Bulletin 781
VARIATIONS IN NATURAL RESOURCE
INVESTMENTS AND INCOME GROWTH AMONG
AREAS OF THE SOUTHEAST
B. R. Eddleman and James C. Cato
Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
University of Florida, Gainesville
J. W. Sites, Dean for Research
In cooperation with the
U. S. Department of Agriculture
July 1976
VARIATIONS IN NATURAL RESOURCE INVESTMENTS
AND INCOME GROWTH AMONG AREAS
OF THE SOUTHEAST
B. R. Eddleman and James C. Cato
B. R. Eddleman was formerly professor of food and resource
economics, University of Florida. James C. Cato was formerly
agricultural economist, NRED-ERS-USDA, and is assistant profes-
sor of food and resource economics, University of Florida.
AGRICULTURAL EXPERIMENT STATIONS
Institute of Food and Agricultural Sciences
University of Florida
J. W. Sites, Dean for Research
In Cooperation With
Natural Resource Economics Division
Economic Research Service
U. S. Department of Agriculture
Contributing project to Southern Regional Research Project S-71
-"Income and Employment Effects of Public Investments in Na-
tural Resources."
TABLE OF CONTENTS
Page
LIST OF TABLES ----------------------------------------iii
LIST OF APPENDIX TABLES ------------------------------- iii
LIST OF FIGURES ----------------------------------------iv
INTRODUCTION _------------------1
DELINEATION AND CLASSIFICATION OF AREAS---------- 2
NATURAL RESOURCE INVESTMENTS ------------------------ 7
Corps of Enginers Investments------------------------------ 8
Tennessee Valley Authority Investments --------------------- 11
Farmers Home Administration Investments -------------------11
Agricultural Conservation Program Investments ---------------11
Flood Prevention Watershed Program Investments -------------15
Small Watershed Program Investments --------------------- 15
MEASURES OF INCOME CHANGES -------------------19
STATISTICAL TECHNIQUES ------------------------ 20
RESULTS -------------------------- ---34
Correlations of Natural Resource Investments with
Income Changes ----------------------------------- 34
Resource Investment Impacts on Income Changes --------------41
SUMMARY AND CONCLUSIONS ------ ---------------- 52
Summary ----------------------------------- 52
Conclusions --_------- ---------_------ 54
APPENDIX ---------------------------------- 57
LITERATURE CITED ----- ---------------------------57
ii
LIST OF TABLES
Number Page
1. Values of the discriminant functions for grouping
counties into homogeneous areas, 1950 -------------------------- 3
2. Values of the discriminant functions for grouping
counties into homogeneous areas, 1960 ------------------------- 3
3. Average values for selected characteristics of counties
comprising area groups ---------------------_________________ 6
4. Correlation coefficients between natural resource investments
and income changes by areas of the Southeast, 1950-1970 -------- 36
5. Correlation coefficients between natural resource investments
and income changes by areas of the Southeast, 1960-1970 -------- 38
6. Partial regression coefficients for impact of natural resource
investments on income change by areas of the Southeast,
1950-1970 ----- ---------------------- 42
7. Partial regression coefficients for impact of natural resource
investments on income change by areas of the Southeast,
1960-1970 ----------------------------- 46
LIST OF APPENDIX TABLES
Number Page
1. Correlation coefficients between total water investments
and income changes by areas of the Southeast, 1950-1970
and 1960-1970 ----------------__--_____---- __--__ ---- 57
LIST OF FIGURES
Number Page
1. Delineation of counties in nine Southeastern states into
four groups using 1950 data -------------------------------- 4
2. Delineation of counties in nine Southeastern states into
four groups using 1960 data -------------------------------- 5
3. Corps of Engineers civil works new work construction,
1950-1970 --------------------------------------- 9
4. Corps of Engineers civil works new work construction,
1960-1970 -------------------------------------------------- 10
5. Tennessee Valley Authority investments in water resource
projects and steam generating plants, 1950-1970 ------------ 12
6. Tennessee Valley Authority investments in water resource
projects and steam generating plants, 1960-1970 --------------- 13
7. Total loans and grants for community water and sewer systems
and recreational facilities by the Farmers Home Administration,
1962-1970 --------------------------------------------------- 14
8. Total payments in the Agricultural Conservation Program by
the Agricultural Conservation and Stabilization Service,
1950-1970 --------------------- 16
9. Total payments in the Agricultural Conservation Program by
the Agricultural Conservation and Stabilization Service,
1960-1970 --------------------------------------------------- 17
10. Total construction investments in the Flood Prevention
Watershed Program by the Soil Conservation Service,
1954-1970 ---------------------------------------- -----18
11. Total construction expenditures in the PL-566 Small Watershed
Program by the Soil Conservation Service, 1958-1970 --------- 21
12. Change in per-capita income, 1950-1969 ----------------------- 22
13. Change in per-capita income, 1959-1969 _--------------------- 23
14. Change in total personal income, 1950-1969 ------------------- 24
15. Change in total personal income, 1959-1969 -------------------- 25
16. Change in total earnings, 1950-1969 -------------------- 26
17. Change in total earnings, 1959-1969 --------------------------- 27
18. Change in total farm earnings, 1950-1969 _-------------------- 28
19. Change in total farm earnings, 1959-1969 ---------------------- 29
20. Change in private manufacturing earnings, 1950-1969 ---------- 30
21. Change in private manufacturing earnings, 1959-1969 ---------- 31
22. Change in private construction earnings, 1950-1969 -------------32
23. Change in private construction earnings, 1959-1969 -------------- 33
VARIATIONS IN NATURAL RESOURCE INVESTMENTS
AND INCOME GROWTH AMONG AREAS OF THE
SOUTHEAST
B. R. Eddleman and James C. Cato
INTRODUCTION
Considerable emphasis has been placed on the role of natural
resource investments in stimulating growth in income of recipi-
ent areas. Natural resource investments can influence economic
growth in recipient areas, particularly in those areas where other
sufficient conditions for economic growth exists. Because all
areas do not respond to investment stimuli in the same way, the
effects of natural resource investment programs will differ
among areas with diverse characteristics. Little is known about
the relationships between natural resource investments and in-
come growth in recipient areas of the Southeast. This study is
one phase of a regional research project to evaluate the impact
of natural research investments on regional income, employment,
and composition of economic activity.
Relationships are examined between various kinds and
amounts of natural resource investments and changes in income
of recipient areas that were differentiated into four groups of
relatively homogeneous areas. Several natural resource endow-
ment, economic, social, and demographic characteristics were
used to delineate the four groups of areas in the Southeast. Then,
the relationship of variation in income changes among counties
within each group of areas and natural resource investments
was examined. The time periods under consideration were the
two decades between 1950 and 1970 and the decade of 1960 to
1970.
Data were from both primary and secondary sources, and in-
cluded measures of natural resource endowments, population,
demographic characteristics of the population, employment, in-
come, and expenditures for various types of natural resource
investment programs. Observational units were counties. Six
separate measures of income changes were. considered. These
measures differentiated among the potentially affected indus-
tries of agriculture, manufacturing, and construction, as well as
changes in per capital income, total personal income, and total
earnings. Six types of natural resource investments were consid-
ered. These included new civil works construction expenditures
by the Corps of Engineers, Tennessee Valley Authority (TVA)
investments, Farmers Home Administration (FHA) expenditures
for community water and sewer systems and recreational facili-
ties, Agricultural Conservation Program (ACP) payments by the
Agricultural Stabilization and Conservation Service (ASCS),
Flood Prevention Watershed (FPW) construction expenditures
by the Soil Conservation Service (SCS), and Public Law-566
Small Watershed construction expenditures by the SCS.
DELINEATION AND CLASSIFICATION OF AREAS
The 755 counties of the nine contiguous southeastern states1
were delineated into four groups of relatively homogeneous areas
for the years 1950 and 1960. The procedure used in delineating
the groups of areas was a multiple discriminant analysis as out-
lined by Martin (4).2 Tintner (5) provides an additional discus-
sion of the discriminant analysis technique. Variables used in
grouping the counties into areas and the values for these vari-
ables and for the discriminant functions are presented in Tables
1 and 2. The coefficients in the tables indicate the linear com-
bination of the differences of the mean values of the county
variables which best discriminates3 between the four sets of
variables.
The values in Tables 1 and 2 reveal that the largest weight in
the discriminant functions generally has been given to the
median age variable followed by the education and total land
area variables. This seems to indicate that median age, education
levels and size of county are possibly more important than other
characteristics in distinguishing between the county groups.
Each of the discriminant functions in Tables 1 and 2 was statis-
tically significant at the 5 percent level based on the F-distribu-
tion test of the variance ratio.
Geographic distributions of the four groups of areas for 1950
and 1960 are mapped in Figures 1 and 2, respectively. Group 1
areas are the most urban-oriented of the four groups with an
average of 253 people per square mile in 1960. Almost 63 percent
of the population in these areas was classified as urban (Table
3). The highest proportion of persons 25 years old and over with
a high school education is found in these areas. The proportions
'The nine southeastern states include Alabama, Arkansas, Florida, Georgia,
Louisiana, Mississippi, North Carolina, South Carolina, and Tennessee.
'Numbers in parentheses refer to literature cited.
3The term "best discriminates" refers to the linear function of the dif-
ferences of the means which discriminates most successfully in that its
square is maximum relative to its variance.
Table 1.-Values of the discriminant
into homogeneous areas, 1950.
functions for grouping counties
County Groups
County Variablest
1 2 3 4
Total population .1039 .0712 .0654 .0597
Urban population -.0236 -.0204 -.0105 -.0204
Percent persons 25
years old and older
with high school education 2.4974 1.4132 1.3552 .8112
Median age 5.7647 5.6168 5.0051 5.1455
Total employment -.2503. -.1758 -.1792 -.1350
Agricultural employment .2812 .3161 .5215 .2810
Manufacturing employment .1893 .1546 .1300 .0989
Total land area 2.0249 1.6931 1.5074 1.1374
Land in farms .0182 -1.2297 -.1117 -1.6018
Value of farm products sold -.1192 -.1135 -.1627 -.0984
tAdditional variables that were removed from the 1950 functions because
of relatively small weights included value added by manufacture, retail trade
sales; wholesale trade sales; personal, business, and repair services receipts;
total crop land; and total woodland.
Table 2.-Values of the discriminant
into homogeneous areas, 1960.
functions for grouping counties
County Groups
County Variablest
1 2 3 4
Total population -.0005 -.0023 .0005 -.0022
Urban population .0078 -.0005 .0058 .0014
Percent persons 25
years old and older
with high school education 2.7407 1.9299 1.6742 1.4358
Median age 2.1706 2.2668 1.8126 2.0138
Total employment -.0270 -.0085 -.0308 -.0108
Agricultural employment .0366 .0623 .2753 .0978
Manufacturing employment .0524 .0495 .0438 .0359
Total land area 2.0478 1.8295 1.3909 1.2217
Land in farms 1.5937 -.0600 2.2323 -.2205
Value of farm products sold -.0168 -.0173 -.0356 -.0157
tAdditional variables that were removed from the 1960 functions because
of relatively small weights included value added by manufacture; retail trade
sales; wholesale trade sales; personal, business, and repair services receipts;
total crop land; and total woodland.
of employment in agriculture and land in farms are relatively low
in these areas.
Figure i.-- Delineation of counties in nine Southeastern
states into four groups using 1950 data.
Group 1 Urban
Group 2 Urban-rural
Group 3 | Densely settled rural
Group 4 | Sparsely settled rural
9. r i i
F;gure 2. --
Delinea;ton of count ies n nine Sou[heasc[ern
Estates into four groups us;ng 1960 data.
Group I Urban
Group 2 Urbar-rural
Group 3 Densel, ,etcled rural
Group 4 Sparsely settled rural
Table 3.-Average values for selected characteristics of counties comprising area groups.
Area Population Urban Schoolingtt Agricultural Land in
Group Density Populationt Employmentttt Farmstttt
Persons/sq. mi. Percent Percent Percent Percent
.......................................... 1950 ................ .........................
1 185 56.6 31.6 12.6 33.9
2 49 25.0 18.1 29.1 37.2
3 51 17.9 15.8 49.6 47.6
4 36 8.3 11.5 43.3 39.3
......................................... 1960 ................ .........................
1 253 62.7 41.0 6.7 28.3
2 59 30.1 27.1 12.3 27.4
3 52 25.6 22.8 29.4 43.4
4 37 13.6 19.8 21.9 32.5
tPercent of total population classified as urban.
ttPercent of total population 25 years of age and over with a high school education.
tttPercent of total labor force employed in agriculture.
ttttPercent of total land area in farms.
Most of northeast Florida and practically all of central and
southern Florida are included in group 1 (Figures 1 and 2). In
the other eight southeastern states, major metropolitan areas
and their coterminous urban fringe, as well as major cities that
are relatively isolated from these major urban centers, comprise
the areas of group 1.
Group 2 areas have considerably less urban orientation than
the areas in group 1. These areas are generally clustered adja-
cent to group 1 areas, but they are much more sparsely populated
and exhibit more of a rural structure than group 1 areas. Group
2 areas averaged 59 persons per square mile in 1960 with only
30 percent urban population. The proportion of employment in
agriculture was about double that of group 1 areas; however, the
proportion of land in farms averaged slightly less than for group
1 areas in 1960 (Table 3). Group 2 areas cover large parts of
Arkansas, Louisiana, western North Carolina, northwestern
South Carolina, northwestern Florida, and isolated parts of the
other southeastern states.
Group 3 areas are the more densely settled rural areas with
an average of 52 persons per square mile in 1960. These areas
contain cities as central places that are smaller than the more
urban-oriented groups. In 1960 slightly over one-fourth of their
population was urban. Group 3 areas exhibit the highest average
proportions of employment in agriculture and land in farms.
These areas cover large parts of Alabama, Mississippi, the
coastal plains of North Carolina and South Carolina, and the tier
of counties in Arkansas and Louisiana that lie just west of the
Mississippi River. Isolated counties and multi-county areas in
group 3 are found in the other southeastern states.
Group 4 areas are sparsely populated rural areas with small
towns as central places. Areas included in this group averaged
37 persons per square mile in 1960 with only 13.6 percent of the
population classified as urban. Over one-fifth of the labor force
was employed in agriculture and an average of almost one-third
of the land areas was in farms. Group 4 areas are found exten-
sively in Georgia, northeast and southeast Louisiana, and west-
ern and east central Tennessee.
NATURAL RESOURCE INVESTMENTS
The impact of the six types of public investment in natural
resources has both a time and space dimension. 'This study fo-
cuses on the space dimension of the impact among the four ma-
jor groups of areas in the Southeast for a specified time span.
Total project investments over the designated periods of time
were used to delineate areas on the basis of whether or not a
county received any investments, or received an amount greater
than or less than the average county investment for all recipient
counties in the Southeast.
Corps of Engineers Investments
Investments by the Corps of Engineers in new civil works
construction over the period 1950 and 1970 were obtained for
each county from the various district offices which administer
portions of the nine-state region. These investments covered
projects categorized into multipurpose, navigation, flood control,
beach control, and recreation projects. The major portion of ex-
penditures was for flood control and navigation with a very small
proportion allocated to beach control and recreation.
Counties were mapped on the basis of total investment in the
projects. The categories included no investment, from one dollar
to the mean county investment for all of the region, and invest-
ments greater than the average county investment for the
region. The geographic distributions of investments by the Corps
for 1950-1970 and 1960-1970 are mapped in Figures 3 and 4, re-
spectively. Comparisons between these two periods show little
change in the geometric classifications. Countries that were
major recipients during the long 1950-1970 period (i.e., counties
receiving greater than the average county investment for all
recipient counties) also were major recipients in the latter 1960-
1970 period. Additionally, very few new counties were added to
the Corps program during the 1960-1970 period, while the con-
struction phase was completed in some of the counties during the
1950-1960 period.
Florida, Louisiana, Arkansas, and the northwest portion of
Mississippi contain a substantial majority of the recipient coun-
ties (Figures 3 and 4). Other identifiable concentrations of re-
cipient areas include counties in western Alabama, the Coastal
Plains of North Carolina and South Carolina, isolated areas of
Tennessee and Georgia, and a portion of northwest South Caro-
lina.
Comparisons of Figures 3 and 4 with Figures 1 and 2, respec-
tively, reveal that a substantial portion of the Corps invest-
ments were concentrated in group 1 and group 2 areas, the ur-
ban-oriented counties. This is particularly noticeable for counties
in Florida, Arkansas, and parts of Louisiana.
Figure 3. -- Corps of Engineers new civil works
construction, 1950-1970.
None
$1 to $7,931,700
$7,931,800 and greater
I~~*~X
L
E"rn~x
Figure 4. -- Corps of Engineers new civil works
construction, 1960-1970.
Tennessee Valley Authority Investments
Investments by the TVA were for multipurpose, navigation,
flood control, and steam generating projects. Annual investment
data were obtained from agency officials.
The geographic distributions of TVA investment expenditures
are shown in Figures 5 and 6. Counties receiving TVA invest-
ments were concentrated in northern Alabama and eastern Ten-
nessee. Other areas receiving investments included a few coun-
ties in western North Carolina and central Tennessee, as well as
Union County in north Georgia. The locations of investments
were little changed between 1950-1960 and 1960-1970 with the
most noticeable change being completion of investment projects
in some of the eastern Tennessee and western North Carolina
counties during the period 1950-1960.
Farmers Home Administration Investments
Data on loans and grants by the FHA for community water,
sanitary sewer, and solid waste disposal systems were obtained
from the state directors of the various FHA offices. This pro-
gram provides financial assistance to communities in developing
essential new public service facilities and in expanding existing
facilities. The time period of investments by the FHA for these
purposes was 1962-1970. Included in the investment data were
relatively small additional grants to assist in the projects by other
federal agencies such as the Economic Development Administra-
tion. It is possible that some of these grants had not actually
been expended, particularly during the latter years of the time
period.
Figure 7 shows the geographical distribution of FHA loans
and grants for the period 1962-1970. A substantial number of
counties in each state received FHA assistance with the notice-
able concentration of major recipient counties in Mississippi. In
terms of those counties not participating in the FHA program,
the most noticeable concentrations are in rural areas of north-
east Florida, the west Florida panhandle, northern Georgia, and
parts of south Georgia and southeast Louisiana.
Agricultural Conservation Program Investments
Investments by the ASCS constitute a joint effort by the pub-
lic sector, farmers, and ranchers, to share the cost of establishing
needed conservation measures. These agricultural conservation
programs include practices to protect, improve, and renew soil,
Figure 5. --
Tennessee Valley Authority investments in
water resource projects and stream generating
plants, 1950-1970.
None
$1 to $62,112,000
$62,113,000 and greater
Figure 6. --
None
Tennessee Valley Authority investments in
water resource projects and steam generating
plants, 1960-1970.
LiI
$1 to $49,949,000
$49,950,000 and greater
Total loans and grants for community water
and sewer systems and recreational facilities
by the Farmers Home Administration, 1962-1970.
$800,000 and greater
--
ROMMINW-
water, woodland and wildlife resources of private land owners.
Data were obtained from annual state ASCS reports and com-
puter printouts during the 1950-1970 period. Expenditure infor-
mation included cash payments to farmers and allowances paid
to vendors for conservation materials furnished to farmers.
The geographical distributions for ACP payments for each of
the periods are shown in Figures 8 and 9. Almost all of the coun-
ties in the nine southeastern states participated in the program
during the 1950-1970 period. Major recipient counties were con-
centrated in Mississippi, eastern Arkansas, the western and west
central areas of Tennessee, northwest and southwest Louisiana,
northern and southern Alabama, and portions of North Carolina
and South Carolina.
Flood Prevention Watershed Program Investments
The Flood Prevention Watershed Program by the SCS was
concentrated in only two of the nine southeastern states. These
projects were located in northern Mississippi (Yazoo and Little
Tallahatchie Flood Prevention Watersheds) and northern Geor-
gia (Coosa Flood Prevention Watershed) as shown in Figure 10.
The programs involved construction of flood prevention struc-
tures, reforestation, and other soil conservation practices. Data
on annual investment levels for the two areas were obtained from
:state offices of the SCS for the period 1954-1970.
Small Watershed Program Investments
Construction expenditures by county were available only for
the period 1958-1970 for the Small Watershed Program and were
obtained from the personnel in each of the state offices of the SCS.
Data were tabulated from 239-B forms which provided actual
dates of construction expenditures for each project. Investments
were then allocated to each county based on project location as
identified on maps obtained from the SCS.
The Small Watershed Program is designed to aid in the solu-
tion of several types of problems. Reduction of floodwater dam-
ages to cropland, residences, businesses, and the protection of
health and lives of people from floods receive major importance.
Other potential and existing problems that this program at-
tempts to alleviate include erosion and sediment damage, im-
proper drainage, and irrigation needs. Recreation, fish and wild-
life enhancement, and improvements in the economic and social
well-being of people also have received emphasis, particularly in
recent years.
Total payments in the Agricultural Conservation
Program by the Agricultural Conservation and
Stabilization Service, 1950-1970.
Total payments in the Agricultural Conservation
Program by the Agricultural Conservation and
Stabilization Service, 1960-1970.
Figure 10. --
None
Total construction investments in the Flood
Prevention Watershed Program by the Soil
Conservation Service, 1954-1970.
I I
$1 to $1,802,100
$1,802,200 and greater
The geographical distribution of PL-566 Small Watershed Pro-
gram investments during the period 1958-1970 is mapped in
Figure 11. Counties in each of the nine southeastern states have
received these types of investments. The most noticeable concen-
trations of recipient counties are in the coastal plains of North
Carolina, northern South Carolina and Georgia, central Mississ-
ippi, and the tier of counties east and west of the Mississippi
River in Arkansas, Mississippi, and Louisiana. Other recipient
areas include portions of northwest Arkansas, western Tennes-
see, northwest and southwest Louisiana, and central and south-
ern Florida.
MEASURES OF INCOME CHANGES
Six separate measures of income changes for each time period
were developed. Three of these related to total income changes
for all groups of people in the counties and included both aggre-
gate and per capital measures. Changes in per capital income,
total personal income, and total earnings comprised these mea-
sures of income change. Areas were delineated on the basis of
county income changes less than the average change and the
mean income range or above. These income changes are mapped
for each of the study periods in Figures 12 through 17.
General observations can be made concerning the geographical
incidence of changes in incomes. There was little difference in
the groups of counties that were major recipients of income
during the two periods. That is, counties that had above average
changes in per capital income, total personal income, or total
earnings during the 1950-1970 period also exhibited the greatest
absolute growth in these measures of income during 1960-1970.
In general, the largest changes in these measures of income oc-
curred in counties classified in group 1 and group 2 areas, the
urban oriented counties and areas clustered adjacent to them.
The remaining three measures of income change were included
to differentiate among industries that potentially could be di-
rectly affected by the particular types of investments considered
in the study. They included measures of changes in total farm
earnings, total private manufacturing earnings, and total private
construction earning. Geographical distribution of changes in
these industry income measures are shown in Figures 18 through
23.
Group 3 and group 4 areas tended to exhibit substantial farm
income changes in much of North Carolina, South Carolina,
Georgia, Alabama, and Mississippi during 1950-1970. In contrast,
a large number of group 2 counties in Arkansas and Louisiana
exhibited declines in total farm earnings during this period. In
Florida, the citrus and winter vegetable areas in the central
and southern portions of the state showed the greatest gains in
farm earnings. Similar patterns of the counties in farm earnings
occurred during 1960-1970. Some of the counties that experienced
declines in farm earnings during 1950-1960 did show gains dur-
ing 1960-1970.
Counties having the greatest gains in private manufacturing
earnings during both periods were generally found in the urban
oriented group 1 and group 2 areas. However, large numbers of
rural counties in group 3 and group 4 areas also experienced ab-
solute increases in private manufacturing earnings. A similar
pattern of changes in private construction earnings occurred
during both time periods.
STATISTICAL TECHNIQUES
The basic method for evaluating the variations among changes
in income and natural resource investments entailed the use of
historical cross-sectional data for the 755 counties in the South-
east Region. Two general time-dated analyses were conducted
in which 1950-1970 or 1960-1970 changes in natural resource in-
vestments were related to each of the six measures of income
change for the four separate groups of areas. Income data were
measured in absolute dollar changes and natural resource invest-
ment data were measured in absolute total dollar amounts ex-
pended during the specified time period.
Statistical techniques used were simple correlation analysis
and multiple linear regression analysis. The simple correlation
coefficients merely estimate the degree of closeness of the linear
relationship between a particular type of natural resource invest-
ment and a particular measure of income change. To ask whether
the relation between two such variables is close or loose may be
sufficient at an early stage. But the more important questions
are: How much does income change for a given amount of in-
vestment? How accurately can income changes be predicted from
investment levels? These types of questions can more appropri-
ately be analyzed with regression techniques.4
4A study by Cato (1) and an accompanying publication by Cato and Eddle-
man (2) examine these types of relationships in much greater detail than
provided in this publication.
Figure 11. -- Total construction expenditures in the PL-566
Small Watershed Program by the Soil Conser-
vation Service, 1958-1970.
None
$1 to $350,700
$350,800 and greater
Figure 12. -- Change in per-capita income, 1950-1969.
Less than $1,752
$1,752 and greater
Figure 13. -- Change in per-capita income, 1959-1969.
Less than $1,334 l-
$1,334 and greater
I
~: r~r~~~,~~lPnlS~~ ~g~BBB~h', i' \'
~
\
r:go-: \"
\ ,\. ~BLI ~ r .
- t
Figure 16, -- Change in total earnings, 1950-1969.
Less than $76,933,400
$76,933,400 and greater
Figure 17. -- Change in total earnings, 1959-1969.
Less than $54,899,900
$54,899,900 and greater
W2~~o
Figure 18. -- Change in total farm earnings, 1950-1969.
Negative
$0 to $1,448,500
$1,448,600 and greater
III
s\v
tK^/
N".'
Y-7lUO
IN^
:1
-7
Figure 19. -- Change in
Negative
$0 to $1,534,000
$1,534,100 and greater
total farm earnings, 1959-1969.
EZK
Figure 20. -- Change in private manufacturing
earnings, 1950-1969.
Data not available
Less than $26,380,700
$26,380,700 and greater
Figure 21. -- Change in private manufacturing
earnings, 1959-1969.
Data not available I
Less than $19,251,800
$19,251,800 and greater
Figure 22. Change in private construction
earnings, 1950- 1969.
Data not available ..
Less than $7,117,500
S7.17,500 and greater
c us n a
as^ zern assacr
;'oo sch\~"h4~~i'l~~ E: M ?. pt~8B~l
wr" a """
cowan se am0 ac
v sa" o w 1 I0 E c e a
au a wo
Au n r oso r aruo a rca n a -t
~~~u~~~~f~~~~Crmu :: Lacao\8~~oL~~lPBBY~~ 2
Is LL t eo M or
inl
Figure 22. -- Change in private construction
earnings, 1950- 1969.
Data not available .. -
Less than $7,I17,500 ~"
S7.117,500 and greater
Figure 23. -- Change in private construction earnings,
1959-1969.
RESULTS
Correlations of Natural Resource Investments with Income Changes
The simple correlation coefficients between measures of in-
come change and each of the investment categories are given for
each of the four groups of areas for 1950-1970 and 1960-1970
in Tables 4 and 5, respectively. None of the investments consid-
ered were highly correlated with any of the measures of income
change for either time period (simple correlation coefficients of
.382 or less). Additionally, many of the investments were nega-
tively correlated with the measures of income change. However,
with the exception of FHA investments in group 1 urban areas,
most of the negative correlations were not statistically signifi-
cant at the 10 percent level of confidence.
Investments by the Corps of Engineers were associated with
increases in total personal income, total earnings, private manu-
facturing earnings, and private construction earnings in group 1
areas during both periods. All of these correlation coefficients
were statistically significant at the 5 percent level of confidence.
Additionally, Corps investments were significantly correlated (at
the 1 percent level of confidence) with increases in total personal
income, total earnings, and private construction earnings in
group 4 rural areas during the longer 1950-1970 period. Private
manufacturing earnings appeared to be negatively associated
with Corps investments in group 2 areas during both time
periods.
No strong correlations between investments by the TVA and
the measures of income change were found. Group 3 areas ex-
hibited the only significant positive correlations (at the 5 per-
cent level of confidence) of TVA investments with per capital
income changes during the 1950-1970 period and with private
construction earnings during the 1960-1970 period.
Investments by the FHA were significantly correlated with
measures of income change in each of the four groups of areas
with the exception of group 4 rural areas during 1950-1970.
Negative and significant correlations with changes in per capital
income and total earnings occurred during both time periods in
group 1 urban areas. Also a negative and significant correlation
(at 10 percent level of confidence) with changes in total personal
income resulted for the 1960-1970 period. In group 2 areas signi-
ficant correlations (at the 10 percent level of confidence or
above) were obtained for FHA investments and increases in
total personal income and private manufacturing earnings, which
occurred during both time periods, and increases in total earn-
ings during the 1950-1970 period. Group 3 rural areas exhibited
the largest number of significant correlations (at the 10 percent
level of confidence or above) of income increases with FHA in-
vestments. These correlations were with increases in total per-
sonal income, total earnings and private manufacturing earnings
during both time periods and with increases in private construc-
tion earnings during the 1950-1970 period. Increases in total
personal income and private manufacturing earnings were signi-
ficantly correlated (at the 5 percent level of confidence) with
FHA investments in group 4 rural areas during the 1960-1970
period.
Investments by the ASCS in the ACP program were found to
be significantly related to many of the measures of income in-
creases in group 2, group 3, and group 4 areas. In group 1 urban
areas negative correlations between ACP payments and the vari-
ous measures of income changes were obtained, except for total
farm earnings during the 1960-1970 period. In group 2 areas
significant correlations with increases in per capital income, total
personal income, total earnings, private manufacturing earnings,
and private construction earnings occurred for the 1960-1970
period. In group 3 areas significant positive associations between
ASCS payments and changes in total personal income, total
earnings, and private manufacturing earnings resulted for both
periods. Also, a significant and positive correlation with changes
in total farm earnings resulted for the 1960-1970 period. Signi-
ficant correlations between ASCS investments and increases in
per capital income in group 4 rural areas occurred for the 1950-
1970 period. Significant associations with increases in total per-
sonal income, total earnings, total farm earnings, and private
manufacturing earnings resulted for the 1960-1970 period.
Measures of investments in the Flood Prevention Watershed
program by the SCS were available only for the longer 1950-1970
period. In general, the investments were negatively correlated
with each of the measures of income change in all areas. The
only significant correlation resulted with changes in per capital
income in group 3 rural areas, and this association was a nega-
tive correlation.
PL-566 investments were not found to be very highly cor-
related with changes in income. Negative correlations occurred
for group 1 urban areas during both time periods for all income
change measures except total farm earnings. The only significant
correlations occurred with increases in total farm earnings dur-
Table 4.-Correlation coefficients between natural resource investments and income changes by areas of the Southeast,
1950-1970.
Area Group and
Type of Income Change Corps TVA FHA ASCS FPW PL-566
Group 1 (Urban)
Per Capita income .172* .042 -.206** -.126 -.127 -.002
Total personal income .309*** -.056 -.137 -.137 -.059 -.065
Total earnings .314"** -.048 -.180" -.114 -.057 -.063
Total farm earnings .048 -.041 -.033 -.010 -.022 .013
c Private manufacturing earnings .193** .013 -.069 -.062 -.058 -.096
Private construction earnings .382*** -.036 -.093 -.091 .165 -.154
Group 2 (Urban-rural)
Per Capita income -.084 .085 .010 .103 .012 -.048
Total personal income -.088 .010 .121* .104 -.016 -.018
Total earnings -.086 .013 .121* .103 -.013 -.014
Total farm earnings .082 .013 -.017 -.018 -.062 -.066
Private manufacturing earnings -.155** .053 .233*** .197*** -.002 -.046
Private construction earnings .098 -.054 .084 -.006 .064 .094
continued
Table 4. continued-Correlation
the Southeast, 1950-1970.
coefficients between natural resource investments and income changes by areas of
Area Group and Corps TVA FHA ASCS FPW PL-566
Type of Income Change
Group 3 (Densely settled rural)
Per Capita income .034 .188*" .064 .077 -.174** .097
Total personal income -.039 .099 .138* .189"* -.117 .014
Total earnings -.039 .097 .147** .186"* -.112 .016
Total farm earnings .114 .051 -.059 -.001 .050 .175**
Private manufacturing earnings -.001 .077 .296*** .189*" -.081 .120
Private construction earnings -.031 .059 .207** .173" -.101 -.074
Group 4 (Sparsely settled rural)
Per Capita income -.010 .023 -.073 .107" -.061 .053
Total personal income .180*** -.023 -.004 .101 -.022 .048
Total earnings .163*** -.021 -.001 .098 -.023 -.048
Total farm earnings .062 -.041 -.020 .048 .022 .079
Private manufacturing earnings .008 -.021 .008 .046 .039 .044
Private construction earnings .263*** -.035 -.020 .008 -.035 -.008
*Significant at 10% level.
S*Significant at 5% level.
**"Significant at 1% level.
Table 5.-Correlation coefficients between natural resource investments and income changes by areas of the Southeast,
1960-1970.
Area Group and
Type of Income Change Corps TVA FHA ASCS PL-566
Group 1 (Urban)
Per capital income .070 -.010 -.212** -.118 -.031
Total personal income .334*** -.059 -.188* -.145 -.079
Total earnings .338"** -.058 -.229** -.130 -.082
Total farm earnings .114 -.041 -.012 .051 .030
co Private manufacturing earnings .267 -.041 -.163 -.080 -.087
co
Private construction earnings .369*** .008 -.203 -.155 -.137
Group 2 (Urban-rural)
Per capital income -.048 -.007 -.004 .118" -.004
Total personal income -.080 -.032 .121" .321*** .046
Total earnings -.085 -.032 .113 .297*** .044
Total farm earnings .011 -.024 .013 .021 .030
Private manufacturing earnings -.119" -.017 .182*** .354*** .007
Private construction earnings .027 .011 .051 .165"* .029
continued
Table 5. Continued-Correlation coefficients between natural resource investments and income changes
the. Southeast 1960-197f0
by areas of
Area Group and
Type of Income Change Corps TVA FHA ASCS PL-566
Group 3 (Densley settled rural)
Per capital income
Total personal income
Total earnings
Total farm earnings
Private manufacturing earnings
Private construction earnings
Group 4 (Sparsely settled rural)
Per capital income
Total personal income
Total earnings
Total farm earnings
Private manufacturing earnings
Private construction earnings
.012
.079
.067
.104
.028
-.005
.012
.006
.004
.069
-.021
.086
.111
.106
.113
.044
.061
.265""*
.019
.018
.012
-.049
-.004
-.023
-.036
.132"
.129"
.091*
.157*
.095
.012
.123"*
.096
-.077
.140*
-.036
.096
.209***
.198**"
.223***
.173**
.109
.094
.274***
.244***
.170"**
.217**"
.127
-.069
-.026
-.036
.087
-.032
.078
.070
.082
.073
.007
.066
.137"
"Significant at 10% level.
*Significant at 5% level.
-"Significant at 1% level.
ing the 1950-1970 period in group 3 areas and with increases
in private construction earnings in group 4 areas during the
1960-1970 period.
All of these correlations between natural resource investments
and the measures of income change were based on the total
number of county observations in each group of areas. Thus,
counties receiving no investment in a particular type of natural
resource project were included among the observations for each
of the area groups. Correlation coefficients were also computed
in which counties were included only if a particular type of nat-
ural resource investment had been made in the county. The re-
sulting correlation coefficients were essentially unchanged with
respect to their size, algebraic signs, and statistical significance
for all measures of income change and investments by the Corps,
FHA, ASCS, and SCS in PL-566 projects. This was due to the
relative large number of counties receiving these types of in-
vestments during both time periods.
In the case of TVA investments and FPW programs by the
SCS, substantial changes in the size of the correlation coeffi-
cients resulted from a considerable smaller number of county
observations being included in the analysis. Since only 33 coun-
ties received TVA investments and 32 counties received FPW
investments during 1950-1970, the correlation coefficients were
computed for all groups of areas combined as well as for group
2, 3, and 4 areas.' For both types of investments only per capital
income changes were significantly correlated (at the 10 percent
level of confidence) with dollar amounts of project expenditures.
The correlations were positive for TVA investments and negative
for FPW investments. For TVA investments the statistically
significant associations with changes in per capital income had
correlation coefficients of .642, .780, .505 and .553 for the group
2, group 3, group 4, and the combined group of all counties,
respectively. Statistically significant correlation coefficients of
-.478 and -.618 resulted for FPW investments in the combined
groups and group 4 rural areas, respectively. These correlation
coefficients indicate only a moderate level of association between
the investment expenditures and per capital income changes over
the 1950-1970 period. None of the correlation coefficients be-
tween TVA investments and the measures of income change for
only those counties receiving the investments were statistically
,There were only four and two county observations in group 1 urban areas
for TVA and FPW investments, respectively. All other area groups had
nine or more county observations.
significant (at the 10 percent level of confidence) for the 1960-
1970 period."
In summary, a number of the measures of income change were
positively related to investments in natural resources. That is,
investments in natural resources were associated with increases
in incomes among counties. Although the simple correlation co-
efficients were relatively low, (values of .382 or less) the results
indicate that natural resource investments have differential im-
pacts on local economies depending upon the natural resource
base and the economic structure of the recipient area. The fol-
lowing section provides measures of the size of these impacts in
the various groups of areas.
Resource Investment Impacts on Income Changes
Partial regression coefficients reflecting the impact of natural
resource investments on the various measures of income change
for each group of areas during 1950-1970 and 1960-1970 are
given in Tables 6 and 7, respectively. The values in these tables
may be interpreted as the dollar change in the particular income
category over the designated time period resulting from a one
dollar investment in the specified type of natural resource proj-
ect. Very little of the total variation among counties for each
area was explained by the investment variables. In no case did
the investment variables account for more than 20 percent of the
total variation in income changes, and except for group 1 urban
areas, the coefficient of determination (R2) rarely exceeded a
value of .10.
While the total variation in income changes accounted for by
the investment categories was low for all areas, differences ex-
isted among the various groups of counties with respect to the
statistical significance of the investment variables. Some of the
investments that were significant (at the 10 percent level of
confidence or above) for one group of counties were not impor-
tant for other groups. Intercorrelation among the investment
categories was not an apparent problem. The largest simple cor-
relation coefficient among the investment variables had a value
"Members of the S-71 Technical Committee who reviewed the results re-
quested that additional correlation analysis be carried out with the four
types of water resource development investment data aggregated into a
"total water investment" variable. The results were very similar to those
shown in Tables 4 and 5. The correlation coefficients from this analysis
are given in Appendix Table 1.
Table 6.-Partial regression coefficients
the Southeast, 1950-1970.t
for impact of natural resource investments on income change by areas of
Area Group and
Type of Income Corps TVA FHA ASCS FPW PL-566 Constant R-
Change
Group 1 (Urban)
Per capital
income
Total personal
S income
Total earnings
Total farm
earnings
Private
manufacturing
earnings
Private
construction
earnings
-.07"
38.16"*" -6.74
30.76*** -4.70
.22
4.49*
-.18
-194.88 -229.68
-231.51 -107.94
-2.23
.27 -20.76
3.44*** -.21
-4.57
-.10
-13.89
-7.26
-.08
-109.48
-89.27
-2.44
-30.37
1,116.03
-.02
-419.32
-352.43
2,124.00 .09
1,525,938.00
1,161,689.00
2.22 12,523.00 .01
-116.67 254,323.00 .05
-51.49 75,088.00 .20
continued
Table 6. Continued-Partial regression coefficients for impact of natural
areas of the Southeast, 1950-1970.t
resource investments on income change by
Area Group and
Type of Income Corps TVA FHA ASCS FPW PL-566 Constant R2
Change
Group 2 (Urban-rural)
Per capital
income -.00
Total personal
income
Total earnings
-1.50
-1.32
.06 15.08
.07 13.10
Total farm
earnings .18
21.27
18.30
-.04
-.00
-19.28
-15.48
-3.48
1,738.00 .03
-4.57 107,356.00 .03
-2.10 90,077.00 .03
-3.72
2,821.00 .02
Private
manufacturing
earnings
Private
construction
earnings
-.87**
9.88*** 13.46*
-.51
6.75
2.84
-4.92 20,080.00 .10
-5.71" 7,683.00 .05
continued
Table 6. Continued-Partial regression coefficients for impact of natural resource investments on income change by
areas of the Southeast, 1950-1970.t
Area Group and
Type of Income Corps TVA FHA ASCS FPW PL-566 Constant R2
Change
Group 3 (Densely settled rural)
Per capital
income .00 .003*"* .03 .04 -.10* .11 1,583.00 .09
Total personal
income -.68 .81 31.00 59.49"* -52.68"" -57.30 -1,893.00 .08
Total earnings -.57 .69 29.07" 49.69** -44.86"* -48.57 -3,202.00 .08
Total farm
earnings .30 .04 -2.11 -.91 1.85 19.23"* -3,336.00 .05
Private
manufacturing
earnings .16 .14 17.11*** 8.46 -12.08"* 10.66 -3,211.00 .13
Private
construction
earnings -.00 .03 2.30* 2.85 -3.79 -10.22 423.00 .09
continued
Table 6. Continued-Partial regression coefficients for
areas of the Southeast, 1950-1970.t
impact of natural resource investments on income change by
Area Group and
Type of Income Corps TVA FHA ASCS FPW PL-566 Constant R2
Change
Group 4 (Sparsely settled rural)
Per capital
income .00 .00 -.11 .11* -.134 .12 1,538.00 .03
Total personal
income 1.30"** -.01 -2.11 10.18 -5.03 12.80 30,922.00 .04
Total earnings 1.04** -.01 -1.44 8.71 -4.88 11.01 25,600.00 .04
Total farm
earnings .09 -.04 -.68 .90 .67 4.05 -2,907.00 .01
Private
manufacturing
earnings .02 -.02 .01 1.74 -4.51 4.96 9,540.00 .01
Private
construction
earnings .13"** -.01 -.19 .13 -.68 .10 2,755.00 .07
tThe regression coefficients are interpreted as the dollar change in income over the time period per dollar of investment in the
natural resource project.
*Significant at 10% level.
"Significant at 5% level.
-* Significant at 1% level.
Table 7. Partial regression coefficients
Southeast, 1960-1970.t
for impact of natural resource investments on income change by areas of the
Area Group and
Type of Income Corps TVA FHA ASCS PL-566 Constant R2
Change
Group 1 (Urban)
Per capital
income
Total personal
^ income
Total earnings
Total farm
earnings
Private
manufacturing
earnings
Private
construction
earnings
.00 -.00
211.03***
160.47*"*
1.54
-28.84
-21.59
-.30
-.08*
1,654.00 .06
-920.09 -1,724.78
-917.97** -1,009.28
-2.77
27.11"** -3.16 -146.20
14.08"** 5.62
-61.28
15.31
-114.24
-108.94
-2,061.94
-1,745.94
10.17
-385.88
-206.67
5,954,984.00
4,531,418.00
31,830.00 .02
925,300.00 .11
391,448.00 .19
continued
Table 7. Continued-Partial regression
areas of the Southeast, 1960-1970.t
coefficients for impact of natural resource investments on income change by
Area Group and
Type of Income Corps TVA FHA ASCS PL-566 Constant R2
Change
Group 2 (Urban-rural)
Per capital
income
Total personal
income
Total earnings
Total farm
earnings
Private
manufacturing
earnings
Private
construction
earnings
-.00
-6.48
-6.43
.01
-3.34
.26
-.00
-2.49
-2.19
-.04
-.02
9.82
8.31
.09
-.74 16.27
.07 -.11
.16
620.57***
512.72"*
1.10
263.49***
23.33*
.00
74.48
72.79
1.70
-4.12
-1.09
1,289.00 .02
255,883.00
217,051.00
2,746.00 .00
37,669.00 .14
16,237.00 .03
continued
Table 7. Continued-Partial regression
areas of the Southeast, 1960-1970.t
coefficients for impact of natural resource investments on income change by
Area Group and
Type of Income Corps TVA FHA ASCS PL-566 Constant R2
Change
Group 3 (Densely settled rural)
Per capital
income
Total personal
income
Total earnings
Total farm
earnings
Private
manufacturing
earnings
Private
construction
earnings
.00 -.01
1.74 56.50*
1.60 47.15*
3.85
2.83
2.91
.31 22.16**
.27** -1.73
.05
137.95"**
109.65"*
14.96"**
37.28**
5.30
-.10
-85.05
-86.16
11.34
-32.16
1.73
1,270.00 .03
265,640.00 .08
210,079.00 .07
815.00 .07
59,832.00 .06
9,252.00 .08
continued
Table 7. Continued-Partial regression
areas of the Southeast, 1960-1970.t
coefficients for impact of natural resource investments on income change by
Area Group and
Type of Income Corps TVA FHA ASCS PL-566 Constant R2
Change
Group 4 (Sparsely settled rural)
Per capital
income .00 .00 -.01 .15 .09 1,154.00 .01
Total personal
income .44 .92 19.50 228.28*** 47.01 76,729.00 .08
Total earnings .32 .67 10.67 190.96**" 38.26 57,954.00 .06
Total farm
earnings .27 -.08 -4.90** 16.10*** -.98 983.00 .05
Private
manufacturing
earnings -.38 -.08 17.01 78.50*** 17.56 13,845.00 .06
Private
construction
earnings .44 .02 -3.48 15.32* 16.11 2,810.00 .05
tThe regression coefficients are interpreted
natural resource project.
*Significant at 10% level.
Significant at 5% level.
* *Significant at 1% level.
as the dollar change in income over the time period per dollar of investment in the
of .37, and in very few cases did these correlation coefficients
exceed an absolute value of .25. None of the diagonal elements
of the inverted correlation matrix revealed any combinations of
highly correlated independent variables.
Investments by the Corps of Engineers should result in in-
creased incomes due to the large size of many projects that are
undertaken. Positive and statistically significant effects occurred
for changes in total personal income, total earnings, and private
construction earnings in the most urban counties (group 1 areas)
and the most rural counties (group 4 areas) during the 1950-
1970 period. These income effects result from both the initial
construction phase and the construction and economic activity
resulting from project utilization. The impacts were much larger
for the urban counties than for the more rural counties as indi-
cated by the relative sizes of the partial regression coefficients
for Corps investments in Table 6. Similar results occurred for
Corps investments in group 1 urban areas during the 1960-1970
period. Also, during both periods Corps investments were signi-
ficantly and positively related to changes in private manufactur-
ing earnings in group 1. urban areas. Changes in each of the
income measures per dollar of Corps investments were from four
to six times as large during the 1960-1970 period than during
the 1950-1970 period in group 1 urban areas. Group 4 rural areas
did not exhibit significant impacts for Corps investments on in-
come changes during the 1960-1970 period.
Corps investments were not very important in affecting in-
come changes in group 2 and group 3 areas for either time
period. The relationships for group 2 areas were generally nega-
tive for both time periods with the only significant positive im-
pact occurring for private construction earnings during the 1950-
1970 period.
The only statistically significant relationships found for TVA
investments were in group 3 counties for both time periods.
Many TVA investments occurred in group 3. A significant regres-
sion estimate resulted for per capital income increases during
1950-1970 and for increases in private construction earnings dur-
ing 1960-1970. In the case of group 1 and group 4 areas, the
regression coefficients for TVA investments generally had nega-
tive signs. These general negative relationships also occurred for
group 2 areas during the 1960-1970 period. However, based on
a 10 percent level of confidence none of these negative coeffici-
ents can be interpreted as being significantly different from a
value of zero.
Investments by the FHA for community water, sewer, and
recreational facilities are generally made in rural areas. Invest-
ments in these facilities are generally made in rural areas. In-
vestments in these facilities can often have the effect of influ-
encing firms to locate in a particular community because of the
advantages these systems afford to the firm and its employees.
FHA investments appeared to have the greatest impact on the
densely settled rural areas (group 3 areas). Positive relation-
ships were generally obtained for both group 2 and group 3 areas
during both time periods. The regression coefficients were sta-
tistically significant for changes in private manufacturing earn-
ings for group 2 and group 3 areas during 1950-1970, and group
3 areas during 1960-1970. Also during this later 1960-1970 period,
changes in total personal income and total earnings were impor-
tantly affected by FHA investments in group 3 areas. During the
longer 1950-1970 period both total earnings and private construc-
tion earnings increased as a result of FHA investments in group
3 areas. Negative regression coefficients were generally obtained
for FHA investments in group 1 urban areas during both time
periods. However, only in the case of per capital income changes
during both periods and changes in total earnings during 1960-
1970 were any of these estimated regression coefficients signifi-
cantly different from zero (at the 10 percent level of confidence).
In group 4 rural areas a significant negative regression coeffici-
ent was obtained for changes in total farm earnings during 1960-
1970, perhaps reflecting a movement of farm operators and
laborers to towns and other employment opportunities as com-
munity service facilities were improved.
Program investments in the ACP of the ASCS are generally
oriented to rural counties. This program provided the most uni-
form coverage over the nine-state region of any of the invest-
ment programs considered in this analysis. Positive and signifi-
cant regression coefficients were obtained for many of the mea-
sures of income change for group 2, 3, and 4 areas during both
periods of analysis. Group 1 urban areas reflected negative re-
gression coefficients for ASCS investments on all measures of
income change except total farm earnings during the 1960-1970
period. However, none of the estimated regression coefficients
were significantly different from zero (at the 10 percent level
of confidence) for either time period in this group. Increases in
total personal income, total earnings, and private manufacturing
earnings resulted from ASCS investments in group 2, 3, and 4
areas during 1960-1970. In group 3 and 4 areas these investments
also importantly impacted total farm earnings and in group 4
areas, private construction earnings during 1960-1970. ACP in-
vestments resulted in increased private manufacturing earnings
in group 2 areas, increased total personal income and total earn-
ings in group 3 areas, and increased per capital incomes in group
4 areas during the 1950-1970 period.
Flood Prevention Watershed investments generally exhibited
negative regression coefficients of income changes in each of
the four groups of areas during 1950-1970. The only significant
coefficients reflected decreases in per capital income, total per-
sonal income, total earnings and private manufacturing earnings
from these investments in group 3 areas. Since these measures
of income change were generally reflecting declines in the recipi-
ent counties during this period, investment in FPW probably had
the important impact of slowing down these decreases in income
levels.
PL-566 Small Watershed investments did not appear to be
very important in explaining income changes among the recipi-
ent areas. A significant and positive impact on total farm earn-
ings was obtained for group 3 areas during the 1950-1970 period.
Negative regression estimates were obtained for many of the
measures of income change in the other three groups of areas
for both time periods. However, except for private construction
earnings in group 2 areas during 1950-1970, none of the esti-
mated regression coefficients were significantly different from
zero (at the 10 percent level of confidence).
Many of the parameter estimates from the analysis were not
statistically significant at the 10 percent level of confidence.
However, an examination of the standard errors of the estimates
revealed that the regression coefficients were generally consider-
ably larger than the standard error, particularly for those in-
vestments that had a positive impact on income changes. Thus,
if the criterion for judging the importance of the various natural
resource investments on income changes is a ratio of the regres-
sion coefficient to the standard error greater than one, consider-
able evidence is available that natural resource investments
increase the income of recipient areas.
SUMMARY AND CONCLUSIONS
Summary
The relationships between natural resource investments and
changes in income of recipient areas were examined in this
study. Counties of the nine contiguous southeastern states were
delineated into four groups of relatively homogeneous areas
based on a number of natural resource endowment, economic,
social and demographic characteristics. Then county variations
in income change for each group of areas were related to ex-
penditures for natural resource projects. Six separate measures
of income changes were included that differentiated among the
potentially affected industries of agriculture, manufacturing,
and construction, as well as changes in per capital income, total
personal income, and total earnings. The resource investments
included new civil works construction expenditures by the Corps
of Engineers, TVA investments, FHA expenditures for commu-
nity water and sewer systems and recreational facilities, ACP
payments by the ASCS, Flood Prevention Watershed construc-
tion expenditures by the SCS, and PL-566 Small Watershed con-
struction expenditures by the SCS.
Emphasis was placed on identifying differences in program
impacts among the four types of areas in the Southeast. Results
indicated that investments by the Corps of Engineers for flood
control and navigation projects had the greatest impact on in-
comes in the most urban-oriented areas when compared to the
other three types of areas. Some evidence was provided that
Corps investments importantly increased incomes of recipient
counties in group 4 areas, the most sparsely settled rural areas
during the 1950-1970 period.
Investments by the TVA for multipurpose navigation, flood
control and steam generating projects were made in a small num-
ber of counties and were importantly associated with per capital
income increases in recipient counties of group 2, 3, and 4 areas.
The strongest impact in the most urban oriented counties (group
1 areas) was in private construction earnings.
Loans and grants by the Farmers Home Administration for
community water, sanitary sewer, solid waste disposal systems
and recreational facilities had the greatest income increasing im-
pact in group 3 areas (the densely settled rural areas) as com-
pared with the other three groups of areas. Some evidence
existed to indicate that FHA investments importantly increased
incomes in group 2 areas (the areas generally clustered around
the more urban oriented counties of the Southeast). Income
changes in group 1 urban areas were generally negatively asso-
ciated with FHA investments. However, few investments were
made in this group of counties during the 1950-1970 period.
Thus, the relationships obtained for group 1 areas must be
viewed with considerable caution.
Investments in the ACP program by the ASCS for conserva-
tion measures importantly impacted incomes in group 2, 3 and 4
areas during the 1960-1970 period. The greatest effect on income
increases were in group 2 areas (the rural areas clustered around
the more urban oriented counties) and group 4 areas (the sparse-
ly settled rural counties).
The two Flood Prevention Watershed investments in northern
Mississippi and northern Georgia were most important in influ-
encing incomes in group 3 areas (densely settled rural counties).
The relationships were generally negative indicating that in-
comes in these areas were decreasing even though investments in
flood prevention structures, reforestation and other soil conser-
vation measures were being made. The most important impacts
of these investments were to slow the rate of decline in income
of the recipient areas. That is, in the absence of the investment
programs income declines might have been greater than those
actually experienced.
Investments in the PL-566 Small Watershed Program by the
SCS were not found to be importantly affecting income changes
in the recipient areas. The impact on farm earnings was generally
to increase this measure of income change. In the case of group
3 areas (the densely settled rural counties) PL-566 Small Water-
shed investments importantly increased total farm earnings in
the recipient counties.
During both the 1950-1970 and 1960-1970 periods, low simple
correlation coefficients (correlation coefficient of .382 or less)
between natural resource investments and incomes of recipient
areas were obtained. However, many of these correlation coeffi-
cients were statistically significant at the 10 percent level of
confidence. The regression analysis revealed that many of the
natural resource investments were importantly affecting income
changes in the recipient areas as judged by the significance of
the regression coefficients at a 10 percent level of confidence.
However, very little of the total variation in income among coun-
ties for each type of area was accounted for by the nat-
ural resource investment variables (R2's of .19 or less).
Conclusions
Natural resource investments were found to be importantly
related to many of the measures of income change in the study
region. More importantly, considerable evidence was obtained to
indicate that natural resource investments have differential im-
pacts on local economies depending upon the natural resource
base and economic structure of the recipient area. The income
impacts per dollar of investment in natural resources were dif-
ferent among the four groups of areas. In this regard, natural
resource investments do have an important effect on the geo-
graphical distribution of income among regions of the Southeast.
Investments in natural resources accounted for a relatively
small amount of the total variation in income changes among
counties in the Southeast. Other important causal variables
which were not included in this particular analysis require exam-
ination for their effects on income changes.7 Assessment of the
rate of natural resource investments in stimulating income
growth in recipient areas will remain a difficult task without
knowledge of the relationships between natural resource invest-
ments and other important stimuli on changes in income of the
areas. The results of this study indicate that natural resource
investments should not be solely justified on the basis of im-
provements in regional income.
Some of the income measures were decreasing over the speci-
fied time periods in many of the counties comprising the four
groups, particularly in the group 3 and group 4 rural areas. Yet,
during these time spans, investments were being made in natural
resource projects in these counties. It is quite probable that
without these investments the income declines would have been
much larger than actually recorded. In this manner, natural re-
source investments may have had the effect of reducing the rate
of decline in income of some counties, and in some cases to even
reverse the direction of change.
From a long run viewpoint, some of the resource investments
may be acting to maintain the current productivity of natural
resources that were substantially developed in an earlier time
period. Incomes of landowners and secondary recipients in the
Mississippi Delta, for example, have probably been enhanced by
resource investments over long periods of time. The time lags in
7Some of these variables include other types of resource investments,
changes in other factor supplies, changes in product demands, changes in
firm technology and changes in the supplies of firm operators. Studies by
Cato (1), and Cato and Eddleman (2) consider some of these factors as
well as the impact of natural resource investments on employment changes
in the Southeast.
response by product users are not captured adequately by the
time-dated cross-sectional analysis carried out in this study. This
represents the most important limitation on the study. Research
on particular types of natural resource investments in case study
areas using time-series is required before further insights into
these relationships can be obtained."
"These types of analysis were carried out by project investigators partici-
pating in the Southern Regional Project S-71-"Income and Employment
Effects of Public Investments in Natural Resources." An evaluation of the
results has been prepared by Jansma (3).
APPENDIX
Appendix Table 1.-Correlation coefficients between total water invest-
ments and income changes by areas of the Southeast, 1950-1970 and
1960-1970.
Type of Income Change Group 1 Group 2 Group 3 Group 4
1950-1970
Per Capita income .189" .056 .247*** -.001
Total personal income .153 -.005 .081 .094
Total earnings .174 -.005 .080 .089
Total farm earnings .043 -.088 .004 -.049
Private manufacturing earnings .177 .008 .121 -.084
Private construction earnings .379**" -.014 .271** .053
1960-1970
Per Capita income .053 -.028 .147 .029
Total personal income .152 -.007 .110 -.030
Total earnings .159 -.011 .118 -.034
Total farm earnings .046 -.001 -.013 .016
Private manufacturing earnings .150 .015 .081 -.079
Private construction earnings .416*** .124 .270** -.024
"Significant at 10% level.
"Significant at 5% level.
"*Significant at 1% level.
LITERATURE CITED
(1) Cato, James. "The Effect of Resource Investment Programs on Labor
Employment." Unpublished Ph.D. dissertation, University of Florida,
Gainesville, December 1973.
(2) Cato, James and B. R. Eddleman. The Effect of Resource Investment
Programs on Labor Employment in the Southeast. Florida Agricul-
tural Experiment Station Bulletin 785. Gainesville: 1976.
(3) Jansma, J. Dean. "Measuring the Impact of Rural Development Acti-
vities on Employment, Income and Economic Structure: With Emphasis
on Natural Resource Investment". Staff Paper 10, Food and Resource
Economics Department, University of Florida, Gainesville: 1975.
(4) Martin, James E. "Use of Discriminant Analysis and Factor Analysis
In Delineating Regions and Relating Income and Employment to
Natural Resource Investments." Paper presented at the Methodology
Workshop of the Southern Land Economic Research Committee, Uni-
versity of Florida, Gainesville, March 20-21, 1969, (Mimeo).
(5) Tintner, G. Econometrics, 1st Science Edition. New York: John Wiley
and Sons, Inc., 1965.
APPENDIX
Appendix Table 1.-Correlation coefficients between total water invest-
ments and income changes by areas of the Southeast, 1950-1970 and
1960-1970.
Type of Income Change Group 1 Group 2 Group 3 Group 4
1950-1970
Per Capita income .189" .056 .247*** -.001
Total personal income .153 -.005 .081 .094
Total earnings .174 -.005 .080 .089
Total farm earnings .043 -.088 .004 -.049
Private manufacturing earnings .177 .008 .121 -.084
Private construction earnings .379**" -.014 .271** .053
1960-1970
Per Capita income .053 -.028 .147 .029
Total personal income .152 -.007 .110 -.030
Total earnings .159 -.011 .118 -.034
Total farm earnings .046 -.001 -.013 .016
Private manufacturing earnings .150 .015 .081 -.079
Private construction earnings .416*** .124 .270** -.024
"Significant at 10% level.
"Significant at 5% level.
"*Significant at 1% level.
LITERATURE CITED
(1) Cato, James. "The Effect of Resource Investment Programs on Labor
Employment." Unpublished Ph.D. dissertation, University of Florida,
Gainesville, December 1973.
(2) Cato, James and B. R. Eddleman. The Effect of Resource Investment
Programs on Labor Employment in the Southeast. Florida Agricul-
tural Experiment Station Bulletin 785. Gainesville: 1976.
(3) Jansma, J. Dean. "Measuring the Impact of Rural Development Acti-
vities on Employment, Income and Economic Structure: With Emphasis
on Natural Resource Investment". Staff Paper 10, Food and Resource
Economics Department, University of Florida, Gainesville: 1975.
(4) Martin, James E. "Use of Discriminant Analysis and Factor Analysis
In Delineating Regions and Relating Income and Employment to
Natural Resource Investments." Paper presented at the Methodology
Workshop of the Southern Land Economic Research Committee, Uni-
versity of Florida, Gainesville, March 20-21, 1969, (Mimeo).
(5) Tintner, G. Econometrics, 1st Science Edition. New York: John Wiley
and Sons, Inc., 1965.
This public document was promulgated at an annual cost of
$2,333.45 or a cost of 770 per copy to provide information on the
effects of investments in natural resources on regional income,
employment, and composition of economic activity in the nine-
state area of the Southeast.
I University of Florida1
|