overseas (via the use of) electronic transmission. is making (companies) even more vulnerable. Work that they have to have for (their) company to function, they are placing outside the realm of their control. In terms of control, what do you do if youre not -Chamot, op. cit. This argument is also presented in The Electronic Sweatshop: The Use and Misuse of Work Stations in the Home, a presentation given by Dennis Chamott and John 1,. Zalusky to the National Executive Forum: Office Work Stations in the Home, November 1983. Chamot, interview, Barbara Hutchinson, Director, Womens Bureau, AFI,CIO, Washington, DC, personal interview, Washington, DC, NOV. 19, 1984.
PAGE 232
Ch. 8Off-Shore Office Work l 225 duplicating that system here, and that system shuts down? 55 Off-shore offices would be very easy to shut down, because they tend not to be very large, do not have a major impact on host country economies, and are heavily dependent on vulnerable communication and transportation networks. Foreign Government Officials The direct impact of off-shore data entry on increased employment in developing countries may be significant and quite rapid. One company began interviewing prospective employees in Barbados in May 1983, commenced training on August 1, and began production in October of the same year. The operation presently employees 275 people, 80 of them managerial, technical, and supervisory personnel. This suggests that keying operations, especially larger ones, may provide significant job opportunities not only for clerical workers, but also for more highly trained workers. The same firm let approximately $1 million in contracts to local firms to refurbish its building. Thus, business opportunities and jobs may also be created in other economic sectors. Reactions of foreign government officials to off-shore office work are very positive. Those countries that already host such operations are pleased with the results, while those that do not yet have such operations want them. In an informal survey, representatives from more than 20 Caribbean and Central American countries unanimously claimed that such investment projects would aid their economies and would be welcomed. 56 Barbados has gone so far as to single out information services as a sector to receive major emphasis for development. At the other end of the spectrum, representatives of some Ibid. Answers were in response to informal interviews conducted by the contractor at the Miami Conference on the Caribbean, Miami, Dec. 3-7, 1984. Representatives of approximatel~ 20 Caribbean and Central American countries were queried, Astriab, op. cit. countries know little about the phenomenon, but are eager to learn more about it. The work is labor intensive, requires only a moderate capital outlay, and can generate employment rapidly. It is a clean industry, without the heavy equipment, large space requirements, and pollution often associated with other industrial enterprises. The industry provides, at the very least, rudimentary training in computer usea rare opportunity for workers in less developed nations. Finally, it establishes a foundation on which further advancement in computer-related industries such as software development, technical services, and data transmission may grow. The Director of the Barbados Industrial Development Corporation (BIDC) elaborates: 1 know from experience in Barbados that we can offer a lower cost location (for the industry), and answer the needs of many industrial companies in the United States. I know too that the countries in the Caribbean can benefit from increased employment, and in the case of off-shore keypunch operations, this can be fairly rapid. And I also think we can benefit from technical education which would ultimately set us on the path to higher levels of technology in the computer services industry .57 Barbados has targeted information services as an industry group that is expected to have a major impact on the objectives of the Barbados Industrial Development Corporation and the Barbadian economy. According to its 1983-87 Development Plan: The Information Services industry can be expected to generate substantial employment in the short term linkages with the more technically proficient computer industry will be sought as a means of expanding the sectors contribution to the economy Local participation in the Information Services Industry can come about through the establishment of service bureaus to perform offshore data processing for North American companies. 58 Gollop, op. cit. Ilarbados Industrial De\relepment Corp., Industrial De\elopment Plan, 1983-1987, 13ridgetown, Barbados, 1983.
PAGE 233
226 l Automation of Americas Offices Development officials interviewed for OTA all agreed that foreign data-entry workers view their jobs and status in a considerably more positive light than their American counterparts. While keying in the United States is sometimes pictured as a low status, boring job, foreign workers often consider it a gateway to opportunities that did not formerly exist (as do many minority workers in the United Statessee chapter 12). Over-dependence of countries on information services is not seen as a problem at this point because the industry is presently so small. In addition, many developing nations opt for diversification of industrial development as a key ingredient of a more stable economy. They want to avoid heavy dependence on traditional industries such as tourism or agriculture, with their characteristic seasonal fluctuations in income and employment. Economic Development Organizations Development organizations are in essence the middlemen in regard to the development of off-shore office work. They view the expansion of the sector in the much larger context of economic development, are supportive of its growth, and reaffirm the arguments put forth by U.S. companies and foreign governments in favor of such growth. The sheer number of jobs (off-shore data processing) creates is the main attraction of the industry in the view of developing countries, says one official of a private, nonprofit development organization .59 This development expert claims that negative reactions, if any, in developing countries usually come from the more educated sectors and the labor movement, or, as he put it: people who are not worried about having a job. 60 Those citizens of developing countries who view U.S. investment as exploitative are those who are generally of higher economic status Gordon Hunt, Director, Investment Services, Caribbean/Central American Action, Washington, DC, personal interview, Washington, DC, Dec. 18, 1984. This claim was also affirmed by Gooch, op. cit. Ibid. and are not in need of the job opportunities that result from such investment. Some experts favor investment in developing country private sectors as opposed to just public-infrastructural development projects. According to this argument, in many cases developmental aid programs have not achieved their intended purposes. For example, while good roads may indeed have been built, private businesses along those roads are necessary for the general advancement of the economy. While the wages paid to workers in these businesses may be extremely low by U.S. standards, this should not be considered exploitative. Rebutting arguments from U.S. labor that the flow of clerical work out of the United States should be stopped, development organizations point to the great impetus of increased international economic interdependency and the disadvantages of impeding capital flows. Business goes where business can be done, as one development organization official said, The flow of capital cannot be stopped unless one wishes to control the economy, and this runs against the whole tradition of free trade. 62 Answering U.S. labors claim that wage competition is the only focus when selecting business sites, a development official countered that it has never been a case of just wage rates: If that were the case, there wouldnt be a bit of work done but in China and India. 63 While this is, of course, an overstatement, higher profits are not the only motivation for many businesses off-shore. For some, it may simply be a matter of survival. Viewing the exploitation argument in a different light: You will certainly get cases of exploitation, but it is far more exploitative to say (that we will) keep those jobs here, and pay an unacceptably high level of wage which will make our products unacceptable anywhere other than in our own economy. Plus, by doing so ] Ibid. Ibid. Ibid.
PAGE 234
Ch. 8Off-Shore Office Work l 227 we will starve (the people in less developed countries). 64 There has been massive legal and illegal immigration to the United States from the Caribbean. Many countries in the region have their second largest, and in some cases their largest population centers in the United States. Some see it in the best interest of the United States to create jobs for people in their own countries to stem this tide of immigration: Ibid. Do you employ those people in their own countries or do you employ them (in the United States)?, because thats where theyre all coming. Do they come up here and work illegally for sweatshop wages or do they stay in their own countries, where they would rather be to begin with if they could make a decent wage? 65 Ibid. PUBLIC POLICY Off-shore office work does not have a significant influence on the U.S. economy at present. Only a few dozen firms and a few thousand employees are involved. Nevertheless, it could grow. If domestic wage rates remain comparatively high and telecommunication and transportation costs continue to fall, market forces will encourage more firms to investigate the off-shore alternative. Growth could be quite rapid, at least in the shortto medium-term future. In the long term, technological advances in input technology, especially optical scanning, are likely to undermine the cost advantages of off-shore data-entry work. Thus, off-shore offices are likely to be a temporary phenomenon, unless some other technological advances or economic conditions make it feasible to move other types of office work off-shore for example, telephone reservations or order processing. Assuming that it is temporary, the question remains how U.S. public policy should deal with off-shore offices. The alternatives include prohibiting, regulating, or encouraging it. Regulating or Prohibiting Off-Shore Offices Methods for limiting the growth of off-shore offices could include such government actions as: ISSUES imposing regulation on the types of data that can be imported, imposing restrictions on the hardware and software used off-shore, using taxes to discourage off-shore work or encourage that work be done domestically, and/or imposing restrictions on the use of private telecommunication lines. Regulation of Data Many nations have imposed restrictions on transborder data flows for the purpose of protecting the privacy of their citizens, for example, prohibiting transmission of name-linked information to countries with less stringent privacy laws. The United States has data protection laws designed to restrict the uses of name-linked data and assure access to the data by persons referenced in them. These laws do not, however, address the front-end issue of how or where information may be entered into electronic databases. Nor does U.S. law restrict the flow of data across international borders. It is not clear that privacy protection is necessarily a good motivation for U.S. regulation of off-shore offices. Even if it were, the enactment of privacy protection laws governing transborder data flows would affect other industries, such as banking, and the internal
PAGE 235
228 l Automation of Americas Offices operation of many multinational corporations, as well as off-shore keyers. A more direct approach to discouraging establishment of off-shore offices is to simply designate data entry as a phase of processing that U.S. companies must, in virtually all circumstances, carry out within our own borders if the data is destined for eventual sale. This would, in effect, constitute a local content requirement. Because such a requirement implies that the data being entered is a commodity, this type of restriction may not be applicable to certain types of datae.g., revenue, personnel, payroll information, and the like-reserved for the exclusive use of a business, and not meant to be sold. Buy national requirements could fill this gap by mandating that off-shore data entry could not be resorted to unless domestic alternatives were unavailable. Brazil and Canada for example, have imposed regulations that require certain phases of data processing to be conducted within their own borders. 66 If viewed in a broad context, however, imposition of such regulations may serve to stimulate reciprocal actions by other countries. Thus, more harm could be done to the U.S. data processing industry than could be counter-balanced by benefits of restricting a particular phase of data processing to our own shores. Restrictions on Hardware and Software In lieu of limiting the kinds of data that could be entered off-shore or requiring that data entry be carried out domestically, requirements that U.S. equipment be used could be imposed. In essence, this would constitute another buy national requirement. This might effectively preclude the establishment of keying operations in some countries, since, as previously mentioned, foreign communications authorities often monopolize telecommunication equipment markets and require that their locally manufactured equipment be used. However, many countries do not even have a computer Joan E. Spero, Barriers to International Information Flows, Telecommunications, November 1983, p. 68. industry, and U.S. computers and modems are already usually the equipment of choice for off-shore firms. Taxation Alternatives In theory, duties could be imposed on the data that is keyed off-shore and subsequently imported into the United States by physical or electronic means; the data could be treated as a commodity or primary good imported for consumption. Alternatives might be to levy duties on the labor value added to data keyed off-shore, or to impose a trigger price tariff that would keep the price of off-shore keying above domestic levels. In practice, however, levying any such tariffs would be fraught with difficulty. First, this shotgun approach to regulation would probably affect all data importers, of which offshore keyers area very small minority. In addition, a method for determining the value of data would have to be developed. Value might be related to proposed end use or might be based simply on the basis of the volume of data imported (e.g., a duty might be levied for every kilobyteone thousand characters-of data). The latter seems inherently inequitable, since some kinds of data are much more valuable then others. The former seems nearly impossible to enforce, unless customs officials are going to play back and analyze every computer tape. Imposing duties on data imported via telecommunication presents even more onerous technical and political problems. The sheer volume of transmissions to the United States and the many routes they may follow present logistical problems of inestimable proportions for monitoring. In the future, the increased use of direct satellite links between off-shore installations and domestic offices will make the imposition of border controls on data flows extremely difficult, if not impossible. 67 In any case, it seems clear that on-site monitoring by government authorities and/or wiretapping on a grand scale would be necessary -Ibid.
PAGE 236
Ch. 8Off-Shore Office Work l 229 to police electronic data flows. Even if this were technically feasible, it is doubtful that revenues accrued by a government could justify the efforts made to collect them. Putting technical and economic considerations aside, moves toward monitoring data flows in the United States would undoubtedly encounter many political obstacles. The practice would cast a clear shadow of authoritarianism, and have a very inhibitive effect on business and communications in general. A far greater purpose than stifling a minor industry would obviously be in order if the practice were adopted. In sum, while the objective to curtailing off-shore keying could readily be attained through tariff legislation, a more tightly focused approach would priate. seem more approRestrictions on Dedicated Telecommunication Lines Limiting the availability of leased private telecommunication lines to off-shore keying operations or prohibiting their use for this purpose would unquestionably deter the establishment of such operations, because it would directly effect one of the lifelines of the industry. Since leased lines appear to be the most cost effective way to transmit large volumes of data across international borders, off-shore keyers relying predominantly on telecommunication capabilities to operate would be acutely effected by any such restrictions. Regulations of this sort could, of course, be applied only to the U.S. side of any communication link (e.g., the down-link from a satellite sending data to the United States, a terrestrial cable in U.S. territory, etc.), since the foreign ends of such links are out of American jurisdiction. Nonetheless, this strategy could be effective since the regulations could be applied specifically to off-shore data-entry firms without impinging on the rest of the data processing industry. Restricting access to dedicated lines might be accomplished by specifying a limitation on the number of lines available to off-shore keyers in different regions. Implicit limitations might be achieved by imposing stiff tax penalties or surcharges on data-entry firms leasing such lines. Several countries already use this strategy to protect their internal data processing industry and it is quite effective in curtailing undesired activities. In Japan, the Ministry of Posts and Telecommunications, through the international telecommunication authority, KDD, kept two large U.S. data processing firms out of the countrys market by first denying them dedicated lines, and then restricting their use so that the firms could not market all of their services. The Deutsche Bundepost in West Germany requires users of leased lines to use the public-data network. In 1982, the authority declared that it would permit the use of international leased lines only if some phase of data processing were conducted in Germany before data was transmitted out of the country. 69 Brazil has enacted many restrictions on the use of leased lines. For instance, firms may not use them to access databases located outside the country. Short of limiting access or prohibiting offshore data-entry companies from using leased lines, legislation could require telecommunication carriers to impose usage-sensitive rates on these firms. A number of foreign governments have expressed interest in usage-sensitive rates, since they fear the loss of revenues that could result from reduced use of their public networks. One international data processing service estimated that the introduction of usage-sensitive rates would raise its operating costs by 700 percent. 72 In the face of such cost increases, many off-shore data-entry firms would probably have to close up shop. Regulations on leased lines could, therefore, offer a relatively clean approach toward discouraging off-shore keying; impeding other types of data flows could be avoided because restrictions could be tightly focused. To be Spero, op. cit., p. 68. Ibid. Ibid. ,Jussawal]a and Cheah, op. cit.. p. 292. Spero. op. ci~., p. 68.
PAGE 237
230 l Automation of Americas Offices truly effective, however, such restrictions would probably have to go hand-in-hand with some form of customs regulations so that control could be extended over firms who import data recorded on magnetic media. Encourage or Do Nothing About Off-Shore Office Work Encouraging off-shore offices is discussed in the same section as a do nothing alternative because the current status of telecom munication and customs regulation already favor the growth of this industry. Inaction will undoubtedly assure the continued export of data processing work for so long as the marketplace provides the needed incentives. Further encouragement might be achieved through increased activity of international development agencies. For example, the Overseas Private Investment Corporation could make more loans available and otherwise step up efforts to help U.S. firms find suitable locations and establish keying operations off-shore.
PAGE 238
Chapter 9 The Automation of Federal Government Offices
PAGE 239
Contents Page Federal Procurement and Acquisition of Office Automation . . . . 234 The Three Phases of Federal Office Automation . . . . . . 234 Laws and Policies . . . . . . . . . . . . . 235 The Federal Inventory of Office Automation . . . . . . . 238 Acquisition Strategies and Problems . . . . . . . . . 239 The Dilemma of Procurement Policy . . . . . . . . . 241 What Will Office Automation Mean for Federal Office Productivity? . . 242 Evaluations of productivity . . . . . . . . . . 242 Incentives . . . . . . . . . . . . . . 243 Implications for the Federal Work Force . . . . . . . . 244 Size of the Work Force . . . . . . . . . . . 244 Changes in the Mix of Federal Jobs . . . . . . . . . 246 Shifts Among Occupational Categories and Grade Levels . . . . 246 Unanswered Questions. . . . . . . . . . . . 250 Part-Time and Temporary Workers . . . . . . . . . 251 Women in the Federal Work Force . . . . . . . . . 252 Quality of Worklife. . . . . . . . . . . . . 252 Labor Management Relations and Federal Unions . . . . . . 254 International Comparisons . . . . . . . . . . . 255 Effects of Office Automation on Public Services . . . . . . . 256 A Case Study of Expanding Responsibilities . . . . . . . 256 Changes in the Nature of Delivered Services . . . . . . . 258 Effects on Governance . . . . . . . . . . . . 258 Figures Figure No. Page 9-l. Paid Civilian Employment of the Federal Government, 1881-1983 . . 245 9-2. Changes in the Distribution of Federal Employees by Occupation, 1975 and 1983 . . . . . . . . . . . . . 247 9-3. Change in Distribution of Federal Employees by Grade Levels, 1974-80 . 248 9-4. Distribution of Federal Employment by Grade Level 1975 and 1983 . 249 9-5. Full-Time White-Collar Employment in the Federal Work Force by Grade Level and Sex, 1974-75 and 1983 . . . . . . . 253
PAGE 240
Chapter 9 The Automation of Federal Government Offices Automation of Federal Government offices is generally keeping pace with automation in the private sector. The effects are likely to be at least as significant in government offices as in corporate offices. But the forces that drive change are not the same, and the consequences will not necessarily be the same. Government is not business, although it is often argued that it should be business-like in its approach to delivering services. Federal office automation is preceding on the reasonable assumption of significant benefits. Large investments of public resources are involved, and most Federal employees will be affected. Thus, a close look at the potential consequences of Federal office automation is merited, The Federal Government is in effect the Nations largest office. It occupies 2.6 billion square feet of office space; it has 332 accounting systems and over 100 payroll systems; and it employs about 1,7 million white-collar workers. 1 Opinions vary widely as to how well and how rapidly the Federal office is being automated. One trade journal concluded that government is pioneering some leading-edge office automation programs and in many respects is ahead of the private sector because the Reagan Administration is emphasizing automaIt is in fact surprisingly difficult to ascertain how many white-collar Federal employees there are, or indeed how many Federal workers there are in all, at any one time. The Office of Personnel Management, the Bureau of Labor Statistics, and the Bureau of the Census all publish figures, but they are nevrer the same figures. The total given varies widely depending on the time of year or month; or whether it is a monthly average or an estimate or a survey; or whether and how the count treats temporary workers, part-time workers, intermittent workers, postal workers, census takers, CIA and NSA workers, congressional employees, judicial-system employees, student interns, fellows, etc.; and on how many of the approximately 106 pay plans OMB chooses to count. All of these are traps for the unwary analyst. tion in the effort to increase efficiency. Some experts say with equal confidence that government offices are behind the state-of-the-art. Comparisons based on many case studies however indicate that while some large corporations are far ahead of Federal agencies in using information technologies, the government is at least keeping up with the private sector as a whole. Among major institutional sectors it has been the largest user of computer-based information systems. 3 Some agencies are behind their closest private sector counterparts. For example, the Federal Reserve Board does not compare with leading financial institutions in terms of either advanced applications or the proportion of critical work that is automated. On the other hand, some agencies are at the frontier in specific office automation applications; International Revenue Service (IRS) is one example. Most agencies are in the mainstream in terms of penetration and in terms of advanced applications of hardware and software. Federal agencies were among the first institutions to adopt large computers. In the last few years, they have been adding small computers and word processors to augment their large-scale data processing. For the next few years, a major trend will be the linking of microprocessors, mainframe computers, and other devices into integrated office systems; and the networking arrangements that will connect office to office, headquarters to field Office Administration and Automation, September 1984, p. 56. John Leslie King and Kenneth L. Kraemer, Information Systems and Intergovernmental Relations, Public Sector Performance, Trudi C. Miller (cd. ) (Baltimore: The Johns Hopkins University Press, 1984), p, 103, This judgment is based on materials supplied by Federal agencies and comparisons drawn from the literature and from OTA contractor case studies; and specifically, on the conclusions of an OTA contracted report by The Educational Fund for Individual Rights. 233
PAGE 241
234 Automation of America's Offices offices, and Federal agencies to external systems and databases. This section looks briefly at the effects of office automation that can be detected now, and the effects to be expected over the next 15 years. After a discussion of Federal procurel ment and acquisition policies with regard to office automation, the rest of the chapter looks l in succession at the following questions: l l l l Are there major problems in Federal acl quisition of information technology? Will automation make Federal offices more efficient, or more productive? If so, can that benefit be translated into l lower labor costs, and lower Federal budgets? What are the potential effects on the size and structure of the civil service, and what are the implications for recruitment, classification, and retention of Federal workers and for budgetary and personnel policies? What are the implications for career expectations and opportunities of Federal white-collar employees? How will automation affect the quality of their working environment? Will automation affect the relationship of government and citizenswill it change the availability or quality of government services? Could it affect the exercise of authority, accountability, responsibility, and the quality of decisionmaking? FEDERAL PROCUREMENT AND ACQUISITION OF OFFICE AUTOMATION It is the declared policy of the present Administration that information is an economic resource and should be managed efficiently. The emphasis has been on reducing the cost of information-handling rather than on increasing information services. Federal agencies have had wide latitude in making decisions about office automation. The policy has been to keep governmentwide requirements and restrictions to a minimum. Critics charge that this has led to uncontrolled proliferation of small computers, and that the lack of compatibility among them is preventing the realization of expected benefits of automation. On the other hand, overly detailed and rigid specifications in procurement of major information systems, including local area networks, is said to preclude vendors from finding innovative ways to meet government needs, and to result in purchase of equipment that is already far behind the state-of-the-art when it is installed. The Three Phases of Federal Office Automation The adoption of mainframe computers in the late 1950s 5 led almost immediately to development of large centralized-computer centers. The Office of Management and Budget (OMB) and the General Services Administration (GSA) were soon given governmentwide authority over automatic data processing (ADP) policy and standards. During the 1960s and 1970s many of the major activities of government became dependent on computers for management functions such as planning, program control, financial and payroll operations, procurement control, auditing and inspection, and other government functions. The acquisition process was framed around centralized ADP The first general data processing computer, UNIVAC I, was acquired by the Bureau of the Census, in 1951.
PAGE 242
Ch. 9The Automation of Federal Government Offices 235 and the communication functions that were developing in parallel. In the 1980s however the dominant theme in Federal office automation has become the spread of end-user computing. It is difficult to distinguish the effects of these two phases of computerization since decentralized computing does not replace, but is superimposed on, centralized computing. Now personal computers and word processors are often networked and part of an integrated system. Federal procurement policy, and accounting and inventory practices also blur distinctions between these two kinds of office automation. Most Federal agencies now have word processing, spread-sheet packages, automated document transmission, and calendaring. The other most frequently used functions are electronic filing and computer graphics. Many agency headquarters communicate by computer with their field offices nationwide. Personal computers are being used more and more by managers and professionals as well as by support staff. The National Academy of Public Administration, in a 1983 report, assumed that this reduces the load factors and utilization rates of mainframe computers, and said that this cast serious doubt on the future role of many of the large central computer service centers that have been built up. over the last 10 to 15 years. 8 But experts generally do not believe that ADP centers will be superseded by distributed processing. Rather, as small and large computers are linked, the ADP center will be the locus and guardian of the agency data to which all managers and professionals will increasingly have access and make contributions. International Data Corp., Procurement Information Management Service, Federal Acquisition Strategies for Office Automation, research paper for Continuous Information Services Clients, March 1983. According to an office automation survey compiled by the Information Management Assistance division of the Office of Information Resources Management, General Services Administration, November 1984. Thirty-one departments and agencies responded to the survey. National Academy of Public Administration, I?e~ritzziizing Federal Management: Managers and Their Overburdened S.Fstems, November 1983. Some large Federal computers are, however, already obsolete or will soon become so. 9 Obsolete, in this case, does not mean that they are no longer functioning, but merely that more cost-effective technology is available. The old systems require repeated patching and modification, maintenance costs are high, and spare parts sometimes not available. Older computers sometimes have limited on-line processing capacity because they were designed for 24-hour operation and not for the peaks caused by many end-users. When these computers are replaced, databases often have to be converted, and new software packages developed or adopted. This is expensive. The incompatibility of equipment from many vendors is also causing problems. Laws and Policies The Paperwork Reduction Act of 1980, Public Law 96-511, was a milestone in government information management. In addition to reducing the paperwork burden imposed on business by government, the act was aimed at improving efficiency and effectiveness in the use of information. It promulgated the concept of information resources management (IRM), meaning the integrated management of all basic information-handling activities and functions. The sections of the act dealing with information-resources management cover everything from conventional libraries to centralized ADP sytems, and have a direct effect on office automation. The act required each agency to appoint an information-resources manager. 11 It charged The Grace Commission criticized the government for allowing its data processing systems, which in the early 1970s were state-of-the-art, to fall behind; according to the Commission about 60 percent of the governments then 17,200 computers were in need of replacement or significant upgrading. However, in a recent review of 100 major systems, GSA concluded that only 5 percent are totally supported by obsolete ADPE systems, and that obsolescence is not as extensive as has been claimed. See Assessing ADPE Obsolescence in Major Federtd S.ywtems, U.S. General Services Administration, February 1985. The act built on the recommendations of the Commission on Federal Paperwork (1975-77). The act specified that these officials should be at the level of Assistant Secretary. Since the number of these positions (continued)
PAGE 243
236 l Automation of Americas Offices OMB, assisted by GSA, with reviewing information resources management in each agency at least every 3 years. OMB has largely delegated this responsibility to GSA. In practice, each agency conducts its own review, for which GSA provides a voluminous handbook or set of guidelines. The agency reviews are then reviewed by GSA. OMB has stated two basic tenets that govern its approach to information-resources management: 13 l Information is an economic resource and should be managed in the same way that other economic resources are managed. l Information-resources management should entail the management of the total information life cycle from collection to dissemination. OMB provides guidance on all matters of budget allocation and procurement in Federal agencies, but this guidance is not specific enough to materially affect the acquisition of office automation equipment. 14 GSA does provide guidance to agencies on this subject, although agencies still make their own basic decisions. Twenty years ago, in 1965, the Brooks Act (Public Law 89-306) gave GSA sole authority to procure ADP systems. Although this auin each agency is limited by law, the responsibility was generally given to the Assistant Secretary for Administration or the equivalent. These are now referred to as Senior Designated Officials or SDOS, and there is typically an Office of Information Resources Management under them, with a director who acts as the IRM and represents the agency on the Interagency Committee on Information Resources Management. There is an Office of Information and Regulatory Affairs in OMB that coordinates OMB responsibilities for implementation of the Paperwork Reduction Act. Letter of delegation of June 13, 1983, and Temporary Regulation 10. Executive Office of the President, Office of Management and Budget, Improving Government Information Resources Management, March 1982, p. I, hereafter cited as EOP/OMB 1. MB guidance on procurement (which comes from the M or management side of OMB, and specifically from the Office of Federal Procurement Policy or OFPP) is of course to be distinguished from OMB review of agency budget requests on the B side of OMB, which certainly does affect the level of office automation procurements. DOD computers and related devices were exempted if the equipment was a mission critical computer resource. DOD has argued that if any office automation system is tied into thority is premised on large-systems procurement it remains the basic authority for purchase of office automation equipment. If office equipment contains data processing components it must be bought under GSA standards and rules governing competitiveness in procurement. Procurements of microcomputers and word processors, 16 when under $300,000, are handled under a GSA schedule (Schedule C), which means that there are simplified procedures for competitive bids from already authorized vendors. When the total value of a procurement exceeds certain thresholds (for ADP equipment in general it is now $2.5 million) GSA usually issues to the agency a Delegation of Procurement Authority, based on information provided by the agency about the justification for the procurement. The agency then draws up specifications and goes through its own competitive bidding procedure. (GSA can withdraw this delegation or change the thresholds.) GSA acquisitions do not account for all, or even for the preponderance of, Federal-agency microcomputer acquisitions. The Competition in Contracting Act 17 that took effect April 1, 1985, created simpler procedures for using GSAs ADP Schedule in microcomputer purchases under $300,000. GSA has an approved list of 45 microcomputers, available from 36 suppliers at special Federal rates. In addition, a GSA Computer Store carries 15 brands (not necessarily on the list); this is designed to encourage the selection by end users rather than leaving the choice to an agencys central-purchasing agent. GSA has published several attractive booklets of a mission-critical computer resource system then it is also exempt from GSA procurement regulations. Statement of Undersecretary of Defense DeLauer on Mar. 4, 1983, according to International Data Corp., op. cit., p. 22. The question of whether this applies to word processors has been under review several times and the outcome is not clear. ) Word processors were not included under ADP equipment until late 1983 (SPMR Temporary Reg. F500, Oct. 25, 1983). Part of the Deficit Reduction Act of 1984. The CIC Act, aimed at increasing competitiveness in procurement, makes it more difficult to buy from a preselected sole source, but defines as competitive, awards that are made under a GSA MultiAward Schedule Program such as the ADP program. An agency may also exclude a specific vendor in the interest of maintaining alternative sources.
PAGE 244
Ch, 9The Automation of Federal Government Offices 237 advice to help agencies in buying and managing microcomputers. 18 GSA also has responsibility for most of the many common-user telecommunication facilities used by Federal civil agencies. If an agency wants to make a major change in common-user services (e.g., Wide Area Telephone Service, or WATS, lines) to implement office automation communications, GSA must approve.] The GSA authority under the Paperwork Reduction Act (and by delegation from OMB) was until recently exercised through two different services. The Automated Data and Telecommunications Service, which dealt with ADP equipment, and part of the National Archives and Record Service, which dealt with records management, micrographic, and word processing equipment, have been merged to create the Office of Information Resources Management. 20 In April 1984, a number of policies and regulations related to information resources and technologies were consolidated in a Federal Information Resources Management Regulation (CFR Pt. 41, ch. 201). Amendments to the Paperwork Reduction Act have been introduced in Congress aimed at strengthening some aspects of the law. 21 Other agencies have roles in office automation procurement. The Institute for Computer Science and Technology in the National Bureau of Standards develops standards for ADP and communications equipment, develops technical guidelines, and prepares a yearly forecast of developments in computer technology, including office automation. 22 The General office of Information Resources Management, U.S. General Services Administration, Managing End User Computing in the Federal Government, June 1983, and End User Guide to Bu.ving Small Computers, August 1984, DOD manages its own procurements of communications technology. Another part of GSAs National Archives and Records Service became the independent National Archives and Records Administration. -S. 2433 (amendments of 1984); H.R. 2718 (amendments of 1983); hearings were held in the House in April 1983 and in the Senate in April 1984. -Recent budget cuts have abolished the planning office of ICST, leaving in doubt the question of whether these forecasts will be done in the future. Accounting Office (GAO) has general auditing power over all government expenditures and has repeatedly evaluated office automation acquisition programs. The Presidents Private Sector Survey on Cost Control (the Grace Commission), which was highly critical of government management procedures, asked, Can improvement of information systems create cost savings and efficiencies and facilitate managerial decisionmaking throughout the Federal Government? 23 The report concluded that it could, and that this offered the opportunity for savings and revenue of $15.2 billion over 3 years, an estimate that included both office-microelectronic equipment and large systems. The survey team said that the acquisition process was inadequate, characterized by excessive procedural steps, a confusing array of policies and directives, lack of qualified personnel, and deficient training and supervision. The survey team recommended stronger, centralized, govemmentwide policies for information-technology acquisition and management, but not necessarily less discretion for the agencies, although it is not clear how both objectives can be preserved. The Administration has stressed the importance of planning, and OMB, GSA, and the National Bureau of Standards together prepare a 5-year plan, updated every 2 years, together with guidelines to assist agencies in planning. A new OMB circular, 85-12, will provide agencies with further guidelines to be used in planning. Agency managers, however, often express a feeling of futility in doing long-range planning for information systems because their budgetary constraints are constantly changing. Nevertheless, the desire to link personal computers and word processors to mainframes and minicomputers, and to other devices, is pushing Federal agencies toward planning and coordination of equipment acquisition, probably more effectively than could be done by instituting government-wide requirements. The Executive Committee to carry out the survey was established by Executive Order 12369, June 30, 1982. See: Presidents Private Sector Survey on Cost Control, Management Office Selected Issues, vol. VII 1, Information Gap in the Federal Government, winter 1983.
PAGE 245
238 Automation of Americas Offices The Federal Inventory of Office Automation Expenditures for information technology for Federal agencies grew from about $10.4 billion in 1983 to $13.9 billion in fiscal year 1985, increasing 19 percent in the first year and 13 percent in the second. 24 This is a much larger growth than that for Federal expenditures as a whole. About a third of these expenditures are for defense-related agencies. The government has been spending about 1.4 percent of its budget on information technology; this is perhaps somewhat less than the rate of spending by private corporations in the services and manufacturing sectors, 25 in spite of the greater intensivity of whitecollar work in government (about 80 percent of the Federal work force are white-collar workers, compared to about 55 percent of the total civilian work force). The average length of time in service for Federal computers in 1982 was just under 7 years, and decreasing as old systems are replaced. The purchase of equipment (capital investment) accounts for only about 19 percent of these expenditures, compared to 36 percent for commercial information services (ADP, etc.). The rest is for equipment leasing or rental and for personnel costs. Federal policy has generally been to encourage purchasing rather than leasing; it is more cost effective, and the proportion of systems that are leased has declined from 36 percent in 1970 to 12 percent at present. 26 But some procurement specialists question this strategy since leasing would make state-of-the-art technology more readily available. .Office of Management and Budget, General Services Administration, and Department of Commerce, A Five-Year Plan for Meeting the Automatic Data Processing and Telecommunications Needs of the Federal Government, vol. 1, April 1984. Hereafter cited as Five Year Plan, 1984. Thomas G. Cody, How Senior Execs View Info Technology, Government Computer News, July 1984, p. 64. In 1960, over 80 percent of computer systems used in the Federal Government were leased; by 1970 this had dropped to 36 percent, from 1977 to 1983 it was about 9 percent, and since then it is rising again, to 12 percent in 1984. Five-Year Plan, 1984, pp. 1-3. In recent years, the largest absolute increases in expenditures for information technologies have been in the Department of Defense (DOD), the Department of Energy, and the Department of Health and Human Resources. But a number of small agencies that had lagged behind have had bigger percentage increases in order to catch up with the pace; for example, the Office of the U.S. Trade Representative (USTR), the Department of Justice, and the Securities and Exchange Commission. DOD is by far the largest user of office automation among Federal agencies with about 67 percent of the stock. DOD however also has the largest share of all Federal white-collar workers (about 40 percent of them). One measure of the extent to which an agency has automated is the comparison between its share of total Federal office automation, and its share of the Federal whitecollar work force. The ratio of DODs share of automation to its share of the white-collar work force is a modest 1.68 compared to 8.24 for the Department of State, 3.41 for the Environmental Protection Agency, and 3.34 for the National Aeronautics and Space Administration. 27 In 1982, the Federal Government had in its computer inventory about 13,667 major systems with over 20,000 central processing units. This inventory is not complete nor accurate. Moreover, GSA will no longer attempt to list systems costing less than $50,000, and has recently discontinued its tracking of communications use and costs. It is therefore impossible to say how much office automation equipment the government owns. In early 1983, the General Service Administration said that there are estimates of 82,000 word proces q Internat~onal Data Corp., op. cit. These percentages are for 1982-83; more recent figures are not available. GSA, Automatic Data Processing Inventory, April 1984. GSA guidelines require agencies to report all systems with a CPU but agencies nevertheless differ on their interpretations of the guidelines-e. g., some do not include word processing systems and some do. An on-line version of the inventory is being developed, but it will not include systems costing less than $50,000. GAO concluded in March 1985 that GSAs data base of the governments inventory of computer equipment has been inaccurate for some time. (U.S. General Accounting Office, Effective Management of Computer Leasing Needed to Reduce Government Costs, IMTEC-85-3, Mar. 21, 1985, p. 11 1).
PAGE 246
sors in government offices 29 but these esti mates were based only on old rules-of-thum b about governments share of computer pur chases. Most agencies are not sure how many personal computers or small word processor s they have. Accounting and inventory catego ries differ across agencies in how they catego rize leasing arrangements, paced payments etc 30 Agency IRM officials nevertheless com plain about surveys or audits aimed at clarify ing these questions; audits are seen as a heavy burden that detracts from more productiv e work Much of the office automation equipmen t now being bought is low cost; decisions abou t personal computers are in this sense not much different from decisions about desks and type writers. Over-elaborate regulations and con trols could be needlessly costly. Organization s need the freedom to experiment in order t o identify the most useful technology for them However, monitoring and inventory could b e done without affecting the range of choice Current procedures for inventory of office automation equipment make it difficult to assess the status, l;evel of capital investment, and rate of investment. Projections of future office automation, future costs, and future benefits are therefore unreliable, and effects hard to measure. This is a problem for agency planners and decisionmakers trying to assess the cost-effectiveness of automation.
PAGE 247
240 l Automation of Americas Offices Noncompatibility Some of the most difficult decisions about acquisitions strategies involve the value of compatibility of small systems and devices. Some Federal executives insist that it is most important to get office equipment that is useful now; diversity will help the agencies determine what kinds of automation are most successful. Noncompatibility is a minor problem, they say; by the time it becomes cost-effective to tie systems together, it will be time to buy new up-to-date equipment. The experience and learning gained meanwhile will carry over. This seems to be a minority position. The tying together is already going on, and for most agencies noncompatibility is certainly a problem. However, this problem may be an unavoidable cost of competitive procurement. Moreover, cumbersome attempts at coordination and control during the first few years of personal computers would almost certainly have greatly delayed the automation of Federal offices and put the government far behind the private sector in the pace of adoption. Because the government is a major market for office automation, it is often urged to force the development of industry-wide standards, or to develop its own standards. Many experts think however that since the development of voluntary standards is proceeding, it would be preferable in the long run for government not to force this issue by intervening more strongly. There appears at present to be little pressure for such intervention. Inappropriate Choices Many day-to-day problems come about because of the selection of equipment by managers who do not understand the mechanics or flow of work in their own offices. The support staff is often not consulted, although they could bring to the decisionmaking valuable information that is otherwise not available. Formal description of work procedures often bears only nominal relationship to the real process of moving a form, a letter, or a report out the door. How smoothly the movement proceeds is affected by a myriad of details from how the office furniture is arranged, to what else has to be done at the same time. Hardware, software, or auxiliary furniture can improve this workflow or it can disrupt it, depending on characteristics that may appear unimportant or irrelevant to those who are not actually doing the work. The Complexity of Options The procurement of networking technologies involves much higher costs and longer-lived systems than the choice between stand alone devices. Mistakes are more serious. Even in procuring simple systems, the procurement options are becoming complex. To get a telephone system for a new office, the Federal executive must now choose between buying and leasing. If the budget only has operating funds, not capital investment funds, lease financing is necessary .33 Privatization An agencys evaluation of its needs for computers and related equipment has also been complicated by confusion over Administration policy about contracting with private organizations for information-related services. OMB Circular A-76 requires the contracting out of services that can be performed more cheaply by the private sector. Many agencies have contracted for data entry or word processing to relieve the load on old systems and avoid the need for new systems. Some have become dependent on outside sources for training and support of office automation instead of developing the capabilities that they will need in the future. Recent studies of productivity factors resulted in revision of A-76 to emphasize 14 major categories of services in contracting out. These include ADP, data entry and keypunch, audiovisual, and mail and file services. Agencies are now required to consider three options internal performance, use of another government agency, and outside contractors. They Patrick J. Keogh, Chief of the Economic Analysis Branch of GSA, Deregulation is Challenge to Procurement Officials, Government Computer News, November 1984.
PAGE 248
Ch. 9The Automation of Federal Government Offices l 241 are still encouraged to contract out as many professional procedures and the increasingl y services as possible, if those services woul d demanding safeguards and checks imposed b y require more than 10 Federal employees 3 4 Congress to ensure fair competition OMB recently reported that cost comparisons are now going against privatization in about half of the activities reviewed, a much higher percent than in the past. This has been advanced as an indicator that government i s becoming more efficient in carrying out it s activities The Dilemma of Procurement Policy In their own audits and reviews, agencie s frequently identify problems such as thos e GAO listed above. For example, the Inspec tor General of a major agenc y 36 in a 198 5 post-installation review of major system s listed a number of failures: no clarification over who has control over the project, no determi nation of total-user requirements, poor plan ning and design, lack of coordination between two user groups, and acquisitions occurrin g at field level, resulting in incompatibility. Mistakes of this kind can and should be avoided as better procedures are developed But at the same time, for larger systems, agency procurement procedures may be too elaborate and too rigid. 37 For advanced Systems, including the computer networks that are becoming a high priority goal in most agencies, the procurement procedures often resul t in specifications that are precise, elaborate and rigid. This creates major problems. Th e agency contracts officer finds himself caugh t in a tension between the agency or progra m officers need to solve old problems or achieve new goals with advanced technology on th e one hand, and on the other hand, customar y Eric Fredell, OMB Restricts Contracting Out in A-76 Revision, Government Computer News, November 1984. 5 Myron Struck, Workers Break Even in Cost Studies, The Washington Post, Apr. 11, 1985, A19. This was typical of audits and reviews shown to OTA by agencies; there is no intent here to single out a specific agency for criticism or otherwise, and the agency is therefore not identified. Based on discussions at workshops for Federal agency office automation specialists held by OTA in October 1984 and July 1985, and on many other written and spoken communications from Federal officials. To guard against violations of established safeguards, the contracts officer tends to insist on elaborate specifications before request ing bids. As a result, the development of the technology may outrun the procurement cycle, and equipment may be behind the state of-the-art by the time it is installed. Vendor s have no opportunity to propose alternativ e specifications that could provide more inno vative ways of meeting agency needs Overspecification is particularly likely to occur with information technologies because con tracting officers usually are not experts in th e technology and are unable to rely on their own professional judgment about how detaile d specifications must be. Contracting officer s are themselves at severe risk if there are too many challenges to their actions on ground s of inadequate competition and this make s them even more cautious 3 8 An additional complication with competitiveness in regard to information-technolog y procurement is the desire for compatibilit y with existing equipment. If the procuremen t is decided on the grounds of compatibility, i t may be faulted for noncompetitiveness; i f greater competition results in procurement s that require additional high expenditures t o compensate for noncompatibility, the decisio n may also be faulted Overspecification, according to industry experts, can result in procurement specifications that reflect obsolete approaches to technica l problems, or specifications that no vendor can exactly meet, although several vendors ma y have alternative approaches that would solv e the technical problem, perhaps at less cost than is entailed in meeting the government speci fications Federal contracts officers have a warrant, giving them sign-off authorit~ on contracts up to a specific dollar limit. If there are serious challenges to their decisions about competitive contracts they may not only get a poor performance evaluation, but can lose their warrant, or have their sign-off authority reduced, which would seriously damage their future career outlook.
PAGE 249
242 l Automation of Americas Offices In some cases, attempts to make procure policies that give agency managers wide discrements more competitive and to increase ac tion in formulating acquisition policies, has kept countability result in procedures that seriously pace with private sector office automation, and obstruct the accomplishment of agency responthat the opportunist y for increased productivity sibilities. will not be unduly limited by procurement-rein spite of these problems, OTA concludes lated problems. that office automation, under Administration WHAT WILL OFFICE AUTOMATION MEAN FOR FEDERAL OFFICE PRODUCTIVITY? Nearly all observers agree that Federal office automation is increasing agency produc tivity. But because there is no agreement o n the definition of productivity, this benefit i s hard to measure and document Based on past experience and analogy wit h the private sector, the gain in productivity i s probably far below that which is eventuall y possible. A mixture of old and new technol ogy is rarely completely effective. Worker s using the new technology must cope with ex isting paper-based data, through procedure s built up around older work activities, coordi nating with other offices that have not bee n automated, in a bureaucratic structure orga nized to fit the earlier patterns of workflo w and task sequencing, and in a workplace envi ronment that was not planned with the ne w technology in mind Moreover, the training that is given to workers using the new technology is clearly inade quate. Most of them are learning on their own, and from each other. This takes time to ac complish, and also cuts into the time that i s spent in direct production of output, both fo r learners and teachers. Managers, too, are stil l struggling to learn new techniques for man aging the automated office. (Federal trainin g is also discussed in chapter 3. ) Evaluations of Productivity The General Accounting Office estimated the cost of the government work force as $81 billion in 1980 and on that basis calculated a potential savings of $12 billion, or 15 per cent, through the use of office automation 39 It is not clear how this estimate was calcu lated, but the 15 percent estimate recurs fre quently in agency projections of increased pro ductivity The first round of agency reviews of thei r information-resources management, as re quired by the Paperwork Reduction Act, be gan in 1982. These reviews varied in scope and objective; some were designed to assess sys tem performance, some evaluated the effec tiveness of systems in direct support of spe cific program areas, and some concentrate d on the effects on administrative responsibili ties and budgets 4 0 The reviews were intended to address progress in reduction of the paperwork burden, im proved delivery of services, elimination of ac tivities that duplicate private sector sources budget savings, improved productivity, im proved technology, and improved managemen t controls. But few cited specific productivit y gains that could be quantified. These statu s reports illustrated why it is difficult if not impossible to specify, either in advance or in retrospect, the direct productivity y benefits of office automation. First, information resource s management goes beyond office automatio n AS quoted by the U.S. General Services Administration, Office of Information Resources Management, Managing End User Computing in the Federal Government. OMB then targeted and monitored 66 of the reviews, being conducted by 26 agencies. Executive Office of the President, Office of Management and Budget, Improving Government Information Resources Management: A Status Report, March 1983. Hereafter cited as EOP/OMB 2. See also EOP/OMB 1, cited previously.
PAGE 250
Ch. 9The Automation of Federal Government Offices 243 and some of the gains come from reorganization or increased management attention. Sec ondly, in most cases, automation is not a one time, self-contained event and comparison o f clear-cut before and after measurements ar e not possible Many agencies have provided OTA with sys tem plans and evaluation reports in which they made estimates or projections of productivit y gains from office automation. For example a natural resource agency recounts cuttin g staff in one field office from 370 to 120 when much of the work was automated by a ne w computer system. There are also many esti mates of future savings. A science-oriente d agency estimates that 10 percent of the wor k hours of administrative personnel and 3 per cent of those of its scientist s will be saved Another agency expects time savings of 1 0 percent for administrative personnel and 2 0 percent for professionals, and the eliminatio n of some technical positions. These estimate s are persuasive in context, but they use so many different ways of estimating or aggregatin g that they cannot be evaluated or compared Although not called out as a criteria fo r evaluation in the status reports, there appea r to be many cases where information technol ogy has resulted, or can result, in services to the public that would not be possible withou t it 4 2 But in the interest of reducing the Fed eral deficit, there has been more emphasis on attempts to reduce or constrain the growt h of government services than on expandin g them OMB, while specifying productivity as a n evaluation criteria, has not defined it nor pro vided a metric for it. OMB says only that with the technolog y the Federal Governmen t should be able to utilize available resource s in a more efficient and effective manner. 4 3 Fifteen percent of the 20 percent of their time that the scientists spend on administrative duties. For example, the automation of the Occupational Safety and Health .Administration technical Data Center allows 0S11A to respond yearly to several thousand inquiries by go\rernment agencies, workers and their unions, employers, public interest groups, and the general public. lEOP 021111 1, p. 12. Productivity could mean more or better work performed or services delivered; it could mean doing the same work at less cost The Merlin Experiment One example of productivity assessment is an Office of Personnel Management (OPM ) evaluation of the General Service Administra tions Merlin System, installed in GSAs West ern Region in July 1983 44 Merlin is compose d of 84 personal computer workstations linke d by a local area network and used by mana gerial, technical, and senior clerical employ ees. The OPM study was based on systemati c measurement and evaluation of work befor e and after the installation. In 39 of the 47 categories of tasks studied, output increased re markably: OPM rated this a strong overal l improvement The volume of output per em ployee hour went up in 14 of 19 categories Cost per unit of output was reduced in 79 percent of the categories. In some categories i t was possible to compare the same kinds of output with and without use of Merlin, in the same time period; 10 of the 17 categories showe d improved unit time with Merlin. OPM als o found a reduction in contract preparation time improved time and project management, an d a marked reduction in time spent in meet ings, due to the systems communications ca pabilities. OPMs figures appear to indicat e about a l-percent cost saving for the organiza tion as a whole (across about 4.4 person year s spent in all tasks done by the organization) .45 No information is given about the cost of Merlin or its anticipated lifetime Incentives In the private sector, increasing the output and/or the quality of the product should increase the organizations market share an d ultimately its profits. This incentive does no t U.S. Ge~eral Services Administration, Office of Information Resources Management, Merlin Impro\ements, January 1985. App. D of the cited report gives total time and total cost for sample 1 (preMerlin) as 9,200.6 hours and $132,160,10 or S 14.36 per hour; for sample 2 (post-Merlin) as 9,763.0 hours and $138,731.30 or $14.21 per hour.
PAGE 251
244 l Automation of Americas Offices operate in government. On the contrary, a government organization that performs its wor k at less cost typically finds its budget reduced for the following year, usually with no reward for the managers who achieved the cost re duction. Instead, if they have fewer employ ees they may lose status and find their chances of a higher grade level significantly prejudiced .46 The always present possibility of across-the board budget cuts encourages Federal execu tives to keep staff levels above the minimum that is necessary so that cuts can be absorbed For example, a 1980 U.S. General Accounting Office report concluded that productivity rates achieved by Federal payment centers varied by 600 percent, and added that the GAO auditors were told by payment center managers that there were strong disincentives against raising productivity, of the kind described here. GAO, Improving the Productivity of FederaJ Payment Centers Could Save Millions, FGMSD-80-13, Feb. 12, 1980. without damaging their ability to get the work done. New York City undertook to automate its municipal offices under a strong drive to re duce costs and increase government revenues 47 and high level officials say that progress is being made toward these goals. The sam e pressures are operating within the Federa l agencies, and there is every reason to antici pate that the results will be much the same Productivity as measured by output and b y labor costs is likely to increase; whether thi s is accompanied by improvement in the qual ity of government services is much more un certain According to a case study of municipal office automation done for OTA; see app. B for a summary. IMPLICATIONS FOR THE FEDERAL WORK FORCE If office automation does increase produc tivity in Federal offices, it could allow Fed eral whit~collar employment to be reduced (o r it could allow government services to be ex panded without a proportionate increase in em ployment). Automation may also : l change the nature of the work and the skill requirements of some jobs, l effect the number of people needed in some job categories l create new jobs and occupations, an d make other jobs obsolete Such effects would change the relative number of people needed in the major categorie s of Federal jobsprofessional, administrative technical, and clerical; and also affect the dis tribution of people at each grade level and salary level 48 Such changes in the structure o f the Federal work force, as well as the size o f the work force, will strongly affect total labo r costs and thus the cost of government .-. In the Federal civil service, employees hold a grade level in one or the other of approximately 100 pay plans used by the executive branch. About two-thirds of them are covered by the largest pay plan, the General Schedule (GS). This has grade levels from 1 (the lowest) to 18. The employees grade level determines the range of salaries that he or she can obtain. These potential changes should be considered in work force projections, in planning, and in developing policies about personnel recruit ment, rewarding and promoting, retention, and retirement. There is no indication that OPM OMB, or other central agencies of governmen t are studying the implications of such changes Since Federal offices have been using computers for about 20 years and have been rapidly automating over the last 5 to 10 years, indicators of such change should already be apparent with careful statistical analysis. This analysis has not been done on a systemati c basis. However, OTA concludes that there ar e at least some general indicators that suc h changes are underway Size of the Work Force If office automation is affecting Federal employment levels, then historical growth rate s should be slowing or reversing. Growth of th e Federal work force has in fact slowed, stabi lized, and then been reversed over the pas t 15 years. It is not necessary to argue that of fice automation has been the cause of this de cline. The volume of government employmen t
PAGE 252
Ch, 9The Automation of Federal Government Offices l 245 depends strongly on the political philosophy of the Administration in power and of Congress about the scope of Federal Governmen t responsibility and the services that govern ment should provide. It is also strongly af fected by population growth and economic ex pansion. However, given apolitical imperativ e to reduce the cost of government and henc e to shrink Federal employment, it can be ar gued that office automation has at least al lowed and facilitated a reduction of the work force Federal employment has grown every decade of our history except the 1920s. (See fig ure 9-1. ) It has not, however, kept pace with growth in population or in the economy. Fed eral employment grew by 73s percent durin g the 1930s and the New Deal, 88 percent dur ing the 1940s and world war, 23 percent i n the 1950s, and 24 percent in the 1960s 50 Th e Federal employment has steadily decreased relative to the size of the economy and the national population, from 3.7 percent of U.S. civilian employment in 1960, to 2.9 percent in 1982, and from 13.1 (executive branch) Federal workers per 1,000 population in 1960, to 12.2 in 1982. The Federal proportion of total public sector employees has declined from 33 percent in 1950 to under 18 percent in 1980 (U.S. Department of Commerce, Statistical Abstracts) Bureau of the Census, Historical Statistics of the United States, Part II, Series 308-317. Figure 9.1 .Paid Civilian Employment of the Federal Government, 1881-1983 1881 1891 1901 1910 19201930194019501960 197019801991 Year SOURCES Data from 1881-1960 based on U S Department of Commerce Bureau of Census li~s(orical Abstracts of the U S Co/on/a/ T/rnes to 7970, fart // Series y 308-317 Washington DC 1975 Data from 1970-1983 based on U S Department of Commerce Bureau of Cen sus Stat(st/cal A bsfracfs of fhe Um fed States. N 523, Washl n gton DC 1985 growth in the 1960s was perhaps in part related to the Vietnam War, but it was charac terized by increases in the number of posta l workers and the number of general adminis trative, clerical, and office service workers in nondefense agencies 51 Thus, it was probabl y related to the trend toward an information intensive economy and to political initiative s such as the War on Poverty But from 1970 to 1980, Federal employment growth slowed; it grew only 2 percent in that decade. During this period the Federal wor k force waslike the countrys general labo r forcebecoming more thoroughly white-collar The ratio of white-collar to blue-collar Federa l workers changed from 2.9: 1 in 1960, to 4:1 in 1980. Yet at the same time the growth of th e Federal whit~collar work force has also slowed; it grew more than 13 percent from 1960-6 5 and again from 1965-70, but only 7 percen t from 1970-75 and 3 percent from 1975-80. This change was at least coincident with the speed ing up of office automation, which began about 1974 Every 2 years from 1960-74, the number of General Schedule and Merit Pay employees (about two-thirds of all Federal white-collar employ ment) 53 grew about 3 percent; from 1974-84, the average growth rate each 2 years was only 0.3 percent. This may also be an indicator that office automation was having a n effect Such indicators do not of course establish a cause-and-effect relationship. Nevertheless the cumulative evidence that automation o f white-collar work can at least allow employ --Department of Defense employment increased 15 percent from 1960-70, but it was 43 percent of total Federal ci~ilian employment in 1960 and only 41 percent in 1970. (Stat. Abstr. 1984, table 535.) But General Administrative Clerical, and Office Services increased 20 percent and postal workers increased by 26 percent. Stat. Abstr. 1984, table 535, p. 333. There are over 100 pay codes or schedules for Federal executive branch employees, although some co~er only a handful of workers. A few, such as those used only for Foreign Service employees, the Tennessee Valley Authority, or the Veterans Administration, co~er thousands of employees. However, the GSGM pay codes cover about 67 percent of all white-collar workers, and a higher percentage of those who are administrative or clerical. 52-641 0 85 9
PAGE 253
246 l Automation of Americas Offices ment growth to be constrained or reverse d while the amount of information-handling in creases, is persuasive Changes in the Mix of Federal Jobs There are now 8 percent fewer people in the jobs grouped under General Administrative, Clerical, and Office Services than there were in 1975 54 A closer examination of this group which contains about one-quarter of all Fed eral workers, also suggests that office auto mation is bringing about basic changes in Fed eral office work and in Federal jobs For example, four job titles within this grouping have disappeared since 1975 55 coldtype composing machine operation, dictatin g machine transcription, electric bookkeepin g machine operation, and calculating machin e operation. Together they accounted for 8,50 0 jobs in 1975, although even then they wer e on the way to becoming obsolete jobs. In 1975, there were already 42,500 computer operators specialists, and aides; that number has grown by 9,000 5 Some job titles have changed as the technology associated with them changed-for ex ample, card punch operations has become dat a transcription. The number of people who pre sumably spend much of their time typing (now known as keyboarding) may have grown de spite the spread of word processing. In 1975 there were over 7,146 dictating machine tran scribers (a classification that does not no w exist), 72,895 clerk-typists, and 62,373 secre taries. By 1983, there were 11,780 fewer clerktypists but 29,100 more secretaries, a 7-percen t U.S. Office of Personnel Management, Federal Civilian Work force Statistics: Occupations of Federal White-Collar and Blue-Collar Workers, MW 56-18, Oct. 31, 1983, table E, and U.S. Civil Service Commission, Federal Civilian Manpower Statistics: Occupations of Federal White-collar Workers, SM 5611, table F-1. Postal workers have been removed from the 1975 data to make it comparable to 1983 data. The only entirely new job title is equal opportunity specialist, although as noted the title of some job classifications (as identified by code number) has changed. The extent to which people displaced from eliminated jobs move into newly created jobs cannot be answered from this data. However, it appears from OPM print-outs that most of the over 7,000 dictating machine transcribers were women; most of those in computer specialties are men. overall increase. Secretaries, who have mor e diverse duties and make higher salaries tha n clerk-typists, now make up 60 percent of thi s group of workers, as compared to 44 percen t in 1975 Shifts Among Occupational Categories and Grade Levels These changes in and among job classifications imply that there will also be changes in the shape of the Federal work forcethe dis tribution among major occupational catego ries (professional, administrative, technical and clerical) and across grade levels and sal ary levels. Such changes have occurred in th e last decade, although undoubtedly many force s are operating to affect this trend. The distri bution in terms of the major occupational cat egories is squaring up. The proportion of cler ical workers is declining, and the proportion s of technical, administrative, and professiona l workers have been growing (although the shar e made up by professionals is now beginning to shrink slightly). (See figure 9-2. ) Clerical Workers In October 1983, about 25 percent of General Schedule (GS) employees were clerical workers. In 1975, more than 31 percent were clerical workers. The figures for both years exclude postal workers (who were removed from OPMS occupational survey between 1979 and 1981); with them included, more than 45 per cent of white-collar employees were in the clerical grouping in 1975. Ninety percent of Fed eral clerical workers in 1983 were in grad e levels 3 through 6, with another 3 percent in grades 1 and 2 57 As a reference point, there were over 305,000 Federal clerical workers in grades 1-6 in October 1984. At the high end of the scale are secretaries and claims, payroll, legal, and travel clerks and statistical assistants; average grades for these job classifications are between 5 and 6 (roughly $16,000 to $17,000 in 1983) and there are about 110,000 people in such jobs. At the low end of the clerical scale are messengers, clerk-typists, mail and file clerks, and computer clerks, with average grades of 2-4 ($12,000 to $14,000). The breakdown of workers in each category by grade level is based on an OMB computer printout as of Mar. 31, 1983, while the total numbers for each category are for October 1983. This should not, however, cause any major misstatements.
PAGE 254
Ch. 9The Automation of Federal Government Offices l 247 Figure 9-2. Changes in the Distribution of Federal Employees by Occupation, 1975 and 1983 1975 Technical (1 80/0) Admlnl strative (24 %) professional ( 19) Other (8 O.) 1983 Technical (22 O/. ) Adminlstratlv e (28 O/. ) Clerlcal (31 0 /0) Clerical 25%) 0/0 ) Professional (22/0) SOURCES 1975: U S CIVII Service Commlsslon, Federal Ciwlfan Manpower Sfaf~stfcs Occupat~or?s of Federa/ kVh/te.Co//ar Workers, Oct 31, 1974 and 1975, SM 56.11, Washington, DC, 1975, and 1983: U S Off Ice of Personnel Management, Federa/ C/v///an Workforce Stat/sf/cs Occupations of Federa/ Wh/te-Co//ar and B/ue-Col/ar Workers, Oct 31. 1983, MW 56-18, Washington DC, 1983 In internal Federal agency documents an d in anecdotes told by Federal office automa tion managers, it is frequently said that gov ernment clerical workers, once trained in th e use of office automation, are often hired away by private sector industries that offer higher wages. This is alleged to be true especially in field offices in some sections of the country This disparity is largely at lower salary levels; 34 percent of Federal data processing workers made under $23,400 in 1983, compared to 22 percent of those in the private sector, ac cording to figures derived by the Congressional Budget Office. However, 35 percent of Fed eral data processing specialists earned $35,400 or more, compared to 33 percent of those i n the private sector s U.S. Congressional Budget Office, Reducing Grades of the General Schedule Vtork Force, September 1984, p. 43. Technical Workers About 22 percent of GS employees in 1983 (nearly 332,000) were technical workers o r paraprofessionals with such diverse titles a s economics assistant, communications special ist, tax technician, and equal opportunity as sistant titles in all. In 1975, only 1 8 percent of the Federal workers had techni cal ratings. This trend probably reflects th e development of an information society in which scientific and technical information is a ma jor factor in all economic activity. Administrative Workers Another 28 percent of GS workers in 1983 (432,000), were classified as administrative compared to 24 percent in 1975. Nearly 11 percent of all Federal white-collar workers are supervisors; just over 2 percent are called man agers, or top-level executives. (The latter ar e usually in the professional rather than th e administrative category. ) 61 About 90 percen t of administrative workers in 1983 were lo w and middle-level managers in grades 9 through 15, with nearly another 10 percent just below in grades 5 through 8 62 Program managers health systems administrators, and computer systems administrators, for example, hav e average grade levels of 13 or over. Passpor t and visa examiners and social insurance claim s examiners are at the lower end of the rang e and have grade levels under 9. Again, it is the administrative classifications with lowest grad e levels that appear most likely to be reduce d in numbers as some of their tasks are auto mated, if the experience of their closest coun terparts in the private sector, insurance raters and underwriters, can be generalized U.S. Civil Service Commission, Federal Civilian Manpower Statistics: Occupations of Federal WhiteCollar Workers, Oct. 31, 1974 and 1975, SM 56-11, pp. viii and x. The 1975 figures have again been revised to exclude postal workers. See table B of the CSC publication. The grade levels for technical jobs vary widely (the 90 percent range is from GS 4-11) but the a~erage grade is 7.14 (a\erage salary $21,000 in 1983]. U.S. Office of Personnel Management, Federal Civilian Work force Statistics: Occupations of Federal White-Collar and Blue-Collar Workers, Oct. 31, 1983, MW 56-18, p. 8. ( The average grade for administrators is 10.9 (average salary $32,000 in 1983).
PAGE 255
248 Automation of Americas Offices Professionals Just over 22 percent of Federal workers were professionals in 1983, as compared to 19 percent in 1975. Again, this increased proportion is probably related to changes in American society and economy as well as to office auto mation. In 1983, 94 percent of professional s were in grade levels 9-15. Their average grad e level was 11.7 with an average salary o f $36,000. (In 1983-84, however, the proportion of professionals in the Federal work force wa s reduced slightly from that in 1980. ) 6 3 Changes In and Among Grade Levels These changes have squared up the Federal structure, viewed in terms of occupational categories, so that their proportions, which i n 1974 ranged from 18 to 31 percent, are no w more nearly equal, from 22 to 28 percent, with administrative rather than clerical occupation s as the largest group. Paralleling these shifts, the number of workers at each grade level ha s also changed. (See figure 9-3.) In 1974, 41.9 percen t 64 of Federal workers were in the lowe r grades 1-6, where clerical workers are concen ..-. About 3 percent of Federal workers are usually classified as other or unspecified, usually because their jobs are being reclassified during the counting period. The figure for both years excludes postal workers; if they are included for 1974, the percentage becomes 41.6 percent. Figure 9-3.Change in Distribution of Federal Employees by Grade Levels, 1974-80 5 1 4 3 2 1 o 2 :; ~ GS 1-6 GS 7-8 GS 9-14 GS 15+ Grade level SOURCES: u.: +wil Service Commission, federal Civl/ian Manpower Stafh3tics: Occupations of Federal White-Co//ar Workers, Oct. 31, 1974 and 1975, SM 56-11, Washln@on, DC, 1975; and U.S. Office of Personnel Management, Federa/ Civi/ian Workforce Statistics Occupations of Federal White-Co//ar and B/ue-Co//ar Workers, Oct 31, 1983, MW 56-18, Washington, DC, 1983 trated. By 1980, it was 38.6 percent. The percentage of workers in grades 7 and 8 remained much the same (11.4 percent in 1974, 11.5 per cent in 1983). But the proportion in grade s 9 through 14, where most of the professional s and supervisory administrative workers ar e concentrated, has grown from 45.1 to 49.4 percent 65 (Congressional Budget Office (CBO) taking slightly different parameters, says tha t the percentage of people in middle-manage ment grade levels 9 through 15 has grown from 33 to 37 percent since 1974. ) 6 6 The establishment of the Senior Executiv e Service (SES) was intended to increase th e retention of senior Federal executives by pro viding them with the opportunity for highe r rewards (bonus pay) in return for some sacri fice in job security. It is generally agreed tha t it has not worked well; many SES people believe that they have been made more visibl e and therefore the particular target of reduc tions in force (RIFs) and salary caps or freezes, and most of the few bonuses have gone to a relative few of the highest ranking people 6 7 People tend to think of bureaucracies a s shaped like a pyramid. The Federal work forc e is shaped more like a very flattened figure 8 or hour-glass. There are relatively few peopl e in the lowest grades, a large bulge at grade s 5 and 6, a narrow waist representing grade s 7 through 10, another large bulge at grade s 11 and 12, and a rapid drop-off in numbers in grades 13 through 15. The number of positions at grades 16 through 18 and in the Senio r Executive Service (equivalent to grades 1 6 through 18) is limited by law and is less than 0.5 percent of all Federal jobs. (See figure 9 4.) The figure eight is getting squashed down the bulges growing wider, and the top and bottom flattened -. 65 Figures supplied by OPM in advance of the 1984 analysis of Federal civilian work force statistics now in progress, and compared to OPM occupational statistics for 1974 and 1980, see previous cites. U.S. Congressional Budget Office, op. cit., p. x. Sar Levitan and Alexandra B. Noden, Working for the Sovereign: Employee Rdationsin the Federal Government (Baltimore: The Johns Hopkins University Press, 1983), pp. 94ff.
PAGE 256
. Grade level GS 15+ GS 13-14 GS 11-12 GS 9-10 GS 7-8 GS 5-6 GS 3-4 GS 1-2 GS 15+ a GS 13-14 GS 11-12 GS 9-10 GS 7-8 GS 5-6 GS 3-4 GS 1-2 I 400 Ch. 9The Automation of Federa/ Government Offices l 249 Figure 9-4. Distribution of Federal Employment by Grade Level, 1975 and 1983 1975 I I I I I I I I I 300 2!0 2!0 150 100 5 o 50 1 JO 150 200 2!0 3! 0 1983 I 1 I I 1 111111 1 111111 1 350 300 250 200 150 100 50 0 50 100 150 200 250 300 350 Employees (in thousands) The ~r~d~~ 16.1s excluded senior executive service empkwees 36,902 165,930 296,800 178,712 165,538 273,454 282,406 28,745 39,883 181,163 361,904 189,316 170,512 289,732 188,666 12,425 I 400 SOURCES 1975: U S Civil Service Commission, Federal Civilian Manpower Sfati5tlcS. Occupations of Federal Whlfe-Col/ar Workers, Oct 31, 1974 and 1975, SM 56-11, Washington, DC, 1975, p 34 1983: U S. Office of Personnel Management, Federal civilian Workforce Statistics Occupations of Federal White. Co//ar and Et/ue-Co//ar Workers, Oct 31, 1983, MW 56-18, Washington, DC, 1983
PAGE 257
250 l Automation of Americas Offices Unanswered Questions The Question of Overgrading The average grade level for the civilian work force, not surprisingly in view of changes described above, increased from 8.03 to 8.51 between 1974 and 1983. As a result, salaries increased 4 percent (in equal dollars) and about 1.3 billion was added to payroll costs. 68 The Presidents Private Sector Survey on Cost Control, and others, have charged that the Civil Service is overgraded, meaning that grades are too high in comparison with earlier levels and with the private sector. 69 The Congressional Budget Office (CBO) concluded that the growth in the proportion of nonclerical jobs accounted for about two-thirds of the half-grade increase. They attributed most of the additional increase to promotions granted by some managers to compensate for repeated caps. 70 An earlier OPM analysis (1980) said that 60 percent of the half-grade rise was due to changes in the occupational mix of the workforce, citing the disproportional decrease in the clerical work force. 71 Neither the Private Sector Survey report, OPM, nor the Congressional Budget Office explicitly addressed the effect of the increasing professionalization of government work nor of the penetration of information technologies in the office. CBO did acknowledge the changing character of governmental work in identifying the factor of growth in nonclerical jobs, 72 but did not explicitly raise the question of how the declining proportion of clerical jobs might be related to office automation, U.S. Congressional Budget Office, op. cit. The Administration has proposed reducing the number of positions graded GS 11-15 by shifting 8 percent (about 40,000 jobs) to lower grades; this would be done over a period of 4 years, as vacancies occur. This would save $3.9 billion over a 5-year period. See U.S. Congressional Budget Office, op. cit., p. xi. [R. Mark Musell, Reducing Grades of the General Schedule Work Force, LRS84-12344 (Washington, DC: U.S. Congressional Budget Office, 1984). Average Grade Trend 1974 -1980, prepared by James Hall, Program Management Information Section, Work Force Information Division, U.S. Office of Personnel Management (manuscript, unpublished). Ibid, p. x. nor did it point out that the Federal work force is far more thoroughly white collar than the private sector work force. CBO also pointed out that while the government has more middleand top-ranked workers than the private sector, it also pays them much less; Federal and private sector salary distributions were almost identical in proportion going to people at those levels. CBO noted that only 43 percent of Federal jobs were, at that time, below grade 11 compared with 61 percent in equivalent levels in the private sector. Again, it did not specifically note the greater proportion of white-collar work in the public sector. Neither CBO, nor OPM in its own analysis, have given attention to the implications of office automation for future work force size, grade distribution, and payroll costs. Neither has systematically analyzed the forces behind the changing character of governmental work, nor acknowledged that these forces are likely to grow stronger in the future. There appears, in fact, to have been no major executive branch study of the possible effects of office automation on the Federal work force or labor costs. Considering the current emphasis on cutting the costs of government, the high priority given to increased productivity, and the strong trend toward procuring information technology to accomplish these goals, this is surprising. Whether the potential increase in productivity because of office automation, and the consequential possibility of shrinking the size of the work force, will more than compensate for the need for more highly paid technical and professional workers is a question that should be studied thoroughly. To both reduce the clerical work force and artificially hold down average grade level and salary could mean that the government cannot attract and hold the people it needs for excellence. The lack of systematic analysis and planning for changing work force needs makes some Federal workers anxious about their jobs. In discussions and workshops with Federal clerical employees concern is often expressed about what it may mean for job security and
PAGE 258
Ch. 9The Automation of Federal Government Offices 251 the opportunity for promotion. Some of this concern may be unrealistic and exaggerated; it could be alleviated if officials could talk with their workers about current effects and the implications for the future. They cannot honestly do that at present because little or no information is available about the effect that office automation is now having on Federal employment. How Jobs Are Classified As discussed in chapter 4, office automation changes the nature of work and the skills requirements for many jobs. In both the private sector and the public sector, this leads to an expectation that job descriptions and personnel classification schemes will be revised. OPM uses a point/factor rating system in assigning grade levels. 73 Brief definitions of the factors do not mention any consideration of equipment that is used in the work. The first and most heavily weighted factor is knowledge required, defined as the nature and extent of the information or facts that the worker must understand to do acceptable work. It is not clear whether more knowledge is required to operate a word processor than to operate a typewriter, or whether using a spread-sheet program takes more or less special knowledge than using a bookkeeping ledger. In some cases automation requires less knowledge about the process being carried out, in other cases it requires additional abstract thinking or ability at problem solving. It appears that Factor No. 1 could be interpreted so as to accommodate necessary technological skills, and thus jobs could be reevaluated and their grade level moved up or down, if their nature changes because of automation. .. The factors are: 1. knowledge required hy position { 1,850 points, 41 percent of maximum total), 2. supervisory control (650 points), 3. guidelines (and judgment needed to apply them) (650 points), 4. complexity of work (450 points), 5. scope and effect of work product (450 points), 6. personal contacts with nonsupervisors (220 points), 7. physical demands (50 points), and 8. work en~ironment (.50 points). The point range for grades 1-6 is 190 to 1,350 points; for grade 9 1,855-2, 100; for grade 15,055 and up. The maximum total is 4,480 points. Unless this is explicitly recognized, however, it is likely that there will be challenges to the factor evaluation system, especially if workers believe that they are mastering new skills not recognized and explicitly valued in their old job descriptions. Catherine Waelder, General Counsel of the National Federation of Federal Employees, has already said that the factor system is new but its standards and values are old. She points out that if not continuously revised, the system will fail to account for the ability needed to use developing technological resources essential in many professions. . 74 Part-Time and Temporary Workers Government use of part-time and temporary workers is determined by many factors; it may or may not increase because of office automation. The Federal Employees Part-Time Career Act of 1978 was passed after several years of effort by members of the Senate and House of Representatives who were responding to urging of female constituents wanting to protect career opportunities in government while reserving time for at-home child care. Unions objected on the grounds that it would take jobs needed by over three million women actively seeking full-time employment who could not afford to support their families on parttime employ ment. 75 The act required agencies to do feasibility studies, establish annual goals and timetables, and review and monitor part-time work opportunities. Part-time Federal employment did increase by almost 14,000 from 1979 to 1981, to a high of 57,184; but in 1981 in the strong effort to reduce Federal employment, part-time workers were laid off at a higher rate than fulltime employees, and by 1984, the number of part-time workers had declined by roughly 2,000. In a 1982 review of the Part-Time Career Act, GAO found that most agencies were not conforming to its requirements. 76 Statement in IIearing on 11. R. 4,599, Federal l]a~ I;quit.} Act of 1984, op. cit., p. 248. William G. J4hittaker, Federal Legislatitre Interest in Alternative Patterns of Itork: An O\er~iewr, U.S. Congress, I.ihrary of Congress, Congressional Research Service, Mar. 31, 1980. U.S. General Accounting Office, Part-Tinle ll~mp)o.~rn]ent in the Federal Gox-ernment, Jul~T 1 !2, 1982.
PAGE 259
252 l Automation of Americas Offices OPM is now again encouraging part-time and temporary employment on the grounds that it is cost-effective. 77 If office automation allows the work force to be reduced and kept lean, part-time and temporary workers are likely to be used to augment permanent staff when there is a short-term increase in the workload; thus temporary and part-time Federal employment is likely to increase as along-term trend. Since the objective of part-time work is from an employers perspective to allow rapid, flexible adaptation to changes in workload, the number of part-time workers will fluctuate widely .78 Women in the Federal Work Force Several bills now before Congress deal with questions of pay equity and comparable worth in the Federal civil service. Comparisons of job content and skill requirements will be made more complex as office automation changes the nature of the work. While many clerical jobs have traditionally been considered womens work, the degree of gender dominance differs significantly by grade level. (See figure 9-5. ) In the occupational category of General Administrative, Clerical, and Office Services, women predominate at grade levels 1 through 9, and men at grade levels 10 through 15. In grades 2 through 6 the percentage of workers who are female varies from 78 to 90 percent. 79 In grades 7 through 9, the proportion of women ranges from 71 down to 61 percent, and at levels 10 through 15 the proportion decreases with each level, from 41 to 10.3 percent. Thus, while office services jobs are thought of as a group as womens jobs, their grade level distribution is not much different from their distribution in the Federal work force as a whole for women. About 62 percent of female Federal employees are in grades 1 through 6, as compared to 20 percent of male employees. Only 10 percent of female Federal -. ----Since January 1985 agencies have been allowed to keep temporary workers for up to 4 years; the previous limitation was 1 year. Wee ch. 2 for a general discussion of part-time and temporary work as related to office automation. Grade 1, which is held by fewer than 150 workers (many of them messengers) is only 66 percent female. workers are in grades 14 and 15. As a result of their concentration in lower level jobs, women (who hold 46 percent of Federal General Schedule jobs) have an average salary that is 62 percent of mens average. 80 It is estimated that three-quarters of the lowest ranking Federal jobs (by grade level) are predominantly womens jobs. 81 Since office automation is most likely to bring about reductions in lower level Federal jobs, women workers are those most likely to be made redundant. This could lead to a statistical improvement in the position of women in the Federal work force, as a group, since their average grade level should rise as a result. This statistical improvement should not obscure the fact that the job losses will be disproportionately suffered by women. Quality of Worklife Again, the case study of office automation in New York City 82 offers some possible indicators of present and future effects on a civil service work force. In automated municipal departments, new tasks were learned that cut across traditional occupational definitions, but there was little or no evidence of job upgrading in clerical and technical (paraprofessional) positions. Neither old or new job ladders provide a ready path for these workers to move up in the hierarchy of jobs, and some rungs of the job ladder have been removed by automation of intermediate tasks and responsibilities. In some jobs, abstract and conceptual knowledge is becoming more important than traditional skills, but this change has not been accompanied by an increase in pay scales. This In the Federal Wages Schedule (blue-collar workers) women are only 9 percent of the total, and their average wages are 79 percent of mens average. Among all FWS and GS workers (83 percent of all Federal workers) women are 38 percent of all employees and their average salary/wages are 68 percent of mens average. U.S. General Accounting Office, Options for Conducting a Pay Equity Study of Federal Pay and Classification Systems, GAO/GGD-85-37, Mar. 1, 1985. ] Statement of the Honorable Michael I). Barnes of Maryland, Hearings on H.R. 4599, Federal Pay Equity Act of 1984, before the Subcommittee on Compensation and Employee Benefits of the House Committee on the Post Office and Civil Service, 98th Cong., 2d sess., pt. I, Apr. 3-4, 1984. Greenbaum, Pullman, and Syzmanski, op. cit., ch. 1,
PAGE 260
Ch. 9The Automation of Federal Government Offices c 253 Figure 9-5. Full-Time White-Collar Employment in the Federal Work Force by Grade Level and Sex, GS 1-3 1974-75 and 1983 I I 1974-1975 I I I I 1 I I I I I I I I I I ) 350 300 250 200 150 100 50 0 50 100 150 200 250 300 350 4 Work force (in thousands) 1983 I I I 1 I I I I I I I I I I I ] 350 300 250 200 150 100 50 0 50 100 150 200 250 300 350 4[ Work force (in thousands) o 10 SOURCES 1974-75: U S CIVII Service Commlsslon, Federal C/v///an Manpower SfaI/sfics Occupat/or7s of Federal kVh/te-Co//ar Workers, Oct 31, 1974 and 1975, SM 56-11, Washington, DC, 1975 1983: U S Office of Personnel Management, Federal C/v/l/an Work force Staf/sf/cs Occupations of Federal Whife.Co//ar and B/ue-Co//ar Workers, Oct 31, 1983, MW 56-18, Washington, DC, 1983
PAGE 261
254 Automation of Americas Offices is becoming a union issue and apolitical issue Labor Management Relations and in the city. Federal Unions The observations made on quality of worklife and the automated office environment in chapter 5 generally apply to the Federal office as well as to the private sector. Physical conditions are probably not as good in public sector offices as in many private sector offices. Turnover of facilities and office furniture is probably slower so that new technology is even more likely to be inserted into old settings. Furniture, lighting, and space layout is often inappropriate for microelectronic equipment now in use. One issue likely to be controversial is the use of computer monitoring and pacing of work. 83 Federal labor law holds that applying work measurements and using these to evaluate employees are management prerogatives and need not be negotiated with unions representing Federal employees. 84 However, agency management can if they so choose bring the unions into discussion of work standards and measurement. Unions have attempted to raise the monitoring issue at some agencies but these attempts have been rebuffed as nonnegotiable under the management rights clause. Arbitration rulings have upheld agency management when work monitoring was challenged, for example, in the Social Security Administration. See ch. 5, for a general discussion of computer monitorin g of work. Civil Service Reform Act of 1978, Code of Federal Regulations, title V, sec. 7106 (a) (A) and (a) (B), Management Rights. The discussion in this section on work monitoring, where not otherwise cited, is drawn from an OTA contractor report, prepared by the Educational Fund for Individual Rights, Alan Westin, principal investigator, February 1985. The report is based on site visits and interviews in seven Federal agencies between 1982 and 1984, supplemented by a letter of inquiry to 78 Federal agencies, bureaus, and departments in September 1984 to which 44 agencies responded with letters and accompanying documentation. Follow-up telephone interviews were held with 12 agencies that reported significant activity related to work monitoring. In the civil service, managers have less control over subordinates than in the private sector, because of civil service rules and protection. Managers themselves are subject to man)of the same rules, regulations, and restrictions on wages and on bargaining as are their subordinates. Many Federal managers and professionals belong to professional associations that serve much the same functions as unions; 85 for Federal employees, this means lobbying Congress rather than bargaining for pay and benefits, since these are set by legislative action. 8G Federal unions will not be able to mount effective resistance to reductions in employment as these become possible, but they may be expected to monitor changes in job content, job classifications, and compensation as a result of office automation. The position that the largest unions have taken and will take on such issues as stress, risks to health from CRT use, 87 These include the Federal Management Association, the Professional Management Association, and the Senior Executive Association (the latter open only to those in grades 16-18 and the Senior Executive Service). Until 1961, membership in unions by Federal workers was not recognized officially, and there was no provision for collective bargaining. However, about 13 percent of nonpostal workers and a much higher percent of postal and blue-collar workers had already joined. Between 1963 and 1969 the membership more than quadrupled, from 180,000 to 843,000. President Kennedy affirmed the right to join unions and a very limited right to collective bargaining in 1962 in Executive Order 10988, and the area for collective bargaining was slightly enlarged by the Civil Service Reform Act of 1978. But Federal unions are still generally precluded from bargaining about wages and fringe benefits, hiring, firing, work assignment, disciplinary actions, and contracting out. Levitan and Noden, op. cit. Federal employees do not have a right to strike, and may be dismissed and subjected to felony charges if they do strike. The Supreme Court has upheld this ruling, saying that the employees rights under the First Amendment and other Constitutional provisions must be balanced against the right of government to regulate the behavior and speech of its employees, The dismissal of the air controllers when they struck illegally is likely to discourage other strikes for some time. William B. Gould, A Primer on American Labor Law (Cambridge: MIT Press, 1982), p. 171. Thirty-nine States have some form of legislation protecting the right of public employees to organize and bargain collectively. See William B. Gould, op. cit. Federal employees are not covered by the Occupational Safety and Health Act but agency heads are responsible for maintaining standards consistent with the act (Executive Or-
PAGE 262
Ch. 9The Automation of Federal Government Offices l 255 or access to training, will be important. Federal unions have already adopted guidelines or model contract provisions on VDTs, which they distribute to bargaining committees for use in agency negotiations. Their activism may increase in the future. Federal white-collar workers are more likely to be represented by a union than are those in the private sector. About 54 percent of GS workers are covered, although only about onethird of those are dues-paying members. In the national labor force, only 8.5 million whitecollar workers, about 17 percent of the total, are members of unions. 89 Those covered by Federal unions are predominantly women and predominantly in clerical categories. Although figures are not available, it is likely that Federal unions have had a net loss of members in the last several years. International Comparisons The International Labour Organization (ILO) says that in developed and developing countries public sector employment: In respect of information processing, it is most likely that the over-all impact of com puterization lies not so much in the threat of redundancies, but in future limits on growth in clerical and related employ ment. g O der 12196, Feb. 26, 1980); and Federal emplo?ees ha~e the right to compensation for disability, death, or in]uries sustained in performance of their work (Federal Employees Compensation Act, 5 USC sec. 8101 et seq.). About 61 percent of the Federal work force is represented by unions, but this includes 86 percent of Federal blue-collar workers. Dues paying membership is harder to determine; unions are not required to make these figures public, See Sar A. I.evitan and Alexandra B. Koden, iiorking for the So~rereign: Ernp]ovee Relations in the Fecferaf Go}ernment (Baltimore: The Johns Hopkins University Press, 1983), pp. 15-20. Federal whitecollar workers are chiefly represented by the American Federation of Government Employees (34 percent of those who are co~.ered by union bargaining), the National Association of Go\ Fernment Employees (4 percent), the National Federation of Federal Employees (7 percentl, and the National Treasury Employees Union (5 percent). The first two of these are part of the AFL-CIO; the other two are independent. Iienry Le\in, Jobs: A Changing Workforce, Chance, Iol. 16, October 1984, pp. 32-37. International Labour Organization, The Effects of Structural Changes and Technological Progress on Employment in the Public Service, Report I II, Joint Committee on the Public Service, 3d sess. (Geneva: International Labour Organization, 1983), p. 30. The I LO notes that this will have its most serious effects on the prospects of women seeking to return to the work force and on school leavers. Nearly all governments are trying to improve efficiency and hold down costs, the I LO notes, citing evidence that many advanced countries have made cuts in their public work force as a result of information technologies. For example, Canada had a net reduction of 12 percent in employment in government secretarial positions from 1975-80. Within that general category, the number of typists declined by 35 percent and the number of stenographers by 68 percent, while secretaries increased by 17 percent and operators of office-composing equipment increased by 97 percent 91 Evidence of employment effects in centrally planned developed nations are, according to ILO, unreported or ambiguous. In developing countries, wherever statistics are available, they often show a shortage of qualified computer personnel. 92 A technology forecast done for the Federal Republic of Germany by a consulting firm suggested that if the government adopted a strategy of maximum use of information technology, 75 percent of all public sector office jobs could be standardized, and 38 percent automated. The ILO recognizes, however, that owing to cost, lack of planning, a desire to maintain existing work arrangements, or simply bureaucratic inertia such comprehensive plans will materialize only slowly, and a more likely scenario for most countries is the gradual introduction of components and systems that can later be linked, as the need arises. 94 Rapid development of telecommunications technology has affected postal workers in many countries, although employment levels have not fallen dramatically. Rather, they have shown slower growth rates in recent years, and .- Ibid. -11.0, op. cit., p. 28. J. Sleigh, et al., The Manpower Implications of Microelectronic Technolo~ (London: Her Majestys Stationery Office, 1979), pp. 77-79, cited in 11.0, op. cit. .(), op. cit., p. 17.
PAGE 263
256 Automation of Americas Offices in some countries staff reductions or hiring most countries are all public service jobs) and freezes occurred despite increases in the volsome do not. 95 ume of mail traffic. As to the future, some governments expect reductions in postal, telegraph, and telephone occupations (which in ILO, op. cit., p. 42. EFFECTS OF OFFICE AUTOMATION ON PUBLIC SERVICES Productivity increases do not necessarily lead to a reduction in labor force numbers or costs. Complementary or alternative objectives may be to increase the volume of products or services delivered, or to improve their quality. Or, increased productivity may merely mean coping with, instead of being overwhelmed by, an increasing workload. Because our economy increasingly depends on information as a resource and a driver of growth, the office workload is inexorably increasing. The primary objective in office automation is sometimes merely to accommodate this increased workload. 96 One final meaning of increased productivity is the creation of entirely new information-related services for customers, clients, or constituents. There has been rapid change in the global economy, and Americas role in that economy, since the 1960s. There have also been significant changes-particularly during the 1970s in the expectations that our citizens have about involvement in Federal decisionmaking, and the demands that they make on government agencies related to information about public resources and public expenditures. In allowing Federal agencies to meet these demands and to accommodate the increased workload, office automation quite possibly helped to avoid serious problems and disruptions that could have weakened the U.S. position of world leadership. Many managers and professionals report that the quality of analysis, reports, and other forms of decision-related services is improved by office automation, which allows large dataAlan Porter, David Roessner, et al., Office Automation Outlook, 1985-2000, contractor report for OTA, February 1985. bases to be collected, manipulated, analyzed, modeled, and systematically applied to policy formulation and decisionmaking. This improvement is more difficult to document and measure than other forms of increased productivity. It is also possible, however, that office automation may lead to a change in the nature of some government services and even to their deterioration. For example, services could become overly depersonalized and standardized, inflexible, and unresponsive to changing needs, if the technology itself is allowed to drive the design of the service or delivery mechanisms. A Case Study of Expanding Responsibilities The history of office automation in the Office of the U.S. Trade Representative is one illustration of office automation providing the means of meeting new challenges and avoiding disruptive overload in carrying out new government responsibilities. 97 The Trade Act of 1974, authorizing U.S. participation in a new round of multilateral trade negotiations, elevated to cabinet status the position of Special Trade Representative and assigned that officer responsibility for representing the United States in the negotiations, and administration of the Trade Agreements Program under that and other existing acts. --This section is drawn from an OTA contractor report, William Neufeld, Office Automation in the Office of the U.S. Trade Representative, February 1985. OTA is indebted to the Office of the U.S. Trade Representative for cooperation in allowing this study to be made and in facilitating the research by the contractor. The case study is also summarized in app. B.
PAGE 264
Ch. 9The Automation of Federal Government Offices 257 The Office of the Special Trade Representative, which until 1972 had only about 30 employees, had grown to about 126 employees by 1980, when it became the Office of the U.S. Trade Representative. In 1980 it was assigned overall responsibility for formulating and coordinating foreign trade policy among all government agencies, representing the United States in all trade negotiations regarding the General Agreement on Tariffs and Trade (GAIT), East-West trade, international investment, commodity agreements, the Organization for Economic Cooperation and Development, and the UN Conference on Trade and Development. 98 These developments reflected the growing importance of world trade in the U.S. economy. During the 1960s, the small agency had needed more and more data to use in developing positions for multinational negotiations. There was no central database on trade volumes, tariff rates, exports and imports, etc. Data was kept or reported by many different agencies, in different forms, and categories. The Washington headquarters and the Geneva negotiating office relied on different sets of data collected from different sources, which made policy coordination difficult. Planning to build a computerized trade database began in the late 1960s, but not until 1977 did an integrated database begin to function, serving many agencies and lodged for administrative purposes in a large computer system at the National Institutes of Health. 99 The development of the interagency trade database helped pave the way for office automation within USTR. It did not come about piecemeal but was systematically planned and implemented over a 5or 6-year period, to mesh with the large computer system. It has so far cost only about $2 million (including about $1 million investment in hardware). -The P~es[dent in 1984 proposed to Congress the creation of a Department of International Trade and Industry, with USTR under the direct control of the Secretary of that Department. q Renamed the Trade Policy Staff Committee Trade Net in 1984, the system provides a number of member agencies with data directly available through dial-out computer terminals. Through a combination of central computer operations, personal computers, and standalone and shared-logic terminals the USTR staff now has the capabilities of word processing, spreadsheet analysis, on-line data management, graphics, correspondence control, internal electronic mail, telemail, facsimile, and telex. Users can access the general database, and also a magazine containing USTR schedules, recent trade and labor summaries, and economic news. They can call up trade data on bilateral trade balances, imports and exports, domestic and foreign trade actions, and GATT documents. They can communicate through their terminals with the members and staff of the Trade Policy Staff Committee (the nine agency committee that provides analyses and recommendations to the senior level Trade Policy Committee) and of its subcommittees, and can electronically transmit documents to them. They can communicate by telemail with the Geneva office, USTR ambassadors, and staff members on travel, and can use this method of transferring documents between Geneva and Washington. Hard copy original documents can also be facsimile. In interviews, nearly all professionals on the USTR staff and half of senior managers said that they use terminals extensively, for up to 75 percent of their working hours. Professionals who were interviewed say that they can complete a project in half the time it would take without their computers. They also reported that the increased productivity had allowed them to clear a backlog of work that always built up in the past. Senior staff reported that they could ask for work on shorter deadlines. Some members of the professional staff reported that because of electronic communication and transmission of documents, the time required for review and clearance of trade-policy positions has in many cases been cut in half. At least in the perception of many professional staff members, the USTR has been able to take on a continually expanding workload, and maintain or improve the quality and pace of performance at the same time, because of information and communication technologies.
PAGE 265
258 Automation ot Americas Offices The staff has grown little if any during the 5 years in which office automation was implemented. 1 Changes in the Nature of Delivered Services Standardization and depersonalization of government services is a possible outcome of office automation. In the OTA case study of municipal agencies in New York City, 101 referred to earlier, an agency providing social services to disadvantaged clients reorganized the application and recordkeeping process so .. In 1980 after reorganization and expansion of responsibilities, the Office had 116 permanent full-time staff plus 10 Assistant U.S. Trade Representatives, or approximately 126 people. In 1984, it has 122 permanent staff members. But USTR has always used detailees from other agencies, and some consultants and part-time people. When this case study was done in 1984, the inclusion of such people brought the number of people working at USTR to 183; the number of such people on hand in 1980 could not be determined. This material is drawn from an OTA contractor report, The Effects of Office Automation on the Public Sector WorkforceA Case Study, prepared by the Labor Institute (Professor Joan Greenbaum, principal investigator), February 1985. as to use standardized coded forms. When inaccurate or incomplete information is fed into the computers it cannot be detected until late stages of the process, at which time the case has to be put on hold. Case workers say that some clients do not understand computerized forms and throw them away; others are intimidated by them and throw them away. Many of the municipal employees say that the quality of the services they deliver has been degraded. The Federal Government also delivers some services directly to individuals and households; for example, social security checks. With office automation some element of personal attention to the difficulties or foibles of recipients may be lost. The less educated or sophisticated, and those with language limitations, are more likely to be unable to cope with formal, unfamiliar processes. Thus, as the delivery of government services is automated, special procedures may be needed to assure that the access that some people have to services, or the attention paid to their needs, is not inadvertently degraded. EFFECTS ON GOVERNANCE Bureaucracya hierarchy of authority, specialization of functions, and formal channels of communication and command-is the basic structure of government as it is of all complex formal organizations. Bureaucracy is a valuable form of organization because it allows expertise to be harnessed, focused, and directed toward collective ends. Bureaucracies serve both despotic and democratic societies. In a highly technologized society, it can be argued that democracy can exist only because bureaucracy provides a way of using experts without being ruled by experts. Policy can be made by political representatives, and implemented by the hierarchy of specialists under their control and oversight. But as bureaucracy becomes larger and more complex, the specialization of functions at the lower ranks can become extremerationalization of tasks makes the work routine, monotonous, and finally mindless. The managers may spend increasing amounts of time correcting and coordinating, and little time doing anything productive. Some political theorists argue that in the electronic era, elaborate bureaucracies may be unnecessary. Information and communication technologies may offer a way to replace overly large and complex bureaucratic structures with more autonomous units assuming responsibility for an entire job. This prescription would be hard to implement in government without some sacrifice of accountability. Those who want to transform bureaucracies into democratic work groups tend to see formal channels of communication and control as a byproduct of specialization of functions, or to imply that they are antithetical to participatory democracy. But formal communication channels are
PAGE 266
Ch. 9-The Automation of Federal Government Offices 259 also a mechanism for achieving accountability, including accountability of bureaucrats to politically responsible policy makers. Within government, for example, communication typically moves through a series of nodes of authority. 102 The flow of information also involves in many cases a flow of authority. The presence of an authorizing signature on or with the information may give it forcethat is, cause something to happen. It may create or make binding the acknowledgment of an obligation. At least it acknowledges the receipt of the information. The signature, in other words, expresses a personal responsibility or validation of the information or some action associated with or flowing from it. Finding a fully adequate electronic substitute for the handwritten signature, as an easily recognizable and widely accepted validation for the origin and receipt of messages, is only one aspect of the problem and probably a minor one. 103 Electronic sign off may merely replace the buckslip and the paper-based signature of authority and accountability. But a signature is only one example of the set of procedures that in a government bureaucracy are designed to control and standardize behavior. The procedures usually include the designation of particular, formal channels of communication through which information (and authority) must flow. The flow is limited to these designated channels not only for the sake of efficiency and control, but to make it auditable in the future. These procedures not only guarantee control, for those at the top of the bureaucracy, but they also protect the general public; they are the means of making power lawful, of limiting its exercise -The di&ssion here draws heavily on Ronald M. I.ee. .Automating Red Tape: The Performative vs. Informative Roles of Bureaucratic Documents, Office: Technolog~ and People, vol. 2 (The .Netherlands: Elsevier Science Publishers B. V., 1984). pp. 187-204. Many organizations rely on the use of a combination of identification numbers or names and passwords to substitute for a signature. A few are developing techniques based on the concept of public key encryption, a mathematical technique developed in 1977 at the Massachusetts Institute of Technology (E.MMS, vol. 9, No. 13, July 1, 1985). Many people den~ the need for electronic signatures, on a practical basis, saying that messages sent by computer are no different from messages sent by telephone or telegraph. within agreed on boundaries. Some political scientists fear that reliance on telecommunications both within and between organizations (e.g., electronic mail) makes it difficult to control or to monitor the flow of information and blurs the location of both authority and responsibility. For example, changes in a database may be made without an authorizing signature, or instructions may be transmitted to contractors by persons without adequate authority for fiscal commitments. Thus, bureaucratic procedures, including formal channels of communication, are integral parts of the concepts of due process and accountability. But to the extent that information and communication technologies offer new opportunities and perhaps new temptations to ignore, evade, and erode those established channels, care must be taken that they do not diminish the accountability of government officials. The rules and established channels could persist, ossify, but be ignored. The principle of accountability, carried to extremes, leads to inscribed, detailed, and inflexible rules for every process and procedure. Information technology may provide a healthy corrective, making information so generally available that bureaucratic authority is constantly challenged in a productive, rather than a disruptive, way. But when information is electronic rather than paper-based, it propagates more rapidly, is less closely identified with its source, is both more accessible and at the same time more opaque to the casual seeker. It may tend to be used less responsibly. The public could find itself less able than before to identify and hold responsible the propagator, or the user, of information. On the other hand, as information becomes the subject of formal resources management, it is also possible for its use to become so encumbered with safeguards and procedures that access to it is reduced for both the public and the experts. None of these untoward consequences is inevitable. But the pervasiveness of new information technologies does mean that government information-handling is likely to change in unanticipated ways.
PAGE 267
260 l Automation of Americas Offices Over a decade ago Kenneth Laudon pointed out 104 that information about a government organization affects that organization in four ways: it affects the organizations reputation with the public, with its primary constituents, and with its employees (and, although not mentioned by Laudon, with those who oversee the organization in Congress); l it affects the organizations autonomy the more public information there is about the organizations inner workings, the less independent the role it can play in formulating policy; l the loss of full control over internal information tends to make an organization defensive and may lessen its own propensity to self-evaluation; and increased public information constrains the informal accommodations an organization can make with influential actors in its environment (congressional critics, interest groups, other agencies, etc.). In short, when the flow of information within and from an organization becomes less controllable, the organization tends to react in ways that are sometimes desirable, from the standpoint of democratic responsiveness, and sometimes not desirable. It cannot be assumed a priori that information technology makes information management more efficient, from the viewpoint of organization managers; nor can the relationship between information management and effective performance of an agencys mission be fully anticipated. Another major concern is the way in which information technology may be used either to increase governments collection and use of information about individual citizens, or to increase citizens access to information about government. Abuse of the capability to collect, aggregate, and use information about citizens, to violate the citizens privacy, or to make the citizen more visible to government and more susceptible to its control, has been Kenneth Laudon, Computers and Bureaucratic Reform (New York: John Wiley & Sons, 1974). a pervasive concern. These issues are being addressed in another OTA report, an assessment of government information systems. 105 Many enthusiasts for information technology anticipate that it will greatly increase citizens ability to know what government is doing and how, and to use that information to evaluate, influence, and participate in public decisionmaking. Four kinds of information about government are of primary interest to citizens, especially to citizens actively attempting to monitor or to participate in public decisions: 106 l l l l information about decisionmaking, policies, and practices internal to an agency budgeting and progr amming priorities, planning, key appointments and assignments, the locus of authority for specific actions and decisions, the actions of specific officials, etc.; evaluations, reviews, status reports, and other policy-relevant information from actors external to the agency, such as administration officials, congressional oversight committees, etc.; intermediate and final outcomes of specific cases and decisions, especially individual client problems such as social security eligibility or income tax appeals, and other personal data; and information about the costs of agency decisions and actions to citizens-paperwork burden, personal records disclosure, inequities in application of regulations, etc. Citizen access to the first two kinds of information has been widened by the Freedom of Information Act and other laws. Particularly when organized into public interest groups, citizens are generally able to obtain (Government Information Systems, forthcoming from OTA Communication and Information Technologies program later in 1985. wThis analysis draws on that of Donald A. Marchand and Mark E. Tompkins of the University of South Carolina, presented before the panel on Information Control Policy at the Annual Conference of the American Society for Public Administration, Phoenix, AZ, 1974, and later published as Information Management and Use in Public Organizations: Some Impacts on Citizen Participation, State and Local Government Review, September 1981, p. 103.
PAGE 268
Ch. 9The Automation of Federal Government Offices l 261 such information readily. Information technologies make it easier and less costly for an agency to supply such information when demanded by citizens. Some agencies are giving citizens direct access to such data through information systems. The citizen has much less access to information about the status or progress of his or her individual dealings with government agencies as a client; for example, in appealing income tax decisions. Here the individual is dependent on the agency to give up information, and has no means, technological or otherwise, to access it directly. But the agency can use information technology to collect, or at least to aggregate, process, use, and disseminate information about the client. In this case, the clients information-processing relationship with the public agency is likely to be onesided. the client must reveal information about himself to get services from the public organization. 107 The thrust of this argument is that information technology improves the agencys abilit y to get or use information about the citizen as client, but does not improve the client ability to know about the progress or outcome of his or her own case as it is handled by an agency. But even the first kind of benefit cannot always be assumed; that is, that agencies will have better data about clients. As already noted, New York City social services workers reported that many of their clients ignored, failed to understand, or were intimidated by standardized forms, and also that erroneous information fed into the automated system could not be caught and corrected until late in the process. 108 In this case information tech nology may have degraded the client information going to the agency, without improving the citizan/client control over it or access to it. The fourth kind of information, about the direct and indirect costs of government action to the public, is diffuse and hard to make specific. Citizens have difficulty in figuring out exactly what a change in policy or procedures would cost them in taxes, additional paper-Marchand and Tompkins, op. cit., p. 106. The Effects of Office Automation on the Public Sector 14rorkforce-A Case Stud.~, op. cit. (OTA contractor report). work, narrowing of options, or political power. The same is true in many cases for those who would benefit from government action. Thus, the poor, the disadvantaged, or the small businessman who are targets of government programs, may not even know that they are or would be affected. Information technology could be of help to both citizens and government agencies in making this kind of data available. 109 Office automation could improve the ability of agencies to target the recipients of their services through conducting surveys and analyses, improved special census counts, and collection and analysis of information about clients as a group. It can be designed to improve the handling of client correspondence and to allow customized response to inquiries. However, information technology can also be implemented in such a way that relationships between government and constituents are further standardized and dehumanized, or information delivered to constituents becomes so highly technical and jargon-laden as to become useless to those who are not highly educated or even to those who are not specialists in fields such as contract law, taxes, or health delivery. Mechanisms such as forms for requesting information from, or delivering information to, a citizen are often designed for the convenience of the computer, or data processor, rather than for clarity to the citizen. This creates a real possibility that information technology, unless used with care and judgment, could further increase the disadvantages confronting those who are already educationally and economically disadvantaged. The effects of information technology on the delivery of government services, and more importantly, on the quality and equity of governance, must be carefully monitored so that corrections can be made if needed. W This is not the conclusion of the authors cited above, Marchand and Tompkins, who conclude on the contrary that information technology is of little direct help to either the agency or the citizen as victim. They argue that because information technology and the ability to access and use it is not equally available to all, it has distributive effects on the ability of citizens to use information about government to monitor and influence policy; current information policy has regressive distributive effectsit benefits the well-to-do more than the disadvantaged.
PAGE 269
Chapter 10 Office Automation in State and Local Governments
PAGE 270
Contents State Government Office Automation Plannin g and Procurement . . . . The Extent of Automation in Local Governments. . . . . . . . Perspectives on the Effects of Public Office Automation . . . . . . The Effects of Local Government Automation . . . . . . . Examples of Office Automation in State and Local Government . . . . The State of South Carolina . . . . . . . . . . . New York City.. . . . . . . . . . . . . . . Unanswered Questions About Information Technology and Governance . . . Figure 265 267 269 269 271 271 275 279 Figure~No. Page 10-1. An Example of a Statewide Integrated Communication Network for South Carolina . . . . . . . . . . . . . 273
PAGE 271
Chapter 10 Office Automation in State and Local Governments There are 50 State and about 78,200 local government units in the United States. To make general statements about their activities is always difficult; one of the benefits of a Federal system is that the diversity of their approaches to problem-solving provides a living laboratory for public policy formulation. The workload in State and local government has increased more rapidly in recent decades than that of the Federal Government. Between the end of World War II and 1980, State and local government employment increased steadily to a high of 13.3 million. It has since slightly declined, as pressure for increased administrative efficiency and cost reduction has grown. As a percent of total public sector employment, State government employment increased from 16.5 percent in 1950 to 23 perThis includes about 3,000 counties, 19,000 municipalities, 17,000 townships, 15,000 school districts, and 26,000 special districts (when they are fiscally and administratively separate from other governmental units). At the beginning of World War II there were about 156,000 local government units, or 49 percent more. The big reduction was the elimination of over 93,000 school districts during the 1950s and 1960s, by consolidations. cent in 1979; local governments share grew from 50 to 59 percent. Federal budget cuts and the phase out of Federal grant programs could bring about a massive transfer of administrative responsibility from central to State and local levels of government. This provides a further strong incentive for seeking greater efficiency. These governments are automating their offices at different paces and following different strategies, and the consequences will be different for each governmental unit. Much of the information now available comes from the limited number of case studies of individual jurisdictions. Many of these case studies have a narrow perspective, being focused on public administration criteria of cost-effectiveness. Relatively little literature is available as yet dealing with effects of office automation on governance, policy, or political responsiveness. U.S. Department of Commerce, Bureau of the Census, Public Emploj~ment in 1983, GE83-No. 1, STATE GOVERNMENT OFFICE AUTOMATION PLANNING AND PROCUREMENT Some State governments have gone much further than others in computerizing their offices. Tax and finance operations, personnel records, and routine recordkeeping are however thoroughly automated in almost all States. It is common for a number of State agencies to share a data processing center, but there are usually several or many processing centers within a State government. In planning, procurement, management, and degree of control over information processing there is wide diversity. Some State governments have realized that computers represent the possibility of a fundamental change in the way public affairs are administered, and have established long-range plans and strategies for effective use of the technology. Other States have not gotten beyond treating office automation as merely the latest marginal improvement in office equipment. Some have let of265
PAGE 272
266 l Automation of Americas Offices fice automation happen to them, unplanned and incrementally, and are now struggling to rationalize these changes. This is reflected in procurement strategies. Some States have governmentwide master plans for investment in information technologies while some expect planning to be done on an agency-by-agency basis as part of the budgetary process. Many States, after some floundering, have developed procurement controls and standards for large computer systems, only to be taken by surprise by the infiltration of small computers and word processors, which generally fall below the threshold cost that requires centralized procurement approval. Minnesota, for example, has highly centralized planning and control of information systems, while in New York State the responsibility is almost completely decentralized. Other States fall along the continuum between these poles. 3 Most States do, however, have some central office or division that coordinates information-technology procurement, if not longrange planning. This office is usually within a department of administration or general services, or in the department of finance. Some States, however, give the responsibility for such coordination to a special commi ssion. For example, in Florida, the Information Resources Commission, made up of the Governor and six major State officials, reviews agency information-technology plans. In Texas, the Automated Information Systems Advisory Panel, with a similar role, has both public and private sector members. State and local governments are regarded as the largest single market for the microcomputer industry in the next few years. 4 During the early period of acquisition of government computer systems, States were concerned over Many of the examples in this section are taken from an OTA contractor report, A Comparative Review of Information Technology Management Practices in State Government, prepared by the Institute of Information Management, Technology, and Policy of the University of South Carolina, John C, Kresslein, research analyst, December 1984. J. Robert Ippolito (Director, Division of Electronic Data Processing, Florida Department of General Services), Computer Technology Procurement: Can It Be Standardized for State and Local Governments? State and Local Government Review, vol. 13, No. 3, September 1981, p. 85. their ability to maintain their control over public purchasing (especially physical specifications, terms, and conditions of price competition) because of the powerful market positions of a few firms. In this large but disaggregated and dispersed market, it was difficult for State governments to make effective demands regarding physical specifications of systems or to specify the terms and conditions under which they would do business with computer system vendors. The Council of State Governments warned in a 1975 report that: There are factors at work which defeat certain long-standing principles of public purchasing. The effects of oligopoly are multiplying. 5 This led the American Bar Association to develop a Model Procurement Code for State and Local Governments, which was published in 1979. Many States have since developed their own standard contracts and agreements. As State officials have become more knowledgeable and experienced with computer technology the problem has eased. But more recently, stand-alone word processors and personal computers (PCs) have introduced a new element of uncertainty, since their cost is low enough that their purchase may not require centralized approval. It is common for State agencies and local governments to find that equipment from several different vendors have found their way into government offices with no central plan foror even inventory oftheir spread. For example, one State government report acknowledges that thousands of PCs have found their way into large organizations by the end-user endaround, that is, users avoiding bureaucratic purchasing procedures by buying their PCs below required dollar review levels. At the State level as in large corporations, microcomputers and stand-alone word processors are almost always superimposed on an existing pattern of centralized computing. A The Council of State Governments, State and Loczd Government Purchasing, Lexington, Kentucky, 1975, p. 7. State of South Carolina, Division of Information Resource Management, Personal Computers in State Government, May 1984, p. 2.
PAGE 273
Ch. 10Office Automation in State and Local Governments l 267 report on personal computers in the government of South Carolina notes that this has caused a rethinking of management strategy: In 1979, when the Division of Information Resources Management began its original office automation efforts, the PC was not a viable end-user device for integration into the States office system model. However, we now see the computing landscape changing to a three level model in which: Level 1 is the corporate mainframe node; Level 2 is the departmental (minicomputer node; and Level 3 is the personal computer node. These levels more closely match the organizational hierarchy. At the time of this writing (1984) DIRM has begun efforts to provide management direction and support for the PC and is developing strategies to integrate the PC into its overall State Plan on Technology. (Emphasis added,) This State government anticipates, according to the report, that by the end of the decade there may be one PC for every State government office worker. State office automation is not limited to computers in the State House or capital city; increasingly, it is including sophisticated systems to provide an interactive network between dispersed nodes of government services. At least two States, California and South Carolina, are planning digital backbone networks to link together centers of government activity around the State. Alaska has an audioconferencing network connecting 70 sites across the State and 17 full-time information offices that are also networked. Michigan has a distributed network throughout the capital complex, with a terminal in every senators and representatives office. State legislatures, as well as administrators, are using computers. Most States have electronic data processing systems for statute search and retrieval and reporting on the status of bills before the legislature. g Ibid. See the case stud~ cm South Carolina government use of computers later in this chapter. Linda Schulte, A Survey of Computerized Legislative Information Systems, La% Library Journal, \ol. 72, winter 1979, pp. 99-129. THE EXTENT OF AUTOMATION IN LOCAL GOVERNMENTS By the mid-1970s most large counties and municipalities were also using computer systems for high-volume data processing. 9 By 1980 these governments had an average of 31 operating applications and as many as 30 under development, and were spending an average of $500,000 yearly for automated data processing. The systems tended to be used first and most pervasively for departments of finance. Recordkeeping and printing were other James ,N. Danziger, The Use of Automated Information in IJocal Government, American Beha\rioral Scientist, vol. 22, No. 3, January-February 1979, pp. 363-392. A survey by the Urban Information Systems Research Group of the Pubhc Policy Research organization indicated that more than 90 percent of cities with populations of 50,000 and larger, and 90 percent of counties of o~rer 100,000 people had computer systems, while less than half of those smaller had them. Kenneth I.. Kraemer, William H. Dutton, and Alana Northrop, The Afanagement of Information Systems (New York: Columbia University Press, 1981), p. 8, widely sought applications. 1 More recently, however, a large number of sophisticated applications have been developed for use by local governments, both in the processing of data for routine operations and in the analysis of information for broader purposes of management and planning. 11 Larger governments usually bought or leased equipment and developed in-house technical staffs, often developing their own software. Others depended on contractual agreements for specific computer applications with a corn -. Robert E. Sellers, Miniand Microcomputers in Local Government: Their Application and Their Impact, State and Local Go\wrnment Re~iew, September 1981, p. 91. tJohn Leslie King and Kenneth L. Kraemer, Information Systems and Intergovernmental Relations, Public Sector Performance. .+! Conceptual Turning Point, Trudi C. Miller (cd. ) (Baltimore: The Johns Hopkins University Press, 1984), pp. 102-130.
PAGE 274
268 l Automation of Americas Offices puter services/time-sharing company. In a few cases local governments acquired computer systems but hired a facility management concern to operate them, or two local governments (e.g., a city or county and a school system) developed a jointly owned and operated computer center. End-use computing is now spreading in the government agencies of large cities. 13 Public Technology, Inc., surveyed its member cities and counties with a population over 400,000 in 1983 and found that among the 50 respondents, 75 percent had microcomputers. By then, 93 percent had one or more minicomputers, and over half had mainframes. A similar survey in 1984 indicated that 77 percent of 84 cities responding (80 percent response) had microcomputers, and 87 percent planned to have them by 1984. There are already hundreds of local government applications programs available for microcomputers. But despite the low cost of small computers, most small local governments have not yet begun to use them. A survey in 1983 of 162 local governments within the State of South Carolina found that 81 percent were using computers of some kind, but this usually represented centralized computing, or large-volume data processing. While 3 of the 6 large cities were using microcomputers, only 3 of 27 counties, 1 of 12 medium size cities, and 6 of 53 small cities had acquired them. 15 The International City Managers Association (ICMA) surveyed 5,808 cities of all sizes in 1982, and found that only 13.2 percent of Sellers, op. cit., p. 93. See Special Symposium on Microcomputers in Local Government, A Public Administration Review, vol. 44, No. 1, January-February 1944, especially James R. Grieseemer, Microelectronics and Local Government: New Economies and New Opportunities, p. 57. C-TAC Survey Reveals Members Information-Handling Resources, Public Technology, vol. 5, No. 11, November 1983. The 1984 data was reported in Joey P. George, Who Does the Buying? Government Data Systems, August-September 1985. State of South Carolina, The South Carolina Local Government Survey of the Use of Computer and Communications Technology, prepared by the Institute of Information Management, Technology, and Policy, University of South Carolina, 1983. them were using microcomputers, although about one-third planned to buy one or more during the next 2 years. 16 Most of those that had computers were using them for word processing and financial management. Of the more than 19,000 U.S. municipalities, over three-quarters have populations of less than 6,000. These small towns or villages usually provide the same basic services as larger citiesutilities, fire and police protection, recreation, taxing functions, usually school systems, and often pl anning departments. In addition, they have many responsibilities under Federal programs, such as administering revenue sharing and block grants. 17 They too are trying to cut costs and increase productivity. In spite of this obvious need it appears that small local units, especially those in rural areas, are not rushing to computerize their offices .18 However, it is possible that the purchase of microcomputers by small governments has accelerated in the last 2 years beyond expectations. As late as 1982, lack of familiarity with computers was probably the major factor impeding their purchase by small cities; until recently there were likely to be few computer vendors in small towns and this market was not being aggressively pursued. This may have changed considerably by 1985. Donald F. Norris and Vincent J. Webb, Microcomputers: Baseline Data Report (Washington, DC: International City Managers Association, July 1983), as reported in Donald F. Norris, Computers and Small Local Governments, Public Administration Review, vol. 44, No. 1, January-February 1984, pp. 7off. Sellers, op. cit., p. 91. Another survey in 1982 looked at cities and counties in seven Mountain and Plains States. Of 75 cities with populations of 2,500 to 50,000, 68 percent were using computers; of 75 counties under 100,000, 36 percent. Smaller communities were less likely to have computers than larger ones percent of those with 5,000 or fewer people compared to 76 percent of those with more than 10,000 people. Those with city managers were much more likely to have computers than those with mayor/council governments, and metropolitan counties were more likely than rural counties to use them. About onethird of those who were already using computers, but only 16 percent of the nonusers, planned to buy more in the next 2 years. Over three-quarters of those that were using computers had a minicomputer or bookkeeping system and only 22 percent had microcomputers (7 percent of all the cities surveyed). Most were using their computer system for payroll and accounting, budgeting, and utility bills. Ninety percent had only one computer, and only 1 percent had as many as four. Norris, op. cit., 1984.
PAGE 275
Ch. 10Office Automation in State and Local Governments 269 PERSPECTIVES ON THE EFFECTS OF PUBLIC OFFICE AUTOMATION Early commentaries on State and local government use of computers tended to be theoretical rather than empirical, and often took one of three basic perspectives: 19 l Information technology will rationalize not only operating procedures but the political/policy process. l Adoption of information technology will be driven partly by the technological imperative and partly by the self-seeking actions of a technical elite, uncontrolled by humane values and unresponsive to public attitudes. Information technology is no different from other technology and other kinds of information resources, and will have no unique or discernible effects on government. The first of these perspectives tended to be found in public administration literature, while the second tended to come from political scientists, perhaps reflecting the different relative values that the two disciplines place on efficiency and responsiveness. Enthusiasts often saw automation as a way of achieving the goals of the old reform movement, and claimed that technology would increase productivity, cut costs, improve decisionmaking, allow better management control of operations, improve job satisfaction, and allow streamlined governments to offer more and better services. Critics feared that it would lessen the responsiveness of bureaucrats to citizens, put a technological elite in charge of the local political process, shift power from elected representatives to hired managers, deskill jobs, and greatly increase the costs of government. The Effects of Local Government Automation Some local governments were indeed taken aback by the costs of office automation because they had not anticipated that the lifecycle costs of the equipment, with the supDanziger, op. cit. port and training required, would so far exceed investment costs. The first phase of computerization, at least, may not have reduced labor costs but rather stimulated the hiring of clerks because it required input of massive amounts of data. 2o There were widespread expectations of a shortage of trained personnel in local government 21 Other effects were mixed but generally favorable. In some jurisdictions administrative control was improved, and workers reported increased job satisfaction. There are indications in many cities that office automation tends to reinforce, rather than change, existing patterns of bureaucratic and political power. 22 Some social scientists make this point more strongly, concluding that computerization of local government has benefited those who already have broad power and control in the local hierarchy. They usually argue that K. I.. Kraemer, J.N. Danziger, and tJ.1,. King, 4I.ocal Go\ernment and Information Technology in the United States, Local Government and Information Technology>, In formatics Studies No. 12, OECD, Paris, 1978. -J. L. King, Local Government [Jse of Information Technology: The Next Decade, Public Administration Rmriew, vol. 42, No. 1, January-February 1982, p. 31. --For example, a study of 42 cities from 1975 to 1979 found evidence for improved administrative control and operating performance, especiall~r where computers were used for routine tasks, for example, tax recordkeeping and traffic-ticket processing. In applications, such as support of police detect iies investigations or planning and policy anal?wis, the results were mixed or marginal, or the evidence was ambiguous. P: ffects on job satisfaction and the office work en~ironment were said to be beneficial; the researchers said that this was often an unplanned, unanticipated benefit. The same survey concluded that there was some e~idence that office automation reinforced existing pat terns of bureaucratic power rather than changing them. 1 n most cases, a trend toward centralized, professionalized management in local government seemed to be strengthened as a result of automation. The perspective of the research group in this case was clearly that of public administrators who placed high \alue on rationalizing government activities. They found that the most successful implementations of automation were alwa~rs linked with the most advanced and sophisticated technolo~ and a highly professionalized work force; the sociotechnical approach, emphasizing user-friendly equipment, human relationships, and worker in~oliement in planning and decisionmaking, often worked best from the perspective of indi~iduals, but at the cost of some sacrifice of efficiency. *Evaluation of Information Technology in I.ocal C~o\ernments, 1975-1979. a sur~e~ by the urban information systems research group of the Public Policq Research Organization, reported in Kraemer, et al.. op. cit., pp. 27ff.
PAGE 276
270 l Automation of Americas Offices there is a need for explicit democratization of government computer strategies. 23 Local governments collect and use information for two broad categories of activitymanagement of routine government operations, and planning and analysis. 24 Most of the information systems and applications in use are for operational purposes-management of government revenues, payrolls, etc. But local governments must collect and use information for scores of activities; two middle-size cities, in 1980, identified over 300 distinct operational information systems (not all then computerized). 25 This operational data, in summarized or aggregated form, is also useful for planning and management, especially status reporting, performance analysis, and compliance analysis. Computer programs and models have been developed for these purposes, including, for example, fiscal impact analysis, land use, transportation analysis, urban development planning, and expenditure forecasting models. Many of these are in the public domain and others are readily and inexpensively available; but relatively few governments have used them, and these are primarily the larger municipalities. It maybe that most local government organizations do not yet have the expertise required to use these tools, but it may also be that modeling does not fit the informal, pragmatic, personalized mode of decisionmaking characteristic of government close to the grass roots of democratic government. 26 James N. Danziger (University of California-Irvine), Computers and Politics: High Technology in American Load Go}Ternments (New York: Columbia University Press, November 1981). This analysis draws heavily on the conceptual framework provided by King and Kraemer, op. cit., see especially pp. 106-110. King and Kraemer, op. cit., p. 107. An analysis by Professors King and Kraemer of the University of California-Irvine concludes that in terms of producing improvements in performance, most attempts to use these models for planning and management have been practical failures, because of several problems: l l l Information collected for operational use is generally not aggregated at the appropriate level for, or organized for, planning and management purposes. I.ocal governments collect relatively little data appropriate for assessment of output. Models must be fitted to the expected analyses, and this level of analytical competence is not always readily available. Behavioral and political constraints: Data useful for performance and compliance analysis often depends on self-evaluation by government officials and this affects their behavior in collecting/reporting the There are many examples of local governments using information and communications technologies in innovative ways that can be copied and adapted by other communities. For example, Jacksonville, FL, is using electronic printing and a local-area network to keep its city ordinances up to date. Amendments passed by the city council are immediately incorporated in the code, and supplements are printed four to six times yearly, whereas formerly the one supplement per year required 6 months and up to $50,000 to prepare. The Office of General Counsel reports that it has reduced costs by eliminating the need for a large staff of editorial lawyers, printers, typesetters and other personnel. Boston has required its licensee for cable television to link with a coaxial cable network all public buildings, including municipal offices, schools, and fire and police buildings. The system provides electronic mail, and the citys 29 word processor systems will be part of the network. During the last municipal election a direct link was setup between the ballot counting at City Hall and local television and newspaper offices. 27 San Diego developed a computer-based planning and management system to determine the most cost effective location for the citys service units in order to reduce travel to and from work sites. By reassigning crews and equipment among existing stations the city will realize an immediate net cost saving of over $445,000. 28 Fort Collins, CO, installed an automated information system for use by police officers that allowed it to eliminate one position in the Records Department, freed room for three additional workstations, and increased the reliability and integrity of criminal justice information. The effect of the advent of personal computers in local government offices, even if it is now occurring more rapidly than was ex dataespecially if it is to be used to evaluate performance. The assumptions used in constructing models may be unrealistic in a specific situation and location, given the diversity among American communities. Models tend to be used to rationalize or defend decisions already made for political or other reasons rather than to help make decisions. (King and Kraemer, op. cit., p. 108.) -Reported in The Government Office, Office Administration and Automation, September 1984. h Information provided by Public Technology, Inc., to OTA from a list of PTI computer-technology award winners, November 1984.
PAGE 277
Ch. 10Office Automation in State and Local Governments 271 pected in 1982, will not be apparent for some uneven rate, and in many governments with time, especially since they are likely to come little attention at the highest level of maninto use without systematic planning, at a very agement. EXAMPLES OF OFFICE AUTOMATION IN STATE AND LOCAL GOVERNMENT In the absence of a large body of crossjurisdictional comparisons, some brief case histories are presented as illustrative of what is taking place in government offices. They are not necessarily typical, but they provide a picture of some of the possible outcomes of office automation for States and municipalities. The State of South Carolina South Carolina falls about midway among the 50 States in size (it ranked 24th in terms of population, in 1980) and in government revenue per capita (38th in 1980). Like other States, South Carolina began to acquire computer systems during the 1960s and 1970s. In 1981, South Carolina was spending as much as 6 percent of its operating budget for telecommunications, data processing, and office automation, but without any systematic plan, strategy, or control over these expenditures .29 In buying equipment or services for data procThis section relies heavily on work done for agencies of the State (especially the Division of Information Resource Management of the State Budget and Control Board) by Donald A. Marchand, John C. Kresslein, and others at the Institute of Information Management, Technology and Policy, College of Business Administration, University of South Carolina. Reports, to be referenced at points in this section as South Carolina (number), include: 1. 2. 3. 4. 5 ,. 6. 7. Information Resource Alanagement: A Statewide Strate~, April 1983. The South Carolina Local Government Sur\e~ of the Use of Computer and Communications Technology, November 1983. Implementing Information Resource Jlanagement in State Government: The South Carolina Experience, Research Report RR-84 -l(K-84-8), May 1984. Personal Computers in State Go\rernment, May 1984. A hfanagers Guide for Implementing Information Resource ,$ fanagemtmt (IIiM) in a State ,4gencj, ,June 1984. Office Automation in the Office of the Go~rernor, tJanuary 1985. Initial E}aluation of Prcductivit} Benefits Achie\ed From S0,4PS, Progress Report No. 2, tJuly 1983. However, neither specific State agencies, the Institute, the university, or the authors cited abete are responsible for the interpretation and analysis in this section, which is solel~ that of OTA analysts. essing some State agencies had underestimated the costs by as much as 50 percent. In a decade, expenditures for managing data had more than tripled; and outlays for telecommunications in 1980 were 38 times greater than in 1960. 30 The Governor, the State Budget and Control Board, and leaders of the State legislature agreed that something must be done. The time was right since the legislature was then working on a new Model Procurement Code. The consensus across branches of government was important, because power in this State government is particularly fragmented and diffuse. The heads of major State agencies are elected separately rather than appointed by the Governor, and therefore tend to have their own unique relationships with legislators as well as their own constituencies. The Governor consequently must often depend on powers of persuasion to initiate any change within the State bureaucracy. With a powerful consensus among State leaders that some way must be found to cut the burgeoning costs of information management, it was possible to take action. The State administration began efforts to develop a comprehensive information resources management strategy. Key officials throughout government were interviewed, and State agencies were asked to develop productivity objectives. On this basis a preliminary State Plan on Technology was developed, and later a Master Plan. Agencies were asked to prepare their own 3year plans for use of information technologies. The Department of General Services was reorganized to create a new Director of Information Resource Management and a division by that name (DIRM) under the Budget and Control Board. Another important step was South Carolina (5), p. 5.
PAGE 278
272 l Automation of Americas Offices the creation of a new series of personnel classifications under the title Information Resource Management, to encourage the appointment of such specialists in each agency. The State Plan, as it developed, encompassed office automation, telecommunications, and large-scale data processing technology. The backbone of the planned technology is to be a statewide microwave network for voice, data, and video transmission. (See figure 101.) This network, which is planned for completion in 1986, will cost an estimated $16 million, but is projected to save the State $100 million over the next 20 years. The State did not set out initially to establish its own communication system, university consultants say, but sought instead the most cost-effective alternative with the goal of controlling communication costs. 31 The planners sought and got bids from 25 private industry vendors but none could provide what was wanted. Another aspect of the State Plan was a systematic approach to office automation (meaning, in the beginning, large-scale computer applications). An 18-month pilot project was begun in August 1981. Task forces and study groups were formed to consider procurement and standards, ergonomic issues, training needs, user perceptions, and contingency planning and security needs. The major objectives in office automation were: l l l l automation of formerly manual tasks to reduce manpower requirements, direct time-saving applications and work redistribution to reduce time and manpower requirements, time-saving applications providing opportunity to do additional work, and quality improvements in office products. The first pilot project was purchase of three IBM 8100s and 100 peripherals for the Department of General Services in mid-1982. In the following year further advanced data processing applications were installed, and technical assistance was given to other State agencies. The Model Procurement Code passed by South Carolina (4), p. 15. the legislature in 1981 provided strong induce ments for agencies to comply with standards formulated by the Department of General Services new DIRM. DIRM must oversee any procurement of over $2,500. The State would coordinate long-term contracts under which agencies would receive a significant discount on computer-related purchases. Office automation standards were developed in 1982 and subsequently revised. By September 1983, an electronic network linked the Governors Office with nine critical agencies within the State House Complex the office of the Executive Director of the State Budget and Control Board, the Health and Human Services Finance Commission, the Department of Agriculture, the Division of Human Resources Management, the Department of Social Services, the Office of the Secretary of State, the Department of Insurance, the Division of General Services, and the Division of Information Resource Management. For automating the Governors Office two minicomputers were linked to host computers at DIRM and the University of South Carolina; there are 22 workstations, 3 data processing terminals, 7 letter-quality printers, a line printer, and a data processing printer. A recent assessment concluded that the result was a 53 percent cumulative time saving in actual time to complete all tasks studied (approximately one-half person year), with an average time saving per task of 5.3 percent or approximately 112 person hours per year. 32 The improvement in time to complete each task was estimated to be about 86.5 percent. Three tasks that were done externally are now done internally (answering employment application letters, approximately 10 per day; preparing proclamations; preparing 200 news release envelopes, about twice each week) with an improvement in turnaround time of about 94 percent. Nine of twelve typewriters have been relinquished for redistribution to other offices. South Carolina (6). The specific tasks studied were property management certification, grant report, audit report, grant award letters and forms, consolidated vouchers (math calculations), updating office personnel lists, preparing governors schedule, repetitive letters, commission letters, educational statistical report.
PAGE 279
. Ch. 10Office Automation in State and Local Governments 273 L m I r a) (5
PAGE 280
274 l Automation of Americas Offices More qualitatively, the assessment found that the successful adoption of the Governors Education Improvement Act in 1984 was partially achieved through the efficient and effective use of the office automation capabilities that were used for creating the text, completing 100 major revisions, preparing the related speech by the Governor with 18 major revisions, and performing mass mailings to constituents, related to the act. The Governors Executive Assistant was quoted as saying that without office automation capabilities the act would not have been adopted in that fiscal year, or would have contained a lesser program content. The assessment concluded that: (0)ffice automation is indeed initially more costly than conventional or semiautomated techniques. It also is extremely difficult to precisely measure the quantitative productivity improvements and equally as difficult to determine the return on investment However, (automation resulted) in the capability to produce and distribute information faster and more accurately and to make more accurate and timely decisions Indeed, the. technologies helped improve the level of service and the relationship of the Office of the Governor to the citizens of this State. 33 The Legislative Services Group also automated (with 200 word processing terminals connected to a mainframe) and terminals were put in the offices of legislators. Through the system, legislators can access and track bills and amendments and search the State Code of Laws. Word processing, data processing, key-word search, and calendaring are also provided. Automating State offices was not accomplished without significant problems. Those identified at an early stage of the process were those found in nearly every large organization functions beyond word processing tended to be underused because people had not been taught how to use them; there was sometimes a mismatch between capabilities provided and those needed; and distribution of terminals was sometimes not appropriate so that people queued to use them in some places while other terminals were idle. But four-fifths of the clerical staff and over half of managers and professionals said that the new systems had improved not only their productivity but their attitudes toward their work. 34 By 1983 it was clear that personal computers were rapidly being added to State offices. Their prices often fell below the threshold at which there is central oversight. There was concern about the potential lack of compliance with office automation standards, which could cause compatibility problems later, yet the flexibility that PCs offer pointed them to becoming the primary office automation workstation, university consultants said. 35 A survey of 10 agencies in early 1984 showed at least 500 microcomputers in use, and that number was likely to double during fiscal year 1985. The Comptroller Generals Office was using two personal computers to design a financial reporting system to be used by small State agencies. 3 Several agencies reported that there was a high demand for personal computers, and staff members were dissatisfied and frustrated by the necessity for sharing them. No local area networks (LANS) were yet in place in these agencies and use of shared hard disks was just beginning. But the two State universities were planning campus-wide networks and four of the agencies were interested in developing them. The State now requires personal computers to be bought under a term contract, so that all agencies enjoy the volume discount and guarantees of support negotiated by the State. It is officially anticipated that by the end of the 1980s there may be one microcomputer for every State white-collar worker. DIRM has purchased software packages for PCs that pro vide English-language-like query capability and facilitate the downloading of data from mainframes to PCs. PCs and mainframes will be linked using leased or dial-up lines to allow South Carolina (7). South Carolina (7), p. 26. South Carolina (4). South Carolina (6), p. 2.
PAGE 281
Ch. 10Office Automation in State and Local Governments l 275 end-users controlled, preauthorized access to central government databases. Several State agencies and institutions (e.g., the Department of Mental Retardation, the Tax Commi ssion, a State technical college, and the State university medical school) are systematically developing their own long-range, information-resource management plans. Throughout the process of automating government offices, this State government has used university experts to monitor and report on progress and problems in achieving its stated objectives. A broader assessment from the perspective of effects on delivery of government services, governmental responsiveness to citizens, and on the locus and exercise of governmental power is to be undertaken but is not now available. New York City The Mayors Office of Operations in New York City cooperated with OTA in studying the effects of public sector office automation by encouraging and facilitating a case study by The Labor Institute of automation in three municipal departmentsthe Department of Finance, the Department of General Services, and the Human Resources Administration. 37 New York City has for several years been under intense pressure to reduce government operating costs and increase revenues. Major reorganizations aimed at increasing productivity took place in city government departments both before and during office automation. The case study focused on the consequences of automation for the clerical, managerial, and professional work force, and on their perceptions of how it affected both them and the services that they deliver to the public. Many of the workers who contributed their insights to this study said that in spite of some early misgivings, they like office automation. However, there is at the same time a high level ~;ffects of Office Automation on the Public Sector llorkforce: A Case Stud~, of Yew York Cit.}, done under an VI*4 contract by The I.abor Institute ( New York): ,Joan Crreenbaum, principal in~estigator, with Cydne~ Pullman and Sharon Sz~zmanski, February 1985. of dissatisfaction. Someboth clericals and professionals-felt that their jobs had been degraded, and their own interest in the job eroded. Most said that their own productivity had increased; and some said that the quality, quantity, and timeliness of services delivered by their work units had improved. But in other cases, services have been depersonalized, standardized, and routinized; accountability of individuals, if not of the government as a whole, has been decreased. The clerical workers in the three departments are group clerks, office aides, office associates, technical support aides, and word processor operators. The clerical work that has been automated was all done manually until 1982, but now requires personal computers, word processors, Automated Forms Systems, and other systems. 38 For example, at the Bureau of Child Support (in HRA) a clerical work unit formerly typed between 400 and 600 letters each week. Now a clerk enters a code to select forms to be sent to a client from nine forms that are generated by an automated system. Clerks at the Income Maintenance Center once hunted for paper files on particular clients as they were needed by eligibility specialists; now the files are called up on a cathode ray tube (CRT) and the needed data is printed out. personal computers are used in the Real Property Transfer Tax Unit of the Department of Finance to call up information on property sales history and determine the market value of property. Word processors are used for tax billing, as well as for generating reports, mailings, memoranda, and other documents in all of the departments. In one department the most proficient typists were removed from their old work units and placed in a central production unit. Twenty-eight clerical workers took part in workshops and group interviews for this case study. The~r are predominantl~ women, and predominantly black and Hispanic. host ha~.e completed high school and a few have completed college; they ha~re worked for the cits from 2 to 10 years, and have been using computerized office systems for at least a year. The clerical workers spent most of their 7-hour workda~s using the s~sLems, and usuall~ hate two 20-nlinute breaks, in addition to 1 hour for lunch.
PAGE 282
276 l Automation of Americas Offices One goal of city government is reduction of the clerical work force and this has been occurring. Managers report that the objective is to reduce the clerical work force about 15 to 20 percent through attrition. The effects of office automation cannot always be clearly separated from the effects of reorganization that either preceded or accompanied it, but much of that reorganization was itself for the purpose of making better use of the new technology. One work unit manager reported that the automated systems have already resulted in a 15 to 20 percent reduction in clerical staff over 3 years. Clerical workers said that when automation began, they were fearful of losing their jobs and afraid that they could not learn the new systems. Minimal formal training was provided by vendors. Most of the workers learned from coworkers, and then taught others. Now the workers generally approve of the automation, saying that it freed them from tedious manual work and allowed them to learn new skills, and that they can now produce better quality work, at a faster rate. They also say, however, that they still need more and better training and that they are not being adequately compensated for what they perceive as their newly learned and higher level skills. Because the new systems can handle a larger volume of work in a shorter period of time, most of the workers work more steadily and take fewer breaks. 39 Those doing data entry in many cases say that their control over their work has decreased because it is paced by the machine. However, those in one kind of decentralized word processing pool (clusters now have relatively more variety in what they are keyboarding and also have to prioritize it, so that they feel they have more control over their work. 40 Their union, AFSCME, has negotiated two 20-minute breaks for VDT users in some work units, but the workers say that they do not always take them because they are too busy. Some, however, say that they take informal breaks when necessary to relieve eyestrain and back discomfort. In two word processing units, one a pool and one a cluster, some similar effects were found, including rotation of jobs and prioritization of work; but for different reasons. The cluster is a smaller unit where management encourages worker involvement in everything from selecting equipment to organizing the work. In the word processing pool the increased control that the workers have is due to a virtual lack of supervision. As a result, workers must work as a team to coordinate and organize the units work. Most of the New York City clerical workers say that even though they like the new systems, they also suffer from increased stress. They report increased eyestrain, backaches, and headaches. Many of the new systems have been installed in old offices with unsuitable furniture and lighting, and poor ventilation; excessive noise from printers, and wires or cables stretched haphazardly around the floor, add to the risks to health and comfort. Almost all of the clerical workers do more work than before automation. Some say that they now have a better understanding of the whole picture, that is, what the agency does and how. The sharing of information and cooperative learning that has occurred has generally increased the interaction between coworkers. But frequently this interaction is forced; it is an attempt to overcome problems caused by lack of supervisory coordination, lack of formal training, and problems with the equipment. The only way that work gets done is by the workers going into a huddle and finding some way to get around the problem. The workers overwhelmingly say that they have more interest in their work now, although as they fully master the new process some anticipate that it will become boring. In New Yorks city government, paraprofessionals, who have less formal training than professionals, can perform some of the work of professionals and assist them in other tasks. Some paraprofessionals are now using computers and related devices, although usually less intensively than the clerical workers. Typically, they may review computerized files or scan data for 2 or 3 hours a day interspersed with other duties and activities. 41 For example, in the Bureau of Child Support (HRA) preassignment investigators are now reviewing forms, generated by computers, to track down information about absent parents who are evading child support payments. In the Income Maintenance Office (also HRA) eligibility specialists use computers to review client Nine paraprofessionals participated or were interviewed for this case study, They were predominantly black and Hispanic women between the ages of 30 and 50, all high-school graduates, most with some college training, They had worked for the city from 5 to more than 15 years and are now using automated data-entry and or data-retrieval s}stems in their work.
PAGE 283
Ch. 10Office Automation in State and Local Governments l 277 files for recertification of public assistance eligibility. The eligibility specialists also have extensive direct contact with clients by telephone; they help them with emergencies such as evictions, fires, or illness. These paraprofessionals rate the work they do with computers for about 10 hours a week as the most satisfactory of their tasks. They report that their productivity has increased and the quality of the output improved after automation. The workload has also increased; because more information is available they are now required to submit more reports, and their deadlines are tighter. Although many feel that they have a better sense of the whole picture and more control over their work, many feel that they are overworked, undercompensated, and that the quality of services delivery has not necessarily improved. The preassignment investigators in the Bureau of Child Support also rate their reviewing of files by computer as the most satisfying of their tasks. They have almost no direct contact with clients, using forms almost exclusively in processing cases. But while they find reviewing these forms more satisfying than their other major task of sending out form letters, automation has not, in their opinion, improved the quality of the services. The paraprofessional workers also complain of eyestrain, backpain, and headaches, even though they spend much less time in front of CRTs than do the clerical workers. Professionals in these three city departments vary widely in the extent to which they use computers or other forms of office automation, but their jobs have also been strongly affected. For example, in the Department of General Services, analysts in the Commissioners Office of Management and Analysis use Wang VS terminals or IBMs linked to a -Seven of the nine paraprofessionals who contributed directly to this case study have had to get stronger eyeglass prescriptions since they began using a VDT. Fifteen professionals participated in workshops and interviews. They included 9 men and 6 women (80 percent were white), and their ages ranged from 20 to 40, with the average in the mid-20s. All have college degrees and the majority some graduate-school training. A few have worked for the city more than 15 years, but most less than 4 years. mainframe computer for data input or inquiry, or word processing. Tax auditors in the Tax Examinations Unit use IBM PCs for data inquiry and input, and caseworkers in the Bureau of Child Support use IV Phase terminals for similar purposes. Before automation, tax auditors in the Real Property Transfer Tax unit handled all aspects of a tax audit. Particular cases and case followup were assigned to one auditor. As automation is being introduced, this unit, with a staff of seven, is being merged into the much larger Examinations Unit, which now handles eight kinds of municipal taxes. This represents the beginning of a larger reorganization in which all tax auditors in the unit will be trained to handle all eight types of taxes. Auditors acknowledge that the computer generates more information, quicker, and thus provides more control over the status of each case. Yet the auditors are dissatisfied with the overall process. With the reorganization of the work unit, the tax audit procedure has become fragmented and more clericalized. Some of the auditors are frustrated with what they perceive as a decreasing need for their professional training and judgment. Analysts in the Office of Management Planning and Analysis say that the improved quantity, quality, and accessibility of data has resulted in improved services from their office. Unlike other employees participating in the case study they do not see an increase in work because of computers. Rather, the computers have changed their perceptions of their jobs. The expansion in the amount, speed, and wider range of reliable information has given the analysts more varied and creative possibilities for solving problems. They feel that their influence has increased as they go beyond identifying problems and become more involved in working out solutions. But caseworkers in the Bureau of Child support think that the services provided by their unit have definitely not been improved by Automated Forms Service. Some of their clients, they report, ignore all standardized forms, others are intimidated by computerized forms and will not respond to them. The caseworkers say that the data fed into the sys~2644 O 85 10
PAGE 284
278 Automation of Americas Offices tern is often inaccurate, and that, thanks to automation, errors are not caught until late in the progress of a case, which must then be held up until the errors can be corrected. Also the system has generated increasing numbers of small procedures that not only increase the amount of work but also the chance for errors. These caseworkers, before the automated forms service was adopted, wrote or typed numerous letters. They no longer perform such clerical duties, 44 but paradoxically their position is being deprofessionalized. Since they are no longer solely responsible for deciding what letters to send out, their control over a case throughout its history is not as complete. Their job title is being changed from caseworker to eligibility specialist. A college degree is no longer required, and the new job title does not have the range of possible grade levels that went with the old title, so that opportunities for promotion are decreased. This job degradation was not caused by office automation, but office automation was used to foster an ongoing process. The professionals as a group have mixed perceptions about other effects of automation on their work. They generally report that interactions with coworkers increased, with sharing of ideas about uses of the systems. They continue to pace their own work to suit themselves and the needs of a particular project, as professionals generally can. Some report an increase in petty supervision, apparently because supervisors can review more drafts of reports and more easily ask for changes. Some see the increased use of data (because more is available) as an increase in their workload, while others perceive this in terms of more options for solving problems. Most think that their ability to perceive the whole picture of their agencys work is enhanced; yet some say this is marred by increased uncertainty as to the long-range effects of computerization on government. Both caseworkers and auditors, however, feel that they have lost some control over their work, because it has been fragmented. Caseworkers, by being relieved of clerical aspects of their work have also been relieved of knowledge about its progress and outcome. Crucial steps are taken by other people who do not know the whole story. Auditors do not always complete an audit they have initiated, as it may be passed on to others. Both the caseworkers and these auditors say that their overall job interest has decreased. Managers, those top-level administrators who set broad policies and exercise overall responsibility for their execution, have also been affected by office automation, although they may or may not use any microelectronic devices themselves. 45 Some of the top-level managers in these departments use word processors for memos, notes, drafting materials, etc. Some use electronic mail and messaging. Most also use database inquiry from time to time; a few reported using spreadsheeting, graphics, list files, and typesetting functions. These managers used a computer from 2 or 3 hours a week to as many as 15 hours. Mostly they had learned by watching others use the systems, but a few had taken courses. Buying microcomputers for managers to use was said to be much harder to justify to the city than buying large systems for clerical use. Most of the purchases were approved on an experimental basis. But all of the managers said that computers increased the speed and improved the quality of their work, and allowed them to do new kinds of work. Almost all also said that computers increase the amount of work to be done. The managers, however, were more concerned about the effect of office automation on their departments than on their own work habits. Some were eager to use office automation to reduce the work force and their operatTheir work was rationalized and restructured to relieve them of some of the paperwork, prior to automation. The clerical workers who took o~er the paper work became overloaded, and Automated Forms Service was brought in to relieve this problem. r Seven managers were interviewed in-depth; they include a department commissioner, an assistant commissioner, a deputy general counsel, and the directors of four major offices within departments. They included 4 men and 3 women, between the ages of 28 and 50, who have worked for the city between 1 and 15 years. Their own use of computers varied from none, to working (at home) for several years on a computer.
PAGE 285
Ch. 10Office Automation in State and Local Governments l 279 ing costs. Others felt that the combination of civil service rules and union resistance would keep that from happening in the near future, but that other changes set in motion by automation would nevertheless have major impacts on the delivery of government services. Said the Commissioner of a large municipal department: UNANSWERED QUESTIONS TECHNOLOGY AND Since the city civil service ensures that jobs and people stay, I feel that this offers us room to experiment with using computer tools for new uses. I see the line between clerical and managerial workers blending. A clerical worker is not just a clerk and a manager is not just a manager. ABOUT INFORMATION GOVERNANCE In all public sector offices, office automation is increasingly tied to and part of larger information systems, and the effects on quality of governance must be considered as a whole. In State and local governments this is particularly important because this is the level of government most likely to impact directly on individuals and households on a daily basis. To the extent that this assessment has considered public sector offices, it has been directed almost entirely to the effects of office automation on government itself, its productivity, effectiveness, and efficiency. There has been only peripheral attention to the ultimate effects on the constituents and clients of government offices. A study of 42 cities, in the first phase of office automation from 1975 to 1979, found improved administrative control and operating performance, increased job satisfaction, and improvements in the offices as working environments. 46 The survey also produced evidence that office automation reinforced existing patterns of bureaucratic power, and in most cases, this meant a trend toward centralized, professionalized management at the expense of the power of elected officials. This was one of very few cross-jurisdictional studies and also one of even fewer that looked at effects on power relationships within government. 1: taluatif)n of I nf{)rrnation Technolog~ in I.ocal (lo~ernments, 1 W1979, a sur~e} of the urhan information s~:stems rest~arch ~rr(jup of the I)uhl ic Polic~ Research ( )r~aniz at Ion, t-e ported in Kt-aemer, et al,, op. cit., pp. Yiff, To date, the sparse and fragmented literature on government office automation and on the broader topic of government information systems, and the case studies reviewed above, suggest some further questions that should be thoroughly studied. If information and communication technologies can, as now appears highly probable, increase the efficiency of State and local governments and decrease their operating costs, what can be done to help the citizens of small as well as large governmental units enjoy these benefits? How can office automation be implemented and managed so that it improves, rather than degrades, the work life of civil servants and thus attracts to government service capable and dedicated people? Can office automation improve the effectiveness and responsiveness of government, as well as its efficiency? Several aspects of governance on which office automation has a direct bearing were beyond the scope of this assessment, but are particularly worth further consideration. One is the ability of governments to gather essential information needed to carry out their responsibilities effectively. A second is the ability of citizens to know and understand what their governments are doingi.e., access to information. A third is the ability of citizens to withhold some personal information from governmenti.e., civil liberties. Finally, there are questions about how information technologies affect the nature and the equitable distribution of government services.
PAGE 286
280 l Automation of Americas Offices This asses |