Citation
The Effect of resource investment programs on labor employment

Material Information

Title:
The Effect of resource investment programs on labor employment
Creator:
Cato, James Carey ( Dissertant )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1973
Language:
English
Physical Description:
180 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Agriculture ( jstor )
Counties ( jstor )
Employment ( jstor )
Farms ( jstor )
Financial investments ( jstor )
Manufacturing industries ( jstor )
Mathematical variables ( jstor )
Prices ( jstor )
Supply ( jstor )
Wages ( jstor )
Dissertations, Academic -- Food and Resource Economics -- UF
Food and Resource Economics thesis Ph. D
Watersheds -- Economic aspects ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
This study examined the importance of investments in human and natural resources along with several other variables in explaining employment changes among counties comprising the four-state region of Mississippi, Alabama, Georgia, and Florida over the time period 1960 to 1970. Counties in the study area were delineated into urban and non-urban counties according to their human and natural resource endowments. A theoretical economic model was developed that explains changes in employment and firm numbers brought about by exogenous shifts in the supplies of resources, demand for products, supplies of other factors, firm production possibilities, and shifters of the number of firms in an industry. Empirical analysis was undertaken to determine the importance of each exogenous shifter on industry employment changes. The industries studied included agriculture, construction, textile mill products and other fabricated textile products, food and kindred products, transportation, furniture and fixtures, lumber and wood products, electrical equipment, durable, and non-durable products manufacturing. Changes In factor supplies included changes in per pupil education expenditures. Corps of Engineers investments, Soil Conservation Service investments in the Public Law 566 Small Watershed Program, Agricultural Stabilization and Conservation Service payments In the Agricultural Conservation Program and loans and grants for community water and sewer systems made by the Farmers Home Administration. Changes in county product price indexes, allotment reductions, farm operator age, wage rates, technology indexes, number of firms, and alternative wage and employment opportunities represented the other types of exogenous shifters. General equations for empirical analysis were specified for each industry. A two-equation model was used for agriculture with the number of farms considered endogenous. Single equation models were specified for the remaining Industries. Changes in the number of firms were considered exogenous In these models. Two-stage and ordinary least squares were used as empirical estimation procedures. Results indicate that employment changes emanating from changes in the exogenous shifters differed quite substantially among industries and the county groups considered. Most employment effects were generally consistent with expectations. Logical explanations were normally apparent for employment effects that differed from initial expectations. It was observed that some Investments Influenced employment in a particular industry and yet were not Important In other Industries. Location of industries was also important. Effects differed between the urban and non-urban counties for some Industries. The most satisfactory results were obtained for agricultural employment and farm number changes Results indicated that any attempt to stimulate employment in an area with investments in human and natural resource should take into consideration not only the agricultural, urban, and non-urban characteristics of the area, but the type of industry employment most evident in the area.
Thesis:
Thesis (Ph. D.)--University of Florida, 1973.
Bibliography:
Includes bibliographical references (leaves 175-179).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by James Carey Cato.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
030063903 ( AlephBibNum )
37876239 ( OCLC )
ACG8110 ( NOTIS )

Downloads

This item has the following downloads:


Full Text

RF

THE EFFECT OF RE50URCF INVEF>TMENT PROGRAMS
ON LABOR EMPLOYMENT





dft By
JAMES CAREY CATO









tON PRESENTED T IHE GR,DUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
1 FULFILLMENT OF THE REQUIRFjIENTS FOR TIE
DEGREE OF DOCTOR OF PHILOSOPHY






SITY OF











ACKNOWLEDGEMENTS


The author wishes to express appreciation to the Food and

Resource Economics Department, University of Florida, and the Economic

Research Service and Soil Conservation Service, United States Depart-

ment of Agriculture, for making this research possible. Sincere

appreciation is acknowledged to Dr. B. R. Eddleman, Chairman of the

Supervisory Committee, for his guidance and assistance during the

author's graduate program. Gratitude is also expressed to Drs.

J. E. Reynolds, K. C. Gibbs, and R. W. Bradbury who served as committee

members. Dr. W. W. McPherson also provided valuable comments.

Mr. Gene Harris provided valuable programming assistance.

Dr. Neil Cook, Economic Research Service, United States Department of

Agriculture provided excellent comments and guidance.

The author also wishes to express thanks to Mrs. Christine Ward

for her valuable assistance in typing the dissertation and to Mrs.

Phyllis Childress for her clerical and typing assistance in earlier

drafts. Appreciation is also due Miss Wanda Rhea who typed a complete

draft of the dissertation.

The greatest debt is due the author's wife, Diane, and sons

Kyle and Chad, for their unselfish devotion and sacrifice during his

graduate program.













TABLE OF CONTENTS




Page






LISTRO A BLES . . . . . . . . . . . . . vi




LIST OF ObGU EStiv.s. . . . . . . . . . . . ix
ABSTRACTe o. .L. .te.r. ture. . . . . . . . . . x

CHAPTER II

ITHEODUCTICAL .R . .RK. . . . . . . . . . 12





Mathematical Models of Equilibration Process . . . 12
Changes in Employment . . . . . . . 15
Economic Interpretation of Employment Effects . 16
Changes in Firm Numbers . . . . . . . 19
Economic Interpretation of Firm Number Effects . 21
Economic Interpretation of Exogenous Changes . . . 24
Factor Supplies . . . . . . . . . 25
Number of Firm Entrepreneurs . . . . . 32
Product Price . . . . . . . . . 32
Factor Price . . . . . . . . . 33
Firm Production Possibilities . . . . . 34

CHAPTER ill

STUDY AREA AND MODEL SPECIFICATION . . . . . . 3

Stildy Area . . . . . . . . . . . 35
selection OF Employml-ent Categorle! . . . . . 37
Cenera I MoUdel Spec.i f c ion . . . . 42
S trurctu ra I E S Lma tion . . . . . . . 45
REdvced Form Estim~atlon . . . . 45 .. M







Construction . . . . . . . . .
Manufacturing . . . . . . . .
Model Estimation Procedure. . . . . . .
Measurement of Variables and Empirical Expectations .
Employment . . . . . . . . . .
Firm Numbers . . . . . . . . .
Factor Supplies . . . . . . . .
Education investments (X1). . . . .
Corps of Engineers'natural resource
investments (X2)' . . . .
Soil Conservation Service PL-566
investments (X ). . . . . .
Agricultural Stabilization and Conservation


Service ACP investments (X4). .
Farmers Home Administration
investments (X ) . . . .
Crop allotment ( ) . . . .
Product Demand . . . . . . .
Agricultural product price (FP ).
Manufacturing product price (PPk)
Factor Price . . . . .
Agriculture wage rate (FP ) .
Manufacturing wage rate (FPk) . .
Technology . . . . . . . .
Agricultural technology (Z ).
Manufacturing technology () .
Farm Operator Supplies . . . . .
Agricultural wage opportunity (WW )
Agricultural employment opportunity
Farm operator age (WA) . . .
Manufacturing Labor Supplies . . .


* . i i
. . .
* . .
* . .
. *.
* .
* i *
. *.
. . .





(WE ) .
. . .


Manufacturing wage opportunity (WWV)..
Manufacturing employment opportunity (WE,)
Number of manufacturing firms (Mik) .


CHAPTER IV

ANALYSIS OF RESULTS . . . . . . . . .


Agriculture . . . . . . . . . .
Type of Equation . . . . . . .
Endogenous Variables . . . . . .
Exogenous Shifters . . . . . . .
Education investments . . . . .
Corps of Engineers' investments . .
Small Watershed Program investments .
Agricultural Conservation Program
investments . . . . . . .
Farmers Home Administration investimnts
Crop allotment . . . . . .
Agricultural product price . . .


I ( j





Page

Agricultural wage rate. . . . . .. 96
Agricultural technology . . . . .. 97
Agricultural wage opportunity . . ... 98
Agricultural employment opportunity ... . 99
Farm operator age . . . . . . . 100
Group differences . . . . . ... 100
Construction. .................. .. .. 100
Dependent Variable . . . . . . . . 102
Exogenous Shifters . . . . . . . . 102
Education investments . . . . . . 102
Corps of Engineers' investments. . . ... 103
Small Watershed Program investments ... 103
Agricultural Conservation Program
investments ............... 104
Farmers Home Administration investments . 104
Construction wage rate. . . . . .. 105
Construction wage opportunity . . . .. 105
Construction employment opportunity . . 106
Group differences . . . . . ... 06
Nondurable Manufacturing. . . . . . . ... 106
Dependent Variables. . . . . . . ... 107
Exogenous Shifters . . . . . . ... .11
Education investments . . . . ... .1
Corps of Engineers' investments . . ... 112
Small Watershed Program investments ... 113
Agricultural Conservation Program
investments . . . . . . . . 113
Farmers Home Administration investments .. 114
Manufacturing product price . . . . 114
Manufacturing wage rate . . . . .. 116
Manufacturing technology. . . . . ... 116
Manufacturing wage opportunity. . . ... 118
Manufacturing employment opportunity. ... 1!9
Number of firms . . . . . . . 1!9
Group differences . . . . . . . 120
Durable Manufacturing . . . . . . . . 121
Dependent Variables. . . . . . . ... 126
Exogenous Shifters . . . . . . . 126
Education investments . . . . . . 126
Corps of Engineers' investments . . ... 127
Small Watershed Program investments .... 127
Agricultural Conscrvation Program
investments . . . ... . . . . 128
Farmers Home Administratior. investrients . 128
Manufacturing product price . . . ... 129
Manrufscturing wage rate . . . . .. 131
Manufacturing technology. . . . ... !. 32
Manufacturing wage opportunity. . . . 134
Manufacturing employment opport'.it-. ... 13'
Number of fir rms ...... ............ 136
Group differences . . . . . ... 137




Page

CHAPTER V

SUMMARY AND CONCLUSIONS. . . . . . . . . ... 138

Summary . . . . . . . . ... . . .138
Conclusions . . . . . . . . ... . . 141
Limitations . . . . . . . ... . . 146
Need for Further Research . . . . . . . 147

APPENDIX A

SPECIFICATION OF AREA ADJUSTMENT MODEL . . . . . 149

APPENDIX B

MEANS AND STANDARD DEVIATIONS OF VARIABLES . . . ... 166

BIBLIOGRAPHY. . . . . . . . . . . . . ... 175

BIOGRAPHICAL SKETCH . . . . . . . . ... .... .180












LIST OF TABLES



Table Pagge

1 Industry identification and employment rankings for
the f ou r-s ta te a rea ( 1967) . . . . . . . 4o

2 Expected effect of changes in predetermined
variables on agricultural employment and


3 Expected effect of changes in predetermined
variables on manufacturing employment . . . . 55

14 Structural form and reduced form coefficients for
change in agricultural employment (E ) and number
of farm firms (N1I), all counties, 1910 to 1970 . . 77

5 Structural form and reduced form coefficients
for change in agricultural employment . . . . 79

6 Structural form and reduced form coefficients for
change in agricultural employment (El) and number of
farm firms (N 1), nonurban counties, 1560 to 19-70. .. 81

7 Regression equations for construction employment
change (E ) for all counties, urban counties,
and nonurarin counties, 1960 to )q70 . . . . . 101

8 Regression equations for textile mill product and
other fabricated textile products employment
change (E3) for all counties, urban counties, and
nonurban, counties, 11960 to 1S70 . . . . . 108

9 Regression equations for food and kindred products
emplo, ment change (Ej,) for all counties, urban
counties, and nonurban counties, 1960 to 1970 . . log

il0 Regression equations for nondurable manufacturing
em plIo yment change (END) for )IlI counties, urban
counties, and nonurban counties, 1960 to 1-570 110

11 Regression equatio,-ns for transportation products
,-mploymerit change ( 'E5) for all counties, urban
countie-, and nonurban couitic-sl 10,60 to, 970 . . 122


Vii





Table Page

12 Regression equations for furniture and lumber
and wood products employment change (E6) for
all counties, urban counties, nonurban counties,
1960 to 1970. . . . . . . . . ... ... 123

13 Regression equations for electrical equipment
products employment change (E7) for all
counties, urban counties, and nonurban
counties, 1960 to 1970. . . . . . . . . 124

14 Regression equations for durable manufacturing
employment (ED) for all counties, urban counties,
and nonurban counties, 1960 to 1970 . . . . .. 125

15 Definition and interpretation of terms
in equation (11). . . . . . . . . . 158

16 Definition and interpretation of terms
in equation (12). . . . . . . . . ... 162

17 Approximation of terms to represent a
change in residual return . . . . . . . 165

18 Means and standard deviations of employment change
variables for each county group of observations . . 167

19 Means and standard deviations of firm number change
variables for each county group of observations . .. 168

20 Means and standard deviations of factor supply
variables for each county group of observations . . 169

21 Means and standard deviations of product price
change variables for each county group of
observations. . . . ... . . ...... 170

22 Means and standard deviations of wage change
variables for each county group of observations . . 171

23 Means and standard deviations for technology change
variables for each county group of observations . .. 172

24 Means and standard deviations for wage opportunity
change variables for each county group of
observations . . . . . . . . . . 173

25 Means and standard deviations for employment
opportunity change variables for each county
group of observations . . . . . . . . 174


v ; ;












LIST OF FIGURES


Figure Page

I Illustration of the equilibration process th
resulting from a change in supply of the n
factor. . . . . . . . . ... .... 26

2 Illustration of the equilibration process in
the nth factor market resulting from a change
in supply of the nth factor . . . . . . . 27

3 Illustration of the equilibration process in
the firm entrepreneur market resulting from a
change in supply of the nth factor. . . . . ... 28

4 Illustration of the equilibration process in
the labor market resulting from a change in
supply of the nth factor. . . . . . . .. 29

5 Grouping of counties for the four-state study
area. . . . . . . . . ......... 38





Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy

THE EFFECTS OF RESOURCE INVESTMENT PROGRAMS
ON LABOR EMPLOYMENT

By

James Carey Cato

December, 1973


Chairman: Dr. B. R. Eddleman
Major Department: Food and Resource Economics

This study examined the importance of investments in human and

natural resources along with several other variables in explaining

employment changes among counties comprising the four-state region of

Mississippi, Alabama, Georgia, and Florida over the time period 1960

to 1970. Counties in the study area were delineated into urban and

nonurban counties according to their human and natural resource endow-

ments. A theoretical economic model was developed that explains

changes in employment and firm numbers brought about by exogenous

shifts in the supplies of resources, demand for products, supplies of

other factors, firm production possibilities, and shifters of the

number of firms in an industry.

Empirical analysis was undertaken to determine the importance

of each exogenous shifter on industry employment changes. The

industries studied included agriculture, construction, textile mill

products and other fabricated textile products, food and kindred

products, transportation, furniture and fixtures, lumber and wood

products, electrical equipment, durable, and nondurable products

manufacturing. Changes in factor supplies included changes in per pupil






education expenditures, Corps of Engineers investments, Soil Conservation

Service investments in the Public Law 566 small Watershed Program,

Agricultural Stabilization and Conservation Service payments in the

Agricultural Conservation Program and loans and grants for community

water and sewer systems made by the Farmers Home Administration. Changes

in county product price indexes, allotment reductions, farm operator

age, wage rates technology indexes, number of firms, and alternative

wage and employment opportunities represented the other types of

exogenous shifters.

General equations for empirical analysis were specified for

each industry. A two-equation model was used for agriculture with the

number of farms considered endogenous. Single equation models were

specified for the remaining industries. Changes in the number of firms

were considered exogenous in these models. Two-stage and ordinary

least squares were used as empirical estimation procedures.

Results indicate that employment changes eminating from changes

in the exogenous shifters differed quite substantially among industries

and the county groups considered. Most employment effects were generally

consistent with expectations. Logical explanations were normally

apparent for employment effects that differed from initial expectations.

It was observed that some investm,-ents influenced employment in a parti-

cular industry and yet wiere not important in other industries. Location

of industries was also important. Effects differed between the urban

and nonurban counties for 5ome industries. The most satisfactory

res-ults were obtained for agricultural employment end farm number




changxst





Results indicated that any attempt to stimulate employment in

an area with investments in human and natural resource should take into

consideration not only the agricultural, urban, and nonurban character-

istics of the area, but the type of industry employment most evident

in the area.












CHAPTER I

INTRODUCTION



Investments in natural resources usually are for the expressed

purposes of conserving, developing, or managing the nation's supply

of soil, water, timber, mineral, and marine resources. Many public

investment programs in natural resources such as those associated

with the Tennessee Valley Authority (TVA) and the Small Watershed

Development Program administered by the USDA's Soil Conservation

Service contained explicit development objectives. These objectives

were concerned with alleviating depressed regional economic conditions

or improving the incomes of specific groups of people.

Senate Document 97 [1], issued in 1962, also made explicit a

national policy of natural resource investments for purposes of

increasing income and employment in particular regions. The Appala-

chian Regional Development Act of 1965 [2] provided for the construction

of water resource projects to stimulate economic growth of the region.

Guidelines concerning principles and standards for the planning of

water and related land resource use, issued for review by the Water

Resources Council [3] in 1971, gave added emphasis to the role of water





This program, created by che Watershed Pr,:te.tion and Fiood
Prevention Act of 195L4, with its An.-ndnints, is zomnonly referred tc
as Pub ic Law 566.







resource investments in the development of a regional economy. This

orientation in policy has given added emphasis to natural resource

development programs and projects as instruments for dealing with

regional economic problems. Many other programs have evolved that focus

on goals of community improvement by concentrating on such areas as

increasing local employment and income, increasing public revenues, and

improving the quality of the environment.

Local employment and income of an area depend on many factors

other than investments in natural resources. Any explanation of employ-

ment and income changes occurring within a region requires analysis of

the many variables which interact to determine these changes. Identi-

fication and measurement of these complicated interdependencies are

necessary in order to assess previous or prospective effects of the

various programs in influencing the level of employment and income.

Changes in investment levels that shift the supplies of critical resources

often occur concurrently with changes in the demands for products,

supplies of other resources, firm production possibilities, and the

number of firms. An important element is the consideration of how

equilibration in product and factor markets is affected by programs

designed to change the supplies of resources and, in turn, how changes in

product and factor prices affect the level of output, resource employment

and income within the recipient region. Since similar investments in

heterogenous regions might have different effects on employment, dis-

similarities among regions need to be considered. These differences

could exist in the form of differino resource base or differing industrial

structure.

Planners and decision-makers need information on the effective-

ness of the small watershed and other natural reso r'ce projects in




3


fostering employment and income growth. Knowledge concerning the

relationships between natural resource investment and the other impor-

tant stimul i on changes in employment and income in a regional economy

is vital. Back [(41 has pointed out that assessment of the role of

natural resource investments in stimulating growth of a regional economy

will be a difficult task without information of this type.

Competition for the federal dollar between supporters of various

programs is very keen. Areas faced with employment and income problems

should utilize scarce resources in programs that would result in the

greatest rates of change in employment and income. Thus, information

on the effectiveness of natural resource investment programs in meeting

specified objectives is critical for future program planning. The

results of this evaluation will indicate that either small watershed

and other natural resource projects satisfy income and employment object-

ives or that program reform is necessary if objectives continue to

include income and employment goals. This research, chough conducted

for a sub-region of the Southeast, should provide answers applicable to

the entire Southeastern region. It is sufficiently broad to be indica-

tive of the national Small Watershed Program and otHer resource develop-

ment programs-


Objectives


The general objective of thi study is to evaluate the

effect iveness of ";ie Iel 4aterrshecl Program and other natural resource

investments in aceeaigepomn rwhin local recipient areas

The study include th-e Tour-szatc rec1Qn f Mississippi, Alabama, Geords.,








and Florida over the time period 1960 to 1970. More specifically, the

objectives are to:

1. Develop an economic model to explain changes in employment

and in the number of firms brought about by exogenous

shifts in the supplies of resources, demand for products,

supplies of other factors, firm production possibilities,

and shifters of the number of firms in an industry.

2. Empirically apply the model to selected industries in

order to determine the importance of changes in the above

factors on changes in employment and in the number of

farm firms.


Review of Literature


Previous studies concerned with the estimation and explanation

of the effects of investments in natural resources on employment,

income, and output are quite varied in purpose, objective and scope.

All previous work can be grouped into three basic categories consisting

of (1) case studies of individual projects and their impacts on local

areas, (2) studies proposing various procedures that could be used in

evaluating project effects, and (3) studies that attempt to determine

the effect of water resource investments over large multi-county or

multi-state areas. Some of these studies are reviewed in this section

to provide a cross section of previous work.

The first group consists of case studies of various small

watershed projects and the impact of the investment program on the

local economy and/or the sectors they were intended to benefit. Jarni.ia

and Back [53 estimated the local secondary effects of the construction







of watershed project structures for upstream flood protection in Roger

Mills County, Oklahoma. Through the use of an input-output analysis

they estimated income multipliers which were used to determine the

effect on the county's economy of increases in agricultural and

recreational income as a result of the watershed program. They found

that for each $100,000 in gross receipts to farmers in the county, there

was an estimated net (disposable) income to farmers of $26,867. This

$100,000 increase also generated $77,845 in gross receipts to other

sectors of the local economy and a net income of $16,457 to these

sectors. The local gross receipts multiplier of farm income was 1.78

and the local net income multiplier was 1.62. Net income receipts

from gross recreational receipts were about one-half those of gross

agricultural receipts. Gray and Trock [6] in an evaluation of the Green

Creek Watershed Project in Texas used traditional benefit-cost analysis

to compare the actual benefit-cost ratio derived from post-project

evaluation with the ratio estimated in the watershed work plan. They

found that the project failed to provide benefits in excess of costs

over the first eight years of the project. The ratio of benefits to

costs determined by the study was .919:1 as compared with an estimate of

1.49:1 made in the watershed work plan.

Input-output analysis was used by Kasal [7] to estimate the

local economic impacts in a four-county area as a result of five

Colorado watershed projects. Local net income arising from project

expenditures as well as benefits to the project users were estimated.

Kasal compared the original benefit-cost ratios for the projects with

those derived through the use O.: multipliers from the input-output








analysis and found that the local economic impact of all five projects
2
exceeded that estimated in the watershed work plan. Clonts is using

a similar analysis to measure the economic impact of the Cheaha Creek

Watershed on the local economy of Talledega County, Alabama. Cato

and Eddleman are evaluating the secondary impacts from the Taylor

Creek Watershed Project in a six-county area of South Florida. Work

in this study has been directed toward estimating the net increases

between the local impact area and the rest of the U. S., and among

groups within the local area. Several other studies not mentioned

were concerned mainly with benefits to primary project users, land use,

and increases in farmland values.

The second category of studies has been concerned primarily with

the introduction of suggested methodology for use in the evaluation of

water resource investments. The first of these methods as suggested by

Eidman [8] has had one empirical application. Eidman's presentation

described the linkages or interdependence of the various sectors and

subsectors of Southwestern Oklahoma with the use of a simple five-

equation economic model that employed economic base multipliers and

regression multipliers. This model was designed to explain employment

and income changes as a result of resource investments. Mazuera [9] used

the model to determine the secondary impact of using water in the Sugar

Creek Watershed floodwater prevention structures for irrigation develop-

ment and found that the project generated secondary effects amounting to




2
H. R. Clonts, personal communication.


"Work currently in progress.





7


less than a I percent change in both population and total income in the

watershed project area.

The work of Bromley et @1. [101 attempts to outline the role

of economic logic and method in analyzing the consequences of water

resource investments. Their work does not offer a concrete method

of project evaluation but rather is concerned with the many questions

raised in such an undertaking. A later study by Gibbs and Loehman [Hl]

deals with the evaluation of resource investment projects in terms of

multiple public goals. This study offers a model that might be used

to predict regional economic effects of resource investments which the

authors claim presents a more realistic view of the regional economy

than input-output or multiplier analysis.

The last category of studies has also been concerned with the

estimation of the impacts of water resource development at the local

level, but includes studies concerned with a much broader geographical

area than the local project area. Many of these studies used county

observations in various forms of econometric models to examine resource

investment effects oihile others examined regional differences in the

effects of resource investments. Haveman and Krutilla [121 looked at

both the national and regional effects of twelve types of water resource

projects with respect to their influence on various occupational and

industrial categories. Their analysis based on input-output models

constituted an effort to sort out the demands which public expenditures

for water resource developments imposed on the economy.

Water availability in relation to regional economic groWO) was

assessed by Howe L131 who determined that water deficit area~s did riot




8


were not guaranteed rapid growth. Howe's study suggested that water

resource developments are likely to be poor tools for accelerating

regional economic growth if markets, resources, and other factors

considered vital to development are lacking. Howe did not consider the

effect of water resource development on small regions.

Wiebe[14] attempted to evaluate the effectiveness of water

resource investment projects in alleviating regionally depressed

economic conditions. This study of the Tennessee River Watershed

consisting of 125 counties in parts of seven states suggested that (1)

residents in counties close to water resource investment projects enjoyed

a greater per capital income in the long run than did those living farther

away, (2) investments in water resources were in the long run associated

with increases in employment in counties removed from the project site

while nearer counties were associated with long run decreases in

employment, and (3) investments were not associated with an increase in

the level of living for people in low income and less educated groups

living near investment areas as compared to similar groups living in

areas removed from the investment site.

Another analysis by Cox et al. [15] was specifically designed to

assess broad based economic growth emanating from multipurpose projects

by the application of multiple regression analyses to many socioeconomic

indicators. Counties in the 13 Northeastern states of the United States

in which large .-:ater resource development projects had been constructed

between 1948 and 1958 were examined in 1960 to determine if changes had

occurred as a result of the projects. They concluded that there was no

relationship between project size and economic grc'-..th and that the

selection of project sites was biased toward urban areas where there is








a greater aprior likelihood of economic growth. They concluded that

it was dubious whether water resource projects served as a stimulus to

economic growth for the area studied. An index of economic growth

based on numerous income and employment variables were used to measure

economic growth.

Boxley and Harmon have recently worked on a study to determine

the relationship between Public Law 566 watershed investments and

economic growth in the Southeastern United States since 1959. Economic

growth was measured as income changes. They have attempted to use a

modified form of the shift-share analytical technique for the period

1959-1968. Tentative conclusions are that (1) there was no measurable

relationship between watershed investment and economic growth, and (2)

the selection of watershed sites for development appeared random when

measured in terms of rates and types of economic growth underway in the

study area. They have expressed limitations, however, as to the appro-

priateness of the statistical techniques used and thus offer these

conclusions in a qualified manner. They do point out that other

elements of capital infrastructure are probably necessary for economic

growth to occur.

Cato and Eddleman 5 are also concerned with the relation of

changes in income over th, last two decades to the level of investment

in natural resources for the same time period. Counties of a nine-state

region of the Southeastern United States have been delineated into

four groups on the bas i!s of the ir natural I and human resource endowments



4Robert F. Eoxley, and Marie Harmort, personal communication.







and level of economic activity. Multiple regression and correlation

analysis have been used to examine the effect of the level of natural

resource investment on changes in various income measures for the

four groups. The various types of water resource investments being

considered that are of interest to this study include water projects of

TVA, Corps of Engineers, and investments under the Small Watershed and

Flood Prevention Watershed Programs of the Soil Conservation Service.

Tentative results have indicated that with the exception of some invest-

ments by TVA and the Corps of Engineers, little effect is felt in local

income changes as a result of these water development projects.

Conflicting patterns have emerged as to the type of investment recipient

area that realizes the greatest impact.

This survey of the literature concerning the effect of water

resource investment on economic growth leads to two observations.

Either water resource investments are poor tools for stimulating economic

growth, or the methodology for measuring these effects falls far short

of accomplishing the goal for which it was designed. Many of the studies

cited have failed to consider the importance of interdependencies among

other important factors within a region which affect employment and

income changes. Changes in investment levels that shift the supplies of

critical factors, i.e., investments in water resources, often occur

concurrently with changes in the demands for products, supplies of other

resources, firm production possibilities, and shifters of the supply of

firms. Equilibration in product and factor markets is affected by

programs designed to change the supplies and/or productivity of resources

and this in turn causes changes in product and factor markets which affect

the level of employment and income within the recipient region.







A general approach that could be used in considering these
additional changes has been suggested by Tolley and Schrimper [ 16] and
Schrimper [ 17] This approach simultaneously considers aggregate and
micro adjustments in product and factor markets. Application of a
variant of the general model was performed by Schrimper [181 to deter-
mine the extent to which changes in various eognenous factors suggested

by the general model explained changes in the number of farms between
1954 and 1959 for six comparable groups of farms among states as well
as- among counties within North Crl in and. Ilio s
Eddleman and Cato [191 in a current study are using the same
variant of the general model as Schrimper to examine factors affecting
differential rates of change in the number of farms among counties in
Florida for the time period 1959-1969. Eddleman [20] has also proposed
the use of another variant of the general model to analyze the effects
ofivsmns nrsuc evlpetpogaso eina mlyet











CHAPTER II

THEORETICAL FRAMEWORK



All previous work using variants of the general model developed

by Tolley and Schrimper [16] has been concerned with measuring the rate

of change in employment or farm numbers. That is, employment and farm

number changes as well as changes in the exogenous variables were meas-

ured as percentage changes. The two-equation model developed in this

study is concerned with explaining absolute changes. The first equation

of the model explains changes in employment as a function of exogenous

changes in the prices of products, prices of factors having perfectly

elastic supplies, shifters of the supply of factors assumed to have other

than perfectly elastic supply functions for the region, shifters of firm

production possibilities, and changes in the number of firms. The second

equation of the model explains changes in the number of firms as a

function of these same exogenous variables and exogenous shifters of

firm supply functions in each industry.



Mathematical Models of Equilibration Process


Changes in the production of products and utilization of

resources in a region can be observed at several levels of aggregation

with each level providing a different insight about the impact of changes.

This dynamic process centers around three focal points. At the firm

level, product supplies and factor demands can be influenced by changes







in production possibilities or product and factor prices. The second

area concerns the aggregate effects of the micro adjustments reflected

in firms' product supplies and factor demands. Interaction of variation

in the number of firms with changes initiated at the micro or macro level

represents the third area.

Variation in product demands and factor supplies emanating from

the market level or from more disaggregated levels of specific types of

product demanders or resource suppliers can also be considered. For this

model, however, specification of the determinants of local employment and

firm number changes are under consideration and only variation at the ag-

gregate level for these two types of functions are considered as being

important to overall adjustments in a region's economy. Consequently,

product demand and factor supply functions, except for one critical fac-

tor (the nth factor), were assumed perfectly elastic.

The general regional model of adjustments consists of three basic

types of components. These are (1) product supplies and factor demands

for all firms in individual types of industries, (2) aggregate product

demand and factor supply functions, and (3) the number of firms in each

industry. Changes in any of the exogenous components can have substantial

effects on regional adjustments at all levels of aggregation because of

the relationships among the various functions. Primary causes of this

interaction are changes in the prices of products and factors which occur

at the aggregate level but still have an important impact on product out-

put and resource input decisions made by individual firms. Firm

adjustments emanating from changes at the aggregate level may result in


The complete od is specified in Appendix A.
The complete nodel is specified in Appendix A.







changes in the number of firms which lead to additional changes at the

aggregate level involving still further price adjustments. An equilib-

rium would exist when all product and factor prices and the number of

firms in the industry are consistent with their total demand and supplies.

For small areas or regions such as a county or group of counties,

the majority of product and factor prices might be realistically assumed

fixed since regional adjustments within the area would not likely have a

significant influence on the total market for these commodities. Feed-

back effects associated with adjustments would operate primarily within

factor markets since nearly all product prices, with the exception of

those products produced for local markets, would be mostly exogenous to

an individual region unless it accounted for a significant proportion of

the total national production of a commodity. Governmental price support

activities could also contribute to fixing some agricultural product

prices faced by farm firms in any given area. The probable existence of

fixed product prices in the analysis of regional adjustment or equilibra-

tion allows one to assume that the product demands and the supplies of

all but the critical factors are perfectly elastic for a region in which

firms in k different types of industries exist and produce m different

kinds of products with n-m different kinds of factors. Firms within

each of the k industry groups are assumed to have similar production

possibilities and be operated by basically the same type of entrepreneur.

The remaining part of this chapter is concerned with the discus-

sion of both changes in employment and firm numbers. First, an equation

explaining the effects of the various shifters on employment is discussed.

Then. a similar discussion of changes in firm numbers is outlined and fi-

nally, a discussion of both equations in a simultaneous context is given.








Changes in Employment

Using the assumptions discussed above, the change in demand for

labor (employment) in a g iven industry, k, cap in expressed as 2


t
(2.1) dQ E b S dX
Lk f 1 Lkn A f] f



+ d b III dp.
Lkj Lkn *j



+ d b d nki dp i
i M+l I Lki Lkn 1)

V
+ [d b d
h Lkh Lkn Zh


+ I q Lk b Lkn (71) nk d" ]k

where

b Lkn = change in quantity of lalww, demanded in the k th

industry associated with iD one-unit change ii

price of the n th factor.

A = difference in total quantity of the n th factor

demanded and supplied at z price one-unit abovc

the equilibrium price for the n th factor.

Sf = change in total quantity soopp] ied of the nth factor

associated with a one-unii change in the f th exog-

enous shifter for the n th actor supply,

2 The complete model from which equation (1.1) is derived and a
detailed treatment of the derivation and c-quMbsAtton process associated
with the exogenous shifters k presented in App**^x A. Table 15,
Appendix Agives a math-_matlcal def inition of t4-terms in the equatlon

am -







d = change in quantity of labor demanded in the kth
Lkj
industry associated with a one-unit change in
th
price of the j product.
th
d = change in quantity of the n factor demanded in
nkj
the kth industry associated with a one-unit change
.th
in price of the j product.

dLki = change in quantity of labor demanded in the kth

industry associated with a one-unit change in
.th
price of the i factor.
th
d k = difference in total quantity of the n factor
nki
supplied and demanded resulting from a one-unit

change in price of the i factor.

dLkh = change in quantity of labor demanded in the kth

industry associated with a one-unit change in the
th
h exogenous shifter of firm production

possibilities.

dnkh = change in quantity of the nth factor demanded in

the kth industry associated with a one-unit change

in the hth exogenous shifter of firm production

possibilities.

q = quantity of labor demanded by a firm in the kth
Lk
industry.

qnk = quantity of the nth factor demanded by a firm in

the kth industry.


Economic Interpretation of Employment Effects

Any discussion of the equilibration process on the demand for

labor associated with the exogenous shifters in the model must necessarily







be preceded by a discussion of the economic interpretation of the

various types of terms found in equation (2.1). Two terms appear in

each coefficient of the equation. The first of these is


b = oLk
Lkn k b-P

which represents the change in the total quantity of labor demanded in

the kth industry associated with a one-unit change in price of the nt
th
factor. Nk represents the total number of firms in the k industry,
bLk
while --- represents the change in quantity of labor demanded by a firm
Sn
in the kth industry associated with a one-unit change in price of the

n factor.

The second term is

r bq nk S
A = nk n
S1k ,P BP
k = 1 n n
th
which represents the amount by which the total demand for the n factor

would be less than the supply at a price one-unit above its equilibrium
r bqnk
level. The portion NW p represents the number of firms in the
k= k 6P
th n th
k industry times the change in the quantity of the n factor demanded

by a firm in the kth industry associated with a one-unit change in price
th n
of the n factor summed over all industries. The portion .-p- represents
th n
the total change in the quantity of the n factor supplied in the area

associated with a one-unit change in price of the th factor. The

absolute value of A (or -A) represents the increase in demand or decrease

in supply for this factor consistent with a one-unit change in price of

the factor. The negative reciprocal can then be interpreted as the ap-

proximate change in price for each unit increase in demand or decrease in

supply of the nt factor resulting from exogenous shifters in the model.







The magnitude of this reciprocal would ultimately depend on the

elasticities of the demand and supply functions involved since the elas-

ticities would determine the magnitude of the quantity change associated

with a given price change. Assuming constant elasticities within the

relevant range,(--) would represent the decrease in price for each unit

decrease in the demand or increase in the supply of this factor.

Three coefficients contain other terms that are very similar. The

terms dLkj d Lki, and dLkh represent the direct change in quantity of

labor demanded in the kth industry associated with a one-unit change in

the jh product price, i factor price, or h shifter of firm pro-

duction possibilities, respectively. Similarly, dnkj and dnkh represent

the change in quantity of the n factor demanded in the kt industry
th
associated with a one-unit change in price of the j product and in the

hth shifter of firm production possibilities, respectively. The term,
th
d ki represents the difference in total quantity of the n factor

supplied and demanded associated with a one-unit change in price of the
.th
i factor. The mathematical construction of these terms is similar to

that discussed for bLkn with the appropriate notational changes made as
L kn
shown in Table 15, Appendix A.

Interpretation of these terms becomes even more meaningful when

discussed simultaneously. For example,(- is the approximate change in
th
price of the nth factor associated with each unit change in total demand

or supply of the nth factor, and dnkj is the change in total quantity of

the nth factor demanded in the kth industry associated with a one-unit

change in price of the jth product. Their product, () dkj, represents

the total price change for the nth factor associated with a one-unit

change in price of the jth product. Final multiplication of this term by




19


b Lkn, which represents the change in quantity oF labor demanded in the k th

industry associated with a one-unit change in price of the n th factor,

gives b () d which may be interpreted as the total change in the
Lkn A nkj'
quantity of labor demanded in the k th industry resulting from a one-unit

change in the j th product price operating through the n th factor market,

This might be referred to as the indirect effect on labor demanded in the

k th industry resulting from an exogenous change in the j th product price.

The direct effect of this price term would be indicated by d Lk and would

complete the coefficient of dP.j. All other terms in each coefficient

would be interpreted in a similar manner.


Changes in Firm Numbers

Assuming only one critical factor, the change in the number of

firms in the k th industry can be expressed as3



(2.2) dN = S dX
k kn\/g f


+ +aD d]
d I kd kn B dd


+ 1akj + a kn t U dP

n-I
a .- a L, dP.
M Iki kn B

V
+ [Ck + a kn(i Mh dZh
h =


3The complete model from w,,hich equation (2.2) is dertved and a
detai -1ed treatment of the der iva ti on and equ ilIibr~im process assoc iated
w 'ith the exogenous shifters is presented in Appendix A. Table 16,
Appendix A,9i.,ies &. mathem-atica I def in it ion of the terms i n the equation.







where

akn = change in the number of firms in the k industry

associated with a change in residual returns

brought about by a one-unit change in the price
th
of the n factor.

B = excess demand for the nt factor associated with

a one-unit change in price of the nth factor.

Sf = change in total quantity supplied of the nt

factor associated with a one-unit change in the

f exogenous shifter of the n factor supplies.

Skd = change in the number of firms in the kth industry

associated with a one-unit change in the dth

exogenous shifter of firm entrepreneur supply.

Dd = change in total demand for the nth factor

associated with a one-unit change in the dth

exogenous shifter of firm entrepreneur supply.
th
akj = change in the number of firms in the k industry

associated with a change in residual returns

brought about by a one-unit change in price of

the jth product.
th
U. = change in demand for the n factor associated
J
th
with a one-unit change in price of the j product.

aki = change in the number of firms in the kth industry

associated with a change in residual returns

brought about by a one-unit change in price of

the ith factor.
the i factor.







th
L. = difference in total quantity of the n factor

that would be demanded and supplied resulting

from a one-unit change in price of the it

factor.

ckh = change in the number of firms in the kth industry

associated with a change in residual returns

brought about by a one-unit change in the hth

exogenous shifter of firm production possibilities.

Mh = change in demand for the nth factor associated

with a one-unit change in the hth exogenous

shifter of firm production possibilities.


Economic Interpretation of Firm Number Effects

Interpretation of the terms in equation (2.2) must also begin

with a discussion of those terms appearing in each coefficient and those

whose definitions are quite similar. Schrimper [18] provides a detailed

discussion of coefficients from a somewhat similar equation expressed in

terms of elasticities and percentage changes. Equation (2.2) represents

the sum of effects associated with shifters of factor supplies, shifters

of firm entrepreneur supply, product demands, factor supplies, and

shifters of firm production possibilities on the number of firms. Each

of these shifters operates through either a direct shift in the number of

firms or through changes in firm residual returns resulting indirectly

from these changes and from changes in use of the critical factor.

Two terms appear in each coefficient of the equation. The first

of these is

r kbN r ,b
a = -- qe N:' = -
ekn n ne e nP
e = 1 e e= n







as derived in Appendix A and shown in Table 16. This term represents the

change in the number of firms in the kh industry associated with change

in residual return brought about by a one-unit change in price of the

nth factor. The second term is defined as

B r 'q nk nn r o
B N k bP P E qnk kn
k = n n k = I

which represents the excess demand for the nth factor associated with a one-

unit change in price of the nth factor. The first portion of B is

identical to A as defined for equation (2.1). The second portion takes

into account some of the feedback effects on the quantity of the nt

factor demanded resulting from changes in the rnuber of firms that are

associated with a change in residual return brought about by a one-unit

price change for the nh factor. The reciprocal of this term, -) can

then be interpreted in the same manner as ( The negative reciprocal,

- ) can be interpreted as the approximate change in price of the n

factor for each unit decrease in total demand or increase in supply of

the nth factor after adjustments resulting from ,xegenous shifters in the

model.

The term Xkd represents the change in the number of firms asso-

ciated with a one-unit change in the dth shifter of firm entrepreneur

supply. Three additional terms that are very similar in nature are akj'

aki, and ckh. These terms are similar in mathematical construction and

interpretation to akn. They are interpreted as tOe change in the number

of firms associated with a change in residual return brought about by a

one-unit change in j product price, i factor price, and h shifter

of firm production possibilities, respectively. Each of these terms is

treated mathematically in Appendix A and Table 16.




23


The term D d may be interpreted as the change in demand for the n t+I

factor associated with a one-unit change in the d th shifter of firm entre-

preneur supply. Two terms that are similar to D d in that they represent

demand and supply relationship for Q n under alternative price changes are

U i and L Price changes in the j th product are considered in U The

r 0 tq nk
first part of U E N represents the change in demand for the
j, k = I k

n th factor associated with a one-un;t change in price of the j th product.

r
The second part, E q a a takes into account the feedback effects on
k = I nk kj' th
changes in the number of firms of a change in j product price as it is

felt through changes in the residual return. Both parts of this term

would have positive signs.

6S n r )qnk
The first part of L E No represents the
P i k k 6P

difference in supply and demand for the n th factor at its initial equi-

librium price after a one-unit increase in price of the i th factor,

assuming no change in firm numbers. The second part,

r 0 th
q nk a ki' takes into account the effects on demand for the n factor


resulting from changes i-n the number of firris brought about by a one-unit

change in P i as felt through a change in residual rturn. Both these

terms would tend to operate in the same -direction, The first part of L

differs from the f irst part of li i sinc-O factor suppies were assumed to

depend on possibly more than one factor price but b-- iadependent of
th ,
product prices. 'ExistIng relationship5 between the n factor supply

and changes in i th es, would requlate this movement to some

degree.







The last term, Mh, is also composed of two parts. The first

r bq
part, E N n, represents the change in quantity demanded of the
k = 1 h

nth factor associated with a one-unit change in the hth exogenous shifter

of firm production possibilities. The second part takes into account the
th
change in the number of firms as it affects the n factor demand. This

feedback effect is reflected through residual return changes resulting

from a one unit change in the hth shifter of firm production possibilities.

Changes in the hth shifter could be either output increasing and/or input

decreasing as explained and footnoted in Appendix A and Table 16.

Again, as in equation (2.1), the product of each term in the co-

efficients gives total meaning to each coefficient. For example, in

the coefficient of dXf, the product of (g) S would be interpreted as

the total price change for the nth factor associated with a one-unit

change in Xf, since Sf is the change in quantity of the nth factor sup-

plied as a result of a one-unit change in Xf and () is the n factor's

price change for each unit change in the quantity. Further multiplica-

tion by akn, which represents the change in the number of firms associated

with a change in residual return brought about by a one-unit change in

Pn, would then give the total change in the number of firms associated

with a one-unit change in Xf. The multiplicative effects of terms in

each of the other coefficients would be interpreted similarly.



Economic Interpretation of Exoqenous Changes


Changes in the demand for laborwvre expressed as a function of

changes in shifters of factor supplies, product prices, factor prices,

firm production possibilities, and the number of Firms. Changes in the








number of firms were then expressed as a function of the same shifters ot

factor supplies, product prices, factor prices, firm production possi-

bilities and shifters of firm entrepreneur supply. Each equation, (2.1)

and (2.2), was derived from the same basic model as shown in Appendix A.

Considered simultaneously, these equations then explain the changes oc-

curring in an area economy as the result of changes in the exogenous

variables. These processes are discussed separately in the following

sections.


Factor Supplies

Changes in factor supplies in each equation are represented by

dXf. All changes in factor supplies occur in the supply of the factor

assumed to be other than perfectly elastic. Equilibration resulting

from an exogenous shift in factor supplies is demonstrated graphically

in Figures 1 through 4. Figure 1 demonstrates the entire process while

Figures 2, 3, and 4 outline in more detailed form the actual changes oc-

curring in three of the five segments of Figure 1.

The effect of a change in supply of the nt factor is indicated

in Figures 1 (b) and 2 as the shift from S to S' with the resultant in-
n n
0
crease in Qn to Q' and decrease from P to P'. Shown in Figure 2 is the
n n n n
total change in quantity of the nt factor indicated by

bS 1 S
bn dXf and the total price change as indicated by i- n dX.
ff

Since (-) represents the change in P associated vith a one unit change
An
in Qn (- would be negative as indicated since Q increases and the de-
th
mand curve for the n factor is negatively sloped. This decrease in P
n
th
would then cause an increase in residual returns to firms in the k in-

dustry as reflecLed in the shift from Rk to R: in Figures 1(d) and 3.
k K








P.

Qj k




P.
J








P.





P.


eQe Q /unit
n n n time

Sk









SR
*s \^


o e unit 1 /unit
Q, Q Q Q. u N F l /
ik ik k ik time k k k k time
(c) (d)
PLk

Lk


QLk QLk QLkQLk


-Sk
SSLk



Lk
/unit
Lk time


Illustration of the equilibration process resulting
from a change in supply of the nth factor.


o e u n i t 0
jk jk j Lk jk time n


Figure 1.




27











S
n




nS









nn e


nn


q dX d
A nk kk
nb
D

D


Figure 2 Illustation ofthe equlibratin poesi h t




















He
k

k

TI k


B bX F
f


a k (- )


bS
-XF dXf
f


I1


a (k 1 n dXf

Figure 3. Illustration of the equilibration process in the firm
entrepreneur market resulting from a change in supply
of the nth factor.


S k




29














PL
L k'







PL S Lk








D) 0 D D"
Lk Lk D L k Lk


Q Q QQ Q ,un i t
L k Lk L k QLk / Lie

qi kd k m

q 6S
N Lk I n d



k jP \A nk
n

Figure 4, 1Mustration of the- equilibration process in the- Ihbor
mark-et resu It ing f rom a change- i n supply of the nt
factor.







The effect of this shift would be the movement from Nk to NI as
k k


SSbs
indicated by akn (- -- dXf and the movement from ik to Hii as in-

BS
dicated by (- B) --j dXf in Figure 3. Resulting from an increase in the
-, f

th
number of firms would be an increase in the supply of the j product by

the kth industry, Sjk to Sk with the new quantity measured as Qjk as

shown in Figure l(a). Accompanying this increase would be an increase

in demand for the i factor shown in Figure 1(c) as the shift from Dik

to Di with Qi indicating the new quantity demanded. The final initial
ik ik
result would be an increase in the demand for labor as shown by the shift

from Dk to D- in Figures l(e) and 4. This total shift comes from two
Lk Lk
sources. The first is a direct result of the shift in Xf as shown by

bq bS
S Lk 1- n dX and indicated by the movement from Qk to Qk in
k 7P A N F Lk Lk
n f

Figure 4. The remainder of the increase in labor demand comes from an

increase in the number of firms resulting from increases in residual re-

turns through lower prices of P This increase is given by qL dNk as

the movement from Q to Q- in Figure 4.
Lk Lk
Entrance of new firms to the kth industry would ultimately in-

crease the demand for the critical factor as indicated by the shift from

D to De in Figures 1(b) and 2. The resultant shift in Q' to Qe is in-
n n n n
dictated in Figure 2 by q0k dN. Resulting from this would be a bidding
nk k'
upward of Pn to Pe as shown by (- + q dNk in Figure 2. Since

represents the decrease in price for each unit decrease in demand, an
e
increase in demand, i.e., the movement from D to D would cause a
n n

Drice increase as shown. Increases in P would then result in the exit
n
of some firms from the kh industry through a decrease in the residual







return to these industries. This decrease in return is shown by the

shift from R to Re in Figures 1(d) and 3. The effect of a price in-
k e
crease through -) then causes the movement from nk to Hk as

indicated in Figure 3. A similar response in the movement from Nk to

Ne is also shown. Resultant shifts would also occur in Sk and D.
k k ik
e e
Equilibrium positions would be indicated by S.k and De with equilibrium
jk ik
quantities indicated by Q and Q in Figures 1(a) and k(c),
jk ik
respectively.

The final movement requiring discussion concerns that in the

labor market. Changes in P resulting from increases in demand for the
n
th
n factor also cause a final shift in the labor demand function as in-

dicated by the shift from DL to De in Figures 1(e) and 4. Again
Lk Lk
since -~ represents the decrease in price of the n factor associated

with decreases in demand, the price increase occurring as a result of

the increased demand from new firms as felt through (- .) would partially

affect the earlier increase in the quantity of labor demanded as in-

o LK 1o o
dicated through the term, Nk bP- ) q dNk, with equilibrium quantity
n

demanded at Qek as shown in Figure 4.

A similar analysis could be performed for each of the remaining

four exogenous shifters in the model. Each discussion would center

around the equilibration process as reflected through the appropriate

terms in equations (2.1) and (2.2). However, due to their similarities

and the complete treatment given dXf, only a brief discussion is given

for the remaining shifters in the following sections.







Number of Firm Entrepreneurs

The initial effect of an increase in an exogenous shifter of the

number of firm entrepreneurs would, of course, be an increase in the num-

ber of firms in a given industry. Consequently, the residual return to

each firm in the industry would decline as the result of both increases

in the number of Firms and increases in demand for the nt factor which

would increase its price. The quantity of the jth product produced would
.th
initially increase along with the quantities demanded of the i factor

and labor. Ultimately, price increases for the nt Factor would result

in declines in residual returns to firms and lead to fewer numbersof firms.

Quantities of the jth product produced would then decline, along with the
.th
quantity demanded of the i factor and labor. Equilibrium quantities

would be expected to remain above their initial values as the result of

the initial upward shift outweighing secondary changes due to the reduc-

tion in the number of firms. The actual quantities of factors utilized

would depend on their substitutability for the nth factor and the increase

in the number of firms. Even though the nt factor and labor might pos-
th
sibly substitute, as the price of the n factor increased, the overall

effect should be an increase in quantity demanded of all factors as

well as in the number of firms.

Product Price

Increases in product prices (upward shifts in perfectly elastic

product demands) would directly affect the residual return to firms and

cause firm numbers to increase. As firm numbers increase, the demand

for the nth factor would increase, and since its supply function is

other than perfectly elastic its price would increase. Simultaneous in-

creases would also occur in the quantities of labor and the i factor
demanded. Ultimately, as the price of the nth factor increased the
demanded. Ultimately, as the price of the n factor increased the








residual return to firms would decrease accompanied by a decrease in the

number of firms. This would cause a decrease in the quantities demanded

of labor, the i th factor, and the n th factor. Equilibrium conditions

would be expected to result in an increase in the quantity of labor de-

manded as well as in the number of firms from the initial levels befor

the product price change.


Factor Price

Changes in factor prices, representing a shift in factor supplie

assumed to be perfectly elastic, operate in an opposite manner than prod

uct price changes. As factor prices increase, the residual return to



firms. Initially the quantity of the jth product produced, quantities



,crease. These decreases would be offset somewhat by their substitut-

ability for the i th factor. As the price of the n th factor declines a

the result of smaller quantities demanded, existing firms would expei

ence increased residual returns and, thus, new firms would be enticed

to enter the industry thus increasing the demand for the n th factor, h

i th factor, and labor, as well as increasing the quantity of product

produced. Depending on the elasticity of substitution among factors n

the relative magnitudes of the various changes occurring throughout

equilibration process, equilibrium increases in the quantities of lao

and the n th factor demanded might occur simultaneously with the decrt se

in the quantities of product output, i th factor demanded, and the nume







Firm Production Possibilities

Changes in firm production possibilities which are output

increasing and/or input decreasing result in higher residual returns to

firms and thus increases in firm numbers. Entrance of new firms would

increase the output of product, quantity of both labor and the th

factor demanded and bid up the price of the nt factor through increases
th
in its demands. Increases in the n factor price would ultimately lead

to a decline in returns and cause an offsetting effect by reducing the

number of firms. This decrease would then lead to a decrease in output,
.th
and decreases in quantities demanded of the i factor and labor.

Equilibrium quantities, however, would be expected to be greater than

the initial quantities.












CHAPTER III

STUDY AREA AND MODEL SPECIFICATION



One of the major objectives of this research was to determine

the importance of natural resource investments and other types of invest-

ments in influencing employment changes in recipient areas. Particular

interest was given to investments in critical resources including water

and other forms of capital investment. For research such as this to be

applicable in other areas some level of confidence must be maintained

that knowledge acquired through research in a particular geographical

area is transferable to other areas. If there is some question as L'o

the applicability of research results among areas then the information

becomes quite limited in use. Since regions differ in geographic, in-

stitutional, sociological, and economic characteristics one cannot

conclude that natural resource investments stimulate employment in gen-

eral without some regard to these different regional characteristics.



StudyArea


The four-state region of Mississippi, Alabama, Georgia, and

Florida containing 3,75 counties was chosen as the study area. The area

was delineated into togroups of homogenous subareas. Counties in the

four states were classified into two groups on the basis of a set of ten

variables depicting the county's human and natural resource endow'ments,

and its urban, industrlil, and agricultural structure. These ten







variables were 1960 measures of (1) population, (2) urban population,

(3) percent of persons 25 years old and over with a high school education,

(4) median age, (5) total employment, (6) total agricultural employment,

(7) total manufacturing employment, (8) land area, (9) land in farms, and

(10) value of farm products sold.

Discriminant analysis was used in this delineation. Counties were

grouped initially into two groups. A linear function of the differences

of the means between the two groups was found which discriminated most

successfully between the two groups. A general mean was then derived by

substituting the means of all variables into the function. The function

was then used to derive a mean for each county using county observations

on each variable. Counties having values above the general mean were

placed in one group and those having means below the general mean were

placed in the second group. The process was repeated until the number of

misclassified counties was minimal. Computational procedures of the tech-

nique are outlined b/ Martin [21]. Additional discussion of the discrim-

inant analysis technique is also provided by Tintner [22, pp. 96-102].

Several other variables were used with the above ten in grouping the

counties, but their relative lack of importance in the discriminant func-

tion excluded them from consideration in the final delineation process.

The most important variables in the delineation process were median age,

education level, land area, agricultural employment, and total employment.

Characteristics of the two groups were as expected. For example, the ur-

ban counties had higher average educational levels and higher average

total employment. The nonurban group had higher average agricultural

employment levels.

Some judgement was warranted as to the meaningfulness of the de-

lineations by persons familiar with the four-state region. Members of







the Southern Land Economics Research Committee from each of the four

states were asked to evaluate the initial classifications for their

state. These evaluations were used to adjust the mathematical delinea-

tions and resulted in the final grouping shown in Figure 5. About

one-fourth of the counties (91) represent the more urban-oriented group

while the remaining counties (284) exhibit a more rural orientation.

Analyses were carried out for three groupings of counties. All

counties in the four-state region constituted one group. The urban

counties and nonurban counties formed the remaining two groups. Effects

of similar investments in human, natural, and capital resources in these

dissimilar areas should provide results that would be expected in other

areas. For example, information on the effects of investments in urban-

oriented versus the rural-oriented areas could provide guidelines on

what types of investments should stimulate employment in the same types

of areas in other regions of the country.



Selection of Employment Categories


Some categorization of employment is necessary to insure a

meaningful analysis. Employment is reported on the basis of either oc-

cupational or industrial classifications. Occupational categories

include such classifications as the various types of laborers, crafts-

men, and professional workers, while the industrial classification

distinguishes among the various industries such as agriculture, construc-

tion, textile products, and metals. The various types of investment in


This committee consists of representatives from eleven southern
land-grant universities, several federal agencies, and is supported in
part by the Farm Foundation, Chicago, Illinois.



















































a,













n


0
1-
%- %-
O



En n n


4rJ L










0 0 -


ct
U U UP
a)






a, a O
Ci-


0 0
U-











L.1




3L 3
U U-In







human, natural, and capital resources could affect employment in each

occupational or industry category in a different way. Industrial classi-

fication lends itself to a more meaningful analysis from the standpoint

of the theoretical framework outlined in Chapter II. Also secondary data

on industrial classification are more readily available. The industrial

classification was, therefore, used to distinguish among employment cate-

gories in the analysis.

Industry selection was made primarily according to the number of

employees. Agriculture, construction, five individual manufacturing in-

dustries, and durable and nondurable manufacturing were selected for

analysis (Table 1). These industries are likely candidates for growth-

generating activities with growth in the remaining industries being

dependent on them. For this reason, and since data of the nature needed

for this study are not readily available for the more service-oriented

industries, the service industries such as wholesale and retail trade,

finances, and communications were not included in the analysis. Also,

the initial employment effects of investments in natural resources would

most likely be felt in the industries selected.

Industries three through seven accounted for approximately 57

percent (762,098 employees) of all manufacturing employment in the four-

state region in 1967 [23, pp. 1.7-1.11, 10.7-10.11, 11.9-11.13, 25.6-25.9].

The importance of construction and agriculture is shown in Table 1 where

these industries rank one (460,771 employees) and three (286,528 em-

ployees), respectively, when compared to employment in the individual

manufacturing industries [24, Table 54]. Employment data for construction

and agriculture were for 1969.






Table 1. Industry identification and employment
four-state area (1967)


rankings for the


Standard Total
Industry industrial four-
identification Industry classifica- state
for this tion code employment
study (SIC) rankb

SAgriculture 15-17 3

2 Construction 01, 07-09

3 Textile mill products and
other fabricated textile
products manufacturing 22, 23 2

4 Food and kindred products
manufacturing 20 4

5 Transportation equipment
manufacturing 37 5

6 Furniture, lumber, and wood
products manufacturing 24, 25 6

7 Electrical equipment
manufacturing 36 7

D Durable products
manufacturing 24-26, 32-34,
36, 37

ND Nondurable products
manufacturing 20, 22, 23, 28

Several industries were grouped together to make data from the
U.S. Census of Population and the U.S. Census of Manufactures comparable
for later uses in the study.

Calculated from: U.S. Bureau of the Census, Census of
Manufactures, Area Statistics: 1967 (Washington, U.S. Government
Printing Office), Vol. III, Part I, pp. 1.7-1.11, 10.7-10.11, 11.9-11.13,
25.6-25.9 and Census of Population: 1970 (Washington, U.S. Government
Printing Office), Vol. I, Parts 2, 11, 12, and 26, Table 54.







Water usage of each industry was also examined since two of the

natural resource investment programs included in the study are water

oriented. Water use data for construction and agriculture are not avail-

able. Agriculture should benefit from drainage, flood control and an

increased irrigation water supply. Analysis of national rank in water
2
use for manufacturing industries indicated that some of the largest

water using industries have not been included in the analysis. The pe-

troleum and coal industry is the third largest industrial water use in

the U.S. However, this industry provided only .3 percent of the total

four-state employment and thus was not included. Manufacture of stone,

clay,and glass products, nonelectrical machinery, and rubber and plastics

products accounted for only 5.5 percent of total four-state employment.

Additionally, these industries ranked low in total U.S. water use. Al-

though the stone, clay, and glass industry provides products required in

building water control structures, the combination of low employment and

inadequate county data provided sufficient basis for excluding these in-

dustries. Inadequate county data also resulted in the exclusion of both

primary and fabricated metal products, paper.and allied products, and

chemicals and allied products, although chemicals and fabricated metals

manufacturing rank one and two nationally in water use.

The five individual manufacturing industries included in the

analysis ranged from fifth to twelfth among the rankings for all indus-

tries in national water use. Water investment projects of the Soil

Conservation Service included in this study do not influence major in-

dustrial users of water. However, certain projects carried out by the



National water use data are available in the Census of
Manufactures [25, Vol. 1, pp. 7.16-7.17].






Corps of Engineers could influence the location and expansion of these

industries. Water use by these industries is important in assessing

employment changes due to investments in water resource programs.



General Model Specification


Economic interpretation of changes in the exogenous shifters in-

volved in equations (2.1) and (2.2) in the previous chapter made it

apparent that interaction occurs in both equations simultaneously.. Esti-

mation of the effects of exogenous changes on employment and firm

numbers could be considered using a system of equations. Figure I rep-

resents the simultaneous changes occurring as the result of a change in

resource supplies. Total interaction as the result of all changes in

the exogenous variables can be represented for each industry by a system

of equations. The availability of n county observations for each in-

dustry gives the following general system for the kth industry.


(3.1) dQLkt = 10 + PlldXft + 12dPkt + +13 ikt 14d hkt


+ lldNkt + Ulkt t = 1, 2, . ., n



(3.2) dNkt = 20 + 21dXft + 22dPkt + 23dPikt + dZhkt


+ 25dWdkt + U2kt t = 1, 2, . ., n

where the subscript t represents county observations, and

dQLkt = total change in the quantity of labor demanded

in the kth industry endogenouss).
th th
dXft = total change in the f shifter of the n

factor supply (exogenous).







dP. = total change in product demands in the kt
j kt
industry (exogenous).

dPikt = total change in factor supplies in the kth
ikt
industry for all factors other than the nt

factor (exogenous).

dZh = total change in firm production possibilities
hkt
for firms in the kth industry (exogenous).

dN = total change in the number of firms in the kt
kt
industry endogenouss).

dWdkt = total change in shifters of firm entrepreneur

supply for firms in the kth industry (exogenous).

The B coefficients and the parameters of the distribution are un-

known which leaves the problem of obtaining estimates of these parameters.

By rearranging equations (3.1) and (3.2) as


(3.3) -dQLkt + ylldNkt + ,0 + PlldXft + 1 dP + 3 dPkt


14dZhkt = Ulkt


(3.4) -dNkt + 20 + PdXft + 22dP.kt + 23dP.ik + 4dZhkt


+ 25dWdkt = U2kt


Estimates derived using the following model are consistent and
asymptotically efficient if the following assumptions are made:

E [(kt)] =


E[U(kt) U(kt+s)] if s
t U t = otherwise

E X(kt) U(kt 0







they can conveniently be written in matrix notation as


(3.5) -1 YIl dQLkt 010 oil 3 12 $13 314 0

0 -kt 20 21 [22 :23 24 25



SdXf dPjkt ikt hkt dkt
-U
U 2kt

and finally condensed to


(3.6) rY(kt) + X(kt) = U(kt)

where


r= D -_(kt) = dN



B o10 11 12 313 14 0

L20 21 B22 F23 e24 H25



X(kt)= dXft dPjkt dPikt dZhkt dWdk1


Ikkt
U(kt) =


The subscript k indicates that this particular system is for the

k industry. Each industry under study would be represented by an equa-

tion of this general form. Equation (3.6) presents the simultaneous

linear equation model in its structural form. Each of these sources pro-

videsa detailed discussion of simultaneous equation models, their

assumptions, limitations, and possible problems which might arise in

their application.







Structural Estimation

It is necessary that an equation in a set of simultaneous

equations be just or over-identified before estimates of the parameters

can be obtained. The order condition for identification can be used to

determine whether a given equation is identified. If A is equal to the

number of predetermined (exogenous) variables that do not appear in an

equation and B refers to the number of endogenous variables in the equa-

tion, then the following conditions determine the identification of the

equa t ion.

A = B 1 just identified

A > B 1 over identified

A < B 1 under identified


Reduced Form Estimation

The complete structural system may be written so that each endog-

enous variable is expressed as a function of all exogenous variables

appearing in the system. This representation of the system is referred

to as the reduced form and brings out the explicit dependence of the de-

pendent variables on the predetermined variables and the disturbances.

Premultiplication of equation (3.6) by F-1 and rearrangement gives

-1 -1
(37) Ykt= X(kt) + U(kt)


or (3.8) Ykt = nx(kt) + v(kt)


where -r1 I
10 11 . 15
-1
n =- r'P=

20 fl21 . 25

'Goldberger L26, pp, 352-372] and Johnston [27, pp. 241-252]
provide detailed discussions of the identification problem.







V1 V
-Ikt
V(kt) U (kt) V2kt
2kt

Estimates from the reduced form may be derived directly by

ordinary least squares since their disturbances are linear combinations

of the structural disturbances.5 However, in general practice the struc-

tural parameters are estimated first and then the reduced form parameter
-l
estimates are obtained by use of i = -Ir P.



Individual Industry Models


A two-equation system was specified for agriculture according to

the theoretical framework outlined in equations (2.1) and (2.2), and the

general estimation model outlined in equations (3.1) and (3.2). Single

equation models were specified for construction and the various manu-

facturing industries. These models are presented below with each

variable coded similarly to equations (3.1) and (3.2). That is, changes

in factor supplies are represented by X, changes in shifters of firm

supply by W, changes in product demands by PP, changes in factor prices

by FP, and changes in firm production possibilities by Z. Data sources

for each variable are given in the latter sections of this chapter and

the mathematical derivation is given in Appendix B.

5The assumptions for the reduced form disturbances are

E [V(kt)] = E [U(kt = 0

E/ V E 1 -1 =F for
E (kt) (kt) (kt) (kt)
t = t
= G otherwise
r [(kt v (k) = o
(kt) (kt)






Agriculture

The system of equations representing the agricultural industry is
6
(3.9.1.1) Eli = l + rl Xri + 711PP li + 811FPli


+ l911 li Z + 10, 11GRPi + ,11N i + Illi


6
(3.9.1.2) N; = 3012 + 1 X 7PPi + 2FP
r = 1

+ 912 li+ 10,12GRPi + 1,12 li + 1 i 1 22WE i


+ 13,12WAli + 112i

where the first subscript is the parameter number, the second

subscript is the industry number, the third subscript is the

equation number where

Eli = Change in agricultural employment for the period

.th
1960-1970 in the ith county.

Xli = Change in federal and state expenditures per pupil for

primary and secondary education during the period 1960
.th
to 1970 in the i county.

X2i = Change in total construction expenditures in water

development projects by the Corps of Engineers during
.th
the period 1960 to 1970 in the i county.

X3i = Change in total construction expenditures in the PL-566

Small Watershed Program by the Soil Conservation Service
th
during the period 1960 to 1970 in the i county.

X4i = Change in total investment in the Agricultural Conserva-

tion Program (renamed the Rural Environmental Assistance







Program in 1971) by the Agricultural Stabilization and

Conservation Service during the period 1960 to 1970 in
.th
the i county.

X5i = Change in total loans and grants for community water and

sewer systems and waste disposal systems made by the

Farmers Home Administration during the period 1960 to
.th
1970 in the h county.

X6i = Change in acreage of allotment crops due to reduction in

allotments between 1959 and 1969 weighted b/ the propor-

tion of the total value of the allotment crop to total
.th
value of crops and livestock in the i county in 1959.

PPIi = Change in price index of agricultural commodity groups

during the period 1959-61 to 1969-71 weighted by the

proportion of the value of the commodity group to the
.th
total value of crops and livestock in the i county in

1959.

FP = Change in the average annual wage per hired farm worker
.th
during the period 1959 to 1969 in the i count/.

Zli = Change in the Southeast index of agricultural output per

man-hour for corrmodity groups during the period 1959-61

to 1969-71 weighted b/ the proportion of the value of the

commodity/ group to total value of crops and livestock in
.th
the th county in 1959.

GRP. = Intercept shifter dummy variable

ith GRP. = I when urban-oriented county
Sw0 when rural-oriented county

N i = Change in the number of farms during the period 1959 to
1969 for the ith county
1969 for the county.







WW1i = Change in total annual nonagricultural wage payments

during the period 1960 to 1970 per agricultural
th
employee in 1960 for the i county.

WEli = Change in total nonagricultural employment during the

period 1960 to 1970 per agricultural employee in 1960
th
for the i county.

WAli = Change in the number of farm operators who were 55 or

more years of age during the period 1959 to 1969 in
.th
the i county.

U ll U i = Disturbance terms.
111i 1121
The Greek letters B and y represent the parameters to be estimated.

Variables endogenous to the system are Eli and N i. Equation (3.9.1.1)

is over-identified and equation (3.9.1.2) is just-identified.


Construction

Data availability on the construction industry precluded the

calculation of changes in the product price and technology variables.

Data on firm number changes also were inadequate. However, since the

construction industry would be one of the more important in evaluating

the primary employment effects of investments in natural resources a

single equation model was formulated in this industry. The equation for

the construction industry is

5
(3.9.2.1) E2 = 02 + X FP + 72GRPi
2r r2 ri 62FPi 72

+ 82"'+12 i + 92Ei + 2i

where

E 2 = Chanye 'n construction employment for the period 1960

to 1970 in the ith count%,.







Xli to X5i = Same as in equation (3.9.1.1).

FP2i = Change in average annual wages per construction

employee during the period 1958 to 1967 for the
.th
I county.

GRP. = Same as in equation (3.9.1.1).

WW2i = Change in total annual wage payments during the period

1960 to 1970 for all nonagricultural and nonconstruction
.th
employees per construction employee in 1960 for the i

county.

WE2i = Change in total nonagricultural and nonconstruction

employment during the period 1960 to 1970 per con-
.th
struction employee in 1960 for the i county.

U l2 = Disturbance terms.


Manufacturing

The remaining seven industries are manufacturing industries. A

two-equation model did not yield results comparable to those obtained

for agriculture. The number of manufacturing firms is much smaller and

the size of firms (measured in terms of the number of employees) is

generally larger than for the agricultural industry. Estimation of

changes in the number of manufacturing firms did not add to the interpre-

tive and explanatory power of the system of equations. Much of the

impact of changes in the exogenous variables for manufacturing industries

is transmitted through employment changes within existing firms rather

than change in the number of firms. Therefore, changes in the number of

firms in each manufacturing industry was treated as an exogenous variable

and a single-equation model was specified for each manufacturing industry.

This model is of the form







0 9.k.1) Eki Ok + rk Xri + 6k PPki +57k FPki
r = 1

+ 8kZk.+ GRP. + a W.+B W
kk 9k i lO~k'M + 11,kk ki


+ $12,k 14ki + ki

k = 3, 4, ,7, Do ND

where

E ki = Change in k th industry employment for the period 1960

to 1970 in the i th county.

X li to X 5i = Same as in equation (3.9.1.1).

PP ki = Change in price index for each SIC three-digit level

industry commodity group in the k th industry during

the period 1959-61 to 1969-71 weighted by the proportion

of value added by the coqimodity group to total value

added by the commodity group to total value added in

the k th industry in 1958 for the i th county.

FP ki =Change in average annual wages per production worker

for each SIC three-digit level industry group in the

k th industry during the period 1958 to 1567 weighted

by the proportion of production 'worker employment in

each Subgroup to total production worker employment in

the k th industry for the 1 th county.

Z ki =Change in the index of output per man-hour for each StC

three-digit level industry group in the k th industry

during the period 1959-61 to 1969-71 weighted by the

proportion of value added by the commodity group to

total value added in the k th industry for the it







GRP. = Same as in equation (3.9.1.1).

WWmi = Change in total annual nonagricultural and non-kth in-

dustry wage payments during the period 1960 to 1970 per
th th
k industry employee in 1960 for the i county.

WEki = Change in total nonagricultural and non-kth industry

employment during the period 1960 to 1970 per kth in-

dustry employee in 1960 for the i county.

Nki = Change in the number of firms in the kth industry during
th
the period 1959 to 1969 for the i county.

U ki = Disturbance term.

The symbols 0 and ,, represent the parameters to be estimated.



Model Estimation Procedure


Two-stage least squares was used to estimate the structural

parameters of equations having the form of (3.9.1.1) in the two-equation

system for agriculture. Ordinary least squares was used to estimate

equations having the form of equation (3.9.2.1) in each of the uses of

the two-equation models. This statistical procedure was also used for

all single-equation models. Estimates of the reduced form coefficients

for the two-equation models were derived from the structural parameter

estimates.








A computer program written by William James Raduchel [28] was used.

7Ibid.








Measurement of Variables and
Empirical Expectations


Changes in employment and firm numbers as defined in equations

(3.9.1.1) and (3.9.1.2) were considered as endogenous variables. In the

model that forms a system of equations, the endogenous variables are con-

sidered functions of some or all predetermined (exogenous) variables.

In the first equation employment changes were also considered as a

function of the endogenous variable for firm number changes. Predeter-

mined variables represent changes in exogenous shifters of product

demand, factor price, critical resource supply, firm production possibil-

ities or firm entrepreneur supply functions. Each of these general types

of shifters is represented by one or more variables in each equation.

Each variable, the type of shifter it represents, and the expected ef-

fects of changes in these variables on agricultural employment and the

number of farms are given in Table 2. A similar illustration of the ef-

fects on construction and manufacturing employment is given in Table 3.

Detailed discussion of each variable follows in the remaining sections

of this chapter.


Employment

Changes in employment for each county were computed for the time

period 1960 to 1970 as reported in the 1960 Census of Population [29,

Table 85] and the 1970 Census of Population [24, Table 123]. Employment

is reported in this source using basically the same industry categories

as suggested by the Standard Industrial Classification (SIC) of industries.

It was necessary to combine employment reported in some industries to con-

form to the industry classification shown in Table 1. Changes in

employment were used as dependent variables.














C

rUC










- 0
1 -
0.1.- 43

















x v
1U 0
U



















43
z



































a
U
a
L












0 4
43-




43' -


X I'
LUi 0
U





z





U
c3


I U


L

,4


10

1a 0



U
&- -








UU U
aa a





u u


i I I + 0
4-


43
a a
3 L I-
-0 0

U
4 -
a


*


0 0 k
u
a 4

- I- 4s









U u U








u u
3 04
a a

43 3 4
V. 3 4
43 3 4
3 3 4

- -


43
a a

0 0 .-
- 3
a


434 43 43 43 4



n a u u U
nx m C3 .


U U U
CC C




a -



U U U
c= c= .=
-= -=


a a


L. .. ...
U U U
vl r l





- k- Ia



o u u
c c 4
- -


-- -r \ .
X x x 0. C .


C0 0 3 C



2 0 U C
43 > 43 4 0. &
~ ID L

S- 0 04 -
E i 0 o 0 -
S 4 E 3 o
43 0 0 r 0 0 3
r 4 l C 'l 3 : C



0 a c 6 >. 60 0
D- c D 0 0.
^ U .-fl.D E3 0.0- U. 03 L E .



U 0 u U U 4 4 3 ;
Su c E3 a 1 -


W,
&&~ ~ G. I* u-fc


a a a


0 0 0















a +
v n
u u
v e a u















a











'
-- k-



U
a a a e










L.
U




0
93

U

















- --a a
S ro
.0 .0 .0 U
U_


0





CC
LO















(1U

a L O
fu





u -









U U ,)



) 0





-' L-
4 ,


S-o 0 0








S"
U 0

o --











1- 0 -
> (a fu





























r' o
0 C *-


IU L





C L.
(U a)


























VI
113 (UT





O '- o












L L










CD O O
C C



(U C


0 0







C- 3 U
n o Q






dmo
O / /
ct 0
[1 Q
(0 3


2
a co
v 43
- 43 -



- U
Q 0


0









i








55







u + +
lu
x c
0
u






a 10 1
0 0 v 0
1 1 1 1 c 43
4C u u v u u
z 4C
-tl -V c
0
4C
m

E v U lu
I
10 10 -C 10
v -u u u
v u
C





ID 10 11
I 1 10 m
c v lu c
I I
u u u
c I Q u u c u
lz Lrl


4"

ND v c v
m
0 c ID
c
u u u
o 41






ot
L u v

ru
4-J



t4-





Ln
Q)
Ln
cc










lu C., 7

CL -U C, c


2
4-1
u c -0 lu
ol
V
4-J
0)
r L







4A
0








A







Firm Numbers

Changes in the number of manufacturing firms for the period

1958 to 1967 were obtained from data reported in the 1958 Census of

Manufactures [30, State Table 7] and the 1967 Census of Manufactures

[23, State Table 9]. Firm numbers are reported for each two-digit SIC

industry in each county. Some industries were combined to conform to

the classification of industries used in this study.

Changes in the number of farms from 1959 to 1969 for each county

were obtained from the 1959 Census of Agriculture [31, County Table l]

and the 1969 Census of Agriculture [32, County Table 1]. Changes in the

number of farm firms were used as both predetermined and endogenous

variables. This distinction is discussed in a later section.


Factor Supplies

Changes in the supply of critical factors in a county should in-

fluence that county's labor employment. Investments that increase the

supply of a given factor would be expected to cause the price of the

factor to decline initially. A price decline would entice users to sub-

stitute more of the factor for other production inputs including labor.

With perfectly elastic product demands and a reduced price for the

critical factor, residual returns to firm operators should increase. In

the absence of any substantial barriers to entry of new firms into the

industry, new firms would be established, and this would increase in-

dustry output. Entrance of new firms would result in an upward shift in

the demand for all factors, including labor, causing the prices of those

factors having upward sloping supply functions to rise. Existing firms

would also expand output. Residual returns to firm operators would sub-

sequently fall, resulting in a cutback in factor employment. At the new







equilibrium, the quantity of labor employed would be greater or less

than the initial quantity, depending on the relative sizes of the demand

and supply elasticities as pointed out in the theory section. It would

also depend on the elasticity of substitution of labor for the other

factors whose prices varied. Labor could probably not be easily sub-

stituted for many of the natural resources considered in this analysis.

Therefore, investment programs to increase the supplies of these natural

resources should also increase the amount of labor required to complement

industry expansion resulting from either the initial project construction

phase or from users of the project.


Education investments (X1).--Federal and state expenditures per

pupil for education were considered to be exogenous to a county. Changes

in these expenditures over the period 1959-1960 to 1969-1970 were in-

cluded as a measure of exogenous shifts in the investment in a county's

human resources. Expenditure estimates were obtained from state educa-

tion agencies in Alabama [33, 34], Mississippi [35, 36], and Georgia [37,

38], and the Florida Statistical Abstract [39].

Increases in education expenditures should partially reflect an

increase in the number of persons in the recipient county attaining a

given educational level as well as an increase in the average productivity

level of the labor force. Unless outmigration from the counties of

people receiving the education occurs, an upgraded labor force should

attract potential employers and eventually result in increased employment

within the recipient counties.

The effect of increases in education levels or agricultural

employment and farm numbers should be opposite their effect on construction






and manufacturing employment. Since agricultural labor draws heavily

from the unskilled labor force it seems likely that increased educational

opportunities would reduce this labor force and lead to outmigration from

the more rural counties to more attractive job alternatives in urban areas.

This would lead to fewer agricultural workers and probable consolidation

of farms to gain operational efficiencies. Increased educational oppor-

tunity should, therefore, yield negative coefficients for agricultural

employment and farm numbers and positive coefficients for construction

and manufacturing employment as shown in Tables 2 and 3.


Corps of Engineers' natural resource investments (X ).--For this
2-
variable as well as the other types of natural resource projects, total

project expenditures by county over the period 1960 to 1970 were used as

the independent variable. Interpretation of the estimated coefficients

for these investment variables should provide insights into two components

of the analysis. First, the empirical significance of the various types

of natural resource investments on local employment and firm numbers can

be shown, and second, the relative importance of the various natural re-

source investment categories in influencing employment and firm numbers

can be appraised.

Investments by the Corps of Engineers in civil works and new

work construction were obtained For each county from the various district
8
offices which administer portions of the four-state area. Investments


Corps of Engineer personnel providing data through personal
communications were:

J. W. Dement, Chief, Engineering Division, Memphis
District, U.S. Army Corps of Engineers, Memphis,
Tennessee.
W. T. Moore, Chief, Engineering Division, Savannah









projects categorized into multipurpose, navigation, flood control, beach

control, and recreation projects. The major portion of expenditures was

for-flood control and navigation with a very small portion allocated to

beach control and recreation. Due to the large amount of the invest-

ments going into flood control and navigation projects, no distinction

among the above investment categories was made. Some expenditures by

the Corps for construction projects along the Mississippi River could not

be allocated to counties. Therefore, these investments were not in-

cluded in the analysis. Investments occurred in 148 of the 375 counties

comprising the four-state area.

Investments by the Corps of Engineers should also have con-

trasting effects on agricultural and manufacturing employment. Improved

flood control would be beneficial to agricultural areas by making more

,land available for use. This would bring about a two-fold reaction.

Both expansion of existing farms and the entrance of new Farms would

occur with the probable result being decreased residual returns to each

firm as the price of land is bid upward. The probable consequence would


District, U.S. Army Corps of Engineers, Savannah,
Georgia.
Powell Williams, Jr., As~st. Chief, Engineering Division,
Mobile District, U.S. Army Corps otF Engineers,
Mobile, Alabama.
George Marsh, Acting Chief, Engineering Division,
Jacksonville District, U.S. Army Corps of Enigineers,
Jacksonville, Florida,
J. L. Smith, Chief, Construction Division, New Orleans
District, U.S. Army Corps of Engineers, New
Orleans, Louisiana.
K. E. McLaughlin, Comptroller, Vicksburg District, U.S.
Army Corps of Engineers, Vicksburg, Mississippi.
F. P. Gaines, Chief, Engineering Divi lon, Nashville







be further expansion of the larger more established farms with the

overall effect being a reduction in farm numbers. It follows that the

larger farms might also operate with a smaller total labor force though a

more efficient operation realized as the result of the larger farm size.

The consequence would be reductions in farm numbers and agricultural em-

ployment as indicated by the coefficient sign in Table 2.

In contrast to the effect on agriculture, construction and manu-

facturing employment would likely increase as the result of Corps of

Engineers' investments. Employment would certainly increase in the re-

cipient area during the initial construction phase of the project.

Initial effects might also be felt in manufacturing provided that local

area materials were used. More importantly for manufacturing, however,

would be the effect occurring during the life of any project. Improved

transportation facilities, protection from flood damage, etc., would en-

courage the entrance of new firms and resultant employment increases.

Since firm consolidation in manufacturing does not occur as readily as in

agriculture, the indirect effect of decreases in firm numbers and employ-

ment is probably not large enough to offset the initial positive gains in

the recipient area. A positive coefficient for construction and manfac-

turing employment would be expected as indicated in Table 3.


Soil Conservation Service PL-566 investments (X ).--Construction
3-
expenditures by county from 1960 to 1970 for the Small Watershed Program

were obtained from personnel in each of the four-state offices of the

Soil Conservation Service. Data were tabulated from 239-B forms which


9Soil Conservation Service data were provided through personal
communications from:







gave actual dates of construction expenditures for each project.

Investments were then allocated to each county based on project location

as identified on maps prepared by the Soil Conservation Service. A total

of 111 counties received these types of investments during the study

period. Investments were measured in thousands of dollars.

The Small Watershed Program is designed to aid in the solution

of several types of problems. Of major importance among these is the re-

duction in floodwater damages to cropland, residences, businesses, and

protection of the health and lives of people from floods. Other poten-

tial and existing problems that this program attempts to alleviate

include erosion and sediment damage, improper drainage, and irrigation

needs. Recreation, fish and wildlife enhancement, and improvements in

the economic and social well-being of people have also received atten-

tion. These latter categories have been given increased emphasis in

recent years. PL-566 investments by the SCS which result in improve-

ments for the local recipient areas should provide conditions that

affect agricultural and manufacturing employment and firm numbers within

the local areas in a manner quite similar to that of Corps of Engineers'

investments. Reduction of floodwater damage to cropland should in

total reduce the number of farms and agricultural employment as indicated

in Table 2. Although some new farms might become established




Barbara Kennedy, Accounting Technician, Soil
Conservation Service State Office, Auburn,
Alabama.
Robert Salsman, Financial Manager, Soil Conserva-
tion Service State Office, Jackson, Mississippi.
George Adair, Accounting Technician, Soil Conserva-
tion Service State Office, Athens, Georgia.
Gertrude Griffin, Accounting Technician, Soil
Conservation Service State Office, Gainesville,
Florida.
I







the consolidation effect into larger farms to take advantage of

improvements made possible by the PL-566 project should be greater.

Construction employment should increase in the recipient area

during both the initial and secondary project phases and ultimately as a

result of the project in a manner similar to that discussed for Corps of

Engineers' investments. Also to be expected is an increase in manufac-

turing employment as indicated in Table 3. Increased output resulting

from project expenditures should provide a base for more manufacturing

employment in conjunction with increased enhancement for manufacturing

firm location resulting from the reduction in floodwater damages to

residences, businesses and through the effect of other firm location

attributes improved by the investment project.


Agricultural Stabilization and Conservation Service ACP

investments (X)..--Investments by the Agricultural Stabilization and

Conservation Service constitute a joint effort by the public sector,

farmers, and ranchers to share the cost of establishing needed conserva-

tion measures. These conservation programs include practices to

protect, improve, and renew soil, water, woodland, and wildlife resources

of private landowners. Data for the analysis were obtained from annual

state ASCS reports during 1960 to 1970 for Alabama [40], Mississippi

[41], and Florida [42]. Expenditure information included cash payments

to farmers and allowances paid to vendors for conservation materials

furnished farmers. Data for Georgia were taken directly from computer

printouts.0 A total of 374 counties had participants in the ACP



10Personal communication from the Data Division, Agricultural
Stabilization and Conservation Seivice, U.S. Department of Agriculture,
Washington, D.C.







Program during the study period. Investments were measured in thousands

of dollars.

Conservation measures which make land more available could also

influence the expansion of existing farms as well as encourage new farms.

The trend historically has been toward larger firms. Since this is a

cost-sharing program it seems logical to expect the larger farmers to

take advantage of this program opportunity and expand operations even

more. The indirect effect of entrance of new farms should be more than

offset by the consolidation of existing farms leading to a decrease in

total agricultural employment and farm numbers. Conservation measures

that remove land from production would provide a similar circumstance.

This negative overall effect is indicated in Table 2.

Since this type program requires some construction activity and

manufactured input which usually are purchased locally, a positive

effect on construction and manufacturing industry employment would be

expected as indicated in Table 3. Increased output would also be ex-

pected to result from the application of conservation measures leading

to a need for more processing and support facilities which in turn

should have some positive effect on manufacturing employment. Any

negative feedback effect on manufacturing employment should not be large

enough to affect the initial positive effect.


Farmers Home Administration investments (X) .--Loans and grants

for community water, sanitary sewer, arid solid waste disposal systems

were also considered to be investments that would influence employment

and farm numbers in each county as indicated in Tables 2 and 3. This

program provides financial assistance to communities in developing es-

sential new public service facilities and in expanding existing







facilities. Data for these investment loans in thousands of dollars were

obtained from the various state directors of the Farmers Home Administra-

tion.1 A total of 278 counties received financial assistance from FHA

during the study period.

Services and facilities provided by this type program are neces-

sary before a community can expand with regard to attracting new industry

and in turn services to support these industries. Communities demon-

strating adequate services will likely attract new industry and thus

expand employment in construction and manufacturing industries. Expan-

sion of existing firms in the community might also occur, and this

further supports the positive employment coefficient demonstrated in

Table 3. This program does not specifically influence a production in-

put used in agriculture such as land or water in the same manner as the

four earlier programs. A similar negative effect on agricultural employ-

ment and farm numbers as discussed for the earlier programs would be

expected. As community facilities become available and the community

begins to develop its manufacturing base, job alternatives for agri-

cultural employees and farm operators become more available. Smaller

farms are soon consolidated with the displaced operators assuming other

types of employment. Fewer employees are then required because of more

efficient operations and the negative effect occurs.


Crop allotment (XI).--Changes in crop allotments represent the

effect of shifts in a perfectly inelastic factor supply on the number of



State Directors, Farmers Home Administration, providing data
through personal communication were: S. B. Wise, Jackson, Mississippi;
John N. McDuffie, Atlanta, Georgia; John A. Garrett, Montgomery, Alabama;
and William Shaddick, Gainesville, Florida.









would beepced to increase the market price ofteallotment, or of

land, leading to lower residual returns to farm operators and ultimately

to a reduction in the number of farm firms. As the number of farms de-

cline aggregate demand for allotments would decline and thus lower their

market prices. The magnitude of changes in the number of farm firms and

consequently in agricultural employment due to a decrease in allotments

would depend on farm operators' responsiveness to changes in their

residual returns, the amount of allotments used in the production

process, and the actual level of operator returns.

Reduction s in acreage allotments between 1959 and 1969 for all

allotted crops were computed for all counties having acreages of these

crops. Annual reports for 1959 and 1969 from the Agricultural Stabiliza-

tion and Conservation Service in Alabama [401, Mississippi [411, Georgia

[433 and Florida [421 provide data on allotted crop acreages. County

reductions were weighted according to total value of sales of each

crop as a proportion of total value of crop and livestock sales in the

county in 1959 as derived from data available in the 1959 Census of

Agriculture [311. Declines in harvested acres in each county were also

computed using the same data sources that provided information on al lot-

merit reductions. The smaller of these two changes was then selected as

the effective cut in allotments. Allotment reductions w,ere not con-

sidered relevant for a county if its harvesting acreage in 10,59 was less

than the county's acreage allotment for the selected crop in 1969.
Posiivecoeficent ,oul beexpcte as how inTabe 2







Product Demand

Use of a product price as an indicator of product demand is

based on the assumption of perfectly elastic demand functions at the

county level. Producers in both agricultural and manufacturing in-

dustries at the county level are assumed to be price takers and thus

face perfectly elastic demand functions. An increase in product price

would initially increase residual returns to firm entrepreneurs. This

would have the effect of enticing new firms into the industry and

ultimately an increase in the quantity of resources employed in the pro-

duction process including labor. Also, existing firms would expand

output and hence increase their demands for production factors. In-

creases in demand for factors with upward sloping supply functions would

lead to increased factor prices and reduced returns to firm entrepreneurs.

Consequently, firm numbers would decline and this would result in a re-

duction in labor employment. Existing Firms would also reduce their out-

put and cause reductions in resource demands. The net change in labor

employment would depend on the relative magnitude of both the initial

and indirect effects of these changes in output of existing firms and in

the number of firms. A product price variable was not included for

the construction industry since output is not easily defined in terms of

a product with an established market price.


Agricultural product price (PPl).--Changes in price indexes were

computed for each of the seven major agricultural commodity groups pro-

duced in the study area using three year averages centered on 1959 price

indexes [44] and 1969 price indexes [45]. These changes were weighted

by the 1959 value of each commodity group as a proportion of the total




67


value of crops and livestock produced in each county. The resulting

measure was a weighted change in product prices faced by farm producers

at the county level. Value of products sold was obtained from the 1959

Census of Agriculture 131, County Table 51. A prior! specification of

the net result of increases in agricultural product prices on employment

is difficult. It is quite possible that increases in Agricultural

product prices could lead to a reduction in agricultural employment and

farm numbers due to farm consolidation. Since some agricultural opera-

tions do allow fairly easy entry, the opposite effect could occur under

certain conditions.


Manufacturing product rice (PPk).----Changes in national whole-

sale price indexes as reported by the Bureau of Labor Statistics [46] for

three-digit level (SIC) industries between 1959-61 and 1969-71 were used

in computing a county product price for each two-digit level industry.

For each industry, county price changes were obtained by weighting the

change in the three-digit national wholesale price indexes by the-1958

value added by manufacturing for each three-digit level industry as a

proportion of the total value added for the two-digit industry in the

county. Data used to calculate value added for each industry were ob-

tained from published data made available by the US. Bureau of the

Census [30, 471 .

Expected effects of product price increase in the manufacturing

industries are also difficult to specify. Existing firms wo 1d be ex-

pected to expand output and new firms enter the industry as the result

of a product price incerase. This assumes there ae no barriers to

entrance. Increases in employment should occur. Output increases







expansion being offset. With manufactured products, unlike agricultural

commodities, some apparent downward inflexibilities of prices would help

support in part a conclusion that the indirect effect of decreasing em-

ployment would not completely offset the direct effect leaving a positive

overall effect on manufacturing employment. It does remain possible that

decreased residual returns to firms as the result of entry by new firms

would be substantial enough to cause an actual decline in employment.

Both alternatives are indicated to demonstrate the effect of product

price increases on manufacturing employment in Table 3.


Factor Price

Changes in the price of factors whose supply is assumed to be

perfectly elastic would affect the net returns to firm entrepreneurs and

consequently the number of firms. Indirectly, the level of labor employ-

ment would be affected. As the price of factors having perfectly elastic

supply functions increased, labor as well as other production factors

would be substituted for these inputs to the extent possible. This would

result in an increase in the price of all factors having upward sloping

supply functions. As these factor prices increased, residual returns to

firm entrepreneurs would decrease and consequently the number of firms

would decline. A reduction in firm numbers would decrease factor de-

mands, resulting in a decline in factor prices. Increases in residual

returns to firm entrepreneurs would entice some new firms into the in-

dustry with resultant increases in labor employment. Employment levels

under new equilibriumr conditions would depend on the relative magnitudes

of these various changes. The more inelastic the supply function of

critical factors, other than labor, the larger the price increase will







be for that factor as demand for it increases. Consequently, labor

whose supply function is more elastic would be substituted for the

higher priced factor with a resultant employment increase.


Agriculture wage rate (FPI1.--A proxy variable was used as the

annual wage rate for agricultural employees. Total expenditures for

hired farm labor in 1959 and 1969 for all farms in each county were

divided by the total number of hired farm laborers working 150 days or

more each year in that county to obtain an annual wage per worker. The

change in this wage was then computed. These data were obtained from

the 1959 and 1969 Censuses of Aqriculture [31, 32]. Employment effects

of increases in hired farm labor wage rates should be negative as shown

in Table 2. Wage increases would result in higher factor costs to op-

erators. This would encourage substitution of other factors for labor.

Smaller farms would not be able to make sufficient substitutions and

would not be able to compete with larger and more efficient farm opera-

tions. Farm numbers would then decline through consolidation and

expansion of existing farms.


Manufacturing waqe rate (FPk).--Changes in average annual produc-

tion worker wage rates between 1959 and 1970 for each two-digit level

manufacturing industry in each county were used for manufacturing wage

rate changes. Data were obtained from the 1959 and 1970 County Business

Patterns for each state [48, 49]. If the two-digit level industry wage

was not reported for a county due to disclosure problems, the change in

average annual production worker wage for all manufacturing industries in

the county was used. Increases in an industry's wage rates would be ex-

pected to result in a decrease in labor employment within the industry as







indicated by the negative coefficient sign in Table 3. Other production

factors would be substituted for labor as the price paid to labor in-

creased.


Technology

Changes in technological forces that affect agriculture and manu-

facturing industries should have an effect on the amount of labor

employed. Similar to the other types of shifters discussed previously.

technology changes would also affect factor demand, product supplies,

and the number of firms. Increases in technology that were output in-

creasing and nonlabor input decreasing would cause the quantity of

products produced to increase with subsequent decreases in the use of

inputs. Prices of those inputs having inelastic supply functions would

decrease since demand for them would decline. Further substitution of

the lower priced inputs for labor could cause a decrease in employment.

If technology changes had been labor decreasing, the quantity of labor

would have decreased initially. The indirect effect of these changes

would be an increase in the number of firms concomitant with an increase

in residual returns as a result of the change in technology. flew firms

would then increase the demand for all factors and reduce firm residual

returns. Equilibrium quantity of labor demanded could be either smaller

or larger than the initial quantity demanded depending on the degree of

factor supply inelasticity, substitutability of labor for the other

factors used, and the magnitude of changes in the number of firms.


Agricultural technology (Z.).--Changes in output per man-hour in

agriculture were used as indicators of changes in agricultural technology.

Changes in the index of output per man-hour for six major commodity groups







in the Southeast were computed using three-year averages centered on

1959 and 1969 [50, p. 8]. These changes were then weighted by the 1959

value of each commodity group produced as a proportion of the total value

of crops and livestock produced in each county obtained from the 1959

Census of Agriculture [31, County Table 5]. The resulting measure was

a weighted change in labor productivity for each county. Trends in out-

put per man-hour and advances in agricultural mechanization suggest that

technology increases in agriculture are likely to be labor decreasing.

A negative effect on agricultural employment should be indicated as sug-

gested in Table 2. Similar effects would be expected on farm numbers.

Technology advances should enable the operation of larger farms with

resultant decreases in farm numbers.


Manufacturing technology (Zk).--Changes in technology for each

manufacturing industry were computed in a manner similar to that for

agriculture. Changes in national output per man-hour indexes between

1960 and 1970 for three-digit level industries were used in computing a

county technology change variable for each two-digit level industry.

These indexes are published by the Federal Reserve System [51]. Changes

in the national output per man-hour indexes for the three-digit level

industry were weighted by the 1959 value added of each three-digit level

industry as a proportion of the total value added by the two-d;git level

industry in the county. The same value added data used in calculating

industry product price was used in the weighting procedure. Technology

changes in the manufacturing industries have employment effects similar

to those in agriculture. The negative effect indicated in Table 3 im-

plies that technology changes are probably labor decreasing. A technology








variable for the construction industry was not included since output per

man-hour indexes for construction were not available.


Farm Operator Supplies

Changes in farm operator supplies affect agricultural employment

in various ways. Several variables used in this study are quite unique

with respect to the types of shifters discussed earlier. These shifters

are thought to affect farm operator supplies which in turn affect agri-

cultural employment. Changes that increase the number of farms indirectly

cause increases in the amount of products produced and factors used in-

cluding labor. Ultimately, the price of factors having less than

perfectly elastic supply functions would increase with concomitant de-

creases in the number of farms, quantity of products produced, and

quantity of labor employed. Equilibrium employment levels would depend

on the relative magnitudes and effects of the described changes. In

general, declines in farm operator numbers should cause declines in

agricultural employment.


Agricultural waqe opportunity (WW1).--Wages in industries other

than agriculture represent changes in the opportunity cost to farm op-

erators of remaining in present employment as a result of changes in

wages in other employment alternatives. Initially, wage increases in

employment alternatives would decrease the number of farm operators re-

naining in agriculture. As the larger farms realize greater residual

returns some increase in farm numbers might occur. This effect should

be minimal with an overall decline in farm numbers expected as indicated

in Table 2. The movement to fewer, larger, and more efficient farms

should then cause a negative effect on agricultural employment as

indic.i-red in Table 2.







Change in agricultural opportunity wages between 1960 and 1970

in each county was determined using data on employee wages obtained from

County Business Patterns [48, 49] and the Census of Population [29].

Change in annual nonagricultural wages between 1959 and 1970 per agri-

cultural employee in 1960 was used as the indicator of wage opportunity

for agricultural employees. Changes in county unemployment levels would

have provided an alternative measure for this variable.


Agricultural employment opportunity (WEl).--Increases in employ-

ment opportunity in alternative employment situations would be expected

to decrease the number of farm operators remaining in agriculture in a

manner similar to that of increases in wages in employment alternatives.

Changes in employment alternatives were calculated using employment data

obtained from the 1960 and 1970 Censuses of Population [24, 29]. Changes

in nonagricultural employment between 1960 and 1970 per agricultural em-

ployee in 1960 indicate employment opportunities for agricultural workers

and farm operators.


Farm operator age (WA).--Farm operator age represents the change

in the number of farmers who were 55 or more years of age during the

period 1959 to 1969. This variable is intended to reflect the relative

effects of potential operator retirements on the number of farms during

1959 to 1969. The greater the number of farmers who are reaching an

older age, the greater should be the decline in farm numbers and employ-

ment during the entire period. This variable was calculated for each

county from the 1959 Census of Acriculture [31, County Table 5] and the

1969 Census of Agriculture [32, County Table 3]. A positive coefficient

sign would be expected as shown ir. Tablr- 2. Declines in the number of







older farm operators would be expected to cause declinesin farm numbers.

Implicit in this is the assumption that farm consolidation occurs rather

than operator replacement.


Manufacturing Labor Supplies

Changes in the supply of labor available to a particular manu-

facturing industry certainly affect employment in that industry.

Shifters of labor supplies would logically cause changes in the number

of firms in the industry which would in turn affect employment. Changes

in the number of manufacturing firms for a given manufacturing industry

were considered exogenous, however, and shifters of manufacturing in-

dustry labor supplies are discussed below as a direct effect on

manufacturing industry employment.


Manufacturing wage opportunity (WWl)k.--Wages in manufacturing in-

dustries other than the industry of present employment (kth) represent

changes in the opportunity cost to employees of remaining in present em-

ployment as a result of changes in wages in other employment alternatives.

Initially, wage increases in other industries would entice employees to

leave their present industry if their skills were transferable. Their

present industry might bid wages upward and regain to some extent but an

overall negative effect would be expected as indicated in Table 3.

Change in opportunity wages between 1960 and 1970 in each county

was determined for construction and manufacturing using data on employee

wages obtained from County Business Patterns [48, 49] and the Census of

Population [29]. It was hypothesized that employees would not be moving

into the agricultural industry because of its low average wage level.

lhe char.o in annual nonagricultural wages between 1959 and 1970 per




75


construction worker in 1960 in each county was used to indicate the wage

opportunity for construction industry employees. A similar measure was

c31culated per manufacturing employee in each county.


Manufacturing employment opportunity (WE. --increases in employ-

ment opportunity in alternative employment industries would be expected

to af f ec t the nuribe r of em p 1 oyee5 i n the 9 iven i ndu s t ry i ri a wa nne r

similar to that of increases in wages in other industries. Changes in

alternative manufacturing employment were calculated for each county

using employment data obtained from the 1960 and 1970 Census of

Population [24, 201 by computing the change in all employment other than

agriculture and industry of present cmployment per worker in industry of

present cfTiployment. Increases in employment opportunity in other manu-

facturing industries should decrease employment in the industry of

present employment as shown in Table 3. Growth in industries that are

complementary in nature would be expected to positively affect employ-

ment in each other.


Number oF nianuacturinq firms (N Changes in the number of

manufacturing firms were used as predetermined variables in the con-

struction and manufacturing models. An increase in the n6mber c firms

in general would be expected to bring about an increase in employment.

Some cases would exist where intrafirm expansion could bring about an

employment increase while firm -.unibers were declining. In general, a

pos i t; ve s i gn s hou I d be expec ted f or th i s coef f t c tent as s hown 1 n Table

3. Data sources for firm numbcr5 were outtltned earlier.














CHAPTER IV

ANALYSIS OF RESULTS



Parameter estimates in equations identical to those presented in

Chapter III were made for agriculture, construction, and the various

manufacturing industries. Equations (3.9.1.1) and (3.9.1.2) were used

for agriculture. Equation (3.9.2.1) was used for construction and equa-

tion (3.9.k.l) was used for the manufacturing industries. All equations

were estimated for each of the three groupings of counties. Counties

were excluded if no employment was reported in both 1960 and 1970. In

this chapter a comparison is made of parameter estimates obtained from

estimating the relationships for each of the three groups. Effects of

the predetermined variables on employment in each industry and farm firm

numbers for agriculture are discussed. General comparisons of the

results obtained for all three groups are made.


Agriculture


Tables 4 through 6 contain the parameter estimates for the two-

equation models used for agriculture. Each table presents three equa-

tions for one of the three groupings of counties. Table 4 presents the

results for all counties, Table 5 the results for the urban counties and

Table 6 the results for the nonurban counties. !n the following dis-

cussion each agricultural equation in the tables is not referred to

separately as the poramete: estimateF are discu'-,ed. Each para.neter and




77
Table 4. Structural form and reduced form coefficients for change in
agricultural employment (El) and number of farm firms (N,),
all counties, 1960 to 1970.
Endogenous var abea

PredtermnedStructural Derived reduced
Preetrmiedform coefficients form coefficients
variales aAgricultural Number of Agricultural




Constant -8.69 14.63 3.03

Education (X 1) -.1476 .2055* .0169
(.2906) (.1073)

CE (X 2) -.0033 -.0005 -.0037
(.0038) (.0014)
PL-566 (x 3) -.0741 .0314 -.o489
(.1003) (.0307)
ACP (X 4) -.5671 --3 8 8 2* -.8770,


FHA (X 5) -.0527 -.-05 96 -.1004
(.0309) (.0105)
Allotment (Y .1577 07 63** .2187
(.0747) (.0268)
Product price (PP 1) 2.4475 7.4334-***,, 8.3083
(3-9030) (1-3790)

Wages (FP,) -.0070 .0031 -.0045
(.0093) (.0030)
Technology (Z 1) 1.9176 -1.0211* 1.1002
(1.4180) (.5400)

Wage opportunity (WWJ) ---1.1679 -.9350
(2.6060)

Employment opportuni ty
(WEI) ---.4192 -.3356
(1.0110)







Table 4 (Continued)

Endogenous variables
Structural Derived reduced
Predetermined form coefficients form coefficients
variables Agricultural Number of Agricultural
employment farm firms employment
(El) ( 1ll)C E(E1)



Number of farms (I1) .8006
(.1251)

Dummy (GRP) 297.4100 29.5200 321.0400
(58.2700) (23.4800)

-2
R2 .82

R -- .81

aComplete variable definitions can be found in Chapter III.

Change in agricultural employment was estimated with two-stage
least squares. Figures in parentheses for this equation are asymptotic
standard errors. Levels of significance are not indicated since they are
approximations.

CChange in number of farm Firms was estimated by ordinary least
squares with figures in parentheses indicating standard errors. Since
this equation is just-identified and contains all predetermined vari-
ables the structural coefficients are identical to the derived reduced
form coefficients.

-:'Significant at 10 percent level.

-'-Significant at 5 percent level.


-''.-"'Significant at 1 percent level.






Table 5. Structural form and reduced form coefficients for change in
agricultural employment (EI) and number of farm firms (I1),
urban counties, 1960 to 1970

Endoqenous variables
Predetermined Structural Derived reduced
Predetermined
form coefficients form coefficients
a
variables
Agricultural Number of Agricultural
employment farm firms employment
(E )b (N )c (El)


Constant


268.80


-78.21


263.30


Education (X1)


CE (X2)


PL-566 (X3)


ACP (X4)


FHA (X5)


Allotment (X6)


Product Price (PPI)


Wages (FP1)


Technology (Z1)


Wage opportunity (WW1)


Employment opportunity
(WEl)

Farm operator age (WA)


.8879
(.8316)

-.0058
(.0104)

.3388
(.2186)

-.8410
(.2107)

-.0682
(.0595)

.1843
(.3383)

-14.4830
(9.5980)

.0343
(.0427)

-1.1216
(3.2520)


.399 1-':
(.1619)

.0018
(.0021)

.0804:-
(.0433)

2482'"
(.0427)

.03 1 '1 :
(.0121)

.0580
(.0688)

4.9352 -
(1.9490)

.0131
(.0088)

-.1356
(.6920)

-.3720
(2.5290)


.1350
(.7820)

1.911 1--':1
(.0965)


.9159


-.0057


-.8584


-.0703


.1884


-14.1364


.0352


-1.2014


-.0261



.0095


.1342







Table 5 (Continued)

Endogenous variables
Structural Derived reduced
Predetermined .
riForm coefficients form coefficients
variables
Agricultural Number of Agricultural
employment farm firms employment
(E1)b (N1)c (El)


Number of farms (NI) .0702
(.2503)

R2 .93
-2
R -- .92


Complete variable definitions can be found in Chapter III.
b
Change in agricultural employment was estimated with two-stage
least squares. Figures in parentheses for this equation are asymptotic
standard errors. Levels of significance are not indicated since they are
approx imat ions.

CChange in number of farm firms was estimated by ordinary least
squares with figures in parentheses indicating standard errors. Since
this equation is just-identified and contains all predetermined variables
the structural coefficients are identical to the derived reduced form
coefficients.

"'Significant at 10 percent level.

"-Significant at 5 percent level.


*;,-,-Significant at 1 percent level.







Table 6. Structural form and reduced Form coefficients for change in
agricultural e7pioyment (EI) and nu'-Iber of farw firml (10.
nonurban counties, 1.960 to 1970

Erdo2enous variables a

Predetermined Sructural Derived rQduced
varlAbles form coefficients form coefficients
Agricultural Number of Agricultural
criployment farm firms C'mployment
(E I )b (N I )c (El)


Constant -47o.6o 58-59 -390-70

Education (X 1) -.3154 .1246 -,1454
(.2515) (.1312)

CE IX .0007 -.0021 -.0022
2 (.0034) (-OC17)

PL-566 (x 3 -.1474 -.Clog -.1623
.0959) (.0481)
ACP (X 4) -.2659 .4153-,,-,' -8324
(.C940) (.0352)

FHA (X 5 .0456 0 7 9 6 C63,0
(.0334) (.0143)
Allotment (x 6) ..0821 0 7 7 1884
(.0595) (.0290)
Product price (PP 4.284o 8 5 8 9 16.OOZ9
0-5910 0-7430)
Wages (FP -.CO58 .0020 -.003C
(-oO73) (.0037)
Technology (Z 1) 7.;C522 -i.4802'- 5.0328
(1.498o) (-7969)

Wage opportunity WW,) .1759 .2400
(4,1390)

Employment opportunity
(Wy 9.09,93* 12.4128
(47030)

Farm operator age (WA) I 4 0 2 7 1-9137
(.01947)




mola,







Table 6 (Continued)

Endoqenous variables
Structural Derived reduced
Predetermined form coefficients form coefficients
variablesa Agricultural Number of Agricultural
employment farm firms employment
(El)b (N1)c (El)


Number of farms (NI) 1.3643
(.1322)

R2 .80

-2
R -- .80

aComplete variable definitions can be found in Chapter III.

b
Change in agricultural employment was estimated with two-stage
least squares. Figures in parentheses for this equation are asymptotic
standard errors. Levels of significance are not indicated since they are
approximations.

Change in number of farm firms was estimated by ordinary least
squares with figures in parentheses indicating standard errors. Since
this equation is just-identified and contains all predetermined vari-
ables the structural coefficients are identical to the derived reduced
form coefficients.

*Significant at 10 percent level.

**-Significant at 5 percent level.


'--'-Significant at 1 percent level.








the comparisons that are made can be found in the three tables. Means

and standard deviations of each variable are given in Appendix B.


Type of Equation

Tables 4 through 6 contain both structural equations and derived

reduced form equations. Each structural equation for changes in

employment was estimated by two-stage least squares. The equation is

over-identified in each model and contains two endogenous variables.

Coefficients of these equations can be interpreted as the direct effect

on changes in agricultural employment of a one-unit change in the pre-

determined variable. Figures in parentheses are asymptotic standard

errors. These can be examined in relation to each coefficient to provide

an approximation of the statistical reliability for each coefficient

using standard normal test procedures.

The equation for changes in the number of farm firms was esti-

mated by ordinary least squares. It is one of the two structural

equations of each two-equation model. The equation contains only one

endogenous variable (change in the number of farm firms) and all of the

predetermined variables in the model.

Each table also contains a derived reduced form equation for

employment changes. Reduced form equations express an endogenous

variable as a function of all exogenous variables in the model. Coeffi-

cients in this equation can be interpreted as the partial derivative of

the endogenous variable with respect to any predetermined variable with

all other predetermined variables held constant at their mean values.

This coefficient indicates the total effect of a change in the

endogenous variable after taking into account the interdependencies

among all current predetermined variables. This coefficient can be









referred to as a multiplier in contrast to a structural coefficient which

indicates only a direct effect. For this particular model the structural

form equation and reduced form equation for changes in the number of farm

firms are identical. The structural equation contains all predetermined

variables in the model expressed as a function of changes in the number

of farms. This is also the reduced form equation of the model by

definition.


Endogenous Variables

Changes in agricultural employment and in the number of farm

firms from 1960 to 1970 represented the two endogenous variables in the

agricultural equations. The trend in agricultural employment during

this time period was generally downward. Although some counties did

experience increases the average change was negative. The urban counties

experienced the smallest average decline per county in agricultural

employment while the nonurban counties had the largest average decline

(Appendix B, Table 18). Urban counties had fewer agricultural employees

and had already experienced larger declines in some earlier time period.

Employment decline in the urban counties averaged slightly less than

one-half the decline in the nonurban group. Average decline for all

counties was approximately twice that of all urban counties.

The number of farm firms per county also declined for each of the

three groups with the magnitude of the declines for each group having the

same ordering as employment declines. Relative differences were not as

large as experienced in employment. Large variation among counties was

indicated for both farm number changes and agricultural employment

changes (Appendix B, Tables 18 and 19).





85


A decline in farm numbers would be expected to cause a decree

in the number of agricultural employees. Fewer farms would cause a

decline in farm operators and decrease employment. Larger and more efi

cent farms should also require fewer employees. This relationship wra

supported without exception by positive coefficients for changes in ith

number of farm firms when this variable was included in the employment

change equation. Standard errors for this coefficient were relatively

small in relation to the coefficient for both the all county group and

nunurban group. The larges efec of Ua decriease in one farm firmon

agricultural employment occurred for the nonurban group of counties %,,hl

the smallest effect was indicated for the urban group. Since the struc

tural form equation for changes in the number of farm firms was estimate

by ordinary least squares the coefficient of determination (R 2) becomes

a valid measure of the proportion of total variation explained by the

predetermined variables in the system. The exogenous variables explaied

80 percent of the total variation in farm numbers in the nonurban count

group, 82 percent using all counties, and 93 percent using urban counts

Examination of corrected R 21s showed very little decline in explanatory

power when the results were adjusted for degrees of freedom.

While explan-ntory power wshigh for all groups a difference

existed among the various groups of counties with respect to the statis-

tical significance of the predetermined variables. Some of the variabe

that were important for one group of counties were not important for

other groups. Multicollinearity was never a serious problems. Very fe

simple correlations among the independent variables exceeded an absolut

value of .5 and in no cases were these coefficients extremely large.
Idntfiaio o tevaiale n hih h smpe oreato








coefficients exceeded .5 will be made in later sections when each inde-

pendent variable is discussed.


Exogenous Shifters

All predetermined variables included in Tables 4 through 6 were

discussed in Chapter III. Each estimated coefficient reflects the effect

of the shifter it represents according to the type of equation in which

the coefficient is found.


Education investments.--Changes in per capital education expendi-

tures did not appear to be a very important variable affecting agricul-

tural employment changes. Negative coefficients were found for all

counties and the nonurban group of counties. The coefficients indicate

that a one dollar increase in per pupil education expenditures was

accompanied by changes in agricultural employment of the magnitude of

the estimate. Changes in per pupil expenditures were smaller for the

two groups having negative coefficients than for the urban group

(Appendix C, Table 20). The urban group indicated a positive effect from

education expenditures. A more skilled work force resulting from higher

education levels would be expected to migrate to urban areas to realize

their employment potential. This migration effect is supported by the

negative coefficients for the nonurban counties. It appears possible

that the migration effect was large enough to support a positive coeffi-

cient for the urban group. Education expenditures appeared most impor-

tant in the nonurban counties upon examination of the standard errors of

each coefficient.

Somewhat different results were obtained for changes in the

number of farm firms. Increases in per capital education expenditures








were significantly associated with increases in farm numbers for both the

urban group of counties and the group consisting of all counties while a

positive but nonsignificant coefficient was experienced in the nonurban

group of counties. The opposite relationship was expected. Increases in

per pupil education expenditures would be expected to enable farm opera-

tors to attain more skills and take advantage of other employment alter-

natives in a given area. This would be particularly true with operators

of smaller and more marginal farms. The strong positive relationship for

the urban counties suggests one possible effect that may be occurring.

Higher educational expenditures in the urban counties coupled with higher

incomes may encourage an increase in the number of small and part-time

farms. This could occur even though the total number of farms might

decline for a given area.

Reduced form coefficient signs and magnitudes indicate the total

effect on employment as a result of education expenditures. The effect

was positive for all counties and for the urban counties group. For the

nonurban group where the direct negative effect on employment was largest

and the positive effect on farm numbers was small the total effect was

negative.

A correlation coefficient of -.50 between changes in per pupil

education expenditures and technology change was noted in the urban group

of counties. This must represent some form of spurious correlation.

Increased technology levels would normally be associated with increases

in educational skills.


Corps of Engineers' investments.--Investments by the Corps of

Engineers demonstrated a negative effect on agricultural employment

change for all groups except the nonurban group. A major portion of









investments by the Corps in the four-state study area was for flood

control. Effective flood control should make more land available for

agricultural use. Expansion of existing farms to benefit from this

should enable more efficient labor saving operations with resultant

declines in agricultural employment. Displacement of existing farms by

expansion would also contribute to employment declines. It appears that

this occurred in the two groups having negative coefficients. Any

employment increase due to new farms in the recipient area was offset by

reduction in employment as the result of farm expansion. The standard

error for the positive coefficient associated with the nonurban group

was quite large. Average investment in the nonurban counties was approx-

imately one-half that in the urban counties.

The effect of a one thousand dollar change in investments on farm

numbers was not significant for any of the three groups. Negative effects

occurred for all counties and the nonurban group. Fewer farms would be

expected if existing farms expand to take advantage of more flood pro-

tected land.

Examination of the geographical pattern of Corps of Engineers'

investments gives further indication as to why positive coefficients

occurred for employment in the nonurban group and for farm number changes

in the urban group. A large proportion of these investments occurred

in the delta area of Mississippi and in an area in east-central Alabama

which are predominately nonurban areas. These areas have traditionally

been areas where large numbers of small farms were in existence. Flood

protection provided by Corps of Engineers' projects could have made new

land available which was suited for mechanized agricultural use. This

undoubtedly created a movement to larger farms through both the effect




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E4YSU3AIR_6UIAYK INGEST_TIME 2017-07-14T22:17:51Z PACKAGE UF00097568_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

THE EFFECT OF RESOURCE INVESTMENT PROGRAMS ON LABOR EMPLOYMENT By JAMES CAREY CATO A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL O:" THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY CF FLORIDA 1573

PAGE 2

ACKNOWLEDGEMENTS The author wishes to express appreciation to the Food and Resource Economics Department, University of Florida, and the Economic Research Service and Soil Conservation Service, United States Department of Agriculture, for making this research possible. Sincere appreciation is acknowledged to Dr. B. R. Eddleman, Chairman of the Supervisory Committee, for his guidance and assistance during the author's graduate program. Gratitude is also expressed to Drs. J. E. Reynolds, K. C. Gibbs, and R. W. Bradbury who served as committee members. Dr. W. W. McPherson also provided valuable comments. Mr. Gene Harris provided valuable programming assistance. Dr. Neil Cook, Economic Research Service, United States Department of Agriculture provided excellent comments and guidance. The author also wishes to express thanks to Mrs. Christine Ward for her valuable assistance in typing the dissertation and to Mrs. Phyllis Childress for her clerical and typing assistance in earlier drafts. Appreciation is also due Miss Wanda Rhea who typed a complete draft of the dissertation. The greatest debt is due the author's wife, Diane, and sons Kyle and Chad, for their unselfisn devotion and sacrifice during his graduar.e program. I i

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS II LIST OF TABLES vi? LIST OF FIGURES }x ABSTRACT x CHAPTER I INTRODUCTION 1 Objectives 3 Review of Literature k CHAPTER I I THEORETICAL FRAMEWORK 12 Mathematical Models of Equilibration Process 12 Changes in Employment 15 Economic Interpretation of Employment Effects. . . 16 Changes in Firm Numbers 19 Economic Interpretation of Firm Number Effects . . 21 Economic Interpretation of Exogenous Changes 2k Factor Supplies , 25 Number of Firm Entrepreneurs 32 Product Price 32 Factor Price 33 Firm Production Possibilities 34 CHAPTER I I I STUDY AREA AND MODEL SPECIFICATION 35 Study Area 35 Selection of Employment Categories. .... 37 General Model Specification 42 Structural Estimation. ..... .... kS Reduced Form Estimation i+5 individual Industry Models 46 Agriculture. , 47 f I i

PAGE 4

62 Page Construction ^9 Manufacturing. 50 Model Estimation Procedure' 52 Measurement of Variables and Empirical Expectations . . 53 Employment 53 Firm Numbers 56 Factor Supplies 56 Education investments (X.) 57 Corps of Eng ineers' natural resource Investments (X^) 58 Soil Conservation Service PL-566 Investments (X_) 60 Agricultural Stabilization and Conservation Service ACP investments (X,) Farmers Home Administration investments (X-) 63 Crop allotment {x/} 64 Product Demand 66 Agricultural product price (FP,) 66 Manufacturing product price (PP.) 6? Factor Price 68 Agriculture wage rate (FP,) 69 Manufacturing wage rate (FPJ ^9 Technology 70 Agricultural technology (Z,) 70 Manufacturing technology (Z.) 71 Farm Operator Suppl ies 72 Agricultural wage opportunity (WW.) 72 Agricultural employment opportunity (WE^ . . 73 Farm operator age (WA) 73 Manufacturing Labor Supplies 7^ Manufacturing wage opportunity (WW ) 7^ Manufacturing employment opportunity (WE^) . . 75 Number of manufacturing firms (N.) 75 CHAPTER IV ANALYSIS OF RESULTS 76 Agriculture 76 Type of Equation ^3 Endogenous Variables 8-4 Exogenous Shifters 86 Education investments 86 Corps of Engineers' investments 87 Small Watershed Program investoients 89 Agricultural Conservation Program investments 9' Farmers Home Administration investments ... 92 Crop allotment 9^ Agricultural product price 95 iv

PAGE 5

Page Agricultural wage rate 96 Agricultural technology 97 Agricultural wage opportunity 98 Agricultural employment opportunity 99 Farm operator age 100 Group differences 100 Construction 100 Dependent Variable 102 Exogenous Shifters 102 Education investments 102 Corps of Engineers' investments 103 Small Watershed Program investments 103 Agricultural Conservation Program investments 104 Farmers Home Administration investments . . . 104 Construction wage rate 105 Construction wage opportunity 105 Construction employment opportunity IO6 Group differences I06 Nondurable Manufacturing I06 Dependent Variables 107 Exogenous Shifters Ill Education investments Ill Corps of Engineers' investments 112 Small Watershed Program investments II3 Agricultural Conservation Program investments 113 Farmers Home Administration investments , . . 11^ Manufacturing product price 1 l4 Manufacturing wage rate 116 Manufacturing technology 116 Manufacturing wage opportunity 118 Manufacturing employment opportunity 1 I9 Number of firms 119 Group differences 120 Durable Manufacturing 121 Dependent Variables 126 Exogenous Shifters 126 Education investments 126 Corps of Engineers' investments. ....... 127 Small Watershed Program investments 127 Agricultural Conservation Program investments 128 Farmers Home Administration investments ... I28 Manufacturing product price 12S Manufacturing wage rate ........... I3I Manufacturing technology. .,.....,,. 132 Manufacturing wage opportunity. ....... 134 Manufacturing employment cpporfjni ty 135 Number of firms 136 Group differences .....,.,, !37

PAGE 6

Page CHAPTER V SUMMARY AND CONCLUSIONS I38 Summary 138 Conclusions 1^1 Limitations 1^6 Need for Further Research 147 APPENDIX A SPECIFICATION OF AREA ADJUSTMENT MODEL 1^9 APPENDIX B MEANS AND STANDARD DEVIATIONS OF VARIABLES 166 BIBLIOGRAPHY 175 BIOGRAPHICAL SKETCH I8O VI

PAGE 7

LIST OF TABLES Table Page 1 Industry identification and employment rankings for the four-state area (I967) ^0 2 Expected effect of changes in predetermined variables on agricultural employment and farm numbers 5^ 3 Expected effect of changes in predetermined variables on manufacturing employment 55 4 Structural form and reduced form coefficients for change in agricultural employment (Ei) and number of farm firms (N,), all counties, 19bO to I97O 77 5 Structural form and reduced form coefficients for change in agricultural employment 79 6 Structural form and reduced form coefficients for change in agricultural employment (E^) and number of farm firms (N,), nonurban counties, 19dO to 1970. ... 81 7 Regression equations for construction employment change (E2) for all counties, urban counties, and nonurban counties, I96O to I97O 101 8 Regression equations for textile mill product and other fabricated textile products employment change (E3) for all counties, urban counties, and nonurban counties, I96O to I97O IO8 9 Regression equations for food and kindred products employment change ( E/^) for all counties, urban counties, and nonurban counties, I96O to 1970 109 10 Regression equations for nondurable manufacturing employment change {^\^q) for all counties, urban counties, and nonurban counties, I96O to 197C 110 il Regression equations for transportation products employment change (Er) for all counties, urban counties, and nonurban counties, I96O tc 1970 122 vii

PAGE 8

Tab 1 e Page 12 Regression equations for furniture and lumber and wood products employment cinange (E^) for all counties, urban counties, nonurban counties, i960 to 1970 123 13 Regression equations for electrical equipment products employment change (Ey) for all counties, urban counties, and nonurban counties, I96O to 1970 124 ]k Regression equations for durable manufacturing employment (Eq) for all counties, urban counties, and nonurban counties, I96O to I97O 125 15 Definition and interpretation of terms in equation (11) 158 16 Definition and interpretation of terms in equation (1*2) 162 17 Approximation of terms to represent a change in residual return 165 18 Means and standard deviations of employment change variables for each county group of observations .... 167 19 Means and standard deviations of firm number change variables for each county group of observations .... 168 20 Means and standard deviations of factor supply variables for each county group of observations .... 169 21 Means and standard deviations of product price change variables for each county group of observations 170 22 Means and standard deviations of wage change variables for each county group of observations .... 171 23 Means and standard deviations for technology change variables for each county group of observations .... 172 2k Means and standard deviations for wage opportunity change variables for each county group of observations 173 25 Means and standard deviations for employment opportunity change variables for each county group of observations 1 74 • Vl I <

PAGE 9

LIST OF FIGURES Figure Page 1 Illustration of the equilibration process , resulting from a change In supply of the n factor 26 2 Illustration of the equilibration process in the n^*^ factor market resulting from a change in supply of the n^" factor 27 3 Illustration of the equilibration process in the firm entrepreneur market resulting from a change in supply of the n^" factor 28 k Illustration of the equilibration process in the labor market resulting from a change in supply of the n^h factor 29 5 Grouping of counties for the four-state study area 38 IX

PAGE 10

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECTS OF RESOURCE INVESTMENT PROGRAMS ON LABOR EMPLOYMENT By James Carey Cato December, 1973 Chairman: Dr. B. R. Eddleman Major Department: Food and Resource Economics This study examined the importance of investments in human and natural resources along with several other variables in explaining employment changes among counties comprising the four-state region of Mississippi, Alabama, Georgia, and Florida over the time period I960 to 1970. Counties in the study area were delineated into urban and nonurban counties according to their human and natural resource endowments. A theoretical economic model was developed that explains changes in employment and firm numbers brought about by exogenous shifts in the supplies of resources, demand for products, supplies of other factors, firm production possibilities, and shifters of the number of firms in an industry. Empirical analysis was undertaken to determine the importance of each exogenous shifter on industry employment changes. The industries studied included agriculture, construction, textile mill products and other fabricated textile products, food and kindred products, transportation, furniture and fixtures, lumber and wood products, electrical equipment, durable, and nondurable products manufacturing. Changes In factor supplies inciudeo' changes in per pupii

PAGE 11

education expenditures. Corps of Engineers investments, Soil Conservation Service investments in the Public Law 566 Small Watershed Program, Agricultural Stabilization and Conservation Service payments In the Agricultural Conservation Program and loans and grants for community water and sewer systems made by the Farmers Home Administration. Changes in county product price indexes, allotment reductions, farm operator age, wage rates, technology indexes, number of firms, and alternative wage and employment opportunities represented the other types of exogenous shifters. General equations for empirical analysis were specified for each industry. A two-equation model was used for agriculture with the number of farms considered endogenous. Single equation models were specified for the remaining Industries. Changes in the number of firms were considered exogenous In these models. Two-stage and ordinary least squares were used as empirical estimation procedures. Results indicate that employment changes eminating from changes in the exogenous shifters differed quite substantially among industries and the county groups considered. Most employment effects were generally consistent with expectations. Logical explanations were normally apparent for employment effects that differed from initial expectations. It was observed that some Investments Influenced employment in a particular industry and yet were not Important In other Industries. Location of industries was also important. Effects differed between the urban and nonurban counties for some Industries. The most satisfactory results were obtained for agricultural employment and farm number changes . XI

PAGE 12

Results indicated tiiat any attempt to stimulate employment in an area with investments in human and natural resource should take into consideration not only the agricultural, urban, and nonurban characteristics of the area, but the type of industry employment most evident in the area. XI I

PAGE 13

CHAPTER I INTRODUCTION Investments in natural resources usually are for the expressed purposes of conserving, developing, or managing the nation's supply of soil, water, timber, mineral, and marine resources. Many public investment programs in natural resources such as those associated with the Tennessee Valley Authority (TVA) and the Small Watershed Development Program administered by the USDA's Soil Conservation Service contained explicit development objectives. These objectives were concerned with alleviating depressed regional economic conditions or improving the incomes of specific groups of people. . Senate Document 97 [l]f issued in 1962, also made explicit a national policy of natural resource investments for purposes of increasing income and employment in particular regions. The Appalachian Regional Development Act of 19^5 L2] provided for the construction of water resource projects to stimulate economic growth of the region. Guidelines concerning principles and standards for the planning of water and related land resource use, issued for review by the Water Resources Council [3] in 1971, gave added emphasis to the role of water This program, created by the Watershed Protection and Fiood Prevention Act of 195^t-j with its Aniendments, is coiTimoniy referred to as Pub i ic Law 566.

PAGE 14

2 resource investments In the development of a regional economy. This orientation in policy has given added emphasis to natural resource development programs and projects as instruments for dealing with regional economic problems. Many other programs have evolved that focus on goals of community improvement by concentrating on such areas as increasing local employment and income, increasing public revenues, and improving the quality of the environment. Local employment and income of an area depend on many factors other than investments in natural resources. Any explanation of employment and income changes occurring within a region requires analysis of the many variables which interact to determine these changes. Identification and measurement of these complicated interdependencies are necessary in order to assess previous or prospective effects of the various programs in influencing the level of employment and incomeChanges in investment levels that shift the supplies of critical resources often occur concurrently with changes in the demands for products, supplies of other resources, firm production possibilities, and the number of firms. An important element is the consideration of how equilibration in product and factor markets is affected by programs designed to change the supplies of resources and, in turn, how changes in product and factor prices affect the level of output, resource employment and income within the recipient region. Since similar investments in heterogenous regions might have different effects on employment, dissimilarities among regions need to be considered. These differences could exist in the form of differing resource base or differing industrial structure. Planners and decision-makers need information on the effectiveness of the small watershed and other natural resource projects in

PAGE 15

fostering employment and income growth. Knowledge concerning the relationships between natural resource investment and the other important stimuli on changes in employment and income in a regional economy is vital. Bacl< [k] has pointed out that assessment of the role of natural resource investments in stimulating growth of a regional economy will be a difficult task without information of this type. Competition for the federal dollar between supporters of various programs is very keen. Areas faced with employment and income problems should utilize scarce resources in programs that would result in the greatest rates of change in employment and income. Thus, information on the effectiveness of natural resource investment programs in meeting specified objectives is critical for future program planning. The results of this evaluation will indicate that either small watershed and other natural resource projects satisfy income and employment objectives or that program reform is necessary if objectives continue to include income and employment goals. This research, though conducted for a sub-region of the Southeast, should provide answers applicable to the entire Southeastern region. It is sufficiently broad to be indicative of the national Small Watershed Program and other resource development programs. Object ives The general objective of this study is to evaluate the effectiveness of the S.noll Watershed Program and other natural resource investments in accelerating employment growth in local recipient areas. The study includes the four -state region of Mississippi, Alabama, Georgia,

PAGE 16

and Florida over the time period I960 to 1970. More specifically, the objectives are to: 1. Develop an economic model to explain changes in employment and in the number of firms brought about by exogenous shifts in the supplies of resources, demand for products, supplies of other factors, firm production possibilities, and shifters of the number of firms in an industry. 2. Empirically apply the model to selected industries in order to determine the importance of changes in the above factors on changes in employment and in the number of farm firms. Review of Literature Previous studies concerned with the estimation and explanation of the effects of investments in natural resources on employment, income, and output are quite varied in purpose, objective and scope. All previous work can be grouped into three basic categories consisting of (I) case studies of individual projects and their impacts on local areas, (2) studies proposing various procedures that could be used in evaluating project effects, and (3) studies that attempt to determine the effect of water resource investments over large multi-county or multi-state areas. Some of these studies are reviewed in this section to provide a cross section of previous work. The first group consists o^ case studies of various small watershed projects and the impact of the investment program on the local economy and/or the sectors they were intended to benefit. Jansma and Back [5] estimated the local secondary effects of tho construction

PAGE 17

of watershed project structures for upstream flood protection in Roger Mills County, Oklahoma. Through the use of an input-output analysis they estimated income multipliers which were used to determine the effect on the county's economy of increases in agricultural and recreational income as a result of the watershed program. They found that for each $100,000 in gross receipts to farmers in the county, there was an estimated net (disposable) income to farmers of $26,867. This $100,000 increase also generated $77,845 in gross receipts to other sectors of the local economy and a net income of $16,457 to these sectors. The local gross receipts multiplier of farm income was I.78 and the local net income multiplier was 1.62. Net income receipts from gross recreational receipts were about one-half those of gross agricultural receipts. Gray and Trock [6] in an evaluation of the Green Creek Watershed Project in Texas used traditional benefit-cost analysis to compare the actual benefit-cost ratio derived from post-project evaluation with the ratio estimated in the watershed work plan. They found that the project failed to provide benefits in excess of costs over the first eight years of the project. The ratio of benefits to costs determined by the study was .919:1 as compared with an estimate of 1 •49: 1 made in the watershed work plan. Input-output analysis was used by Kasal [7] to estimate the local economic impacts in a four-county area as a result of five Colorado watershed projects. Local net income arising from project expenditures as well as blenefits to the project users were estimated. Kasal compared the original benefit-cost ratios for the projects with those derived through the use of multipliers from the input-output

PAGE 18

analysis and found that the local economic impact of all five projects 2 exceeded that estimated in the watershed work plan. Clonts is using a similar analysis to measure the economic impact of the Cheaha Creek Watershed on the local economy of Talledega County, Alabama. Cato 3 and Eddleman are evaluating the secondary impacts from the Taylor Creek Watershed Project in a six-county area of South Florida. Work in this study has been directed toward estimating the net increases between the local impact area and the rest of the US., and among groups within the local area. Several other studies not mentioned were concerned mainly with benefits to primary project users, land use, and increases in farmland values. The second category of studies has been concerned primarily with the introduction of suggested methodology for use in the evaluation of water resource investments. The first of these methods as suggested by Eidman [8] has had one empirical application. Eidman's presentation described the linkages or interdependence of the various sectors and subsectors of Southwestern Oklahoma with the use of a simple fiveequation economic model that employed economic base multipliers and regression multipliers. This model was designed to explain employment and income changes as a result of resource investments. Mazuera [9] used the model to determine the secondary impact of using water in the Sugar Creek Watershed floodwater prevention structures for irrigation development and found that the project generated secondary effects amounting to 2 H. R. Clonts, personal communication. 3 Work currently in progress.

PAGE 19

less than a 1 percent change in both population and total income In the watershed project area. The work of Bromley et al . [lO] attempts to outline the role of economic logic and method in analyzing the consequences of water resource investments. Their work does not offer a concrete method of project evaluation but rather is concerned with the many questions raised in such an undertaking. A later study by Gibbs and Loehman [ll] deals with the evaluation of resource investment projects in terms of multiple public goals. This study offers a model that might be used to predict regional economic effects of resource investments which the authors claim presents a more realistic view of the regional economy than input-output or multiplier analysis. The last category of studies has also been concerned with the estimation of the impacts of water resource development at the local level, but includes studies concerned with a much broader geographical area than the local project area. Many of these studies used county observations in various forms of econometric models to examine resource investment effects while others examined regional differences in the effects of resource investments. Haveman and Krutilla [12] looked at both the national and regional effects of twelve types of water resource projects with respect to their influence on various occupational and industrial categories. Their analysis based on input-output models constituted an effort to sort out the demands which public expenditures for water resource developments imposed on the economy. Water availability in relation to regional economic growth was assessed by Howe L13] vvho determined that water deficit areas did net experience drawbacks to economic growth and that v/ater surplus areas

PAGE 20

were not guaranteed rapid growth. Howe's study suggested that water ' "*"" resource developments are likely to be poor tools for accelerating regional economic growth if markets, resources, and other factors considered vital to development are lacking. Howe did not consider the effect of water resource development on small regions. Wiebe[l4] attempted to evaluate the effectiveness of water resource investment projects in alleviating regionally depressed economic conditions. This study of the Tennessee River Watershed consisting of 125 counties In parts of seven states suggested that (1) residents in counties close to water resource investment projects enjoyed a greater per capita income in the long run than did those living farther away, (2) investments in water resources were in the long run associated with increases in employment in counties removed from the project site while nearer counties v;ere associated with long run decreases in employment, and (3) investments were not associated with an increase in the level of living for people in low income and less educated groups living near investment areas as compared to similar groups living in areas removed from the investment site. Another analysis by Cox et al . [ 1 5^ was specifically designed to assess broad based economic growth emanating from multipurpose projects by the application of multiple regression analyses to many socioeconomic indicators. Counties in the 13 Northeastern states of the United States in which large v.-ater resource development projects had been constructed between 19^8 and 1958 were examined in I960 to determine if changes had occurred as a result of the projects. They concluded that there was no relationship betv;een project size and economic growth and that the selection of project sites v;as biased toward urban areas wliere there is

PAGE 21

a greater a priori likelihood of economic growth. They concluded that it was dubious whether water resource projects served as a stimulus to economic growth for the area studied. An index of economic growth based on numerous income and employment variables were used to measure economic growth. k Boxley and Harmon have recently worked on a study to determine the relationship between Public Law 566 watershed investments and economic growth in the Southeastern United States since 1959Economic growth was measured as income changes. They have attempted to use a modified form of the shift-share analytical technique for the period I959-I968. Tentative conclusions are that (1) there was no measurable relationship between watershed investment and economic growth, and (2) the selection of watershed sites for development appeared random when measured in terms of rates and types of economic growth underway in the study area. They have expressed limitations, however, as to the appropriateness of the statistical techniques used and thus offer these conclusions in a qualified manner. They do point out that other elements of capital infrastructure are probably necessary for economic growth to occur. Cato and Eddleman are also concerned with the relation of changes in income over ths last two decades to the level of investment in natural resources for the same time period. Counties of a nine-state region of the Southeastern United States have been delineated into four groups on the basis of their natural and human resource endowments Robert F. Eoxle/ and Marie Harmon, personal communication. V/ork currently in progress.

PAGE 22

10 and level of economic activityMultiple regression and correlation analysis have been used to examine the effect of the level of natural resource investment on changes in various income measures for the four groups. The various types of water resource investments being considered that are of interest to this study include water projects of TVA, Corps of Engineers, and investments under the Small Watershed and Flood Prevention Watershed Programs of the Soil Conservation Service. Tentative results have indicated that with the exception of some investments by TVA and the Corps of Engineers, little effect is felt in local income changes as a result of these water development projects. Conflicting patterns have emerged as to the type of investment recipient area that realizes the greatest Impact. This survey of the literature concerning the effect of water resource investment on economic growth leads to two observations. Either water resource investments are poor tools for stimulating economic growth, or the methodology for measuring these effects falls far short of accomplishing the goal for which it was designed. Many of the studies cited have failed to consider the importance of interdependencies among ^ other important factors within a region which affect employment and income changes. Changes in investment levels that shift the supplies of critical factors, i.e., investments in water resources, often occur concurrently with changes in the demands for products, supplies of other resources, firm production possibilities, and shifters of the supply of firms. Equilibration in product and factor markets is affected by programs designed to change the supplies and/or productivity of resources and this in turn causes changes in product and factor markets which affect the level of employment and income within the recipient region.

PAGE 23

11 A general approach that could be used in considering these additional changes has been suggested by Tolley and Schrimper [ 16] and Schrimper [l?]This approach simultaneously considers aggregate and micro adjustments in product and factor markets. Application of a variant of the general model was performed by Schrimper [l8] to determine the extent to which changes in various exogenous factors suggested by the general model explained changes in the number of farms between 195^ and 1959 for six comparable groups of farms among states as well as among counties within North Carolina and Illinois. Eddleman and Cato [l9] In a current study are using the same variant of the general model as Schrimper to examine factors affecting differential rates of change in the number of farms among counties in Florida for the time period 1959-1969. Eddleman [20] has also proposed the use of another variant of the general model to analyze the effects of investments in resource development programs on regional employment.

PAGE 24

CHAPTER II THEORETICAL FRAMEWORK All previous work using variants of the general model developed by Tolley and Schrimper [16] has been concerned with measuring the rate of change in employment or farm numbers. That is, employment and farm number changes as well as changes in the exogenous variables were measured as percentage changes. The two-equation model developed in this study is concerned with explaining absolute changes. The first equation of the model explains changes in employment as a function of exogenous changes in the prices of products, prices of factors having perfectly elastic supplies, shifters of the supply of factors assumed to have other than perfectly elastic supply functions for the region, shifters of firm production possibilities, and changes in the number of firms. The second equation of the model explains changes in the number of firms as a function of these same exogenous variables and exogenous shifters of firm supply functions in each industry. Mathefnatical Models of Equilibration Process Changes in the production of products and utilization of resources in a region can be observed at several levels of aggregation with each level providing a different insight about the impact of changes, This dynamic process centers around three focal points. At the firm level, product: supplies and factor demands can be influenced by changes 1 '

PAGE 25

13 in production possibilities or product and factor prices. The second area concerns the aggregate effects of the micro adjustments reflected in firms' product supplies and factor demands. Interaction of variation in the number of firms with changes initiated at the micro or macro level represents the third area. Variation in product demands and factor supplies emanating from the marl
PAGE 26

changes in the number of firms which lead to additional changes at the aggregate level involving still further price adjustments. An equilibrium would exist when all product and factor prices and the number of firms in the industry are consistent with their total demand and supplies. For small areas or regions such as a county or group of counties, the majority of product and factor prices might be realistically assumed fixed since regional adjustments within the area would not likely have a significant influence on the total market for these commodities. Feedback effects associated with adjustments would operate primarily within factor markets since nearly all product prices, with the exception of those products produced for local markets, would be mostly exogenous to an individual region unless it accounted for a significant proportion of the total national production of a commodity. Governmental price support activities could also contribute to fixing some agricultural product prices faced by farm firms in any given area. The probable existence of fixed product prices in the analysis of regional adjustment or equilibration allows one to assume that the product demands and the supplies of all but the critical factors are perfectly elastic for a region in which firms in k different types of industries exist and produce m different kinds of products with n-m different kinds of factors. Firms within each of the k industry groups are assumed to have similar production possibilities and be operated by basically the same type of entrepreneur. The remaining part of this chapter is concerned with the discussion of both changes in employment and firm numbers. First, an equation explaining the effects of the various shifters on employment is discussed. Then, a similar discussion of changes In firm numbers is outlined and finally, a discussion of both equations in a simultaneous context Is given.

PAGE 27

15 Changes in Employment Using the assumptions discussed above, the change in demand for labor (employment) in a given industry, k, can be expressed as (2.1) dQ = z 'ikn U/ dX. m + Z J = 1 '^Lkj ^kn (j) %• dP. n-l + E i = m+1 '^Lki " \kn a; "^nki dP. V + i: h = 1 '^Lkh " ^Lkn \a) ^nkh dZ. where ^Lk Lkn ii) %^ dN. Lkn = change in quantity of labsr demanded in the k industry associated with a one-unit change in th price of the n factor. th A = difference in total quantity of the n factor demanded and supplied at a price one-unit above the equilibrium price for tthe n factor. th Sr = change in total quantity suupplied of the n factor associated v/ith a one-unit change in the f exogenous shifter for the n factor supply. The complete model from which equation (l. 1) is derived and a detailed treatment of the derivation and equilibration process associated with the exogenous shifters is presented in Appendix A. Table 15, Appendix A, gives a mathematical definition of the terms in the equation.

PAGE 28

16 d = change in quantity of labor demanded in the k industry associated with a one-unit change in t*h price of the j product. d , . = change in quantity of the n factor demanded in nkj the k industry associated with a one-unit change in price of the j product. d, , . = chanqe in quantity of labor demanded in the k Lki ^ industry associated with a one-unit change in price of the i factor. d , . = difference in total quantity of the n factor nki supplied and demanded resulting from a one-unit change in price of the i factor. d, , , = chanqe in quantity of labor demanded in the k Lkh ^ industry associated with a one-unit change in the h exogenous shifter of firm production poss ibi 1 i ties. d ,, = change In quantity of the n factor demanded in nkh ^ the k industry associated with a one-unit change t"h in the h exogenous shifter of firm production poss ib i 1 i t ies . q° = quantity of labor demanded by a firm in the k i ndustry. th q°, = quantity of the n factor demanded by a firm in ^nk ^ ' *.u I th . , . the k inaustry. Ec onomic Interpretation of Employment Effects Any discussion of the equilibration process on the demand for labor associated with the exogenous shifters in the model must necessarily

PAGE 29

17 be preceded by a discussion of the economic interpretation of the various types of terms found In equation (2.1). Two terms appear in each coefficient of the equation. The first of these is ^kn " \-br which represents the change in the total quantity of labor demanded in f'h f h the k industry associated with a one-unit change in price of the n f-L. factor. N° represents the total number of firms in the k industry, while— rrrepresents the change in quantity of labor demanded by a firm in the k*" industry associated with a one-unit change in price of the n factor. The second term is r bq , bS k = 1 n n which represents the amount by which the total demand for the n factor would be less than the supply at a price one-unit above its equilibrium r &qn|^ level. The portion z N: — tttrepresents the number of firms in the k = 1 •< bP^ k industry times the change in the quantity of the n factor demanded by a firm in the k industry associated with a one-unit change in price th ^^n of the n factor summed over all industries. The portion -.p represents th " the total change in the quantity of the n factor supplied in the area associated with a one-unit change in price of the n factor. The absolute value of A (or -A) represents the increase in demand or decrease in supply for this factor consistent with a one-unit change in price of the factor. The negative reciprocal can then be interpreted as the approximate change in price for each unit increase in demand or decrease in supply of the n factor resulting from exogenous shifters in the model.

PAGE 30

18 The magnitude of this reciprocal would ultimately depend on the elasticities of the demand and supply functions involved since the elasticities would determine the magnitude of the quantity change associated with a given price change. Assuming constant elasticities within the relevant rangeY— ) would represent the decrease in price for each unit decrease in the demand or increase in the supply of this factor. Three coefficients contain other terms that are very similar. The terms d, ... d, ,., and d, ., represent the direct change in quantity of Lkj'Lki' Lkh'^ ^ ^ ' labor demanded In the k Industry associated with a one-unit change In the j product price, i factor price, or h shifter of firm production possibilities, respectively. Similarly, d , . and d represent the change In quantity of the n factor demanded In the k industry associated with a one-unit change in price of the j product and in the fh h shifter of firm production possibilities, respectively. The term, d ... represents the difference in total quantity of the n factor nki' "^ supplied and demanded associated with a one-unit change in price of the I factor. The mathematical construction of these terms is similar to that discussed for b, , with the appropriate notatlonal changes made as Lkn shown in Table 15, Appendix A. interpretation of these terms becomes even more meaningful when discussed simultaneously. For example, f-^j is the approximate change in price of the n factor associated with each unit change in total demand or supply of the n factor, and d , . is the change in total quantity of n K J the n factor demanded in the k Industry associated with a one-unit change in price of the j product. Their product, f-^j d ,., represents the total price change for the n factor associated with a one-unit th change in price of the j product. Final multiplication of this term by

PAGE 31

19 ''ikn' ^^'"^^ represents the change in quantity of labor demanded in the k industry associated with a one-unit change in price of the n'*^ factor, gives b^^^^ ^J d^^., which may be interpreted as the total change in the quantity of labor demanded in the k*^ industry resulting from a one-unit change in the j product price operating through the n^^ factor market. This might be referred to as the indirect effect on labor demanded in the k industry resulting from an exogenous change in the j ^^ product price. The direct effect of this price term would be indicated by d , and would Lkj complete the coefficient of dP All other terms in each coefficient would be interpreted in a similar manner. 1?h Changes in Firm Numbers Assuming only one critical factor, the change in the number of th "3 firms in the k industry can be expressed as"^ (2.2) dN. = I " f = 1 (i) ^n W Sg dX, s + I d = 1 m + Z J = 1 \d^\Ai) '6 a + a, kj kn (i) "j dW dP. J n-1 I i = m+1 \i \n (?) h dP. V + ^ h = 1 c, . + a, kh kn (i) -h M. dZ. The complete model from which equation (2.2) is derived and a detailed treatment of the derivation and equilibrium process associated with the exogenous shifters is presented in Appendix A. Table 16 Appendix A.gives a mathematical definition of the terms in the equation

PAGE 32

where 20 a, = change in the number of firms in the k industry associated with a change in residual returns brought about by a one-unit change in the price of the n factor. B = excess demand for the n factor associated with a one-unit change in price of the n factor. S, = change in total quantity supplied of the n factor associated with a one-unit change in the f exogenous shifter of the n factor supplies. \, , = change in the number of firms in the k industry f h associated with a one-unit change in the d exogenous shifter of firm entrepreneur supply. D, = change in total demand for the n factor d associated with a one-unit change in the d exogenous shifter of firm entrepreneur supply. a, . = change in the number of firms in the k industry associated with a change in residual returns brought about by a one-unit change in price of the j product. U. = chanqe in demand for the n factor associated J with a one-unit change in price of the j product, a, . = chanqe in the number of firms in the k industry ki ^ associated with a change in residual returns brought about by a one-unit change in price of 4-u -th ^ , the I factor.

PAGE 33

21 L. = difference in total quantity of the n factor that would be demanded and supplied resulting from a one-unit change in price of the i factor, c,, = change in the number of firms in the k industry associated with a change in residual returns brought about by a one-unit change in the h exogenous shifter of firm production possibilities, M, = change in demand for the n factor associated with a oneunit change in the h exogenous shifter of firm production possibilities. Economic Interpretation of Firm Number Effects Interpretation of the terms in equation (2.2) must also begin with a discussion of those terms appearing in each coefficient and those whose definitions are qu i te s imi lar . Schrimper [l8] provides a detailed discussion of coefficients from a somewhat similar equation expressed In terms of elasticities and percentage changes. Equation (2.2) represents the sum of effects associated with shifters of factor supplies, shifters of firm entrepreneur supply, product demands, factor supplies, and shifters of firm production possibilities on the number of firms. Each of these shifters operates through either a direct shift in the number of firms or through changes in firm residual returns resulting indirectly from these changes and from changes in use of the critical factor. Two terms appear in each coefficient of the equation. The first of these is r bNj^ IbH ^n = ^ , "FT
PAGE 34

22 as derived in Appendix A and shown in Table 16. This term represents the f h change in the number of firms in the k industry associated with change in residual return brought about by a one-unit change in price of the n factor. The second term is defined as / r bq , 5S \ r n f TiiO nk n\ ^ B = Z N. — r^ -r^E q , a, , , k bP bP , , nk kn \k = 1 n n/ k = 1 which represents the excess demand for the n factor associated with a onef h unit change in price of the n factor. The first portion of B is identical to A as defined for equation (2.1). Hhe second portion takes into account some of the feedback effects on the quantity of the n factor demanded resulting from changes in the ntmber of firms that are associated with a change in residual return broi!ig.ht about by a one-unit price change for the n factor. The reciprocal of this term, f— j , can then be interpreted in the same manner as (-7) . The negative reciprocal, (— • j , can be interpreted as the approximate change in price of the n factor for each unit decrease in total demand or increase in supply of the n factor after adjustments resulting from exogenous shifters in the model. The term X , , represents the change in the number of firms associated with a one-unit change in the d shifter ©f firm entrepreneur supply. Three additional terms that are very similar in nature are a, ., a,., and c, , . These terms are similar in mathemaitical construction and ki kh interpretation to a, . They are interpreted as the change in the number of firms associated with a change in residual retuurn brought about by a one-unit change in j product price, i factor price, and h shifter of firm production possibilities, respectively. Each of these terms is treated mathematically in Appendix A and Table 16. th

PAGE 35

23 The term D , may be interpreted as the change in demand for the n factor associated with a one-unit change in the d shifter of firm entrepreneur supply. Two terms that are similar to D, in that they represent demand and supply relationships for Q, under alternative price changes are U, and L.. Price changes in the j product are considered in U.. The first part of U., I N° ^p , represents the change in demand for the I I 1 K Or. -" k = I J n factor associated with a one-unit change in price of the j product. r The second part, E '^°\,^\.-* takes into account the feedback effects on changes in the number of firms of a change in j product price as It Is felt through changes in the residual return. Both parts of this term would have positive signs. The first part of L., -r-r z N° . " | , represents the difference In supply and demand for the n factor at its Initial equilibrium price after a one-unit increase in price of the i factor, assuming no change in firm numbers. The second part, r E q i,3i,., takes into account the effects on demand for the n factor ki n K K I = I resulting from changes in the number of firms brought about by a or.e-unit change in P. as felt through a change in residual return. Both these terms would tend to operate in the same direction. The first part of L. differs from the first part of U. since factor supplies were assumed to depend on possibly more than one factor price but be Independent of product prices. Existing relationships between the n factor supply and changes in i factor prices would regulate this movement to some degree.

PAGE 36

24 The last term, M, , is also composed of two parts. The first h part, E N° — r-=— , represents the change in quantity demanded of the k = 1 '' °^h n factor associated with a one-unit change in the h exogenous shifter of firm production possibilities. The second part takes into account the change in the number of firms as it affects the n factor demand. This feedback effect is reflected through residual return changes resulting from a one unit change in the h shifter of firm production possibilities. Changes in the h shifter could be either output increasing and/or input decreasing as explained and footnoted in Appendix A and Table 16. Again, as in equation (2.1), the product of each term in the coefficients gives total meaning to each coefficient. For example, in the coefficient of dX,, the product of ( — j S, would be interpreted as the total price change for the n factor associated with a one-unit change in X,, since S, is the change in quantity of the n factor supplied as a result of a one-unit change in X, and (— j is the n factor's price change for each unit change in the quantity. Further multiplication by a, , which represents the change in the number of firms associated kn with a change in residual return brought about by a one-unit change in P , would then give the total change in the number of firms associated n' ^ with a one-unit change in X,. The multiplicative effects of terms in each of the other coefficients would be interpreted similarly. Economic interpretation of Exogenous Changes Changes in the demand for labor^'^re expressed as a function of changes in shifters of factor supplies, product prices, factor prices, firm production possibilities, and the number of firms. Changes in the

PAGE 37

25 number of firms were then expressed as a function of the same shifters of factor supplies, product prices, factor prices, firm production possibilities and shifters of firm entrepreneur supply. Each equation, (2.1) and (2.2), was derived from the same basic model as shown in Appendix A. Considered simultaneously, these equations then explain the changes occurring in an area economy as the result of changes in the exogenous variables. These processes are discussed separately in the following sections. Factor Suppi ies Changes in factor supplies in each equation are represented by dX-. All changes in factqr supplies occur in the supply of the factor assumed to be other than perfectly elastic. Equilibration resulting from an exogenous shift in factor supplies is demonstrated graphically in Figures 1 through k. Figure 1 demonstrates the entire process while Figures 2, 3, and k outline In more detailed form the actual changes occurring in three of the five segments of Figure 1, The effect of a change in supply of the n factor is indicated in Figures 1 (b) and 2 as the shift from S to S' with the resultant i nn n crease in Q to Q and decrease from P to P'. Shown in Fiqure 2 is the n n n n ^ total change in quantity of the n factor indicated by 5S^ ' l\^^n r— — dX, and the total price change as indicated by ( — r I ttt" dX,. 0X„ f '^ ^ V A/ bXf Since f — j represents the change in P associated with a one unit change in Q , i-Tj would be negative as indicated since Q increases and the deth mand curve for the n factor is negatively sloped. This decrease in P n would then cause an increase in residual returns to firms in the k' industry as reflected in the shi<'t from R to rT in Figures 1(d) and 3. K K

PAGE 38

26 p. 'jk Q., !k J k J k J k 1 1 me n n n n time \

PAGE 39

27 f r 1 X ^S f 1 (j) %V\ Figure 2. Illustration of the equilibration process in the n factor market resulting from a change in supply of the n^h factor. th

PAGE 40

28 (i) w'/'f 'kn (-i) bS [ "bX, dXf Figure 3. Illustration of the equilibration process in the firm entrepreneur market resulting from a change in supply of the n^*^ factor.

PAGE 41

29 ^Lk 1 1 me "k bP \ A/ ^nk [ Figure h IllustraHon of the equilibration process in the l_abor market resulting from a change in supply of the n factor.

PAGE 42

30 The effect of this shift would be the movement from N, to N,' as k k indicated by a, (-r ) tt;— dX, and the movement from 11, to n' as i n' kn \ B / bXr f k k 1 ^^n dicated by /•^JTT^r' ^X, in Figure 3. Resulting from an increase in the number of firms would be an increase in the supply of the j product by the k industry, S.. to SI. with the new quantity measured as 0.1. as J K J K J K shown in Figure 1(a). Accompanying this increase would be an increase in demand for the i factor shown in Figure 1(c) as the shift from D., ^ ik to D". , with Q,"., indicating the new quantity demanded. The final initial I k ik ^ ^ ' result would be an increase in the demand for labor as shown by the shift from D, , to D'' in Figures 1(e) and k. This total shift comes from two Lk Lk ^ sources. The first is a direct result of the shift in Xas shown by bq. , 1 \ ^^ N? —— (TlTV^ dX^ and indicated by the movement from Q., , to Q,'. in k OP \ Ay oX, f Lk Lk n T Figure k. The remainder of the increase in labor demand comes from an increase in the number of firms resulting from increases in residual returns through lower prices of P . This increase is given by q j^ dN, as the movement from Q' to Q."'' in Figure k. Entrance of new firms to the k industry would ultimately increase the demand for the critical factor as indicated by the shift from D to D® in Figures 1(b) and 2. The resultant shift in Q" to Q.® is inn n ^ n n dicated in Fiqure 2 by q°, dN, . Resulting from this would be a bidding ^ ' nk k upward of P" to P^ as shown by (t) "^ l'^'^i^ '" f"''gure 2. Since ("^ j represents the decrease in price for each unit decrease in demand, an increase in demand, i.e., the movement from D to D , would cause a ' n n price increase as shown. Increases in P would then result in the exit n of some firms from tfie k ' industry through a decrease in the residual

PAGE 43

31 return to these industries. This decrease in return is shown by the shift from R' to R. in Figures 1(d) and 3The effect of a price increase through f— ) then causes the movement from 11' to IT, as indicated in Figure 3. A similar response in the movement from N' to N, is also shown. Resultant shifts would also occur in ST, and DT, . k jk I k e e Equilibrium positions would be indicated by S.. and D., with equilibrium quantities indicated by Q... and Q.. in Figures 1(a) and 1(c), respectively. The final movement requiring discussion concerns that in the labor market. Changes in P resulting from increases in demand for the ^ n ^ n factor also cause a final shift in the labor demand function as indicated by the shift from D^J to D , in Figures 1(e) and k. Again since (~z) represents the decrease in price of the n factor associated with decreases in demand, the price increase occurring as a result of the increased demand from new firms as felt through [-7) would partially affect the earlier increase in the quantity of labor demanded as indicated through the term, N, — r-r— (-"j) q i,'^^,, with equilibrium quantity n demanded at Q., , as shown in Figure k. Lk ^ A similar analysis could be performed for each of the remaining four exogenous shifters in the model. Each discussion would center around the equilibration process as reflected through the appropriate terms in equations (2.1) and (2.2). However, due to their similarities and the complete treatment given dX^ , only a brief discussion is given for the remaining shifters in the following sections.

PAGE 44

32 Number of Firm Entrepreneurs ' The initial effect of an increase in an exogenous shifter of the number of firm entrepreneurs would, of course, be an increase in the number of firms in a given industry. Consequently, the residual return to each firm in the industry would decline as the result of both increases in the number of firms and increases in demand for the n factor which would increase its price. The quantity of the j product produced would initially increase along with the quantities demanded of the i factor and labor. Ultimately, price increases for the n factor would result in declines in residual returns to firms and lead to fewer numbersof firms. Quantities of the j product produced would then decline, along with the quantity demanded of the ! factor and labor. Equilibrium quantities would be expected to remain above their initial values as the result of the initial upward shift outweighing secondary changes due to the reduction in the number of firms. The actual quantities of factors utilized would depend on their subs ti tutab i 1 i ty for the n factor and the increase in the number of firms. Even though the n factor and labor might possibly substitute, as the price of the n factor increased, the overall effect should be an increase in quantity demanded of all factors as well as in the number of firms. Product Price Increases in product prices (upward shifts in perfectly elastic product demands) would directly affect the residual return to firms and cause firm numbers to increase. As firm numbers increase, the demand for the n^ factor would increase, and since its supply function is other than perfectly elastic its price would increase. Simultaneous increases would also occur in the quantities of labor and the i factor demanded. Ultimately, as the price of the n factor increased the

PAGE 45

33 residual return to firms would decrease accompanied by a decrease in the number of firms. This would cause a decrease in the quantities demanded t'h f h of labor, the i factor, and the n factor. Equilibrium conditions would be expected to result in an increase in the quantity of labor demanded as well as in the number of firms from the initial levels before the product price change. Factor Price Changes in factor prices, representing a shift in factor supplies assumed to be perfectly elastic, operate in an opposite manner than product price changes. As factor prices increase, the residual return to firms becomes smaller, thus directly causing a decrease in the number of firms. Initially the quantity of the j product produced, quantities of the n factor demanded, and quantity of labor demanded will decrease. These decreases would be offset somewhat by their substitutability for the i factor. As the price of the n factor declines as the result of smaller quantities demanded, existing firms would experience increased residual returns and, thus, new firms would be enticed to enter the industry thus increasing the demand for the n factor, the i factor, and labor, as well as increasing the quantity of product produced. Depending on the elasticity of substitution among factors and the relative magnitudes of the various changes occurring throughout the equilibration process, equilibrium increases in the quantities of labor t'h and the n factor demanded might occur simultaneously with the decreases in the quantities of product output, i factor demanded, and the number of firms.

PAGE 46

3^ Firm Production Possibilities Changes in firm production possibilities whicii are output increasing and/or input decreasing result in higher residual returns to firms and thus increases in firm numbers. Entrance of new firms would increase the output of product, quantity of both labor and the i factor demanded and bid up the price of the n factor through increases in its demands. increases in the n factor price would ultimately lead to a decline in returns and cause an offsetting effect by reducing the number of firms. This decrease would then lead to a decrease in output, and decreases in quantities demanded of the i factor and labor. Equilibrium quantities, however, would be expected to be greater than the initial quantities.

PAGE 47

CHAPTER I I I STUDY AREA AND MODEL SPECIFICATION One of the major objectives of this research was to determine the importance of natural resource investments and other types of investments in influencing employment changes in recipient areas. Particular interest was given to investments in critical resources including water and other forms of capital investment. For research such as this to be applicable in other areas some level of confidence must be maintained that knowledge acquired through research in a particular geographical area is transferable to other areas. If there is some question as to the applicability of research results among areas then the information becomes quite limited in use. Since regions differ in geographic, institutional, sociological, and economic characteristics one cannot conclude that natural resource investments stimulate employment in general without some regard to these different regional characteristics. Study Area The four-state region of Mississippi, Alabama, Georgia, and Florida containing 375 counties was chosen as the study area. The area was delineated into two groups of homogenous subareas. Counties in the four states were classified into tv/o groups on the basis of a set of ten variables depicting the county's human and natural resource endowments, and its urban, industrial, and agricultural structure. These ten 35

PAGE 48

36 variables were I96O measures of (1) population, (2) urban population, (3) percent of persons 25 years old and over with a high school education, (4) median age, (5) total employment, (6) total agricultural employment, (7) total manufacturing employment, (8) land area, (9) land in farms, and (10) value of farm products sold. Discriminant analysis was used in this delineation. Counties were grouped initially into two groups. A linear function of the differences of the means between the two groups was found which discriminated most successfully between the two groups. A general mean was then derived by substituting the means of all variables into the function. The function was then used to derive a mean for each county using county observations on each variable. Counties having values above the general mean were placed in one group and those having means below the general mean were placed in the second group. The process was repeated until the number of misclass if led counties was minimal. Computational procedures of the technique are outlined by Martin [2l]. Additional discussion of the discriminant analysis technique is also provided by Tintner [22, pp. 96-102], Several other variables were used with the above ten in grouping the counties, but their relative lack of importance in the discriminant function excluded them from consideration in the final delineation process. The most important variables in the delineation process were median age, education level, land area, agricultural employment, and total employment. Characteristics of the two groups were as expected. For example, the urban counties had higher average educational levels and higher average total employment. The nonurban group had higher average agricultural employment levels. Some judgement was warranted as to the mean ingf ulness of the delineations by persons familiar with the foiir-si.ate regio.i. Members of

PAGE 49

37 the Southern Land Economics Research Committee from each of the four states were asked to evaluate the initial classifications for their state. These evaluations were used to adjust the mathematical delineations and resulted in the final grouping shown in Figure 5. About one-fourth of the counties (91) represent the more urban-oriented group while the remaining counties (28^) exhibit a more rural orientation. Analyses were carried out for three groupings of counties. All counties in the four-state region constituted one group. The urban counties and nonurban counties formed the remaining two groups. Effects of similar investments in human, natural, and capital resources in these dissimilar areas should provide results that would be expected in other areas. For example, information on the effects of investments in urbanoriented versus the rural-oriented areas could provide guidelines on what types of investments should stimulate employment in the same types of areas in other regions of the country. Selection of Employment Categories Some categorization of employment is necessary to insure a meaningful analysis. Employment is reported on the basis of either occupational or industrial classifications. Occupational categories include such classifications as the various types of laborers, craftsmen, and professional wori
PAGE 50

38 in

PAGE 51

39 human, natural, and capital resources could affect employment in each occupational or industry category in a different way. Industrial classification lends itself to a more meaningful analysis from the standpoint of the theoretical framework outlined in Chapter 11. Also secondary data on industrial classification are more readily available. The industrial classification was, therefore, used to distinguish among employment categories in the analysis. Industry selection was made primarily according to the number of employees. Agriculture, construction, five individual manufacturing Industries, and durable and nondurable manufacturing were selected for analysis (Table 1). These Industries are likely candidates for growthgenerating activities with growth In the remaining industries being dependent on them. For this reason, and since data of the nature needed for this study are not readily available for the more service-oriented Industries, the service Industries such as wholesale and retail trade, finances, and communications were not included in the analysis. Also, the initial employment effects of investments In natural resources would most likely be felt In the Industries selected. Industries three through seven accounted for approximately 57 percent (762,098 employees) of all manufacturing employment In the fourstate region In 196? [23, PP. 1.7-1.11, 10.7-10.11, 11.9-11.13, 25.6-25.9]. The importance of construction and agriculture is shown in Table 1 where these Industries rank one (460,771 employees) and three (286,528 employees), respectively, when compared to employment in the individual manufacturing industries llh. Table 5^1 • Employment data for construction and agriculture were for I969.

PAGE 52

40 Table 1. Industry identification and employment rankings for the four-state area (196?) Industry ident if i cation for this study Industry Standard Total industrial fourclassificastate tion code employment (SIC) rank*^ 1 2 3 k 5 6 7 ND Agr icul ture Construction Textile mill products and other fabricated textile products manufacturing Food and kindred products manuf actur i ng Transportation equipment manuf actur i ng Furniture, lumber, and wood products manufacturing Electrical equipment manufacturing Durable products manuf actur i ng Nondurable products manuf actur i ng 15-17 01, 07-09 22, 23 20

PAGE 53

41 Water usage of each industry was also examined since two of the natural resource investment programs included in the study are water oriented. Water use data for construction and agriculture are not available. Agriculture should benefit from drainage, flood control and an increased irrigation water supply. Analysis of national rank in water 2 use for manufacturing industries indicated that some of the largest water using industries have not been included in the analysis. The petroleum and coal industry is the third largest industrial water use in the U.S. However, this industry provided only .3 percent of the total four-state employment and thus was not included. Manufacture of stone, clay, and glass products, nonelectrical machinery, and rubber and plastics products accounted for only 5.5 percent of total four-state employment. Additionally, these industries ranked low in total U.S. water use. Although the stone, clay, and glass industry provides products required in building water control structures, the combination of low employment and inadequate county data provided sufficient basis for excluding these industries. Inadequate county data also resulted in the exclusion of both primary and fabricated metal products, paper. and allied products, and chemicals and allied products, although chemicals and fabricated metals manufacturing rank one and two nationally in water use. The five individual manufacturing industries included in the analysis ranged from fifth to twelfth among the rankings for all industries in national water use. Water investment projects of the Soil Conservation Service included in this study do not influence major industrial users of water. However, certain projects carried out by the 2 National water use data are available in the Census of Manufactu res [25, Vol. ! , pp. 7.16-7/17].

PAGE 54

^2 Corps of Engineers could influence the location and expansion of these industries. Water use by these industries is important in assessing employment changes due to investments in water resource programs. General Model Specification Economic interpretation of changes in the exogenous shifters involved in equations (2.1) and (2.2) in the previous chapter made it apparent that interaction occurs in both equations simultaneously.. Estimation of the effects of exogenous changes on employment and firm numbers could be considered using a system of equations. Figure 1 represents the simultaneous changes occurring as the result of a change in resource supplies. Total Interaction as the result of all changes In the exogenous variables can be represented for each Industry by a system of equations. The availability of n county observations for each industry gives the following general system for the k industry. <3-') '\kt = ^0 -^ Pll^^ft ^ Pl2^^-kt " ^3''ikt " ^lAkt ^Yn^N^t^^lkt ^= '• 2 n <3-2)
PAGE 55

k3 dP.,^ = total change In product demands in the k J kt 3 r industry (exogenous). dP., = total change in factor supplies in the k industry for all factors other than the n factor (exogenous) . dZ,. = total change in firm production possibilities for firms in the k industry (exogenous). dN, ^ = total change in the number of firms in the k kt ^ industry (endogenous). dW ,, = total change in shifters of firm entrepreneur dkt ^ supply for firms in the k industry (exogenous). The P coefficients and the parameters of the distribution are un3 known which leaves the problem of obtaining estimates of these parameters. By rearranging equations (3.1) and (3-2) as (3.3) -dQ^^^ + v,,dN^^ Hp,^ + p,,dX^^ + P^2^P.^^ + P,3dP.^^ ^^4^^hkt = "ikt (3.^) -dN^^ + P^o ^ P2,dX^^ . P^Z^P.^^ . P23dPjkt ^ ^24^^hkt ^ ^zs'^dkt = "Zkt 3 Estimates derived using the following model are consistent and asymptotically efficient if the following assumptions are made; E P(kt) ^(k.t.s)1 = '^ ''\l = ° L ^ ^ * 'J = C otherwise [^kt) ^(kt)] E iX/. .X U,, .xl =

PAGE 56

kk they can conveniently be written in matrix notation as (3.5) -1 dQ Lkt dN kt ^10 ^11 ^12 ^13 ^1^ ^20 ^21 ^22 ^23 ^2^ ^25 [' • f— 1 ^^f ^''jkt ^^kt ^^hkt ^dkt and finally condensed to ] lkt 2kt (3.6) rY(^^) . px(,^j = U(,^) where r = 1 Y 11 (kt) dQ dN Lkt kt P = '^10 ^1 ^12 ^13 ^^ °20 ^21 ^22 ^23 °24 ^25 '(kt) = Q ^^ft ^^-kt ^^kt ^^hkt ^dk^ '(kt) u lkt U 2kt The subscript k indicates that this particular system is for the k industry. Each industry under study would be represented by an equation of this general form. Equation (3.6) presents the simultaneous linear equation model in its structural form. Each of these sources providesa detailed discussion of s imul ti^ncous equation models, their assumptions, limitations, and possible problems which might arise in their appl icat ion.

PAGE 57

45 Structural Estimation It is necessary that an equation in a set of simultaneous equations be just or overi dent if led before estimates of the parameters can be obtained. The order condition for identification can be used to determine whether a given equation is identified. If A is equal to the number of predetermined (exogenous) variables that do not appear in an equation and B refers to the number of endogenous variables in the equation, then the following conditions determine the identification of the equation. A = B 1 just identified A > B 1 over identified A < B 1 under identified Reduced Form Estimation The complete structural system may be written so that each endogenous variable is expressed as a function of all exogenous variables appearing in the system. This representation of the system is referred to as the reduced form and brings out the explicit dependence of the dependent variables on the predetermined variables and the disturbances. Premul t ipl ication of equation (3.6) by r and rearrangement gives (3.7) Y kt ^'P^kt) ^ r"' U (kt) or ^3-^^ \t = "^(kt) ^ ^kt) where .. 1 n = r P "lO "ll . . . "l5 '20 n2i . . . n25 "^Goldberger [26, pp. 352-372] -and Johnston [27, pp. 2^1-252] provide detailed discussions of the identification problem.

PAGE 58

46 "(kt) = ^kt) Ikt '2kt Estimates from the reduced form may be derived directly by ordinary least squares since their disturbances are linear combinations of the structural disturbances. However, in general practice the structural parameters are estimated first and then the reduced form parameter estimates are obtained by use of 11 = -T p. Individual Industry Models A two-equation system was specified for agriculture according to the theoretical framework out 1 ined in equations (2.1) and (2.2), and the general estimation model outlined in equations (3.') and (3.2). Single equation models were specified for construction and the various manufacturing industries. These models are presented below with each variable coded similarly to equations (3-1) and (3.2). That is, changes in factor supplies are represented by X, changes in shifters of firm supply by W, changes in product demands by PP, changes in factor prices by FP, and changes in firm production possibilities by Z. Data sources for each variable are given in the latter sections of this chapter and the mathematical derivation is given in Appendix B. ^The assumptions for the reduced form disturbances are E t^.oVo^-^f'^'Sko'c-'Skt))']'^"'^^-"' = otherwise = n for t = t

PAGE 59

^7 Aqr icul ture The system of equations representing the agricultural industry Is (3.9.1.1) E,,M„,,*^f ^P^,, X^,^p^,,PP,,.P3,jFP,, ^^911^11 *PlO,ll'=''Pi *Y|,,,,H,,.^„,, (3.9.1.2) ^i=Poi2\i,Pr,2'
PAGE 60

48 Program In 1971) by the Agricultural Stabilization and Conservation Service during the period I96O to 1970 in the I county. X^. = Change in total loans and grants for community water and sewer systems and waste disposal systems made by the Farmers Home Administration during the period I96O to 1970 in the i county. X/-. = Change in acreage of allotment crops due to reduction in allotments between 1959 and 1969 weighted by the proportion of the total value of the allotment crop to total value of crops and livestock in the i county in 1959* PP, . = Change in price index of agricultural commodity groups during the period 1959-61 to I969-7I weighted by the proportion of the value of the commodity group to the total value of crops and livestock in the i county in 1959. FP, . = Change in the average annual wage per hired farm worker during the period 1959 to I969 in the i county. Z,. = Change in the Southeast index of agricultural output per 1 I ^ man-hour for commodity groups during the period 1959-61 to 1969-71 weighted by the proportion of the value of the commodity group to total value of crops and livestock In the i county in 1959. GRP. = Intercept shifter dummy variable ity ity . , _ 1 r when urban-oriented count with bRP. Q "^when rural-oriented count N, . = Change in the number of farms during the period 1959 to 1 I ^ 1969 for the i county.

PAGE 61

hS WW.. = Change in total annual nonagr icul tural wage payments during the period I96O to 1970 per agricultural employee in I96O for the i county. WE,. = Change in total nonagr icul tural employment during the period I96O to 1970 per agricultural employee in I96O for the i county. WA. . = Change in the number of farm operators who were 55 or more years of age during the period 1959 to I969 in 4.1. -th the I county. U,i,U,,„. = Disturbance terms. 1 1 1 I , 112 I The Greek letters P and y represent the parameters to be estimated. Variables endogenous to the system are E. . and N... Equation (3.9.1.1) is overident if led and equation (3.9.1«2) is jus tident if ied. Construction Data availability on the construction industry precluded the calculation of changes in the product price and technology variables. Data on firm number changes also were inadequate. However, since the construction industry would be one of the more important in evaluating the primary employment effects of investments in natural resources a single equation model was formulated in this industry. The equation for the construction industry is (3.9.2.1) E^. -^02\l, ^2^i"P62^^1i^ v^ *-^82^2i ^ V2i ^h2i where E . = Change in construction employment for the period I96O to 1970 in the i * county.

PAGE 62

50 X.. to Xj-. = Same as In equation (3.9.1.1). FP^. = Change in average annual wages per construction employee during the period 1958 to 196? for the .th ^ I county. GRP. = Same as in equation (3.9.1.1)* WW . = Change in total annual wage payments during the period i960 to 1970 for all nonagr icul tural and nonconstruct ion employees per construction employee in I96O for the i county. WE„. = Change in total nonagr icul tural and nonconstruction 2i employment during the period I96O to 1970 per construction employee in I96O for the I county. U,^. = Disturbance terms. Manuf actur ing The remaining seven industries are manufacturing industries. A two-equation model did not yield results comparable to those obtained for agriculture. The number of manufacturing firms is much smaller and the size of firms (measured in terms of the number of employees) is generally larger than for the agricultural industry. Estimation of changes in the number of manufacturing firms did not add to the interpretive and explanatory power of the system of equations. Much of the impact of changes in the exogenous variables for manufacturing industries is transmitted through employment changes within existing firms rather than change in the number of firms. Therefore, changes in the number of firms in each manufacturing industry was treated as an exogenous variable and a single-equation model was specified for each manufacturing industry, This model is of the form

PAGE 63

51 (3.9. k.l) E^, =0^^. z ^^\rhu'\'. " V^i r = I ^^8k^i ^P9k^''l ^^10,k^^ -^^l.kW^ki ^^}2,k\\ ^^k\ k = 3, ^, . . .. 7, D, ND where E, . = Change in k industry employment for the period I96O to 1970 in the i county. [. . to Xr . = Same as in equation (3.9.1.1). PP. . = Change in price index for each SIC three-digit level industry commodity group in the k industry during the period 1959-61 to I969-7I weighted by the proportion of value added by the commodity group to total value added by the commodity group to total value added in the k industry in 1958 for the i county. FP, . = Change in average annual wages per production worker for each SIC three-digit level industry group in the k industry during the period 1958 to I967 weighted by the proportion of production worker employment in each subgroup to total production worker employment in , , th , , , . th the k industry for the 1 county. Z, . = Change in the index of output per man-hour for each SIC three-digit level industry group in the k industry during the period I959-6I to I969-7I weighted by the proportion of value added by the commodity group to total value added in the k industry for the i county.

PAGE 64

52 GRP. = Same as in equation (3.9.1.1). th WW . = Change in total annual nonagr icul tural and non-k industry wage payments during the period I96O to I970 per k industry employee in I96O for the i county. WE. . = Change in total nonagricul tural and non-k industry th employment during the period I96O to 1970 per k i ndustry employee in I96O for the i county. N, . = Change in the number of firms in the k industry during the period 1959 to I969 for the i^ county. U, . = Disturbance term. ki The symbols P and y represent the parameters to be estimated. Model Estimation Procedure Two-stage least squares was used to estimate the structural parameters of equations having the form of (3.9.1.1) in the two-equation system, for agriculture. Ordinary least squares was used to estimate equations having the form of equation (3.9.2. 1) in each of the uses of the two-equation models. This statistical procedure was also used for all single-equation models. Estimates of the reduced form coefficients for the two-equation models were derived from the structural parameter estimates. A computer progran written by William James Raduchel [28] was used, ^ibid.

PAGE 65

53 Measurement of Variables and Empirical Expectations Clianges in employment and firm numbers as defined in equations (3.9.1.1) and (3.9.1.2) were considered as endogenous variables. in the model that forms a system of equations, the endogenous variables are considered functions of some or all predetermined (exogenous) variables. In the first equation employment changes were also considered as a function of the endogenous variable for firm number changes. Predetermined variables rapresent changes in exogenous shifters of product demand, factor price, critical resource supply, firm production possibilities or firm entrepreneur supply functions. Each of these general types of shifters is represented by one or more variables in each equation. Each variable, the type of shifter it represents, and the expected effects of changes in these variables on agricultural employment and the number of farms are given in Table 2. A similar illustration of the effects on construction and manufacturing employment is given in Table 3. Detailed discussion of each variable follows in the remaining sections of this chapter. Employment Changes in employment for each county were computed for the time period I96O to I97O as reported in the I960 Census of Population [29, Table 85] and the I97O Census of Population ilk. Table 1231. Employment is reported in this source using basically the same industry categories as suggested by the Standard Industrial Classification (SIC) of industries, It was necessary to combine employment reported in some industries to conform to the industry classification shown in Table 1. Changes in employment were used as dependent variables.

PAGE 66

5^ in iV J3 E c E tK-D C ID C 0) E >. o Q. E 0)
PAGE 67

55 a. <4X B uj o o C 0) E >. _o E <1) Ol c u c E c o in a. tJ J3 —

PAGE 68

56 Firm Numbers Changes in the number of manufacturing firms for the period 1958 to 1967 were obtained from data reported in the 1958 Census of Manufactures [30, State Table 7] and the 19^7 Census of Manufactures £23, State Table 9]. Firm numbers are reported for each two-digit SIC industry in each county. Some industries were combined to conform to the classification of industries used in this study. Changes in the number of farms from 1959 to I969 for each county were obtained from the 1959 Census of Agriculture [3I, County Table l] and the I969 Census of Agriculture [32, County Table l]. Changes in the number of farm firms were used as both predetermined and endogenous variables. This distinction is discussed in a later section. Factor Suppl ies Changes in the supply of critical factors in a county should influence that county's labor employment. Investments that increase the supply of a given factor would be expected to cause the price of the factor to decline initially. A price decline would entice users to substitute more of the factor for other production inputs including labor. With perfectly elastic product demands and a reduced price for the critical factor, residual returns to firm operators should increase. In the absence of any substantial barriers to entry of new firms into the industry, new firms would be established, and this would increase industry output. Entrance of new firms would result in an upward shift in the demand for all factors, including labor, causing the prices of those factors having upward sloping supply functions to rise. Existing firms would also expand output. Residual returns to firm operators would subsequently fall, resulting in a cutback in factor employment. At the new

PAGE 69

57 equilibrium, the quantity of labor employed would be greater or less than the initial quantity, depending on the relative sizes of the demand and supply elasticities as pointed out in the theory section. It would also depend on the elasticity of substitution of labor for the other factors whose prices varied. Labor could probably not be easily substituted for many of the natural resources considered in this analysis. Therefore, investment programs to increase the supplies of these natural resources should also increase the amount of labor required to complement industry expansion resulting from either the initial project construction phase or from users of the project. Education investments (X .). — Federal and state expenditures per pupil for education were considered to be exogenous to a county. Changes in these expenditures over the period 1959I96O to I969-I97O were included as a measure of exogenous shifts in the investment in a county's human resources. Expenditure estimates were obtained from state education agencies in Alabama [33, 34], Mississippi [35, 36], and Georgia [37, 38], and the Florida Statistical Abstract [39]. Increases in education expenditures should partially reflect an increase in the number of persons in the recipient county attaining a given educational level as well as an increase in the average productivity level of the labor force. Unless outmigration from the counties of people receiving the education occurs, an upgraded labor force should attract potential employers and eventually result in increased employment within the recipient counties. The effect of increases in education levels or agricultural employment and farm numbers should be opposite their effect on construction

PAGE 70

58 and manufacturing employment. Since agricultural labor draws heavily from the unskilled labor force it seems likely that increased educational opportunities would reduce this labor force and lead to outmigration from the more rural counties to more attractive job alternatives in urban areas, This would lead to fewer agricultural workers and probable consolidation of farms to gain operational efficiencies. Increased educational opportunity should, therefore, yield negative coefficients for agricultural employment and farm numbers and positive coefficients for construction and manufacturing employment as shown in Tables 2 and 3. Corps of Engineers' natural resource investments (X ).--For this variable as well as the other types of natural resource projects, total project expenditures by county over the period I96O to I97O were used as the independent variable. interpretation of the estimated coefficients for these investment variables should provide Insights Into two components of the analysis. First, the empirical significance of the various types of natural resource investments on local employment and firm numbers can be shown, and second, the relative Importance of the various natural resource Investment categories In influencing employment and firm numbers can be appraised. Investments by the Corps of Engineers in civil works and new work construction were obtained for each county from the various district offices which administer portions of the four-state area. Investments o Corps of Engineer personnel providing data through personal communications were: J. W. Dement, Chief, Engineering Division, Memphis District, U.S. Army Corps of Engineers, Memphis, Tennessee. W. T. Moore, Chief, Engineering Division, Savannah

PAGE 71

59 were measured in thousands of dollars. These investments covered projects categorized into multipurpose, navigation, flood control, beach control, and recreation projects. The major portion of expenditures was for flood control and navigation with a very small portion allocated to beach control and recreation. Due to the large amount of the investments going into flood control and navigation projects, no distinction among the above investment categories was made. Some expenditures by the Corps for construction projects along the Mississippi River could not be allocated to counties. Therefore, these investments were not included in the analysis. Investments occurred in 1^8 of the 375 counties comprising the four-state area. Investments by the Corps of Engineers should also have contrasting effects on agricultural and manufacturing employment. Improved flood control would be beneficial to agricultural areas by making more land available for use. This would bring about a two-fold reaction. Both expansion of existing farms and the entrance of new farms would occur with the probable result being decreased residual returns to each firm as the price of land is bid upward. The probable consequence would District, U.S. Army Corps of Engineers, Savannah, Georg ia. Powell Williams, Jr., Asst. Chief, Engineering Division, Mobile District, U.S. Army Corps of Engineers, Mob i le, Alabama. George Marsh, Acting Chief, Engineering Division, Jacksonville District, U.S. Army Corps of Engineers, Jacksonville, Florida. J. L. Smith, Chief, Construction Division, New Orleans District, U.S. Army Corps of Engineers, New Orleans, Louisiana. K. E. McLaughlin, Comptroller, Vicksburg District, U.S. Army Corps of Engineers, Vicksburg, Mississippi. F. P. Gaines, Chief, Engineering Division, Nashville District, U.S. Army Corps of Engineers, Nashville, Tennessee.

PAGE 72

60 be further expansion of the larger more established farms with the overall effect being a reduction in farm numbers. It follows that the larger farms might also operate with a smaller total labor force though a more efficient operation realized as the result of the larger farm size. The consequence would be reductions in farm numbers and agricultural employment as indicated by the coefficient sign in Table 2. In contrast to the effect on agriculture, construction and manufacturing employment would likely increase as the result of Corps of Engineers' investments. Employment would certainly increase in the recipient area during the initial construction phase of the project. Initial effects might also be felt in manufacturing provided that local area materials were used. More importantly for manufacturing, however, would be the effect occurring during the life of any project. Improved transportation facilities, protection from flood damage, etc., would encourage the entrance of new firms and resultant employment increases. Since firm consolidation in manufacturing does not occur as readily as in agriculture, the indirect effect of decreases in firm numbers and employment is probably not large enough to offset the initial positive gains in the recipient area. A positive coefficient for construction and manfacturing employment would be expected as indicated in Table 3. Soil Conservation Service PL-566 investments (X.) .--Construction — — — — — ' ^—yexpenditures by county from I960 to 1970 for the Small Watershed Program were obtained from personnel in each of the four-state offices of the Soil Conservation Service. Data were tabulated from 239-B forms which q Soil Conservation Service data were provided through personal communications from;

PAGE 73

61 gave actual dates of construction expenditures for each project. Investments were then allocated to each county based on project location as identified on maps prepared by the Soil Conservation Service. A total of 111 counties received these types of investments during the study period. Investments were measured in thousands of dollars. The Small Watershed Program is designed to aid in the solution of several types of problems. Of major importance among these is the reduction in floodwater damages to cropland, residences, businesses, and protection of the health and lives of people from floods. Other potential and existing problems that this program attempts to alleviate include erosion and sediment damage, improper drainage, and irrigation needs. Recreation, fish and wildlife enhancement, and improvements in the economic and social well-being of people have also received attention. These latter categories have been given increased emphasis in recent years. PL-566 investments by the SCS which result in Improvements for the local recipient areas should provide conditions that affect agricultural and manufacturing employment and firm numbers within the local areas in a manner quite similar to that of Corps of Engineers' investments. Reduction of floodwater damage to cropland should in total reduce the number of farms and agricultural employment as indicated in Table 2. Although some new farms might become established Barbara Kennedy, Accounting Technician, Soil Conservation Service State Office, Auburn, Alabama. Robert Salsman, Financial Manager, Soil Conservation Service State Office, Jackson, Mississippi. George Adair, Accounting Technician, Soil Conservation Service State Office, Athens, Georgia. Gertrude Griffin, Accounting Technician, Soil Conservation Service State Office, Gainesville, Flor Ida.

PAGE 74

62 the consolidation effect into larger farms to take advantage of improvements made possible by the PL-566 project should be greater. Construction employment should increase in the recipient area during both the initial and secondary project phases and ultimately as a result of the project in a manner similar to that discussed for Corps of Engineers' investments. Also to be expected is an increase in manufacturing employment as indicated in Table 3. Increased output resulting from project expenditures should provide a base for more manufacturing employment in conjunction with increased enhancement for manufacturing firm location resulting from the reduction in floodwater damages to residences, businesses and through the effect of other firm location attributes improved by the investment project. Agricultural Stabilization and Conservation Service ACP investments (X, ) .-I nvestments by the Agricultural Stabilization and Conservation Service constitute a joint effort by the public sector, farmers, and ranchers to share the cost of establishing needed conservation measures. These conservation programs include practices to protect, improve, and renew soil, water, woodland, and wildlife resources of private landowners. Data for the analysis were obtained from annual state ASCS reports during I96O to I97O for Alabama [40] , Mississippi [4l], and Florida [^2]. Expenditure information included cash payments to farmers and allowances paid to vendors for conservation materials furnished farmers. Data for Georgia were taken directly from computer printouts. A total of 37^ counties had participants in the ACP Personal communication from the Data Division, Agricultural Stabilization and Conservation Service, U.S. Department of Agriculture, Washington, D.C.

PAGE 75

63 Program during the study period. Investments were measured in thousands of dol lars . Conservation measures which make land more available could also influence the expansion of existing farms as well as encourage new farms. The trend historically has been toward larger firms. Since this is a cost-sharing program it seems logical to expect the larger farmers to take advantage of this program opportunity and expand operations even more. The indirect effect of entrance of new farms should be more than offset by the consolidation of existing farms leading to a decrease in total agricultural employment and farm numbers. Conservation measures that remove land from production would provide a similar circumstance. This negative overall effect is indicated in Table 2. Since this type program requires some construction activity and manufactured input which usually are purchased locally, a positive effect on construction and manufacturing industry employment would be expected as indicated in Table 3. Increased output would also be expected to result from the application of conservation measures leading to a need for more processing and support facilities which In turn should have some positive effect on manufacturing employment. Any negative feedback effect on manufacturing employment should not be large enough to affect the initial positive effect. Farmers Home Administration investments (X^-) . — Loans and grants for community v^ater, sanitary sewer, and solid waste disposal systems were also considered to be Investments that would Ivifluence employment and farm numbers In each county as indicated in Tables 2 and 3. This program provides financial assistance to comniuni ties in developing essential new public service facilities and in expanding existing

PAGE 76

64 facilities. Data for these investment loans in thousands of dollars were obtained from the various state directors of the Farmers Home Administration. A total of 278 counties received financial assistance from FHA during the study period. Services and facilities provided by this type program are necessary before a community can expand with regard to attracting new industry and in turn services to support these industries. Communities demonstrating adequate services will likely attract new industry and thus expand employment in construction and manufacturing industries. Expansion of existing firms in the community might also occur, and this further supports the positive employment coefficient demonstrated in Table 3. This program does not specifically influence a production input used in agriculture such as land or water in the same manner as the four earlier programs. A similar negative effect on agricultural employment and farm numbers as discussed for the earlier programs would be expected. As community facilities become available and the community begins to develop its manufacturing base, job alternatives for agricultural employees and farm operators become more available. Smaller farms are soon consolidated with the displaced operators assuming other types of employment. Fewer employees are then required because of more efficient operations and the negative effect occurs. Crop allotment (X ^) .--Changes in crop allotments represent the effect of shifts in a perfectly inelastic factor supply on the number of State Directors, Farmers Home Administration, providing data through personal communication were; S. B. Wise, Jackson, Mississippi; John N. McDuffie, Atlanta, Georgia; John A. Garrett, Montgomery, Alabama; and William Shaddick, Gainesville, Florida.

PAGE 77

65 farms and employment in agriculture. A decrease in allotment acreage would be expected to increase the market price of the allotment, or of land, leading to lower residual returns to farm operators and ultimately to a reduction in the number of farm firms. As the number of farms decline aggregate demand for allotments would decline and thus lower their market prices. The magnitude of changes in the number of farm firms and consequently in agricultural employment due to a decrease in allotments would depend on farm operators' responsiveness to changes in their residual returns, the amount of allotments used in the production process, and the actual level of operator returns. Reductions in acreage allotments between 1959 and I969 for all allotted crops were computed for all counties having acreages of these crops. Annual reports for 1959 and I969 from the Agricultural Stabilization and Conservation Service in Alabama [^O] , Mississippi [4l], Georgia [43] and Florida [^2] provide data on allotted crop acreages. County reductions were v;eighted according to total value of sales of each crop as a proportion of total value of crop and livestock sales in the county in 1959 as derived from data available in the 1959 Census of Agr icu 1 ture [3I]. Declines in harvested acres in each county were also computed using the same data sources that provided information on allotment reductions. The smaller of these two changes was then selected as the effective cut in allotments. Allotment reductions were not considered relevant for a county if its harvesting acreage in 1959 was less than the county's acreage allotment for the selected crop in 1969Positive coefficients would be expected as shown in Table 2.

PAGE 78

66 Product Demand Use of a product price as an indicator of product demand is based on the assumption of perfectly elastic demand functions at the county level. Producers in both agricultural and manufacturing industries at the county level are assumed to be price takers and thus face perfectly elastic demand functions. An increase in product price would initially increase residual returns to firm entrepreneurs. This would have the effect of enticing new firms into the industry and ultimately an increase in the quantity of resources employed in the production process including labor. Also, existing firms would expand output and hence increase their demands for production factors. increases in demand for factors with upward sloping supply functions would lead to increased factor prices and reduced returns to firm entrepreneurs, Consequently, firm numbers would decline and this would result in a reduction in labor employment. Existing firms would also reduce their output and cause reductions in resource demands. The net change in labor employment would depend on the relative magnitude of both the initial and indirect effects of these changes in output of existing firms and in the number of firms. A product price variable was not included for the construction industry since output is not easily defined in terms of a product with an established market price. Agricultural product price (PP ,). — Changes in price Indexes were computed for each of the seven major agricultural commodity groups produced in the study area using three year averages centered on 1959 price indexes [hk'] and I969 price indexes [^Sl. These changes were weighted by the 1959 value of each commodity group as a proportion of the total

PAGE 79

67 value of crops and livestock produced in each county. The resulting measure was a weighted change in product prices faced by farm producers at the county level. Value of products sold was obtained from the 1959 Census of Agriculture [31, County Table Sl . A pr ior i specification of the net result of increases in agricultural product prices on employment is difficult. It is quite possible that increases in Agricultural product prices could lead to a reduction in agricultural employment and farm numbers due to farm consolidation. Since some agricultural operations do allow fairly easy entry, the opposite effect could occur under certain conditions. Manufacturing product price ( PP . ) .--Changes in national wholesale price indexes as reported by the Bureau of Labor Statistics [46] for three-digit level (SiC) industries between 1959-61 and I969-7I were used in computing a county product price for each two-digit level industry. For each industry, county price changes were obtained by weighting the change in the three-digit national wholesale price indexes by the 1958 value added by manufacturing for each three-digit level industry as a proportion of the total value added for the two-digit industry in the county. Data used to calculate value added for each industry were obtained from published data made available by the U.S. Bureau of the Census [30, ^7] . Expected effects of product price increases in the manufacturing industries are also difficult to specify. Existing firms would be expected to expand output and new firms enter the industry as the result of a product price increase. This assumes there are no barriers to entrance. Increases in employment should occur. Output increases should ultimately cause price declines with some of the earlier

PAGE 80

68 expansion being offset. With manufactured products, unlike agricultural commodities, some apparent downward inflexibilities of prices would help support in part a conclusion that the indirect effect of decreasing employment would not completely offset the direct effect leaving a positive overall effect on manufacturing employment. It does remain possible that decreased residual returns to firms as the result of entry by new firms would be substantial enough to cause an actual decline in employment. Both alternatives are indicated to demonstrate the effect of product price increases on manufacturing employment in Table 3. Factor Price Changes in the price of factors whose supply is assumed to be perfectly elastic would affect the net returns to firm entrepreneurs and consequently the number of firms. Indirectly, the level of labor employment would be affected. As the price of factors having perfectly elastic supply functions increased, labor as well as other production factors would be substituted for these inputs to the extent possible. This would result in an increase in the price of all factors having upward sloping supply functions. As these factor prices increased, residual returns to firm entrepreneurs would decrease and consequently the number of firms would decline. A reduction in firm numbers would decrease factor demands, resulting in a decline in factor prices. Increases in residual returns to firm entrepreneurs would entice some new firms into the industry with resultant increases in labor employment. Employment levels under new equilibrium conditions would depend on the relative magnitudes of these various changes. The more inelastic the supply function of critical factors, other than labor, the larger the price increase will

PAGE 81

69 be for that factor as demand for it increases. Consequently, labor whose supply function Is more elastic would be substituted for the higher priced factor with a resultant employment increase. Agriculture wage rate (FP .).--A proxy variable was used as the annual wage rate for agricultural employees. Total expenditures for hired farm labor in 1959 and I969 for all farms in each county were divided by the total number of hired farm laborers working 150 days or more each year in that county to obtain an annual wage per worker. The change in this wage was then computed. These data were obtained from the 1959 and I969 Censuses of Agriculture [3I, 32]. Employment effects of increases in hired farm labor wage rates should be negative as shown in Table 2. Wage increases would result in higher factor costs to operators. This would encourage substitution of other factors for labor. Smaller farms would not be able to make sufficient substitutions and would not be able to compete with larger and more efficient farm operations. Farm numbers would then decline through consolidation and expansion of existing farms. Manufacturing v/age rate (PP .). — Changes in average annual production worker wage rates between 1959 and 1970 for each two-digit level manufacturing industry in each county were used for manufacturing wage rate changes. Data were obtained from the 1959 and 1970 County Business Patterns for each state [^8, '+91. If the two-digit level industry wage was not reported for a county due to disclosure problems, the change in average annual production worker wage for all manufacturing industries in the county was used. Increases in an industry's wage rates would be expected to result in a decrease in labor employment within the industry as

PAGE 82

70 indicated by the negative coefficient sign in Table 3. Other production factors would be substituted for labor as the price paid to labor increased. Technology Changes in technological forces that affect agriculture and manufacturing industries should have an effect on the amount of labor employed. Similar to the other types of shifters discussed previously, technology changes would also affect factor demand, product supplies, and the number of firms. Increases in technology that were output increasing and nonlabor input decreasing would cause the quantity of products produced to increase with subsequent decreases in the use of inputs. Prices of those inputs having inelastic supply functions would decrease since demand for them would decline. Further substitution of the lower priced inputs for labor could cause a decrease in employment. If technology changes had been labor decreasing, the quantity of labor would have decreased initially. The indirect effect of these changes would be an increase in the number of firms concomitant with an increase in residual returns as a result of the change in technology. New firms would then increase the demand for all factors and reduce firm residual returns. Equilibrium quantity of labor demanded could be either smaller or larger than the initial quantity demanded depending on the degree of factor supply inelasticity, substi tutab i 1 i ty of labor for the other factors used, and the magnitude of changes in the number of firms. Agricultural technology (Z .) .--Changes in output per man-hour in agriculture were used as indicators of changes in agricultural technology. Changes in the index of output per man-hour for six major commodity groups

PAGE 83

71 in the Southeast were computed using three-year averages centered on 1959 and 1969 [50, p. 8], These changes were then weighted by the 1959 value of each commodity group produced as a proportion of the total value of crops and livestock produced in each county obtained from the 1959 Census of Agriculture [3I, County Table 5]. The resulting measure was a weighted change in labor productivity for each county. Trends in output per man-hour and advances in agricultural mechanization suggest that technology increases in agriculture are likely to be labor decreasing. A negative effect on agricultural employment should be indicated as suggested in Table 2. Similar effects would be expected on farm numbers. Technology advances should enable the operation of larger farms with resultant decreases in farm numbers. Manufacturing technology (Z ,). — Changes in technology for each manufacturing industry were computed in a manner similar to that for agriculture. Changes in national output per man-hour indexes between i960 and 1970 for three-digit level industries were used in computing a county technology change variable for each two-digit level Industry. These indexes are published by the Federal Reserve System [5I]. Changes in the national output per man-hour indexes for the three-digit level industty were weighted by the I959 value added of each three-digit level industry as a proportion of the total value added by the two-digit level industry in the county. The same value added data used in calculating industry product price was used in the weighting procedure. Technology changes in the manufacturing industries have employment effects similar to those in agriculture. The negative effect indicated in Table 3 implies that technology changes are probably labor decreasing, A technology

PAGE 84

72 variable for the construction Industry was not Included since output per man-hour indexes for construction were not available. Farm Operator Suppl les Changes in farm operator supplies affect agricultural employment in various ways. Several variables used In this study are quite unique with respect to the types of shifters discussed earlier. These shifters are thought to affect farm operator supplies which In turn affect agricultural employment. Changes that Increase the number of farms Indirectly cause increases in the amount of products produced and factors used including labor. Ultimately, the price of factors having less than perfectly elastic supply functions would Increase with concomitant decreases in the number of farms, quantity of products produced, and quantity of labor employed. Equilibrium employment levels would depend on the relative magnitudes and effects of the described changes. in general, declines in farm operator numbers should cause declines in agricultural employment. Agricultural wage opportunity (VAJ .) .--Wages In Industries other than agriculture represent changes In the opportunity cost to farm operators of remaining in present employment as a result of changes in wages in other employment alternatives. Initially, wage increases in employment alternatives would decrease the number of farm operators remaining in agriculture. As the larger farms realize greater residual returns some increase in farm numbers might occur. This effect should be minimal with an overall decline In farm numbers expected as indicated in Table 2. The movement to fewer, larger, and more efficient farms should then cause a negative effect on agricultural employment as indicated in Table 2.

PAGE 85

73 Change in agricultural opportunity wages between I96O and 1970 in each county was determined using data on employee wages obtained from County Business Patterns [48, 49] and the Census of Population [29]. Change in annual nonagr icul tural wages between 1959 and 1970 per agricultural employee in I96O was used as the indicator of wage opportunity for agricultural employees. Changes in county unemployment levels would have provided an alternative measure for this variable. Agricultural employment opportunity (WE .). — Increases in employment opportunity in alternative employment situations would be expected to decrease the number of farm operators remaining in agriculture in a manner similar to that of increases in wages in employment alternatives. Changes in employment alternatives were calculated using employment data obtained from the I96O and I97O Censuses of Population [24, 29]. Changes in nonagr icul tural employment between I96O and 1970 per agricultural employee in i960 indicate employment opportunities for agricultural workers and farm operators. Farm operator age (WA) .--Farm operator age represents the change in the number of farmers who were 55 or more years of age during the period 1959 to 1969. This variable is intended to reflect the relative effects of potential operator retirements on the number of farms during 1959 to 1969. The greater the number of farmers who are reaching an older age, the greater should be the decline in farm numbers and employment during the entire period. This variable was calculated for each county from the 1959 .Census^ of Acri culture [3I, County Table 5] and the 1969 Census of Agricu lture [32, County Table 3]. A positive coefficient sign would be expected as sh'';wn in Table 2. Declines in the number of

PAGE 86

7^ older farm operators would be expected to cause declines In farm numbers. Implicit in this is the assumption that farm consolidation occurs rather than operator replacement. Manufacturing Labor Supplies Changes in the supply of labor available to a particular manufacturing industry certainly affect employment In that industry. Shifters of labor supplies would logically cause changes in the number of firms in the industry which would in turn affect employment. Changes in the number of manufacturing firms for a given manufacturing industry were considered exogenous, however, and shifters of manufacturing industry labor supplies are discussed below as a direct effect on manufacturing industry employment. Manufacturing wage opportunity (W , ) . — Wages In manufacturing industries other than the Industry of present employment (k ) represent changes in the opportunity cost to employees of remaining In present employment as a result of changes In wages in other employment alternatives, Initially, wage increases In other Industries would entice employees to leave their present industry if their skills were transferable. Their present Industry might bid wages upward and regain to some extent but an overall negative effect would be expected as indicated In Table 3. Change In opportunity wages between I96O and 1970 In each county was determined for construction and manufacturing using data on employee wages obtained from County Business Patterns [48, ^9] and the Census of Population [29]. It was hypothesized that employees would not be moving into the agricultural industry because of its low average wage level. The change in annual nonagr Icul tural wages between 1959 and 1970 per

PAGE 87

75 construction worker in I960 in each county was used to indicate the wage opportunity for construction industry employees. A similar measure was calculated per manufacturing employee in each county. Manufacturing employment opportunity (WE , ).-Increases in employment opportunity in alternative employment industries would be expected to affect the number of employees in the given industry in a manner similar to that of increases in wages in other industries. Changes in alternative manufacturing employment were calculated for each county using employment data obtained from the I96O and 1970 Census of Population [2^, 291 by computing the change in all employment other than agriculture and industry of present employment per worker in industry of present employment. Increases in employment opportunity in other manufacturing industries should decrease employment in the industry of present employment as shown in Table 3. Growth in industries that are complementary in nature would be expected to positively affect employment in each other. Number of manufacturing firms (N .) .--Changes in the number of manufacturing firms were used as predetermined variables in the construction and manufacturing models. An increase in the number of firms In general would be expected to bring about an increase in employment. Some cases would exist where intraf irm expansion could bring about an employment increase while firm numbers were declining. In general, a positive sign should be expected for this coefficient as shown in Table 3. Data sources for firm numbers were outlined earlier.

PAGE 88

CHAPTER IV ANALYSIS OF RESULTS Parameter estimates in equations Identical to those presented in Chapter III were made for agriculture, construction, and the various manufacturing industries. Equations (3-9-1.1) and {}.3.].2) were used for agriculture. Equation (3-9.2.1) was used for construction and equation (3-9'k'l) was used for the manufacturing industries. All equations were estimated for each of the three groupings of counties. Counties were excluded if no employment was reported in both I960 and 1970. In this chapter a comparison is made of parameter estimates obtained from estimating the relationships for each of the three groups. Effects of the predetermined variables on employment in each industry and farm firm numbers for agriculture are discussed. General comparisons of the results obtained for all three groups are made. Aqr icul ture Tables k through 6 contain the parameter estimates for the twoequation models used for agriculture. Each table presents three equations for one of the three groupings of counties. Table 4 presents the results for all counties. Table 5 the results for the urban counties and Table 6 the results for the nonurban counties. !n the follov^ing discussion each agricultural equation in the table'i is not referred to separately as the parameter estimates are discussed. Each parameter and

PAGE 89

77 Table 4. Structural form and reduced form coefficients for change in agricultural employment (Ei) and number of farm firms (N,), all counties, I960 to 1970.

PAGE 90

78 Table k (Continued)

PAGE 91

79 Table 5. Structural form and reduced form coefficients for change in agricultural employment (Ei) and number of farm firms (Ni), urban counties, I96O to I97O

PAGE 92

80 Table 5 (Continued) Endogenous variables Predetermined var iables Structural Derived reduced form coefficients form coefficients Agricultural Number of Agricultural employment farm firms employment (E,)b (N,)c (E,) Number of farms (N.) .0702 ' (.2503) R^ -.93 R^ — .92 Complete variable definitions can be found in Chapter ill. Change in agricultural employment was estimated with two-stage least squares. Figures in parentheses for this equation are asymptotic standard errors. Levels of significance are not indicated since they are approximat ions . Change in number of farm firms was estimated by ordinary least squares with figures in parentheses indicating standard errors. Since this equation is justident if ied and contains all predetermined variables the structural coefficients are identical to the derived reduced form coeff ic ients . "Significant at 10 percent level. "^-Significant at 5 percent level. -A-A-vsignif icant at 1 percent level.

PAGE 93

Table 6. Structural form and reduced form coefficients for change in agricultural employment (E^) and number of farm firms (Ni), nonurban counties, I96O to 1970

PAGE 94

82 Table 6 (Continued) Endogenous variables Structural Derived reduced Predetermined form coefficients form coefficients variables Agricultural Number of Agricultural employment farm firms employment (E,)b (N^)c (Ep Number of farms (N.) 1.3643 ' (.1322) R^ — .80 R^ — ,80 Complete variable definitions can be found in Chapter III. Change in agricultural employment was estimated with tvra-stage least squares. Figures in parentheses for this equation are asymptotic standard errors. Levels of significance are not indicated since they are approximat ions . Change In number of farm firms was estimated by ordinary least squares with figures in parentheses Indicating standard errors. Since this equation Is jus ti dent if led and contains all predetermined variables the structural coefficients are identical to the derived reduced form coefficients. *Slgnif leant at 10 percent level. *^'>Signif leant at 5 percent level. ***SignIf icant at 1 percent level.

PAGE 95

83 the comparisons that are made can be found in the three tables. Means and standard deviations of each variable are given in Appendix B. Type of Equation Tables k through 6 contain both structural equations and derived reduced form equations . Each structural equation for changes in employment was estimated by two-stage least squares. The equation is overident i f led in each model and contains two endogenous variables. Coefficients of these equations can be interpreted as the direct effect on changes in agricultural employment of a one-unit change in the predetermined variable. Figures in parentheses are asymptotic standard errors. These can be examined in relation to each coefficient to provide an approximation of the statistical reliability for each coefficient using standard normal test procedures. The equation for changes in the number of farm firms was estimated by ordinary least squares. It is one of the two structural equations of each two-equation model. The equation contains only one endogenous variable (change in the number of farm firms) and all of the predetermined variables in the model. Each table also contains a derived reduced form equation for employment changes. Reduced form equations express an endogenous variable as a function of all exogenous variables in the model. Coefficients in this equation can be interpreted as the partial derivative of the endogenous variable with respect to any predetermined variable with all other predetermined variables held constant at their mean values. This coefficient indicates the total effect of a change in the endogenous variable after taking into account the i nterdependenc ies among all current predetermined variables. This coefficient can be

PAGE 96

8^ referred to as a multiplier in contrast to a structural coefficient which indicates only a direct effect. For this particular model the structural form equation and reduced form equation for changes in the number of farm firms are identical. The structural equation contains all predetermined variables in the model expressed as a function of changes in the number of farms. This is also the reduced form equation of the model by def inl t ion. Endogenous Variables Changes in agricultural employment and in the number of farm firms from I960 to 1970 represented the tvjo endogenous variables in the agricultural equations. The trend in agricultural employment during this time period was generally downward. Although some counties did experience increases the average change was negative. The urban counties experienced the smallest average decline per county in agricultural employment while the nonurban counties had the largest average decline (Appendix B, Table 18) . Urban counties had fewer agricultural employees and had already experienced larger declines in some earlier time period. Employment decline in the urban counties averaged slightly less than one-half the decline in the nonurban group. Average decline for all counties was approximately twice that of all urban counties. The number of farm firms per county also declined for each of the three groups with the magnitude of the declines for each group having the same ordering as employment declines. Relative differences were not as large as experienced in employment. Large variation among counties was indicated for both farm number changes and agricultural employment changes (Appendix B, Tables 18 and 19)-

PAGE 97

85 A decline in farm numbers would be expected to cause a decrease in the number of agricultural employees. Fewer farms v/ould cause a decline in farm operators and decrease employment. Larger and more efficient farms should also require fewer employees. This relationship was supported without exception by positive coefficients for changes in the number of farm firms when this variable was included in the employment change equation. Standard errors for this coefficient were relatively small in relation to the coefficient for both the all county group and nonurban group. The largest effect of a decrease in one farm firm on agricultural employment occurred for the nonurban group of counties while the smallest effect v;as indicated for the urban group. Since the structural form equation for changes in the number of farm firms was estimated by ordinary least squares the coefficient of determination (R ) becomes a valid measure of the proportion of total variation explained by the predetermined variables in the system. The exogenous variables explained 80 percent of the total variation in farm numbers in the nonurban county group, 82 percent using all counties, and 93 percent using urban counties. 2 Examination of corrected R 's showed very little decline in explanatory power when the results were adjusted for degrees of freedom. While explanatory power was high for all groups a difference existed among the various groups of counties with respect to the statistical significance of the predetermined variables. Some of the variables that were important for one group of counties were not important for other groups. Mul t icol 1 inear i ty was never a serious problem. Very few simple correlations among the independent variables exceeded an absolute value of .5 and in no cases were these coefficients extremely large. Identification of the variables in which the simple correlation

PAGE 98

86 coefficients exceeded -5 will be made in later sections when each independent variable is discussed. Exogenous Shifters All predetermined variables included in Tables k through 6 were discussed in Chapter III. Each estimated coefficient reflects the effect of the shifter it represents according to the type of equation in which the coefficient is found. Education investments . — Changes in per capita education expenditures did not appear to be a very important variable affecting agricultural employment changes. Negative coefficients were found for all counties and the nonurban group of counties. The coefficients indicate that a one dollar increase in per pupil education expenditures was accompanied by changes in agricultural employment of the magnitude of the estimate. Changes in per pupil expenditures were smaller for the two groups having negative coefficients than for the urban group (Appendix C, Table 20). The urban group indicated a positive effect from education expenditures. A more skilled work force resulting from higher education levels would be expected to migrate to urban areas to realize their employment potential. This migration effect is supported by the negative coefficients for the nonurban counties. It appears possible that the migration effect was large enough to support a positive coefficient for the urban group. Education expenditures appeared most important In the nonurban counties upon examination of the standard errors of each coefficient. Somewhat different results were obtained for changes in the number of farm firmsIncreases in per capita education expenditures

PAGE 99

87 were significantly associated with increases in farm numbers for both the urban group of counties and the group consisting of all counties while a positive but nonsignificant coefficient v;as experienced in the nonurban group of counties. The opposite relationship was expected. Increases in per pupil education expenditures v/ould be expected to enable farm operators to attain more skills and take advantage of other employment alternatives in a given area. This would be particularly true with operators of smaller and more marginal farms. The strong positive relationship for the urban counties suggests one possible effect that may be occurring. Higher educational expenditures in the urban counties coupled with higher Incomes may encourage an increase in the number of small and part-time farms. This could occur even though the total number of farms might decline for a given area. Reduced form coefficient signs and magnitudes indicate the total effect on employment as a result of education expenditures. The effect was positive for all counties and for the urban counties group. For the nonurban group vjhere the direct negative effect on employment was largest and the positive effect on farm numbers was small the total effect vias negative. A correlation coefficient of -.50 between changes in per pupil education expenditures and technology change was noted in the urban group of counties. This must represent some form of spurious correlation. Increased technology levels would normally be associated with increases in educational skills. Corps of Engineers' investment s -1 nvestments by the Corps of Engineers demonstrated a negative effect on agricultural employment change for all groups except the nonurban group. A major portion of

PAGE 100

88 investments by the Corps in the four-state study area was for flood control. Effective flood control should make more land available for agricultural use. Expansion of existing farms to benefit from this should enable more efficient labor saving operations with resultant declines in agricultural employment. Displacement of existing farms by expansion would also contribute to employment declines. It appears that this occurred in the two groups having negative coefficients. Any employment increase due to new farms in the recipient area v/as offset by reduction in employment as the result of farm expansion. The standard error for the positive coefficient associated with the nonurban group was quite large. Average investment in the nonurban counties v;as approximately one-half that in the urban counties. The effect of a one thousand dollar change in investments on farm numbers was not significant for any of the three groups. Negative effects occurred for all counties and the nonurban group. Fewer farms would be expected if existing farms expand to take advantage of more flood protected land. Examination of the geographical pattern of Corps of Engineers' investments gives further indication as to why positive coefficients occurred for employment in the nonurban group and for farm number changes in the urban group. A large proportion of these investments occurred in the delta area of Mississippi and in an area in east-central Alabama which are predominately nonurban areas. These areas have traditionally been areas where large numbers of small farms were in existence. Flood protection provided by Corps of Engineers' projects could have made new land available which was suited for mechanized agricultural use. This undoubtedly created a movement to larger farms through both the effect

PAGE 101

89 of new farms and consolidation of the smaller farms into larger units. Increased agricultural activity as a result of the various project investments could have easily attributed to a fewer number of farms and larger number of agricultural employees in these areas as suggested by the coefficient signs. The standard errors for this variable given in Appendix B, Table 20, point out the wide variation of these investments among the counties in each group. Examination of the reduced form coefficients indicate that the total effect on agricultural employment was negative in all three areas as expected. This total effect takes into consideration the direct effect of the investments on agricultural employment and the indirect effect as felt through changes in the number of farms. Small Watershed Program i nvestments .-Investments by the Soil Conservation Service in the Public Law 566 Small Watershed Program yielded average investment levels per county that were fairly uniform in all three county groups (Appendix B, Table 20). This program also exhibited the lowest average level of investment per county of any of the land or water related investments. The nonurban and all counties groups both exhibited negative effects on employment for each thousand dollars of investment. Employment effects of this investment program were identical to those of Corps of Engineers' investments. Negative effects for these two groups indicate that farm expansion may be occurring in these areas with a concurrent reduction in agricultural employment. Increases in output as the result of this expansion may then explain the positive effect in the urban area where more processing employment and agricultural services employment becomes required. The major portion of PL-566 investments in Florida occurred along the lower

PAGE 102

90 West coast. Central Florida, the area immediately above Lake Okeechobee and in extreme V/est Florida where the areas of greatest growth have occurred during the past decade. These areas contain primarily urban counties. Similar observations on project location, but to a smaller degree, could be made about the remaining urban counties. This may have influenced the positive coefficient. As with Corps of Engineers investments the standard deviations for each group in the Small Watershed investment category point out the wide differences that exist on a county to county bas is . Positive coefficients were observed in the farm numbers equation for coefficients in the urban group and with all counties. The urban group coefficient demonstrated low statistical significance. Since one of the major stated purposes of this investment program is to prevent floodwater damage, it appears possible that previously flood-prone land became available for new farms in these groups. Declines in farm numbers occurred in the nonurban group. Farm expansion in this group may have offset any response from new farms as the result of the investment. The total effect on employment as demonstrated by the reduced form coefficients was similar in sign to the direct effect in all cases. That is, the effect on agricultural employment of PL-566 investments as felt through changes in the number of farms was not sufficient to influence a sign change from the direct effect. Small Watershed Program investments are similar to Corps of Engineers investments in that they are somewhat long term in nature although possibly not as long as the average Corps investment. Some effects of flood control structures and channelization projects may not yet be measurable. Some of the investments considered during the latter

PAGE 103

91 part of the study period represented the beginning of certain projectsOnly the effects of the construction phase may have been measured. These effects vvould be felt in other sectors such as construction. Since investments in the actual construction phase of any project began only as late as 1958 in the study area It is likely that some effects on agricultural employment and farm numbers had not occurred. Agricultural Conservation Program investments . — Program investments in the Agricultural Conservation Program provided the most uniform coverage over the four-state area of any of the investment programs analyzed. Only one county did not receive an investment and the average investment per county did not vary to a large degree among the three groups. Large levels of county variation were noted. All three groups demonstrated a negative effect on changes in employment as expected. Each group indicated very low standard errors when compared to the estimated coefficients. The negative effect on agricultural employment of a one thousand dollar increase in ACP program investments thus seemed quite important. Since this is a cost sharing program with farmers and is intended to introduce various conservation measures the negative effect on employment is logical in that measures taken with regard to land stabilization and resource improvement may have enabled the use of labor saving production practices. In some cases this program would remove land from production in further support of the negative effect. The effect on farm numbers was also negative in each group with all groups demonstrating a high level of significance. This supports the hypothesis that larger farms with the more responsive operators have taken advantage of the cost sharing program to improve their

PAGE 104

92 production practices. This may have enabled these operators to expand operations and consolidate smaller units into their ov/n operations since the marginal units are not as likely to be able to benefit from the cost sharing program. The total effect on employment is given by the reduced form coefficients. Since all coefficients were negative the total employment effect represents a larger negative effect than either the direct investment effect or the effect as felt through reductions in farm numbers . The importance of this variable is also indicated by its correlation with changes in the number of farms and alternative wage opportunities. Correlation with farm number change in the employment equation was -.68 for all counties, -.64 for the urban group, and -.69 for the nonurban group of counties. Correlation with alternative wage opportunities was -.56 for the nonurban group. Farmers Home Administration investments .-Investments by the Farmers Home Administration represented a somewhat different type of investment than those discussed previously. Those discussed earlier represented actual construction expenditures or payments made that were spent in the year of allocation. Investments by the Farmers Home Administration for community water and sewer programs and recreational facilities represented loans and grants made during the time period under study and Included additional grants for the projects made by other federal agencies. It is possible that some expenditures of project loans and grants made during the latter segments of the time period had not actually occurred. Data sources were inadequate to determine those expend I tures .

PAGE 105

93 Loans and grants made during this time period resulted in negative employment effects for all counties and the urban group. The size of the standard error of the coefficient for the all counties group indicated some importance for this variable. Investments in this program were fairly widespread over the four-state region with Mississippi receiving a large portion of the larger loans and grants particularly throughout the upper and central portions of the state. Alabama received heavy coverage with many of the grants being smaller in magnitude than those received in Mississippi. Coverage in Georgia and Florida was more isolated than in Mississippi and Alabama although most projects were larger when compared to those in Alabama, Some variation was detected among counties in this investment category. Heavy coverage in the nonurban counties might explain the positive coefficients demonstrated for this group. Availability of community water and sewage facilities could have influenced the location of agricultural related firms other than farms and thus agricultural employment. A high level of association was determined for changes In the number of farms in all areas. Employment effects were negative for each group. Community improvement through water and sewer facilities may have caused some movement of farm operators from the farm to the communities and other employment alternatives. These alternatives may have developed because of improved community service facilities. The effect on employment of farm number changes in the nonurban group was large enough to cause a total negative employment effect. The other groups had negative effects both directly and indirectly. Since both this program and the Agricultural Conservation Program are focused especially on agricultural areas and rural communities. It appears

PAGE 106

9^ likely that the general downward trend in agricultural employment and farm numbers played an important role in causing the large number of negative effects for these two investment programs. There was some negative correlation betv/een farm number changes and Farmers Home Administration loans and grants. The nonurban group had a correlation coefficient of -.56. Crop al lotment .--The effect of allotment reductions on agricultural employment was as expected as indicated by the positive coefficients In each of the three groups of counties. Positive coefficients for this variable indicate movement in the same direction since both the dependent variables and allotment reductions moved in a general downward trend. A small standard error was associated with the coefficients for all counties and for the nonurban group. This indicated some importance for allotment reductions in explaining employment decreases. Florida experienced large reductions in allotted acres of cotton, tobacco, and peanuts during the study period. Alabama experienced major reductions in peanut allotments while Georgia went through a period where major declines were felt primarily in tobacco and cotton. Mississippi experienced major declines in cotton allotments. Since these crops use large amounts of labor a negative employment effect is expected from allotment reductions. Some counties suffered allotment declines in excess of ^,000 acres. Large variations in allotment reductions among counties were observed as indicated in Appendix B, Table 20. Positive coefficients were also experienced for the effect of allotment reductions on the number of farms. A high level of association was observed for all counties with the nonurban group demonstrating a slightly lower level of association. This marked a consistent pattern

PAGE 107

95 of behavior since these groups are heavily weighted with nonurban counties where the major proportion of allotment reductions occurred. It becomes evident that reductions in allotments increased their market value leading to lov/er operator returns. The estimated coefficients then indicate that operator responsiveness has been in some cases to exit from agriculture due to the lower returns and the inability to secure allotments. Those farms most responsive to allotment reductions were most likely the marginal units and those having large acreages of allotted crops experiencing a decline. Consequently, operator responsiveness to exit from farming as a result of allotment reductions complemented the direct effect of these reductions on agricultural employment. The total negative effect on employment as the result of allotment declines is thus demonstrated in the reduced form coefficients to be greater than the initial direct effect. This added reduction in employment is due to the effect of the reduction in farm numbers. Agricultural product price . — Changes in product demand as measured by changes in the index of product prices did appear to be important in explaining employment changes. The nonurban counties and the group containing all counties demonstrated positive effects on employment as the result of increases in product prices. Increases in prices appear to have caused increases in product output with resultant increases in agricultural employment. Price effects for the urban group were negative. Agricultural employment would be expected to respond differently in the nonurban areas where it is more responsive to changes in product demand. Product price effects on farm numbers were positive in all three groups. Their effect on farm numbers seemed to be very important due to

PAGE 108

96 the high degree of significance of the coefficients. As product prices have increased it appears that the effect of new farms has outweighed any consolidation effect that would reduce farm numbers. The importance of product demand points out the responsiveness of agriculture to product price changes. Product price changes are probably more important to small marginal farms because of the small volume of products sold. This might explain the increase in farm numbers rather than a decrease as one might expect through farm consolidation. The total effect on employment determined by the reduced form equation showed no directional changes although the positive effect became much larger for all counties and the nonurban group of counties. Three of the seven commodity groups used to derive product price indexes showed a decline in the decade under study. Field crops, fruits and nuts, and poultry and poultry products showed declines while vegetables, forest products, dairy products, and livestock and livestock products experienced fairly substantial increases. The derived price index using the seven commodities demonstrated an average increase for all three groups with the urban counties demonstrating the largest price increase (Appendix B, Table 21). Agricultural wage rate .--Changes in factor prices as demonstrated by increases in the hired wage rate did not appear important in influencing employment changes. Wage increases did appear to cause employment increases in the urban counties. Higher wage rates were associated with employment declines in the other two groups. This indicates some substitution possibly occurred for the higher priced labor input.

PAGE 109

97 Mechanization is the normal substitute for labor employment in agriculture. More mechanization is also normally associated with increases in farm size. Wage rate increases v»/oulcl then be expected to cause declines in farm numbers. This particular circumstance was not indicated in the farm number equation as was expected. Wage rate coefficients were positive although no levels of acceptable significance were indicated. Hired wage rate increases might have encouraged enough entrepreneurs to enter farming to offset any consolidation effect occurring as a result of higher costs. This effect should be quite minimal. The total effect on employment after consideration of both the direct wage increase effect on employment and the effect through farm number changes was quite similar to the initial direct employment effect. Substantial amounts of variation v;ere noted among counties for changes in hired wage rates. Agricultural technology . — Output per man-hour increases for agriculture appeared to increase agricultural employment for all counties and the nonurban group. These effects seemed important upon examination of the standard errors for these two coefficients. Technology increases apparently were responsible for the use of more labor employment. This may have resulted from large output increases which in turn may have resulted in agricultural employment increases in those agricultural industries servicing the farms. Many commodities grown in the four-state study area are labor intensive. Technology increases in these commodities could have easily been responsible for increases in agricultural employment . A somewhat different pattern emerged with respect to farm number changes and output per man-hour. Technology advancements caused a decline

PAGE 110

98 in farm numbers for all three groups with low significance levels associated with all counties and the nonurban counties. It was anticipated that output per man-hour increases would reduce farm numbers through two effectsOne is the ability of larger farms to take advantage of technological innovations and both increase output and farm size and the second, given that the demand for agricultural commodities is inelastic, output increases by the larger producers may affect immediate prices to such degree that smaller farms are forced to consolidate. The total effect on employment after consideration of the farm number changes did not alter the direct technology effect. This total effect on employment further substantiates the earlier observation that any reductions of onfarm employment due to technology are more than offset by increases in on-farm and off-farm agricultural employment required to handle increased output in certain commodities. Changes in output per man-hour among the various commodity groupings varied substantially. Those representing livestock and livestock products experienced the largest increases. Output indexes for all livestock and livestock products, dairy, and poultry doubled in the ten-year period studied. All field crops indicated increases of approximately 50 percent while increases for vegetables and fruits and nuts were very small. Consequently, the value of each county's major commodities demonstrated major influence on the index for that county. Average change for all three groups was fairly uniform. Variation among counties within each group was also fairly small. Agricultural viage opportuni ty .-I ncreases in the opportunity cost of remaining as a farm operator should influence operators making a low return on their farming investment to seek a higher wage

PAGE 111

99 alternative. Changes in nonagr icul tural wages did produce this result for all groups except the nonurban group although the effect on farm numbers of this change did seem Important. The positive effect for the nonurban group may be explained by the relative lack of alternative employ:r:3nt opportunities in the nonurban areas. Average increases in opportunity wages were fairly consistent for all groups with changes in the nonurban group the smallest. The total effect on agricultural employment as felt through changes in the number of farm firms v/as similar to the effect on farm numbers. Agricultural employment opportunity . — Changes in nonagr icul tural employment opportunities established much the same pattern as wage opportunities in affecting farm numbers and employment. Negative effects on farm numbers v/ere observed only for the group containing all counties. The nonurban group of counties did indicate a low level of significance. Off-farm employment opportunities v;ere most available in the urban counties (Appendix B, Table 25) where an increase in off-farm employment suggested that an increase in farm numbers would occur. This may be reflected from an Increase in part-time farmers and the increasing availability of part-time jobs for marginal operators which prohibited full exit from farming. The nonurban area demonstrated the lowest average increase in off-farm employment opportunities yet indicated that an increase in farm numbers occurred as the result of the employment opportunities. Consideration of the effect of employment opportunities on agricultural employment yielded the same directional effect in each group as v;as determined for changes in the number of farms.

PAGE 112

100 Farm operator age . --Changes in the age of farm operators was highly associated v/i th changes in the number of farms. Reductions in the number of operators v;ho were 55 years of age and older occurred in each grouping of counties (Appendix B, Table 25). Decreases in the number of farms as the result of decreases in the number of older farm operators give further support to the observation that many farms are marginal. Farm consolidation is occurring rather than each older operator through retirement or death being replaced with another operator on a one-to-one basis. The total effect on agricultural employment is the same. Coefficients for each group in the reduced form equation indicated that reductions in the number of older operators reduced total agricultural employment through farm number reductions. Group differences . — A zero-one intercept shifter was included in the group consisting of all counties to determine if differences existed between the urban and nonurban counties within the groupings. The nonurban group was used as the base group. Group differences did not appear to be very important in the farm number equation. Differences between the urban and nonurban counties did exist for changes in agricultural employment since the coefficient for this variable was approximately five times the magnitude of the standard error. Construct ion Parameter estimates in each construction employment equation for each of the three groups are given in Table 7This equation is identical in theoretical construction to equation (3.9-2.1) and differs from those used for agriculture and manufacturing in terms of included variables. These equations were estimated by ordinary least squares.

PAGE 113

Table ~j . Regression equations for construction employment change (E„) for all counties, urban counties, and nonurban counties, i960 to 1570^ , J J ^ • ui b All Urban Nonurban Independent varjables counties counties counties Constant -279-60 -240,70 127.10 Education (X.) CE (X^) PL-566 (X^) ACP (X^) FHA (Xc) Wages (FP^) Wage opportunity (WW ) Employment opportunity (WE ) Dummy (GRP) 2 R^ R ;2 Number of observations .9453* (.5175)

PAGE 114

102 Dependent Variable The change in the number of construction employees in each county between .I960 and 1970 was used as the dependent variable. Changes in construction employment demonstrated the largest average increases in the urban group. Construction employment increase in the nonurban counties was about one-eighth the size of the increase in the urban counties. Variation among counties within groups was large as indicated by the standard deviations shown in Appendix B, Table 18. The amount of explained variation was not particularly high for construction employment. The urban counties exhibited the largest amount of variation explained (45 percent). Corrected coefficients of determination were not substantially different. Exogenous Shifters Lack of data prevented the use of the product price, technology, and number of firms variables in the construction equation. These variables were fairly important in agriculture and some of the manufacturing equations discussed later. Their omission probably contributed to the low percentages of total variation explained. These three variables should have been very important in the construction industry during the past decade. Education i nvestments .--Changes in per pupil education expenditures demonstrated significance for the all county group and the nonurban group while also exhibiting opposite signs. The positive effect shown for the urban group and all county group is an indication that a higher skilled work force is contributing to the construction industry. This would occur through the use of more skilled employees within the

PAGE 115

103 industry as well as through growth in other industries that the construction industry servesNegative effects for the nonurban counties support the hypothesis offered for agriculture that education expenditures influence outmigrat ion for the nonurban counties with the resultant decrease in construction employment. Corps of Engineers' investments -I nvestments by the Corps of Engineers should be very significant in stimulating construction employment in a given area due to the large size of many projects undertal
PAGE 116

104 commuting work force rather than the employment of local workers. Employment of this nature would not be reported in the county where the project was located. Any secondary construction employment generated in the local project area as the result of the initial project expenditure was either of minimal importance or not yet experienced during the study period. Some knowledge of project expenditures in Florida indicate that use of contractors outside the project area occurred. Agricultural Conservation Program investments . — This program includes conservation practices such as planting permanent land cover, trees and other similar land stabilization measures as well as land leveling, reservoir construction, drainageways , pipelines, terracing and the installment of sediment control structures. Since it is done on a cost share basis the participating land owners and farm operators probably are able to implement some of the practices using their own equipment. Larger measures would undoubtedly require the use of the heavy construction industry. The Indicated employment effects of this program on construction were negative for both the all county and urban county groups with low levels of significance indicated. Since this program operates on a relatively small project basis it appears likely that any contract work done is performed by contractors based outside the county of project implementation. Larger farm operators with special equipment also do custom construction work. No increases in construction employment would be felt if this occurs. Farmers Home Administration i nves tments .--Local water and sewer development projects should also be expected to create employment in the

PAGE 117

105 construction industry. Hov/ever, a negative impact was demonstrated for all three groups with the nonurban group indicating some significance. Since most of these projects are located in rural communities, the same comments about outside contractors must be made. The initial construction employment impact is not felt in the county or community receiving the sewer or water system. Any economic growth or development generated in the system was not substantial or the time period was too short to capture the effects. Construction wage rate . --Wage increases in any given industry represent an increase in factor prices to that industry and would be expected to decrease employment. This would occur only if other factors could be substituted for labor. Wage increases in the construction industry were, very important in positively affecting employment as demonstrated by high significance levels for all three county groups. Large amounts of labor used in construction activity coupled with the apparent inability to substitute other production factors for labor are two important reasons why a positive effect was demonstrated. Labor could also be a critical factor in the construction industry. With a downward sloping demand function and an upward sloping supply function, a shift in demand for construction labor would result in increased wages and greater employment. Construction wage opportuni ty •-I ncreases in opportunity wages were highly significant with the negative sign as expected in all three groups. It is apparent that the opportunity cost of remaining employed in the construction sector became quite large for many workers who transferred their sl
PAGE 118

106 Average change in nonagr icul tural wages per construction worker was fairly consistent except for the urban counties where the average was slightly less than one-half that of the other two groups. A correlation coefficient of -.5^ was noticed between wage opportunity and the dummy variable for group differences. Construction employment opportuni ty .-I ni t ial a priori expectations indicate that employment opportunities in alternative industries would decrease employment in any given industry. This did not occur in construction where all three groups demonstrated a significant positive effect on construction employment from increases in alternative employment opportunities. All groups demonstrated levels of significance with the urban counties demonstrating the highest. This positive effect again points out the complementary effect of this industry. That is, as the economy of any area expands through the influence of any sector, accompanying expansion must be felt through the construction sector. increases in employment opportunities thus require increases In construction employment. Average values for these variables were fairly consistent in all three groups. Group differences . — Differences between the urban and nonurban counties for the all county groupwere indicated with a high significance level. This variable indicates a much higher level of construction activity for the urban counties. Nondurable Manufacturing Analysis of the equation used for manufacturing pointed out several distinctive patterns and differences between the nondurable and

PAGE 119

107 durable manufacturing IndustriesBecause of these differences discussion of the manufacturing results is broken into two sections with each equation estimated identical to (S'S-k.l) in theory. Industry three (textile mill products and other fabricated textile mill products) and industry four (food and kindred products) were the two individual nondurable manufacturing industries included. In addition, data from the chemicals and allied products industry were included in the calculation of variables used for the nondurable industry (ND) . This equation represents the total effect of each exogenous shifter on nondurable manufacturing employment from these industries. Tables 8, 9, and 10 give the nondurable manufacturing equations and parameter estimates. All equations were estimated with ordinary least squares . Exogenous variables explained as much as 83 percent of the total variation in the urban counties in the textiles equation. Explanatory power for all counties was 72 percent with 32 percent of the variation explained for the nonurban counties. Explanatory power fell substantially in the food and kindred products equation where only kO percent of the variation was explained in the urban group. Explanatory power was quite low for all counties and in the nonurban group at 28 percent and 10 percent, respectively. Explanatory power for the total nondurable equation was betv^een that of the individual industry equations. Corrected coefficients of determination were not substantially different. Dependent Variables Average employment change patterns between industries varied quite substantially. Average employment change for the textiles and other fabricated textile products was quite consistent in all three

PAGE 120

108 Table 8. Regression equations for textile mill products and other fabricated textile products employment change (Eo) for all counties, urban counties, and nonurban counties, I96D to 1970^ Independent variables

PAGE 121

109 Table 9. Regression equations for food and kindred products employment change (E/^) for all counties, urban counties, and nonurban counties, I96O to 1970^ Independent variables All counties Urban counties Nonurban counties Constant 269.80 277.60 82.52 Education (X,) CE (X^) PL-566 (X^) ACP (X^) FHA (X ) Product price (PP.) Wages (FP.) Technology (Z.) Wage opportunity (WWM) Employment opportunity (WE,) Number of firms (N,) Dummy (GRP) r2 -.1210 (.1447)

PAGE 122

110 Table 10. Regression equations for nondurable manufacturing employment change (Ej^rj) for all counties, urban counties, and nonurban counties, '156O to 1970^ Independent variables

PAGE 123

in groups (Appendix B, Table 18) . However, standard deviations were quite large which indicates isolated county employment patterns for textiles and fabricated textile productsAverage employment change for food and kindred products was negative for all county groups except the nonurban counties. The urban counties showed a decline of substantial magnitude when compared to the other areas (Appendix B, Table 18) . Comparison of the standard deviations also indicated some variation among counties. Total employment change for the nondurable industries group also varied among the delineated groups. The nonurban group of counties demonstrated the largest average increase while the urban group experienced the smallest increase of slightly more than one-third of the nonurban increase (Appendix B, Table I8) . Large variations between counties were indicated for the urban counties. Exogenous Shifters Sufficient data were available on all the nondurable industries to include all independent variables. Each equation discussed is identical in theoretical construction and variables included. Education investments . — Changes in per pupil education expenditures demonstrated a negative effect in the three nondurable equations for all groups. A low level of association was noted for the group of all counties in both the textiles and nondurable equations. Nondurable manufacturing industries employ a high proportion of nonskilled laborers since many of the jobs are in processing plants and mill type work. Traditionally, the typesof firms In these industries, particularly in the textile industry, have attempted to locate and relocate in areas

PAGE 124

112 where a large quantity of low-co5t labor was available. This observation might explain two apparent trends indicated in nondurable employment patterns. First, increases in education expenditures have in general contributed negatively to employment changes in nondurables which might indicate the migration of a more highly educated v/ork force out of these particular Industries and the areas in which the industries are located. Second, the average increase in nondurable employment for nonurban counties was larger than the other two areas indicating a possible movement toward a lower cost work forceThese should be offsetting movements depending on the rate of education expenditures and the number of nondurable firms in nonurban areas. Corps of Engineers' investments .--The effects of Corps of Engineers' investments were consistent along industry lines rather than county groups. Investments by the Corps indicated a negative effect in all three groups for food and kindred products with high significant relationships demonstrated in the group of all counties and the urban county group. Positive effects were indicated for textiles and the total nondurable equation in all groups. A significant effect was demonstrated for all counties and the urban counties for the nondurable equation. The two groups demonstrating significant relationships received the largest average investments. Consistency of coefficient signs for these equations seems to indicate consistent differences in employment patterns for the various industries. A negative effect on agricultural employment and farm numbers by Corps of Engineers' investments was demonstrated in an earlier section. It is of interest that the same negative effect is indicated for food and kindred products manufacturing since this industry is highly

PAGE 125

113 dependent on agriculture. It appears that firm consolidation may also be occurring in food and kindred products with resultant decreases in employment. Nondurable manufacturing in general apparently increased as the result of Corps of Engineers' investments. Nondurable industries use large amounts of water and should benefit as indicated from increased supplies of raw products made available from flood protection. Small Watershed Program inves tments .-I nvestments in the Small Watershed Program indicated a consistent positive effect for all three groups in the three nondurable equations. The only negative coefficient appeared in the food and kindred products equation for the nonurban counties. None of the coefficients showed a significant level of association. Investments in this program would be expected to influence nondurable manufacturing employment through secondary and tertiary effects. That is, enough time must elapse for a program investment to increase the output of fiber and raw food products in an area so that the processing of these additional products will increase employment. It is apparent that either this effect did not occur or that inadequate time passed to measure the full effect. It also remains possible that processing occurred outside the investment county and the total effect was not measured. Agricultural Conservation Program investments .--Conservation measures available through this program would be expected to have similar effects on manufacturing employment as the Small Watershed Program. Some time delay would be involved before conservation measures could influence food and fiber output of sufficient degree to appreciably affect nondurable manufacturing employment. In consideration of these

PAGE 126

factors, coefficient signs and association levels were consistent with those experienced with the Small Watershed Program. All effects were positive with the exception of the urban counties for the total nondurable equation and for the nonurban counties in the food and kindred products equation. The only sign difference between this investment category and the Small Watershed Program was the negative sign found in the urban group. No significant levels of association were determined. Farmers Ho me Administration investments .--Communi ty water and sewer system investments could conceivably influence manufacturing firms to locate in a particular community because of the advantages these systems might afford both the firm and its employees. Time delay problems in measuring these responses to an initial investment also exist with this factor supply category. Response to water and sewer system loans and grants was generally positive with several exceptions. Negative effects were noticed for food and kindred products in the urban counties and for textiles in the nonurban counties. The pattern of no significant coefficients evident in the last two investment categories also occurred with water and sewer system grants. Manufacturing product price . — Increases in product price would normally be expected to increase employment. Existing firms would be expected to increase output and new firms v/ould be encouraged to enter the industry in an attempt to benefit from higher residual returns to firms. Too many new firms could cause residual returns to decrease with an offsetting reduction in firm numbers and employment. This offsetting decline could be more substantial than the initial increase. All product price indexes used I958 = 100 as the base.

PAGE 127

115 It appears likely that the indirect effect on new firms, i.e., a reduction in firm numbers and employment due to reduced returns as the result of expanded output, was large enough to offset the initial positive effect. All product price increases in the nondurable equations resulted in negative employment effects in all three groups. High levels of association were indicated for food and kindred products for the group of all counties and the urban county group. A correlation coefficient of .5 was noted between manufacturing product price and manufacturing technology in the urban group. Average product price increases were fairly consistent in all three groups for both textiles and food and kindred products. There was slightly more variation in the urban group. tntracounty variation within each group was fairly small. Product price changes in the textile and fabricated textile products industries ranged from fairly substantial increases to some minor decreases. Textile mill product prices rose 8.9 points while fabricated textile product prices increased by 1^.^. The largest increase in textiles of 78.3 occurred in narrow fabric mills products while decreases of 7-5 occurred in both knitting mills products and floor coverings. Fairly uniform increases of ]k.k occurred in the fabricated textile industry with only a few products showing a decrease of 11.0. Food and kindred product prices showed a larger increase in the price index changing by 24.2 points. Within the industry, dairy products experienced the largest increase of 31-5 while miscellaneous food products increased the least by 17.3 index points. Average price change for all nondurables was 21.9 index points. Chemical product prices which

PAGE 128

116 were included in the total nondurable computation experienced both increases and decreases among its products. Manufacturing wage rate . — Wage increases would affect employment in either of two ways. For industries where other factors could not be substituted for labor in the production process, wage increases would be associated with increases in labor employment. Conversely, in industries where labor was replaceable by other factors, v;age increases would cause a decline in employment. Wage increases could reduce firm residual returns to such degree that the firm would relocate or exit from the industry. The effect of wage increases was negative in all three groups for both textiles and the total nondurable equation. Negative effects for food and kindred products v/ere experienced in the nonurban group. None of the wage effects exhibited significant levels of association. Average wage changes did vary somewhat among the three groups. The urban counties demonstrated the largest wage increases for all three nondurable categories. Wage increases for the three industries were consistent and county variation within groups minimal (Appendix B, Table 22) . Manufacturing technology . — Technology increases would generally be expected to reduce employment in a given area. The expected effect of increases in output per man-hour through the use of new equipment and/or different production techniques would be a decrease in the demand for labor and consequently a negative effect on employment. Negative effects did occur for textiles in all three groups although none of the coefficients were significant. A somewhat different pattern emerged for food and kindred products and all nondurable products. Positive

PAGE 129

117 employment effects were demonstrated for food and kindred products in the group of all counties and the urban group. The urban group coefficient indicated a lov; level of significance. Positive coefficients v;ere indicated for all nondurables in the group consisting of all counties and the nonurban group. It seems apparent that in food and kindred products and in all nondurable products that technology increases may have been labor increasing. Output expansions were sufficient to overcome any labor decreasing effects of the technology. 2 Technology or output per man-hour index average increases were fairly uniform among groups for both textiles and food and kindred products. Hov;ever, the increase was about three times larger for food and kindred products than for textiles (Appendix B, Table 23). For all nondurables, the urban group of counties showed a larger increase than either textiles or food and kindred products. More county variation was indicated for textiles and all nondurables than for food products where intracounty variation v;as quite small. Changes in the output per man-hour indexes for three-digit manufacturing industries which were used to compute county indexes did vary substantially in some cases. Changes in textile mill products indicated a decrease in the wool weaving and finishing mill products index of 3^ points while all other industries exhibited increases with the maximum increase of 55 in floor covering mills products. An interesting observation here ts that this particular industry suffered price decreases during the same period. Apparel and fabricated textiles had three decreases with the maximum decrease of 41.5 points occurring in women's 2 Indexes developed used 1958 = 100 as base.

PAGE 130

118 and misses' outerwear. A maximum increase of almost 76 points occurred in hats, caps, and miliinery products. Food and kindred products experienced increases in all industries with the maximum of 53-5 occurring in beverages. Chemicals which viere included with other nondurables for the total nondurable equation exhibited substantial increases in all industries with the maximum of 104.7 occurring in gum and wood chemicals. Manufacturing wage opportuni ty . --Changes in wage opportunities demonstrated a varying pattern of effects both among groups and nondurable industries. A negative effect on employment v;as indicated for all counties and the nonurban group for textiles with the latter group demonstrating a significant level of association. Since average wage changes in this industry were relatively lower than in other industries, it seems apparent that the opportunity cost of remaining in textile employment was large enough to induce employees to take advantage of alternative wage opportunities. A positive effect was demonstrated in all groups for food and kindred products although no significant levels of association vyere apparent. Alternative wages may not have been high enough to entice labor employment to leave the industry. Since this is a low-skilled industry, it is more likely that employment skills were not sufficient for employees to leave this industry and transfer their skills to another industry. The pattern emerging for all nondurables was negative with the exception of the urban group. It appears that for nondurable employment in general the opportunity cost of remaining employed was large enough to influence some degree of movement to other job alternatives.

PAGE 131

119 Manufacturing employment opportunity . — The effect of increases in employment opportunities v;as mixed with no significant levels of association noted. Negative effects were indicated for textile products employment in the nonurban group and for all nondurables in the all county group and urban group. Food and kindred products had positive effects in all groups. This same relationship was indicated for the wage opportunity variable giving further support to the hypothesis that skills in this industry are prohibitive in the movement of labor employment from this industry. The negative coefficients mentioned did indicate that some movement in labor employment was occurring. Average employment opportunity increases were by far the largest for textiles in the urban group with almost 190 new jobs per textile worker. Rates for the all county group and nonurban group were slightly less than one-half and one-fourth that of the urban group, respectively. Job opportunities for food products were fairly consistent among all three groups. Increases in job opportunities for all nondurables were lower than those of the individual industries as would be expected. Group averages were fairly consistent in size (Appendix B, Table 25). Number of firms . — Increases in the number of firms would be expected to increase employment in that industry. It is possible that firm consolidation to a fev/er number of larger size firms could indicate a reduction in firm numbers and yet an increase in industry employment. It was apparent for nondurables both individually and in total that firm number changes and employment changes moved in the same direction. The effect of changes in firm numbers was positive and highly associated with employment changes for each group in all three nondurable equations.

PAGE 132

120 Average change in textile firm numbers ranged from slightly greater than one in the nonurban group to slightly less than two and one-half in the urban counties with the average change in employment following basically the same pattern. Average change in firm numbers for food and kindred products was negative for all groups with the urban counties demonstrating the largest decrease. These two groups experienced the largest decreases in food products employmentThe smallest decrease in firm numbers occurred in the nonurban group. This group was the only one experiencing a small employment increase. This indicates a movement to larger firms in the nonurban counties. Positive coefficients for this industry indicate similar directional movements in firm numbers and employment since both have been decreasing. It appears likely that for the nonurban group firm number decreases have not been great enough to cause an average decrease in employment. Small increases in the number of total nondurable firms were indicated for all three groups (Appendix B, Table 19). The positive effect of firm numbers on employment would be expected since total nondurable employment generally increased for all groups. There were large variations among counties in firm number changes. Group differences . — A significant difference between the urban and nonurban groups was indicated for both textiles and food and kindred products. No significant difference was indicated for the total nondurable equation. Grouping of the two individual industries and chemicals to construct the total nondurable equation aggregated out of existence individual industry differences between the urban and nonurban groups. Average employment change for all three industries was smaller in the urban group since a negative coefficient was determined for this variable.

PAGE 133

121 Durable Manufacturing Three individual industries and the total durable manufacturing industry were included for analysis. Industry five was the transportation equipment industry. Lumber and wood products and furniture and fixtures comprised industry six while electrical equipment and supplies comprised industry seven. Four additional industries were combined with these to constitute the total durable manufacturing equation. These four industries were paper and allied products, stone, clay, and glass products, primary metal products, and fabricated metal products. Each equation was identical to equation (3-9-k.l) in theoretical construction with a slight exception in industries five and seven. Product price data for transportation and electrical equipment were inadequate for the inclusion of this variable in the equations. This differentiation should be noted as general comparisons of these equations are made with the other two equations. Estimated coefficients for the four durable equations are given in Tables 11, 12, 13, and 14 with all equations estimated by ordinary least squares. The exogenous variables explained as high as 24 percent of the total variation in employment change for the transportation equipment equation for all counties. Explanatory power fell in the other two groups with 22 percent explained for the urban counties and 17 percent for the nonurban group. Explanatory power was even less for industry six where maximum percentage explained of 22 percent occurred for the urban group. Explanatory power was relatively higher for the remaining two groups. Fifty-five percent was explained in the all county group for transportation equipment. The urban county group experienced explanatory power of 54 percent with the nonurban group at 30 percent. The urban

PAGE 134

122 Table 11. Regression equations for transportation equipment products employment change (Er) ^or all counties, urban counties, and nonurban counties, I960 to 1970 Independent

PAGE 135

123 Table 12. Regression equations for furniture and lumber and wood products employment change (Eg) for all counties, urban counties, and nonurban counties, I96O to 1970^ Independent

PAGE 136

124 Table 13. Regression equations for electrical equipment products employment change (Ey) for all counties, urban counties, and nonurban counties, I960 to 1970^ 1 ndepsndent

PAGE 137

125 Table 1^. Regression equations for durable manufacturing employment change (Eq) for all counties, urban counties, and nonurban counties, I96O to 1970^ Independent variables

PAGE 138

126 county group had the largest amount of variation explained for all durables at 39 percent. The all county group was next at 28 percent while the nonurban group was I6 percent. Dependent Variables Substantial variation was indicated among groups in average employment change for the various durable industries. Transportation equipment average employment change was the largest in the urban counties with a change of over 600 employees. Average change for all counties was slightly over 200 employees while the nonurban groups change was onesixth that of the urban group. Average employment change for lumber and wood products and furniture and fixtures was negative for all three groups with very little variation among groups in magnitude of change (Appendix B, Table 18) . Electrical equipment average employment change also varied substantially with the urban counties showing the largest increase of 370 employees. The smallest increase of 50 employees occurred in the nonurban counties. Total durable manufacturing average employment change was the largest in the urban group with a change of 1,059 employees. Exogenous Shifters Exogenous shifters included in the durable equations were identical in method of derivation to those used in the nondurable equations. Product price was not included in transportation equipment and electrical machinery equations because of inadequate data. Educat ion i nvestments .-I ncreases in per pupil education expenditures demonstrated a consistently positive effect on employment with only two exceptions. These occurred in the nonurban groups for both trans-

PAGE 139

127 portation equipment and all durable manufacturing. The only significant level of association occurred in lumber products employment for the group containing all counties. The effect demonstrated for durables v;as generally opposite that experienced for nondurables. It appears evident that durable manufacturing industries benefit directly from a more educated and higher skilled work force as opposed to nondurables. Even though employment decreased in lumber and furniture the education effect was positive. Corps of Engineers i nves tments .-I nves tments by the Corps of Engineers produced a more varied response for durable manufacturing than was noticed for nondurable. Positive and significant effects were indicated for durable manufacturing in both the all county group and urban counties. No other significant levels of association v/ere obtained. Negative coefficients were indicated for electrical equipment manufacturing employment for both the all county group and the urban group. Increases In water supplies, protection from flooding, and improved transportation are all products of Corps of Engineers investments. These improvements should increase manufacturing activity in a given area. The mixed response for durable manufacturing employment indicates that the effect Is highly dependent on the industry considered and area In which the industry Is located. Small Watershed Program investments . — Investments in the Small Watershed Program also gave a somewhat contrasting effect in durables as opposed to nondurables. Most groups demonstrated a generally negative effect as compared to the positive effect demonstrated In nondurables The only group showing a significant level of association was the nonurban group for durable manufacturing. This occurred despite relatively

PAGE 140

a e 128 small employment changes for durable manufacturing in the nonurban group. The negative coefficents indicate that possible improvements in factors of production in an area lead to intensification of agricultural activity rather than manufacturing activity. investments of these types v;ould also be expected to influence nondurable manufacturing employment to a larger extent than durable manufacturing employment. The large numbers of positive employment effects for nondurable manufacturing as compared to durable manufacturing employment support this hypothesis. Agricultural Conservation Program i nves tments .--The same general pattern emerged with investments in this program for durables as with the Small Watershed Program. The majority of coefficient signs indicated negative effect on employment in contrast to the generally positive ffect indicated for nondurable manufacturing. No significant levels of association were noted. Agricultural Conservation Program investments would also be expected to influence both agricultural activity and nondurable manufacturing activity in a given area at the expense of durable manufacturing activity. A negative effect v/as indicated for transportation equipment employment for all groups. Lumber and wood products also exhibited a negative effect for all three groups. The nonurban group demonstrated negative effects for electrical equipment. Total durable employment was affected negatively for all groups except the nonurban group. Farmers Home Administration investments . — Loans and grants made by the Farmers Home Administration seemed to be more important in durable manufacturing than in nondurable manufacturing employment. The effect on transportation equipment employment was negative for all groups with a low level of significance determined for the nonurban group. High

PAGE 141

129 levels of association viere indicated in all groups for the lumber and furniture products industry v^/ith all employment effects positive. Lumber and wood products and furniture manufacturing are usually located in rural areas near the raw product source. Investments in community water and sewer systems which are often made in rural areas seemed to have been important in influencing employment in this industry. Although this was a generally declining industry in terms of people employed, loans and grants for water and sewer systems may have been important in increasing local area employment. Positive effects were demonstrated for electrical equipment employment for all groups except the urban group with a high level of association noted for the nonurban group. Total durables demonstrated a pattern somewhat similar to transportation in that all effects were negative except for a significant positive effect in the nonurban group. The importance of this program for the nonurban counties seemed to be consistent for all durable industries, although the nonurban counties as a group did not receive average loans and grants as large as some of the other groups. Only one industry indicated a negative effect. All four industries examined did exhibit some level of significance in the nonurban group of counties. Since individual counties in this group received substantial Investments it seems likely that they influenced the entire group. This program was insignificant for the nonurban group in the nondurable employment equation. This program has been in effect longer in many of the counties which constitute the nonurban group than for the urban counties. Manufacturing product price . — Product price increases seemed to influence employment in durable manufacturing similarly to the effect in

PAGE 142

130 nondurables . The transportation and electrical equipment equations did not have the product price variable included. Price increases in the lumber and wood products industry demonstrated a highly significant negative effect for both the all county group and urban county group. The nonurban group also showed a negative effect. Average product price increases were fairly consistent among groups. Total durables also demonstrated negative effects in all groups with very high levels of association indicated for all counties and the nonurban group. There were some differences in average product price increases for all durables with the nonurban counties showing the largest change 3 of 25.3 points and the urban counties a change of 16. 5 points. The other three groups varied between changes of approximately 20 to 23 index points (Appendix B, Table 21). Intracounty variation was quite low. Changes in the product price index for lumber and furniture products were fairly consistent with the group averages varying from iG.h to 29.2. Price changes in the lumber and wood products industry did vary somewhat within the industry. Plywood, millwork, and wooden containers experienced the smallest change of 16. 5 while the raw products of logging contractors experienced the largest change of 37-9 index points. Price changes in furniture and fixtures were fairly consistent with an increase of slightly greater than 2k in all three-digit level industries except for miscellaneous furniture and fixtures which increased 37.^ index points. Price changes for the remaining four industries used in determining price changes for all durables were fairly consistent except for paper and allied products. This industry showed a price All product price indexes used I958-IOO as the base.

PAGE 143

131 decrease for building paper and paper board and fairly small increases in several other industries. Only papermill products, excluding building paper, demonstrated an increase on the same magnitude as the other three tv/o-digit level industries. Stone, clay, and glass, and both fabricated and primary metals showed price increases generally around 20 index pointsProduct price increases were correlated with technology changes in the all county group. The level of correlation was -.53Manufacturing wage rate .--The effect of wage changes on employment was consistent along industry lines for durables as it was for nondurables. Wage increases seemed very important in influencing employment changes for transportation equipment manufacturing. All three groups demonstrated significant positive effects. Even though wages did increase substantially in transportation when compared to other industries, particularly nondurables, it is apparent that labor could not be replaced by other production factors. The effect on lumber and furniture products employment was negative for all groups except the urban group. Average wage increases in this industry were consistent with those in the other durable industries. This indicates that labor employment has been replaced by other resources as the result of the higher wages. Positive employment effects were demonstrated for both electrical equipment and all durables for all three groups. Levels of significance were indicated in the group containing all counties for electrical equipment and in the all county group and urban group for all durables. It seems evident that v.age increases were not sufficient to cause labor replacement or that labor could not be replaced in the physical production process for these two industries.

PAGE 144

132 Average wage Increases for durables were generally larger than those in nondurables. Electrical equipment wages changed the greatest amount. Average annual wage change for electrical equipment in the urban group was approximately $4,270. This was much higher than any other group. Changes in nondurables normally ranged from approximately $2,100 to $2,800 with the larger changes usually occurring in the urban group of counties. Intracounty variation was normally small. Manufactur ing technology .--Changes in technology gave the expected effect in each durable equation. Increases in technology contributed to decreases in employment for all three groups in the transportation equation with a low level of association noted for the group consisting of all counties. A more definite effect was demonstrated for lumber and furniture products where all groups demonstrated highly significant negative effects. This industry suffered a decline in average employment for each group in the decade under study with the apparent addition of either output increasing and/or labor input decreasing capabilities to their production processes. This consistent negative effect also substantiates earlier observations about the apparent effect of wages and prices on employment in this industry. The remaining two industries were consistent in sign with a priori expectations but did not demonstrate significant levels of association. Average changes in output per man-hour indexes were fairly consistent within each industry. Average technology changes in the Indexes developed used 1958=100 as base.

PAGE 145

133 transportation equipment industry were extremely consistent among groups with a change of about l3 index points. Substantial index variation was seen v/ithin the industry. Changes in the index varied from slightly less than six index points in railroad equipment to 36 points in motorcycles and bicycles. Average technology changes for lumber and furniture manufacturing were approximately 25 index points for most of the groups. Intraindustry variation was substantial as lumber and wood products had three-digit industries experiencing changes as low as six points (miscellaneous wood products) and as high as 67-6 points (wooden containers). Furniture and fixtures had one industry that declined slightly over three points (office furniture) while increases in the remaining industries were rather low with miscellaneous furniture showing the largest increase of slightly less than 16 points. Electrical equipment output per man-hour changes also were erratic with changes ranging from a decline of almost 2k points in radio and television receiving equipment to an increase of almost 37 in household appliances. The remaining industries which were used for total durables also exhibited substantial variation. Paper and allied products had one major increase of slightly over 100 points in building paper and paperboard with the remaining industries increasing approximately one-third this much. Stone, clay, and glass exhibited increases ranging from less than one in products of purchased glass to over 55 in hydraulic cement. Primary metals had two industries that experienced decreases in the technology index with the largest decrease occurring in nonferrous foundries products. The largest increase of over k3 points occurred in primary nonferrous metal products. Fabricated metals had the largest of three decreases to occur in metal stampings where 16.4 points were

PAGE 146

13^ lost. The largest increase of 56.2 points occurred in nonelectrical plumbing and heating equipment. This increase v/as approximately double that of any other industry within fabricated metals. Much of this variation was removed v/hen county changes were computed according to their value of output v;i th intracounty variation within each of the three groups not excessively larger. Manufacturing wage opportuni ty . --Effects on employment of increasing wage opportuni ties were very consistent along industry lines for durable industries. The effect on the transportation industry of changes in wages in other industries was negative for all groups with high levels of significance noted for each group. This consistently negative effect makes it apparent that the opportunity cost of staying in textiles was large enough to encourage employees to take advantage of higher wage alternatives. The effect in lumber and furniture was positive for all groups. No significiant coefficients v;ere observed. Low-skilled labor employment in this industry may not have been able to take advantage of alternative employment opportunities. Wage opportunity effects for electrical equipment were consistently negative for all groups but did not seem to be significantly important in explainiing employment changes for that industry. Electrical equipment is a relatively high wage paying industry. This reduces the number of opportuni tiies to leave the industry to obtain higher wages. Average wage opportunities for the industry were lower than for any other industry with only one minor exception in the urban group for transportation equipment. This indicates why the effect was in the expected direction but at nonsignificant levels. Wage opportunity effects for all durables were also negative for all groups with only

PAGE 147

135 the nonurban group not highly significant. This consistent effect seems to Indicate that employees in durable manufacturing are influenced by alternative wage opportunities and that the opportunity cost of remaining in durable manufacturing employment is sufficiently large to influence employment changes. Average increases in v/age opportunities were consistent in magnitude among groups for all four durable industries used. The nonurban group demonstrated the largest increase (Appendix B, Table 2k). Manufacturing employment opportunity . — The effects of Increases in employment opportunity v/ere more mixed than wage opportunity but were generally consistent in sign with wage opportunity effects. Negative effects were demonstrated for transportation equipment manufacturing for all groups except the nonurban group with some significant association noted for the group containing all counties. These effects complement the observations made in opportunity wages that activities in other Industries affect employment In transportation equipment manufacturing. Electrical equipment manufacturing demonstrated a similar effect to transportation in that all groups except the nonurban group exhibited a negative effect. No significant levels of association were noted. Positive effects were prevalent in both lumber and furniture and total durables. Highly significant positive effects occurred in both the all county group and the nonurban group for lumber and furniture products employment. Positive effects were also demonstrated for durables in all groups except the nonurban group with significant effects noted in the urban group and the all county group. Average employment opportunities were also smaller for these groups with the exception of

PAGE 148

136 the urban group. It appears that for these two Industries employment Increases were complementary to employment Increases In other industries. Electrical equipment experienced the largest average employment opportunity with the urban group demonstrating the largest increase of 256 employees for each electrical employee. The other groups averaged about two-thirds' this rate of increase. The next largest average employment opportunity occurred in transportation with a maximum increase of 178 employees per transportation employee occurring in the urban group. Furniture and lumber average employment opportunity was lower with the urban counties demonstrating an average of approximately 64 employees. Number of fi rms .--Changes In the number of firms were also Important in explaining employment changes in durable manufacturing. Increases in the number of transportation equipment manufacturing firms contributed positively to employment with highly significant effects noted for both the nonurban group and the group consisting of all counties . A contrasting situation was determined for lumber and furniture products manufacturing where all effects were negative. Average firm number changes were fairly consistent for all groups. Movement to a larger number of firms In this industry seemed to decrease employment. It becomes apparent that technology increases have demonstrated major influence in this industry and that the larger number of firms actually need fewer employees. Positive and highly significant effects were demonstrated for all three groups in electrical equipment manufacturing although average firm number increases were fairly small. Coefficient size for increases in firm numbers indicates that new firms in this industry may be quite

PAGE 149

137 large in terms of the number of employees and that the addition of one new firm occupies an important position in determining employment patterns in the area of its location. Positive effects viere also demonstrated for all durables in each group except the nonurban group. The all county group was the only group demonstrating any significance with a high level of association indicated. Group differences . — Significant differences between the urban and nonurban counties were noted for two of the industries included in the durable manufacturing analysis. Highly significant differences were indicated for both the transportation and lumber and furniture industries. No significant differences were noted for electrical equipment and all durables. Inclusion of the urban counties would generally cause a lower intercept for lumber and furniture and a larger intercept for the other three durable equations. This demonstrates the importance of these individual industries in the two diverse groups of counties.

PAGE 150

CHAPTER V SUMMARY AND CONCLUSIONS Summary The general objective of this study was to examine the importance of investments in human and natural resources along with several other variables in explaining employment changes within county groups for the four-state region of Mississippi, Alabama, Georgia, and Florida during the time period I960 to 1970. A theoretical economic model was developed and used in an attempt to explain changes in employment and firm numbers brought about by exogenous shifts in the supplies of resources, demand for products, supplies of other factors, firm production possibilities, and shifters of the number of firms In an industry. Empirical analysis attempted to determine the importance of each exogenous shifter on employment in agriculture, construction, durable and nondurable manufacturing industries. Comprehensive examination of previous analyses indicated that either water resource investments are poor tools for stimulating economic growth or inadequate methodologies have thus far been used in attempts to measure investment effects. This work represents an attempt to both improve existing methodology considering the Importance of Interdependencies among other important factors In a region receiving resource investments and to use the model to empirically demonstrate Its applicability and determine the effect of actual resource investments. 138

PAGE 151

139 The theoretical model developed was designed to explain absolute changes in e.npIoym,enc as a function of exogenous changes in the prices of products, prices of factors having perfectly elastic supplies, shifters of the supply of factors assumed to have other than perfectly elastic supply functions for a region, and shifters of firm production possibilities. Changes in the number of firms were explained as a result of the same exogenous variables and exogenous shifters of firm supply functions for agriculture. Complete mathematical derivation of the theoretical model was presented. The model consisted of three basic types of components with two equations representing each component. These components were (I) product supplies and factor demands for all firms In individual types of industries, (2) aggregate product demand and factor supply functions, and (3) the number of firms in each industry. Economic interpretation of both employment effects and firm number effects was presented. Each term and combination of terms in each of the coefficients were defined. The effect on changes in employment and changes in firm numbers as the result of changes in each of the exogenous factors was examined in detail. Graphical illustration of the equilibration process resulting from a change in supply of the th . n factor was given along with a mathematical discussion of the same process. Effects of changes in the number of firm entreprenuers , product price, factor price, and firm production possibilities also received detailed mathematical discussion. The area chosen for this study included the four states of Mississippi. Alabama, Georgia and Florida. Counties within these states were delineated into urban and nonurban counties using ten variables

PAGE 152

1^0 designed to classify the counties according to their human and natural resource endov/ments, and their urban, industrial and agricultural orientations. Three groups of counties v/ere selected for empirical analysis. One group consisted of 91 urban-oriented counties. A second group consisted of 28^ nonurban-or iented counties. The third group contained all 375 countiesIndustry classifications were also determined for analysis. Nine industry classifications were specified for analysis primarily on the basis of total employment in the study area. Some consideration was given to water use criteria. These industries were agriculture, construction, textile mill products and other fabricated textile products, food and kindred products, transportation, the combination of furniture and fixtures and lumber and wood products, electrical equipment, and industry classifications to represent both durable and nondurable manufacturing products. General equations for empirical analysis were specified for each industry. The model for agriculture was a twoequation model. Both two-stage least squares and ordinary least squares were used in estimating these equations. Some deviation from the theoretical model occurred for the remaining industries where changes in the number of firms were considered an exogenous variable. For these industries single-equation models were used to estimate the functional equations for employment changes. Counties were used as the units of observation. Exogenous shifters used in the models to explain employment and firm number changes were of five types. Changes in factor supply included changes in state and federal per pupil education expenditures. Corps of Engineers' investments, investments by the Soil Conservation

PAGE 153

]k] Service in the Public Lai-/ 566 Small Watershed Program, Agricultural Stabilization and Conservation Service payments in the Agricultural Conservation Program, and loans and grants for community water and sewer systems made by the Farmers Home Administration. Reductions in crop allotments were also included for agriculture. Changes in product demand were indicated by changes in derived county product price indexes. Changes in factor prices were indicated by changes in the annual wage rate for employees in each particular industry in a county. Changes in firm production possibilities were computed using changes in derived county technology indexes for each individual industry. Shifters of entreprenuer supply were considered to be variables that indi-cated alternative wage and employment opportunities. Farm operator age was also included in the agricultural equations. Conclus ions Effects on employment and firm number changes as the result of the various exogenous shifters differed quite substantia 1 ly among industries and among the individual groups considered. It was concluded from these variations that the exogenous shifters do have certain effects on employment and firm number changes depending on the type of shifter, the area of employment location and the type of industry under consideration. For example, certain factor supply changes were very consistent in their effect on employment and firm number changes as well as cons istently , inf luenc ing employment changes in given county groups. Changes in per pupil education expenditures demonstrated both positive and negative effects on agricultural employment and farm number changes as v/ell as in construction employment changes. The effect on

PAGE 154

nondurable rrwnufactur ing was consistently negative as opposed to a consistent positive effect for all durable manufacturing industries. Significant effects v/ere isolated and demonstrated no definite pattern. PosFtive effects were generally expected. The negative effects apparent in nondurable manufacturing employment and agricultural employment may have been the result of higher education expenditures which allowed outmigration from these industries to take advantage of newly acquired skills. These industries typically require lower skilled employees. Corps of Engineers' investments generally demonstrated a negative effect on agricultural employment and farm numbers. This may have been the result of larger and more efficient farms made possible by the investmentsThe effect on construction was positive as expected with high levels of significance noted. Variation in consistency patterns for the martufacturing industries was noted. Food and kindred products manufacturiTig employment demonstrated consistent negative effects. Total nondurables demonstrated positive effects. Significant coefficients were observed for food and kindred products employment, nondurable manufacturing efsployment and durable manufacturing employment in the urban counties ar?d the group consisting of all counties. Negative effects for this irpvestment category could have been the result of a time period too short to measure the employment effect. SmaSI Watershed Program investments also demonstrated both positive Bu
PAGE 155

1^3 Engineers investments on these employment categories. Few levels of significance v/ere noted. Investments by the Soil Conservation Service in the Agricultural Conservation Program demonstrated consistently negative effects on agricultural employment and farm number changes as well as in construction employment changes. Positive effects were noted in nondurable manufacturing and negative effects were observed for durable manufacturing employment. The negative effects on farm numbers were highly significant. This indicates that the larger farms were able to take advantage of this cost-share program and in effect consolidate the smaller farms with a resultant decline in farm numbers and agricultural employment. Farmers Home Administration loans and grants exhibited the same pattern of behavior for agriculture and construction as did the Agricultural Conservation ProgramThe effect in nondurables was also similar with some difference observed in durable manufacturing. Lumber and furniture products demonstrated highly significant positive effects with positive effects also occurring in electrical equipment. The nonurban group of counties demonstrated significant effects for all four durable industries. It is also quite possible that inadequate time elapsed during the study period to allow for adequate employment response to this investment category. Allotment decreases demonstrated the effect in both employment and farm number changes that a priori expectations would dictate. Decreases in employment and farm numbers \vere associated with allotment reductions. Some levels of significance v/ere noted for farm number changes in the urban counties and the all county group. Allotment reductions were significant in the all county group.

PAGE 156

Product demand shifts as measured through increases in product price generally increased employment in agriculture. Farm number increases were also observed with high levels of significance for all county groups. Negative effects on employment for all manufacturing industries were demonstrated for all groupings of counties. High levels of association were noted for food and kindred products, lumber and furniture products, and all durables. A reduction in firm numbers and employment due to reduced returns resulting indirectly from output increases, new firms entering the industry, and consequent reductions in residual returns was apparently large enough to offset the initial increases . Increases in agricultural wages produced generally negative effects for agricultural employment and positive effects on farm number changes. The effect of an increase in wages on construction employment was positive and highly significant. Nondurable manufacturing generally demonstrated a negative effect while durables were normally positive with several high levels of significance. Positive employment changes resulting from increases in wages indicate a situation where labor cannot easily be replaced in the production process. Decreases in labor employment v/ould indicate that other factors were being substituted for higher priced labor. The effect of changes in firm production possibilities or technology on employment was generally consistent with a priori expectations. Technology increases which expand output cause more agricultural employment to be required in industries that process the output. It is apparent that this type of effect occurred. Farm numbers did decline as

PAGE 157

the result of Increases in technology. Employment decreases were also observed for both nondurables and durables with some exceptions occurring for food and kindred products employment. High levels of significance were noted for transportation and lumber and furniture products employment. This indicates the importance of labor displacing technology being introduced into these two industries. Shifters of firm entreprenuer supply were generally very consistent with a priori expectations. Increases in wage opportuni ties demonstrated jiegative effects on agricultural employment and farm number chang es with the nonurban counties providing the exception. The negative effects noticed for construction were highly significant. Positive effects were generally consistent for food and kindred products and lumber and furniture products manufacturing employment. This indicates the Inability of low skilled employees In these industries to take advantage of other employment alternatives. Negative effects were Indicated for all other manufacturing employment industries. High significance levels were demonstrated for transportation and durable manufacturing employmentEmployment opportunity changes were consistent In most cases with wage opportunity effects on employment changes. Effects were not as apparent for agricultural employment and firm numbers as was expected. Positive employment effects were generally indicated for food and kindred p roducts , lumber and furniture products, and durable manufacturing products. Employment increases indicate that expansion In other Industries was complemented by expansion in the Industries demonstrating positive coefficients. Farm operator age was consistent with a priori expectations for all groups. Decreases in the number of farm operators 55 years of age

PAGE 158

1^6 or older were significantly associated with decreases in the number of farms . Changes in the number of farms and manufacturing firms were important in explaining employment changes. Increases in the number of firms in each industry led to increases in employment with the exception of the lumber and furniture industry. Most of the coefficients were significant. Group differences betv;een the urban and nonurban groups of counties were important for some industries. Significant differences were noticed for the construction, textiles, food and kindred products, transportation, and lumber and furniture industries. Limi tat ions Several limitations concerning this study should be discussed. The first concerns data problems. When an attempt is made to define variables as used for this study sources of adequate data are normally a major obstacle. Data were used from various sources to overcome this. Sources included both investment data provided from agency files and published data of the type found in census publications. The time period under study was I960 to 1970. Data sometimes were reported for periods that did not coincide with this period. For example, data from some census sources v;ere for the time period 1959 to 1969This is a problem that is always present in a research attempt of this nature. There is also some deviation from the theoretical definitions of variables. Some effects may also go unmeasured. Private investments in the same areas were not measured. These investments also should

PAGE 159

1^7 affect employrr,ent changes. Regardless of these shortcomings, the best possible data sources v/ere utilized and data difficulties should not appreciably affect the results offered. Another major limitation concerns the application of the research results. Employment or firm number effects that occur as the result of a resource investment may not always occur for each ind iv idual project. This analysis suggests that in general the results as reported will occur for the given type of area. No definite statement can be made that the same effect will absolutely occur for an individual project. Need for Further Research Definite suggestions can be made for future research. The model presented in this study is based on economic theory as derived from the basic economic model. It seems apparent that the equations used for the empirical analysis do an average job of explaining employment changes for some industries and an excellent job in others. One major improvement which could be made is the conversion of the model presented to a time series analysis model. Incorporation of the county data into a time series analysis should provide a model that would do an even better job in explaining employment and firm number changes. Use of this type model would allow the computation of multipliers which would be quite useful In any economic analysis. The major problem that would be encountered is the availability of adequate time series data. Indexes of county product price, technology, and several other variables would have to be derived on an annual basis. Sufficient data at this time are not available to allow these calculations. A second suggestion centers around the use of a mul tiindustry simultaneous equation model. This model would be better suited to time

PAGE 160

148 series analysis. Simultaneous equation models would allow for the development of interindustry relationships and provide more insight into the complementary and subs t i tutabi 1 i ty relationships of labor employment among the alternative industries.

PAGE 161

APPENDIX A SPECIFICATION OF AREA ADJUSTMENT MODEL

PAGE 162

150 In this model the product demands and factor supplies of all but the n factor are assumed to be perfectly elastic for a region in which k different types of industries exist. A. Product Demands (0 P:,, = k,. j = 1, 2, . . ., m B. Product Supplies (2) Qj.^ = N^qj.^(P.Z) j = 1. 2 m where Q = aggregate quantity of the j product supplied j k by firms in the k industry in the area. N, = number of firms in the k industry, q..(P,Z) = quantity of the j product supplied by a J "^ firm in the k industry as a function of the prices of all products and factors (p) and the exogenous shifters of product supply and factor demand functions (Z). C. Factor Demands r (3) Q. = N,q..(P,Z) i = m + 1, m + 2, . . ., n 1 , ,'k^ik' k = I where Q. = aggregate quantity of the i factor demanded in the area. q..(P,Z) = quantity of the i factor demanded by a firm in the k industry as a function of the prices for all products and factors (p) and the exogenous shifters of factor demand functions (Z).

PAGE 163

151 D. Factor Suppl ies (^) P. = k. i = m + 1, m + 2, . . ., n -1 II and for the n factor (5) Q^ = S^(P,X) p = m + 1, m + 2, . . ., n where Q = aggregate quantity of the n factor supplied in the area. S (P,X) = aggregate supply of the n factor as a function of all factor prices (P) and the exogenous shifters in the factor supply function ( X) . E. Firm Entrepreneur Supply (6) N^ = P,^(n^,W) e = 1, 2, . . ., r where Pu.(^ »W) ~ number of firm entrepreneurs willing to operate a k' e firm in the k industry as a function of the residual returns from all types of industries (n ) and the exogenous shifters of firm entrepreneur supply (W) . Additional comments on this function are given in the discussion of the total differential of this equation. F. Industry Entrepreneur Residual Return Function m n i = 1 • j_i 'k I J ' I = m+1 where n = total residual return to all firms in the k industry.

PAGE 164

152 th q., = quantity of the j product produced by a firm in J '^ t u the k industry P. = price of the j product, q = quantity of i factor used by a firm in the k i ndus try . r -th , , P. = price of I factor. I Area adjustments imply various types of changes. Conversion of the above equations into their total differential form allo\f6for a more thorough study of the nature of equilibration. A. Product Demands All area product demand functions are assumed to be perfectly elastic and all product price changes exogenous. (la) dP. = dk. dk. autonomous for j = 1, 2, J J J . . . , m B. Product Suppl ies m bq n bq^ v 5q > j ~ ' J i=m+I 1 h-l h ^^jk'^^ C. Factor Demands r m ^^• I, f" n ^*^ • k k=lj=l J J k=li=m+l I r V ^'^•i. '' + E Z N.-TT^dZ, + Z q.,dN, , , , , k 5Z. h I _ , ik k k = 1 h = 1 h k = 1 Expanding the second term of (3a) and rewriting the equation in terms of the n factor gives

PAGE 165

153 r m ^"^i, rn-I ^^ <. (3b)dO = E I N.-r^P. + E E N,-^dP. r bq r ^ ^^ i. + 2 N.-r^dP + E E N.r-^ dZ. ,,k0P n , ,, ,koZ, h k=l n k=lh=l h + E q ,dN, D. Factor Suppl ies All area factor supply functions are assumed to be perfectly 1"h elastic except for the n factor, thus Ika) dP. = dk. dk. autonomous for i = m+1, II I m + 2, . . ., n-1 and for the n factor n 1 5S bS t bS I = m+1 I n f = 1 f E. Firm Entrepreneur Supply r bp. s bP. e = 1 e d = 1 d interpretation of the first term in equation (6a) implies that a change in the number of firms in an industry is a function of the residual return to firm entrepreneurs in all industries. Since most major industries produce quite diverse products this is probably not true for most industries. In these situations this term would be zero with f h only the residual return in the k industry of interest influencing changes In the number of firms in that Industry. To avoid misinterpretation of terms in later equations this term is left in general form as presented.

PAGE 166

Since N. & equation (6a) can be written as r ON s !)N (6b) dN^ = I ^dn^ . E — dw^ e = 1 e d = I d 15^ F. Industry Entrepreneur Residual Return Function n i = m+1 m (7a) dn. = 5:Nkqik^P; ^N^qikdR. m a = 1 m &q;kN. n ^^ik^ j = 1 J '' I = m+ I a dP n £ + a = m+l V + E h = 1 = 1 J ^P m Z P i = 1 a ^i i 5P I = m+i a J ^z^ j_k k ^^ik^^k I = m+l h dP f a dZ. Using the assumption that firms act as profit maximizers in a competitive market, the third term (enclosed in braces) in equation (7a) equals zero. This assumption means that the value of quantity changes of commodities produced induced by a small price change equals the value of quantity changes in factor utilization caused by the price change at an equilibrium position of the firm. This relationship follows from the discussion by Hicks [52, p. 322] of the stability conditions of production. By letting ?)R. = P,5q,.^lk + P bq .'^2I< + . . . + P„&q^i, "^^ ^ change in total m mk returns in the k industry and ^^k = ^m+l^Vl.k^m+l.k ^ W^V2,kV2,k + • • • + V^k^mk = change in total costs in the k industry, then equation (7a) can be wr i tten as

PAGE 167

155 m (7b) dn^= _ lyjk'^j J V + E h = 1 n-1 Z N,q.,dP. N,q , dP ,, k IK I k nk n t " m+ I &Z. 6Z. h n dZ, Differential equations (la), (2a), (3b), and (5a) indicate the changes in the quantity of products and factors supplied and demanded resulting from changes in prices, exogenous shifters, and the number of firms in a given industry. Equilibrium conditions would imply that changes in the quantity demanded and supplied of any factor or product were always equal. Since all product and factor prices except for the n factor are considered to be exogenous ly determined in this model, the equilibration process centers on the demand and supply of the n factor. Changes in product and factor prices and exogenous shifters of firm supply also can alter the residual return received by firm entrepreneurs. Equations (6b) and (7b) bring the functions related to the number of firms in an industry into the equilibration process. The equilibration process that will ultimately lead to an equation explaining changes in employment as a function of product and factor prices, shifters of firm production possibilities, shifters of factor supply, and changes in the numbers of firms is described below. Also derived is an equation explaining changes in firm numbers as a function of product and factor prices, shifters of firm production possibilities, shifters of factor supply, and shifters of firm entrepreneur supply. Equating total differentials for the n factor demand (3b) and factor supply (5a) and solving for dP gives (8) dP = f (dP., dP., dZ. , dX^, dN,). n J ' I ' h f k j = >, 2, m

PAGE 168

156 i=m+l,m+2, ..., n-1 h = 1 , 2 V f = 1 2 t k = 1 , 2, . . ., r EquatVon (3a) expresses the total change in the quantity of the i factor demanded in the area. This equation can be written in terms of labor demanded as r m ?)q , r n1 bq. , k=lj = l J-" k=1i=m+l I r 5q,, r V bq r + I N.-TF^dP + Z E N,-r^dZ. + Z q, , dN, , , k CP n , , , , k 6Z, h , , Lk k k=l n k=lh=l h k=l Substitution of equation (8) into (9) gives an equation of the form (10) dQj_ = f(dX^, dP^., dP., dZ^, dN^) f = 1, 2, . . ., t j = 1 , 2 , . . . , m i=m+l,m+2, ...., n-1 h = 1, 2, . . ., V which expresses the total change in the quantity of labor demanded in the area as a function of exogenous changes in shifters of factor supplies, product prices, factor prices except for the n factor, shifters of firm production possibilities, and the number of firms. Finally, equation (10) can be written for the k industry as

PAGE 169

157 t (11) dQ, . = E m Ykn (ij^f dX, dP. J r "I n-1 + Z i = m+1 V + E h = 1 L^Lkh ^kn(i) ^nkh] dZ. ' [^Lk \kn (i)
PAGE 170

158 Table 15. Definition and interpretation of terms in equation (11) Term Def in i t ion I nterpretat ion Change in quantity of labor demanded in the k industry associated with a one unit change in price of the n factor. ''ikn

PAGE 171

Table 15 (Continued) 159 Term Def in i t ion Interpretat ion nkh th Change In quantity of the n factor demanded In the k'-'' industry associated with a one unit change in the h^'^ shifter of firm production possibilities. *rik ^nk th Quantity of the n factor demanded for a firm in the k^^ industry. The superscript o indicates initial values,

PAGE 172

160 Letting dn = dn. for the k industry and substituting the industry return function (7b) into (12) gives the total price change of fh the n factor as a function of all exogenous variables as expressed by (13) dp = f(dP., dP., dZ, , dX,, dW.) n J ' n f " j = 1 » 2 m i=m+l,m+2, ...,n-l h = 1 , 2, . . . , V f = 1, 2, . . ., t d=l,2, . . .,s. Then, by substitution of (13) into (7b) we get (14) dn^= f(dP^, dP.. dZ^, dX^, dW^) j = 1 , 2 , . . . , m i =m+,l,m+2, . . ., n1 h = 1 , 2, . . . , V f = 1, 2 t d=l,2, . . .,s. Finally by letting dn = dll for the k industry and substituting (14) into (6b) we can derive an equation which expresses the change in the number of firms in the k industry as a function of all exogenous variables. This equation may conveniently be written as (15) dN = E ^ f = 1 (i) \n W ^f dX, s + Z d = 1 m + Z J = » \d -^ \ \} ^ \ n (i) ^d n (i) ^• dW dP.

PAGE 173

161 n-l

PAGE 174

162 Table 16. Definition and interpretation of terms in equation (12)' Term Def i n i t ion I nterpreta I: ion 'kn r 5N, .,0 r 5N. „ k N k ,bn ne ,5P e = 1 e e = 1 n Change in the number of firms in the k^'^ industry associated v/ith a change in entrepreneurial residual return brought about by a one unit change in price of the n^h factor. b kd 5N, Change in the number of firms in the k^'' industry associated with a one unit change in the d^"" shifter of firm entrepreneur supply. i • r 5N, .,0 r X -J
PAGE 175

Table 16 (Continued) 163 Term Def in i t ion Interpretat ion bS r 6X, Change in total quantity supplied of the n*-" factor associated with a one unit change in the f-'^ shifter of factor suppl ies . l< = 1 Change, in total demand for 4-1-, the n factor associated with a one unit change in the d*^^ shifter of firm entrepreneur supply. i N,' + J; q , a, . k= 1 1^ ^P; k= 1 ""^ "^J Change in total demand for f h the n '' factor associated with a one unit change in price of the j ^^ product. L. bS n bP. r ,o nk . , k bP. r Z k = 1
PAGE 176

]6k q . = q . + Aq . le le le i =m+ 1,m+2, . . ., n1 Aq . = change in q. resulting from AP ^je ^ ^je ^ n Aq. = change in q. resulting from aP t h N = number of firms in the e industry, e ' Making the above substitutions and rearranging terms gives An = AP q N + AP Aq N + e n ne e n ne e m U E P.Aq. = 1 J J^ i = m+1 P.Aq. le N e. Assuming that the bracketed term is close to zero based on Hick's stability condition of production discussed earlier in the appendix, and assuming that Aq is small relative to q , the term reduces to ne ^ne AIT ^ Ap q° N° e n ne e 5N^ Since a, = -r-fr-q N , appropriate substitution would allow a, to be kn 5n ^ne e' ^^ ^ kn e written as -Tp— and interpreted as the change in the number of firms associated with a one unit change in P operating through a reduction in entrepreneurial residual return. Similar substitutions could be performed for the terms a, ., a, ., '^ kj k I and c,, . Table 17 summarizes the appropriate approximation for All for each term. Equation (15) is similar to that used by Schrimper [18] with two exceptions. Schrimper's model was concerned with the rate of change expressed as percentages and had elasticities incorporated into the

PAGE 177

165 Table 17. Approximation of terms to represent a change in residual return Term Def ini tion Representation for ati 'kn ON, k o 1,0 binned e AP°q N° n ne e i • 5N, Oil ^je e e -' AP.q? N° J je e 'kf bN. k .,0 on ^ le e e AP.q? N° lie e kh biT bR bC e e bZ, " bZ, An T a This effect would occur through either an output increasing and/or input decreasing change in firm production possibilities. Since both could occur the representation is written in a similar manner to the actual definition. coefficients whereas equation (15) is expressed in terms of absolute changes and each coefficient consists of combinations of partial derivatives as shown in Table 16. A complete discussion of the similar equation and its terms can be found in Schrimper [18, pp. 116-133].

PAGE 178

APPENDIX B MEANS AND STANDARD DEVIATIONS OF VARIABLES

PAGE 179

167 CL O c o u j: u 0) V) 0) <0 L. > c c « >^ O "a. E 0) tf>

PAGE 180

168 Q. 3 o I. en c o u o 0) 0) J3 > E c E in C o >

PAGE 181

169

PAGE 182

170 o. o CT> C 3 o o JZ o lU o I/) 01 c (D -C o o o O a. c o Q>

PAGE 183

171 m C o > o 11o a. o CD c 3 o o x: u a> o u1/) (U ro I. ITJ > CT) C 03 0) 3 in c o to > •o -D t — O IS\ LA -— C3A CM « • u\ O {r\ CM — -f OO « * CO J(T\ Lf\ CO — CM — CM CM • « -Jr^ -3-3" — CM — O J« • o^oo moo LA CM -JOO -T OO

PAGE 184

172 o 1c o o s: o (U >_ o lU 03 > en c •0 o "o c o 0) o in c o > 10 c nj 03 •J c o XI c *-> > c

PAGE 185

173 CL 3 o u CTl C O o o TO (U VO
x> (Q u (0 > o en c m c 3 4-> 1. o o. CL o V 01 TO i_ O in c o

PAGE 186

17^ Q. o c o U u O in 0> m > 0) O) c u c 3 o o. a. o c (U e >^ o O in c o < LLl \o m >

PAGE 187

BIBLIOGRAPHY 1. U. S. Water Rssources Council. Policies, Standards and Procedures in the Formulation of Plans for Use and Development of Water and Related Land Resources . Senate Document 97Washington, D. C: Government Printing Office, 1962. 2. Appalachian Regional Development Act . 89th Congress. 1st Session, Public Law 89't, Statute 79-5Washington, D. C: Government Printing Office, 1965. 3U. S. Water Resources Council. Proposed Principles and Standards for Planning Water and Related Land Resources . Notice of Public Review and Hearing. Federal Register, Part 11, Volume 36, Number 2^5. Washington, D. C: Government Printing Office, December 21, 1971k. Back, W. B. "Estimating Contributions of Natural Resource Investments to Objectives in Regional Economic Development," American Journal of Agricultural Economics , Vol. 51, No. 5 (December, 1969), pp. 1442-14^+8. 5. Jansma, J. D. and Back, W. B. Local Secondary Effects of Watershed Projects: A Case Study of Roger Mills County, Oklahoma . ERS-I78, Economic Research Service, U. S. Department of Agriculture and Department of Agricultural Economics, Oklahoma Agricultural Experiment Station, Oklahoma State University, Washington, D. C, May 21, 1964. 6. Gray, R. M. and Trock, W. L. An Economic Evaluation of the Green Creek Watershed Project Departmental Technical Report Number 2, Department of Agricultural Economics and Sociology, Texas Agricultural Experiment Station, Texas A & M University, College Station, Texas, December, I968. 7. Kasal, J. "An Analysis of Economic Impacts of Five Colorado Watershed Projects." Natural Resource Economics Division, Economic Research Service, U. S. Department of Agriculture, Washington, D. C: 1970 (mimeographed). 8. Eidman, Vernon R. "Estimating the Secondary Impact of Watershed Projects." Paper presented at an Economics Workshop of the Department of Agricultural Economics, Oklahoma State University and Economic Research Service and Soil Conservation Service, U. S. Department of Agriculture, Oklahoma State University, Stillwater, Oklahoma, May 13, I97O (mimeographed). 175

PAGE 188

176 9. Mazuera, Oscar E. Economic Impact of Irrigation Development: Sugar Creek Wacershed. Ol-.lahoma . Doctoral Dissertation, Department of Agricultural Economics, Oklahoma State University, Stillwater, Oklahoma, 196910. Bromiey, D. W. , Schmid, A. A., and Lord, W. B. Pub 1 ic Water Resource Project Planning and Evaluation: Impacts, Incidence, and Institutions . Center for Resource Policy Studies and Programs Working Paper Number One, School of Natural Resources, University of Wisconsin, Madison, Wisconsin, September, 197111. Gibbs, K. C and Loehman, ET. Impacts of Watershed Projects . Economic Report kO. Food and Resource Economics Department, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida, November, 1972. 12. Haveman, Robert H. and Krutilla, John V. Unemployment, Idle Capacity, and the Evaluation of Public Expenditures: National and Regional Analysis . Baltimore: Johns Hopkins Press, 1968. 13. Howe, Charles W. "Water Resources and Regional Economic Growth in the United States, 1950-1960," Southern Economic Journal , 34 (April, 1968) , 477-^9. 14. Wiebe, J. E. "Effects of Investments in Water Resources on Regional Income and Employment," in Proceedings of the Symposium on Social and Economic Aspects of Water Resources Development . Urbana, Illinois: American Water Resources Association, 1972. 15. Cox, P. T., Grover, C. W., and Siskin, B. "Effect of Water Resource Investment on Economic Growth," Water Resources Research , Vol. 7, No. 1 (February, 1970, 32-38. 16. Tolley, G. S. and Schrimper, R. A. "Feasible Ways for Relating the Micro and Macro," Price and Income Policies . Agricultural Policy Institute Publication No. 17, North Carolina State University, Raleigh, North Carolina (April, 1965), 115-12917. Schrimper, R. A. "Wal ras ian-Hicks ian Analysis of Interrelationships in Agricultural Adjustments," Price and Income Policies . Agricultural Policy Institute Publication No. 17, North Carolina State University, Raleigh, North Carolina (April, 1965), 141-152. 18. Schrimper, R. A. Micro-Aggregated Theory of Agricultural Adjustments with Application to Farm Number Changes . Doctoral Dissertation, Department of Economics, North Carolina State University, Raleigh, North Carolina, I966.

PAGE 189

177 19. Eddlemsn, 3. R. and Cato, J. C. Factors Affecting Rates of Change in Farm Nurribers Among Counties in Florida. 1959-1969 Cooperative study of the Food and Resource Economics Department, Universlfy of Florida, Florida Agricultural Experiment Stations and Natural Resource Economics Division, Economic Research Service, U. S. Department of Agriculture. (in process.) 20. Eddleman, B. R. "Estimating the Effects of Resource Development Programs on Regional Employment," American Journal of Agricul tural Economics , Vol. 51, No. 5 (December, 1969), pp. 1^3^-1^441. 21. 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, University of Florida, Gainesville, Florida, March 20-21, 1969; (mimeographed) . 22. Tintner.G. Econometr ics . 1st Science Edition. New York: John Wiley and Sons, Inc., 196523US. Bureau of the Census. Census of Manufactures, Area Statistics: 1967 Vol, III, Part I. Washington, D. C: Government Printing Office, 1968. 2k. U. S. Bureau of the Census. Census of Population: 1970. Vol. I, Parts 2, 11, 12, and 26. Washington, D. C: Government Printing Office, 1972. 25. U. S. Bureau of the Census. Census of Manufactures, Summary and Subject Statistics: 1967 Vol. 1. Washington, D. C: Government Printing Office, I968. 26. Goldberger, Arthur S. Econometric Theory . New York: John Wiley and Sons, Inc. , 1964, 27. Johnston, J. Econometric Methods . New York: McGraw-Hill Book Co.. 1968. 28. Raduchel, William James. The Regression Analysis Program for Economists . Technical Paper No. 10, Harvard institute of Economic Research, Harvard University, Cambridge, Massachusetts, May 1972. 29. U, S. Bureau of the Census. Census of Population: I960 , Vol. I, Parts 2, 11, 12, and 26. Washington, D. C: Government Printing Office, 1962. 30. U, S. Bureau of the Census. Census of Manufactures, Area Statistics : 1958. Vol. Ill, Part 1. Washington, D, C: Government Printing Office, I960.

PAGE 190

178 31. U. S. Bureau of the Census. Census of Agriculture: 1959 Vol. I, Counties, Parts 28, 29, 3?, and 33Washington, D. C: Government Printing Office, 1961. 32. U. S. Bureau of the Census. Census of Agriculture: 1969 Vol. I, Area Reports, Parts 28. 29, 32, and 33, Section 2, County Data. Washington, D. C: Government Printing Office, 1972. 33Alabama State Board of Education. Annual Report: I960. Montgomery, Alabama: Division of Administration and Finance, I960. 3^. Alabama State Board of Education. Annual Report: 1970Montgomery, Alabama: Division of Administration and Finance, 197035Mississippi State Department of Education. Current Expenditures Per Pupil by School Districts and Related Information: 1959-60 . Jackson, Mississippi: Division of Administration and Finance, I960. 36. Mississippi State Department of Education. Current Expenditures Per Pupil by School Districts and Related Information: 1969-70. Jackson, Mississippi: Division of Administration and Finance, 1970. 37Georgia Department of Education. Current Expenditures or Cost Per ADA Child. I960 (mimeographed). 38. Georgia Department of Education. Georgia Education Statistics , 1969-70 . Atlanta, Georgia: State Superintendent of Schools, 1970. 39. Bureau of Economic and Business Research. Florida Statistical Abstract, 1972 . Gainesville, Florida: University of Florida Press, 1972. 40. U. S. Department of Agriculture, Agricultural Stabilization and Conservation Service. Alabama ASCS Annual Report: 1959-1970 Montgomery, Alabama: Agricultural Stabilization and Conservation Service. k] . U. S. Department of Agriculture, Agricultural Stabilization and Conservation Service. Mississippi ASCS Annual Report: 1959-1970 . Jackson, Mississippi: Agricultural Stabilization and Conservation Service. k2. U. S. Department of Agriculture, Agricultural Stabilization and Conservation Service. Florida ASCS Annual Report: 1959-1970. Gainesville, Florida: Agricultural Stabilization and Conservation Service.

PAGE 191

179 43U. S. Department of /Agriculture, Agricultural Stabilization and Conservation Service. Georgia ASCS Annual Report: 1959-1969 . Athens, Georgia: Agricultural Stabilization and Conservation Serv ice. kk. U. S. Department of Agriculture. Agricultural Statistics: I960 . Washington, D. C: Government Printing Office, I960. ^5' U. S. Department of Agriculture. Agricultural Statistics: 1970. Washington, D. C: Government Printing Office, 1970. 46. U. S. Bureau of Labor Statistics. Monthly Labor Review. Vol. 83, No. 10: Vol. Sk, No. 10; Vol. 85, No. 10; Vol. 93, No. 10; Vol. Sk, No. 10; Vol. 95, No. 10. Washington, D. C: Government Printing Office, 1960-62, 1970-72. ky . U. S. Bureau of the Census. Location of Manufacturing Plants by County, Industry, and Employment Size: 1958 . Part 5. Washington, D. C: Government Printing Office, I96I. kS. U. S. Bureau of the Census and U. S. Bureau of Old-Age and Survivors Insurance, Cooperative Report. County Business Patterns, First Quarter, 1959 Parts 6B and 7. Washington, D. C: Government Printing Office, I96I. 49. U. S. Bureau of the Census. County Business Patterns, 1970 . Parts 2, 11, 12, and 26. Washington, D. C: Government Printing Office, 1971. 50. U. S. Department of Agriculture, Economic Research Service. Changes in Farm Production and Efficiency, Labor: indexes of Farm Production per Hour, by Livestock and Crop Groups, for Each Farm Production Region, 1950-71 Statistical Bulletin No. 233, Supplement IV, 1972. Washington, D. C: U.S. Department of Agriculture, 1972. 51Board of Governors of the Federal Reserve System. Industrial Production: 1971 Edition . Washington, D. C: Federal Reserve System, November, 1972. 52. Hicks, J. R. Value and Capital . Second Edition. Oxford: Clarendon Press, 1961.

PAGE 192

BIOGRAPHICAL SKETCH James C Cato was born in Vernon, Texas, on October 28, 19^2, the son of Willis E. and Mildred Cato. He g'-aduated from Lockett High School, Vernon, Texas, in 1961. In June, 1967, he received the Bachelor of Science degree in agricultural economics at Texas Tech University, Lubbock, Texas. A Master of Science degree in agricultural economics was received from Texas Tech University in June, 1968. He served as an instructor in agricultural economics at Texas Tech University in 1968 and early 1969He enrolled in the Graduate School of the University of Florida in 1969He was employed by the Economic Research Service, United States Department of Agriculture from 1969 until 1973Since June, 1973. he has been employed as Graduate Research Associate and Extension Marine Economist with the Department of Food and Resource Economics, University of Florida, Gainesville, Florida. He married Diane Lynette Richardson, from Vernon, Texas, on December 28, I965, and has two sons, Kyle and Chad. He is a member of Alpha Zeta, Gamma Sigma Delta, Omicron Delta Epsilon, American Agricultural Economics Association, Southern Agricultural Economics Association, and the American Economics Association. 180

PAGE 193

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Si 9'e^ zn -yy: B. R. Eddleman, Chairman Associate Professor of Food and Resource Economics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Oi i/. E. Reynolds' ssociate Professor of Food and Resource Economics I certffy that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. K. C. Gibbs Assistant Professor of Food and Resource Economics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. R. W. Bradbury Professor of Economics"

PAGE 194

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1973 College of (^ifgr icu 1 ture Dean, Graduate School