ABSTRACT
APPRAISAL OF ALTERNATIVE MODELS FOR ESTIMATION OF
INDIRECT EMPLOYMENT AND PROFIT GAINS RESULTING
FROM PUBLIC WATER PROJECT INVESTMENT
By
William Augustus Ward

In June of 1969 the United States Water Resources Council issued
a revised statement of policies and procedures for evaluating Federal
water resource projects. The primary effect of the new procedures
was to greatly broaden the concept of project benefits. In particular,
the benefits previously defined as "primary" and "secondary" or "in—
direct" were both lumped together into categories called "contributions
to national income" and "regional development." The research reported
herein dealt with the appraisal of secondary or indirect benefits.

The literature review revealed three categories of secondary
benefits: 1) The customary variety consisting of the multiplier
impacts "inducedéby" and "stemmingefromv output changes; 2) ex-
ternalities such as economies of scale and technological spillovers;
and 3) dynamic secondary benefits sudh as socially engineering the
human inputs. The latter two categories have been the least

widely used and seemingly the most difficult to appraise.

William Augustus Ward
2
It was decided to focus upon the appraisal of the "customary" variety
which applied the aggregate demand multiplier concept.

The literature review also revealed that most indirect benefit
analyses violate both the compensation principle and the Opportunity
cost concept. The multiplier analyses surveyed showed a marked ten-
dency to under—value the opportunity cost Of labor and to over-
estimate the real income changes. In most cases the benefits es-
timated exceeded the change in the recipients ability to pay.

In the study, argwrents against placing a zero opportunity cost
on human and capital resources were presented. It was argued that
_ only the changes in quasi-rents to capital and the chgges in labor
wages stimulated by the project are benefits to peOple in the project
region. An attempt was made to develOp an indirect benefit appraisal
model which reflected these concepts. rIhe model used national input-
Output coeficients and secondary data regarding wages and profits by
sector and region.

The conditions and assumptions under which the proposed model
would be Operable were spelled out. A total of nineteen conditions
and assumptions were identified. Key among these were: 1) the
Opportunity costs of the factors involved, 2) the state of the national
and local economies, 3) the source and alternative allocation of the
Federal funds expended, and 14) the inter-regional and inter—job mobility
of labor. Five of the nineteen assumptions were not specific to the
proposed model, but were applicable to indirect benefit analysis in
general. Two more related to input—output approaches in general.

Several had the effect of pre-specifying parts of the result.

William Augustus Ward
3
The model was applied to four different water resource projects.
The results were compared.with the indirect benefits estimated for the
same projects by eight other teams of analysts. The teams represented
three major water resource agencies and four universities. Numerical
estimates of indirect benefits differed widely between the teams of
analysts. These differences were found to be largely attributable to:
l) differing concepts of What constituted indirect benefits, and
2) differing assumptions regarding the macro-economic framework within
which the projects were being appraised. Some of the indirect benefit
categories applied by the teams of analysts were: 1) Opening up new
farming Opportunities in the western United States so as to re-distribute
population, 2) raising per capita incomes in the regions effected (ine
tensive growth), 3) increasing the total_income Of the region even if the
increase occurred without increasing the incomes Of individuals already
in the region, and 4) increasing the stability of income. Some key iflh
plicit assumptions regarding the macro-economic policy framework related
to: 1) other programs to deal with income and.employment maintenance,
and 2) the sources of the Federal funds spent and the methods used to
collect those funds.
The study indicates that great difficulty will be had in applying
the expanded benefit categories to Federal water project appraisal.
The difficulty boils down to fOur key tasks:
1. Making the indirect income concepts and methods of estimation
fit into the display framework selected by the water Resources
COuncil:
a. The accounts display income changes for the project region
and for the nation as a.whole and do not facilitate the display
of changes in income to individuals, within or without the region,

b. The accounts do not facilitate the display of income losses
to other regions;

William Augustus Ward
u

2. Choosing the effects to display —- e.g., regional employment,
regional income, new farming opportunities, net national income,
and income stability - and selecting the weights to be assigned
to each for project-by—project comparisons. This is a value judge—
ment which must be made either by the analyst or by the political
process;

3. Finding the data to apply existent methods and developing
improved methods for estimating the indirect impacts;

A. Developing the skills within the field staffs Of the water
resource agencies to handle the complex concepts and estimation
procedures involved in estimating indirect effects.

APPRAISAL OF ALTERNATIVE MODELS FOR ESTIMATION OF
INDIRECT EMPLOYMENT AND PROFIT GAINS RESULTING
FROM PUBLIC WATER PROJECT INVESTMENT

By

William Augustus Ward

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree Of

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

7 ‘ ACKNOWLEDGEMENTS

There are many people who made contributions to this
study. To enumerate all Of them would be impossible.
Nevertheless, certain people made contributions over the
life of the project which should be acknowledged. Among
these are Dr. Larry Conner and Dr. Daniel Chappelle. In
particular, thanks go to Dr. Allan Schmid who served as
Chairman Of the guidance committee and thesis committee.

Finally, a special debt of gratitude is owed to my
wife, Kaye, whose continued encouragement played a major

role in the completion of this study.

TABLE OF CONTENTS

Chapter
I. INTRODUCTION . . . . . . . . . . . . . . . .
Towards the Appraisal of Secondary
Benefits. . . . . . . . . .
Objectives. . . . . . . . . .

II. SECONDARY BENEFITS DEFINED: A REVIEW. . . .
The Customary View - The Multiplier . .

Indirect Benefits as a Pecuniary
thernality. . . . . . . . . .

Primary and Secondary Benefits:
Mutually Bxclusive?.

Secondary Benefits - The "Externality'

Approach . . . . . . . . . .

"Dynamic" or Developmental Secondary
Benefits . . . . . . . . . .

Summary Of Secondary Benefit
Classifications. . . . . . .

III. SECONDARY BENEFITS, ECONOMIC RENT, AND
THE COMPENSATION ABILITY PRINCIPLE. . .

Wage Changes and Economic Rent. . . . .

Economic Rent and the Compensation
Ability Principle . . . . .

A Working Definition of Secondary
Benefits. . . . . . . . . . .

IV. SOME APPROACHES TO SECONDARY BENEFIT
EVALUATION: A REVIEW . . . . . . . . . .

iii

Page

10

12

14

17

22

24

35
39

47

50

SS

iv

Chapter Page
Secondary Benefits and Economic
Impact Analysis . . . . . . . . . . 55
Impact Analysis and "Customary"
Secondary Benefits. . . . . . . . . 56
Multiplier Analysis . . . . . . . . 60
Keynesian Multipliers . . . . . . . 62
"Induced" and "Induced-by": A Note
on Terminology. . . . . . . . . . . 69
The Economic Base Multiplier. . . . 70
The Comparative Projection
Approach . . . . . . . . . . . 78
Index Of Internal Purchases . . . . 84
From-to Analysis. . . . . . . . . . 86
Input-Output Analysis . . . . . . . 89
Output Multipliers . . . . . . 94
Type I (Direct-Indirect)
Income Multipliers. . . . 97
Type II (Direct-Indirect-Induced)
Multipliers . . . . . . . 103
Input-Output and Secondary
Benefits: An Overview. . 108
Linear Programming. . . . . . . . . 113
Regional Linear Programming. . 113
Inter-regional Linear
Programming . . . . . . . 114
River Basin Planning Models. . 114

Input-Output as a Linear
Program . . . . . . . . . 115

Secondary Benefits and the Multiplier:
A Summary . . . . . . . . . . . . . 116

Chapter Page
V. ESTIMATING SECONDARY FACTOR INCOME

CHANGES . . . . . . . . . . . . . . . . 125
Foreword. . . . . . . . . . . . . . . . 125
Estimating Wage Changes . . . . . . . . 127

The Wage Component of Value
Added. . . . . . . . . . . . . 128
The Wage Rent Coefficient . . . . . 134
Estimating Profit Changes . . . . . . . 136

The Profit Component of Value

Added. . . . . . . . . . . . . 136

Rounds of Expansion and Benefit
Creditability . . . . . . . . . . . 142
Associated Investments. . . . . . . 143
Inelasticities of Supply. . . . . . 144
Adjusting the Multiplier. . . . . . 145
Regional Coefficients and Purchases . . 147
Differing Requirements. . . . . . . 148
Localizing Purchases. . . . . . . . 149
National and Local Industries. 153
Localization Ratio . . . . . . 155

Applying the Model. . . . . . . . . . . 157

The Trenton Channel Project . . . . 157
The Computation Procedure . . . . . . . 158
Induced-by Benefits . . . . . . . . 158
Stemming-From Benefits. . . . . . . 164
Construction Impacts: Induced-By . 169

Trenton Channel Project Secondary
Benefits Summarized . . . . . . . . 171

vi
Chapter

The Underlying Assumptions of the
Model . . . . . . . . . . . . . . .

Assumptions Common to Multiplier
Analysis. . . . . . . . . . . . .

Assumptions Common to Input-Output
MOdeISO O O O O O O O O O O O I 0

Additional Assumptions from Conventions
in Rent Model . . . . . . . . . .

VI. A COMPARATIVE ANALYSIS OF TWELVE SECONDARY
BENEFIT APPRAISALS. . . . . . . . . . .

Overview . . .
Trenton Channel Project

Michigan State University
Test Team. . . . . . .

Corps Of Engineers Test Team.
Stonewall Jackson Reservoir .

Corps Of Engineers Test Team.

Cornell University Test Team.
Poteau Watershed Project.

Corps of Engineers TSSt Team.

Soil Conservation Test Team . . . .
Mountain Home Project . . . . .

Corps of Engineers Test Team. . . .

University Of Wisconsin Test Team .

Effects of Differing Approaches on
Secondary Benefit Estimates . . . .

Secondary Benefits Compared to
Primary Benefits . . . . . . .

Regional Versus National Input-Output
Coefficients . . . . . . . . . .

Page

172

174

176

176

182
182
193

194
197
202
203
210
219
220
225
234
235
244

249

254

257

vii
Chapter Page
Summary . . . . . . . . . . . . . . . . 259

VII. SUMMARY AND CONCLUSIONS. . . . . . . . . . . 262

Review of the Study Objectives. . . . . 262
The Estimation Of Secondary Benefits. . 264

The Rent Model. . . . . . . . . . . 264
Recommendations for Further Study . . . 267

Some Further Refinements of the

MOdelo I O O I I O I O O O O O 269
Construct Local Input-Output

Models. . . . . . . . . . 269
Detail of Job Requirements . . 271

Use Linear Programming to
Restrain Gains. . . . . . 271

SOURCES CITED. . . . . . . . . . . . . . . . . . . . 274
APPENDIX . . . . . . . . . . . . . . . . . . . . . . 281

LIST OF TABLES

Table Page

1. After-Tax Profits Per Dollar of Sales
Compared With Before Tax Profit
Rates Reduced by Fifty Per Cent . . . . . 184

2. Trenton Channel Project: Outline of Test
Team Approaches to Secondary Benefit
Estimation. . . . . . . . . . . . . . . . 185

3. Stonewall Jackson Reservoir: Outline Of
Test Team Approaches to Secondary
Benefit Estimation. . . . . . . . . . . . 187

4. Poteau Watershed Project: Outline of
Test Team Approaches to Secondary
Benefit Estimation . . . . . . . . . . . 189

5. Mountain Home Project: Outline of Test
Team Approaches to Secondary
Benefit Estimation . . . . . . . . . . . 190

6. Estimates of Regional and National
Secondary Benefits: Comparison of
Rent Model and Approaches Used
by Others . . . . . . . . . . . . . . 192

7. Primary and Secondary Benefit Estimates
from Eight Appraisals: Regional
and National Accounts . . . . . . . . . . 255

viii

LIST OF FIGURES

Figure Page

1. Wage Determination and Rent Accrual
Where the Elasticity of Supply
Equals Zero (ES= 0) . . . . . '4' . . . 40

2. Wage Determination, Surplus and

Rent Accrual Where the Elasticity
of Supply is Variable [Eé = f(Qs)]. . . 44

ix

CHAPTER I
INTRODUCTION

In its report1 to the Water Resources Council in June
of 1969, the Special Task Force on Evaluation Procedures
presented a greatly expanded concept Of benefits to be con-
sidered in the evaluation Of Federal water projects. The
motivation for the report and the revision Of benefit con-
cepts was reportedly the passage in recent years of several
laws relating to the Federal water program.2 As a result of
these new laws, it was felt that the general guidelines pre-

sented in Senate Document 973

(the currently-governing poli-
cies and procedures document) would need to be updated to
encompass the "broadened" national Objectives signalled by
Congressional enactment Of the new laws.

The broadened national Objectives which were to be
fostered by the Federal water program were classified as 1)
national income Objectives, 2) regional development objec-
tives, 3) environmental objectives, and 4) well-being Ob-
jectives. It was proposed that each of these be made ex-
plicit considerations in water project evaluations, and
that the value Of each project be judged on the basis of

its contribution to one or more of these Objectives. It

was suggested that the Old distinction between "primary” and

2
”secondary" or "indirect" benefits be abandoned and that in
their place the concept of "contributions to the national
income Objective" be adopted. This was felt to be "con-
sistent with the evaluation principle requiring the identi-
fication and measurement of all beneficial effects . . ."
It was proposed that those benefits formerly referred to as
'Secondary" and "indirect" now be discussed in terms of
'production externalities" and "resource unemployment" bene—
fits,4 and that, rather than be treated only in side calcu-
lations, any contributions to national income from these
sources should be accounted for in the national income ob-
jective account. Production externalities were considered

to prevail when

There exist economies of scale in production for
industries affected by the project output when such
economies do not exist in any activities which may
contract. Under this condition production units may
~ be Operated at a more efficient level or interrelated
processes may be better coordinated including the
elimination of bottlenecks to the efficient expansion
gfother actTVities. _5' ‘—

 

 

Resource unemployment benefits could be expected to prevail

when

Otherwise unemployed or underemployed resources (labor,
fixed capital and resource capital) may be used or
better employed as a result of the developments oc-
Casioned by the project. This may come about as a
direct result of project construction, Operation,
maintenance and replacement or as an indirect result

of the use Of project outputs. This latter effect
occurs when the expansion Of industries directly or
indirectly using project outputs draw heavily upon
similarly unemployed or underemployed resources.

Though a potentially major addition to what was the

old "national efficiency" or primary benefit account was

3

being made, the national income Objective account was still
considered to be a principal Objective measure of project
value in light Of which contributions to other Objectives
were to be considered. The addition of environmental con—
sideration, regional development considerations, and con-
siderations concerning the well being Of people were added
as additional Objectives to be evaluated as competitors
wfith national income Objectives. However, the national in—
come objective account was also greatly expanded by the
addition of "secondary" impacts to that category of benefits.

The current status of theory, research methods and
data sources concerning the evaluation of secondary effects
was such that the Task Force acknowledged the difficulty
surrounding the evaluation of these impacts, especially that
of production externalities. Thus, while some general
recommendations concerning the analysis of unemployed re-
source benefits were made, no attempt was made to present
a format for the analysis of production externalities.

The "bottleneck" thesis forms the basis for much Of
the analysis Of secondary impacts, since it allows the
‘assignment of cause for secondary output expansion to the
water project (which supposedly eliminates the bottleneck).
While not all water resource agencies formerly included
the evaluation Of secondary benefits due to bottleneck
elimination in their analyses, explicit consideration was
often given to the elimination of production bottlenecks as

a rationale for public investment. The proposed evaluation

4
procedures would allow for the inclusion of the expansionary
effects Of such elimination as explicit project benefits.

By expanding the concept Of income effects to in-
clude indirect or secondary impacts, the Task Force sought
to comply with newly-enacted legislation which made re-
gional development an explicit goal Of the Appalachian
water resource program. Unemployed resource benefits
were expected to occur largely in the case of projects
built in depressed regions. Benefits from scale economics
would be expected to occur more readily in local economies
xflfich were Operating at very low levels Of output. The
employment gains from breaking production bottlenecks would
be greatest where large volumes of resources were unem-
ployed.

The division of the income effects into a Regional
Ikwelopment Account and a National Income Account made the
nation's regional deveIOpment objectives explicit and made
it possible to explicitly trade-Off national income for
regional income. However, it also made necessary techni-
ques for estimating both regional and national impacts of
water resource projects. In the past, only the Bureau of
Reclamation Of the United States Department of the Interior
(referred to hereinafter as the Bureau) had attempted to
evaluate indirect income effects of water projects. In do-
ing so, the Bureau considered that Congress had assigned
it the mission of developing the arid regions of the West-

ern United States; thus, regional development benefits were

5

treated as synonomous with national benefits. Transfer of
income to the region was a primary goal of the Bureau's
program, and little attempt was made to estimate the net
national impact or to determine the efficiency of the
transfer.7

The Bureau developed a "model" for evaluating the in-
direct effects of its reclamation efforts. The model con-
sisted of a set Of coefficients intended to represent the
value added in processing outputs and producing the pur-
chased inputs for the major agricultural commodities grown
in the region. These coefficients were used in estimating
the indirect income generated by the increased agricultural
output on farms using project-supplied water. The Bureau's
approach to benefit evaluation has been widely criticized.8
The Bureau's model would not appear to be a generally ac-
ceptable means for dealing with the expanded concept of

benefits under the new proposed procedures.

Towards the Appraisal 2£_Secondary Benefits

 

Many other "models" for appraising secondary benefits
have been proposed in recent years. These models can gener-
ally be divided into three groups: 1) multiplier models,

2) externality models, and 3) dynamic models (these three
groups will be discussed more fully in Chapter II). The
most popular has been the multiplier model (these are dis-
cussed in Chapter IV). The Bureau's approach has been

classified as a multiplier model.

6

Perhaps the most misused concept in economics has been
that of the multiplier. As a result, many prominent econo-
mists view with disdain the proposals to use multiplier con-
cepts in the appraisal of Federal water projects. This is
unfortunate since some aspects of multiplier analysis offer
great promise in appraising one kind Of secondary benefit
vflfich most economists would agree needs to be better ap-
praised: the real economic costs of factors used and Of
factors directly stimulated to use by Federal water projects.
The problem of appropriately shadow-pricing resources is
closely tied to the basic static multiplier concept. A
mmjor purpose of this study is to demonstrate the link be-
tween the very simple economic concepts of 1) Opportunity
costs, 2) economic rent, and 3) multiplier analysis, and to
develop a model which utilizes these concepts in appraising

static indirect benefits.

Objectives

 

The Objectives of the present study are

1. To survey, evaluate, and synthesize the current status
of theory regarding secondary or indirect benefits;

2. To identify the elements in Objective one above which
can be forged into a generally applicable model which
a) can be applied to all or most Federal water projects,
b) uses generally available secondary data,
c) makes explicit the assumptions and conditions under

which it is Operable;

7
To demonstrate and test the model by applying it to
four test case projects;
TO compare the model test results with the results ob-
tained by other test teams and to identify the reasons
for differing test results from the several analyses;
To evaluate the applicability of the model in light of

the results obtained in meeting objective four.

8

Footnotes

 

1. Procedures for Evaluation of Water and Related
Land Resource PrOJects: Report Ofithe Special TaEk Force
on EValuation Procedures to the water Resource'CounC1l.
(wash1ngton: June 1969)’.

 

 

 

 

2. Ibid., pp. 6-10.

3. United States Senate, 87th Congress, 2nd Session,
Policies, Standards and Procedures in the Formulation,
Evaluation, andIReview of Plans for Use andeevelppment of
Water and Related Land Resources, (Washington: 1962).

4. Procedures for Evaluation . . . , Op cit, 39-41.
5. Ibid., 42.

 

6. Ibid., 43.

 

7. George S. Tolley, "Impact of Public Resource
Development on Agricultural Production and Income," Pro-
ceedingsof the Aggicultural Economics and Rural SOCTOIogy
Section, Assoc1atlon of Southern AgricfilturaIIWO—kers,

8. See especially Tolley, Op cit.

CHAPTER II
SECONDARY BENEFITS DEFINED: A REVIEW

The term "secondary benefits” as generally used in
the water resources field refers to "The increase in the
value Of goods and services which indirectly result from
the project under conditions expected with the project as
compared to those without the project."1 While primary
benefits include only the value Of outputs which are pro-
duced directly by the project, secondary benefits include
the increased outputs (values) which result from the re-
percussions of the project outputs upon the rest of the
economy. Thus, while the analysis of primary benefits is
by nature micro-economic, the analysis of secondary bene-
fits has been considered macro-economic in approach.2

The concept of secondary benefits found in the lit-
erature is not monolithic. Much controversy persists among
economists regarding the nature, existence and means of
evaluation. A large part Of the disagreement stems from
the existence of more than one type of secondary benefit.
Most authors have not satisfactorily delimited the cate-
gory of benefits they were discussing nor the assumptions

under which the subject was being treated. Often authors

10
have mixed into their discussion more than one type of bene-
fit. The resulting confusion and disagreement between
authors has been extensive.

Basically, one can divide secondary benefits into
three different kinds: 1) the "customary" variety relating
largely to multiplier effects, 2) the "externality" variety
which relate mostly to economies of scale and the effects
of "spillovers", and 3) "dynamic" secondary benefits which
relate to changes in the shape of production functions,
changes in "the human factor" and basic institutional ef-
fects. These three categories are not clearly defined and
adhered to, as the following sections will show. For exam-
ple, it is not clear whether "technological externalities"
should go into the second or the third category. Further,
the treatment of aggregate demand multiplication could be
also viewed as a pecuniary spillover, which could place it
in the second category. Nevertheless, the tripartite
division is helpful in interpreting the literature regard-
ing secondary benefits. It is especially helpful in under-
standing the basis of the disagreements between different

authors.
The Customary View - The Multiplier

The customary approach to secondary benefits has
generally divided them into two classes, following largely
the convention of the Bureau Of Reclamation: 1) those

"induced-by" the increased factor requirements occasioned

 

11

by the project, and 2) those "stemming-from" the added pro-
cessing, etc., of the direct project outputs.3 The first
class can be sub-divided into a) those "induced-by" factor
requirements for constructing, operating and maintaining
the project, and b) those "induced-by" the complementary
relationship between the water-based goods and other factors
in the production of the direct benefits. An example of
the latter is the induced sales of fertilizer as a result
of increased water availability due to an irrigation proj-
ect. The ”induced-by" and "stemming-from" effects can be
classified as demand effects and supply effects,-respec-
tively.4 The former occur through the transmission Of
(derived) demand from the project outputs through the pro-
ducers of the factors which go into intermediate products
to meet those demands. The latter occur through the in-
creased availability Of factors to later stages in the
marketing process. Ostensibly, since the supply of factors
at the project-level stage has increased, the supplies of
outputs at successive stages will increase. Both of these
effects generate household income at each stage in the form
of wages and profits. The indirect benefit is the change
in real income realized as a result of the project.

Clark, Grant and Kelso (referred to hereafter as the
Panel) in their report5 to the Bureau of Reclamation advise
that, since supply and demand act jointly to determine in-
come, stemming-from effects and induced-by effects should

not be added in estimating the value Of secondary benefits

12

from the national vieWpOint. They argue that stemming-from
effects are a special form of supply effect which pre-
supposEs a demand for the increased output. Further, the
impetus of the two effects occur at different times in the
life of the project. Part of the aggregate demand effect
(induced) occurs as a result of construction expenditures,
while the supply effect (stemming) occurs after project
Operation has begun and project outputs are forthcoming,
often with somewhat of a lagged response.6 Also, since the
factors demanded by the project differ from the factors
produced by it, traditional macro-economic aggregate supply
and demand analysis tends to Obscure the multitude Of inter-
actions and transactions necessary to establish a new equi-
librium. The resource movements involved in those pecuniary
interactions make the determination of net income effects

very difficult to get at.
Indirect Benefits as a Pecuniary Externality

McKean7 discusses indirect benefits as a form of
pecuniary spillover or externality. He concludes that, in
the presence of full employment, the net national benefits
from such spillovers will be zero; the incidence of rents
will shift so that some individuals benefit and some lose,
but the net change in income in the absence of technological
spillovers or economies of scale will be zero. However, if
resources are unemployed or under-employed, there exist

possibilities for indirect benefits to occur. He does not

13
view such existence as a sufficient cause for a project to
create indirect benefits, however. Neither does their
existence solve the problem Of separating out the net
changes in real income due to employment of previously un-
employed factors from shifts in the incidence of scarcity
rents, where such changes may be widespread and individually
very small.

McKean argues that, generally, the project analyst
should include, in the national account, only technological
spillovers, and should only consider pecuniary spillovers
(to McKean, indirect benefits) if involuntary unemployment
of a very long range nature is expected to exist in the
absence of the project. He argues that in most situations
the time lag between project analysis and construction is
such that unemployment which existed at the analysis stage
might well have disappeared by the time construction begins.8

The McKean definition of indirect benefits appears to
only include aggregate demand effects and pecuniary ex-
ternalities. He treats technological externalities and
economies of scale under a second group of effects. "The
position taken [by McKean] . . . is that the two general
types of spillover -- technological and pecuniary -- are,

"10 He

in principle distinguishable and mutually exclusive.
does not believe that secondary benefits generally prevail
for the nation as a whole after all project effects are

netted between all resource owners and all regions. Only

under very special conditions of severe unemployment and

l4
factor immobility will any net positive effects occur for
the nation as a whole, and those will be much smaller than

traditionally claimed.11

The Secondary Benefit Controversy

There is some disagreement among economists concerning
the question Of the existence of secondary or indirect bene-
fits. While that portion Of economic theory is so under-
deVeloped as to allow for much controversy,12 perhaps some
part of that disagreement revolves around the confusion
surrounding the inclusiveness of the term ”secondary bene-

13 take a

fits." As pointed out above, McKean and others
very restrictive view of the term. The tendency among this
group is to think only in terms of aggregate demand analysis
which, under certain conditions, can lead to an increase in
national income. Another group, to which the Water Resource
Council Task Force seems to belong,14 view secondary bene-
fits as inclusive Of all possible economic effects beyond
the valuation of the direct project outputs, including

scale economies, technological externalities, aggregate
demand effects, aggregate supply effects, etc. Thus, the

latter group tend to look more favorably upon the possi-

bility of secondary benefits in water resource projects.
Primagy and Secondarnyenefits: Mutually Exclusive?

While the difference in inclusiveness of the defini-

tion Of secondary benefits contributes to the controversy

15
over real net national effects, it is not the only factor.
An additional element is the disagreement over the sort of
assumptions one should make about the future. Some analysts
assume that general fiscal policy and public programs will
be managed so as to promote steady growth, full employment,
and factor mobility. Under these assumptions, long range
resource unemployment problems disappear (or are obfuscated).
Thus, only the demand multiplier effects Of the construction
expenditures are perhaps relevant; and the length of the
planning period is such that fiscal management might have
already eliminated that source of unemployment benefits.15
The necessity for these assumptions is seen as the realiza-
tion that, in the absence of steady growth and "adequate"
levels of aggregate demand, the direct project outputs will
constitute excess supply; thus, the assumptions which give
value to the primary benefits are interpreted as negating
the possibility of indirect benefits.

The difficulties Of the unemployment-adequate demand
dichotomy which plague the analyst under the growth-
employment-mobility assumption can be handled through a
minor alteration in that assumption. The difficulty lies
in the all-or-nothing interpretation of the assumption-~i.e.,
that one can not have both a growth in demand and unemploy-
ment or immobility. Such a situation can exist, however.
Since the growth in demand is the principal factor upon
which the valuation of primary outputs is based, then one

can have both primary and secondary benefits prevailing, if

16

there is unemployment, as a result of factor immobility.

Thus, the assumption Of factor immobility has come to play

a large role in the justification for and evaluation of net

secondary benefits. Some discussion has developed over

the extent to which factor immobility should serve as a

basis for secondary benefit analysis. On the one hand, it

is realized that less than perfect mobility contributes to
market imperfections which can allow unemployment in some

areas or markets to coexist with inflationary demand in

crther areas or markets. Practically all analysts agree

tliat unemployed or underemployed factors constitute a

vwaste Of resources. However, the disagreement comes from

‘tlle interpretations concerning the temporal nature of the

Unemployment. Some argue that, if the unemployment and

market imperfections are short-run in nature, then the
(:1W3ation of indirect impacts in those areas via public in-

VeIStment would thwart the Operation of the incentive system

L1Don which the market system Operates.16 The effects of

SuCh a policy in the long run would be to undermine the
r“arket system. Others interpret many of the market imper-
:Ef3ctions as long run imperfections. Thus, the policy pre-

SSCFTiption is to pinpoint such areas and dose in public in-

\’€>Stments designed to create indirect employment Of the

()1iherwise long-range unemployed.

17

Secondary Benefits - The "Externality" Approach

Margolis argues that the Bureau Of Reclamation program
cannot be justified on the basis of primary benefits alone.
Ruther, the traditional "stemming-from" and ”induced-by"

sort Of approach, when properly done, would not in most

cases lead to a justification Of most irrigation projects.
However, when one considers the possible external economies
due to denser settlement, economies of scale due to larger

lxocal markets, improvements in transportation, etc., the

Irregram of the Bureau becomes more justifiable. Thus,

Nkargolis concludes, "The proper framework within which to

(iirscuss secondary benefits is the theory of external econo-

nflires."17 He would not limit the discussion to pecuniary ex-

17—el‘it‘na1ities but would include other forms Of externalities

as well.

The Water Resource Council Task Force, in effect,

1nStitutionalizes the inclusion of externalities in the

(1‘3ifinition of secondary benefits.18 They apparently accept

1:}IGB judgement that net national secondary benefits do not

()Slcnir from project investments in the absence of unemploy-

Inern: or external economies. Thus, "secondary benefits will

theSiult in contributions to national income when either of

1:}1€> two general conditions are expected to prevail." These

gel'leral conditions are:

l. Externalities--There exist economies of scale
in production for industries affected by the project
output when such economies do not exist in any activi-
ties which may contract. Under this condition

18
production units may be Operated at a more efficient
evel or interrelated processes may be better coordi-

nated including the elimination of bottlenecks to the
efficient expansion of other activities.

2. Resource unemployment--Otherwise unemployed
or underemployed resources (Tabor, fixed capital and
resource capital) may be used or better employed as
a result of the developments occasioned by the project.
This may come about as a direct result of project con-
struction, Operation, maintenance, and replacement or
as an indirect result of the use Of project outputs.
This latter effect occurs when the expansion Of in-
dustries directly or indirectly-using project outputs
drawn upon'igmilarly unemployed or underemployed

I

TESOUI‘CBS .

Under the "Externalities",heading the Task Force lists
‘two general conditions under which the public investments
Inay generate net secondary benefits. The first is that of
ciifferent economies of scale between contracting and ex-
IDanding industries. In terms of benefits, the real dif-
fierence between the net values of the outputs would appear
tile relevant concept. The Task Force realized the diffi-
Clulty of determining and measuring benefits due to scale
economies. and did not present a format for evaluating
tfllem.20

The second category of externalities relates to the
eilimination of bottlenecks. Secondary benefits would exist
jfif the project directly or indirectly stimulated the ex-
IPEInsion of output Of some factor which had effectively
Llfiimited the expansion of other activities. Such a situation
might occur in a water supply project in which the water was

‘1 critical input into one product which was a critical in-

I3111: into a second product. In this case, the provision

19

of water might stimulate the expansion of both activities,
given the existence of effective demand for the outputs of
both. The degree to which the water-related product con-
tributed to such expansion and the extent to which one
might credit the added income to the project would depend
not only upon the marginal rates of substitution between
the "critical" factors and other inputs but also upon the
assumptions made regarding alternative means of providing
those factors, and the development of substitutes (in-
cluding a substitute technology), etc. All of the ex-
Ipansion would be creditable only under the very restrictive
Eissumptions of 1) zero substitutability between factors
13rovided at all levels of expansion, 2) no alternative
Ineans of providing the initial critical factors, and 3) no
Ilew'technology which would effectively substitute for the

Ciritical factors would develop during the life of the

Irroject. In addition, since the bottleneck elimination is

1-argely a supply effect, it is assumed that effective demand
will exist for the outputs over the relevant period.

In actuality, the bottleneck elimination benefits

Efre very much akin to the traditional "stemming-from" bene-

fits in the Bureau of Reclamation procedures. For the stem-

ming benefits to be valid, the same process of reasoning
and assumptions outlined above would be necessary.
In the case of the Bureau, the basic project output

:155 irrigation. The principal direct benefits come from in-

C1‘eased agricultural output. There is little doubt that

20
agricultural outputs are the basic factor in a food proces-

sing operation and that little factor substitution is

possible, except between agricultural commodities. The

same is true for succeeding operations to the final market.

Thus, the Bureau procedure has been largely in keeping with

that part of the requirements. However, a basic issue has

been overlooked. That is, does a shortage of agricultural

outputs exist which serves as a bottleneck to the expansion

(of food processing and succeeding activities? The apparent

zrnswer in light of the agricultural surpluses is that, in

‘teamw of national agricultural output, such a bottleneck

(ices not exist. Whether or not such a bottleneck exists

ill terms of regional expansion is perhaps another issue,
llowever -- an issue which belongs in the discussion of the
regional development effects of indirect benefits, since
1311ch.expansion in the project region would be expected to
53$! offset by a contraction in some other farming region.21
(3’1 the other hand, if there did exist a shortage of agri-
<2llltural products, nationwide and if that shortage did
ZEFDrce the Operation of food processing plants at inef-
:Eficient levels, then net national indirect benefits could
t3€3 realized as a result of such irrigation projects. Thus,
‘tlle validity of the principle stands, in spite of the im-
I)‘JT‘Oper manner in which it has been applied on occasion.
The Resource Unemployment benefits rightly belong

under the customary variety of secondary benefits discussed

€113<rve which view secondary impacts as largely a phenomenon

21

of changes in the level and structure of aggregate demand.
While the externality benefits largely presuppose an ade-
quate level of effective demand for the final outputs, re-
source unemployment benefits assume a deficiency in demand.
As discussed above, the assumption is not necessarily that
the overall level of demand is deficient. Rather, the
deficiency might be the inability of the market to transmit
those demands to the unemployed factors, or in the structure
of those demands relative to the structure of productive
czapacity. However, the benefits from changing the struc-
‘ture of aggregate demand to ease the unemployment problems
vvould possibly be traded for some decrease in incentives
tcn resource mobility.22 It would also alter the structure
£Ind.incidence of income flows. Indirect benefits from
'Vresource unemployment" would occur only if the resources
wOuld not otherwise have found employment. While the re-
structuring of aggregate demand via Federal expenditure
IDCXlicy can lead to an acceleration of economic growth,
ECkstein urges resistence to ”pressure group expenditures"
‘Vllich would reduce factor mobility and reduce the long run
I‘ate of real growth.23 To guard against the possibility
(’13 such long-run negative effects, many economists argue
‘tllat only resource unemployment of a very long run nature
Should be considered.
Secondary benefits due to resource unemployment apply
r10t only to those arising from the employment of otherwise

llrbemployed labor but also from the employment of otherwise

22

unemployed plant capacity (capital), as well as under-
employed labor and capital. Under-employment occurs when

a factor is employed, but not in its highest-valued use.

If the project directly or indirectly leads to full employ-
ment (or less under-employment) of such factors, then the
change in returns to those factors would constitute a

secondary benefit (if, in the process, a decline in the

employment of other factors is not caused). Secondary bene-

fits might result from fuller utilization of otherwise un-

enmfloyed fixed assets,24 in which case the benefits would

<:onsist of the returns above variable and user costs stimu-

Slated by the project.

"Dynamic" or Developmental Secondagy Benefits

A third class of secondary benefits has developed in

IVacent years, though it has not joined the current contro-

‘VWBrsy regarding the other two varieties. "Dynamic Secondary

Benefits" is the title which has been bestowed upon them.

'Ijlis variety of benefits has to do with the "dynamic social

engineering aspects of resource development."25 It finds

its basis in such things as the improvement in the labor

f7<Drce and in technological advance. Thus, it is much more

difficult to conceptualize than the customary variety.
Also falling under this third heading is the socio-

€3<Zonomic effects of moving the center of population in the

IJIlited States westward.26 The Bureau of Reclamation con-

Si~ders such a movement to be a part of the public mandate

23

under which it Operates.27

Thus, the Bureau has had as a
part of its "established procedures" in project evaluation
a policy of crediting a given dollar value to each new
farming opportunity created by its projects. Ostensibly,
the Western United States constitutes a more "desirable"
area in which to live than other sections of the country.
Thus, real welfare Of the populace can be improved by
creating income-earning Opportunities in the West allowing
people to move to more ”desirable" residential areas.
Another category of benefits sometimes placed under
this class of secondary benefits is that relating to re-
source conservation. It is argued that Federal natural
resource projects,by providing more and longer duration ad-
ditions to capital stock than the market would demand,
creates secondary additions to the real welfare of those
who favor such policies and also, ostensibly, to future
generations who would choose a higher rate of long term in-
vestment and a lower rate Of consumption by previous gen-

28 Such benefits

erations, if such a choice were available.
are not easily analyzed and, consequently, are not normally
included in primary benefit calculations, though some have
argued in the past that the discount rate used in Federal
water project evaluations constitutes just such a considera-
tion. While this rate has been much lower than the market
rate in the past, the discount rate used for water project

evaluation in the future will be much closer to the market

rate of interest.

a
(s,
r

1"

1‘-

‘”l ‘
Q4

 

24

The ”dynamic" secondary benefit category has tended
to be a catch-all for the multifarious benefit categories
which were difficult to evaluate. The fact that conserva-
tion and population dispersion are not necessarily "primary"
goals of the Federal water program has created the impres-
sion that they are "secondary" benefits. Only in so far as
they contribute to changing the long-run aggregate pro-
duction function or to "socially engineering” the human
factor should they be considered "dynamic" secondary as-
pects of project development. Thus, only remotely do many
of these factors fit into the dynamic secondary benefit

category.
Summary of Secondary Benefit Classifications

The literature on secondary benefits discussed above
demonstrates substantial controversy regarding the concept.
The greatest controversy surrounds the customary variety
of benefits. Most authors believe the existence of un-
employed factors to be necessary for net national secondary
benefits to occur. Unemployment could result from either
a deficiency of the overall level of aggregate demand or
from resource immobilities. Some authors believe that only
the former factor should be dealt with in public projects,
since interference with the latter would lead to a distor-
tion of the incentive system. They believe that in the long
run the factors would reallocate to accomodate the changed

structure of prices, if an adequate level of overall demand

25

were maintained. Others believe that public interference
should depend upon the length of time required to reallocate
the factors, especially the human ones. They are willing

to forego some of the incentive effects to relieve unem-
ployment and resultant human suffering. All of these deal
with secondary benefits from changing the level or the
structure of aggregate demand, and they tend to view sec-
ondary benefits as arising primarily from this source.

The "externality" approach includes both supply and
demand aspects, though most authors do not make this ex-
plicit. The "bottleneck" thesis is a good example of the
case in which both derived supply and derived demand are
thwarted by the shortage of a commodity. The case of scale
economies is largely a supply effect which presumes a de-
mand for the marginal output from the scale effects. Scale
economies represent the point at which the externalities
approach begins to blend into the "dynamic" secondary bene-
fit category.

"Dynamic" secondary benefits are viewed largely as a
supply effect. They operate through changes in the quality
of resources and through changing the institutional and/or
technological milieu in which the resources are operative.
Frustration with past "theories” of growth and with aggre-
gate demand multiplication as a means of economic develop-
ment has led to increasing attention being given to the

"dynamic" factors.

26

The interaction of long-run supply and demand under
conditions of changing technology and institutions has
created problems in dealing conceptually with secondary
benefits. The domination of economic growth theory by
Keynesian analysis in recent years has led to confusion as
to what constitutes real growth and how it should be
fostered. Many secondary benefit discussants have dealt
largely with aggregate demand effects while making dif-
fering assumptions about aggregate supply and its respon-
siveness. This factor alone has led to substantial con-
fusion and controversy. Others dealt with nebulous con-
cepts regarding changes in productive capacity while as-
suming adequate administration of aggregate demand effects
of public projects without adequately specifying their as-
sumptions regarding the method of financing by the public
sector. This has contributed to further disagreement as
to the "net" secondary effect of public projects. The fol-
lowing section deals with the effects Of differing assump-

tions regarding the means of financing public projects.

Secondary Benefits: The Budget and the Economy

 

Much of the discussion Of the "netness" of secondary
benefits has centered around the relationship of the project
to the overall Federal budget. Quite obviously, the aggre-
gate demand multiplier effects of the construction, opera-
tion and maintenance expenditures will depend greatly upon

the method used to finance them. Keynesian multiplier

27
theory makes it very clear that the multiplier effects Of
a tax-financed public investment are smaller than those of

a deficit-financed public investment.29

If the project
is assumed to be tax-financed, then it appears reasonable
that the funds taxed away would also have had multiplier
effects associated with them in the private sector which
should be subtracted from the multiplier effects Of the
government investment to determine the net multiplier ef-
fects of the tax-financed expenditure. However, it can be
demonstrated that the multiplier effects of tax-financed
public investment in the aggregate exceed those of the
private sector (in the absence of a budget surplus policy).
The systematic difference between the two multipliers has
led to the "Balanced Budget Theorem" which states, in
effect, that government expenditures financed wholly by
taxes will have a net multiplier effect upon the level of
aggregate demand. The difficulty, however, is that under
a very complicated taxing structure (including built-in
stabilization mechanisms) such as that of the United States
the exact size of that net multiplier at the margin of ex-
penditures is somewhat difficult to determine.30
Many economists have argued that the secondary bene-
fits (especially those induced-by the construction, and
operation and maintenance expenditures) created by the
project are at least partially offset by l) the contrac-

tions in the private sector caused by the taxing, or 2)

secondary impacts which would have resulted from any other

28
expenditure of those fiscal dollars. Thus, it is argued
that secondary impacts should not be included in project
analysis. Underlying such an argument is either 1) an
implicit assumption that the secondary impacts of all of
the alternatives are equal to those of the project, or 2)
a belief that the effects of each are so difficult of
determination and estimation that their inclusion would in-
validate the analysis. (A third argument for not including
them relates to the negative effects of distorting incen-
tives, discussed above.)

Eckstein argues that net national indirect benefits
result only under very limited circumstances. Those cir-
cumstances relate basically to the state of the economy
at the time. Induced-by benefits would be expected "In a
time of depression, when national output is far below
capacity and there are idle resources diffused through the
economy . . ."31 These would come largely through the
multiplier effects of the construction expenditure, since
"Induced effects from the operation of the project are
likely to be either very small or non-existent during a

d - "32
epre551on.

Because commodities are already in over-
supply, the increased output from the project will offset
the induced effects of the project, since the income pay-
ments to the project-related producers will replace in-
come payments to the project-related producers will re-

place income payments to other producers. However, if the

products produced are surplus agricultural commodities

29

which will be absorbed by the Federal Budget, then there
will be an additional induced benefit from Operation of the
project, but it will be due to the increase in the govern-
ment deficit resulting from the price support program.
Thus, Eckstein concludes," . . . the Operation of projects
does not create induced benefits in depression other than
the effect of the price-support program, but . . . the
construction of projects does induce substantial amounts."33
While Eckstein does not make the distinction between tax-
financed and deficit-financed projects in his discussion,
the application of his analysis of induced effects can be
made to both alternatives, the difference being in the
magnitudes of the net multiplier effects at the construction
stage. In his discussion of secondary benefits during
periods of economic balance, Eckstein does distinguish
between the two means of finance but concludes that the
net effects would be to stimulate an acceleration of
prices, thus worsening the situation.

If one accepts the counsel of the Panel of Consul-
tants to the Bureau of Reclamation that stemming benefits
are supply effects while induced benefits are demand ef-

fects,34

then one should expect the benefits from the two
categories to occur in obverse fashion over the business
cycle. That is, it would appear that stemming (supply)
effects would contribute to real income during inflation,

while induced (demand) effects would contribute to real

income during recession. Eckstein agrees that ". . .

30

stemming benefits are very unlikely in depression, are a
possibility during inflation if the specific commodities
are in short supply, and can only be granted for periods of
economic balance in those instances where the premises
of mobility can be denied because of extraordinary cir-
cumstances. The routine calculation of stemming benefits,
therefore, is not warranted."35
Eckstein does not make explicit his assumptions con-
cerning the state of the budget in each of the cases dis-
cussed. He largely ignores any off-setting effects from
private sector contractions due to taxing or bond sales.
Thus, it is not clear in his treatment of induced benefits
from project operation during depression, for instance, as
to why there will be no positive real income effects. If
the operation and maintenance costs are tax-financed
(financed from the budget rather than from project bene-
fits, which will not be recaptured by the relevant agency),
there appears to be no reason for assuming that the project
output will merely replace non-project outputs in the
market. In this case, the balanced budget multiplier should
have raised the level of aggregate demand so that at least
part of the new output could be purchased without reducing
the demand for non-project outputs. That is, the overall
level of demand should have increased rather than merely
shifted in incidence. Thus, the real income effects of
operation and maintenance expenditures for most Federal

water projects should be to raise the level of aggregate

31

demand. In times of recession, this would appear to have
real welfare implications.

McKean points out that a difficulty in predicting
the secondary effects of a project upon unemployment lies
in the fact that the analyst has no foreknowledge of the
rest of the Federal Budget. It is important to consider
what would have happened to taxes and expenditures in the
absence of the project.36 Thus, the problem of piece meal
sub-optimization in terms of secondary benefits is no less
substantial than that in terms of primary benefits.37 He
notes elsewhere, however, that "In many problems of choice,
the size of the budget or the scale of the mission is
fixed by higher authority."38 Thus, the question of secon-
dary alternatives forgone is obviated, at best. On the
other hand, Buchanan counsels that the expenditure-taxing
decisions THEE be considered to be made simultaneously,"
. . . the overall size of the budget cannot be separated
from its composition. The two decisions are inter-

39 Thus, the question of secondary benefits is

dependent."
not independent of the taxing process nor of the overall
composition of the budget, including other mutually-
exclusive alternatives. The determination of the real
net changes in the welfare of the peOple involved cannot
be made in isolation from these considerations. The

"involved" people include both those paying for and re-

ceiving the benefits from the public projects.

.xc . EL -

«IN. . ., s. x ..

\

a.~
5...»

.f-
—‘

. ‘

\Lr‘

.\

32

Footnotes

 

1. Procedures for Evaluation --, Op cit, p. 40.

2. Development of Water Resources in Appalachia,
Main Report - Part IV (draft). United States ArmyCOrps
of Engineers, Office of Appalachian Studies, December 1968.

3. Julius Margolis, "Secondary Benefits, External
Economies, and the Justification of Public Investment,"
The Review of Economics and Statistics, Volume 39 (1957),
pp. 284-291.

4. John M. Clark, et al, Rpport of Panel of Consul-
tants on Secondary or Indirect Benefits of Water-Use
Promects to Michael W. Straus,rCOmmissioner Bureau of
Rec.amation, June 26, 1952.

 

 

 

 

5. Ibid.
6. Ibid., pp. 20-21.

7. Roland McKean, Efficiency in Government Through
Systems Analysis: With Emphasis on—Water Resources Develqp-
ment. (New York: 1966)

 

8. 1219', p. 161.

9. .LEiE°’ pp. 154-55.
10. lEl§°: p. 143.
ll. LElE°: p. 159.

12. See, Julius Margolis, "Secondary Benefits, Ex-
ternal Economics, and the Justification of Public Invest
ment," The Review of Economics and Statistics, XXXIX (1957),
284-91.

13. See Roland McKean, "Criteria Of Efficiency in
Government Expenditures," Federal Expenditures Policies for
Economic Growth and Stabilization, U} 8. Congress (wash-
ington: 1957); reprinted in Public Finance and Fiscal
Policy: Selected Readin 5, ed} by J. SCherer and J. A.
Papke (Boston: 1966), - 04; John F. Due, Government
Finance: Economics of the Public Sector (Homewood: 1968);

TH} H. Stoenever and E. N. castle, "Input-Output Models
in Water Resources Research," Journal of Farm Economics,
(Nov. 1965), 1572-79;

U. 8. Congress, Economic Analysis and the Efficiency of
Government, Report of the Sub-committee on‘Economy in

 

 

 

 

33

Government of the Joint Economic Committee, 915t Cong.,
lst Session, (Washington: 1970), 25-28.

14. See Procedures for Evaluation---, op cit;
Margolis, op cit;
Allen V. Kneese, Water Resources . . . Development and
Use, (Kansas City: 1959).

 

 

15. See particularly McKean, Op cit; and Margolis,
Op cit, and

16. Otto Eckstein, "Federal Expenditure Policy for
Economic Growth, " Public Finance and Fiscal Polipy:

Selected Readings, ed_by J. SEherer and J’. A. Papke, op cit,
S71-81.

 

l7. Margolis, op cit, 284.

18. Procedures for Evaluation ---, 42-43.

 

19. Ibid.

 

20. Ibid., p. 101.

21. For example, Tolley, o cit, concludes that the
effect Of the Bureau program has Eeen to transfer income
from Southeastern farming regions to Southwestern farming
regions.

22. Otto Eckstein, "Federal Expenditure Policy for
Economic Growth," Public Finance and Fiscal Poliqy:
Selected Readipgs, edited by J. Scherer and J. A. Papke
(Heston: 1966), 574-81.

 

23. Ibid., 578.

24. See Clark Edwards, "Resource Fixity, Credit
Availability and Agricultural Organization," unpublished
PhD. thesis, Michigan State University, 1958.

25. Kneese, Op cit, 27.

26. Ibid.

27. Margolis, Op cit, 284-91.

28. Kneese, op cit.

29. J. M. Keynes, The General Theory of Employment,
Interest and Money (London: 1964),“lZS-28.

 

30. E. Cary Brown, "The Personal Income Tax as an
Automatic Stabilizer," Tax Revision Compendium, Hearings

34

before the Committee on Ways and Means Of the U. S. Congress
(Washington: 1959), 2357-62.

31. Eckstein, Water Resource Development: The
Economics of Project Evaluation (Cambridge: 1965), 206.

 

32. Ibid., 207.

 

33. Ibid., 207-8.

 

34. J. M. Clark, et al, "Report of Panel of Consul-
tants on Secondary or Ind1rect Benefits of Water-Use
Projects to Michael W. Straus, Commissioner, Bureau of
Reclamation," June 26, 1962.

35. Eckstein, loc cit, 211.

36. McKean, Efficiency in Government . . . , 161.

 

37. McKean, "Criteria of Efficiency . . .", 99.
38. Ibid., 101.

39. James M. Buchanan, The Public Finances (Home-
wood: 1965), 232-3.

 

CHAPTER III

SECONDARY BENEFITS, ECONOMIC RENT, AND THE
COMPENSATION ABILITY PRINCIPLE

The basic issue in the discourse concerning second-
ary benefits is dual in nature: first, are there income
gains to some parties which in total exceed the losses to
other parties? And, secondly, are those estimated changes
in the income of the gaining parties net, such that they
would be able to pay for the privilege of receiving them?
The first aspect of the issue relates to the reality of
net national secondary benefits, while the second relates
to that of net local (or to the individual) secondary bene-
fits. The implications Of the latter should make it ap-
parent that, while much attention has been given to the
existence of offsetting income declines in other regions,
there is a very strong possibility that the generally-
cited gains might not represent even net local gains. This
is true because of a common failure to include all of the
relevant opportunity costs which exist at the local level.
Agreement between most parties to the discourse concerning
indirect benefits has been accomplished in only the very
special circumstance of substantial factor unemployment.

Where substantial unemployment of a long-run nature has been

35

36

evident, most authorities have agreed that indirect bene-
fits should be considered. The common practice has been
to treat the entire payment to the otherwise unemployed
factor as a benefit (or, inversely, to subtract it from
costs). There are two difficulties with this approach,
however. First, it is very restrictive and only admits of
the existence of indirect benefits in one special case.
Secondly, the benefit gains cited in such cases are prob-
ably overvalued because of the implicit zero-valuing of the
opportunity cost to those factors realizing the gains. In
the case of labor, for example, it is doubtful that one
could ever logically expect a zero Opportunity cost to exist
for two reasons: first, in terms of true welfare effects,
such a procedure imputes a zero value to leisure;1 secondly,
as Joan Robinson points out, labor cannot survive in the
alternative in the long run if ”transference earnings" are
zero.2 Thus, even in the restrictive case upon which agree-
ment has been achieved, the question of adequately handling
the problem of the opportunity costs in terms of real wel-
fare changes has not been resolved. The obvious existence
of some degree of indirect benefits in those circumstances
has led protagonists to allow the consideration of indirect
benefits in the extreme case, even though many have felt
that the true opportunity costs were something greater than
the zero value conventionally imputed to them.

For indirect benefits to be relevant and helpful in

comparing projects for investment decision purposes, it is

37
important that some consistency in treatment be incorporated
into project analyses. All public projects produce some
kind of indirect economic effects. In many cases, the in-
direct effects merely take the form of transfers in which
one party benefits at the expense of another. From a
national viewpoint, perhaps no new ppp income results.
From a regional viewpoint, net changes might well occur,
however. In other cases, the incomes of certain individu-
als or regions might change in form without changing in
total. For example, an individual who had previously
managed a bowling alley might change to marina management
as a result of the public water resource investment.
Whether or not he benefited from the shift would depend
upon the changes in his net returns. If he realized
$10,000 net income per year in the bowling alley business
and $12,000 in the marina business, the net benefit would
be $2,000 per year (not $12,000). By the same token, he
(or one of his employees) might have realized a decline in
income from the shift. Such would have been the case had
the project caused a shift in recreation consumption away
from bowling and into boating. His shift into marina
management might have still been his best alternative,
even though a lower income alternative than his pre-
project bowling enterprise. In this case, the relevant
benefit to the individual would be negative. In either
case, a benefit evaluation procedure which did not deduct

the income from the pre-project bowling alley would greatly

38

over-value the positive income effects. In each case, the
key to coming up with the apprOpriate benefit figure lies
in determining the relevant Opportunity costs of the
factors involved in the shifts. As discussed above, the
implicit assumption in much of the discourse concerning
the evaluation of indirect benefits has been that the op-
portunity costs are equal to the new income earned in all
cases except that of substantial long-run factor unemploy-
ment. This convention ignores the possibility of major
net benefits occurring in the case of factor under-
employment. The fact that a secondary impact of a public
project which moves a previously employed laborer from a
job paying $3,000 per year into one paying $9,000 per year
is excluded in some computations, while the employment at
$3,000 per year of a previously unemployed worker is in-
cluded, points up the difficulty in such a procedure.
While one might well argue that the marginal utility
of $3,000 at zero income is very high, one could also make
a very plausible argument that the marginal utility of
$6,000 at a $3,000 income is relatively high. Thus, at
least in the range of such figures as those used above, it
would appear that the marginal utility of income argument
would not constitute a sufficient case for the method in
which all adjustments except that made by the unemployed

are ignored.

39

Wage Changes and Economic Rent

When a worker moves from one job to another, it is
usually a result of a perceived Opportunity to increase
the "net utility" derived from the work experience. In
the simplest case, a worker might shift from one job with
a relatively low wage to an identical job which yielded a
higher wage. Presuming that the "disutility" of work was
the same between the two jobs, the worker would view the
change in wages as an "unearned increment" to his income.
Economists generally refer to such unearned increments as
”economic rent."3 There is general agreement among econo-
mists that changes in economic rent are largely a result
of demand changes. Demand is viewed as the "active deter-
minant" of rent, while supply is viewed as the passive
determinant. Such analysis is based upon the assumption
that the supply of the factor under consideration is per-
fectly inelastic -- that is, regardless of the price of-
fered for the factor, the same quantity would be offered
for sale by factor suppliers. Any price above zero would
yield a "rent" to the factor supplier. The higher the
price, the greater the rent.

In Figure l, the demand and supply of labor for an
individual worker is depicted under the assumptions of
perfectly inelastic supply and three different demand
conditions. The vertical supply of labor curve (SL) in-
dicates that the laborer would work the same number of

hours per year whether his annual salary were zero, $2,000

40

 

 

 

.Annual Worker A
Wage SL
Rate
D3
$2,000 L-un-u- -------
D2
0 LS Quantity of
Work Supplied
D1 Per Year

(Q5)

Figure 1. Wage Determination and Rent Accrual Where the
Elasticity of Supply Equals Zero (E5 = 0).

41

or $5,000. As the demand for the laborer's services shifts
from the level at which an employer would be willing to
pay zero or less for the laborer's full year of work (D1)
to a demand situation in which the employer would be willing
to pay $2,000 for the number of hours of work which this
laborer would supply, the quantity of labor supplied does
not change. Assuming for the moment that there is neither
a "like" nor a ”dislike" for work on the part of the
laborer, the depicted supply curve would indicate that the
marginal value of leisure to this laborer is zero at all
levels of leisure consumption. (A supply curve which
sloped upward and to the right would indicate that the
marginal utility of leisure diminished as more leisure
was "consumed" -- or, alternatively, that the marginal
utility of leisure increased as more hours of work were
undertaken.)

The generally used method of multiplier analysis
treats the benefits to labor in essentially the fashion
depicted in Figure l -- in particular the demand change

illustrated by the shift from D to D2. Prior to the in-

l
crease in aggregate demand indicated by the multiplier im-
pact, the demand for the services of worker A was D1, which
was insufficient to bring about his "full employment," even
though he would have been willing to work for a zero wage.
The aggregate demand change shifts the demand for the

worker's services to D2, which would be sufficient to bring

about his "full-time" employment for $2,000 per year. The

42

$2,000 is treated as an economic rent to the laborer. The
demand for a factor of production such as labor is deter-
mined by the marginal revenue product (MRP) of that factor.
The MRP Of a factor is the dollar value of the addition to
total output which is brought about by an increase in fac-
tor employment. Under pure competition, and in the absence
of technological externalities, the MRP of labor repre-
sents the addition to the value of the output of the
ecOnomy resulting from the employment of one previously
unemployed laborer. Thus, the "rent" realized by the
laborer would also be viewed as a net addition to the value
of the output of the whole economy and would be added to
the "national" benefits of the project.

There are two basic problems with the assumptions
which underlie the approach to multiplier analysis discussed
above. First, it is probably incorrect to assume that
workers neither like nor dislike their jobs. Some workers
appear to enjoy their work a great deal. Many of these
would perhaps perform the same job for less pay, without
decreasing their productivity. Thus, for these workers,
over a particular price range their labor supply functions
might be close to perfectly inelastic. Other workers ob-
viously dislike at least some aspects of their particular
jobs. Many "tolerate" the work performed only because Of
the salary received. The supply function for the services
Of these latter workers probably exhibits a great deal of

elasticity for that particular job over the prevailing

43

price range. Those workers who enjoy their jobs might be
said to be realizing a "surplus" or a ”rent" equivalent to
the difference between the wage they would accept for the
same work and the wage which they are actually paid. The
worker who was "tolerating" his work might be said to
receive no rent, if he would take his services off the
market at any perceivably lower wage rate.

A second factor assumed away in the analysis usually
undertaken by multiplier analysts regards the marginal
value of leisure. In few cases would the marginal value
of leisure be zero at all levels of leisure consumption.
Rather, one should expect the marginal value of leisure to
increase as work time cuts deeper and deeper into leisure
time. Thus, the supply function for labor would appear to
slope upward 32d to the right, with the lepe increasing
as hours worked increased. Such a situation is illustrated
in Figure 2. The intercept of supply function SL depends
upon the marginal value of leisure to the totally unem-
ployed worker. In the case illustrated, the marginal value
of leisure is assumed to be zero for all levels of employ-
ment below ”A" days per year. Because the marginal value
of leisure increases as the quantity of leisure decreases,
the worker must be paid higher and higher wages to induce
him to increase the number of days worked per year. As the
wage rate increases, the worker realizes a "surplus” for
those days on which he would have supplied his services for

a lower payment. Presumably, after some level Of days

Annual
Wage
Rate

44

Worker B

L--------------

‘

  

 

Figure 2.

A Quantity of

 

Work Supplied
Per Year

(Q5)

Wage Determination, Surplus and Rent Accrual
Where the Elasticity of Supply is Variable

[Eé = f(QS].

45

worked per year, the marginal value of leisure will become
very great, and/or there will be no remaining leisure time.
Additional increases in wages will not bring forth a
greater quantity of labor. The increased wages will accrue
to the laborer as ”pure surplus" or ”rent." Thus, only in
the upper wage range would the depicted laborer realize the
”pure rent" described by the multiplier. At lower income
levels, only a "suppliers surplus" would be realized by the
laborer. The extent of the surplus would depend upon the
elasticity of supply of labor.

From the foregoing it would appear that the ”pure
rent” imputed to labor from the usual multiplier analysis
should be expected to be more a phenomenon of fully em-
ployed labor, rather than unemployed labor. That is, when
a change in aggregate demand leads to labor movement to
higher paying jobs, the difference in wages accrues largely
as "rent" to the worker. (The size of the rent component
of the change in wages would depend upon the degree to
which the wage differences resulted from differing psychic
costs between the two jobs. If the psychic costs were
equal, all of the wage difference would be rent.) Only in
those cases in which the economic and psychic costs of
the move were zero or very small would the worker be will-
ing to pay a large part of the income difference for the
opportunity to make the shift.

The occurrence of secondary benefits can relate to

unemployed plant capacity as well as to unemployed labor.

46

If the plant has no alternative use, then the opportunity
cost of the fixed factors would be zero. Thus, the relevant
benefits from induced utilization of the plant would con-
sist of gross revenues minus variable costs. In Marshallian
terminology, this difference constitutes ”quasi-rents."4 For
any such plant which was operating, but at less than "full
capacity", benefits would accrue to any induced increases
in output according to the changes in ”quasi-rents" earned,
including profits.

Secondary benefits from rents at the extensive margin
for plant capacity could occur when the construction of a
new plant was stimulated by the public investment. In this
case, the rents would consist only of pure profits earned
on the new plant investment, since the long-run concept of

5 In the case of fixed plant

rents would be relevant here.
employment as in the case of labor employment, however, con-
sideration must be given to the decline in rents of other
factors. Thus, if construction of the plant discussed above
took place to the exclusion of the construction of some
other plant, the net benefit would consist of the difference
in profits (rents) between the two plants. If public in-
vestment stimulated a shift in investment from lower to
higher profit plants, net economic welfare would presumably
be enhanced. The enhancement would be by the amount of the

difference in pure profits (rents), however, rather than by

the full extent of the new plant profits.

47

Economic Rent and the Compensation Ability Principle

In the evaluation of secondary benefits, as in that
of primary benefits, the ”Compensation Principle" must hold.
That is, the value of the benefits estimated should amount
to no more than the recipients could and actually would pay
for them. Thus, in terms of real welfare changes, the
actual secondary benefits would be equal to the change in
rents realized by factor owners, for that is all that they
would be able to pay. Attempts to make them pay the "full
value" of the new alternative rather than the differential
would be confronted by an inability to do so. In many
cases, it would appear, secondary benefits which had been
thought to be net to the local area would not he opted for
by the local populace, if the project cost were charged to
them. This is true because of the methods by which sec-
ondary benefits have been evaluated, as discussed above.

In secondary benefit evaluation, the compensation
principle can be applied at three different levels. The
most widely discussed of these has been the national level.
It has often been asked whether the nation should be will-
ing to pay for the secondary benefits which have been
claimed in some project analyses. This is, in effect, the
same question as that of the "netness" of secondary bene-
fits. A much less widely discussed level is that or re-
gional or local ability to compensate. The question here

is whether the secondary benefits are net locally. As

48
discussed above, the changes might merely represent reallo—
cations within the area so that some local people gain
and some lose. Whether or not local governmental units (as
representatives of the people) could pay for the changes is
a basic question where measurement of regional development
benefits are considered. The third level is that of in-
dividuals. There is a very complex problem involved at
this level which also permeates the other two via the in-
dividuals involved there. That problem revolves around
the real welfare effects of labor employment. Labor is
dealt with both as the relevant welfare beneficiary and as
a factor input. Supposedly, leisure is a consumption good
just as the products of labor are. A rational laborer-
consumer would be expected to allocate his time between
leisure and labor until marginal values of the two were
equal. One can logically assume that he would do so.
However, it is difficult, if not impossible, to determine
the real change in welfare of a previously unemployed in-
dividual who is employed as a result of a public project.
He, in effect, exchanges a large volume of "leisure goods"
for other kinds of goods. Ostensibly, the leisure goods
had a zero value at the margin; thus, his welfare is en-
hanced by the exchange. To determine how much better Off
he would be, however, one would have to know the total dif-
ference in value. In application, this problem is generally
ignored, and the benefits (welfare gains) are taken to be

the whole of the wages paid. Unless the individual valued

49

leisure at a zero rate at all levels of consumption, the
real welfare gains are obviously over-valued. If asked to
compensate by the amount of the entire wage (or slightly
less) for the privilege of working, few workers would

choose that alternative over leisure. 'The problem is not

so complex when other individual factors are dealt with.

The "consumption good" nature of leisure for idle plant
capacity offers little problem. Most analysts have few
qualms about valuing plant idleness at zero value. (Natural
resources constitute another problem, however -- the prob-
lem of conservation or "Option demand" value. This problem
will not be dealt with herein.) In the case of labor under-
employment, the leisure good problem is not so acute. If
the individual were already working "full time" but at a
lower wage, the wage improvement would not come at the ex-
pense of leisure good consumption. Thus, the welfare ef-
fects would be to increase the total volume of non-leisure
goods while allowing continued consumption of the same
volume of leisure (or more leisure and the same volume of
non-leisure goods).

The invocation of the Compensation Ability Principle
constitutes an attempt to ensure that changes in the real
welfare of people actually occur as a result of a project
in order for those changes to be included in the project
analysis (if the given group is assertedto be all of the
peOple of the region, then the internal transfers should be

nOted).

50

The important consideration is that people as laborers and
as resource owners will be made better off by changes in
resource productiVity. Thus, in many respects, the analysis
of secondary benefits from the employment of non-human
resources conforms very closely to that Of human resources.
The major difference is that the non-human resources do not
constitute the relevant consuming units. Thus, among other
things, "leisure" is valued at zero for those units. (To
the owner of an underutilized plant who must make amortized
payments on the fixed assets, the value of such leisure is
probably viewed as negative.)

In terms of real secondary benefits realized as a
result of the utilization of unemployed plant capacity,
the only changes for which one can find an ability to com-

pensate are the changes in quasi-rents which result.6
A Working Definition of Secondary Benefits

Keynesian multiplier analysis demonstrates the real
income changes which occur in an economy when an infusion
of aggregate demand occurs. The economy to which it is
most applicable is that of the United States in the 1930's
in which economic stagnation had followed many decades of
relative prosperity in which the basic investments had
been made. The aggregate demand multiplier is an accept-
able, and even fairly accurate, device for measuring ag-
gregate demand-induced changes in income where the idle

fixed capacity exists to employ idle workers; given a

51
sufficient level of demand under these conditions, the real
opportunity costs of both of these factors are probably
close to zero, while (because Of price rigidities) their
market prices might be much higher.

The perfect competition model deals-with factors
which in equilibrium are allocated to their highest use.
Under perfect competition, the Opportunity costs of these
factors are equal to their selling prices; factor owners,
possessing perfect knowledge and foresight, undertake
capital investments in the highest yielding industries,
tending to force rates of return to equality; in equilib-
rium, rates of return on all investments equal the Oppor-
tunity cost of capital and no pure profits exist.7

The perfect competition model in which both capital
and labor are fully employed and are earning their Oppor-
tunity costs is the antithesis of the Keynesian model in
which labor and capital are unemployed and the factor
prices do not equal their opportunity costs. Under the
perfect competition model, the price of a good sold to
final demand is exactly equal to the sum of the opportunity
costs of all of the factors used to produce that good;
thus, no "net" value (rent) is realized. In the Keynesian
Model, all factor prices are over-valued relative to their
Opportunity costs; thus, the "net" value produced by a
good sold to final demand exceeds the direct value added
above factor cost and should include the rents realized by

the owners of the factors indirectly used in meeting that

52

final demand. That is, the "real" costs of a good sold to
final demand is composed of the "real" cost of the factors
used to produce that good; the "real" cost of the directly
used factors is composed of the ”real" cost of the factors
used to produce them; and the same is true of all other
"rounds" of production preceding these rounds. Thus, the
cost of the final good is not equal to its market price to
the extent that rents are realized in the production of all
of the factors and sub-factors consumed in its production.

Some of the techniques used in traditional multiplier
analysis allow the tracing of the factor rents realized
from the most basic primary factor up to the good sold to
final demand. Others include these rents but also super-
impose the "induced" effects (see Chapter III) of the re-
spending of these rents. Some techniques look at only the
re-spending of the rents. In all cases, the rents are de-
fined by an assumption of a zero opportunity cost for the
factors used.

Secondary benefits (in the form of multiplier im-
pacts) potentially exist where there are market imperfec-
tions which allow rents to be realized in the production of
final goods or of factors. The traditional approaches to
estimating these rents have used the Keynesian conventions
of zero - costing the resources and/or adding an induced
income component to them. The former is simply a special
case solution, while the latter adds a dynamic component

which damages the validity of what is basically a static

 

'1

53
analysis. By leaving the induced component out and by ap-
propriately shadow pricing the resources used in all rounds
of production leading to final goods, it is possible to
use some of the traditional multiplier techniques to esti-
mate the degree to which market prices over-estimate the
real cost of the goods produced; or, alternatively, it is
possible to estimate the rents which are realized at all
stages of production leading up to the provision of the
final goods. The generation of such rents, it seems to
this author, should form the basis for a workable concept
of indirect income changes.

In Chapter IV, past approaches to secondary benefit
estimation and multiplier analysis are surveyed and the con-
formity to above definition of secondary benefits discussed.
The principal objective is the selection of a format which
best allows the estimation of the above-defined secondary
benefits. In Chapter V, an attempt is made to alter the
selected format so that it will at least provide a con-
ceptual basis for the estimation of the economic rents
which are realized from the higher employment of previously

under-employed factors.

54

Footnotes

 

1. Eckstein, Water Resource Develppment, 216.

 

2. Joan Robinson, The Economics of Imperfect Compe-
tition (London: 1965), 107-7}

 

3. David Ricardo, The Principles of Political Economy,
and Taxation (New York: 1965), 33—45.

 

 

4. Alfred Marshall, Principles of Economics (London:
1964), 349-50.

 

5. In the long-run all costs are variable; thus the
only returns above variable costs would be profits. How-
ever, at long-run equilibrium, pure profits should be zero.
See H. H. Liebhafsky, The Nature of Price Theory (Homewood
1963(, 353-61.

6. M. M. Huffschmidt, et al, "Standards and Criteria
for Formulating and Evaluating Federal Water Resource
Developments: Report of Panel of Consultants to the
Bureau of the Budget" (June 30, 1961), 26.

7. Liebhafsky, op cit, 365-87.

CHAPTER IV
SOME APPROACHES TO SECONDARY BENEFIT EVALUATION: A REVIEW
Secondary Benefits and Economic Impact Analysis

In the past two decades there have been multifarious
attempts at secondary benefit evaluation. Many (or perhaps
most) of the attempts have been applied to other than
natural resource investments. Some of these attempts were
methodology oriented in that they sought to develop, test,
or refine methods for dealing with the problems of second-
ary benefit evaluation. Others represented bonafide em-
pirical attempts at determining the actual secondary bene-
fits from certain economic impacts.

"Economic impact analysis" is the term used in the
economics and Regional Science literature to describe many
of the studies which have attempted to measure secondary
benefits. Impact analysis represents an attempt to deter-
mine the macro-economic effects upon a regional or national
economy of a given exogenous change in a variable affecting
that economy. Several methods have been prOposed for
evaluating the economic repercussions of such changes. The
more important of these will be presented and briefly dis-

cussed in this section.

55

o.

‘9

it

(D

3i!

9

56

Impect Analysis and ”Customary" Secondary Benefits

Contemporary students of secondary benefits generally
distinguish between three varieties of such benefits:
First, the "customary" variety of "stemming from" and "in-
duced by;" second, those due to scale economies; and,
third, "dynamic secondary benefits” which actually change
the form or productivity of the resources involved.1 While
it is probably true that-in terms of the economic develop-
ment aspects of public investment-the latter two perhaps
hold the greatest potential for large-scale impacts upon
the economy, they are by nature so difficult of evaluation
that few attempts have been made to deal with them empiri-
cally. The attempts which have been made have largely
been of the theoretical type, as opposed to the methodologi-
cal or empirical; and as yet little implementable theory
appears to have been developed. While the concepts of "x-
efficiency"2 and "o-efficiency"3 are perhaps valuable con-
tributions to the growing body of theory of "non-marginal
economics," one is hard-pressed at present to devise a
means by which they might be Operationalized. Furthermore,
there exists a great and growing body of such literature
relating to the growth and development of underdeveloped
countries which have realized only very limited operation-
alization by economists working in that area.4 The long-
standing discourse concerning "infant industries" and
"leading sectors" represents an overt attempt to deal with

the problem of achieving scale economies. The discussion

57
of investments in "human capital" constitutes a recognition
of the existence of ”dynamic secondary benefits." However,
no generally acceptable methods for evaluating the impact
of such policies have been developed.

The greatest difficulty posed by the third variety of
secondary benefits is that they constitute not only move-
ments up the aggregate production function but also changes
in that function.5 Thus, traditional production theory
and analysis encounter great difficulty in coping with such
changes. The second variety is perhaps not so difficult to
handle conceptually; however, the prediction of scale econo-
mies is quite difficult when the relevant unit of production
has never been Operated at the scale necessary for their
achievement. To date, the development of scale economies
and other technological externalities has largely been in-
cidental to project development. Analysts still do not
have adequate information regarding this variety of second-
ary benefits to predict with any degree of accuracy their
occurrence. As a result of these and other difficulties,
secondary benefit analysts have spent most of their time
in attempting to measure the first variety of secondary
benefits. This emphasis has been a natural off-shoot of
and has followed closely the "Keynesian revolution" in
economic analysis. For the most part, it has constituted
an attempt to bring macro-economic theory to bear upon the

analysis of economic growth and development.

58

The application of Keynesian economics to the prob-
lems of growth and development which lie at the heart of
the secondary benefit issue has resulted from the failure
of economists to provide anything more closely resembling
a theory of economic growth. Thus, attempts to deal with
secondary benefits have tended for the most part to focus
upon aggregate demand and "the multiplier” and have tended
to ignore the problems of structural and institutional
change involved in the development process.

Very early after its original statement, students of
the General Theory deemphasized its aggregate supply
aspects and began to work in earnest upon the aggregate
demand aspects.6 Comparatively little work has been done
upon the aggregate production function and its determin-
ants.7 As a result, the most familiar and most widely
applied component of the General Theory has been the ag-
gregate demand multiplier. Economists became intrigued
quite early with the multiplier; and during the ensuing
decades the economics literature has been replete with
theoretical and empirical treatments regarding it. While
there have been some discussions regarding "shifting the
aggregate production function," these discussions have
gained nothing of the impetus granted to those concerning
aggregate demand. Aggregate demand maintenance as a
Federal function was given explicit recognition in the
Employment Act of 1946; there is no such explicitly-stated

policy at the Federal level regarding aggregate supply

S9
maintenance, though programs such as that in the water re-
sources field constitute implicit policies Of resource im-
provement.

Given the foregoing as a background from which to
develop, it is understandable that secondary benefit
evaluation would follow the lines it has followed. For
the most part, "secondary benefits" has been synonymous,
with "secondary impacts"; and "secondary impacts" has
generally referred to aggregate demand multiplier effects.
Thus, the approach most often taken in secondary benefit
evaluation has utilized some form of multiplier analysis
in which changes in regional or national demand via the
multiplier effects of the project were taken as the second-
ary benefits. The major exception has been the approach
used by the Bureau of Reclamation in which the aggregate
supply effects ("stemming-from" benefits) were added to the
aggregate demand effects ("induced-by" benefits) in estima-
ting secondary benefits. The problems posed by this ap-
proach in estimating the net secondary benefits from the
national viewpoint have been previously discussed (see
Chapter 11, above). While from a regional viewpoint one
might wish to consider "stemming-from" effects in addition
to "induced-by" effects because of the possibility of net
regional gains via inter-regional pecuniary effects, most
secondary impact analyses have dealt with only the aggregate
demand (induced-by) effects. Thus, the discussion which

follows will be couched basically in the aggregate demand

 

‘—

mm

(D

1

>1

60
multiplier framework. Most of the approaches discussed
below could be used in the estimation of "stemming-from"

effects as well.

 

Multiplier Analyeis

In very general terms, a "multiplier" constitutes an
expression of the relationship between changes in one
variable in association with a given change in another
variable. Thus, one might view a multiplier as simply a
"coefficient" in a functional equation which relates
changes in the dependent variable to changes in the inde-
pendent variable.

As used in economics, "multipliers” generally refer to
the relationship between some welfare-related dependent
variable and an independent "policy" variable which is sub-
ject to manipulation from without. The most common of
the welfare-related variables have been those of employment
and income. Their determinants have ranged from "govern-
ment expenditures" to the "export base" of the relevant
economic unit. There are many such relationships in econo-
mics, with several means for constructing multipliers to
represent most of them. Thus, one should be wary of such
terms as "The Multiplier," which is often encountered in
Federal water resource project reports and, at times, in
other economics literature.

A great part of the attraction which multiplier analy-

sis seems to hold stems, no doubt, from the simplicity and

 

ci-

in

p&\ a

 
 

61
apparent logic of the process by which multipliers are gen-
erally derived. For most multipliers, a knowledge of simple
algebra is sufficient for one to understand the mathemati-
cal relationships involved. By the same token, the same
simple algebra is generally sufficient for the student to
derive the expressed multiplier himself or derive another
on his own. For example, the original Keynesian investment-
income multiplier is sufficiently simple in concept that
normal procedure in economics principles courses is to
have students derive it for themselves. At the same time,
the algebra involved lends to the derivation process an
aura of mathematical logic which tends to create an il-
lusion of a great revelation,8 when, in fact, the "multi-
plier" might be nothing more than another way of stating a
tautology.9 Keynes, though himself accused Of the same
fault by Haberler and others, warns against the obfuscation
of important relationships by "a maze of pretentious and

10 One should interpret apparent mathe-

unhelpful symbols."
matical relationships with some caution lest he interpret

a snapshot specification of an existent situation as a
fundamental long run identity or functional relationship.
The warning concerning the regression fallacy is not new

to economists; indeed, it is included in practically all
introductory texts. Thus, whether one should interpret a
”relationship” between farm income and non-farm income, for

example, as a fundamental multiplier identity is a question

which cannot be answered by simply stating that "total

62

11 The careful

income = farm income + non-farm income."
process by which causation is established is not present in

this statement alone.

 

Keynesian Multipliers

Though the original presentation of the multiplier

12 it was John

concept is generally credited to R. F. Kahn,
Maynard Keynes who brought the concept to fruition in his
General Theory of Employment, Interest and Money. In
Chapter Ten Keynes presented his now famous investment-
income multiplier which expressed a relationship between
equilibrium income and investment. In a closed system in
which investment constituted the only exogenously determined
variable, changes in equilibrium income were given by the
product of investment change and the multiplier, defined

as the reciprocal of one minus the marginal propensity to
consume, or TTMFC' Though Keynes discussed altering effects
upon the multiplier by such factors as foreign trade pay-
ments (later called "leakages") and durable-nondurable

goods composition of purchases, the basic Keynesian multi-
plier remains as originally stated.

Though the original, simplified derivation of the
Keynesian multiplier stressed the effects of exogenous
changes in investment upon the economy, the concept may be
applied to any form of exogenous infusion of expenditures.
Thus, it is particularly useful in looking at the effects

upon regional income of a given level of expenditure in

63
public works originating outside of the region, such as ex-
penditures by the Federal government for construction of a
water resource project.

To estimate the secondary changes in income using the
Keynesian multiplier, one would first have to establish
the marginal prOpensity to consume. If one were largely
interested in regional income changes (as, for example, one
might be in applying the considerations dictated by the
Appalachian Development Act), one would need to know the
regional MPC. Having determined that, the process of
estimating the secondary income changes would be a fairly
simple one. It should be emphasized, however, that the
determination of the MPC is seldom a simple task and would
require a large scale research effort in and of itself.

If the MPC of the region in which the expenditure
occurred were different from that of the nation as a whole,
the secondary income changes would not represent simply
shifts in income between regions; there might also occur
net increases or declines in national income as a result

13 has demonstrated that if

of those shifts. J. S. Chipman
the MPC for all regions were equal, the inter-regional
multiplier would reduce to the national multiplier. Then
there would be no problem of trading off relative changes
in regional income with absolute changes in national in-
come. If the MPC of the project region were greater than

that of the nation as a whole, national income would be in-

creased by virtue of the shift in expenditures. Richardson

 

_-.r

1""

.fiui

9.9
.

‘5'.

9n)
'9

64
notes that, since most underdeveloped regions tend to have
a very high MPC, national income would be maximized if the
increase in government expenditures were concentrated in
those regions.14 While first round expenditures would
perhaps be greater in such regions, "leakages" would also
tend to be greater. Thus, while national income might be
greater, the region in question would probably realize
little of the gains in income beyond the first round of
expenditures.

An obvious difficulty with the simple Keynesian
multiplier is that it does not consider such factors as
imports and other leakages which exist in open systems.
Thus, there are serious limitations to its use in regional
analysis. When other factors such as the industrial com-
position of the region, regional input-output relation-
ships, and local purchasing propensities are considered,
the marginal propensity to consume becomes only one of many
factors operating upon the level of regional income. In-
deed, if the MPC is related to the state of develOpment
while the volume of leakage is related in an inverse
fashion, as would appear to be the case, the simple Key-
nesian multiplier could be quite misleading in terms of re-
gional income effects of exogenous investment.

The degree of aggregation involved in the determina-
tion of a single marginal propensity to consume creates
the kind of problems involved in all such aggregations.

First, there is the problem of different consumption

V -

.u‘

.3

1 ’ v

 

65
patterns between different groups of income recipients. As
a result, it makes some difference as to what groups the
initial income changes accrue. If there is a systematic
bias towards skilled craftsmen, for example, whose incomes
tend to be higher than average, then the relevant MPC might
be quite different from the average propensity to consume.
At least the first round of expenditures should take this
fact into consideration. Secondly, it matters what the
composition of purchases is for both the exogenous infusion
and for the ensuing household purchases. Keynes realized
the difficulties involved with durable goods purchases, for
example, which are normally taken "on order." The lags in-
volved at both the consumption and production stages for
durable goods create difficulties. In the case of Federal
water resource projects a large volume of the expenditures
involved are directed towards durable goods industries.15
The problems created by the differing "velocities of
circulation" (or "duration of the rounds") are substantial
in determining both the magnitude and duration of the sig-
nificant multiplier effects.

As in interpreting all multipliers, one should ask
the question pf exactly what is being multiplied. In this
case, the "welfare-related" variable that one is attempting
to change is income. Whether or not those income changes
represent commensurate changes in real welfare, however,
is not answered by the analysis. This must be answered

from without and in consideration of other factors. This

66
is true because of the manner in which the multiplier oper-
ates. While one often hears the Keynesian multiplier de-
scribed as an "income" multiplier, it multiplies income
directly under only very Special cOnditions. The Keynesian
' multiplier (and most other "income” multipliers) is an ag-
gregate demand multiplier. Thus, for it to be effective,
certain conditions relating to aggregate supply must be met.
In short, aggregate supply must either be perfectly elastic
or it must shift (as a result of the investment, ostensibly)
by the same amount as does aggregate demand for the multi-
plier effectsato'be felt full force without being partially
absorbed through inflation. For the full value of the in-
come payments to represent real changes in welfare to the
recipients, the opportunity costs of those‘factors must
have been zero. Thus, the real welfare effects (secondary
benefits) of public investments are consistently overstated
by the Keynesian multiplier in all except this very special
case.

While practically all secondary impact study reports
begin with a discussion of the Keynesian multiplier concept,
few of them utilize the pure Keynesian version in their
analyses.. The common disclaimer by which the analyst
proceeds to another approach concerns the difficulty of
obtaining data (especially at the local or regional level)
on the marginal prOpensity to consume and the imprecision
of such data, when available. As a result, most impact

studies which purport to utilize the Keynesian multiplier

 

D'.

67
in fact only utilize the Keynesian method of derivation.
That method has been discussed above and some of the pit-
falls delineated. The following example should demon-
strate some of these pitfalls.

In their article on farm-non-farm income linkages,
Mirakhor and Orazem discuss the development of the multi-
plier concept. They quote a definition of the multiplier
by Oskar Lange as ". . . the marginal effect of a change of
one economic variable upon another economic variable, of
which the first is a component." From this definition they
conclude that ". . . if it is possible to divide an eco-
nomic variable into its different components, it is per-
missable to determine the changes that take place in the
original variable as a result of a change in one of its

components."16

The authors then proceed to set up an in-
come identity similar to the Keynesian income identity.
This, however, shows total income as composed Of farm in-
come and non-farm income. From this identity, "multipliers"
are derived which show the relationship between "changes in
farm income" and "changes in total income." In fact, the
multipliers" are nothing more than the tautological state-
ment which Haberler warned against in his discussion of
the Keynesian derivation.

Though Mirakhor and Orazen use correlation analysis
to demonstrate the implied relationship between farm in-

come and non-farm income via their "multiplier," no where

do they show the logical process by which farm income

68
changes "cause" non-farm income changes. As discussed
earlier in this chapter, correlation analysis does not
establish causation between variables: it merely demon-
strates the degree of similarity in the manner in which
they change. Thus, the presentation of statistical analy-
sis with their "farm-non—farm income multiplier" is not
sufficient to rebut the Haberler criticism but rather
further points up the danger of the regression fallacy.

An additional point concerning the pure Keynesian
multiplier bears mentioning in closing this section. That
point regards the part played by the marginal propensity
to consume in determining the equilibrium level of income
in the Keynesian system. If one forgets about the aggregate
supply side of the Keynesian system (as too often multiplier
analysts do), an apparent disagreement among growth theo-
rists and Keynesianites develOpes. This apparent dis-
agreement should serve to illustrate the problems posed
by the uncritical acceptance of multiplier analysis in im-
pact evaluation. The Keynesian multiplier holds that an
inverse relationship exists between the marginal propensity
to save and income growth in the system. That is, the less
saving there is in the economy, the greater will be the
level of equilibrium income. However, most of the growth
models developed by international development economists
made savings (rather than consumption) the prime mover of
growth in the underdeveloped world. Thus, the policy pre-

scription in the growth models is to somehow reduce the

69
marginal propensity to consume and to raise savings out of
current income. The similarity in this dichotemy to that
between the Classicists and the Marginalists over the
determinant of market prices is striking. To one group
the supply problems are the perplexing ones, while to the
other the demand issue is the important one. To all of
those working upon the problems of income determination,
the macro-economic equivalent of Marshall's two-edged
shears analogy would appear instructive.

"Induced" and "Induced-byV: A Note on Terminolpgy.
Often in reading the technical literature on multiplier
analysis and on project appraisal, one encounters the terms
”induced" multipliers and impacts "induced-by" purchases of
inputs. One should note very carefully the context within
which the terms are being used, for they refer to two

vastly different concepts. "Induced-by" impacts are those

 

which result from the purchase of inputs by producers who
realize an increase in output due to a public project.

The "induced-by” analysis traces the increased purchases of
fertilizer and other inputs caused by the completion of an

irrigation project, for example. "Induced" multiplier ef-

 

fects, on the otherhand, are those purchases which occur
through re-spending of household incomes. The Keynesian
multiplier is an induced multiplier. It includes only

those spending effects which occur through the household
sector. Some multipliers include both an ”induced” com-

ponent and an "induced-by" component. An example is the

70
Type II input-output multiplier discussed in a following
section of this chapter. Henceforth in this study, when
reference is made to the multiplier impacts of household
re-spending, these effects will be referred to as Keynesian
induced effects. The effects of input purchases will be

referred to by the hyphenated term induced-by effects.
The Economic Base Multiplier

The economic base multiplier, or "export" base multi-
plier as it is sometimes called, is the most simple and
straightforward of all the regional multipliers.17 The
simple economic base multiplier consists of nothing more
than the ratio of total employment or income in the region
to "basic" employment or income, where "basic" refers to
those activities which serve export markets. This state-
ment alone is sufficient to point up the implicit assump-
tion in economic base analysis that exports constitute
the prime mover of the local economy. Thus, changes in the
”export base" are seen to have certain stimulating effects
upon employment and income in secondary and tertiary activi-
ties; if the basic/non-basic ratio is stable over time
and over stages of development, then the effects of changes
in the export base upon the rest of the regional economy
can be predicted using the economic base multiplier.18

The economic base multiplier has some real advantages
over the Keynesian multiplier in regional analysis. Where-

as the simple Keynesian multiplier makes no distinction

71

between the propensity to consume locally-produced goods
and services and the propensity to import out Of local in-
come, the economic base multiplier does. This relationship
is implicit in the basic-non-basic distinction. Tiebout
demonstrates this by showing that the propensity to consume
locally, multiplied by the income created per dollar of
local sales, yields a product equal to the ratio of non-
basic/total income.19

The simple economic base multiplier discussed above
has been described as a short run multiplier because of
the absence of consideration of investment expenditures.
The long run multiplier can be constructed by simply add-
ing a local investment component to the short run multiplier.
Thus, the long run multiplier becomes the reciprocal of one
minus the sum of the products of the propensity to consume
locally times the income created per dollar of local con-
sumption sales, and the propensity to invest locally times
the income created per dollar of local investment sales, or

1

or OY
1-(PCL 6‘s + PIL OT)

 

As a result of the inclusion of local investment, the
long run multiplier will always be greater than the short
run multiplier. The extent to which it is greater will
depend, of course, upon both the propensity to invest
locally and the income created per dollar of local invest-

ment .

72

Tiebout defines the "long run" as greater than two
years. While some might challenge the arbitrariness of
this distinction, for purposes of evaluating the secondary
impacts of long-range investments such as those of the
Federal water resource program, the Tiebout definition of
the long run creates no difficulties. Such investments so
greatly exceed the time horizons discussed by Tiebout that
one has few qualms about accepting the long run multiplier
as the relevant (me for water resource project evaluation.
The problem, however, is that the data requirements in-
crease greatly when the investment component is added.
Thus, most use only the simple multiplier.

The economic base approach is basically a comparative
statics approach, though (according to Isard)20 Tiebout
maintains that it is a dynamic technique. Thus, even though
one might do a base study of the same community at two
points in time, he would simply have two snapshots of the
local economy rather than one. Therefore, it would appear,
the greatest uses of the base study are for description.

The prediction component of the base study lies
largely in the economic base multiplier. The normal pro-
cedure in predicting future changes in the local economy
is to determine first the ratio of basic to non-basic em-
ployment and/or income. Secondly, the future growth in
basic or export employment and/or income is somehow estab-
lished. Then utilizing the multiplier developed from the

current basic-non-basic economic components, the future

73
changes in non-basic employment and/or income are predicted
from the expected changes in the basic. It is a fairly
simple and straightforward procedure.

In using the economic base approach for predictive
purposes, as one would be doing if he used the multiplier
in an attempt to estimate secondary benefits, the analyst
should understand the weaknesses of the technique. Among
these are the implicit assumptions that "export" (basic)
activities are the prime mover of the local economy and
that the relationship between the basic and non-basic

sectors of the economy are stable over time. Isard,21

Andrews22

and others have pointed out that substantial
evidence exists indicating the relationship is not stable;
thus, the predictive powers of the technique are cast in
doubt. Barkley and Allison classify economic base analysis
as "a somewhat undisciplined body of economic theory" for
its failure to satisfactorily interpret the meaning of

the basic/non-basic ratio and establish logically the

23 It has been widely argued that ex-

causation involved.
ports do not constitute the sole "basic" activity of an
economy and that, as a result, the "export base" multiplier

24 Whether

does not include all of the relevant components.
this explains part of the observed instability in the multi-
plier over time has not been answered.

Another major consideration regarding the use Of the

economic base multiplier is the degree of aggregation in-

volved. The economic base multiplier, like the Keynesian

74

multiplier, is an "average" multiplier. That is, it aver-
ages together the separate multiplier effects of all "basic"
industries into a single multiplier, while the multipliers
relevant to the individual industries might vary widely.
As a result, the multiplier when applied to any change in
the basic sector other than one having an identical in-
dustrial mix to the current one, will be inappropriate.25
In addition, the less diversified the mix of the change,
the less appropriate the multiplier would be, in all
probability. Thus, the changes in basic employment, such
as that stimulated by a water resource project which at-
tracted one firm or group of similar firms, would be amena-
ble to analysis using the economic base multiplier only
under very unusual circumstances.

Barkley and Allison interpret from a passage in
Senate Document 97 a charge to the water resource agencies
to do economic base analyses as a part of the project plan-
ning and evaluation process. They indicate a similar inter-
pretation on the part of N. A. Back of the Corps of Engi-
neers. Thus, they argue, there is a provision in the water
resource development process for economic base studies.

Economic base analysis has been widely used as a
technique for projecting future "demands" or "requirements"
for economic goods. The initial presentation by Homer Hoyt
cast base analysis in just such a role in estimating future

26

demand for housing. It has been picked up and used

fairly extensively by the water resource agencies in

75
projecting future "requirements" for water or for water re-

27 Normally, the procedure involved is to

lated goods.
project the growth in non-basic activity, after making some
assumptions about changes in basic activity, assuming that
water or the water-related goods will be available.28

Then, having established the ”potential" growth with ade-
quate supplies of water, the actual water requirements for
such a level of output are estimated, over and above

those which would actually be available in the absence of
public provision of additional supplies. The excess then
becomes the "requirements" to sustain the "anticipated"
growth. Often the analyst credits the difference in re-
gional income to the provision of that water. In effect,
such a procedure constitutes a one-factor theory of eco-
nomic growth.

The procedure described above is quite similar to
"approach No. l" to secondary benefit evaluation suggested
in the June 1969 Report of the Special Task Force of the
Water Resource Council. The suggested approach would have
the analyst "Establish an achievable level of National in-
come secondary benefits . . ." Then, a "general develop-
ment plan" including the "required" public and private in-
vestments should be established. Lastly, the contributions
of the "specific water resource projects" should be deter-
mined and allocated to each project according to relative

29

investment costs. The problem comes, of course, in

determining the contribution (i.e., the marginal product)

76
of the water service. However, any approach which did not
credit all of the increase in income to the provision of
the water service would constitute an improvement over the
"one-factor theory."

The Bureau of Reclamation has used economic base
analysis in a restrospective attempt to determine the
secondary income producing effects of irrigation projects.30
These attempts, however, were undertaken more in the spirit
of the comparative projection approach (discussed below)
than the traditional economic base multiplier approach in
that they were directed largely at determining the second-
ary effects of previous projects as a guide in predicting
the effects of future projects. In this case, not only the
inter-temporal stability of the basic/non-basic ratio is
subject to question but also the inter-area stability.31
Thus, unless there is some universal, natural and immutable
basic/non-basic ratio, such an approach has all of the pit-
falls of the comparative projection approach.

In the study by Christensen and Matson,33 which uses
the economic base concept to modify a statewide multiplier
for use in local impact analysis, an attempt was made to
evaluate the secondary effects of irrigation develOpment.
The induced effects of an increase in irrigated agriculture
were estimated based on the assumption that agriculture was
a basic industry to the area. Changes in income in the

"non-basic" sectors were estimated using the modified state-

wide multiplier. However, the estimates of the income

fin

.
yiA\
A

3K 7. <

‘A
. If;
...\

77
changes were accepted as the relevant secondary benefits
without any consideration being given to the supply con-
ditions or the opportunity costs of the basic income re-
ceiving units in the analysis. The assumption was im-

plicitly made that intensive income growth would occur to

 

owners of retail and service establishments due to the
fuller utilization of underemployed facilities, while at

the same time extensive growth would occur through the

 

development of other similar facilities. At no point was
it established (even by explicit assumption) that such
facilities were currently under-utilized. Neither was it
established as to whether extensive employment growth would
come from employment of previously unemployed local people
or from immigration. (Neither was it established as to
whether there were unemployed local people). Each of

these are necessary for one to determine whether the esti-
mated impacts are the real monetary changes in the welfare
of the people involved.

In summary, though economic base studies have been
used widely in the water resource develOpment field, there
.is some question as to their applicability to secondary
benefit evaluation. A most basic question is that regard-
ing the stability of the multiplier, not to mention the
3kick of a theoretical basis for its importance. Though
it. deals with the problem of local versus non-local con-
SUantion which limits the Keynesian multiplier, it has all

Of the problems of the Keynesian multiplier in terms of

78
aggregation. In addition, as in the case of the Keynesian
multiplier, the economic base approach does not allow the
analyst to determine the real net changes in welfare of
the people involved. The units which are typically used in
the economic base multiplier are employment, income, and/or
Value added, or some similar measure. The crude estimates
of tlie changes does not give much insight into the basic
quest:ion of "How much better off are the people involved?"
(i.e.,, the people already in the region). If the changes
in t}1ese measures occur through extensive growth (i.e., a
changge in the number of jobs without a change in the kinds
0f jcabs or in average wages) rather than intensive growth

in tlle region, it is quite possible that no one individual

woulcl be any better off, though both regional income and
emplrbyment could have grown substantially.32 In the case
0funder-employment, one is specifically interested in
inteulsive growth (i.e., changes in the composition of

emplxoyment and the level of average wages).

IES_J;pmparative Projection Approach

The comparative projection approach has been much
m01¥3 widely practiced than the discussions of it, per se,
inthe economics literature would indicate. The comparative
prc’liection concept is a fairly simple one, though the
téCihnique itself sometimes becomes a bit involved, and
(then even confused. As its title indicates, the approach

utllizes comparative analyses of similar areas -- one

 

 

(U

rfl

79
(group) having a public project, the other not -- in an
attempt to estimate the differences in income which re-
bound to the advantage of the project region. The approach
represents an attempt to apply the "controlled experiment”
approach Of scientific investigation, the "without-project"
group in this case representing the ”control group." The
differences which exist in the parameters deemed to be im-
portant are noted. If a systematic pattern exists concern-
ing the differences, and if that pattern can not be logic-
ally explained by some other "uncontrolled variable," the
cause of the difference is assigned to the existence of
the public projects. Usually the analyst computes a
"multiplier" which consists of the ratio of the project
area/non-project area income, employment or whatever
parameter or index of parameters one desired to use.

Though not the first use of such an approach, the
presentation of the formalized version of the comparative
projection model can be attributed to John E. Pearson.
Pearson attempted to develop and present "a theoretical
model which would predict with reasonable accuracy the
economic impact of water impOundment on the surrounding

34 A simulation approach was used in which an at-

areas."
tempt was made to simulate the economic growth process of
two water impoundment areas, using a third area which did
not have a water impoundment as a check on the model.

Economic parameters felt to be of economic significance,

and for which data were commonly obtainable, were selected

80

for inclusion in the model. These included area population,
area income, area agricultural income, and distance to a
metropolitan area. Exogenous factors expected to impact
on the local economy were project construction expenditures,
project operation and maintenance expenditures, project-
related associated investments, and visitor recreation ex-
penditures. A model designed to simulate the economic
growth generated by these exogenous factors and composed
of the above economic parameters was described.

Pearson says that ”The model en toto provides a trans-
formation of valid macro-economic theory into a workable

"35 However, he does

simulator of a micro-economic system.
not demonstrate this fact to the reader. Indeed, there is
no theoretical foundation laid in either macro-economic
or micro-economic terms, though there is implicit in his
equations some basic "theory" about how income growth occurs
(for instance, the exponential growth equations of pp. 34-
42).

A similar criticism can be made for other comparative
projection techniques. While one must admit that no
readily acceptable theory of economic growth exists at
present, this admission should not negate the importance
of a theoretical basis for growth projection. Admittedly,
one can learn a great deal about the process of growth by
the case study approach. However, most comparative pro-

jection approachs do not "study" the process, but rather

perform a multiplier or rate of growth "transplant" upon

81
another case. At the least, such a process should be ac-
companied by a detailed explanation of the similarities
between the areas and the reasons the analyst would expect
the area in question to grow in a similar manner to the
studied area.

A similar study to that performed by Pearson was
undertaken by the Business Research Center of the Oklahoma
City University. However, rather than attempting to simu-
late the growth process, the Oklahoma study simply used
linear regression techniques to estimate growth rates in
income, etc., from time series data. The rate of growth
which had been realized by Oklahoma areas having impound-
ments was estimated from time series data and used as the
"with-project" case in analyzing the economic impact of
the Hugo Reservoir. The pre-project rate of growth of
the Hugo Reservoir area was used as the "without-project"
case. The difference was purportedly attributable to the
presence of reservoirs.36

The Bureau of Reclamation has undertaken at least
three such studies of their project areas in an attempt to
estimate the secondary impact of irrigation projects. The
studies by Marts, and Struthers were retrospective analyses
of irrigation development projects, much in the spirit of
the above-discussed studies. The objective in these studies
was also to determine the relationship between water pro-

jects and economic growth.

r9
1' .)

ShOL
ties
anal
is a
cuss

the

82

A common failure of all of the comparative projection
techniques which have been discussed has been the failure
to develop a prOper thread of logic linking the availability
of water or a water service to the process of economic
growth. Perhaps the milieu in which the reports were
written or the audience to which they were directed can
explain this failure. It is understandable that the avail-
ability of water in a water deficit area such as the western
United States might be accepted axiomatically as a (if not
ppe) determinant of economic growth. Nevertheless, a great
deal of credibility is lost on the uninitiated as result
of the lack of an established link. Lacking a water-
related growth theory, the advocate of such an approach
should as a second-best alternative describe the similari-
ties between the areas which might justify the comparative
analysis. In the case of "multiplier transplants," which
is a form of comparative analysis, the analyst should dis-
cuss in detail the factors which influence the level of
the particular multiplier and should give his reasons for
believing these factors -- and, thus, the level of the
particular multiplier -- to be similar between the areas.
The Corps of Engineers Test Team Study of the Mountain Home
Project provides a case in point37 (see Chapter VI). Con-
sidering the very great time limitation under which the
Team worked, the ”multiplier transplant was perhaps the
only means by which regional development effects might be

approximated. Nevertheless, under normal project

“.1

Ala.

83
appraisal circumstances, a great deal more evidence as to
the applicability Of that multiplier should be required of
the analyst.

One should be quite wary of the comparative projection
approach as a means of estimating secondary project impacts.
The comparative growth rates are not developed from "con-
trolled" experiments, though the analyst might strive very
diligently to find comparable areas in which the lack of
a project is the only difference. In reality, comparable
areas can seldom be found. Thus, the analyst must accept
some divergence for his desired control group. The sensi-
tivity of the growth rates experienced under different
industry mixes to certain kinds of public investments has
not been established. The effects of different size com-
munities and the resultant differences in scale economies
might negate the supposed correction for population dif-
ferences -- i.e., the use of per capita income, etc.
Similar precautions apply to the use of "transplanted
multipliers." Differences in industrial mix and structure
and in local consumption and investment propensities can
cause income multipliers from seemingly similar areas to
differ considerably. The amount by which they might differ
could be partially checked through the use of Bromley's
index of internal purchases discussed below. However,
where analytical alternatives are possible such ”trans-

plants" should be avoided.

84

Index of Internal Purchases

Bromley argues that, when one considers the Oppor-
tunity costs Of the technical manpower involved, input-
output analysis has been overdone in recent years. He main-
tains that, for most of the purposes for which local busi-
ness income multipliers are used, acceptable local multi-
pliers can be generated at a much lower cost than by con-
structing input-output models. Thus, he argues for the
adoption of the "index of internal purchases" approach
where the analyst-planner desires a ranking of relative
multipliers between sectors rather than necessarily the
absolute magnitude Of those multipliers. He demonstrates
that, in eight cases for which input-output multipliers
could be compared, the relative rankings of sector multi-
pliers yielded by the index of internal purchases are
identical to those yielded by the input-output tables.38

Very simply stated, the Bromley approach would
arrange the sectors into a four quandrant format according
to the proportions of their purchases and sales which are
local. Those sectors for which the proportions are high
for both purchases and sales would be placed in quadrant I;
in quadrant II would be those sectors which bought from
sectors in quadrants IV and I and used them in producing
items for export; the quadrant III sectors would be those

which import much of their inputs and export most of their

‘39“.
ALL;

85

output; quadrant IV sectors would be those which bought
little locally yet sold a great deal locally.39

Bromley maintains that such a breakdown is sufficient
for most community development planners, since what they
really need to know is for which sectors they should pro-
mote growth to achieve the greatest secondary effect upon
local business. For their purposes, the actual size of
those multipliers is only of secondary importance for in-
ternal planning purposes.

While the index of internal purchases might be suf-
ficient for many local planning purposes, it constitutes
a very imprecise means for estimating secondary benefits.
The index yields an ordinal measure of sector impacts which
can be useful to the local planner. However, for compari-
sons between localities Or between projects justified
largely by primary benefits and those seeking justifica-
tion based upon their secondary impacts, the index is not
quite so helpful. Thus, it is doubtful that the approach,
as it now stands, could serve alone as a means for estima-
ting secondary benefits. However, it is quite possible
that the index could be of use as a check upon other ap-
proaches, such as the more aggregative multipliers dis-
cussed above, in estimating whether the actual sectors
impacted might have multipliers greater than or less than
the aggregate multiplier used. Such a procedure might, at
best, allow for an adjustment upward or downward of the

aggregate multiplier to reflect those differences.

 

’t‘i

86

From-to Analysis

 

From-to analysis is one approach in an increasingly
pOpular area of economic analysis known as "inter-industry
economics." From-to analysis differs only slightly from
input-output analysis, the difference lying chiefly in the
form and extent of data collection. Whereas in the typical
regional input-output study data are collected regarding
both purchases and sales by each sector, in the from-to
study only sales by each sector are recorded. The result
is a much less costly data collection process in the
from-to study. However, a valuable means of cross-checking
is lost in the from-to approach.

40 have argued that for many problems

Kalter and others
in regional analysis, a model which followed output flows
alone would be sufficient. For firms in the region selling
to other firms in the region, the from-to approach would
provide information on the internally produced (regionally)
inputs for the latter and the internal sales of the former.
Thus, for a completely closed region (i.e., a region doing
no trading with the outside world), the from-to formulation
would yield basically the same information as the input-
output approach when all of the inter-firm purchase data
were brought together.

Obviously, the from-to formulation ignores imports,
since sales from extra-regional firms are not included in
the data collection process. If one were interested only

in the intra-regional effects of certain impacts, this

In

(I?

87

exclusion would be of little importance. A more important
disadvantage for regional analysis purposes is the fact
that such a data collection process does not allow one to
estimate the value added coefficients for the producing
sectors. Thus, what is actually developed, if no further
additions are made to the basic data, is a matrix of intra-
regional requirements for the production of the total
sector outputs of the region.

The from-to approach is capable of most of the ana-
lytical uses to which input-output analysis can be put.
The most obvious exceptions, of course, are those concern-
ing inter-regional trade analysis, since the from-to
formulation does not include imports from other regions.
Most of the multipliers which can be derived from input-
output studies and which are relevant to secondary benefit
evaluation can be computed from the from-to study, if an
additional data collection effort is undertaken to provide
the value added components for each sector. Thus, the dis-
cussion of these will be saved for the section on input-
output analysis. Additionally, many of the assumptions
and limitations of from-to analysis are common to input-
output analysis as well and will be reserved for discussion
in that section.

Because from-to analysis includes only locally-
purchased inputs, the analyst assumes implicitly when
utilizing the model for projection that inter-regional

trade patterns will remain stable over time and as regional

88
output changes. Since multiplier analysis is inherently
a projection technique, one should be aware of the trade
stability assumption in evaluating secondary benefits
utilizing multipliers derived from such studies. If a
principal impact of the project under study were to stimu-
late import-substituting industries, the typical from-to
formulation would not allow one to measure the local second-
ary effects.

Jansma and Back41 used the from-to formulation in
estimating the local secondary impacts of small watershed
projects. The most unusual aspect of the study was the
method used in acquiring sales data. Microfilmed records
of checks which were cleared through the only bank in the
community were the basic source of data. Since information
was contained concerning both the payer and the payee, a
matrix could be constructed which included not only in-
ternal transactions but also imports and exports. The
difficulty, however, was that for small owner-operated
businesses, for some professionals and for farms, informa-
tion on household consumption were inseparable from business
inPut purchases. Thus, only induced multipliers could be
constructed, since some household sector expenditures were
implicitly endogenous. As a result, information regarding
inPUt requirements and inter-industry linkages was somewhat
Obscured. In the previously mentioned studies by Kalter,
on the other hand, such a problem was not encountered,

Since firm-by-firm surveys were taken. However, the

 

Cit

Gr:

7-3

89
survey method used by Kalter would require a household
survey, as well as the firm surveys, to provide the informa-

tion implicit in the approach used by Jansma and Back.

Input-Output Analysis

 

1n the input-output model has been called a static
general equilibrium model which depicts the inter-industry
and inter-market flows of goods and services at a given
point in time.42 Though some have questioned the "general
equilibrium" aspects of the above characterization, there
is general agreement that the input-output technique is a

comparative statics analysis.43

That is, the inter-
industry transactions matrix which is basic to input-output
analysis is analogous to a "snapshot" of the economy at a
single point in time.

The basic unit of the input-output model is the
transactions matrix or table. The rows of the table de-
scribe the sales by each industrial sector to other in-
dustrial sectors and to final demand; while the columns
detail the purchases by each industrial sector from other
industrial sectors (and from other regions, if the model is
"open") .

From the transactions matrix, two other matrices are
generated in most input-output studies. The direct re-
quirements matrix gives the breakdown of each dollar of

gross Output by each sector and describes its flows to other

sectors for inputs and to value added. The procedure

 

 

1"

...

u-y.

v».

tll

£11

4,,

'CA‘

;":
1.,

IEL‘

 

90
involved is to simply divide the elements in each column of
the transactions matrix by the adjusted gross output for
that industry.44 The third matrix is the direct and in-
direct requirements matrix, which describes the input re-
quirements, both direct and indirect, which result from
each dollar of gross output by each sector. The direct
and indirect requirements matrix considers the derived
demands which result from a change in the final demand
for the output of one sector and the second and ensuing
rounds of demands which are derived from the demands for
the inputs used to produce the output to meet that change
in final demand. In most cases, when both the direct and
indirect effects have worked themselves out, the output
requirements imposed upon the sector experiencing the
change in final demand will exceed that of the final demand
change alone. This is due to the fact that additional
outputs of that sector will be required as inputs into
other sectors which supply inputs to the former sector
(or perhaps to other sectors which do so). Thus, obviously,
the coefficients in the cells of the direct and indirect
requirements table will be larger than those of the direct
requirements table.

The direct and indirect requirements matrix can be
computed from the direct requirements matrix. One method
of computation would be to simply multiply through the
round after round effects of one sector on another. Since

the round-by-round effects become successively smaller,

 

o '-

n\

 

«\b
.1

.‘~
. is

\u.\
.\s

 

91
probably eight to ten such computations would yield a
fairly close approximation to the "infinite series" sum.
In some cases, such a round by round procedure is used --
generally because the analyst feels that the infinite series
sum is inappropriate for some reason or another and that he

45 How-

should "truncate" the expansions after a few rounds.
ever, in the general case, the full measure of the expan-
sionary effects is computed in which case the round-by-
round method wOuld become a bit laborious, especially for
some of the larger matrices. By inverting the "leontief
Matrix," one can achieve the same results as by multiply-
ing through all of the round effects. This was the basic
eclectic contribution of Wassily Leontief,46 the applica-
tion of matrix techniques to the analysis of interindustry
flows. The Leontief matrix is defined in matrix notation
as [I-A], where A is the direct requirements matrix and I
is an identity matrix of the same dimensions.

The input-output model is composed of a system of
linear equations. As stated above, the equations depict
in snapshot fashion the interindustry flows of goods and
services at a particular point in time. Thus, while the
input-output model provides a very detailed view of the
current structure of inter-industry flows, when used for
projection purposes it has all of the problems which attend
the use of a static model for such purposes. For instance,
the direct requirements coefficients reflect the average

rather than marginal relationships between inputs and

92
outputs. Thus, in using the model to project input re-
quirements for higher levels of output one assumes im-
plicitly that the marginal relationships are equal to the
average. This amounts to imputing a linear homogeneity to
the production functions involved. Further, since the pro-
portions between the inputs used and the outputs of each
sector are fixed in the model, it is implicitly assumed
that no input substitution will occur in any of the pro-
duction procesSes. Stated another way, if used for pro-
jecting, the model assumes that the supply functions for
all factors are infinitely elastic at the prevailing prices.
Thus, one of the ”frictions" which the projection version
assumes away is that of factor supply inelasticities, in-
cluding any market imperfections which might contribute to
those inelasticities. While a real descriptive strength of
the model is its ability to show explicitly the interde-
pendencies between industries, this same explicit statement
of interdependence allows the following seemingly-logical
interpretation of sectoral importance: Since the outputs
of some sector (e.g., the paper clip industry) are used
either directly or indirectly in fixed proportions with
other inputs by all other sectors in the matrix, a shortage
or a blockage in the supply of paper clips would force the
halting of all industrial production. Of course, this is
absurd. Nevertheless, a strictly-interpreted logical

extreme is that such would be the case. Thus, the

93
fundamental economic principle of input substitution is ig-
nored by gross application of the input-output model.

While the input-output model obviously has its weak-
nesses, it is still the most powerful of the regional
models. No other approach allows one to deal with regional
economic impacts at such a disaggregated level. Though
the literature on input-output analysis abounds with dis-
cussions of "the aggregation problem," the aggregation in-
volved in input-output analysis is very slight in compari-
son to that of the economic base model or the Keynesian
model. Thus, input-output analysis allows a great deal
more specificity on the part of the analyst in estimating
the extent and the incidence of secondary economic impacts
on a local community. The disaggregation in the input-
output model allows one to compute many multipliers rather
than a single "average" multiplier. Multipliers for each
industrial sector can be computed. Thus, if the incidence
of the impact is known, the multiplier effects can be esti-
mated with much more precision than is possible with the
more aggregative (and, hence, much less costly to construct)
multipliers such as those from the economic base model.

As pointed out above, the input-output model allows
one to construct a number of multipliers. Thus, the term
"the multiplier” is not even singularly applicable to
input-output multipliers. Some of the most widely used

input-output multipliers are discussed below and their

94
advantages and disadvantages for secondary benefit estima-

tion are noted.
Output Multipliers

The simplest of all input-output multipliers to com-
pute are the "output multipliers.” These simply show the
direct and indirect changes in gross output stimulated by
a $1.00 change in demand for each industrial sector. Thus,
the output multiplier for a particular sector is merely the
sum of the entries in the vector of "direct and indirect
requirement per dollar of sales" for the industry in ques-
tion.

Output multipliers can be instructive as to the
"activity" stimulated by a particular policy, where activity
refers to gross economic output. If one were interested,
for some reason, in maximizing the level of economic ac-
tivity generated by a given volume of exogenous impact, the
output multipliers would serve as a useful guide to the
choice of sectors towards which the emphasis should be
directed. However, one is more often interested in the
income-producing effects of a policy rather than the stimu-
lation of gross activity. Thus, while output multipliers
can yield valuable information as to relative sectoral
interpedence, they can be misleading when used to estimate
income effects.

In general, the larger the volume of input purchases

from other sectors per dollar of sales, the larger will be

95
the output multiplier for a particular sector. At the same
time, however, the volume of such purchases is inversely
related to the value added coefficient for the sector in
question. Thus, the use of output multipliers as a guide
in determining which industrial sectors to emphasize in a
community development plan leads one to adopt a development
strategy which is perhaps diametrically opposite to what
many would consider the "proper" strategy, since the sectors
having the lowest value added per dollar of sales tend to
be those having the greatest sectoral interdependence.
Unless the agglomerative attractions of the sector empha-
sized are great, such a policy can lead to a predominance
of low-wage (or low employment) and perhaps low-profit in-
dustries in the area. Obviously, few local communities
would desire to adopt such a strategy. Yet, in their
input-output study of the Oklahoma economy, Doeksen and

Little47

advised that, since the output multipliers for
"agricultural processing" and "livestock and livestock
products" were relatively large, those sectors should be
emphasized for development purposes.

While the relative size of the output multiplier is
inversely related to the direct income (i.e., value added)
of a sector, it is perhaps positively related to the in-
direct income produced by that sector. Since relatively
larger volumes of activity are stimulated by sectors

having large output multipliers, the possibility of second-

ary net income stimulation occurring is much greater for

96

these sectors, also. However, the income secondarily stimu-
lated also depends largely upon the income-producing capa-
city of the linked industries. Thus, the output multipliers
can not be accepted as dependable guides to the secondary
income-producing capacity of a sector, though there perhaps
exists some degree of correlation between the two. For
such purposes, one is best advised to use one of the income
multipliers discussed below.

For secondary benefit evaluation purposes, one is
more likely to be interested in income than in "activity
generation." Thus, output multipliers are not likely to
be particularly helpful. Additionally, since in secondary
benefit evaluation one is likely to be dealing with
economically-depressed localities, output multipliers can
create a dangerous bias towards perpetuating a poor in-
dustrial mix for the locality. This is true because of the
tendency for from-to models and regional input-output models
to identify the strongest intra-regional linkage effects
in the structural matrix. The problem comes in situations
in which the regional industrial mix includes a pre-

dominance of slow-growing industries.48

The regional
interindustry model shows the greatest interdependence
effects, obviously, for those industries which are "estab-
lished" in the area. Thus, the output multipliers will be
greater for those sectors which are linked to the (slow

growing) sectors which are already there. At least when

the income multiplier is used instead of the output

“In;

E.—

11M

Fe

97
multiplier, those sectors which pay low wages (or employ
few people) are penalized in the relative size of their
interdependence effects. The output multiplier includes

no such penalty.
Type I (direct-indirect) Income Multipliers

A better guide to the income-increasing effects of
sectoral interdependence than that provided by output multi-
pliers is that given by the direct-indirect income multi-
pliers. The basic difference between the two is the em-
phasis of the latter upon income stimulation rather than
activity generation. While the output multiplier shows
the direct and indirect activity generated per dollar of
sales for the sector, the direct-indirect income multiplier
depicts the direct and indirect income changes in compari-
son to the direct income changes. The multiplier which
results is the ratio of direct plus indirect income to
direct income alone.49

The direct-indirect income multiplier was developed
from an input-output study of the St. Louis Metropolitan
area by Werner Hirsch.50 In that same study, Hirsch pre-
sented a second income multiplier which included, in ad-
dition to the direct and indirect income changes, an
"induced" component. The induced component incorporated
the effects of re-spending by households into the multiplier

effect, which the direct-indirect income multiplier ex-

cluded. Hirsch called these multipliers "Type I" and

 

It ~ It \ at} E ;h\ Fly. 1

5“ \u. I
n\. . ‘l « N. .. shun C r . .l‘ W ,P. a »

TE

98

Type II” multipliers, the Type 11 being the multiplier
which included the induced income component. The Type II
multipliers will be discussed later in a following section.

To derive the direct-indirect (Type I) multipliers
two matrixes are needed: First, the direct requirements
matrix with households included in the structural matrix
is needed. In other words, a row must be added to the
transactions matrix to show the wages, salaries and profits
per dollar of direct sales, and a column must be added which
shows the purchases by households from every other sector
and from households (the sums of the column and the row
entries must be equal). Secondly, the matrix of direct
and indirect requirements per dollar of sales (with house-
holds outside of the structural matrix) is needed. To
estimate the direct-indirect income multipliers for a
sector (call it sector A), one would multiply the elements
in the sector A column of the direct and indirect require-
ments matrix by the household row of the direct require-
ments matrix and sum the products (i.e., using matrix pro-
cedures, post-multiply the household row of the direct
requirements matrix by the "A" column of the direct and
indirect requirements matrix); then divide the sum of the

51 The

products by the sector A entry in the household row.
result is the Type I income multiplier for sector A. It
expresses the relationship between direct and indirect

income changes due to sectoral interdependence and direct

changes in income in sector A.

99

As stated above, if one desires a measure of the
relative income-producing sector linkages, the direct-
indirect income multipliers provide a much more reliable
sector ranking device than do the output multipliers. How-
ever, like the output multipliers, the direct-indirect in-
come multipliers do not include the feedback effect from
household expenditures. Thus, these multipliers only show
the buildup in household income which results from the
direct and indirect requirements imposed upon other sectors
(and itself) by changes in the output of sector A.

A point should be made in passing regarding the
relevant change to which the direct-indirect income multi-
plier should be applied. While the input-output tables of
requirements show the input required from other sectors to
allow a one dollar change in output (sales) of sector A,
the direct-indirect income multipliers show the change in
both direct and indirect income which results from a one
dollar change in direct income. Thus, in estimating the
direct and indirect income effects of a one dollar change
in the output of sector A, one should not simply multiply
that one dollar change by the above multiplier. One would
be applying the multiplier to the gross change in output
of sector A rather than to the direct change in income.
Thus, the direct-indirect income changes are much smaller
than is sometimes erroneously estimated using such multi-
pliers. For example, if the household coefficient in

Column A were .20 and the sector A multiplier were 2.0,

100

a one dollar change in export sales of sector A would not
generate two dollars in income, but, rather, it would
generate forty cents in income. The difference comes in
how one defines the initial change in direct income. If
the dollar initial change is, in fact, the direct income
component, then the multiplier should properly be applied
to that figure. However, if the dollar is instead an
exogenous change in export demand, then for regional
analysis purposes the "initial income" aspect of that
dollar is realized in another region; and the initial re-
gion income is only twenty cents. This differs somewhat
from the output multipliers, discussed above, because of
the difference in the base. The base of the output multi-
pliers is formed by the initial dollar change in demand
for the output of that sector. Thus, the output multiplier
should normally be applied to that figure. However, if the
full income effect of that initial dollar expenditure is
not realized in the region, then that dollar should be
subtracted from the effects of the output multiplier. This
differentiation between the initial income incidences was
not considered by Doeksen and Little in their analysis,
however, since they were dealing largely with gross activity
generation instead of changes in local income, per se.

Since the Type I multiplier ignores the effects of
the household re-spending (induced) component, it assumes,
in effect, that the marginal propensity to consume locally

(out of household income) is zero. Thus, if one derived

101
a set of Type II multipliers and added an adjustment for
the propensity to consume locally, the Type II multiplier
would reduce to the Type I multiplier when the MPCL=O. Such
a situation would only exist in the "real world” if the com-
munity in question had no retail sector nor other facilities
for selling to households, so that all consumer spending
occurred outside of the community. For all practical
purposes, the only communities which would fit such a mold
would be industrial parks or city sectors which were zoned
for industrial develOpment only. It is unlikely, however,
that an analyst would define such a region as the relevant
one for study -- especially in the evaluation of secondary
benefits, which should have their basis in real changes in
consumer welfare.

The process of multiplier analysis is a very pre-
carious one wrought with some very great assumptions re-
garding economic behavior. The assumptions made in the
calculation of the indirect requirements matrix provide a
case in point: It is assumed in carrying out those calcu-
lations that all of the factors needed for the envisioned
expansion (both directly and indirectly) will be available
at current prices; further it is assumed that they will be
used in the same proportions as depicted in the tables and
that no thwarting of the derived demands for them will
occur via market imperfections, etc. The net effects of
these assumptions, in all probability, is to bias upwards

the estimates of the real expansionary effects. If, at

102
the same time, an induced component is added, the resulting
projections can greatly exceed the actual multiplier ef-
fects. Thus, if one suspects that the "leakages” from the
endogenous matrix are greater than those depicted, and he
at the same time suspects that the induced effects are over-
stated (the reasons this might be so will be discussed in
the sector on Type II multipliers which follows), then the
rule-of-thumb correction provided by the Type I multiplier
might, in fact, make it a better estimate of the actual
multiplier effects than the Type II multiplier. Thus, one
might choose to use the Type I multiplier as a "more reason-
able" estimate of the true secondary impacts under such cir-
cumstances.

In secondary benefit evaluation, the analyst should
attempt to identify the net secondary changes in real in-
come which accrue to the relevant unit of analysis (i.e.,
to the nation or to the region, depending upon which
"account" he is dealing with). The output multipliers are
better in one respect than are the more aggregative ap-
proaches: They allow the analyst to identify the sectors
experiencing the greatest gross activity changes and, thus,
can help point out the sectors in which supply conditions
can be most critical. However, they are weaker in two
respects: First, they incorporate the gross output con-
cept rather than income; and, second, they do not include
the induced component of income growth. The Type I multi-

pliers, on the other hand, go one step further than the

103
output multiplier in that they utilize the income effects
as the relevant multiplier effects rather than gross output.
However, they do not include the induced income growth com-
ponent. Thus, their applicability is limited largely to
the three following uses: 1) determination of the sectors
having the greatest inter-sectoral income linkages (exclud-
ing the household re-spending component) for purposes of
sector emphasis; 2) situations in which the marginal pro-
pensity to consume locally is assumed to be very low; or
3) as a rule-of-thumb adjustment for "frictions" and leak-
ages in the structural matrix and for leakages from the

induced component.
Type II (direct-indirect-induced) Multipliers

The Type II multiplier differs from the Type I multi-
plier by the addition of the induced component only. The
induced component incorporates into the multiplier the
expansionary effects of the re-spending by households of
incomes earned during the course of the initial and ensuing
output expansions. Thus, the household sector is assumed to
act in a fashion similar to the producing sectors in the
structural matrix -- i.e., it increases its "requirements"
as "sales" (household income) expand.

'In recognition of this link between household ex-
penditures and household income, Hirsch removed the house-
hold sector from the final demand category and placed it

in the structural matrix in computing his Type II

104
multipliers. To do so, he had to determine the distribution
of expenditures by households to each sector in the struc-
tural matrix (including other households); and he had to
determine the part of each sales dollar for each sector
which went to households in the form of wages, salaries,
profits, and interest. Having determined these, Hirsch
re-inverted the Leontief matrix with households included in
the endogenous matrix. Then, to estimate the multiplier
for some sector (Sector A), he took the household coef-
ficient for Sector A from this new matrix and divided it
by the household coefficient for Sector A from the direct
requirements matrix._ The ratio which resulted was the
Type II multiplier.

While, in concept, the addition of the induced com-
ponent offers a more realistic approach to estimating
multiplier effects, there are two new problems created by
its addition. First, since most regional input-output
studies do not include a household survey, the household
column has to be estimated outside of the model. Typically,
the coefficients from national input-output studies are
used in regional studies. Thus, if no adjustments are made,
one is forced to assume that inter-regional consumption
patterns are invariant. To appreciate the compromise which
such an assumption entails one must simply reflect upon the
fact that national coefficients will include items con-
sumed in one part of the nation which perhaps are not even

marketed in the region for which the study is being made --

105
e.g., alcoholic beverages sales are illegal in many south-
ern counties. The fact that homes in Maine incur very sub-
stantial heating costs while many Florida homes do not
even have central heating can only be rationalized by
assuming that the Florida air conditioning costs Offset the
difference. If air conditioning and heating implements are
produced in the same industrial sector, then the problem
is perhaps lessened. Nevertheless, regional coefficients
would normally provide better measures of induced multiplier
effects in the region than do those borrowed from national
tables.

Secondly, the addition of the induced component also
necessitates the additional assumption that the mix of
household purchases remains the same as income increases.
The implicit assumption is that the income elasticity of
demand of all items in the household budget is equal to
unity. Thus, as income increases by 1 per cent, expendi-
tures for all items in the household "market basket" in-
crease by 1 per cent. This is analogous to the assumptions
made concerning the other "producing" sectors in the struc-
tural matrix regarding the linear homogeneity of the pro-
duction functions involved. However, while there has been
an unresolved controversy regarding the fixity of the in-
put coefficients, the difficulty posed by the static con-
sumption function can be resolved by making one rather
palatable additional assumption. That assumption concerns

the manner in which income grows as a result of the

106

expansion. If output (and income) increase as a result of
extensive growth rather than intensive growth, then the
assumption can be made that the consumption patterns of
the "new" money earners will be the same as those of the
"old" money earners at the same level of income. Thus, if
regional output increases as a result of bringing in new
workers at approximately the same wages as the resident
workers, it would appear reasonable that the ”average” con-
sumption patterns would not change very much. However, if
income grew as a result of more intensive use of resident
workers and average incomes increased as a result, then one
should reasonably expect the pattern of consumption to
change.52

While the Type II multiplier includes the induced
multiplier effects of household spending, it does not allow
for the effects of other income-related changes in expendi-
tures which will have further expansionary effects upon
income. Increases in investment and government expenditures
which will occur as a long-run effect have been excluded.
Thus, the Type II multiplier is only a short run multiplier.
To adjust for the long-run induced effects of these two
additional factors one would follow the same procedure as
was followed in incorporating the household re-spending
into the model. Column and row values for each would have
to be estimated, adjusted and placed in the matrix of
direct requirements. Then a new Leontief matrix would be

computed and inverted with households, investment, and

107

government all in the endogenous matrix with the producing
sectors.53 The long-run induced multipliers would be com-
puted in the same manner as the short-run induced (Type
II), the difference being that the household coefficients
\vould be larger as a result of the effects of "capturing"
‘the effects of investment and government spending in the
s tructural matrix.

The difficulty with the long-run induced multiplier
:is that it compounds the already-unrealistic lack of con-
ssideration of leakages from the model. The problems posed
11y the assumptions made concerning the "frictions" and
ileakages in the input-output model have already been dis-
cmissed. Isard54 highlights the problem by pointing out that

an input-output model as such contains no

restrictions on rates of expansion and contraction

and on multiplier effects. The SOphisticated

analyst usually allows for such restrictions;

typically, he establishes the final demand sectors

at such magnitudes tha, being fixed, they auto-

matically set reasonable restraints on rates of

change and multiplier effects. When, however,

the important households sector is removed from

the final demand sectors, unrealistic rates of

change and multiplier effects do result unless

other modifications are made which introduce com-

pensating restrictions or leakages.
(Inite obviously, the removal of two more sectors from final
demand multiplies the problem. The removal of all the
final demand sectors would, obviously, lead to an under-
identified system of equations.55 The long-run induced

multiplier comes very close to creating the latter problem,

in that the only factors left to "explain" the level

108
of the system are exports (and exogenous public and private
investment in the case of regional analysis, where only
As a result, the

local governments are made endogenous).

long-run induced multiplier is generally an unreliable

estimator of the actual long-run multiplier effects.56 Be-

cause of the problems caused by leakages, the (short-run)
'Type II multiplier might, in fact, provide a more reliable

estimate of the long-run multiplier effects than does the

long-run induced multiplier.
Input-Output and Secondary Benefits: An Overview

Because it allows the most detailed analysis of
sectoral interdependence, input-output analysis constitutes
the most effective tool the analyst possesses in evaluating
Shecondary impacts. The biggest problem in applying the
tkschnique is the cost and the time required to construct
tile tables. Where input-output studies exist for the im-
IPElcted region, however, the opportunities for detailed
allalysis of secondary benefits far exceed those Offered
13)’ other techniques.

Input-output analysis is not without its problems as
3J1 analytic technique, though its problems are perhaps
“Hach less severe than many other regional analysis tech-
Iliques. Its greatest weakensses are manifest when it is

Ilsed for projection purposes; multiplier analysis is a

form of projection.

109
Dorfman points out the problem of time involved in
interpreting the matrix of requirements. Since the techno-
logical relationships in the matrix refer to the current
level of output of each industry, the past level of output
of its suppliers, and subsequent levels of output of its
customers, input-output analysis" . . . abstracts from the
time sequence of production and exchange . . ." and, thus,
". . . applies only to a stationary equilibrium, where time
is of no consequence."57 Since the multiplier techniques
include the round-after-round expansions as well, the
‘problem of time is compounded. Thus, the multiplier not
only abstracts from time in dealing with the direct re-
lationships, it also further abstracts from time in han-
<iling the "rounds" of expansions -- i.e., it does not deal
TVith the problem of differing time periods required between
ISectors for a "round" of output to occur. Neither do
iJIput-output (or any other multipliers) differentiate be-
twveen the present values of succeeding rounds of income.
Though input-output is the most disaggregative of

tlle multiplier techniques, there are still some problems
‘laiused by sector aggregation. Thus, in projecting outputs
fFrom a sector, one makes the assumption that the outputs
will be produced by the "average" firm or by a mix of
1firms perfectly representing the mix in the sector. The
aggregation problem is two-fold: Not only is the average

or the mix projected in terms of relative costs but also

110
in terms of products. Thus, one also projects the current
mix of products from that sector.

Dorfman feels that the neglect of substitution is the
most striking divergence of the Leontief model from con-
ventional economic assumptions. The fact that the coeffi-
cients are fixed and assumed to hold for all levels of
output has led some economists to be very critical of the
technique.

Another problem in projecting from input-output models
concerns the handling of investment. It is questionable
whether investment requirements should be included in both
relatively small short-run changes in output and relatively
large long-run changes. Dorfman argues that "The assump-
tion of a unique relationship between the output of an
industry and its purchases from other industries is
plausible only with respect to purchases on current ac-

count."58

Thus, projecting far beyond current outputs
'becomes increasingly precarious. Whether one should use
the same matrix of technical relationships to estimate both
short-run induced and long-run induced multipliers is
questionable, since the long-run multipliers explicitly
consider the induced effects of investment on income with-
out considering the decrease in input requirements gener-
ally signalled by long-run investment changes.59
The greatest strength of the input-output model is

its ability to depict in detail the structural interrela-

tionships within an economy. As a projection tool, the

111
least violence is done to the model when the same mix of
outputs as is currently being produced is projected, and
when the output being projected is not substantially dif-
ferent from the output depicted in the model. Thus, for
secondary benefit evaluation purposes, the model is most
applicable to the "conventional" category of secondary bene-
fits in which changes in the level of aggregate demand are
the principal actors. It is only in this group of secondary
benefits that the economic structure and the productivity
of the resources are treated as constants, while the exo-
genous forces presumed to move the economy to higher levels
of output are changes in demand. Since the other two types
of secondary benefits relate to changes in resource pro-
ductivity and/or changes in economic structure, the input-
output model can not be directly applied to their evalua-
tion. Thus, the secondary benefits of induced change
within the current industrial structure are the only kind
evaluated by input-output multipliers.

Input-output analysis provides a very interesting
picture of the interactions which occur in an economy as a
result of an exogenous impact upon that economy. At the
same time, there exists a tendency on the part of some
to view the problems of under-develOped areas as problems
of inadequate "economic activity." Thus, they view the
generation of "activity" in such economies as being gener-
ally beneficial. Since the input-output model provides a

very good abstract view of economic activity, its

 

112
descriptions are Often used in developing policy prescrip-
tions regarding the generation of increased levels of such
activity. In this regard, its descriptive powers have in
some cases led to misguided policies, as discussed above
in regard to output multipliers. That "economic activity"
in a given area and income to people in a given area are
not synonymous (or even very closely related) is a point
which should be re-emphasized. The fact that direct and
indirect activity in an input-output model is maximized
when value added in all linked sectors is zero should
serve to point up the fallacy involved in using "activity"
as a guide for economic policy. When value added is zero,
net welfare of society is not enhanced by changing the form
of the resources, unless for some reason "activity" itself
is inherently valued.

Whether the analytic tool used to evaluate secondary
benefits is the input-output model or any of the other ap-
proaches discussed herein, the fact remains that the rele-
vant unit to be measured is the change in the welfare of
the peOple involved. Because input-output analysis allows
the greatest degree of disaggregation, it allows the
analyst to come closer to evaluating the real inter-
personal changes than does any of the other techniques
discussed. However, it also allows the analyst " . . .
to lose sight of the complexities and interdependence of
the real world in a maze of pretentious and unhelpful

60

sumbols." Too many users of the technique have perhaps

113
been so enamored with the system of equations and the
mathematical possibilities offered by them that their use
of economic theory in applying the description provided by
the equations has been neglected. It is in the application
of this tool that theory plays its greatest role, not in

the estimation of the structural relationships.61

Thus,

the economics involved in the estimation of secondary bene-
fits using the input-output approach does not end with the
inversion of the Leontief matrix, nor with the computation

of the income multipliers; and it should begin long before

the estimation of the structural equations.

Linear Prpgramming

 

Linear programming is generally defined as a technique

for maximizing some linear objective function, subject to

a set of linear restraints. The technique can be used to
either determine the allocation of a set of resources to
achieve maximal returns or to determine the minimum cost

sep of resources needed to achieve a pre-specified output.
It has been a widely used tool of firm-level analysis.
Linear programming is being increasingly used in analyzing

regional problems as well.

Regional Linear Programming. Linear programming can
be used to estimate the optimum allocation of a regions
resOurces to different products and processes. Just as

the technique can be used to determine the Optimum product

114
mix for an individual firm, it can also be used in estima-
ting the optimum degree of specialization for a particular
region. The technique has been discussed as a tool of
analysis in land use zoning and has been used for such
purposes on some occasions. The data requirements for
regional linear programming are, of course, very great.

Inter-regional Linear Programming. Regional linear
programming, like regional input-output in a developed
country, must deal with categories of inputs and outputs
which are "exogenously defined." The smaller the region,
the larger the number and the greater the importance of
the exogenous variables.

Inter-regional linear programming attempts to make
endogenous to the model the inter-regional effects of in-
put and output decisions. For example, the optimum degree
of specialization and trade for a particular region depends
not only upon the endowment of resources of that region
but also the regions with which its economy is linked.
Thus, one use of inter-regional linear programming would
be the allocation of production and trade between regions.

River Basin Planning Models. The widest use of
linear programming in the water resource field has been
in river basin planning. The greatest resemblance to an
indrect benefit model has been achieved by the Economic
Research Service group. The Economic Research Service

62

model uses a least-cost solution approach to estimate

the benefits from flood control projects where an increase

115
in agricultural output in the flood plain is involved. By
using the projected demand for agricultural commodities
during the project life as a guideline, the Economic Re-
search Service analysts established regional output re-
quirements by commodity. A share of the total "demand" was
allocated to the Wabash River Basin and cropland adjust-
ments were computed for basin lands under flood-prone and
flood-free conditions. The implicit attempt at estimating
inCome shifts between up-land and flood plain land owners
holds valuable prospects in terms of determining who re-
ceives the benefits and who is indirectly hurt by flood
control projects.

Input-Ourput as a Linear Program. The input-output

 

model is very similar to the linear programming model. It
is often described as a linear programming model. The
similarities are basic: 1) The relationships between
inputs and outputs are fixed or linear, and 2) the rela-
tionships between inputs are fixed or linear. When used
for projection purposes, the input-output model is very
similar to a linear program. If overt restraints are
placed upon the projections, the input-output model be-
comes a form of linear programming. Linear programming
solutions to multiplier problems are most applicable where
unemployment exists in a regional economy, but in which
the unemployment is not sufficient to meet the labor re-
quirements indicated by the multipliers derived from

input-output models. By placing labor restraints on the

116
model, the round-by-round computation of income impacts can
be stopped at the point at which the unemployed labor force

has been used up.

 

Secondary Benefits and the Multiplier: A Summary

Of the three types of secondary benefits -- i.e., 1)
conventional, 2) externalities, and 3) dynamic -- the most
widely discussed and the one generally subjected to analysis
has been the conventional variety of "stemming-from" and
"induced-by" changes in economic activity. This has occur-
red largely because this type has been most amenable to
macro-economic analysis. The latter two types -- while
perhaps equally or more important in terms of economic
growth inducement -- have been more difficult of analysis
because they represent changes in the production functions
and in the inherent productivity of the resources involved.
Thus, they are not handled as easily by conventional eco-
nomic theory. While changes in scale economies, for
example, are a familiar part of micro-economic theory, the
implementation of the concept and the measurement of the
effects has not come easily. Neither has the measurement
of output changes due to increasing productivity of labor.
Thus, the basic approach to secondary benefit evaluation
has revolved around the estimation of aggregate changes in
output resulting from the secondary impacts of increased

expenditures due to a public project. The typical approach

117
has utilized the multiplier concept in some form to estimate
these impacts.

"The multiplier" measures the lateral shift in aggre-
gate demand which results from a unit change in exogenous
expenditures. Thus, it does not measure secondary income
changes directly, unless aggregate supply is perfectly
elastic over the relevant range or makes an equivalent
lateral shift. Unless one has reason to believe that one
of these two aggregate supply conditions is being met,
multiplier analysis should not be used to measure benefits
in the absence of at least a cursory analysis of supply
conditions.

The conditions under which the full multiplier effects
Of’a public investment in a water resource project, as they
are typically estimated, would constitute a real net change
in welfare are fairly specific and very operationally
llimiting, though the general approach has been to include
tile full multiplier impact. The full value of the multi-
I>lier effects might be included where l) the conditions of
Sllpply for all factors stimulated to employment by the in-
Vestment are perfectly elastic at prevailing prices, 2)
tlle opportunity costs of those factors in the absence of

tlle investment are zero in all cases -- i.e., the full

fFactor price consists of economic rent, and 3) the outputs
Wl'lich result do not simply substitute for other products
ill the market place; thus they do not unemploy other

ffictors of production. Normally these conditions are not

 

118
present in the general case. Thus, multipliers in general
(whether Keynesian multipliers or input-output multipliers)
exhibit a consistent tendency to over-estimate the real
welfare effects of secondary impacts. That any or all of
the above conditions might partially hold in many cases is
not questioned. However, the typical approach to multi-
plier analysis assumes that all of these conditions hold
fully. Thus, the "secondary benefits" which are generally
estimated using secondary impact analysis are not really
net secondary benefits at all, even from the viewpoint of
the factors being employed. Rather, they are gross changes
(usually long run) in the demand for those factors.

Multiplier analysis is basically a projection techni-
que. All of the uses of the multiplier which were discussed
above deal with the multiplier as a tool for projecting the
expansions in income, employment or economic activity re-
sulting from certain exogenous changes.

In addition to projecting expansion, input-output
multipliers can also be used to prp£e_the incomes received
from current or direct output, since the input-output model
gives a detailed description of the factor payments in-
volved in producing the outputs of a particular industry.
The factor payments are based upon market prices which
might or might not represent the real opportunity costs of
those factors, nor of the resources which might have been
used in producing the factors. In those cases in which

the factor prices differ from the real opportunity cost,

119
the real value added for that sector will be incorrectly
stated. If one knew the real opportunity costs of all
factors, the real economic value added by a particular
sector could be more accurately estimated. That is, the
real shadow price of each output could be estimated by
repricing each factor entry at its real opportunity cost
and recomputing and inverting the Leontieff matrix (with
households outside of the endogenous matrix). The result
could be used in either or both of two ways: 1) To
project the real income from an expected expansion in out-
put due to an expected change in final demand (the estimate
would have all of the problems posed by projecting from
input-output models discussed above), and/or 2) to estimate
the total factor rents (i.e., factor prices minus real
opportunity costs) earned by all of the factors and sub-
factors used in producing the current output. The latter
use would simply constitute a full shadow pricing of the
current outputs. The former would constitute a full

shadow pricing of expected outputs.

120

Footnotes

 

1. Kneese, op cit, 24-9.

2. Harvey Leibenstein, "Allocative Efficiency vs.
'x-Efficiency'," The American Economic Review (June 1966),
392-415.

 

3. Peter Helmberger, "O-Efficiency and the Economic
Organization of Agriculture," in Agricultural Organization
in the Modern Industrial Economy (Columbus: 1969).

4. See Irma Adelman, Theories of Economic Growth
and Development (Stanford: 1961).

 

 

5. Glenn L. Johnson, "A Note on Non-Conventional
Inputs and Conventional Production Functions," in Agr'-
culture in Economic Development, ed. by C. Eicher and L.
Witt (New York: 1964).

6. Edwin Kuh, "Unemployment, Production Functions,
and Effective Demand," Journal of Political Economy, LXXIV,
No. 3, 238-48.

7. Notable exceptions are Ibid; Robert M. Solow,
"Technical Change and the Aggregate Production Function,"
The Review of Economics and Statistics XXX, 312-20; and
Lawrence‘H. Shaw and RObert S. Arden, "Output Effects of
a Changing Composition of Industry, 1947-1965," Ihg_
Review of Economics and Statistics XLX, 134-6.

 

8. Gardner Ackley, Macroeconomic Theory (New York:
1967), 309.

 

9. Gottfried Haberler, "Mr. Keynes' Theory of the
'Multiplier': A Methodological Criticism," Zeitschrift
ffir National 6konomie (1963).

10. Keynes; Op cit, 298.
11. See Abbas Mirokhor and Frank Orazen, "Importance

of the Farm Sector to the Economy: A Multiplier Approach,"
American Journal of Agricultural Economics (November 1968).

 

12. "The Relation of Home Investment to Unemployment,"
Economic Journal (June 1931).

13. The Theory of Inter—sectoral Money Flows and
Income Formation (Baltimore: 1950).

 

121

14. Harry W. Richardson, Elements of Regional
Economics (Baltimore: 1969), 23.

 

15. Robert H. Haveman and John V. Krutilla, Unem-
plgyment, Idle Capacity,_and the Evaluation of Public
Expenditures: _National_and‘Regi6naI'Analysis (Baltimore:
1968), 28—33.

16. op cit, 914.

17. See Walter Isard, Methods of Rggional Analysis:
An Introduction to Regional Science (Cambridge: 1960), I89.

 

18. Charles M. Tiebout, The Communitnyconomic Base
Study, Supplementary Paper Number I6 (nechOfk: December

19. Ibid., 61-2.

 

20. op cit, 199.
21. 32 cit, 200-1.

22. Richard B. Andrews, "Comment re: Criticisms of
the Economic Base Theory," Journal of the American Institute
of Planners XX (1958).

 

23. Paul W. Barkley and Thaine H. Allison, Jr.,
"Economic Base Studies in Resource Administration," Land
Economics XL: 470-9.

24. Hans Blumenfield, "The Economic Base of the
MetrOpolis," Journal of the American Institute of Planners
XXI (1955).

25. C. L. Leven, "Regional and Interregional Accounts
in Perspective," Regional Science Association Papers and
Proceeding, XIII: 127F44.

26. See "Homer Hoyt on Development of Economic Base
Concept,” Land EconomiC§_(May 1954).

 

27. See for example Economic Base Survey of the
Potomac River Service Area, prepared by the Office of’
Business Economies 0f the U. S. Department of Commerce for
the U. S. Army Corps of Engineers (Washington: 1969).

 

28. Barley and Allison, op cit.

29. Procedures for Evaluation . . . , 107-8.

 

30. See M. E. Marts, An Experiment in the Measure-
ment of the Indirect Benefits of'Irrigation, preparedfor

 

122

the Bureau of Reclamation, U. S. Department of the Interior,
Boise, Idaho (June 1950); and Robert E. Struthers, The Role
of Irri ation Development in CommunitygEconomic Structure:
Grand Vallpy Trade Area, COlorado (Washington: 1963).

31. Barkley and Allison, op cit.

32. For a more complete discussion of the "intensive"
versus "extensive" growth problem, see C. M. Tiebout, "An
Empirical Regional Input-Output Projection Model: The
State of Washington," Review of Economics and Statistics
(August 1969).

 

33. Impact of Irrigation Development on Income and
Trade: Eastern and CentraT'South Dakota (Brookings:
February 1969).

 

34. A Study of the Economic Impact of Water Impound-
ment thropgh the Development ofia Comparative Projection
Model, Tec nicaI’Report No.*8, water Resources Institute,
Texas A and M University (August 1967), 4.

 

 

35. Ibid., 42.

36. A Study of the Economic Impact of the Hugo
Reservoir on Choctaw andiPushmataha Counties (Oklahoma
City: 1969).

 

37. A Test of Procedures Proposed byya Task Force
of the WateriResource Council: Special—Study of ihe Mountain
Home Division, Southwest Idaho Water Deveiopment Project,
U. S. Army Corps ofiEngineers,CWa11a Walla District (March
1970).

 

 

 

38. "On the Use and Overuse of Input—Output Models:
A Methodological Hypothesis," unpublished paper, undated.

39. Ibid., 4-5.

40. R. J. Kalter, Estimating Local Secondary Impacts
of Water-Based Recreation Investment USin Interindustry
Anal sis (Madison: 1967), andiAn Interin ustry Analysis
pf the Central New York Re ion,CBullefin 1025 (Ithaca:
1969); Charies L. Leven, egional Income and Product
Accounts: Construction and Applications," in Desi n of
Regional Accounts (Baltimore: 1961), Chapter 6; and W.

Lee Hansen and Charles M. Tiebout, "In Inter-sectoral Flows
Analysis of the California Economy," The Review of Economics
Egg Statistics, XLV, No. 4 (November 1963).

 

 

 

 

 

41. J. D. Jansma and W. B. Back, Local Secondary;
Effects of Watershed Projects: A Case Study_of Roger
fins County, OkTalioma (Stilfivater: 1964).

 

123

42. William H. Miernyk, The Elements of Input-
Output Analysis (New York: 1967)} 30}

 

 

43. H. H. Stoevener and E. N. Castle, "Input-Output
Models and Benefit-Cost Analysis in Water Resources Re-
Search," Journal of Farm Economics (November 1965),
1572-9.

 

44. See Miernyk, op cit, 21.

45. For an example of such a truncation and a dis-
cussion of truncated multipliers, see Walter Isard and
Robert Kuenne, "The Impact of Steel Upon the Greater New
York-Philadelphia Industrial Region," The Review of Eco-
nomics and Statistics (November 1953).

 

 

46. "Quantitative Input-Output Relations in the
Economic System of the United States," The Review of
Economics and Statistics (August 1936).

 

 

47. Gerald A. Doeksen and Charles H. Little, A2
Analysis of the Structure of Oklahoma's Economy by
Districts (Stillwater: 1969).

 

 

48. Which also tend to be low-wage, low-value added
industries. See Harvey S. Perloff, et a1., Re ions Re-
Sources, and Economic Growth (Lincoln: 1960) 63-6.

49. Miernyk, op cit, 47.

 

50. "Interindustry Relations of a Metropolitan
Area," The Review of Economics and Statistics (November

 

51. See Ibid., and Miernyk, op cit, 42-9.

52. Charles M. Tiebout, "Regional and Interregional
Input-Output Models: An Appraisal," Southern Economic
Journal XXIV (1957).

. 53. For a discussion of long-run versus short-run
Input-output multipliers, see Kalter, Estimating_Local
§gcondary Impacts . . ., op cit.

 

54. Isard, Methods of Regional Analysis . . . , 35.

55. Mordecai Ezekial and Karl A. Fox, Methods of
Correlation and Regression Analysis: Linear and'Curvi-
llnear (New York: 1959), 4IS¥33.

 

. 56. Kalter comes to a similar conclusion. See loc
Clt, 77.

 

 

 

Ira; ("a

lp—ril0w—1

124

S7. "The Nature and Significance of Input-Output,"
Review of Economics and Statistics (May 1954), 121-33.

58. Ibid., 124.

59. For a discussion of this point, see H. B.
Chenery, et al., The Structure and Growth of the Italian
Economy (Rome: 1953), T7.

 

60. Keynes skepticism regarding the mathematical
models of some of his colleagues is perhaps relevant to this
discussion:

"Too large a proportion of recent 'mathematical'
economics are mere concoctions, as imprecise as

the initial assumptions they rest on, which allow

the author to lose sight of the complexities and

interdependencies of the real world in a maze of

pretentious unhelpful symbols." The General Theory,

 

298.
61. Dorfman, op cit, 125.
62. Roger Strohbehn, et a1., Analysis of Alternative

Procedures for the Evaluation of Agricultural Flood Control
Benefits (WaShington: 1971).

 

CHAPTER V

ESTIMATING SECONDARY FACTOR INCOME CHANGES

Foreword

 

In Chapter II, the literature regarding secondary
benefits was reviewed. The survey revealed three classifi-
cations of secondary benefits: 1) Customary induced-by and
stemming-from effects, 2) externalities and economies of
scale, and 3) dynamic secondary benefits. It was concluded
that the latter two categories required a great deal of
further study and that, at present, they must be approached
on a case-by-case basis. Even more so than the customary
variety of secondary benefits, the analysis of externali-
ties and dynamic effects requires great perception and
imagination on the part of the analyst. Only the customary
variety of secondary benefits appeared amenable to analysis
using currently available economic knowledge and currently
available secondary data.

In Chapter III, methods of estimating secondary im-
pacts were surveyed. The advantages, disadvantages, and
assumptions of each were discussed. It was concluded that
the major drawbacks to most approaches have been the

assumptions of perfect supply elasticity and zero

125

126
opportunity cost of the factors stimulated to employment
and the tendency to credit all ensuing changes in the level
of income or employment to the initial investment. The
latter tendency gives no credit to the associated invest-
ments required. It was concluded that the detail provided
by the input-output model allowed the analyst the greatest
opportunity to adjust for the effects of these factors. It
has been maintained throughout the preceding sections of
this study that the analyst should apply imagination and
reason to the analysis of customary secondary benefits
rather than faithfully accepting an estimate of "Th3
Multiplier" as an effective measure of real welfare changes..
The purpose of this section is to demonstrate an alterna-
tive approach to such acceptance. The approach utilizes
the input-output model as a starting point and incor-
porates opportunity cost and economic rent concepts in
deriving a set of multipliers which it is believed will
more closely represent the real changes in economic welfare
which result from the operation of "The Multiplier."

The model presented herein is slightly revised from a
version used by Schmid and Ward1 in the Michigan State
University Test Team report on the Water Resource Council's
June 1969 proposed changes in water project evaluation pro-
Cedures. The concepts and approaches are essentially the
Same. The version presented herein is different from the
original in the following respects: 1) The earlier version

USead the coefficients from the 1958 input-output study of

127
the United States, while the newer version uses the 1963
coefficients, 2) the assumptions regarding the treatment
of profits are changed, 3) the approach to estimating
"stemming-from" benefits is changed, and 4) a different
procedure for estimating opportunity cost of labor,
where the local wage rates are lower than the national
wage rates, is used. In addition to demonstrating the
application of the revisions in the earlier model, the
application of the revised model to the Trenton Channel
Project in this Chapter provides a third set of secondary
benefit estimates for use in the comparative analysis of
the Trenton Channel Project in Chapter VI, which evaluates
the sensitivity of secondary benefit concepts to differ-

ences in analytical approaches.
Estimating Wage Changes

Input-output models allow the analyst to trace the
impact of a change in final demand through the economy.
Since "households" is normally one of the sectors included
in the input-output tables, it is possible to estimate the
total effects upon household income caused by these final
demand changes. As discussed in Chapter III, income multi-
pliers derived from input-output studies have generally
presumed that all of the income flows to the household
sector represent real secondary benefits. As discussed
Previously, this convention assumes that the opportunity

COSts of the household sector are zero - that is, that all

128
of the wage, salary and profit earners have no valued a1-
ternatives. It assumes that all returns on capital in
excess of recurrent costs are "pure" profits and ignores
the concept of "normal" returns on investment. It under-
values leisure and assumes that labor earnings are zero
in all possible alternative uses. While this latter
assumption might be justified on some grounds in some under-
deveIOped countries, it is clearly inappropriate for re-

cent conditions in the United States.

The Wage Component of Value Added

 

The input-output model provides the analyst with
snapshot information relating to average relationships in
the economy on a disaggregated basis. It provides much
of the information needed to estimate the 3231 changes in
income which occur as an indirect result of water resource
projects. However, the information in the input-output
tables must be supplemented by additional information.
Part of this information is currently available in second-
ary sources. For example, the household sector (or the
value added sector, in some studies) can be disaggregated
by using data on wages, profits and sales by industry
Which are contained in the "National Income Issues" of the
§grvey of Current Business, published by the Department of
Commerce. ("The National Income Issue" is normally the
July issue.) By dividing the wages and salaries data by

the sales for the particular industry, one can obtain a

129
"wages per dollar of sales" coefficient which can be sub-
stituted in the direct requirements table. The same is
true for profits. Having done this, one could then use
the input-output multipliers discussed in Chapter IV to
estimate a "wage multiplier" or a "profit multiplier."
However, both of these would still deal with the gross con-
cepts of wages and profits discussed in the previous para-
graph. Though both of these can be useful in their own
right, they do not provide appropriate measures of real
wage and profit changes (to even the local area) under
normal United States conditions.

The "wages per dollar of sales" coefficient can
further be disaggregated into an "opportunity cost" com-
ponent and a "rent" component. The rent component could
also be used to generate a "rent multiplier" using the
input-output multipliers discussed in Chapter IV. These
would show the total net changes in income to workers from
a given change in final demand. In order to construct the
"rent coefficients," however, one must have information re-
garding the alternatives available to the workers in each
industry. Unfortunately, there is no generally available
source of such information at present, though a slight
revision in the application forms used by the State Em-
ployment Security Commissions could make such information
available on a continuing basis. By requesting the in-
dustry of previous employment and the wage rate of each

applicant, in addition to the job classification as is now

R

(I)

130

done, these agencies could provide valuable data required
in establishing the alternatives available to the labor
force of each industrial sector. It would also improve
the estimates of unemployment rates by industry groups.
These rates would obviously be even less precise than the
currently estimated rates by job classification. Never-
theless, they could be of substantial aid in determining
which industries could best use an infusion of exogenous
demand. More importantly for analytical purposes, this in-
formation would allow greater precision in dealing with
the problem of supply elasticies, since one could better
forecast the emergence of labor shortages for each industry.

Even in the absence of data sources regarding alter-
native earnings, some logical improvements can be made in
the approaches discussed in preceeding chapters. Though it
is unlikely that one could establish the precise level of
real opportunity costs by logic alone, it is possible to
reduce the range of error apparent in traditional ap-
proaches. For example, it is apparent that zero is too
low an opportunity cost in practically all cases. Any
estimate which is higher and can be logically shown to be
lower than the actual opportunity cost will narrow the
error. Thus, the process becomes one of determining the
highest logically defensible estimate. Since "logically
defensible" involves a subjective judgement, the estimate

should be a conservative one.

131
The most plausible argument relating to alternative labor
earnings looks at the bidding process for labor. As the
demand for the output of each sector shifts, the derived
demand for labor inputs shifts. Under conditions of labor
unemployment, the increased demand for labor can be met
out of the pool of unemployed resources. When the pool is
exhaused, however, firms begin to bid workers away from
other firms to meet continued increases in the demand for
their products. The firms which give up workers do so by
either 1) decreasing their output, or 2) adopting labor-
saving technologies. One might assume that those which
underwent contractions were "marginal" firms which were
earning very low profits (or perhaps were realizing capital
losses). These firms were probably (though notnecessarily)
among the lowest wage firms in the sector. Thus, one would
expect the workers who shifted from the contracting firms
to the eXpanding firms to realize some increment of wage
improvement, though the difference would not be entirely
in the form of real "rent,” since there would be costs
associated with the shift.

In the case of laborers released by labor-saving
technologies, the process is somewhat different. Their
former employers are probably not "marginal" firms. If
they were, the owners would be unwilling to undertake the
capitalization necessary to switch from a high variable-
low fixed cost technology to the high fixed cost technology

suggested by the process of capital substitution. Thus,

132
one should not surmise that these firms are among the lower
wage firms. On the other hand, it is probable that those
workers being released are not the firms' highest salaried
production workers. The process of capital substitution
usually proceeds through the jobs having the lowest skill
requirements. Thus, it is probable that the average wage
of the workers released is lower than the average wage
of all production workers in the firm. Whether it is lower
than the average of all production workers in the industry
is an empirical question. There is reason to believe that
the increased capitalization would occur in the higher
profit firms which also tend to pay higher wages. This
would at least partially offset any tendency of capital to
substitute for lower wage production jobs. In the absence
of an extensive and detailed study of this question, it is
perhaps most defensible to simply assume that the two more
or less offset each other and that the labor released by
capital substitution realized little or no transfer rent,
even when demand for labor is increasing. Thus, the only
rents realized occur through the process of shifting from
marginal to expanding firms.

Wage rates within the "marginal" firms differ from
industry to industry. They normally differ by more than do
the wage rates between marginal and expanding firms within
the same industrial sector. These differences are largely
due to differing capital-labor ratios and skilled-unskilled

labor ratios between the different production processes.

133
While some inter-industry labor transfers can and do occur
under conditions of increasing demand for labor, the dif-
fering labor skill mixes provide some limit to these
transfers. Inter-industry transfers normally occur within
individual job classifications. Wage differences within
the classifications are normally much smaller than the
differences between classifications, especially in union-
scale trades. Thus, the rents realized by a Worker shift-
ing from one industry to another are much smaller than the
difference between the average wages for all production
workers in the two industries would suggest, since the
worker is likely to remain in the same trade.

In the absence of data regarding the shift patterns
of labor, one alternative is to use the average wage from
the sector having the lowest average wages in establishing
the alternative costs of labor. In view of the large inter-
industry wage differentials highlighted above, this con-
vention would appear to greatly understate the real op-
portunity cost of the laborers who change jobs. The skills
of most of the workers will greatly exceed the average
level found in the lowest wage sector. However, this
apparent understatement is ameliorated by the "syphon-like"
process through which the labor market operates: as one
worker moves to higher employment, he is replaced by a
second worker of (marginally) lower skill, who is replaced
by a third whose skill is (marginally) lower than the

second. This process continues until all of the positions

134
are filled or until the limit of the labor supply is
reached. The net effect of the process is to increase the
sum of the earnings of all involved in the "syphon" by the
difference between the new job at the top and the alterna-
tive earnings of the last worker affected. Under condi-
tions of unemployment, the last worker syphoned up would
come from the unemployed; the net results would be the
wages of the new job minus the cost of unemployment. Under
conditions of "full" employment, the alternative at the
bottom of the syphon would be some wage paying job. The
convention adopted here is to assume that the alternative
is employment as an "average" production worker in the
lowest wage industry. This constitutes a very conservative
evaluation of the average opportunity costs, since it
assumes that the syphoning effect will go all the way to
the bottom of the currently employed labor force for every

worker affected by the exogenous impact.
The Wage Rent Coefficient

If one accepts the assumptions outlined above regard-
ing labor opportunity costs, a "wage rent coefficient" for
use in developing a vector of "wage rent multipliers" can
be developed from secondary sources. The data needed are

Ci = total labor compensation in each industry (i)

Li = total employment in industry i

Cs = total labor compensation in lowest average wage

industry, where average wage is defined as
Ci/Li

5X

I I:
«\h t»

ti<

135
Ls

total employment in lowest wage industry

Si

total value of sales in industry i

The "wage rent coefficient" is defined as
Wi (Ci/Li - CS/Ls) Li
Si

Ci - (Li/Ls) Cs
Si

 

The wage rent coefficient can then be used in place of the
household row coefficients in the direct requirements
matrix in generating "wage rent" multipliers in lieu of
the income multipliers. The wage rent multipliers yield
estimates of changes in labor incomes stimulated by the
exogenous change in demand, rather than the Eppgl wages,
salaries and profits paid by the industry as a result of
expansion of output which the traditional "income multi-
pliers" provide.

The reader should be cautioned that the "wage rent
coefficient" is merely an attempt to deal with one of the
problems posed by traditional multiplier analysis - i.e.,
the zero valuing of labor opportunity costs. The estima-
tion procedure outlined in the above formula is based upon
only a gross approximation of the real opportunity costs
involved. Very seldom will the approximation be accurate.
The basis upon which the approximation is accepted is that
the arbitrary assumptions involved in its derivation con-
stitute an improvement over the assumptions of the tradi-

tional multiplier approach. Like the original multiplier

 

5*.

136
presentation, the above derivation is simplistic, and the
mathematical symbols used reveal no great truths in them-

selves.
EstimatingProfit Changes

Business profits are one of the hardest kinds
of income to measure. Any business enterprise
should be able to state accurately the sums that
it has paid out in the course of a year as wages,
interest and rents; it should be able to state
also what income it has paid out to its owners.
But the profits of the enterprise itself are not
definite sums fixed by past transactions. On the
contrary, they are appraisals of net changes in
the position and prospects of the business as a
whole--appraisals that look forward to the un-
certain future as well as back to the irrevocable
past. Like all mixtures of past history and
future anticipations, statements of profits are
necessarily subject to variable margins of un-
certainty.

Wesley C. Mitchell31

The Profit Component of Value Added

In the preceding section, the division of the house-
hold or the value added row of the direct requirement
matrix into a wage component and a profit component was
discussed. The procedure by which the wage component
might be divided into an opportunity cost component and
a rent component was outlined. The use of the rent com-
ponent in devising "wage rent multipliers" as an alternative
to "income multipliers" was discussed.

Much of the discussion regarding the wage component
is applicable to the discussion of the profit component.

At least in concept, one should treat capital in much the

137
the same way as labor. The rents or quasi-rents which
accrue to its employment should be treated as benefits; its
opportunity costs should not be. The principle is, on the
surface, a fairly simple one. In practice, however, the
concept of returns to capital is very difficult to deal
with.

According to orthodox economic theory, quasi-rents
are short-run returns to capital, after deducting variable
costs. Quasi-rents can be divided into payments to cover
fixed costs, and profits. Under pure competition, both
are short-run returns, since in the long run the fixed
costs become variable, and competition will drive all
profits down to the "normal" return on capital. Thus,
secondary benefits from the employment of fixed capital
are theoretically short-run in nature. In the absence of
market distortions which allow idle capacity or excess
profits, no secondary benefits from capital employment
would be possible. One might say that secondary benefits
to capital employment arise out of errors in planning.
This is especially true in the case of excess capacity,
since its existence indicates that either too many plants
or plants of the wrong size were built. Once the investment
is made, it is considered to be irreversible or "sunk," in
the short run. Thus, the surplus of gross receipts above
Variable costs has traditionally been considered to con-

Stitute a net benefit from capacity utilization.

138

In those cases in which plant capacity was being fully
utilized EEQ excess profits were occurring in the industry,
secondary benefits could also accrue. In this case, the
benefits would consist of the surplus of gross receipts
over both variable and fixed costs, since new capital in-
vestment would be required if production were to be ex-
panded. Thus, the "quasi-rents" would consist of only the
”pure profit" component, after deducting the "normal
returns" to capital.

It is possible to estimate the profit per dollar of
sales component in the input-output matrix in the same
manner as the wage component. The National Income Issues

(July issue) of the Survey of Current Business provide data

 

on profits before and after taxes by major industry group.
By dividing the profit data by the sales data, one can ob-
tain an average profit per dollar of sales for the year
in question. There are, of course, problems involved in
using such data in this manner. Some of these are the
same problems which plague the input-output model: The
problems of aggregation. In addition, there are the
problems posed by the accountant's concept of profits
versus the economist's concept.4 Accounting profits are
very difficult to translate into the economic theory
equivalent because of l) the effects of inflation on book
values of capital assets, 2) the difficulty of estimating
"user cost," 3) the difficulty posed by accelerated

depreciation for tax purposes, and 4) the treatment of

139
re-invested profits. All of the tricks available to the
accountant coupled with the competing corporate desires
to show a good profit rate to the stockholders while hold-
ing down tax liabilities make the interpretation of profit
statements a precarious exercise.S Nevertheless, if one
wishes to deal with the returns to capital, he has little
choice but to use the profit data supplied by accountants.
The alternative is to undertake very expensive primary
research (which would be at least as costly as an input-
output study) in an attempt to derive the relevant costs
and state them in the precise concept of the economist.
Some very competent economists have attempted to deal with

this problem in the past.6

The results have largely been
frustration and an equally cloudy "translation" of profit
estimates. The basic reasons for the difficulties ex-
perienced had to do with such things as: 1) Dealing with
yearly "profits" where large year-to-year fluctuations
occurred and the relevant "cycle" was difficult to deter-
mine, 2) separating wage payments from profits in small
firms, 3) valuing the alternative costs of the fixed
assets used where the "economic" costs (especially user
costs) depended upon anticipating future technology and
demand, 4) separating the opportunity cost of capital,
uninsured risk, and pure profit.

Closely tied to determining the level of "profit" is

the problem of defining profit rates and the real Oppor-

tunity cost of capital. Because the fixed cost-variable

140
cost distinction is not nearly so clear in practice as it
is in theory, it is very difficult to determine what con-
stitutes investment and what constitutes recurrent expense.
The simultaneous mix of a building having a life of forty
years, a machine having a life of ten years (ignoring the
rate of use and resultant user cost problems), another
having a life of five years, "variable inputs" taking time
periods ranging from one day to two years to put "on
order" and/or use, and inventory accumulations with ware-
house stays ranging from days to years creates serious
difficulty in operationalizing the fixed cost-variable
cost distinction. Since each of these assets might have
been acquired at different times under different financing
arrangements and "market" rates of interest, it is dif-
ficult to determine the Opportunity cost of the capital
invested in them by evaluating the alternatives at pur-
chase dates. Using the current market value of the plant
as the opportunity cost creates problems also, since the
market value includes such intangibles as company image
and "goodwill" as well as market contacts and purchase
lines for variable inputs. As a result of these factors,
estimates of industry rates of return on invested capital
have been subject to wide margins of error, which has
created great difficulty for analysts attempting to esti-
mate industry profit rates. Thus, while one can readily

be critical of the data on "accounting profits" which are

141
available in secondary sources, one is hard-pressed to find
an alternative which can be effectively demonstrated to be
superior.

If one is willing to accept the profit data found in
secondary sources, he still must decide between profits
before and profits after taxes. Some tax payments are
merely transfers, while others constitute payments to
cover the costs of productive public services which are
consumed by industry (the highway program is a good example
of these). Net profits from the social vieWpoint lie some-
where between the before and after tax profits. One can
not readily say which figure would provide the closer
estimate of profits net of both transfers and real cost of
public services. However, since neither of the profit
figures have been adjusted for risk and "normal” returns
on investment, it is apparent that the profits before tax
estimate greatly overstates the real net profits. At the
same time, the profits after taxes estimate is also too
high, unless the transfer component of the tax payment were
equal to or greater than the level of "normal" returns
expected in the industry. The actual relationship is not
currently known nor easy to determine. It is improbable
that the after-tax profits ". . . are in the form of a
rent which could be taxed away and still allow . . . them
to supply their services."7 Thus, they do not constitute
real benefits to the owners from expanding output. Never-

theless, the after-tax profits per dollar of sales is used

her

35

prc

her

311

cf

142
herein in estimating the changes in profits which occur
as a result of secondary expansions in output. Of the
choices available, these data appear to be most in keeping
with the previously stated principle of attempting to arrive
at the closest logically defensible estimate in reducing
the margin between gross and net changes in income via
secondary effects. It is the belief of the author that
after-tax profits over state'the rents realized by pro-
ducers. Much additional study of specific enterprises is
required in order to answer questions regarding "excess
profits" and scale economies. The convention adopted
herein can only be defended on the grounds that it provides
a more reasonable estimate than does the gross application

of input-output multipliers.

Rounds of Expansion and Benefit Creditability

 

In the immediately preceding sections, it was stated
that the net wage coefficients and the profit coefficients
could be used in lieu of the household direct requirements
coefficients in generating input-output multipliers. Such
multipliers provide valuable information in dealing with
some problems. However, they are, for the most part, in-
appropriate for use in evaluating the secondary benefits

from water resource development.

iv

 

143

Associated Investments

 

Most of the income multipliers generated from input-
output models include both the direct and indirect effects
of final demand changes. That is, they include not only
the first round of transactions but also the infinite
series of rounds which is supposedly generated (see
Chapter IV). Unless unused capacity sufficient to meet
all of these outputs exists, additional capital expendi-
tures will be necessary. To give all of the credit for all
ensuing income changes to the initial investment does great
violence to all of the associated investments which are
required to produce the induced outputs. It, in effect,
assumes that all associated investments are riskless and
that the supply of capital required to undertake them is
perfectly elastic. The erroneous nature of such a con-
vention is illustrated by the following example: The Corps
of Engineers is dredging a channel which will allow a
steel mill to expand its output of steel ingots. The
problem is to estimate the income which will be generated
as a result of increased purchases of inputs by the steel
mill. The analyst used the traditional approach of using
a direct-indirect income multiplier which includes the
direct and indirect transactions of, among other things,
large volumes of limestone. All of the income changes
would be included as benefits under the traditional ap-

proach. However, in order to provide for the greatly

144
increased purchases of limestone additional investments in
quarrying equipment will be required. The credit for these
investments are included in the secondary benefits from
the investment in dredging. To illustrate the point fur-
ther, assume that a second Corps of Engineers navigation
project is required later to provide access to the lime-
stone quarries necessitated by the increased steel output.
The project benefits claimed for the second navigation
project will include a sub-set of those claimed for the
first project. Thus, not only is the credit for the effects
of ensuing investments claimed by the first project, some
of the effects are also double-counted in the Federal navi-

gation program account or other development programs.

Inelasticities of Supply

The input-output multipliers project the current
composition of input purchases and assume that the demands
imposed on the input suppliers can be met at current prices.
That is, projections using the model assume that the supply
of all factors is perfectly elastic at prevailing prices.
Obviously, this is not the case, especially for relatively
large changes in demand such as those caused by large-scale
water projects. These supply inelasticities restrict
the final level of the real multiplier effects. It is
difficult to estimate the exact extent of these restric-

tions, because the relevant supply elasticities are not

known.

145

Some analysts have introduced restraints into the
model to adjust for these "leakages.” Isard and Kuenne8
dealt with the problem by excluding important sectors from
the structural matrix, thus reducing the level of the in-
direct effects which were calculated when the matrix was
inverted. A principal sector which is often excluded is
the household sector, thus eliminating the ”household
induced" component (which is the principal component of
the Keynesian Multiplier, see Chapter IV). A second
method of restraining the multiplier has been to truncate
the multiplier calculation after an arbitrarily selected

number of rounds.
Adjusting the Multiplier

For the most part, "multipliers" provide an estimate
of the absolute upper limit to the income changes which
mighp occur as a result of a given exogenous change. They
assume away the problems of friction, inflexibility, and
"improper" reactions to stimuli that are present in any
economy. Since they are designed for predicting rather
than assigning cause, their use in investment analysis
tends to lead to assigning the credit for all changes to
the initial "exogenous" change occasioned by the invest-
ment.

The problems of "credit” assignment, leakages, and
ancillary Government actions millitate against using the

fUIl value of the direct and indirect effects generated

146
by most multiplier approaches. With each ensuing round of
transactions included in the multiplier, these problems
become more intense, and the justification for truncating
the analysis becomes stronger. Even in the first round of
induced transactions there are likely to be capital costs
and risk undertakings which should receive part of the
credit.

In the absence of knowledge regarding the levels of
leakages, credit, etc., any adjustment of the indirect
effects is necessarily arbitrary. It is the author's be-
lief that in few cases should more than the first round of
transactions be included in the computation. The arbi-
trariness of this judgement is realized, as is the justi-
fication for sometimes carrying the computation a few
rounds further. However, in light of the previously-
discussed problems, it would appear that one is very sel-
dom (if ever) justified in computing the rounds through
to the full "multiplier" value. According to Clark, Grant,
and Kalso, "The stemming-from hypothesis, crediting pro-
duction of raw products as a 'trigger' and causing the chain
of subsequent processes, has limited validity which does not
warrant carrying the computation through to the ultimate
consumer in all cases" (p. 26). The convention used herein
is to use only the first-round direct requirements in com-

puting the wage and profit gains.

147

Regional Coefficients and Purchases

One of the major drawbacks to using input-output
analysis in water resources research has been the lack of
regional input-output studies in other than large metro-
politan areas. In most cases, regional input-output studies
are undertaken largely to provide data from which regional
multipliers might be derived. The ever-increasing desire
for sources from which such multipliers might be derived
has led increasingly to the application of multipliers
and/or study data from other areas deemed to be "comparable"
to the region being analyzed. Input-output data from other
regions having a similar industrial composition have been
widely used in deriving local income multipliers because
of the belief that "comparable" industry composition in-
dicates comparable "leakages" from the local cycle of re-
spending effects. The importance of "leakages" to the
level-and, indeed, to the concept-of the multiplier has
been discussed in previous chapters. The preceding sections
have attempted to show the inapplicability of the "pure"
multiplier concept to the evaluation of secondary effects
of water resource projects. Tracing the inter-regional
trade effects beyond the first round becomes increasingly
difficult; and the assignment of "cause" for gains reali-
zed from extra-regional trade (especially in regions al-
ready enjoying "full employment") becomes increasingly

difficult to justify. Limiting the analysis to one round

148
of transactions largely eliminates the need to measure
round-after-round of leakages from re-spending and, thus,
greatly reduces the advantages of the regional input-
output study over the national one in estimating the local
impacts. In fact, the national input-output model can be
substituted for the regional model, if the analyst is suc-
cessful in dealing with two issues: 1) Rationalizing the
differences between direct requirements coefficients in
the two models, and 2) separating "local" purchases from
”national" purchases so as to determine the extent of the
impact upon local unemployed factors in cases in which the

local unemployment rate exceeds the national rate.

Differing Requirements

 

Actually, the issue of rationalizing differing re-
quirements is not nearly so problematic when the rounds
of transactions are limited, since each iteration would
have compounded any original difference in the coeffi-
cients, especially when "leakages" (intra- as well as
inter-regional) are considered. Since the direct require-

ments coefficients are based largely on technical rela-

 

tionships between inputs and outputs, they do not vary a
great deal from one plant to the other; and in long run
equilibrium, they should be essentially the same. The
greatest difference in direct requirements coefficients
grow out of differing systems of aggregation, rather than

differing production processes. The greatest difference

149

in indirect requirements coefficients grow out of differing

 

specifications of "leakages" between the models.

Localizing Purchases

 

There are basically two reasons for wishing to know
the proportion of purchases which are local: 1) For
estimating the real income effects where the local unem-
ployment rate differs from the national, and 2) for deter-
mining the inter-regional distribution effect of the pro-
ject. Regional input-output studies provide this informa-
tion directly. Additional data are needed in separating
out the regional effects when using a national study.

The secondary outputs to meet the new demands
"caused" by the project might be supplied under any of the
following three circumstances: 1) From firms (inside of
our outside of the region) having excess plant capacity,
2) from marginal outputs from all firms (nationwide) who
are operating at or near capacity, and 3) from new plants
built specifically to supply the new demands. Under the
first and third circumstances, one would expect transport
economies to play a large role in localizing the effects
within the region. Under the third, the location of new
plants would depend upon the plants' "orientation."9
However, as discussed previously, the "credit" for gains
realized from wholly new investments should not be given

to the water project; though new plant investments are of

great importance in designing programs and strategies for

150
promoting the development of depressed regions. The Corps
of Engineers is sponsoring research directed towards as-
sessing the plant location effects of water projects.10
However, until the direct link is proven and the "credit”
assignment problem is solved (i.e., the "marginal product"
of the water project in the location function is deter-
mined), the addition of plant location effects as benefits
in project appraisal only serve to add greater uncertainty
and guesswork to the analysis and to intensify the diffi-
culty of separating the net national gains from the
transfers. The approach used herein is to assume that
any "new" outputs which result from the impact of the

11 will be distributed

project on "national" industries
nationally in the same proportions as are current outputs.
This amounts to an assumption that the national industries
affected by the "stemming-from" impacts are not heavily
input-oriented and that those affected by the "induced-

by" impacts are not "market-oriented." (However, this

does not imply that they are market-oriented and input-
oriented, respectively.) It is also assumed that the
"credit" due the project from ensuing rounds of transactions
is approximately equivalent to that unduly assigned to the
project in the first round of transactions. As previously
discussed, this is very difficult to prove or disprove and
constitutes an arbitrary assumption.

The situation in which project-induced requirements

are met by marginally increasing the output of all plants

151

is more straight-forward though perhaps less realistic
than the other cases. In this case, it is almost defini-
tional to assume that the inter-regional effects are dis-
tributed in the same way as are present outputs. While
it is unlikely that any plant would order one unit of the
same item from every supplier of that item, under some
circumstances the aggregate effect might approach this
situation. The shifting and realigning of purchases which
might occur when a large change in demand occurred in one
locality could sift through much of the affected industry.
It is unlikely, however, that the effects would be in
exactly the same inter-regional proportions as the current
capacity in most industries, largely because of transport
economies. Nevertheless, the possibility of such an
occurrence lends added credence to the convention of
allocating inter-regional effects on the basis of current
proportions.

In those cases in which regional excess capacity
existed in an affected industry, one would be justified
in assuming that all of the induced output which could be
supplied locally would come from local firms. In most
cases, the existence of excess capacity should be known
by the local offices of the water resource agencies. In
many cases, the projects will have grown out of local
support generated by the existence of this excess capacity.
Especially in the non-metropolitan localities, the area

offices should possess the knowledge of the local economy

 

Ana

wt.»

1‘)
(D

152
sufficiently well to ascertain the local existence of ex-
cess capacity in the industries to be affected by the water
project. In these cases, the analyst should not allocate
the impacts according to current inter-area proportions,
at least until the excess capacity of the local firms has
been absorbed. This would be true even in those cases
in which excess capacity existed in other regions, since
one would expect regional firms having excess capacity to
be able to under-bid outside firms on the basis of trans-
port costs and established market contacts.

Of the two reasons, stated at the beginning of this
section, for wishing to know the regional impact, only one
has any impact on the "net benefits" of the project. Only
in those cases in which regional labor and/or plant
capacity are under-utilized relative to the nation is
the total output affected. Where the utilization rates
do not differ greatly from the national rates, such in-
formation is largely secondary or irrelevant. Where local
capacity is already fully utilized, local increases in
output will require increased capital, which should re-
ceive a part of the credit. Thus, the convention of
proportioning the regional share according to existing
proportions probably does little violence to the concept
of "net" local project effects, under conditions of local
"full employment." Where local conditions are such that
the outputs can be met without additional investments, the

“St local benefit concept conforms closely to the national

153
concept, and the localization issue is more important.
Thus, the analyst should use his knowledge concerning local
resource flows to construct the "proper" 10calization pro-
portions for the industries affected. (He should also
utilize his knowledge regarding local employment Oppor-
tunities in attempting to construct a locally-relevant
set of opportunity cost figures for local labor in improv-
ing the wage rent coefficients discussed previously.)
There is presently no widely available, reliable source of
secondary data which would allow the analyst to readily
pinpoint the existence of excess capacity, though there has
been much research directed towards defining and determin-
ing the existence of excess capacity.12 Unfortunately,
the concepts of "capacity" and "capital investment," (and
"profits") are components of capital theory, ". . . the

. . 13
most e1u51ve concept of economics."

It is perhaps
axiomatic that the more "elusive" the concept, the dearer
are the sources of data for dealing with it. The cases of
profits and capacity lend credence to such an axiom.
National and Local Industries. Leontief, in his

14 divided the industries

study of the impact of an arms cut,
into "local" and "National" categories on the basis of the
proportion of sales going to local markets. Those indus-
tries, such as personal services and wholesale and retail

trade, which service basically local markets were classi-

fied as "Local." The industries, such as primary metals

154
and other manufactured goods, which went primarily to
"National" markets were classified as "National" industries.

While there are some difficulties posed by the cate-
gorizing of "Local" and "National" industries, this dif-
ferentiation is helpful in the attempt to separate out the
impacts which are primarily local and those which are
national in getting at the problem of estimating inter-
regional trade flows posed by the absence of a regional
input-output study. As discussed in previous paragraphs,
the analyst's personal knowledge of the local trade flow
structure can be used to supplement this differentiation
where there are ”National" industries which are servicing
primarily local demands (such as a manufacturing concern
which produces inputs which are wholly-used by another
”National" firm in the local area).

Those industries which are classified as Local
seldom sell in National markets. For the most part, their
impacts can generally be considered to be wholly local.
The reverse is not always true of the National industries,
however, as alluded to in the preceding paragraph. While
the National industries are not normally restrained by the
location economies which effect the local industries, there
are many examples of nationally-oriented firms whose out-
puts go largely to the local area. In addition to input
suppliers, there are also nationally-oriented firms sell-
ing to "final demand" which sell largely within the local

area. However, because of the scale economies involved

155
in the production and marketing of the products produced
by such firms, the case of such National firms selling to
local markets is most common in local areas which include

large metropolitan areas.

Localization Ratio. The approach used in the model is to

 

use "Localization Ratios" to estimate the proportion of the
national impacts which will occur locally. In the case of
the "Local" Industries (as defined by Leontief), the
Localization Ratio is assumed to be unity (1.0). That is,
all of the sales which result from the secondary impacts
on those industries are expected to occur locally. For
those industries which are classified as "National," the
Localization Ratio is based upon the proportion of total
national employment in the industry located in the region.
For example, if one-fourth of all employment in auto manu-
facturing in the United States were located in the state
of Michigan, the Localization Ratio for that industry for
the state (defined as the ”region") would be 0.25. Thus,
the model would assume that 25 per cent of all increases in
auto sales would be supplied by production from the state
of Michigan.

Employment proportions obviously are not identical
to capacity proportions, though they are very closelyre-
lated. The convention of using employment as a basis for
estimating localization is based upon two considerations:

1) The general availability of data-relating to employment

156
at sub-national levels in light of the lack of similar
sources relating to other parameters of output capacity for
sub-national areas, and 2) the fact that the major objec-
tive of the model is to estimate wage changes -- thus, the
employment sensitivity of the local industry is more impor-
tant than its output capacity.

Some might find the Localization Ratio an overly-
simplistic tool for estimating the local proportion of the
total impacts.' It admittedly extrapolates the average
localization of employment to the marginal case, and eco-
nomists are constantly cautioning against such procedures.
It is not alone in this regard, however, since many applied
economics techniques (including input-output models - see
Chapter IV) have the same faults. In most cases, such
extrapolations of average relationships are used because
information regarding marginal relationships is impossible
(or expensive -- i.e., uneconomic) to get. Thus, while
economic theory generally deals with relationship at the
margin, economic practice often deals in the application
of data regarding averages. The pervasineness of this
practice among economists does not justify the use of
averages in marginal analysis; however, it does indicate
the difficulty economists have experienced in developing

data sources amenable to marginal analysis.

157

Applying the Model

 

The preceding sections of this Chapter have dealt with
the conceptual basis of the proposed model for estimating
secondary benefits. Following sections of this chapter
deal with the application of the model to the Detroit
River, Trenton Channel Project, of the Army Corps of
Engineers. The application serves two functions: First,
it serves to demonstrate the procedures by which the model
can be applied, and second, it generates an additional set
of numbers to be used in Chapter VI in the comparison of
alternative estimates of the secondary benefits of the

Trenton Channel Project.

The Trenton Channel Prpject

 

The proposed Trenton Channel Project of the Army
Corps of Engineers is a navigation project designed to
provide access to deepwater transport for an extended
portion of the Detroit River navigation channel. While
there will be some side effects on recreation and the en-
vironment, the basic economic impacts come from the pro-

vision of lower cost transport.15

The principal user and
primary beneficiary of the extended channel will be the
McLouth Steel Company, which has indicated that it will
expand its steel production, if the channel is extended.

Ostensibly, an additional 3000 workers would be employed

by the firm, and output of steel in the Detroit area would

158
expand by $40.5 million per year. The principal secondary
impacts of the project would come via the increased pur-
chases of inputs by the steel plant (induced-by), the in-
creased purchases of inputs by users of the steel (stemming-
from), and the purchases of inputs for the construction of
the navigation channel.

The Detroit metropolitan area - the principal "region"
to be affected by the project is not an area of "chronic"
unemployment. Rather, the employment problems of the area
are cyclical because of the heavy dependence of the economy

16 Thus, the local re-

upon durable goods manufacturing.
sources which would be stimulated to employment by the
navigation project should not be viewed as long-term un-
employed resources, and the opportunity costs of those

resources should be deducted from the secondary income

generated, as discussed in preceding sections.

The Computation Procedure

 

Induced-byrBenefits

 

Induced-by benefits are those which result from the
increased purchases of inputs by local industries as a
result of the provision of some water service - in this case
navigation. The estimated increase in direct purchases
of inputs by the McLouth Steel Company is shown in column 3
of Appendix Table l. The estimate is derived by multiply-

ing the $40.5 million of increased steel sales by the

159
Direct Requirements Coefficients in column 2. The Direct
Requirements Coefficients show the proportion of each
dollar of sales of Primary Metals going for the purchase
of inputs from each sector. For example, for every dollar
of sales of Primary Metals, $0-02994 goes for the purchase
of Electricity, Gas and Water Services. Thus, $40.5
million in steel sales results in $1,213,000 in purchases
of Electricity, Gas and Water. The other entries in
column 3 are similarly derived (i.e., $40.5 million x
.00007 = $3,000;$40.5 million x .04657 = $1,886,000; etc.).

The secondary sales shown in column 3 indicate the
first round of "activity" generated by the increase in
steel sales (see Chapter IV). This does not represent
net benefits to anyone, since these are merely payments
for productive services which also have costs associated
with their production.

The proportion of induced sales which go for the
purchase of labor by each affected industry can be esti-
mated by multiplying the "Labor Coefficients" in column 4
by the respective induced sales in column 3. The Labor
Coefficients show the proportion of each sales dollar (for
the industries arrayed on the left of the table) which goes
to purchase labor inputs (i.e., 42.24% of gross receipts
by Electricity, Gas and Water producers are paid out in
wages and salaries. Thus, $1,213,000 in sales results in
$561,000 in payments to labor). The total wages and

salaries generated can be obtained by summing column 5.

160
Again, column 5 does not represent the first round induced
benefits to labor, unless all of the laborers employed had
absolutely no other productive alternative use. As dis-
cussed previously, this is the estimate used for the labor
component in traditional income multiplier approaches.

The number of workers effected can be estimated by
dividing the induced wages in each industry by the average
production worker wage in that industry shown in column 6.
Thus, for Electricity, Gas and Water, an addition of 74
workers (column 7) will be required as a result of the
increased demand posed by the expansion of the steel plant.
These represent "new" jobs only if they will be supplied
by previously unemployed resources. Under "full employment"
conditions, they must be filled by workers from another
industry. In the model, it is assumed that they will be
filled by workers previously employed in the Service in-
dustry. Thus, the "net" wages received by each new worker
in the Electricity, Gas and Water industry will not be
$7,500 as shown in column 6, but, rather, it will be
$7,500 minus the $4,765 which was being earned in the
Service industry; the difference ($2,735) is shown in
column 8. Similarly, the difference between $4,765 and
$6,180 (i.e., $1,415) would represent the increase in wages
for those who shift from Services to the Printing and
Publishing industry. By multiplying the "Wage Rents" in
column 8 by the "Workers Affected" shown in column 7, one

can estimate the total change in wages resulting from the

161
induced effects of the expansion in steel production. This
total for each industry is shown in column 9.

The reader might note that the entry for the Service
industry in column 9 is zero. The analysis says that there
are no induced effects upon the Services industry. However,
column 7 shows an increase of 58 in the demand for Service
workers. While the model assumes that 1,021 workers are
being drawn away from the Service industry and into new
jobs. This should result in a deficit of 1,079 Service
workers relative to the demand for outputs from the
Services sector. The degree to which this represents an
induced disbenefit depends upon the elasticity of supply
and demand for services as compared to that for the other
outputs produced by the workers attracted away from the
Services industry.

Under conditions other than total unemployment, there
will be some cost associated with stimulating factors to
"higher" employment. The convention of subtracting the
earnings in the Service sector from those in the "higher"
employment makes certain assumptions about the utility
realized from the production and consumption of services.
The simplest means of dealing with the welfare changes
posed is to assume that the market for services is com-
petitive, such that the marginal utility from services
consumption is equal to its price. Assuming that the total
utility was captured by service workers in the form of pay-

ments for services, the wage rent estimated in the model

162
would represent the real welfare gains involved in the shift
of the marginal worker. The validity of extending this
"marginal" relationship to all 1,079 workers affected de-
pends upon the slope of the aggregate marginal utility
function for services consumption relative to the marginal
productivity of the 1,079 service workers. The greater
the slope of the function in this range relative to the
marginal product of service labor, the greater will be
the discrepancy between real negative changes in welfare
and the opportunity cost accounting embodied in the price
of service labor. The model assumes that both are hori-
zontal and equal and that the wages paid to service workers
represents the total utility realized in the production
and consumption of services. The net wage gains shown in
column 9, then, represent "real" changes in welfare, under
these assumptions.

The net wage gains shown in column 9 are the esti-
mated gains for the nation as a whole. To estimate the
proportion which accrues to the Detroit area, the "Locali-
zation Ratios" in column 10 are multiplied by the respective
figures in column 9. (The derivation of the Localization
Ratios is discussed in the preceding section of this

17 The Localization Ratio for the Local Indus-

chapter).
tries is set a 1.0, as discussed above. Thus, that

portion of column 11 is identical to column 9. However,
the local proportion of the induced-by effects upon the

National Industries is a much smaller proportion of the

163
total effects -- an average of 2.84 per cent in this case.
As a result, the total "local" effects in terms of wage
changes is estimated to be about 37 per cent of the total
"national" wage effects. Whether this proportion would
be estimated from a regional input-output model of the
Detroit area can be determined only when such a model be-
comes available.

The wage gains estimated in Appendix Table l were
estimated without the use of the wage rent coefficient dis-
cussed above. This was done in order to show the process
through which the wage rent coefficient operates. Using
the wage rent coefficient, columns 4, 5, 6, 7 and 8 could
be replaced by one column (or one "vector," if matrix
procedures are used. See Schmid and Ward, 0p cit, Ap-
pendix). The wage rent coefficient (for Electricity, Gas
and Water, for example) can be computed by dividing
column 8 by column 6 and multiplying the quotient by
column 4 (e.g., 2735/7500 x .4624 = .1686). Multiplying
column 3 by this wage rent coefficient yields column 9.
The wage rent coefficient is used throughout the remainder
of this study rather than the longer procedure used in
Appendix Table 1.

The computation of profit gains is similar to that
of wage gains, except that the "Profit Rents" are obtained
directly by assuming them to be approximated by "profits
after taxes." The induced sales in column 1 of Appendix

Table 2 are multiplied by the average after-tax profits

164

per dollar of sales coefficients of column 2. The result
is column 4 showing the profit gains induced by the steel
expansion. The Detroit portion is then estimated by
applying the Localization Ratios of column 4 to the Profit
Gains. Thus, it is estimated that $208,530 in Profit Gains
accrue in the Detroit area of the $951,061 which accrue to
the nation as a whole.

The sum of the Wage Gains and Profit Gains induced
by the steel Output for the Detroit Region is estimated
at $1,204,728, while that for the nation as a whole is

estimated to be $3,680,783.

Stemminngrom Benefits.

 

Stemming-from benefits are those increases in wages
and profits which result from increased availability of
inputs into other production processes. The readiest
example is that of the increased supply of agricultural
products to food processors as a result of an irrigation
project. Stemming-from benefits have not been as widely
treated in the literature as have induced-by benefits.
This is because (as discussed in Chapter II) stemming-from

18 while traditional

benefits are basically supply effects,
multiplier analysis has dealt with demand effects.

At the national level when wide-scale unemployment
does not exist, a water resource project should produce

stemming-from (supply) effects to help the economy to

absorb the induced-by (demand) effects without simply

165

creating inflation. Ideally, the supply and demand effects
would be commensurate, so that the "multiplier" indicated
real income changes. As argued by Clark, pp_gl, one should
not add together supply and demand in estimating national
secondary benefits, since the two act jointly in deter-
mining equilibrium output. However, at the regional level,
the two effects can be additive, since the increased out-
puts outside Of the region might have grown out of the
inter-regional "leakages" of the induced-by (demand) ef-
fects of the project. However, these "leakages" are not
included as regional induced-by imports. Thus, the
stemming-from (supply) effects allow some of the inter-
regional effects of the project to be recaptured by the
project region. As a result, stemming-from benefits are
only regional in nature and should not be included as
"net" benefits from the national standpoint (i.e., though
they are "real" benefits to the project region, from the
national standpoint they are simply transfers which cancel
out somewhere else). Even from the viewpoint of the region,
the stemming-from effects are net for only those increased
output which are exported from the region, since those sold
internally largely go to satisfy the increased intra-
regional demand created by the "induced-by" effects.

Stemming-from effects constitute special problems for
the analyst, since their validity depends upon the im-

portance of the water project-related factor in further

166
production processes. Of great importance are the elastici-
ties of substitution of the relevant factors and their
elasticities of supply with and without the water project.
As discussed in Chapter II, a serious drawback to most
analyses of stemming-from impacts has been the tendency
to assume the perfect elasticity of all factors, besides
the water project-related factor. It was pointed out in
Chapter II that such an assumption creates an inverse
relationship between the stemming output and the propor-
tion of production costs represented by the water project-
related factor.

The problems of "importance" of factors are dealt with
in the model by adopting a convention assuming that the
proportion of the sales dollar going to each factor indi-
cates its relative importance in the production process.
Thus, a factor whose cost represents 20 per cent of the
sales dollar of some product is twice as "important" in
the production process as a factor which absorbs 10 per
cent. These percentage weights are used as a rough means
of adjusting the stemming outputs to account for inelas-
ticities of other factors and for the tendency to assign
credit in an inverse fashion. The approach used in the
model does have some basis in economic theory, since
normally the "credit" assigned to each input should be re-
lated to its value in producing the final output. If the
input prices reflect their marginal value products (as they

should under perfectly functioning input markets), then

167
the average outlay for water project-related inputs per
dollar of sales is a good approximation of the credit which
should be given to the water project for the stemming-from
output.

In Appendix Table 3, the outputs of steel are dis-
tributed among the steel using industries in the propor-
tions indicated by the 1963 National Input-Output Study.
The distribution coefficients, are arrayed in column 1.
Column 1 indicates that Primary Metals production receives
the greatest share of primary steel output - i.e., approxi-
mately 20 per cent. Multiplying these coefficients by the
$40.5 million increase in steel output yields column 2
showing the expected distribution by industrial purchaser
of the project-related steel output. Primary Metals, for
example, will purchase $8,279,820 of the output (i.e.,
$50.5 million x .204). The value of purchases of steel
by each sector is taken to represent the portion of stem-
ming effects which can be credited to the increased availa-
bility of steel, as a result of the weighting scheme dis-
cussed in the preceding paragraph. Since primary steel
inputs absorb 20.4 per cent of each dollar of sales of
the primary metals industry, the stemming output under the
traditional assumption regarding the elasticity of other
factors would be estimated at $39,730,000 (i.e., $8,279,820/
.20840). The convention of crediting the stemming outputs
in proportion to the input costs results in multiplying

the total output of $39,730,000 by the relative weight of

168

.20840, which gives the original steel sales figure of
$8,279,820. The same is true for each sector affected.
Thus, the sales of steel to each industry shown in column
2 represent also the portion of stemming output creditable
to the increased steel production.

The Profits and Wage Rents are computed in the
same manner for the stemming-from benefits as for the
induced-by benefits, with one exception. Since only the
exports of the increased outputs create net gains to the
region, the local sales of steel must be isolated and ex-
cluded from the computations. Locally-contained demand
effects must be spent upon locally-produced goods in order
to generate the wage and profits estimated in the induced-
by effects. Thus, the local stemming-from effects are
simply the counter part of the induced-by effects.
Stemming-from production is available for export, only
if the local stemming-from effects exceed the local
induced-by effects for a particular sector. For example,
the induced-by effects upon the Printing and Publishing
sector (Appendix Table 1) increases the demand for Printing
and Publishing by $3,000. (All of these demands are as-
sumed to be local). For this demand increase to affect
wages and profits in that sector, an output response must
be made by the sector. Unless the supply of Printing
and Publishing is perfectly elastic, the project-related
demand increase can not be given full credit for the

response, in the absence of a commensurate project-related

169
supply effect. Column 2 of Appendix Table 3 indicates that
the stemming-from (supply) effect on Printing and Publishing
is $3,240. Thus, $240 in "eXportable surplus" Printing and
Publishing services are created. Only the Profit Rents and
Wage Rents which accrue to this $240 in Printing and Pub-
lishing "exports" is included as local stemming-from bene-
fits. The derivation is straight-forward for the "Local
Industries," since the Localization Ratio is 1.0. For the
National Industries, the surplus (deficit) from the na-
tional vieWpoint must be determined first. Then any sur-
plus production is apportioned according to the Localiza-
tion Ratios. For example, the National "surplus" of non-
durable goods amounts to $453,190 (i.e., $809,190 minus
$376,000). Since 0.12 per cent of that industries employ-
ment (capacity) is located in Detroit, $544 of the expor-
table surplus is estimated to accrue to the Detroit Region.
The Wage and Profit Rents on this $544 of "exports" are
included as Detroit Region stemming-from benefits. The
same procedure is followed in calculating the stemming-
from benefits for the other industrial sectors. Those for
which the induced-by effects exceed the stemming-from
effects are assumed to produce no "surplus." Thus, there
is no stemming-from benefit computed for them. The total
of Wage and Profit Rents due to stemming-from effects is
$368,383.

Construction Impacts: Induced-By. In addition to

 

the secondary impacts from project-related output expansions

170
there are also secondary impacts which occur as a result
of project construction expenditures. The construction
effects are basically induced-by effects or aggregate
demand effects. The generation of secondary income via
construction expenditures represents the point at which
water resource develOpment policy comes most directly into
line with the policy of aggregate demand maintenance voiced
in the Employment Act of 1946. While the induced-by and
stemming-from effects from project outputs are to some
extent mutually self-effectuating, the induced-by effects
from construction expenditures depend wholly upon existent
flexibility (i.e., supply elasticity) in the economy to
make their multiplier effects "real." In the absence of
such flexibility, the "balanced budget multiplier" effects
of the tax-expenditure sequence is lost on inflation, and
little or no real income changes result. The induced-by
effects are largely inter-regional changes in incidence
of income, since (from the project-by-project vieWpoint
of local interests - see Schmid and Ward, p. 38-41) the
project region receives a large part of the construction
impact while paying a small part of the cost. Part of the
inter-regional income re-distribution could represent real
gains, if the project region were experiencing under-
employment conditions while the rest of the nation was not.
However, such conditions did not prevail in the Detroit
area during the period relevant to the analysis presented

herein. Thus, in this analysis the construction impacts

171
are treated as inter-regional income transfers, and in-
clusion is made only in the regional account.

The estimation of the secondary impacts induced-by
the construction of the Trenton Channel Project follows
essentially the same lines as the procedure used in esti-
mating the changes induced by the steel output, with two
exceptions: l) The vector of direct requirements is

19 rather than

taken from the Haveman and Krutilla study
the 1963 Input-Output Study, and 2) only the local portion
of the induced-by impacts are counted. The direct require-
ments vector in this case shows the distribution of each
dollar of Federal expenditure for dredging. For example,
10.3 percent of the costs spent in the Non-durable goods
industry. Multiplying this vector by the construction cost
of the project ($31,400,000) gives the estimated impact

by industrial sector. Following the procedures outlined
previously, the Detroit Region impact from the construction
expenditures is estimated in Appendix Tables 5 and 6 to be

$309,310 in Wage Gains and $174,611 in Profit Gains,

respectively.
TrentopyPrpject Secondary Benefits Summarized

The secondary benefits of the Trenton Channel Navi-
gation Project have been estimated in the above sections
from two viewPoints: 1) That of the Detroit Region and 2)
that of the nation as a whole. The estimation procedure

attempted to find means of estimating the real net changes

172

in income which accrued at both levels as a result of the
project. The defects of simplifying the various reactions
and interactions into such a two part framework have been

0

widely discussed.2 Nevertheless, since this convention

was the format suggested in the Water Resource Council Task
Force Report, an attempt was made herein to work within
such a framework.

The secondary benefits from the Detroithegion view-
point consist of $996,198 in Wage Gains induced-by the steel
mill expansion, $208,530 in profit gains stemming-from the
steel mill expansion, $309,310 in Wage Gains induced-by
project construction expenditures, and $174,611 in Profit
Gains induced-by project construction expenditures. The
sum of these effects is $2,055,584.

From the National viewpoint, the secondary benefits
consist of only the Wage and Profit Gains induced-by the
increased steel output. The estimated National Wage Gains

total $2,729,672, while the Profit Gains total $951,061.
The sum of the two is $3,680,733. However, since there
were no unemployed resources in the Detroit area, there

were probably no National indirect benefits at all, after

accounting for declines in other sectors.

The Underlying_Assumptions of the Model

 

Within the scope of the limits defined,
The results have been quite conclusive:
That assuming the assumptions
That have been assumed 21
The problem becomes unobtrusive.

173
The distinguished planner and development economist,
Albert Waterston, has noted that economic planners often
provide plans of little practical value because of the re-

strictive assumptions which are made. He points out that

Because analytical complexity increases with

the number of variables used, planners include

as few variables as possible in their closed
system models and rely on ceteris paribus or
mutatis mutandis to maintain control over the ex—
cluded variabIes. This approach necessarily con-
stricts the limits of possible answers. By posing
questions which are shielded from outside distur-
bance or uncertainty, answers are bound to be
simplistic; by narrowing the scope of problems,
solutions are sure to be circumscribed.2

 

 

The essence of model building is simplification. The

objective is the optimization of the competing desires for

conceptual manageability and prescriptive relevance to a
complex world. This "two-horned dilemma" pervades prac-
tically all of economic analysis, and, unfortunately, its
dangers are not always recognized and appreciated by
economic practitioners. The proliferation of econometric
and mathematical models which purported to "simplify
reality" and/or make economic analysis more "scientific"
during the 1960's bears ample witness to the ever present
danger of "losing sight of reality in a maze or pretentions
and unhelpful symbols."

The assumptions which surround any analysis often
constitute the key to understanding the problem which is
being analyzed. As pointed out above, simplifying assump-
tions are a necessary part of model building, if the

objective of the model is to make understandable the

174
fundamental relationships involved. Yet, the assumptions
themselves constitute real variables (exogenously defined)
whose relationship to the endogenous variables must also
be understood, if the analyst is to draw the full measure
of the heuristic value of his analysis. Sometimes, (as
was suggested in the quotation introducing this section) it
turns out that the assumptions a priori specify the most
important variables in the analysis, and the results be-
come foregone conclusions, thus destroying much of the
value of the "analysis."

The objective of this section is to attempt to deal
with the problems posed by the assumptions made in the
analysis in preceding sections. The explicit assumptions
of the model, most of which were discussed in the text,
will be outlined. An attempt will be made to identify the
implicit assumptions and to spell them out. The underlying
assumptions (as best they can be delineated by the author)
will be presented so that the reader might make his own
analysis of their effects. The result, it is hoped, should
aid both the author and the reader in interpreting the

validity of the model and of its results.
Assumptions Common to Multiplier Analysis

1. There are no negative local indirect impacts
from the acquisition of funds to build the project by the
public sector. The situations in which this would be

true are:

 

175

a) "Money creation" equal to the public expendi-

ture,

b) Bond sales which were paid for from "hoarded"

savings,

c) A balanced budget multiplier equal to the

estimated indirect effect of the project, or

d) All money used for project cost comes from

taxes in pphg: regions

2. The money invested in the project would not other-
wisse: have been spent. Otherwise, the multiplier impact
WOLl].d have occurred anyway.

3. Federal monetary and fiscal policy will continue
to t>e administered such that the employment rates and re-
sultant opportunity costs will not change during the time
Period under analysis.

4. The total supply effects of the stemming-from
i”fl?a&:ts and existent supply elasticities are sufficient
to <2cnnp1ement the induced-by effects at current price
levels. Otherwise, the demand multiplier effects would
Create mostly inflation.

5. Changes in personal income represent "equivalent"
ChaJlges in inter-personally comparable social welfare.
Otherwise, the comparison of income changes is not a good

we 1 f are measure .

176

Assumptions Common to Input-Output Models

1. All of the production functions involved are
liliearly homogeneous: The marginal factor-factor and
fac:tor-product relationships do not change as output ex-
paJldS.

2. The composition of input purchases and the dis-
tiriloution of outputs remain the same as output increases

in each indus try .
.Additional Assumptions From Conventions in Rent Model

1. All purchases of inputs from firms classified
as "Local" come from within the project region. This as-
sumption was implemented through the assignment of locali-
zation ratios equal to unity (1.0) for the "Local" indus-

tries.

2. Purchases of inputs from industries classified
as "PJational" are distributed inter-regionally in the same
proPortion as is employment in those industries. This
assumption was reflected in the localization ratios, which
were derived by dividing employment in the particular in-
duEYtry in the project region by the employment in that in-
dufiStiy'in the nation as a whole. This assumption and the
immediately preceding assumption were made in attempting
t" determine the lpppl impacts using a mppipmpl input-

O'uliput model.

177

3. The "National" industries which are effected by
the: project are not input-market-oriented. This assumption
resnilts from the restraints imposed by the localization
rat:ios also. If the National industries were input-market—
oriJented, the localization ratios would greatly understate
the: proportion of induced-by sales which would come from
the project region.

4. The "Local" industries which are effected by
thi? project are output-market-oriented. All of the sales
of' Isocal industries are assumed to go to the project
regipn.

5. The marginal wages paid equal the average wages
fcxr each effected industry. Thus, each new job created
ill teach industry adds one new worker at the "average" wage
foz‘ that industry.

6. The average alternative cost is equal to the
average production worker wage in the lowest paying in-
‘hlstkry in the region. This assumption is related to the
metihod of computing the Wage Rent Coefficient. Since the
reall- alternatives for each worker was not known, it was
assumed for computational purposes that the alternative
would be at least equal to the average wage in the lowest
wage sector.

7. The surplus of actual profits over the profits
which would be required to bring about the current rate of

‘PTOduction is equal to after-tax profits. This amounted

t0 the assumption that all "take-home" profits were a

178
surplus which could be taxed away without affecting out-
put.

8. The marginal rate of profit rent is equal to the
average rate for each industry affected. The "average
profit rent" was assumed to be equal to the after-tax
profit rate. This rate was used in computing the rates for
any change in induced sales.

9. Declines in employment occur only in jobs whose
overall average wages are equal to those in the lowest
wage industry; declines in production occur only in those
industries which are earning no profits. This assumption
is implicit in the Rent Model, since the model traces only
the expansions and not the contractions which occur in
the system. The subtraction of the "alternative" wages
from the new wages attempts to adjust for these changes on
the employment side. Declines in production are partly
dealt with by the assumption made elsewhere that the demand
change is "new" demand and not a shift. Failure to account
for profit contractions due to labor resource price in-
creases are "justified" by assuming zero profitability of
contracting industries.

10. The total rents which are generated in the first
round of expansion are equal to the portion which is
creditable to the project from all rounds of expansion.

The problem of assigning credit for all rounds of expansion

was "assumed away" by adopting the convention of assigning

179
only the credit for the rents generated in the first round
to the project.

11. The relative cost of each factor in each in-
dustry indicates its relative marginal value product in
that industry. Since the production functions are linearly
homogeneous, the marginal value products do not change as
output increases (and inputs are added in constant pro-
portions). This assumption arose from the convention of
assigning stemming-from credit on the basis of the pro-
portion of total cost in the stemming-from production
process which was represented by the project output.

12. Net marginal social utility in the consumption
of goods produced by the industries which decline is
equal to zero (for all units by which production is cur-
tailed). The welfare of producers in the declining in-
dustries was dealt with in assumption 9, above. However,
the welfare of consumers of those products must also be
dealt with. The simplest procedure is to assume that all
consumers of the products no longer available were 223
realizing any "consumers surplus" from the consumption of
those commodities. That is, the demand for those commodi-

ties was perfectly elastic.

180

Footnotes

 

l. A. Allan Schmid and William Ward, A Test of
Federal Water Project Evaluation Procedures With Emphasis

onCRegional Income and Environmentaliguality (East Lansing:
April 1970).

 

 

 

2. The closest relationship is found between the
proportion of large firms in a particular industry and the
average wage rate for the industry. Generally, the greater
the proportion of large firms, the higher are the average
wages. See Stanley H. Masters, "An Interindustry Analysis
of Wages and Plant Size," The Review of Economics and
Statistics (August 1969).

 

 

3. "The Problem of Measuring Profits: A Preliminary
Note by Wesley C. Mitchell" in Ralph C. Epstein, Indus-
trial Profits in the United States, National Bureau of
Economic Research, Inc. (New York: 1934), 4.

 

4. George J. Stigler, Capital and Rates of Return
in Manufacturing Industries (Princeton: 1963).

 

 

5. U. S. Congress, Profits: Report of a Sub-
Committee of the Joint Committee on the Economic Report
on Profits'Hearinge (Washington: 1949).

 

 

6. For example Stigler, o cit, and Ralph C.
Epstein, Industrial Profits in the United States (New
York: 1934).

 

7. Huffschmidt, et a1., Standards and Criteria for
Formulatin and Evaluatin Federal Water Resource Develep-
ments (Was ington: June 0, 1961).

 

8. "The Impact of Steel Upon the Greater New York-
Philadelphia Industrial Region." The Review of Economics
and Statistics (November 1953).

 

9. That is, whether the firm were "product market
oriented" or "factor market oriented." See Isard, et a1.,
Methods of Regional Analysis, Chapter 7.

 

10. See, for example, Calvin S. Schneider, Proce-
dures Manual for DeterminingPreliminary Empansion Benefits
fer AppalaCHianIWater Prejects, (Lexington: 1968). ‘

 

 

11. "National" industries are those which sell
primarily to national markets.

12. See Edwin Kuh, "Unemployment, Production Functions,
and Effective Demand," Journal of Political Economy (June
1966).

 

181

13. Oskar Morgenstern, On the Accuracy of Economic
Observations (Princeton: 1950).

 

 

14. Wassily Leontieff, et a1., "The Economic Impact-
Industrial and Regional of An Arms Cut," The Review of
Economics and Statistics (August 1965).

 

 

15. Schmid and Ward, op cit.
16. Ibid.

17. The Localization Ratio for Stone, Clay and Glass
products was set at zero, because the principal input
purchased from that sector by the steel industry is lime-
stone, which comes from outside of the region. See Schmid
and Ward, op cit, 29-30.

18. Clark, et a1., op cit.
19. Unemployment, Idle Capacity, and the Evaluation

of Public Expenditures: tNatiOnal and Regional Analysis
(Baltimore: 1968), 20-21.

 

 

20. See A Summary Analysis of Nineteen Tests of
Proposed Evaluation Procedures on’Selected Water and Land
Resource Projects, Report totthe Water Resource CounciI
by the Special Task Force (Washington: July 1970).

 

 

21. William A. Ward, "Economic Research: Prologue
and Epilogue," American Journal of Agmicultural Economics
(February 1970).

 

22. ”Resolving the Three-Horned Planning Dilemma,"
paper presented at a meeting of the American Society of
Planning Officials, New Orleans, March 29, 1971.

CHAPTER VI

A COMPARATIVE ANALYSIS OF TWELVE SECONDARY BENEFIT
APPRAISALS

Overview

In Chapter II, the different concepts and definitions
of secondary benefits were discussed. Three general cate-
gories of such benefits were delineated. In Chapter IV,
the different methods of estimating one of these cate-
gories, the multiplier impacts, were discussed. In Chap-
ter III, the concept of "rent multipliers" was introduced
and discussed as an alternative to the traditional "income
multipliers," and in Chapter V a model designed to estimate
the rents creditable to the initial project was discussed
and used in estimating the secondary rents from the Trenton
Channel Project. In the closing section Of Chapter V,
the underlying assumptions of the model were delineated,
and the role of assumptions in economic analysis were
briefly discussed.

In Chapter VI, the "customary" variety of secondary
benefits from four Federal water resource projects are
looked at. In the case of each project, three different
estimates come from the reports submitted by the Test

Teams which tested the Water Resource Council's proposed

182

183
evaluation procedures as presented in June 1969. The third
estimate comes from the application of the model which was
presented in Chapter V to the respective projects (see
Appendix). The objective of this comparative analysis is
to determine the differences in analysis and in results
which might arise from differing concepts regarding second-
ary benefits (the "customary" variety) and from differing
methodological approaches to dealing with them on the part
of different analysts.

In the sections of this chapter which immediately
follow, the concepts, beliefs and assumptions which are
discernible from the reports to the Water Resource Council
by the Test Teams are discussed. This discussion reveals
significant differences among the Test Teams regarding the
concept of secondary benefits. Several different measures
are proposed by the teams, as well as several methods of
estimating the relative magnitude of the measures suggested.
Widely differing assumptions are discerned regarding the
state of the economy and aggregate demand maintenance.
Other assumptions regarding resource mobility and inter-
regional employment differences are also seen. Tables 2,
3, 4, and 5 outline the assumptions and beliefs of each of
the Test Teams regarding the evaluation of secondary bene-
fits.

In Table 6 the secondary benefit estimates of each of
the Test Teams are presented, along with those generated by

the model presented in Chapter V of this study. As can be

184
TABLE 1

After-Tax Profits Per Dollar of Sales Compared With Before-

Tax Profit Rates Reduced by Fifty Per Cent

Before-Tax

 

 

 

 

Profits/$ After-Tax
Sales/2a -Profits/$ Salesb
Local Industries
Electricity, Gas 8 Water .0891 .0873
Printing 8 Publishing .1016 .0512
Transportation 8 Warehousing .0272 .0135
Wholesale 8 Retail Trade .0121 .0146
Communications .1016 .0943
Services .0235 .0219
Construction --- .0160
National Industries
Metal Ores Mining .0604 .0403
Coal Mining .0263 .0355
Non-durable Goods .0338 .0371
Lumber 8 Wood Products .0283 .0402
Chemicals 6 Allied Products .0338 .0615
Primary Metals .0450 .0547
Fabricated Metals .0339 .0438
Stone, Clay 6 Glass Products .0416 .0464
Non-electrical Machinery .0522 .0533
Electrical Machinery .0399 .0415
Motor Vehicles .0527 .0557
Other Transportation Equip .0296 .0319
Instruments .0747 .0782
Misc-Manufacturing .0309 .0386
aAs used in Schmid and Ward, op cit, p. 35.

bSee Appendix Table 2, Column 2.

 

Trenton Channel Project:

185
TABLE 2

Outline of Test Team Approaches to

Secondary Benefit Estimation

 

Corps of Engineers
Test Team

Michigan State

University Test Team

 

Definition
of Secondary
Benefits

Changes in employment and

income stimulated by the
project. Increased
stability.

'Changes in rents

to factors stimu-
lated to higher
employment by the
project.

 

Measure of
Benefits
Used
(Suggested)

National vs.

EmpIOyment generated in
steel and related sector
and induced employment
in "service" sector.

Alltreal benefits in

DiffErence in
wages generated by
sector and lowest
alternative wages
available to those
workers. "Abnormal"
profits earned on
increased output.
Some regional ih-

 

 

Regional regional account belong come gains are
Benefits in National. Regional transfers from
benefits at expense of other regions.
other regions should not National gains are
be counted as benefits. arrived at only
after netting out
the transfers.
Estimation EmpIOyment Base Multi- Input-output,
Procedure Used plier sectoral multipli-
(Suggested) ers, adjusted for
alternative wages.
EXplicit 1. Extensive growthiin 1. All employment
and Detroit represented in- generated would be
Implicit tensive National grthh supplied by labor-
Assumptions 2. LaBorers employed by ers from Service
and Beliefs plant were earning lower sector ("under- V
of Test wage in other region employment"benefits)
Teams 3. Extensive growth in 2. Expansion of

 

Detroit is good from the
local viewpoint

4. Basic/non-basic ratio
is stable

outputs and inputs
will be linear

3. No supply in
elasticities in
first round

4. Half of profits
are "abnormal"

5. Localization of
impacts would be
equal to current
empleyment ratios

186

TABLE 2 (Continued)

 

Corps of Engineers

Michigan State

 

Test Team University Test Team
Formula and/or Et = 3 0 31) (31h - ash)
value of E - E ° , _ .
multiplier t nb Where. aij - direct
Where: Et = total requirement from in-

Rounds of
Expansion

employment in the

region

Enb = regional
employment

in non-basic (i.e.,
service) industries

Infinite

dustry i per dollar of
sales by industry j,
aih = value added

per dollar of sales by
industry i
aSh = value added

per dollar of sales by
service industry

One round

Stonewall Jackson Reservoir:

187
TABLE 3

Outline of Test Team Approaches
to Secondary Benefit Estimation

 

Definition of
Secondary
Benefits

Relevant
Benefit
Parameter

National vs.
Regional
Benefits

EStimation
Procedure Used

Corps of Engineers
Test Team

Wages, salaries 6

profits from invest-

ments wholly or
partly induced to
locate in region by
project

Employment andw
personal income

Regional =‘National
if resources used
would have been un-
employed; otherwise
no national bene-
fits

Economic Base
Multiplier

Cornell University
Test Team ”

Regional: Change in
real regional income.
National: Payment to
previously unemployed
factors when water
project is clearly a
constraint to growth
Per capita thange in
real income - i.e.,
intensive income

rowth

oth can occur. Re-
gional might be trans-
fers from other
regions, however,

 

(Input-output model,
induced multiplier)

 

ESuggested)
xp 1cit and
Implicit
Assumptions

and Beliefs
of Test Teams

Formula and70r
Value of
Multiplier

ll Basic-non basic

regional income
grows. 2. Alter-
native factor earn-
ings equal zero

3. Perfect factor
markets in other
regions; imperfect
factor markets in
project region

ET and YT

 

Ht ‘ ENB YT ‘ YN13
Where: ET = total

regional employment
ENB = non-basic

(service employment

in the region,

YT = total regional
income

YNB = non-ba51c

(service) income in
the region

ll. Migrants earn no
ratio is constant as rent.

2. Only un-
employed factors can
create real national
gains. 3. Labor
supply and aggregate
supply not perfectly
elastic. 4. "Use ex-
ternalities" exist
but cannot be evalu-
ated

2.36 adjusted to‘l.8
to reflect supply
inelasticities

188
TABLE 3 (Continued)

 

 

Corps of Engineers Cornell University
Test Team Test Team
Rounds of Infinite Truncated at 1/2 total

Expansion

effects to adjust for
supply inelasticities

Poteau Watershed Project:

189
TABLE 4

Corps of Engineers
Test Team

Outline of Test Team Approaches
to Secondary Benefit Estimation

Soil Conservation
Service Test Team

 

Definition of

——r

Growth in income per

Net income effects of

 

Secondary capita (intensive), Second and following

Benefits after deducting in- round of expansion.
flation effects

Relevant EXtensive employment Multiplier effects of

Benefit growth. (Change in difference between

Parameter real income) poultry processing

National vs.
Regional
Benefits

E§timation
Procedure Used
(Suggested)

Explicit and
Implicit
Assumptions
and Beliefs
of Test Teams

Equal, if unemplOyed
resources existed in
project area

Employment BaselMult.

(Keynesian Multiplier)

(Input-output multi-
plier
(Comparative projec-
tion)
1. All resources used
would be unemployed
otherwise. 2.
gate supply not per-
fectly elastic (in-
flation). 3. All
multipliers yield
same results. 4.
Marginal utility of
Regional income =
national average

Aggre-

wages and unemploy-
ment compensation.
National multiplier

= 3.85. Regional

= 1.57 because of
"leaka es."

Multip ier transplant

lTiResource alloca-
tion frictions (not
barriers) existed
which would take 20
years to overcome.

2. Water shortage was
effective "bottle-
neck."

 

Formula and/Or

 

l - . .
Value of (I-MPL suggested) giéigfizij 13§§§’
Multiplier ET = 1 65 d Local "induced by":
W ° use 1.10
Where: MPC =
Marginal Propensity
to consume
ET = total regional
employment
ENB = regional non-
basic (service)
employment
Reunds Of* Infinite Infinite

Expansion

Mountain Home Project:

190
TABLE 5

Outline of Test Team Approaches to

Secondary Benefit Estimation

 

Corps of Engineers
Test Team

University of
Wisconsin Test Team

 

Definition of
Secondary
Benefits

Increase in level and
stability of personal
income. More diverse
economic base. More
jobgepportunities

Change in rents
realized by factors
affected, both posi-
tive and negative

 

Measure of
Benefits Used
(Suggested)

First roufid Of multi-
plier generated by
Team actually used
gross income from
recreation expendi-
tures

Income from first
round of agricul-
tural sales and
purchases. (Change
in ESE income of
factors).

 

National vs.

Regiona1==National.

Partly competitive;

 

 

Regional Unemployed resource partly complementary.
benefits equal in two Net national gains
accounts, though are "probably

minimal."

EstimatiOn Multiplier Transplant Input-output.

Procedure Used for irrigation effects. Sectoral multipliers

(Suggested) Base multiplier for
Recreation develop-
ment, using mis-
specified Keynesian
form

Explicit l} Income multiplier ll} Water availa-

and in region exceeded bility would in-

Implicit that in other regions. crease average size

Assumptions 2. Propensity to con- of farms in project

and Beliefs
of Test Teams

sume locally out of
local income=propor-
tion of visitor ex-
penditure spent
locally. 3. National
benefits are chiefly
local in accrual.

4. Alternative labor
opportunities for
farmers existed,
though they were in
other regions. 5.
Rate of return on
farm capital non-
project alternative

area and open new
farming opportuni-
ties there. 2. New
farmers would come
from contiguous
region a rest of
nation, causing de-
clines in income in
these places.

3. All farmers had
some alternative.

4. Farm income
changes re-spent in
fixed proportions.
5. After first round
of re-spending, in-
come effects could
not be attributed to

y_preject.

191
TABLE 5 (Continued)

 

Corps of Engineers University of
Test Team Wisconsin Test Team

 

Formula and/or

Secondary Tertiary aij(ahj); where:

 

 

Value of sector + sector
Multiplier Income income aij = purchases from
and industry i per dollar
of sales by industry
MPC j, a . = value added
me . .hJ .
in industry 3.
acres irrigated -
ROunds of~ Infinite One round)

Expansion

192

TABLE 6

Estimates of Regional and National Secondary Benefits:
Comparison of Rent Model and Approaches Used By Others

 

Regional Secondary National Secondary

Benefits ($/year)

Benefits ($/Year)

 

Trenton Channel
Project
MSU Test Team
Corps Test Team
Rent Model
High minus low

Mountain Home
Project
U of Wisc.

Test Team
Corps Test Team
Rent Model

High minus low

Stonewall Jackson
Project
Cornell Test Team
Corps Test Team
Rent Model
High minus low

Poteau Project
Corps Test Team
SCS Test Team
Rent Model

High minus low

1,719,000
135,000,000
2 055 584

W

1,415,000
16,649,000
405 000

Irma-fora

55,000
70,600,000
135 000

W

1,584,000
395,000
200 000

mm

0
0
2 831 000

2,453,000
2,972,000
415 000

275571-000—

137,000
7,080,000

0
:,UEU,UUU

368,000
614,000
240 000

m

193
seen, the estimates vary greatly. The estimates presented
by the Test Teams were hurriedly prepared; many of them did
not result from the application of the approaches advocated
by the respective Test Teams. Thus, the magnitudes are not
directly related in all cases to the estimation methods
advocated by the teams. Nevertheless, the estimates are
of significance, since they give an indication of the rela-
tive magnitude of secondary benefits estimates which the
reSpective Test Teams would expect to result from their

advocated estimation procedures.

Trenton Channel Project

 

The Trenton Channel Project has been briefly described
above in Chapter V. No additional description will be pre-
sented here.

The comparative analysis of the Test Team appraisals
of the Trenton Channel Project presented herein differs
somewhat from that of the other projects. In the dis-
cussion which immediately follows, the results from the
Rent Model are compared not only to the results from a
different approach but also to the results from a slightly
different version of the Rent Model. This additional com-
parison allows one to appraise the differences in results
obtained from two appraisals which are similar in concept
but different in coefficients, as opposed to those from two

appraisals using essentially different approaches.

194

Michigan State University Test Team

 

The Michigan State University (MSU) Test Team approach
to estimating the secondary benefits from the Detroit Tren-
ton Channel Project used an earlier version of the model
presented in Chapter V of the present study. The earlier
version was identical in concept to the present version.
Thus, most of the assumptions outlined in Chapter V above
also apply to the version used by the MSU Test Team. The
greatest difference lay in the treatment of rents stemming-
from the increased production of steel. The MSU Test Team
used the full value of the resulting output of the Fabri-
cated Metals industry in estimating the Wage and Profit
Rents stemming-from primary steel production. No other
industries were included in the analysis of stemming-from
impacts, since steel was a relatively minor input to mOst
other industries. Because of the inverse relationship be-
tween relative factor cost and traditional stemming-from
benefits, the MSU Test Team felt that to include the
industries in which steel was a minor input was "to assume
the tail wags the dog."1

The application of the earlier version of the model
by the MSU Test Team differed also in the direct require-
ments coefficients used in estimating the secondary out-
puts. Since the 1963 Input-Output Study was not available
to the Test Team at the time of its analysis, the coeffi-

cients from the 1958 study were used.

195

A third difference between the MSU Test Team version
and the present version of the Rent Model lay in the wage
rates used in evaluating the alternative earnings of local
laborers. The earlier version used national wage rates for
both local and national industries, while the present ver-
sion used local wage rates in assessing the alternatives
to labor in the local industries.

The fourth difference between the MSU Test Team pro-
cedure and that used in Chapter V herein related to the
treatment of Profit Gains. The Test Team used the profits
before taxes in estimating the Profit Gains, but then
”rather arbitrarily" reduced the result by one-half to re-
flect the recognition that these profits were not "net"
in the sense that they could not have been taxed away with-

2 In

out causing a reduction in the level of investment.
fact the "rather arbitrary" reduction of before-tax profits
by 50 per cent was no more arbitrary than the acceptance of
the after-tax profit rate used in Chapter V herein. The
fact that the results are relatively close (Table 1) is
probably due in greatest part to the 48 per cent rate at
which corporate profits are taxed. Consequently, there is
little difference in the results obtained from the Rent
Model under the two different conventions for dealing with
profits.

The absolute difference between the results obtained

from the two versions of the Rent Model appear large

196
(Table 6). However, when compared to the differences which
obtain between analyses based upon differing approaches
in ensuing sections of the present chapter. The differences
obtained from the two versions of the Rent Model are rela-
tively small. While the differences between coefficients
in the two versions caused some difference in the final
estimates obtained, those differences are not so large as
the potential differences obtainable when differing bene-
fit parameters are being estimated -- e.g., gross activity
generated as opposed to employment in the "non-basic"
sectors. Much larger differences are obtained when the
basic assumptions differ between analyses -- e.g., those
relating to the alternative cost of the factors affected
and to the creditablity of ensuing rounds of expansion.

The MSU Test Team differed from the other test teams
in the treatment of the source of funds for financing the
project. The MSU Test Team argued that the taxing and
spending decisions of the Federal government should not
be treated as separate and independent acts, not even from
the local standpoint. The process of building a Corps of
Engineers project in the Detroit area required that taxes
be collected to finance the project. The MSU Test Team
pointed out that project regions typically view the project
expenditures as "new money" in the region: In the aggre-
gate, such cannot be the case, however, since each new
project increases taxes in all regions. The MSU Test Team

was arguing that at least some of the indirect effects of

197
project expenditures would have occurred in the region any-
way, had money not been taxed away from the region to help
build the project. When all projects are viewed together,
the aggregate demand effect simply reduces to the balanced
budget effect, which is a much smaller change than that

suggested by the sum of each project treated in isolation.
Corps of Engineers Test Team

The Trenton Channel Project constituted a Special
case of secondary benefit evaluation. The Detroit metro-
politan area was characterized by a low unemployment rate
and a high level of average income at the time of the test
team reports. The Corps of Engineers Trenton Channel
(CETC) Test Team pointed out that the major economic
problem in the area was the cyclical nature of employment
and income because of the high proportion of durable goods
manufacturing.3 While average wages in these industries
tended to be much above the national average, employment
in these industries was highly sensitive to overall changes
in the economy. It was pointed out that diversification of
the economy to achieve more stable employment was difficult
because the high wages paid by the automobile industry and
related industries tended to keep out the less cyclical in-
dustries which tended to pay generally lower average wages.
The expansion of the McLouth steel plant, which would be
made possible by the project, was not seen as having a

stabilizing effect upon the local economy.

198

The expansion of the McLouth steel plant would be the
generator of most of the benefits of the project, since it
could not be proven that other firms would use the channel.
Thus, the employment effects resulting from the steel mill
expansion constituted the basis for the estimated secondary
benefits.

The primary employment impact would be an increase of
3,000 workers at the steel plant. The test team postu-
lated the effeCts of this additional employment on the
local economy by assuming an employment base multiplier
of 3.0 -- i.e., a ratio of "primary" to "residentiary"
employment of 1.0:2.0. Thus, 6,000 additional jobs in the
services sector would be created. The CETC Test Team also
postulated that stemming-from effects would be realized
in the metal fabricating industry leading to the employ-
ment of an additional 1,500 workers in that industry. The
team assumed that one-half of the McLouth company's out-
put would go to metal fabricators, while the other half
would go to the automobile industry. That part going to the
automobile industry would have no local stemming-from impact
on employment, since the steel would have otherwise been
bought elsewhere and the same level of local employment would
have occurred. The stemming-from employment effects would
also have an effect upon local "residentiary" employment.
The employment multiplier of 3.0 was applied to the 1,500

jobs created in Metal Fabricating, indicating an additional

199
3,000 jobs in the service industry. Thus, the total employ-
ment effects postulated for the region were 4,500 industrial
jobs and 9,000 service sector jobs.

Since the Detroit metropolitan area did not have a
pool of idle labor, the team concluded that these jobs
would be largely filled by migrants from other regions.
Assuming a laborer-dependency ratio for the industrial
sector of l.0:3.0 and for the "residentiary” sector of
l.0:2.0 (since half of these employees would be female),
the team estimated that regional population would increase
by 22,500 persons.

The CETC Test Team concluded that the overall effect
of the project "would be to increase employment and pOpula-
tion in the Detroit area at the expense of other regions
but to the overall benefit of the nation."4 This conclu-
sion could not have been derived solely from the data
presented. There were involved some assumptions regarding
the relative welfare of the prospective migrants in their
community of origin versus their community of destination,
the Detroit area. The test team spoke of "regions" as the
bearers of benefits and costs rather than people as the
bearers. Thus, it was not clear whether the test team was
referring to the migrants alone, or all of the people of
the areas involved. Their statement would imply that the
"collective good" of the Detroit area would increase, while
that of the other regions would decline. Perhaps the

ambiguities surrounding the problem of evaluating real

200
welfare under changing conditions of pOpulation density,
related social anomie, etc., should not be treated herein,
though the test team made some sweeping judgements regard-
ing these factors in coming to its conclusion. Assuming
that the test team spoke only of the migrants involved,
something more concrete can be said about their assumptions
regarding the net income effects of the project. Osten—
sibly, all of the laborers attracted to the Detroit area
would have been earning lower wages in their regions of
origin. Thus, overall national efficiency would be en-
hanced by reallocating these resources to more productive
employment. The regions from which they came would be
made worse off by the loss of these laborers. One could
surmise that the disbenefit would arise largely from the
associated decline in regional production and the taxes
thereon.

The test team seemed to believe that "extensive"
economic growth was good, from the vieWpoint of the Detroit
region. Extensive growth has been defined previously as an
increase in regional employment which occurs without in-
creasing the wages of any people already in the region --
i.e., more workers at the same wage level. In the Trenton
Channel case, the new steel manufacturing employment would
be complemented by additional service employment according
to the existent "basic-non-basic ratio." This is, in
reality, extensive growth. Extensive employment growth can

be very beneficial to the tax base of economic communities

201
of smaller scale which are struggling to establish viable
community services. In the case of a metropolis such as
Detroit, however, it would appear that limited scale is no
longer a problem. Indeed, the problem with the provision
of these services in Detroit may well be that diminishing
returns to scale will occur with continued extensive
growth.

The test team apparently viewed the employment stimu-
lated by the project as "intensive growth" from the
national vieWpoint -- i.e., growth in which average re-
turns to labor rise. Such would be the case if a more
"efficient" allocation of labor occurred by the shift to
employment in the higher-wage Detroit area. If such were
the case, then greater ambiguity is made apparent in the
test team's reference to the relative welfare of "regions."
Ostensibly, the migrants will have been made marginally
better off by this change in residence and employment.
However, their region of origin has supposedly been made
worse off. But, using regional delineations for benefit
assignment, should the before-project or the after-project
populace be used as the basis for comparison? It is not
made clear which is used by the test team. It does make a
difference for the analyst's ultimate concern should be
for human welfare and not for some ill-defined regional

entity.

202

Stonewall Jackson Reservoir

 

The Stonewall Jackson Reservoir is a proposed Corps
of Engineers Project on the West Fork River in northern
West Virginia. The project area is located in Appalachia.
The project itself, as it was initially designed, is pri-
marily a flood control project with some water supply,
recreation and water quality (low flow augmentation)
aspects. Hydro-electric power production from this project
was found to be infeasible. The first cost of the project
was estimated to be $33,954,000 in July 1964 prices. The
annual operation maintenance and replacement costs were
estimated at $164,000, at the same price level.

The principal industries in the project area are coal
mining, glass manufacturing, and livestock production. The
area produces about 20 per cent of the total output of both
coal and agricultural products of the state of West Vir-
ginia. Nevertheless, population out-migration rates have
been very high; and, for all except perhaps the last decade
of the past thirty years, out-migration rates have exceeded
birth-rates, leading to a decreasing population. Largely
as a result of this out-migration, unemployment rates have
slowly declined from very high levels toward (but not yet
reaching) the national average. Average wages in the area
are nearly 20 per cent lower than national average wages.

The project area is located in Appalachia and, thus,

qualifies for the consideration of redevelopment benefits.

 

 

203
As a result, one would expect an extensive analysis of the
secondary impacts of the project. However, very little was
said about such impacts in the initial project report, and
little more has been added by the two test teams in their
analyses. Nevertheless, the two test teams do make ex-
plicit statements concerning the evaluation of secondary
benefits for the project which serve to illucidate their
stances and which allow for an evaluation of their basic

assumptions and presumptions regarding them.

Corps of Epgineers Test Team

The Corps of Engineers Test Team which applied the
proposed new evaluation procedures to the Stonewall Jackson
Project (CESJ Test Team) was composed of Corps of Engi-
neers personnel from the Ohio River Division and the Pitts-
burgh District Offices. Thus, the same offices were in-
volved in producing the CESJ Test Team Report5 as were in-
volved in producing the original Project Report. Therefore,
in applying the new procedures the test team had the added
advantage of familiarity with the project.

Because secondary benefits were not evaluated in the
original Project Report on the Stonewall Jackson Reservoir,
the CESJ Test Team did not have the advantage of having
had data collected for that purpose. The test team indica-
ted that the time limitations were such that they were un-
able to provide any new data for the test evaluation. Thus,

all of the analysis of secondary effects was hypothetical.

204

The CESJ Test Team began their discussion by stating
their concept of secondary benefits. The benefits from
"regional economic development" would consist of wages,
salaries and profits from investments that would be "wholly
or in part induced to locate in the project area." It was
pointed out that, insofar as these might have been located
elsewhere in the absence of the project, the total was not
net to the nation as a whole, though it would be to the
region. However, if the resources which were employed as
a result of the investment locating in the project area
would have been unemployed otherwise, then the wages and
salaries would be net benefits to the nation as a whole.6
Though the CESJ Test Team did not point out the reasoning
behind this conclusion, the implicit assumption is that the
opportunities available to the resources which would have
been employed in the other region in which the investment
would occur in the absence of the project were exactly
equal to the factor payments which the foregone investment
would have made, while the alternatives in the project area
were assumed to be valued at zero. Thus, the investment
must have been attracted away from another region which
had perfectly functioning factor markets and into a region
in which no effective demand existed for the resources which
the investment employed.

The CESJ Test Team saw the basic effects upon employ-
ment as occurring via the construction, operation and main-

tenance of the project and via the provision of flat,

205
flood-free land for commercial and industrial development.
They foresaw additional employment resulting from the de-
mands imposed by recreation users of the lake for gasoline,
restaurant and motel facilities. The test team believed
water supply provision to bear insufficient inducement to
attract industrial location and felt that the greatest
potential for regional employment generation lay in the
provision by the project of flood-free plant sites.

Because the Project Report on the Stonewall Jackson
Reservoir had not included an analysis of the employment
effects via investment attraction, the test team was unable
to say how much of an attraction to new investment would
be created by the project and how much of an effect upon
local employment the project would have. Nevertheless, a
hypothetical rule-of-thumb analysis of the benefits from
such effects was presented.

The test team used a variant of the method develOped
by Spindletop Research Center7 to estimate the number of
potential jobs resulting from the creation of flood-free
industrial sites. The average number of employees per
acre in existent plants in the area was determined and
multiplied by the acreage of suitable plant sites created
by the project. This gave an estimate of the potential new
industrial jobs. The average wages and salaries paid in
these jobs would then be multiplied by the number of new
jobs to determine the total of the new income created by

the potential plant locations. The most likely industries

206
to locate on these sites were judged to be apparels and
electrical machinery manufacturing. This was concluded
from information contained in the economic base study of
the Ohio River Basin Comprehensive Survey. The average
of wages paid in the two industries was taken as the
relevant wage for use in estimating the total potential in-
come change. Then, it was suggested that a multiplier of
2.0 be used in estimating the associated change in service
sector employment and income. The multiplier was taken
from studies of the Appalachian area by Robert R. Nathan
Associates, Incorporated, and Spindletop Research Center.
The total would yield an estimate of the regional develop-
ment benefits. However, only that part of the total which
would accrue to unemployed resources were viewed as national
gains.

Since the out-migration experienced in the area in
previous decades had also decreased the unemployment rate,
it was concluded that unemployment in the area would even-
tually equal the national rate of unemployment. Trend
analysis indicated that this would occur around the year
2000. Thus, unemployment benefits could only be claimed
for that portion which in the absence of the project would
not have out-migrated and found employment in other areas.
For test evaluation purposes it was assumed that the project
would reduce the rate of unemployment in the area to the
national average rate by the year 1975. Those employed by

the project in 1975 would be accounted for in the national

207
account. However, diminishing proportions of those em-
ployed would be included in the national account for ensu-
ing years, and for the year 2000. and thereafter none would
be included.

The CESJ Test Team was perhaps the most skeptical of
all the Federal agency test teams regarding the nature and
existence of secondary benefits for the project which they
were evaluating. Perhaps this was due to the nature of
the project and the area in which it was located. The
project had no irrigation nor hydro-electric capacity
built in. These are the most easily evaluated aspects of
a project so far as secondary income creation is concerned,
since they are both directly used as production inputs and
often are the limiting factors of output expansion. In
most cases the users of these project outputs are already
located in the area or are planning to do so pending their
provision. However, in the Stonewall Jackson Reservoir
case, the other factors were not obviously ready and wait-
ing for the project outputs, and the outputs were not so
obviously strategic to any foreseeable production process.
These factors might have been instrumental in eliciting the
warning which the test team voiced regarding the credit for
expansion benefits. In essence, they warned against the
"one factor theory" of production in saying that credit for
any expansion benefits would be in part due to the new
interstate highway system and the Appalachian corridors,

as well as other associated investment expenditures. Thus,

208
while the provision of flood free land was seen as impor-
tant to the develOpment process, the test team felt that
all of the benefits from economic expansion should not be
assigned to the provision of flood control and urged that
an attempt be made to develop a method for determining
the proportion which should be assigned to each contributing
factor.

The test team felt that underemployment was a "rea-
sonable simple concept" but that it was "very difficult to
measure." However, an attempt was made to measure the ex-
tent of underemployment in the area. "Per capita income
was used as a proxy measure in which the Clarksburg Economic
Area was compared to the United States average."8 The
percentage difference was taken as the measure of under-
employment. The reasons for using per capita income rather
than average wages or some other labor-based unit were not
given; neither was the "reasonably simple concept" of
underemployment held by the test team stipulated. However,
a recognition on the part of the test team of the part
played by educational and attitudinal factors in develOp-
ment was indicated in another section of the report. Thus,
the test team must have believed that even in perfectly
functioning markets the (short run) average wages for the
Appalachian laborers would have been lower than the average
for the rest of the nation. Therefore, the income problem
was not simply one of demand deficiency but also one of

resource productivity; the latter was probably instrumental

209

in causing the former. As a result, the test team tended
to under-emphasize the income producing effect of the aggre-
gate demand changes occasioned by construction, operation
and maintenance expenditures. Whether this was due to a
belief on their part that few local laborers would be em-
ployed and that most project-induced spending would occur
elsewhere, or whether it was due to a down-playing of the
effects of aggregate demand changes on local income could
not be determined. Nevertheless, it was apparent that the
test team gave little consideration to the aggregate demand
effects of the project-related expenditures on the area.

There was some inconsistency on the part of the test
team in that a relatively large multiplier was accepted in
evaluating the secondary income producing effects of employ-
ment opportunities created by flood-free site provision,
while little was said about the secondary effects of con-
struction, Operation and maintenance of the project. The
discrepancy in treatment must have been due to an oversight
rather than a belief that the multiplier effects would have
been 2.0 and zero, respectively.

An economic base multiplier of 2.0 was accepted as
the relevant measure of the secondary income producing
effects of new plant employment in the area. In using such
a multiplier to estimate secondary benefits several assump-
tions must be made. First, it must be assumed that the
relationship between "basic" employment and income and "non-

basic" employment and income in the area will remain

210

approximately the same as area income increases. Whether
or not this would be true for such a region as Appalachia
is perhaps debatable. Secondly, it must be assumed that
those employed in non-basic activities because of the multi—
plier effects of the new basic employment.would not have
been employed otherwise, and that their forced leisure was
of zero value to them. Thirdly, if the secondary benefits
are to be net to the nation, it must be assumed that the
other region in which the plant would have located was ex-
periencing full employment and perfect factor market opera-

tion.

Cornell Universipy Test Team

 

The Cornell University (CU) Test Team Report discus-
sed both the economies of scale type of secondary benefits
and the conventional variety of secondary benefits. How-
ever, the test team preferred to view the former as a
variety of "national efficiency benefits." Only those
changes in income which occurred through the multiplier
effects of the project were viewed as secondary benefits.
While the test team believed that net secondary (multiplier)
benefits could prevail at the national level, it was de-
cided that the multiplier impacts would be evaluated only
in the Regional Income Account of the project analysis, and
it was suggested that any adverse effects upon other regions

should be analyzed and discussed therein as well.9

211

The CU Test Team was very dubious of the analysis of
secondary effects under the "use externalities" definition
provided by the Task Force. They were particularly con-
cerned about the imputation of causation for the increased
outputs to water services, ". . . in view of the essentially
permissive nature of water resources in the regional growth
process. . ." Thus, they felt that the evaluation of such
effects should be included ". . . only if a strong and
prior case can be made that shortage of water services act
as a constraint on more optimal production or that a water
project will bring about technological change." Like the
Corps of Engineers Test Team the Cornell University Test
Team felt that too many other factors were ". . . required
at predetermined levels (water is usually not) to be able
to ascribe reduced labor unemployment solely or partially
to the effect of water development on industry growth."
Thus, the test team recommended that ". . . both of the
approaches, suggested by the Task Force, to measure
'national income secondary benefits' he rejected in the
case where a proposed project is assumed to have impacts
on labor unemployment and immobility through expansion of
industries directly or indirectly using project outputs."10
The reasons for the suggested rejection were delineated as
1) the inability to separate joint effects, 2) the permis-
sive role of water in growth, 3) the possibility of
technical substitution, and 4) the uncertainty involved in

forecasting human resource immobilities.

212

The CU Test Team expressed similar beliefs to those
expressed by the MSU Test Team regarding the credit for
economic expansion. The MSU Test Team felt that some credit
should be given to the water project, though they could not
say how much. The MSU Test Team gave credit for only the
first round of expansion. The position of the CU Test Team
was that me credit should be assigned to the water project.
This extreme position is justifiable only by negating the
projects primary benefits, which depended largely upon
the development of flood-free industrial sites. The CU
Test Team position on the indirect project impacts indi-
cates that the test team did not believe that the primary
project benefits from site creation would materialize.

The CU Test Team felt that indirect impacts upon the
region from recreation development were a distinct possi-
bility. Thus, their approach in dealing with recreation
impacts was different from that dealing with flood control.
It was felt that a multiplier analysis of recreation ex-
penditures could be carried out, since recreation develop-
ment was likely to occur. However, these impacts should
be limited to the regional account.

It was suggested that secondary (multiplier) benefits
of a project only be included in the national income effects
account for the impact on long run unemployed labor put to
work by the expenditure of funds for construction, operation
and maintenance. No other effects should be included, be-

cause of the uncertainty involved in predicting long range

213

resource immobilities and the expansionary effects of proj-
ect outputs on other sectors. Furthermore, in calculating
the multiplier effects of the project expenditures, care
should be taken to determine the origin of the resources
employed. Even though unemployment rates might be high

in the project region, the expansionary effect would not
constitute real net benefits if workers were drawn from
other employment in other regions.

The test team felt that the relevant measure of
secondary benefits was the growth in per capita income
rather than gross regional income or product. Thus,
"extensive" growth would not constitute a real secondary
benefit to the region. However, since one could easily
estimate the per capita measure from the total, it was
suggested that gross income estimating procedures were
permissable as means of getting at the secondary benefit
estimates.

The position of the CU Test Team regarding intensive
growth as the measure of regional development is similar to
the position taken in the present study. The Rent Model is
an attempt to operationalize the concept of intensive
growth through its emphasis upon changes in income to pre-
viously employed laborers. Similar positions were taken by
the MSU Test Team and the University of Wisconsin (UWMH)
Test Team.

Like many of the other test teams, the Cornell Uni-

versity Test Team appeared a bit contradictory in their

214
efforts to separate the national efficiency benefits, the
national secondary benefits, and the regional secondary
benefits. Part of the confusion perhaps was caused by
their attempt to consider secondary benefits only in the
regional income account -- as they interpreted the Task
Force Report to suggest -- while they believed that under
certain conditions net national secondary benefits could
occur as well. Additionally, the attempt to categorize
the project effects into "national efficiency" accounts and
"regional income" accounts rather than primary and second-
ary benefits also seemed to add to the confusion. In large
part, this latter confusion appeared to stem from the fail-
ure to distinguish adequately between regional income
changes due to inter-regional pecuniary effects and those
due to net national secondary changes in output which were
localized in the region. The same difficulty was had by
other test teams.

The test team was skeptical of the analysis of
secondary benefits at the national level due to under-
employment. They felt that under-employment was a very
poorly defined concept and that there was ". . . no re-
source that cannot be put to some 'higher use' given ef-
fective demand." Thus, they recommended that only employ-
ment benefits for unemployed leper be included in the
”national efficiency account." It was considered permis-
sable to include the evaluation of benefits from under-

employed labor in the regional income account, however,

215

since presumably one should include pecuniary effects as
well in that account.11

Again, the CU Test Team position is somewhat extreme
regarding benefits to "under-employed" labor. It is pos-
sible for laborers to achieve "higher" employment without
the changes being purely pecuniary. The conditions under
which such might be the case are perhaps also extreme,
but they nevertheless exist. For example, it is quite
possible that increased availability of a water service
might markedly increase the marginal physical product of
labor, particularly if the two were complementary inputs.
In the case of irrigation water provision, there is good

reason to expect such complementarity. In the extreme

case, if labor were perfectly inelastic in supply and the

 

product being produced were perfectly elastic in demand,
much of the increase in value produced would redound to
labor in the form of "rent." The price of the product
would not change. Purchases of other goods need not de-
cline, since the income generated would be sufficient to
purchase the increased output. Pecuniary ramifications
could be minimized by invoking monetary policy to increase
the nominal money supply equivalent to the change in output.
The CU Test Team position regarding changes in wages
to labor failed to consider both parameters of the derived
demand for labor. The CU Test Team dealt only with the
Keynesian induced demand effects upon the demand for prod-

ucts, which tend to be largely pecuniary in nature. The

216
demand for a factor of production depends also upon its
physical productivity, however, and resource demand changes
which occur through changes in physical productivity need
not be wholly pecuniary.

The test team was very cognizant of the secondary bene-
fit limitations posed by supply inelasticities, especially
that relating to the supply of labor. In essence, labor was
treated as the critical factor in the aggregate production
function. The assumption made by the test team regarding

the aggregate supply function (via the labor supply function)

were made explicit throughout the analysis of secondary bene-
fits. Moreover, an attempt was made to adjust the secondary

employment effects for the probability that not all of the

increased labor income would accrue to previously unem-
ployed laborers. The estimates of the multiplier effects
of recreation expenditures, for example, were cut in half
to reflect this belief.

By interpreting the instructions from the Task Force
in such a manner as to limit the consideration of secondary
benefits to the regional viewpoint, the test team, unlike
the MSU Test Team, circumvented the necessity of overtly
specifying (or implicitly assuming) the fiscal conditions
under which their analysis would be relevant. Except for
the fact that a small portion of the tax-financed cost of
the project would have come from the area, it would make
little difference from the regional vieWpoint how the
project was financed. From the national vieWpoint, how-

ever, it would make a difference; and the acceptance of

217
the Task Force instructions constituted an implicit accept-
ance of the assumptions which underlay them. By evaluating
only the secondary impacts accruing to the project region
they implied that either 1) the costs to other regions and
the nation as a whole were irrelevant either because they
were presumably small or because great value was placed
upon the stimulation of secondary benefits in the project
region, or 2) it was assumed that there were no secondary
impacts (either positive or negative) upon other regions
and that the project financing did not cause other employ-
ment opportunities to contract. The test team did amel-
iorate this implied position somewhat by suggeSting that
the regional account should also show the effects upon other
regions.

In suggesting the exclusion from regional secondary
benefit calculation of those laborers who came from outside
of the region, the test team by implication assumed that
either 1) the laborers were earning equivalent wages before
migrating, or 2) the process of regional population growth
had an equivalent welfare decreasing effect upon the pre—
vious residents of the region. In a perfectly competitive
setting (and ignoring the spatial aspects) the former would
be logically acceptable; however, if sufficient market im-
perfections existed to cause unemployment (or under- .
employment) in the region, it is possible that the adjust-
ments through migration were non-marginal to the laborers

involved. In the case of the latter, the test team might

218
have harbored such a belief, though in no place in their
report did they make it explicit.

In summary, the CU Test Team accepted a fairly re-
strictive definition of secondary benefits. Only the
multiplier impacts were considered "indirect." Their
consideration was limited to the regional account. More-
over, project effects upon resource productivity were con-
sidered to belong in the "national efficiency account."

The multiplier effects of project construction expenditures
were considered to contribute to regional secondary bene-
fits. Secondary effects from project outputs were con-
sidered to be too uncertain and/or not sufficiently
causally-linked to the project to justify their inclusion

as regional or national benefits, even in the presence of
unemployment. Thus, all stemming-from benefits were ex-
cluded from the analysis, and much of the induced-by effects
were omitted (the exception being those induced by project
construction). Changes in the employment of labor which
lead to higher regional per capita income were seen as the
relevant measure of secondary benefits. Direct effects

upon enemployed labor were seen as belonging in the national
efficiency account, while indirect effects upon unemployment
were seen as belonging in the regional (secondary) income

account .

219

Poteau Watershed Project

 

The Poteau Watershed Project was constructed by the
Soil Conservation Service Of the United States Department
of Agriculture. The basic work plan provided for (a) land
treatment measures to reduce flooding, sedimentation, and
erosion, (b) 17 floodwater retardation structures (c) one
multi-purpose structure for flood control and water supply,
(d) 10 miles of channel improvement. The area to be ef-
fected by the project consisted of Scott County, Arkansas,
and Le Flore County, Oklahoma. The project area was
classified as economically depressed and qualified for
the consideration of redevelopment benefits.

The most certain of the anticipated redevelopment
effects of the project were to come from the employment
effects generated by the increased output of a poultry
processing plant. The increased output was to be made
possible by the provision of additional water supply by
the project. Additional income and employment effects
were to come from increased eXpenditures for recreation in
the area by thoseliving outside of the area.

The original evaluation of the project by the Soil
Conservation Service (SCS) included, in essence, two cate-
gories of secondary benefits. First, their consideration
of redevelopment benefits in the form of increased wages
and salaries to those previously unemployed workers who
would be employed as a result of construction and operation

and maintenance of the project constituted one form of

220
secondary benefit. Also, the increased employment due to
the industrial water supply effects were placed in the same
category. Secondly, the SCS increased their primary bene-
fits by 10 per cent and described the addition as "secondary
effects" of the project. The 10 per cent adjustment could
be called a "multiplier" effect. Multipliers are normally
applied to some exogenous change in income or employment,
or whatever is being evaluated. Unless all Of the primary
benefits come in the form of exogenous changes in income,
the effective multiplier used by the SCS is not 1.10,
though there is implicit in the procedure used some sub-

scription to the multiplier concept.

Cqmps of Emgineers Test Team

 

A test team made up of staff members from the South-
western Division and the Tulsa District Offices of the Corps
of Engineers (CEP Test Team) comprised one of the units
which tested the prOposed new evaluation procedures upon
the Poteau Watershed Project. The CEP Test Team did not
attempt a primary research effort to evaluate the secondary
impacts of the project. However, they did present some
very definite ideas about how such evaluation should be
done.12

The CEP Test Team agreed in principle to the in-
clusion of the entire wages and salaries of the workers

employed by the processing plant as project benefits.

However, the team felt that the inclusion of the direct

221
change in wages alone was incomplete: The multiplier ef-
fects of those changes should also be included. In addi-
tion, the team felt that the proper measure of the recrea-
tion benefits of the project was the change in area income
directly and indirectly stimulated by expenditures for
recreation. The multiplier effects of the increased
poultry plant employment were called "redevelopment bene-
fits." The economic effects of the recreatiOn expenditures,
it appeared, were intended to represent the principal
quantitative measure of the recreation benefits.

The term "secondary benefits" was used only in
referring to the multiplier effects. The employment
effects, such as those felt by the poultry processing em-
ployees, were included in the primary benefit account. The
secondary benefits were computed as the multiplier effects
of these and other primary benefits. Secondary benefits
were computed for both the National Income Account and the
Regional Income Account. The argument for including them
in the National Account was two-fold. First, there was the
existence of unemployed and under-employed labor in the
project area. Thus, it was argued, resources put to use
in the area represented real gains to the nation as a
whole. Second, the argument was made that ". . . the de-
cision to develop one region of the country at the expense
of another contains inherent equity and regional considera-
tions, in terms of both income distribution and population

dispersal, which offset the opportunities foregone" (in

222

13 The first of these reasons would appear

other regions).
to be more amenable to economic analysis than the second.
The team was cognizant of the part played by aggre-
gate supply in determining the real income multiplier
effects. They discussed the existence of unemployment in
the area and the effect this unemployment would have upon
the net national income effects of industrial expansion
in the area. Moreover, they felt that increases in per
capita income,"after allowance for inflation," was the
proper measure of progress. Thus, they recognized both
the problem of aggregate supply elasticity and that of
"intensive" regional growth. That is, they viewed the in-
come issue as one of not merely increasing the gross product
of the region by job and worker migration, but of increas-
ing the real incomes of those already residing in the area.
In so far as there were unemployed and under-employed
workers in the area, it was possible to increase area in-
comes without causing a decline in incomes in another area.
The team felt that an income multiplier, "based on
the marginal propensity to consume," should be derived for
the area. The team was unable to derive such a multiplier
in the time allotted for the test, it was pointed out.
However, the team was confident that a reasonably accurate
multiplier could be developed from secondary data. The
use of an input-output study was mentioned, but it was con-
sidered to be too expensive. For purposes of illustration,

a multiplier of 1.65 was accepted for the test. The

223
multiplier was drawn from a recent economic base study done
by the Department of Labor which indicated that ". . . pro-
viding 100 jobs in manufacturing would create 65 additional

jobs in other segments of the economy."14

This multiplier
was applied to the income changes occasioned by the expan-
sion of the poultry processing plant. However, some in-
consistency in their approach resulted from its use, since
the economic base multiplier is an extensive growth multi-
plier, while the team felt that real growth resulted only
from intensive changes in income.

The team felt that multiplier should be estimated for
all project areas. They mentioned some possible means by
which these estimates might be made: "An input-output
study, a study of published data, a historic study of com-
pleted projects, field work, or any combination which will
produce the most reliable result.” "Historic study"
apparently refers to the comparative projection technique.
The team apparently believed that each of these approaches
would yield comparable multipliers.

The team considered both unemployment and under-
employment to be problems in the Poteau project area. Thus,
they believed that the multiplier effects of increases in
the level of regional aggregate demand would be beneficial
to the area. However, in discussing the evaluation of the
multiplier impacts, the team made no distinction between
the evaluation of the impacts upon the two groups. The

opportunity costs in both cases were assumed to be zero.

224
As was pointed out in preceding sections, only in very
limited circumstances would the alternatives be valued at
zero even for those totally unemployed. Certainly, in the
case of the under-employed one should not value the alter-
native at zero.

The team included the increased profits from the
poultry processing operation in the evaluation of project
benefits. They did not describe them as "secondary"
benefits, however. It was their belief that the profits
constituted primary benefits of the project. In evaluating
the secondary project effects, these profits were not
treated in the same manner as were the wages and salaries,
however. The multiplier was not applied to the increased
profits. Consistency would dictate that wages, salaries,
and profits be treated alike in this instance: Ostensibly,
all are exogenous increases in area income, and the second
round effects of the expenditure Of that income should be
similar. However, if the persons to whom the profits
accrued were not area residents, the exclusion of the
secondary effects of the profits from the regional account
would be appropriate. If unemployment existed in the area
in which the profits would finally be re-spent, then the
secondary effects would accrue in the national account,
though not in the regional account.

In summary, the team was cognizant of the role played
by unemployed labor and resultant supply elasticities in

determining the extent of real changes in national income

225
caused by regional changes in aggregate demand.. Further-
more, the team viewed changes in regional employment which
drew labor from outside the area as having possible dele-
terious effects upon other regions. They believed that
net secondary benefits could occur nationally as well as
locally, however, if unemployment existed in the project
area. It was the viewpoint of the team that regional in-
come considerations should prevail, since the Water Re-
source Council had allegedly made population re-dispersion
an overt objective of the water resources program. Thus,
secondary income effects at the local level should be the
factor evaluated. This, in effect, constituted an implicit
assumption that the marginal utility of income in the proj-

ect region was greater than that of the nation as a whole.

Soil Conservation Service Test Team

The Soil Conservation Service Test Team (SCSP Test
Team) and the other teams which tested the proposed pro-
cedures on the Poteau Watershed Project had some advantages
over the other teams involved in the testing in dealing
with the "expansionary" effects of the project. The
poultry processing plant which was a major direct bene-
ficiary of project water had already begun to receive the
increased water supply and had expanded operations as a
result. Thus, "The Test Team found the data available
exceeded that which would normally be available. For exam-

ple, there was an existing industry having unused capacity

226

within the project and it was obvious that lack of water was
the bottleneck for expansion. In addition data was avail-
able for effects after the bottleneck was eliminated."15
As a result of these factors, the SCSP Test Team was able
to estimate, with an unusual amount of certainty, the ef-
fects of the project upon employment via the elimination
of a production bottleneck.

The case of a food processing plant provides an
excellent example of the bottleneck thesis as it relates
to the water resource program. In the processing of most
foods, water is a strategic factor of production. Its
presence or absence strongly influences the location of
food processing plants, and its degree of abundance limits
the extent to which those plants can expand. This indus-
try is somewhat unusual in its degree of dependence upon
water. The "bottleneck clause" in the Task Force Report
is most strongly supported by the food processing example.

The case of the poultry processing plant fits the
bottleneck thesis so well that both the original evaluation
of the Poteau Watershed Project and the SCSP Test Team
Report considered the expansionary effects upon the wages,
salaries and profits from the poultry processing plant as
primary (direct) project benefits. These effects were
classified as "redevelopment benefits" and were grouped with
the wages paid to otherwise-unemployed individuals who were
employed as a result of construction, operation and main-

tenance of the project. This group of benefits, in essence,

227
was composed of the "first round effects" of the project
upon local unemployed.

The second and later round effects of the project upon
exogenous demand were classified as secondary benefits and
were treated separately from the "redevelopment benefits."
The secondary benefits consisted of the multiplier effects
of the increases in exogenous income to the region result-
ing directly from the project. The exogenous changes were
expected to come from 1) per diem expenditures in the
region by project construction workers whose primary resi-
dence was outside of the region, 2) expenditures by visitors
from outside of the region in using the recreation facili-
ties provided by the project and, 3) the increased income
realized from the change in wages and salaries from the
poultry processing operation. The increased profits from
the poultry processing plant were included in the national
account but not in the regional account, since the owners
of the plant resided outside of the region.

In evaluating the "expansion benefits" from the above-
described exogenous changes in regional demand, the SCS Team
selected a multiplier which was based on data prepared by
Clinton Russell, Agricultural Economist, SCS, Little Rock,
Arkansas. The multiplier had a value of 3.85, which the
SCSP Test Team concluded "it appeared that for this portion
of Arkansas, the multiplier should be . . ." However, since
"a leakage of 80 per cent was considered realistic for this

region," the multiplier used in the regional account was

228

lowered to 1.57. The full value of 3.85 was considered to
be the relevant multiplier for evaluating changes in the
national account, since ". . . the leakage would have a
stimulating effect nationally." However, "in the absence
of knowledge concerning the gains that could have been made
had project resources been used elsewhere," the test team
could not make a realistic appraisal of the effect.16

The "redevelopment benefits" of the project were
estimated to have the potential effect of employing all
idle labor in the region. The SCSP Test Team recognized
that no "slack" existed which would allow the multiplier

to create net national gains in the region. Thus, the

multiplier effects were de facto restricted to the re-

 

gional account, though "expansion benefits" were displayed
in brackets in the national account without being included
in the benefit-cost ratio.

In evaluating the employment effects of the project,
the SCS Team assumed that the unemployment rate in the area
would have declined to the national average in 20 years
in the absence of the project. Thus, the "redeveIOpment
benefits" were restricted to the first 20 years of the
life of the project. There is implicit in this assumption
a "belief" regarding the manner in which the inter-regional
reallocation of resources occurs and a "judgement" regard-
ing the amount of lag in the market reallocation process
that society should tolerate. The SCSP Test Team appar-

ently believed that there existed frictions (though not

229

absolute barriers) to the inter-regional flow of resources
and that in this case the frictions were such that a 20
year lag would be required for an "efficient" allocation
to be achieved. The consideration of redevelopment and
secondary benefits indicated that the SCSP Test Team felt
that 20 years was too long, in view of the distributional
and equity considerations involved. This does not mean,
however, that they did not agree with Eckstein and others
that price (wage) differentials are the prime movers of a
market economy. There is a normative question involved
which regards the trade off between maintaining incentives
to better allocation of resources and re-dressing the dis-
tributive (pecuniary) effects of a market which does not
re-allocate resources instantaneously and along a continu-
ous function of alternatives, as marginalist theory depicts
the process.

In evaluating the expansion benefits in the region
from the wages received by those who would be employed as a
result of the project, the SCSP Test Team subtracted the
unemployment compensation payments from the new wages and
applied the multiplier to the differential. This was done
in recognition of the fact that the unemployment payments
constituted exogenous income which would be lost as a result
of the employment. More importantly to the present study,
however, the deduction for the income foregone in this
"first round transaction" constituted a step towards the

same deduction for ensuing rounds.

230

In evaluating the expansion benefits from exogenous
expenditure for recreation, the SCS Team assumed that each
visitor from outside of the region would spend an average
of $5.00 per day in the region. It was assumed that the
profit rate on these expenditures would be about 20
percent. The multiplier was applied to the profits which
accrued to the exogenous recreation expenditures.

By applying the multiplier to only the profits from
the recreation expenditures, the SCS Test Team eliminated
the portion of secondary effects which has been called
"induced-by" effects by practitioners in the water re-
sources field. As was pointed out in Chapter IV, there is
a semantic problem involved in the nomenclature of the dif-
ferent multiplier effects. In the water resource field,
"induced-by" effects are discussed as those increases in
income which occur secondarily as a result of expenditures
for inputs required to meet a specified level of production
(either assuming or due to a specified level of demand for
that production). However, in the Regional Science field
and in Macro-economics literature, "induced" effects are
discussed as those secondary income effects which result
from the respending of income (in the form of wages,
salaries and profits) through the household sector. In
essence, there exist two practically identical names for
the two separate components of the first round dollar. The
water resource field's nomenclature applies to the "cost-

of-production" component of the exogenous expenditure,

231

while the Regional Science-Macro-economics nomenclature
applies to the household re-spending component of incomes
earned. In the report, the SCSP Test Team actually used
the "household induced" component of the expenditure in
applying the multiplier to the recreation expenditure.

The convention used by the SCSP Test Team in evalua-
ting the secondary benefits from recreation warrants a
return to the earlier discussion in the present study re-
garding what the multiplier actually is and what it should
be used to multiply. In the SCSP Test Team report a multi-
plier was adopted which looked to the SCSP Test Team mem-
bers like what the multiplier for the area "should be."
There was no discussion in their report regarding the kind
of multiplier which had been accepted. If the multiplier
actually adOpted was a Keynesian (i.e., "household in-
duced") multiplier, then the application of that multiplier
to the profit component was perhaps the proper application.

The adjustment of the multiplier for "leakages" from
the region provides an additional point for discussion re-
garding the validity of the multiplier. In essence, "the
multiplier" provides an estimate of the total volume of
income over time that a given exogenous change in demand
can sustain. Because of the process by which costs to one
factor become returns to another factor, every dollar which
is respent becomes income to someone. Thus, the limit to
the multiplier is set by the leakages from the stream of

respending. These leakages can occur through hoarding of

232
income by parties to the cycle or by "exporting" part of
the respending -- i.e., by importing goods and services from
outside of the system. Frictions which merely delay the re-
spending do not effect the final value of the multiplier,
since the process is infinite and timeless (i.e., there is
no discounting of the income stream). In essence, all of
the income effects are assumed to occur instantaneously,
as a result. The multiplier which the SCSP Test Team
adOpted and then adjusted for "leakage" must have already
been adjusted for some degree of leakage; otherwise its
value would have been infinite. The degree to which it
had already been adjusted depends upon the kime of multi-
plier which was estimated in the first place. In reality,
the leakages which occur in an economic system (regional or
otherwise) are very difficult to estimate. Furthermore,
the difference that "frictions" which delay responding
actually make when the element of time is considered can
constitute a substantial "quasi-leakage" to the stream.
Thus, conventions such as excluding households from the
endogenous matrix or other such methods are used to place
a limit on the multiplier which is not otherwise provided
in a secondary impact study. Often these adjustments offer
very rough representation of the recognition that this
problem exists. All of this again points up the question
voiced in Chapter IV concerning exactly what it is that the

multiplier shows. The Haberler argument that the multiplier

233

represents merely a mathematical tautology is relevant to
such a discussion.

The SCSP Test Team also included the evaluation of
an "induced-by" component in estimating the secondary bene-
fits of the project. In this instance, they did not use
the multiplier to estimate the secondary income effects.
Rather, they used ". . . percentage estimates of primary
flood prevention and recreation benefits and of the effects
of increased costs of producing added goods as induced by
the project." They used a figure of 10 per cent to reflect
the profits realized at other stages of the production
process. The practice in this regard was very similar to
that used by the Bureau of Reclamation in estimating
"induced-by" benefits. The SCSP Test Team felt that this
was an acceptable procedure under the circumstances, since
"tracing out the income flows through the economy" would
have been ". . . a laborious process and the study was
considered too costly for a comparatively small watershed

project."17

In essence, they assumed that the "output
multipliers" for the sectors affected averaged 1.10.

In summary, the SCSP Test Team enjoyed the luxury of
some hindsight in their test. They tested the new pro-
cedures on a project which provided a classic case of the
water service bottleneck thesis. There was substantial
evidence that unemployment existed and that the effects of

the project upon the unemployed was beneficial. Since,

there had been "slack" in the local economy, there were

234

opportunities for net national secondary effects which
at the same time could help meet regional growth objectives.

The SCSP Test Team made some explicit and implicit
judgements about how resource allocation occurs. They be-
lieved that the process would require 20 years to rectify
the area's unemployment problem, and they judged that 20
years was too long to wait. They used multiplier analysis
to estimate the effects of the project in speeding up this
readjustment process. The multiplier used appeared to be
a Keynesian "induced" multiplier, but no attempt was made
to explain its form or its appropriateness. An adjustment
was made in the multiplier to account for "leakages" of the
effects to outside of the project area. While some adjust-
ment of the multiplier which was accepted was perhaps in-
dicated by its magnitude, the rule-of-thumb adjustment of
an already uncertain figure regenerated earlier questions

regarding the meaning and validity of such analysis.

Mountain Home Project

 

The Mountain Home Project of the Bureau of Reclamation
is primarily an irrigation project. The key feature of the
project is the reclamation of the drylands of the Mountain
Home Desert, though the project includes other objectives
such as electric power production, water supply, flood
control, and fish and wildlife enhancement. The primary
secondary impact is expected to develop from the irrigation

effects, since much of the electric power generated by the

235
project will be used for pumping irrigation water and since
the lake is not expected to have a large recreation impact.

Project construction cost is estimated at $153,700,000.
Corps of Engineers Test Team

The Corps of Engineers (CEMH) Test Team which tested
the proposed procedures on the Mountain Home Project felt
that "no particular problems were raised by the regional
develOpment benefit analysis." The team "used income as
the measure of the benefits." However, they felt that
"more study would be required on multiplier effects of
direct expenditures for general application in project

. "l8
analySis.

The team did nevertheless include multiplier
effects in their report.

The CEMH Test Team believed that only the income
realized by those persons who would have been unemployed
in the absence of the project should be included as net
national secondary benefits. Their report stated that
"if in the absence of the project, comparable employment
would exist elsewhere, no labor resource utilization credit

19 While not elaborating upon this

nationally is realized."
statement, the team appeared to express a viewpoint often
encountered in the theoretical literature, but not often
expressed by Federal water resource agency personnel. The
viewpoint holds that the aggregate demand multiplier ef-

fects created by the project are not net to the nation, if

the funds to build the project were taxed away from other

236
uses. Thus, the project impacts are not necessarily net
national gains, ". . . since equivalent developments un-
doubtedly would take place elsewhere in the nation in the

20

absence of the project." "If in the absence of the proj-

ect comparable employment would exist elsewhere, no labor
resource utilization credit nationally is realized."21
However, the team felt that it was impossible to determine
the net employment differences caused by the project. Thus,
they felt that nothing could be said with any degree of
certainty regarding the net national secondary effects of
the project. Nevertheless, the team did include national
secondary benefits in their evaluation. They assumed that
20 per cent of those employed by the project would have

been otherwise unemployed "anywhere."

The Corps Team was insightful in pointing out other
"secondary benefits" than those of aggregate demand stimu-
lation. These additional benefits consisted of more
stable employment, more diversified economic base, in-
creased investment and an expansion in job opportunities.
The latter two are normally associated with secondary bene-
fits and warrant little further discussion. The former
two, however, constitute something Of a departure from the
general categories of consideration in secondary benefit
evaluation. In fact, to a large degree they must be played
off against multiplier effects in decisions regarding the
prOper development strategy. In general, the local economy

which exhibits the greatest input-output linkages (and,

237
hence the highest local inter-industry multipliers) also
exhibits the least diversified economic base. This depends,
of course, upon the kind of multiplier adopted in the analy-
sis. However, the multipliers which allow the greatest
disaggregation and have the greatest potential of accurate
representation are the input-output multipliers. These
clearly tend to force a trade off between diversification
and multiplier potential. The CEMH Test Team did not
elaborate upon this aspect of the additional group of
secondary benefits, however.

The CEMH Test Team went furthest in the evaluation of
regional develOpment benefits. This was logical, however,
since the team believed that "all national income benefits
generally accrue to the region in which the project is
located or the area of influence for which it is developed."
Thus, the benefits in the regional account included all
of those in the national account. In addition, the re-
gional account included secondary project effects which
exceeded those of the nation as a whole. This was true
because of the convention adopted by the team whereby only
20 per cent of the employment stimulated in the region was
assumed to have been otherwise unemployed "anywhere." In
effect, the assumption was made that the real income multi-
plier in the project region was greater than that in other
adversely effected areas, presumably the rest of the
nation. Given the economic base of the project region,

it is tenuous to assume that the local multiplier exceeded

238
by the amount postulated the average of those for other
localities in the nation.

The local income multiplier effects were based upon
the direct benefits from irrigation, recreation, and fish
and wildlife. The secondary benefits from irrigation were
estimated using a local income multiplier developed in a
research project at Washington State University on the
Columbia Basin Project. The multiplier expressed a rela-
tionship between income generated "in the secondary sector
(project stemming from effects)" and income generated "in
the tertiary sector (project induced by effects)," and the
number of acres irrigated. The team believed the multiplier
transplant to be justifiable because of similarities be-
tween the two areas in terms of economic base, climate,
irrigated area, population density, regional markets, and
transportation facilities. However, since the elevation
of the project areas was 1,500 feet greater than the
Columbia Basin area, the multiplier was adjusted downward
'by 10 per cent. Whether or not the adjustment was appro-
priate is a question which could not be answered without a
great deal of research. Thus, little more can be said
herein than to comment that the transplant has all of the
inherent problems discussed in Chapter II above, and that
the amount of study required to adequately justify and
adjust a proper multiplier transplant is probably not
greatly different from that required to construct a local

multiplier firstjhand.

239

In estimating the secondary benefits from recreation
and from fish and wildlife develOpment, a different ap-
proach was used. The team developed a multiplier from data
on expenditures by visitors from outside of the project
areas of several projects. The data which were used re-
lated to average expenditures per visitor per day and the
proportion of that expenditure which accrued to the project
area. Total expenditures were obtained by multiplying per
person expenditures by the estimated number of visitors
from outside the project area expected each year.

The multiplier was developed using the generalized
Keynesian form and was somewhat analogous to the simple
economic base multiplier, the numerator consisting of the
proportion of total expenditures by outsiders which accrue
locally and the denominator consisting of one minus the
proportion of total expenditures by outsiders which
accrue locally. The difficulty, however, is that both the
numerator and denominator are badly mis-specified. The
team explicitly assumes that the propensity to consume
locally out of local income is identical to the proportion
of visitation expenditures spent locally. Obviously, there

is no a priori reason to assume this to be the case, and

 

there is no basis in theory for linking the two ratios.
Actually, what has been done is to define an income flow
from the recreation user to the local area and back to the
recreation user in the same proportion, then back to the

local area again, etc., ad infinitum. Neither of the above

 

240
two cases is very realistic. The former would provide the
more acceptable alternative, if it could be demonstrated
empirically that the marginal propencity to consume 10-
cally were equal to the proportion of visitation eXpendi-
tures accruing locally.

An additional point which can be made at this juncture
relates to the multiplicand of the multiplier generated by
the Corps Team. The expenditure proportions which are used
relate to gross income to those who provide the goods and
services to the visitors. Thus, notwithstanding the mis-
specification of ensuing rounds in the team's multiplier,
the first round of local income accrual is mis-specified.

In evaluating the direct benefits, the team was very
diligent in attempting to determine the real costs and
returns to fixed factors which would be employed as a re-
sult of the project. The distinction between regional and
national benefits forced them to effectuate certain assump-
tions and beliefs regarding regional and inter-regional
resource mobilities. These same assumptions and beliefs
then came into play as the team evaluated the secondary
impacts of the project.

The CEMH Test Team felt that a reasonable return to
the farm operators' labor should be included as a cost item
in the farm budgets (contrary to Bureau of Reclamation
Practice). They believed that to do otherwise was to as-
Stune that, in the absence of the project, the farmer would

lulve been totally unemployed. To adjust for this omission,

241

they adjusted the net farm income downward by one half.
This reflected their belief that the increased irrigation
water would cause a better utilization of the farmer Opera-
tors managerial ability. The productivity which would be
lost was a cost of the project. However, the team included
it as a cost only in accounting for the national income
benefits. In the regional account the alternative cost of
the farm operators, labor was not included. ‘The differ-
entiation between the two accounts is justifiable only under
certain conditions. If the farm operator in question were
employed at a lower level in some other part of the country
and moved to a project farm and realized a higher income,
then the team's convention was perhaps correct. (Other
similar resource movements could have the same net effect).
Only under such restrictive conditions would the alternative
costs not be felt in the project region as well. Consider-
ing the migration pattern for the region, the direction
of resource flow required to make the team's accounting
correct is not the flow which one would normally predict.

Return to farm capital was treated in similar fashion
to farm Operator's labor. It was deducted as a production
cost in the national account, while it was treated as a
‘benefit in the regional account. The only conditions which
might justify this distinction would be a zero rate of
return on capital in all non-project alternatives in the
IYBgion, while the rate of return in the rest of the nation

“His equal to the rate of return on capital invested in

242
project farms. Such a situation is unlikely. Ordinarily
one would include normal returns to capital as a benefit
only when the capital was already sunk and not transferable.
Then any benefits to its employment should be included in
both accounts.

The treatment of farm labor income was analogous to
that of farm Operator's income. Fifteen per cent of farm
labor income was assumed to go to unemployed'farm labor
and hence was treated as a benefit in the national account,
while all of the wages were included in the regional
account as benefits. One possible justification for this
distinction would be a belief or an assumption that 85 per
cent of those laborers would be mobile between regions
(during the period of project effectiveness) and would
migrate to the jobs which would have existed in other re-
gions in the absence of the project. This, obviously,
would constitute a "slippage" of 15 per cent in the resource
adjustment process. Thus, the team seems to question the
instantaneous adjustment process of micro-economic theory.

The team also evaluated the secondary income created
by the production of electric power. They used multipliers
which related increased output in electro-process industries
to income induced in the local economy by that output. This
was used in conjunction with estimates of output per kilo-
watt for aluminum and other electro-process industries to
estimate the induced income from the electrical output of

the project. The derivation of the multipliers apparently

243
used an input-output type model. The multipliers were
taken from a preliminary study done at the University of
Washington.

In addition to the regional definition problems in-
volved in determining the relevant regional multipliers in
this case, there are the problems of forecasting the users
of the output within that region once the region is defined.
Whether or not the multipliers which were accepted ade-
quately represent the probable pattern of secondary in-
come accrual cannot be answered without a great deal of
research. Again, the multiplier transplant derived from
other areas is a problem. Perhaps of potentially equal im-
portance, however, is the fact that the multipliers were
taken from a preliminary draft of a study which was re-
leased for comment only. It is an ever-present danger of
such transplantation that the adaptor does not fully know
or appreciate the rule-of-thumb adjustments which went into
a model, the range and conditions of applicability, and
the sensitivity of the individual components to slight
differences in industrial composition.

In summary, the CEMH Test Team generally held an
appreciation for the aggregate supply aspects involved in
income multiplication from the national viewpoint. However,
they did not apply the same considerations to the analysis
of regional secondary benefits. Very restrictive and often
untenable assumptions were necessary to rationalize this

differentiation between the two accounts. The team

244
apparently believed that resource mobility at the national
level was at least partially restricted for time periods
conforming to the impact period of the project. They be-
lieved that because of the interation between project ex-
penditures, project funding, and distributional effects of
project outputs, the net secondary effects at the national
level were unknowable. In estimating the regional multi-
plier effects, the team used transplanted multipliers quite
liberally. In one case, the degree of refinement of the
transplant multiplier was not yet acceptable to its devel-
oper. Additionally, several of the multipliers used dealt
with gross income generated rather than net income. In
all, despite the quickness herein to point out their fail-
ures, the team did a commendable job on many counts. They
were at times even insightful, especially in dealing with
the alternative costs of the labor and capital in the
national account and in pointing out the often overlooked
aspects of industrial diversification and employment

stability.

University of Wisconsin Test Team

The University of Wisconsin (UWMH) Test Team report22

dealt largely with the development of a planning-decision
model for choosing among public water resource investments.
An integral part of that model had to do with the secondary
income effects of the project. Though these effects were

not treated as extensively as in some other test team

245
reports, some aspects of the treatment presented by the
UWMH Test Team constituted significant contributions in
dealing with the evaluation of secondary impacts.

Perhaps the most important contribution of the UWMH
Test Team in regard to secondary benefits was their ap-
proach to inter-regional income effects of the project.
Their desire to incorporate this effect into an explicit
decision model led them to attempt to conceptualize an
operational approach to the measurement of the project im-
pacts both in the project region and in other regions of
the country. In doing so, they divided the country into
three regions:

1) The project region - i.e. the state of Idaho,

2) the contiguous region, defined as the states of

Washington and Oregon, and

3) the rest of the nation.

They then attempted to evaluate the primary project impacts
upon the three regions of the country.

The principal primary effect evaluated was that
caused by the increased availability of irrigation water.
It was assumed in the analysis that adjustments would be
made by project area farmers which would increase the
average size of farms (and thus the efficiency of the
operation). This process would at the same time open up
new farming opportunities in the project area, which would
attract new farmers to the area. The result would be an

increase in input purchases in the area, an increase in

246
agricultural income, and an increase in raw materials for
agricultural processing and marketing firms. All of these
would have secondary effects upon other producers in the
project area.

At the same time, however, as the UWMH Test Team
pointed out, the new farmers necessary to take advantage
of the new farming Opportunities provided by the project
must come from somewhere. To illustrate their suggested
approach, the team arbitrarily assumed that half of the new
project farmers would come from the contiguous region and
that the other half would come from the rest of the nation.
In their discussion, they emphasized that the provision of
the new farmers was not costless to the economy and that
the effects upon the economy of the region from which they
would come should also be evaluated. The effects envi-
sioned would be of the same nature as those evaluated for
the project region, except the effects realized by the other
regions would be negative. Thus, declines outside of the
region in farm income, agricultural supply sales and agri-
cultural processing opportunities should be evaluated for
the other regions just as they were being evaluated for
the project region.

The UWMH Test Team also pointed out that those people
who would be attracted to project farm land from within the
region should not be considered as costless resources
either. The team felt that deductions from new farm in-

come should be made to reflect the income given up in

247
other employment by these new farmers. In the case of both
these farmers and those attracted from outside of the
region, the principle was presented that the alternative
cost of the new income should be considered in appraising
the net changes in income.

While the UWMH Test Team Report made some very valua-
ble contributions on other accounts, their most important
contribution to secondary benefit evaluation lay in their
insistence upon the inclusion of the opportunity cost of
the expansion effects. They discussed these opportunity
costs in two contexts: l) The regional economic costs
in the form of income foregone in other regions, and 2)
the personal economic costs in the form of the income
foregone in other employment (in some cases in another
region).

In evaluating the expansion effects of the project,
the UWMH Test Team traced through only the first round of

eexpenditures using a very aggregated industrial classifica-
tzion scheme (the team's approach was similar in many ways
‘CID the Rent Model). The degree of aggregation accepted was
£5€3emingly an item of convenience, since - as the team
.F’C>inted out - they were interested only in demonstrating
ti}1<e principles involved. The reasons for which the team
‘liinnnited the expansion effects to one round were not appar-
6’11 t: in their report. However, members of the team, in
I3€3i1‘sonal conversation with the author, indicated that they

(i:i*<i- not believe that full credit for all ensuing expansions

248
should be credited to the project. Thus, since direct
credit for the effects became increasingly difficult to
assign, they chose to make the arbitrary cut-off after one
round, feeling that to be the most defensible course of
action.

In estimating the secondary expansions in income, the
UWMH Test Team used an input-output type format in alloca-
ting the change in farm income to respending in other
sectors. It was assumed that net farm income changes
would be re-spent in fixed proportions. By the same token,
it was assumed that farm production expenses would exhibit
the same sort of linearity and proportionality as output
expanded. On the basis of these assumptions, the increase
in farm receipts was allocated among other sectors.

Though the UWMH Test Team emphasized the desirability
of including in the analysis of secondary effects the
opportunity costs of the factors used in the project region,
no systematic attempt was made to include those considera-
tions in the analysis. A bid was made by the team for the
provision of information regarding the present occupation
and earnings of workers attracted by the project, as well
as information regarding their previous residence. Addi-
tionally, the team emphasized that information on the
nature and magnitude of expected new jobs in associated
industries would be helpful.

In summary, the UWMH Test Team made substantial con-

ceptual contributions to the evaluation of secondary

249
benefits. Their emphasis of the inter-regional dimension
and of the alternative cost dimension were two of the most
important contributions. However, the team itself failed
to adequately demonstrate in their appraisal example how
the inclusion of the latter might be achieved.

Effects of Differing Approaches on Secondary Benefit
Estimates

 

In the above sections of this chapter, the differ-
ences in beliefs, assumptions and outlooks of different
analysts were discussed. The different concepts and
approaches of the different analysts were outlined. In the
section which follows, the effects of these differences in
concept and approach upon the magnitude of the resulting
secondary benefit estimates will be discussed. The dis-
cussion must of necessity deal with the aggregate effects
of differing approaches, since it is impossible to separate
out the effect of each differing assumption.

For purposes of comparison, the Rent Model was applied
to each of the projects discussed above. The results of
these applications are presented in Appendix Tables 1-22.
There are many problems with the manner in which the model
was applied. The primary benefits to which the model was
applied were those discussed by the respective test teams.
These benefits had to be accepted at face value. Addi-
tionally, because of limitations of time and research

funds, secondary sources had to be relied upon. Because

250
of the small size of the employment base of some of the
regions, differing degrees and kinds of aggregation had to
be resorted to due to disclosure problems in the
secondary data sources. The application of the model to
the other projects was largely mechanical. In this re-
spect, it is probably not vastly different from many of the
other analyses, nor from the application one might expect
to occur too often in agency practice. Thus, while the
author is reticent to vigorously defend the numbers
generated, the numbers serve a heuristic function in the
following comparison of the numerical results yielded by
the several analyses.

The effects of differences in analytical approaches
upon the resulting secondary benefit estimates are re-
flected in a striking manner in Table 6. In the case of
all four projects, the range of estimates between the
lowest and the highest benefit estimates can be measured
in multiples of the lowest estimate. For example, the
Regional Secondary Benefits from the Trenton Channel
Project which were estimated by the Corps of Engineers Test
Team are approximately eighty times larger than those esti-
mated in the other two analyses of the same project. The
pattern is the same for each of the other projects. In
all cases, the range of estimates is substantial. In most
cases, the Rent Model gives the lowest secondary benefit
estimate. It might be noted that the Corps of Engineers

Test Team which reappraised the Trenton Channel Project

251
presented both the highest and lowest secondary benefit
estimates among the twelve appraisals: $135,000,000 in
Regional Secondary Benefits and zero in National Secondary
Benefits. This striking dichotemy largely resulted from
two factors: 1) The test team's inclusion of gross income
as regional benefits, coupled with the acceptance of very
large regional income multipliers, and 2) the test team's
belief that no net national benefits would reSult in the
absence of local unemployed labor (i.e., migrants who came
in to take the newly-created jobs earned no wage rents;
all local secondary benefits were income transfers from
other regions.) The Corps of Engineers test team ap-
praisal of the Trenton Channel Project showed a National
Secondary Benefit estimate of zero. Similar patterns
emerge in the appraisals of the other three projects, the
exception being the case of the Stonewall Jackson Project
appraisal. The Corps of Engineers Test Team which ap-
praised the Stonewall Jackson Project estimated that
Regional Secondary benefits would be approximately ten
times as great as those accruing to the project region.
The Cornell University Test Team in its appraisal of the
same project also estimated that Regional Secondary Bene-
fits would be approximately ten times as great as those
accruing locally. However, the actual magnitude of the two
sets of estimates differed greatly. The very large dif-
ference in the estimates arrived at by these two test

teams can be basically traced to the following factors:

252
The Corps of Engineers Testheam included the provision of
flood-free industrial sites as secondary benefits; the

potential income from the full develOpment of these sites

 

provided the measure of the secondary benefits; it was
implicitly assumed that the provision of flood-free sites
was sufficient to cause plants to locate in the area, and
all of the credit for such location was given to the
project. The Cornell University Test Team, On the other
hand, did not believe that the provision of water supply
nor flood-free sites were sufficient to induce plant loca-
tion. Furthermore, the test team felt that so many other
factors were involved that it would be erroneous to give
the credit to the water project. Thus, no secondary bene-
fits from plant location were assigned to the project by
the Cornell University Team. However, the Cornell Univer-
sity Test Team did credit the project with indirect bene-
fits from recreation development.

The lowest relative variance occurred in the ap-
praisal of the secondary benefits of the Poteau Project,
which was the smallest and perhaps simplest to appraise of
the four projects. Still, the highest estimate of the
Regional Secondary Benefits was almost eight times as
great as the lowest estimate and four times as great as the
other estimate. Much of the difference between the two
higher estimates was explained by the difference in the
size of the income multipliers applied by the two teams.

The difference between the two multipliers related largely

253
to differing assumptions (beliefs) regarding the magnitude
of "leakages" from the multiplier. Both test teams con-
sidered the opportunity cost of the effected factors to be
zero.

The Mountain Home Project ranked third in terms of
the magnitude of the variance of the secondary benefit
estimates. Nevertheless, the highest Regional Secondary
Benefit estimate was approximately thirty-fiVe times larger
than the lowest estimate and twelve times larger than the
other estimate. The difference in estimates between the
three appraisals was in part due to the differences in the
number of rounds of expansion included in the analysis.

The Corps of Engineers Test Team used an infinite-rounds
multiplier, while the other two appraisals included only
the first round of expansion. Additionally, the Corps of
Engineers Test Team assumed the opportunity cost of local
labor to be zero, while the other two appraisals did not.

It is interesting to note that the two test teams
which voiced the greatest skepticism regarding the appraisal
of secondary benefits provided the highest estimates of
secondary benefits among the twelve appraisals. The Corps
of Engineers Team which appraised the Trenton Channel Proj-
ect was particularly skeptical of the possibility of net
national secondary benefits and demonstrated its skepticism
by assigning no secondary benefits to the National Account.
However, the team's Regional Secondary Benefit estimate was

the highest among the twelve appraisals. The second

254
highest estimate was made by the Corps of Engineers Test
Team which appraised the Stonewall Jackson Project. How-
ever, in the Stonewall Jackson Project appraisal, national
(i.e., net) secondary benefits of $7,080,000 were included

in the National Account by the Corps of Engineers Test Team.
Secondary Benefits Compared to Primarnyenefits

In the immediately preceding section, a striking
variance between secondary benefit estimates of the dif-
ferent test teams was observed, with the differences gener-
ally being measurable in multiples of the lowest estimate.
In Table 7, the primary benefit estimates are juxtaposed
with the secondary benefit estimates of the eight test
teams. It might be observed that, while the highest
primary benefit estimate for each project is not routinely
a large multiple of the lowest as in the case of the second-
ary benefit estimates, there is nevertheless substantial
difference between the primary benefit estimates. The
largest difference occurs in the appraisal of the Mountain
Home Project, in which case the Corps of Engineers Test
Team's estimate of the primary project benefits is many
times larger than the estimate accepted by the University
of Wisconsin Test Team, in both the Regional and National
accounts.

In both the Stonewall Jackson Project appraisals and
the Poteau Project appraisals, the higher estimate of

National Primary Benefits is more than 150 per cent of the

255
TABLE 7

Primary and Secondary Benefit Estimates from Eight
Appraisals: Regional and National Accounts

 

 

 

Regional Account National Account
Primary Secondary Primary Secondary
Benefits Benefits Benefits Benefits
Trenton Channel
Project
MSU Test Team 2,623,000 1,719,000 2,623,000 0
Corps of
Engineers
Test Team 2,755,000 135,000,000 2,755,000 0
Mountain Home
Project
Univ. of Wis.
Test Team 232,000 1,415,000 200,000 2,453,000
Corps of
Engineers

Test Team 13,496,000 16,649,000 10,075,000 2,972,000

Stonewall
Jackson Project
Cornell Univ.
Test Team 1,436,000 55,000 1,436,000 137,000
Corps of
Engineers
Test Team 1,967,000 70,600,000 2,702,000 7,080,000

Poteau Project
Corps of
Engineers
Test Team 1,911,000 1,584,000 790,000 368,000
Soil Conser-
vation
Test Team 504,000 395,000 504,000 614,000

256
estimate. The MSU Test Team did not address itself to
an estimate of gross direct benefits but used the original
Corps of Engineers estimate in its tabular presentations.
(The Michigan State University Test Team differed with the
Corps of Engineers Test Team in their treatment of the costs
of the dock and the dredging of berthing areas by the steel
company, however.)23 A similar pattern of variance is
observable in the analyses of the Regional Primary Benefits.
While the variance in the primary benefit estimates is not
so great as that in the secondary benefit estimates, it
would appear to be sufficient to give credence to the state-
ment by the Michigan State University Test Team that "The
biggest problem in water project evaluation is estimation

24 The

of the value of the water product or service."
Cornell University Test Team reflected similar beliefs in
its discussion of benefits from the creation of flood-free
sites. There is apparently much room for disagreement (and
for improvement) in the appraisal of primary benefits. In
comparing the primary benefit estimates to the respective
secondary benefit estimates, one finds that one-half of the
appraisals the Regional Primary Benefit estimates exceed
the respective Regional Secondary Benefit estimates. The
same proportion applies to the estimated National benefits.
However, the appraisals in which the Regional Primary Bene-
fits exceeded the Regional Secondary Benefits are not the

same appraisals as the four in which National Primary

Benefits were larger than the National Secondary Benefits.

257
Again, the Corps of Engineers appraisal of the Trenton Chan-
nel Project is the most striking, in that the Regional
Secondary Benefits greatly exceed the Regional Primary Bene-
fits, while in the National Account the test team estima-
ted that no secondary benefits would be realized. The
Michigan State University Test Team, on the other hand,
estimated National Secondary Benefits which exceeded their
estimate of National Primary Benefits, while the relation-
ship was the reverse in that test team's Regional Account.
Again, a similar pattern prevailed in the other three sets
of appraisals. In all cases but two in each account, the
Secondary Benefit estimates constituted a substantial pro-

portion of the total benefits estimated.
Regional Versus National Input-Output Coefficients

A further test of the Rent Model was made by compar-
ing the results obtained using national input-output coef-
ficients and applying localization ratios with those ob-
tained using local input—output coefficients. Since no
regional input-output models were known to exist for any
of the test project areas, the comparison was applied to a
hypothetical project in a region for which a regional
input-output model did exist. The region selected was the
Charleston, South Carolina, metropolitan area. The input-
output model used was developed by Laurent and Hite.25

The Poteau Project was used as the hypothetical case, and

the Rent Model was applied to the increased sales of

258
processed poultry. Appendix Tables 23 and 24 indicate the
results obtained from the two approaches. The "regionali-
zed" version of the Rent Model using the Charleston input-
output coefficients yielded an estimate of induced-by wage
gains of $69,510. The application of the Rent Model using
national input-output coefficients yielded an estimate of
induced-by wage gains of $168,591. As in the case of the
differing approaches used by different analysts discussed
above, the results of the two applications differed sub-
stantially. Not only was the overall difference sub-
stantial, the differences between estimates for particular
industries were also quite large. Part of the difference
was due to differing technical coefficients between the
Charleston firms and the average for the nation. However,
a larger part was due to the failure of the localization
ratios to adequately determine the portion of the induced-by
sales which would occur locally. The combination of errors
due to differing technical requirements and those due to
inadequacies in the localization ratios can lead to great
divergence.

The application of the national model is probably
most subject to question when projects in small economic
areas are being analyzed. Projects such as the Mountain
Home Project and the Poteau Project would not appear to
be good candidates for the national model. The Trenton
Channel Project, and the other hand, would not appear to be

an unreasonable application, since the national technical

259
coefficients are probably most closely represented by those
in such large metropolitan areas as Detroit. The localiza-
tion ratios are still subject to question even in such

areas as Detroit, however,

Summary

The effects of differing assumptions and approaches
by the different analysts upon the relative magnitude of
the secondary benefits estimated were substantial. The
difference between the highest and the lowest estimate for
each project could be measured in multiples of the lowest.
Additionally, substantial differences were detected in the
estimates of primary benefits, though the discrepancies
were not as great as in the case of the secondary benefit
estimates. There was little conformity in the relative
magnitudes of the primary and secondary benefit estimates
between the two accounts by the eight test teams.

The results of the application of the Rent Model
based on regional versus national input-output coefficients
were similar to those summarized above. The indirect in-
come estimate using the national model was over twice as

large as that obtained from the regional model.

260

Footnotes

 

l. Schmid and Ward, op cit, 23.
2. Ibid., 15.

3. U. S. Army Corps of Engineers, A Special Study of
the Detroit RiverJ Trenton Channel Project (Detroit: March
1970).

4. Ibid., 17.
5. U. S. Army Corps of Engineers, Ap§pecial Study of

the Stonewall Jackson Lake, West Fork River and Tributaries,
West Virginia (Pittsburgh: MarCH 1970).

 

6. 1219-: p. 15-18.

7. Schneider, op cit.

8. U. S. Army Corps of Engineers, loc cit, 20.
9. Ipid., 23-25.
10. lpid.

ll. Ipid.

12. U. S. Army Corps of Engineers, A Special Study of
the Poteau River Watershed Project (Tulsaz‘iMarch 1970).

13. Ibid., 22.

14. Ibid., 13.

15. U. S. Department of Agriculture Soil Conservation
Service, Report of Testipg Special Task Force Evaluation
Procedures, Water Resource Council, for Poteau River Water-
shed, Scott Counpy, Arkansas andee Flore COunty, Oklahoma,
(December 1969), 25.

16. Ibid., 17.

17. Ibid., 10.

18. U. S. Army Corps of Engineers, A Special Study of

the Mountain Home Division, Southwest Idaho Water Develpp-
ment Prpject, (Boise: March 1970), 2.

 

19. Ibid., 16.
20. Ibid., 32-33.

 

261
21. Ibid., 16.

 

22. Daniel W. Bromley, et a1., Procedures for Evalua-

 

tion of Water and Related Land Resource Projects: An
Analysis ofithe Water Resource Councii‘s Task Force Report
(MadiSon: Aprii 1, 1970).

23. Schmid and Ward, op cit, 8.
24. Ibid., 1.
25. Eugene A. Laurent and James C. Hite, Economic-

Ecologic Analysis in the Charleston Metropolitan Area: An
Input- Output Stug (Clemson: 1971).

CHAPTER VII

SUMMARY AND CONCLUSIONS

Review of the Study Objectives

 

This study began with the overall objective of ration-
alizing the difficulties which might emerge with the addi-
tion of secondary benefit concepts to the appraisal of
Federal water projects. This was pursued by attempting to
fulfill the following set of objectives:
1. Surveying, evaluating, and synthesizing of the
current theory and concepts regarding secondary
benefits;
2. Identifying the elements in objective one which
could be forged into a generally applicable
model which
a) can be applied to all or most Federal water
projects,

b) uses generally available secondary data,

c) makes explicit the assumption and conditions
under which it is operable;

3. Demonstrating and testing the model by applying
it to four test case projects;

4. Comparing the test results of the model with
the results obtained by other test teams and

262

263

identifying the reasons for differing test

results from the several analyses;

5. Evaluating the applicability of the model and

appraising the current status and applicability

of secondary benefit concepts.

Objective number one was met in Chapter II, in which

three general

fied: 1) The

categories of secondary benefits were identi-

customary multiplier variety, 2) economies of

scale and externalities, and 3) dynamic secondary benefits.

It was concluded that only the first, the customary multi-

plier variety, was currently amenable to generalized esti-

mation.

Objective two was met in Chapters III, IV and V. The

relationships
economic rent
analysis were
through which

be applied to

between net changes in factor payments and
theory and their applicability to multiplier
discussed in Chapters III and IV; and a model
economic rent and multiplier concepts could

the estimation of the customary multiplier

variety of secondary benefits was presented in Chapter V.

Objectives three and four were met in Chapters V and

VI in which the model was first demonstrated by applying

it to the Trenton Channel Project and then compared to

other approaches by applying it to four other projects

which had been appraised by other analysts. In Chapter VI,

it was found that the model generally provided lower esti-

mates of multiplier benefits than did the approaches with

which its results were compared. The higher estimates

264
yielded by the other approaches were in most instances
largely due to their evaluating factor Opportunity costs
at zero and assigning the credit for infinite rounds of
income expansion to the water project. Other differences
in assumptions and basic approaches which were suspected Of
creating a differential in estimated benefits between
analysts were identified in Chapter VI as well. Some of
the major differences revolved around different assumptions
regarding inter-regional income leakages and different
assumptions regarding the effect on industrial location
of providing water project-related goods and services. It
was found that large differences in secondary benefit esti-
mates occurred between different analysts appraising the
same project. The differences were Often measurable in
multiples of the lowest estimate. Large differences in
primary benefit estimates were also found, though they were
not of the magnitude Observed among the secondary benefit

estimates.
The Estimation of Secondary Benefits

The Rent Model

 

A major goal of this study, as expressed in Objective
two, was to develOp a secondary benefit model which was
generally applicable to water project appraisal and which
would not necessitate primary data collection for its im-
plementation. The study was only partially successful in

that regard. It is the author's belief that the conceptual

265

basis of the model is correct - i.e., that economic rent
concepts should form the groundwork for the appraisal of
the customary multiplier variety Of secondary benefits.
However, there are many weaknesses in the model. One of
the most important of these is its failure to specify the
location and magnitude of the contractions in the adversely
affected areas and sectors. The assumptions regarding the
Services industry and "marginal firms“ only serve as stop-
gap remedies to this problem. It is possible (perhaps
remotely) that the results Of these assumptions adequately
represent reality. However, this is an empirical question
which can not be satisfactorily answered inductively. Un-
fortunately, the research necessary to locate and measure
these contractions would be at least as extensive as that
involved in constructing the input-output model. The
myriad of supply and demand relationships involved would
appear to require a general equilibrium-type model to trace
them. The closest the economics profession has come to
such a model is the inter-regional input-output model,
which is linearly homogeneous in all sectors, with constant
factor-factor ratios. It does not allow for factor sub-
stitution, which would be implied by the contraction of
some sectors simultaneous to the expansion Of others.

A second shortcoming Of the model in its present form
relates to the localization Of economic activity. It is
not sufficient to simply assume that output increases will

be located according to the current location Of capacity,

266
especially at projected output levels which approach or sur-
pass current capacity. The analyst's knowledge Of the local
economy can be utilized to supplement the model when there
is locally available excess capacity. This would perhaps
be a workable alternative in regional economies which were
fairly simple in structure and in which the project analyst
had worked for several years. Where local excess capacity
did not exist, however, specific studies would be needed
to determine where new capacity would locate. The local
analyst would not generally be able to do such studies.

The failure of the study to locate secondary sources
from which the Opportunity cost of labor could be evaluated
dealt a serious blow to the attempt to implement the eco-
nomic rent concept. The convention of using Service sector
wages was at best on arbitrary proxy with little inductive
basis. Thus, establishing the relevant opportunity costs
is again an empirical question which would necessitate a
major primary research effort or, more realistically, a
change in the local system Of reporting unemployment.

The attempt in Chapter V to identify the major assump-
tions of the model was instructive (at least to the author)
in pointing up the myriad of assumptions involved in im-
plementing such a model. Many more underlying assumptions,
and beliefs doubtlessly went unspecified. Some assumptions,
or exogenous specifications, are necessary to "identify"
any model. These "identifying assumptions" alone are Often

sufficient to pre-specify the outcome Of the analysis. The

267
additional assumptions necessitated by the lack of second-
ary data described above make the results Obtained from the
model even more tenuous. Similar skepticism would appear
to be warranted for the other approaches to estimating
secondary benefits which were discussed in Chapter VI. The
widely divergent results yielded by the different analyses
gave an indication of the effects of differences in the
large matrix Of possible assumptions and beliefs between
analysts. While some Of these differences were noted in
the discussion in Chapter VI, it would be impossible to
identify all Of the differences, since many of the basic
ones are epistemological in nature. Thus, even in the
presence Of a "general model" such as the Rent Model,

differences between analysts would persist.
Recommendations for Further Study

The results Of the comparisons made in Chapters V
and VI indicate that there are major deficiencies in the
Rent Model in its present form. The results of the Rent
Model and of the indirect benefit estimates Of the other
analysts were widely divergent. Much divergence was due
to the differing assumptions underlying the analyses.
However, an unknown quantity of divergence was doubtless
attributable to the shear randomness associated with "esti’
mating" indirect income changes using very imperfect tools

Of analysis. The only real consistency added by the Rent

268

Model was its tendency to provide a lower estimate Of the
"total" income changes expected to occur.

The comparison of results obtained from the Rent Model
based on national coefficients and an application using a
local input-output model provided a further test Of the
localization assumptions of the Rent Model. The difference
in estimates yielded by the two approaches is sufficient
to cast serious doubt upon the inter-regional trade assump-
tions underlying the localization ratios. Measuring the
local component Of the total impact is critical, since con-
tributions to regional development have been given in-
creased importance under the new water project appraisal
standards. Additionally, there is greater agreement among
economists concerning the possibility Of positive net
changes in income at the local level than at the national.
The author does not have sufficient confidence in the esti-
mates Of local impact yielded by the model to recommend
its immediate acceptance as a planning model. Many im-
provements are necessary before the Rent Model will yield
estimates which are vastly more dependable than the other
approaches treated critically in Chapters IV and VI. To
say that the Rent Model consistently yields lower esti-
mates of a set of benefits which are almost always over-

stated is not sufficient to recommend its application.

269

Some Further Refinements Of the Model

 

The original Objective of this study was to present
and test an easily applicable model for estimating indirect
benefits of water projects. An inexpensive approach which
utilized currently-existent data was desired. These ob-
jectives were achieved. However, as is usually the case,
achieving Objectives Of keeping down the cost Of informa-
tion development competes with the Objective of having
complete and accurate information. The former Objectives
were probably met at too great a sacrifice of the latter
objectives. The result is probably some degree of "false
economy." In light Of this conclusion the following re-
finements Of the Rent Model are recommended as areas de-
serving further study.

Construct Local Input—Output Models. Local input-

 

output models should be constructed for projects in which
the indirect effects on income are a priori deemed to be
of importance because of the nature of the project and
the project region. The input-output model need not con-
tain all industrial sectors in the region. Only those
impacted in the first three rounds would probably be
sufficient, since there is no household induced component
in the Rent Model to carry the impacts to the other
sectors. Purchases and sales by each sector sustaining
any significant impact should be taken in detail using
direct questionnaires. By beginning with round one and

tracing purchases and sales, it would be possible to

270

determine whp will be affected and by what amounts. If the
impacts are significant and are discernible by the prospec-
tive recipients, the survey should not be stymied by a
failure to cooperate on the part of the effected parties.
Any reticence should be closely scrutinized in light of
the Compensation Ability Principle discussed in preceding
chapters. There is an Observed reticence to provide input
and output information on the part Of firms Operating in
Oligopolistic industries. Such reticence usually yields in
the face of real Opportunities to increase profits. Firms
operating under different industry structures generally do
not hesitate tO supply such information, even when direct
returns to them do not appear to be great.

The task of building an input-output model by tracing
back from the direct impacts would not be great, except
in the case Of multi-product projects in a highly developed
region. In the case of one-product projects in an under-
developed region, the task would be relatively simple.
The author has undertaken similar studies in conjunction
with World Bank project appraisal missions. The Objective
there was to trace the direct and indirect input require-
ments in order to establish the indirect foreign exchange
requirements of the projects under appraisal. Generally,
in small depressed countries as in small depressed regions,
most of the impacts have "leaked" out to more developed
areas by the end of the third round. Tracing through

three rounds for relatively small projects in

271
underdeveloped countries normally takes less than five man-
days for an experienced economist. A similar procedure
in the United States should not be vastly different in this
regard.

Detail of Job Requirements. Additional refinements

 

can be made in Rent Model by breaking the value added down
into greater detail. By utilizing currently available

data from the Department of Commerce Bureau of Labor
Statistics,1 it is possible to break the production

worker requirements of different sectors down into detailed
job categories. Thus, estimates of indirect demands upon
the work force can be converted into the classification
scheme used by State Employment Security Commission Offices.
Supplementing these data with information on multiple job
shifting which is now beginning to become available hOpe-
fully would allow some Of the assumptions regarding source
and alternative earnings Of Wage Rent receivers to be
accurately determined. Additional research of the type
reported by Chilton Research Services2 would help in esti-
mating the sum Of gains via the syphon effect, or "musical
chairs" as the former have chosen to call the trickle down
phenomenon.

Use Linear Programming to Restrain Gains. By con-

 

verting the job requirements into classifications used by
State Employment Security Commissions, it is possible to
estimate a priori the local "bottlenecks" in labor supply

whichnfight limit the indirect expansion from the project.

272
Supplementing the local employment data with the job trans-
formation possibilities observed in the "musical chairs"
studies would allow a determination of the amount of job
switching which might occur. Combining the information
on unemployment by job classification with that on job
switching would allow one to develop a set of equations
which could be used as ”restraints" in determining how
much expansion might occur locally before 13321 labor be-
comes a severely limiting factor. Each restraint could
be sequentially relaxed to estimate the marginal social
product of aiding in the retraining of local labor in
categories not utilized. The same procedure could also
be used to estimate the in-migration necessary for the
indirect impacts to carry through the full local multiplier
effect. Instead of arbitrarily limiting the rounds of
expansion to one round, the multiplier analysis could be
done using a linear programming format in which expansion
was ”restrained" or stopped on the round in which local
unemployment and job switching possibilities were exhausted
in one category. By sequentially relaxing restraints,
estimates of retraining and/or in-migration needs could be
developed. By comparing the difference in gains between
restrained maxima, the marginal product Of retraining

and/or in-migration could be estimated.

273

Footnotes

 

l. Manpower Administration, Occupational Employment
Patterns for 1960 and 1975 (Washington: 1968).

 

 

2. Chilton Research Service, Multiple Job Shifts
Associated With Business Loans (Philadelphia: 1970).

 

 

SOURCES CITED

SOURCES CITED

Ackley, Gardner. Macroeconomic Theory. New York: The
Macmillan Company. i967.

 

Adelman, Irma. Theories of Economic Growth and Development.
Stanford: Stanford University Press. 1961.

Andrews, Richard B. "Comment re: Criticisms Of the
Economic Base Theory." Journal Of the American
Institute of Planners (1958).

 

Barkley, Paul W. and Thaine H. Allison, Jr. "Economic

Studies in Resource Administration." Land Economics
44:470-479 (November 1968).

 

Blumenfield, Hans. "The Economic Base Of the Metropolis."
Journal of the American Institute Of Planners (1955.

Bromley, Daniel. "On the Use and Overuse Of Input-Output
Models: A Methodological Hypothesis." Unpublished
paper. Undated.

Bromley, Daniel W., William B. Lord, Dinesh Sharma, F. A.
Clarenbach, Fred Bennett, Richard Conover, Marvin Feld-
man, Roger Hartung, John Rushing, Fred Sharnke, Thomas
Slater, Mario Villamizor, Maurice Warner, Bruce Webber,
and Eugene Woock. Procedures for Evaluation of Water
and Related Land Resource Prpjects: An Analysis of the
Water Resource COuncilTs TESk FOrce Report. Madison:
The University Of WiSconsin Center for Resource Policy
Studies and Programs. April 1, 1970.

Bromley, Daniel W., A. Allan Schmid, William B. Lord.
Public Water Resource Project Planning and Evalua-
tion: Impacts,iincidence, and Institutions. Madison:
Center for Resource Policy Studies and Programs.
September 1971.

Brown, E. Cary. "The Personal Income Tax as an Automatic
Stabilizer." Tax Revision Compendium. Washington:
United States Government Printing Office. 1959.

Buchanen, James M. The Public Finances. Homewood:
Richard D. Irwin, Inc. 1965.

274

275

Business Research Center. A Study Of the Economic Impact
of the Hugo Reservoir on Choctaw and PuShmataha
COunties. OkiahOma City: Oklahoma CityiUniversity.
1969

 

 

Chenery, Hollis B., P. G. Clark, and Vera Cao Pinna. The
Structure and Growth of the Italian Econo_y. Rome:
United States Mutuai”Security Agency. 1953.

 

Chilton Research Services. Multiple Job Shifts Associated
with EDA Business Loans. Washington: Department Of
COmmerce. 1970

 

Chipman, J. S. The Theory of Inter-sectoral Mongy Flows
and Income Formation. Baltimore: *The JOhns Hopkins
Press. 1950.

 

 

Clark, John M., Eugene L. Grant, and Maurice M. Kelso.
Report of Panel of Consultants on Secondary or
Indirect Benefits Of Water-Use Prgjects to Michael W.
Straus,,Commissionep, Bureau of Reclamation.

June 26, 1952

Doeksen, Gerald A. and Charles H. Little. An Analysis Of
the Structure Of Oklahoma's Economy by Districts.
Stillwater: ‘Oklahoma State University. l969.

 

 

 

 

Dorfman, Robert. "The Nature and Significance Of Input-
Output." The Review Of Economics and Statistics.
36 (May 1954)}

 

Due, John F. Government Finance: Economics of the Public
Sector. Homewood: Richard D. Irwin, Inc. I968.

Eckstein, Otto. "Federal Expenditure Policy for Economic
Growth." Public Finance and Fiscal Policy: Selected
Readin s. Editedgby J. Scherer and J. A. Papke.
Boston: Houghton Mifflin CO. 1966.

Eckstein, Otto. Water Resource Development: The Economics
Of Project Evaluation. Cambridge: Harvard Univer-
sity Press. 1965.

 

Edwards, Clark. "Resource Fixity, Credit Availability and
Agricultural Organization." Unpublished Ph.D. Thesis.
Michigan State University. 1958.

Ezekial, Mordecai and Karl A. Fox. Methods Of Correlation
and Regression Analysis: Linear and Curvilinear.
New York: JOhn Wiley and SOns. i959.

 

 

Haberler, Gottfried. "Mr. Keynes' Theory of the Multiplier:
A Methodological Criticism." Zeitschrift ffir
NationalOkonomie (1963)

 

 

276

Haveman, Robert and John V. Krutilla. Unemployment, Idle
Capacity and the Evaluation of Public Expenditures:
NatiOnal and Regional Analysis. Baltimore: The
Johns Hopkins Press. 1968.

 

Helmberger, Peter. "O-Efficiency and the Economic
Organization Of Agriculture." Agricultural Organi-
zation in the Modern Industrial Econo_y. Columbus:
Ohio State University. 1969}

 

 

Hirsch, Werner. "Interindustry Relations of a Metropolitan
Area." The Review Of Economics and Statistics. 413
360-369 (NOvember 1959).

 

Hoyt, Homer. "Homer Hoyt on Development of Economic Base
Concept." Land Economics. 30:182-186 (May 1954).

 

Huffschmidt, M. M., John Krutilla, Julius Margolis, with

the assistance of S. A. Marglin. "Standards and
Criteria for Formulating and Evaluating Federal Water
Resource Developments." Report Of Panel Of Consul-

tants to the Bureau Of the Budget. June 30, 1961.

Isard, Walter. Methods Of Regional Analysis: An Intro-
duction to Regional Science. Cambridge: The M.I.T.
Press. 1960.

 

Isard, Walter and Robert E. Kuenne. "The Impact Of Steel
Upon the Greater New York-Philadelphia Industrial
Region." The Review of Economics and Statistics
(November I953).

 

Jansma, J. D. and W. B. Back. Local Secondarnyffects of
Watershed Projects: A Case Study of Rpger Mills
COunty, OklaHOma. ’Stillwater: *OklahOma State
University. 1964.

 

 

Johnson, Glenn L. "A Note on Non-Conventional Inputs and
Conventional Production Functions." A riculture and
Economic Development, ed. by Carl K. 1Cher and5
Lawrence Witt. New York: McGraw Hill Book Company.

 

1964.
Kahn, R. F. "The Relation of Home Investment to Unemploy-
ment." Economic Journal (June 1931).

 

Kalter, R. J. Estimating Local Secondary Impacts of Water-
Based Recreation Investment’USin Interindustry
Analysis. Madison: University of Wlsconsin. 1967.

 

Kalter, Robert J., Larry W. Libby, Robert C. Hinman, David
A. Schultz, Leonard A. Shabmon, and Th-mas H. Stevens.
Federal Evaluation Of Resource Investments: A Case
Study. Ithaca: Cornell UniVersity water Resources

 

277

and Marine Center and Department Of Agricultural
Economics. February 1970.

Keynes, John Maynard. The General Theory of Employment,
Interest and Money. London: The Macmillan Company.
1964.

 

Kneese, Allen V. Water Resources . . . Development and
Use. Kansas City: Federal Reserve Bank of Kansas
City. 1959.

Kuh, Edwin. "Unemployment, Production Functions, and
Effective Demand." Journal of Political Economy 74:
288-48 (June 1966).

Laurent, Eugene A. and James C. Hite. Economic-Ecologrg
Analysis in the Charleston Metropoiitan Region: An
rpput-Ourput Study. Ciemson: water Resources
Research Institute: 1971.

 

Leibenstein, Harvey. "Allocative Efficiency vs. 'X-
Efficiency.'" The American Economic Review
56:392-415 (June‘i966).

Leontief, Wassily. "Quantitative Input-Output Relations in
the Economic System Of the United States." The
Review of Economics and Statistics (August 1936).

 

Leven, Charles L. "Regional Income and Product Accounts:
Construction and Application." Design Of Regional
Accounts. Baltimore: The Johnsihepkins Press.

1961.
Leven, Charles L. "Regional and Interregional Accounts in
Perspective." Re ional Science Association Papers

 

and Proceedings 1 Y127¥44 (1964).

 

Liebhafsky, H. H. The Nature Of Price Theory. Homewood:
The Dorsey Press, Inc. 1963.

McKean, Roland. "Criteria Of Efficiency in Government
Expenditures.” Federal Erpenditures Policies for
Economic Growth and Stabilization. washington:
United States Government Printing Office. 1957.

 

McKean, Roland. Efficienoy in Government Through Systems
Analysis: With Emphasis on water Resources Develpp-
ment. New Yoik: John Wiley and Sons, Inc. 1958.

 

Margolis, Julius. "Secondary Benefits, External Economies,
and the Justification of Public Investment." The
Review Of Economics and Statistics 39:284-91 (I957).

 

278

Marshall, Alfred. Principles of Economics. London: The
Macmillan Company. iS64.

 

Marts, M. E. An Experiment in the Measurement Of the
Indirect Benefits Of Irrigation. Boise: U. S.
Department of the Interior. 1950.

Masters, Stanley H. ”An Interindustry Analysis Of Wages
and Plant Size." The Review Of Economics and
Statistics 51:341-5 (August 1969).

 

 

Miernyk, William H. The Elements Of Input-Output Analysis.
New York: Random HOuse. 1965.

 

Mirakhor, Abbas and Frank Orazen. "Importance of the Farm
Sector to the Economy: A Multiplier Approach.”
American Journal of Agricultural Economics 50:913-921
(November 1968).

Mitchell, Wesley C. "The Problem of Measuring Profits: A
Preliminary Note by Wesley C. Mitchell." Industrial
Profits in the United States, by Ralph C. EpstEin.
New YOik: National Bureau of Economic Research.
1934.

 

Morgenstern, Oskar. On the Accuracy Of Economic Observa-
tions. Princeton: ”Princeton University Press. 1950.

 

Pearson, John E. A Study of the Economic Impact Of Water
Impoundment Through the Development Ofia COmparative
PrOjectiOn MOdel. TeEhnical Report Number 8.

CO ege Stafion: Texas A&M University Water Re
sources Institute. 1967

 

Perloff, Harvey 5., Edgar S. Dunn, Jr., E. E. Lampard,
and R. F. Muth. Regions, Resources and Economic
Growth. Lincoln: University of Nebraska Press.

Ricardo, David. The Principles of Political Economy and
Taxation. Lendon: J} MT Dent and Sons Ltd} 1965.

Richardson, Harry W. Elements Of Re ional Economics.
Baltimore: Penguin Becks. 19 97'

Robinson, Joan. The Economics of Imperfect Competition.
London: TheiMacmillan‘COmpany. 1965.

Scherer, J. and J. A. Papke. Public Finance and Fiscal
Poligy: Selected Readings. iBoston: Houghton
Mifflin CO. 1966.

 

279

Schmid, A. Allan and William Ward. A Test Of Federal Water
Project Evaluation Procedures With Emphasis on
RegiOnal Income and Environmental Quality. East
Lansing: Michigan State UniverSity Department of
Agricultural Economics. April 1970.

 

 

Schneider, Calvin S. Procedures Manual for Determining
PreliminaryrExpansiOn Benefits for Appalachian Water
Projects. Lexington: ‘SpindletOp Research, Inc.

Shaw, Lawrence H. and Robert S. Arden. "Output Effects of
a Changing Composition Of Industry, 1947-1965."
The Review Of Economics and Statistics 45:134-6
(February 1968i.

 

Solow, Robert M. "Technical Change and the Aggregate
Production Function." The Review Of Economics and
Statistics 30:312-20 (1957).

 

Stigler, George J. Capital and Rates Of Return in Manu-
facturingrlndustries. Princeton: IPrinceton
University Press. 1963.

 

Stoevener, H. H. and E. N. Castle. "Input-Output Models
and Benefit-Cost Analysis in Water Resources
Research." Journal Of Farm Economics 47:1572-1579

 

(December 1965).

Strohbehn, Roger, Daniel Piper and Robert Boxley. Anal sis
of Alternative Procedures for the Evaluation of
Agricultural Flood Control Bendfits. VOlume II.
sthington: United States Department of Agriculture.
1971.

 

Struthers, Robert E. The Role Of Irrigation Development
in Community EconomiE Structure: Grand Valley Trade
Area,rCOloradO. Washington: U} S? Department of
the‘InteriOr. 1963.

 

 

Tiebout, Charles M. "An Empirical Regional Input-Output
Projection Model: The State Of Washington." The
Review Of Economics and Statistics. (August 1964).

 

Tiebout, Charles M. "An Inter-sectoral Flow Analysis Of the
California Economy." The Review of Economics and
Statistics 45:409-418 (November 1963).

 

Tiebout, Charles M. The Communiry Economic Base Study.
Supplementary Paper Number 16. New York: Committee
for Economic DevelOpment. 1962.

280

Tiebout, Charles M. "Regional and Interregional Input-
Output Models: An Appraisal." Southern Economic
Journal 24:140-147 (October 1957)}

Tolley, George S. "Impact of Public Resource DevelOpment
on Agricultural Production and Income." Proceedings
of the Agricultural Economics and Rural SocioiOgy
Secfion. Association Of'SOuthern Agriculturai

Workers. 1959.

 

 

United States Army Corps of Engineers. A Special Study
Of the Detroit River, Trenton Channel Project.
DetrOit: U. S. Army Corps Of Engineers. March 1970.

 

 

United States Army Corps of Engineers. A Special Study
of the Mountain Home Division, SoutHWest IdahOiWater
Deveippment Project. Boise: ‘U} S. Army Corps of
Engineers. CMarch 1970.

 

 

United States Army Corps Of Engineers. A Special Study Of
the Poteau River Watershed Project. Tulsa: U. 8.
Army Corps of Engineers. *Marchil970.

 

 

United States Army Corps of Engineers. A Special Study Of
the Stonewall Jackson Lake, West FOiR River add
Tributaries. Pittsburghi' U. 8. Army Corps Of
Engineers. March 1970.

 

 

 

United States Army Corps of Engineers. A Test of Pro-
cedures Proposed by a Task Force of_theiWater Resource
COuncil: Special Study of the MOuntainiHOme Division,
Southwest Idaho Water Development PrOject. BOiSe:

U. S. Army COrps of Engineers. 1970.

 

 

 

 

United States Army Corps of Engineers. Development of
Water Resources in Appalachia. Main Report Part IV
(Drafti. 1968.

 

 

United States Congress. Profits: Report Of Sub-Committee
Of the Joint Committee Of the Economic deort on
Profits Hearings. washington: U1 S. Government
Printing Office. 1949.

 

 

 

United States Department of Agriculture Soil Conservation
Service. Repprt Of Testing Special Task Force
EvaluationiProcedures, water Resource COunciIr for
POteau River WatershedJ'SCOtt County, Arkansas amd:

Le Flore County, Oklahoma. Little ROCk: ‘UT S.
Department Of Agriculture. December 1969.

 

 

 

United States Department Of Commerce. Economic Base Survey
of the Potomac River Basin ServicefArea. “Washington:
UT S. Army COrps Of Engineers. I969.

281

United States Water Resources Council. A Summary_Analysis
of Nineteen Tests of Proposed Evaluation Procedures
on Selected Water add Land ResourceiPrOjects.
washington: United States Water Resources Council
July 1970

 

 

 

United States Water Resources Council. Procedures for
Evaluation of Water and Related Land Resource Projects:
Report to the Water Resources Coumcil by the SpeCial
Task Force. Washington: United States’Water Resources
COuncil. 1969.

 

 

 

 

United States Senate. Policies, Standards and Procedures
in the Formulation, Evaluation, and Review Of Plans
for Use and DevelOpment of Water and Related Land
Resources. Senate Document 97, 87th Congress,
2nd Session.

Ward, William A. "Economic Research: Prologue and Epi-
logue." American Journal of Agricultural Economics
52:136 (February 1970).

 

Waterston, Albert. "Resolving the Three-Horned Planning
Dilemma." Paper presented at meeting Of American
Society of Planning Officials, New Orleans, March 29,
1971.

APPENDIX

282

APPENDIX TABLE 1

Trenton Channel Project -- Sales, Labor Compensation
and Net Wage Gains Induced-by Increased Steel Output:
National and Local

 

 

1 2
Increased Direct
Sales of Requirements

Steel Coefficients

($ million) (Steel)

 

Local Industries1

 

 

Electricity, Gas 6 Water .02994
Printing and Publishing .00007
Trans. & Warehousing .04657
Trade .02994
Communications .00417
Finance, Ins. 5 Real Estate .03090
Services .01589
Construction .00704
Sub-Total
National Industries1
Metal Ores Mifiing .04752
Coal Mining .01908
Non-durable Goods .00929
Lumber 8 Wood Products .00152
Chemicals 8 Allied Prod. .01673
Primary Metals .22035
Fabricated Metals .02362
Stone,C1ay 8 Glass Prod. .00185
Non-Elec. Machinery .01999
Electrical Machinery .00689
Motor Vehicles .00314
Other Transportation Equip. .00005
Instruments .00094
Miscellaneous Mfg. .00127
Sub-Total

Total 40.5

283
APPENDIX TABLE 1 (cont'd)

 

 

 

3 4 5 6 _ 7
Induced Labor Induced Average Annual Job ,
Sales Compensation Wa es Production Equivalents
($) Per Sales $ 3 ( ) Worker Required_
Wage/Year -
1,212,570 .4624 560,692 7,500 74.758
2,835 .1979 561 6,180 .090
1,886,085 .5181 977,181 8,125 120.268
1,192,320 .6152 733,515 5,870 124.959
168,885 .4315 72,874 7,500 9.716
1,251,450 .3850 481,808 6,717 71.729
643,545 .4333 278,848 4,765 58.520
285,120 .3049 86,933 7,417 11.720
1,924,560 .4405 847,769 7,556 112.198
772,740 .4405 340,392 7,556 45.049
376,245 .1800 67,724 6,180 10.958
61.560 .3165 19,484 5,162 3.774
.677,565 .2484 168,307 8,136 20.686
8,924,175 .2714 2,422,021 7,873 307.636
956,610 .3490 333,857 7,141 46.752
74,925 .3454 25,879 6,779 3.817
809,595 .3902 315,904 7,830 40.345
279,045 .3877 108,186 7,071 15.299
127,170 .1865 23,717 8,475 2.798
20,250 .3868 7,833 8,477 .924
38,070 .3814 14,520 7,539 1.925
51,435 .3477 17,884 5,663 3.158

284

APPENDIX TABLE 1 (cont'd)

 

 

 

8 9 10 11
Wage Net Localization Detroit
Rent Per Wa e . 1 Wa e
Job Gain Ratios Gaifis
Equivalent
2,735 204,439 1.0 204,439
1,415 127 1.0 127
3,360 404,100 1.0 404,100
1,105 138,080 1.0 138,080
2,735 26,573 1.0 26,573
1,952 140,015 1.0 140,015
0 0 1.0 0
2,652 31 081 1.0 31 081
9 4, 15 944,415
2,791 313,145 .0016 501
2,791 125,732 .0016 201
1,415 15,506 .0012 19
397 1,498 .0001 0
3,371 69,733 .0192 1,339
3,108 956,133 .0349 33,369
2,376 111,083 .0595 6,609
2,014 11,504 ** 0
3,065 123,657 .0492 6,084
2,306 35,279 .0048 169
3,710 10,381 .2709 2,812
3,712 3,430 .1573 540
2,774 5,340 .0228 122
898 2 836 .0065 18
1,733,737 51,733
24:29:63: ggagigg

 

285

APPENDIX TABLE 2

Trenton Channel Project -- National and Local Profit
Gains Induced-by Increased Steel Production

 

 

l 2
Induced After-tax
Sales Profits
(from Appendix Per Dollar
Table l) of Sales

 

Local Industries
‘Eiec., Gas, Wdter 8
Sanitary Services 1,212,570 .0873

 

Printing 8 Publishing
Transp. & Warehousing
Whlse. 8 Retail Trade
Communications
Fin., Ins., 8 Real Est.
Services
Construction

Sub-Total

National Industries

 

Metal Ores Mining
Coal Mining
Nondurable Goods
Lbr. 8 Wood Products

Chemicals 6 Allied Prod.

Primary Metals
Fabricated Metals

Stone,C1ay 8 Glass Prod.

Non-Elec. Machinery

Electrical Machinery

Motor Vehicles

Other Transp. Equip.

Instruments

Misc. Manufacturing
Sub-Total

Total

2,835

1,886,085
1,192,320

168,885

1,251,450

643,545
285,120

1,924,560

772,740
376,245

61,560
677,565

8,924,175

956,610
74,925
809,595
279,045
127,170
20,250
38,070
51,435

.0512
.0135
.0146
.0943
.0219
.0160

.0403
.0403
.0371
.0402
.0615
.0547
.0438
.0464
.0533
.0415
.0557
.0319
.0782
.0386

286

APPENDIX TABLE 2 (cont'd)

 

 

 

3 4 5
Profit Localization Detroit
Gains Ratiosl Profit

Gains
105,895 1.0 105,895
154 1.0 154
25,461 1.0 25,461
17,710 1.0 17,710
15,937 1.0 15,937

-- 1.0 --

14,104 1.0 14,104
4 560 1.0 4,560
133,821 0
77,578 .0016 124
31,152 .0016 50
13,950 .0012 17
2,492 .0001 0
41,697 .0192 801
488,143 .0349 17,036
41,917 .0595 2,494
3,480 0 0
43,173 .0492 2,124
11,579 .0048 51
7,074 .2709 1,916
64 .1573 10
2,972 .0228 68
1 969 .0065 13
757.245 24,755
9519661 2630530

287

APPENDIX TABLE 3

Trenton Channel Project -- Local Sales and Net Wage
Gains Stemming-from Increased Steel Production

 

 

1 2 3
Increased Distribution Sales
Steel Of Steel Stemming
Sales Sales From

($ million)

 

Local Industries1
Elec., Gas, Water 6

 

 

Sanitary Services .00230 93,150
Printing and Publ. .00008 3,240
Trans. 5 Warehousing .00913 369,765
Whlse. & Retail Trade .00186 75,330
Communications 0 0
Fin., Ins., 8 Real Est. .00105 42,525
Services .00048 19,440
Construction .01632 660,960

Sub-Total
National Industries1
—Metai’OresiMining .00194 78,570
Coal Mining .00158 63,990
Non-durable Goods .01998 809,190
Lbr. 5 Wood Prod. .00324 131,220
Chem. 8 Allied Prod. .00922 373,410
Primary Metals .20444 8,279,820
Fabricated Metals .07047 3,025,350
Stone,C1ay 5 Glass Prod. .00479 193,995
Non-elec. Machinery .12340 4,997,700
Electrical Machinery .05695 2,306,475
Motor Vehicles .14026 5,680,530
Other Transp. Equip. .03777 1,529,685
Instruments .00593 240,165
Misc. Manufacturing .00824 333,720
Sub-Total 40.5

Total

288

APPENDIX TABLE 3 (cont'd)

 

 

 

4 5 6 7
Sales Net Localization Local
Induced-by Production R t' 1 Exportable
(from Appendix (3) - (4) a 10 Surplus
Table 1)
1,212,570 ---
2,835 240 1.0 240
1,8886,085 ---
1,192,320 ---
168,885 ---
1,251,460 ---
643,545 ---
285,120 375,960 1.0 375,960
1,924,560 ---
772,740 ---
376,245 433,190 .0012 520
61,560 69,220 .0001 7
677,565 ---
8,924,175 ---
956,610 2,068,350 .0596 123,067
74,925 118,995 0 0
809,595 4,187,700 .0492 206,035
279,045 2,027,475 .0048 9,732
127,170 5,553,530 .2709 1,504,451
20,250 1,527,685 .1573 240,305
38,070 202,165 .0228 4,609
51,435 282,720 .0065 1,838

289

APPENDIX TABLE 3 (cont'd)

 

 

 

8 9
Wage Local
Rent Net Wage

Coefficients Gains
.2140 51
.1090 40,980

9
.0412 21
.0243 0
.1161 14,288
.1527 31,462
.1264 1,230
.0816 122,763
.1693 40,684
.1403 647
.0551 101

9

290
APPENDIX TABLE 4

Trenton Channel Project-Local Profit Gains Stemming-from

Increased Steel Production

 

 

 

 

1 2 3
Local After-Tax Local
Exportable Profits Per Profit
Surplus Dollar of Gains
(from Appendix 5
Table 3) Sales
Local Industries
Elec., Gas, Water 8
Sanitary Services
Prtng. 8 Publishing
Transp. G Warehousing 240 .0135 3
Wholesale 8 Retail Trade
Communications
Fin., Insurance 8
Real Estate
Services
Construction 375,960 .0160 6 015
Sub-Total 6,0I8
National Industries
—_Meta1 Ores Mining
Coal Mining
Nondurable Goods 520 .0371 19
Lbr. 8 Wood Products 7 .0402 *
Chem. 8 Allied Prod.
Primary Metals
Fabricated Metals 123,067 .0438 5,390
Stone,C1ay 5 Glass Prod.
Non-electrical Mach. 206,035 .0533 10,982
Electrical Machinery 9,732 .0415 404
Motor Vehicles 1,504,451 .0556 83,798
Other Transp. Equip. 240,305 .0319 7,666
Instruments 4,609 .0782 360
Misc. Manufacturing 1,838 .0386 71
Sub-Total 108,090
Total 114,700

291

APPEDIX TABLE 5

Trenton Channel Project-Sales and Local Net Wage Gains
Induced-by Project Construction

 

 

 

l 2 3 4
Construction Direct Induced Wage Rent
Cost Require- Sales Coeffi-
($ million) ments ($1000) cients

Agric., For. 6

Fisheries 0 0 0
Mining 0 0 .177
Construction 0 0 .091
Non-durable

Goods Mfg. .103 3,234.2 .074
Lumber 6 Wood

Products 0 0 .091
Stone 8 Clay

Products 0 0 .163
Primary Metals .045 1,413.0 .162
Fabricated Metals .021 659.4 .156
Non-electrical

Machinery .001 31.4 .194
Construction

Machinery .032 1,004.8 .176
Electrical Mach. 0 0 .162
Transp. Equip. .151 4,741.4 .176
Misc. durable Goods 0 0 .103
Transp. 5 Warehousing .017 533.8 .232
Whlse. & Retail Trade .056 1,758.4 .038
Services .028 879.2 0

Total 31.4

292

APPENDIX TABLE 5 (cont'd)

 

 

 

5 6 7
Net Wage Localization Local Net
Gains . 1 Wage

($) Ratlo Gains ($)

0 0 0

0 .001 0

0 .016 0

239,330 .011 2,633

0 .004 0

0 .017 0

228,906 .037 8,470

102,866 .060 6,172

6,092 .050 305

176,845 .050 8,442

0 .005 0

834,486 .111 92,628

0 .042 0

123,841 1.000 123,841

66,819 1.000 66,819

0 1.000 0

309.310

293

APPENDIX TABLE 6

 

 

 

Trenton Channel Project -- Local Profits Induced-by
Project Construction
1 2

Induced Profits per

Sales Dollar of
(from Appendix Sales

Table 5) ($)

Agric., For. 6 Fisheries 0 .030
Mining 0 .077
Construction 0 .028
Non-durable Goods Mfg. 3,234,200 .068
Lbr. 8 Wood Products 0 .057
Stone 6 Clay Products 0 .080
Primary Metal Products 1,413,000 .090
Fabricated Metal Prod. 659,400 .078
Non-electrical Machinery 31,400 .104
Construction Machinery 1,004,800 .090
Electrical Machinery 0 .080
Transportation Equip. 4,741,400 .090
Misc. Durable Goods 0 .062
Transportation 8 thg. 533,800 .054
Wholesale 6 Retail Trade 1,758,400 .024
Services 879,200 .047

Total

294

APPENDIX TABLE 6 (cont'd)

 

 

 

3 4 5
Induced Localization Local
Profits Ratios Profit

($) Gains

($)

0 0 0
0 .001 0
0 .016 0
219,925 .011 2,419
0 .004 0
0 .017 0
127,170 .037 4,705
51,433 .060 3,086
3,266 .050 163
90,432 .050 4,522
0 .005 0
426,726 .111 47,367
0 .042 0
28,825 1.000 28,825
42,202 1.000 42,202
41,322 1.000 41,322

Total

174,611

295
APPENDIX TABLE 7

Stonewall Jackson Reservoir-Sales and Net Wage Gains
Induced-by Recreation Expenditures

 

 

 

l 2
Annual Distribution
Recreation 1 of
Expenditures Purchases11
($1
Food 6 Kindred Products .252
Restaurants .084
Hotels .254
Transportation .211
Misc. Services .199

Total 469,500

296
APPENDIX TABLE 7 (cont'd)

 

 

 

3 4 5 6 7
Induced Proportion Local Wage Rent -8 Local
Sales Spent Induced Coefficients Net Wage
($) Locally11 Sales Gains
($) ($)
118,314 .15 17,747 .2615 4,641
39,438 .48 18,930 .6081 11,511
119,253 .75 89,440 .6081 54,388
99,065 .25 24,766 .4813 11,920
93,431 .40 37,372 .5314 18,859

101,319

297

APPENDIX TABLE 8

Stonewall Jackson Reservoir-Sales and Profit Gains
Induced-by Recreation Expenditures

 

 

 

l 2 3
Local After-Tax Local
Induced Sales Profits-per 5 Profit
(from Appendix Dollar of Sales Gains
Table 5)
Food 5 Kindred
Products 17,747 .0209 371
Restaurants 18,930 .0257 487
Hotels 89,440 .0257 2,299
Transportation 24,766 .0144 357
Misc. Services 37,372 .0373 1,394

 

4,908

298

APPENDIX TABLE 9

Stonewall Jackson Reservoir-Sales and Local Net Wage

Gains Induced-by Project Construction

 

 

 

1 2 3
Construction Direct Induced
Cost Requirements Sales
($'000) ($)
Agriculture, For. 8
Fisheries .000 0
Mining .041 2,053,485
Non-durable Goods .026 1,302,210
Lumber & Wood Prod. .006 300,510
Stone,C1ay 8 Glass
Products .060 3,005,100
Primary Metals .035 1,752,971
Fabricated Metals .077 3,856,545
Non-electrical Mach. .043 2,153,655
Construction Equip. .044 2,203,740
Electrical Mach. .046 2,303,910
Transportation Equip. .006 300,510
Misc. Durable Goods .002 100,170
Transp. G Warehousing .039 1,953,315
Wholesale 8 Retail
Trade .061 3,055,185
Services .027 1,352,295
Total 50,085

299

APPENDIX TABLE 9 (cont'd)

 

 

 

3 4 5 6 7.
Induced Wage Net Localization Local
Sales Rent. Wage Ratiol Net Wage
($) Coeff1c1ents Ga1ns Ga1ns
0 .0000 0 .0057
2,053,485 .1771 363,692 .0057 23,749
1,302,210 .0358 46,619 .0112 522
300,510 .0497 14,935 .0075 112
3,005,100 .0669 201,041 .0463 9,308
1,752,971 .1056 185,114 .1442 26,693
3,856,545 .1199 462,400 .0254 11,745
2,153,655 .1505 324,125 .0148 4,797
2,203,740 .1509 332,544 .0068 2,261
2,303,910 .1239 285,454 .0217 6,194
300,510 .0807 24,251 .0056 136
100,170 .1187 11,890 .0227 270
1,953,315 .1713 334,603 1.0 334,603
3,055,185 .0824 251,747 1.0 251,747
1,352,295 .0000 0 1.0 0

672,137

300

APPENDIX TABLE 10

Stonewall Jackson Reservoir-Local Profit Gains

Induced—by Project Construction

 

 

 

l 2
Induced Localization
(froSaAggendix Ratiol
Table 7)

Agriculture, Forestry

8 Fisheries O .0057
Mining 2,053,485 .0653
Nondurable Goods 1,302,210 .0112
Lbr. & Wood Products 300,510 .0075
Stone, Clay 8 Glass Prod.3,005,100 .0463
Primary Metals 1,752,971 .1442
Fabricated Metals 3,856,545 .0254
Non-electrical Mach. 2,153,655 .0148
Construction Equip. 2,203,740 .0068
Electrical Machinery 2,203,910 .0217
Transportation Equip. 300,510 .0056
Misc. Durable Goods 100,170 .0227
Transp. 8 Warehousing 1,953,315 1.0
Wholesale 8 Retail

Trade 3,055,185 1.0
Services 1,352,295 1.0

Total

301

APPENDIX TABLE 10 (cont'd)

 

 

 

3 4 5
Local After-tax Local
Induced Profits per 5 Profit
Sales Dollar of Sales Ga1ns
0 .0024 0
134,093 .0403 5,404
14,585 .0371 541
2,254 .0402 91
139,136 .0464 6,451
252,778 .0547 13,827
97,959 .0438 4,290
31,874 .0533 1,699
14,985 .0319 478
47,825 .0415 1,985
1,683 .0319 54
2,274 .0327 74
1,953,315 .0135 26,370
3,055,185 .0146 44,606
1,352,295 .0219 29,615

135,490

302
APPENDIX TABLE 11

Poteau Project-Sales and Net Wage Gains Induced-by
Increased Poultry Plant Sales

 

 

 

 

 

1 2 3
Poultry Direct Induced
Plant Requirements Sales
Sales Per Dollar Of ($)
1
($ 000) Sales2
Local Industries1
Elec., Gas, Water 8
Sanitary Services .00667 18,009
Printing 8 Publishing .00606 16,362
Transp. 5 Warehousing .03314 89,478
Whlse. & Retail Trade .03497 94,419
Communications .00257 6,939
Finance, Ins. & Real
Estate .01011 27,297
Government 8 Private
Services .03594 97,038
Construction .00200 5,400
Sub-Total
National Industries1
Agriculture, For. 8
Fisheries .30569 825,363
Mining .00069 22,680
Non-durable Goods .19986 539,622
Lbr. 8 Wood Prod. .00130 3,510
Chemicals 8 Allied Prod. .00757 20,439
Primary 6 Fabricated Mtls. .00268 7,236
Paper 8 Allied Prod. .00049 1,323
Stone, Clay 6 Glass Prod. .00031 837
Machinery 6 Equipment .00253 6,831
Misc. Durable Goods .01739 46,953

Sub-Total
Total 2,700

APPENDIX TABLE 11 (cont'd)

 

 

 

4 5 6 7
Wage Net Localization Local
Rent Wage . 1 Net Wage

Coefficients7 Gains Ratios Gains

($) ($)
.4624 8,327 1.0 8,327
.1979 3,238 1.0 3,238
.5181 46,359 1.0 46,359
.6152 58,087 1.0 58,087
.4315 2,994 1.0 2,994
.3850 10,509 1.0 10,509
.4333 42,047 1.0 42,047
.3049 1,646 1.0 1,646
1 39 y

.0000 0 --- 0
.1771 4,017 .0078 31
.0358 19,318 .0098 180
.0497 174 .0361 6
.0810 1,656 .0059 10
.1169 846 .0055 5
.1168 155 .0111 2
.0669 56 .0065 0
.1429 976 .0036 4
.1187 5 573 .0069 38
34,774 276

304
APPENDIX TABLE 12

Poteau Project-National and Local Profit Gains
Induced-by Increased Poultry Processing

 

 

 

 

 

1 2
Induced After-Tax
Sales Profit Per
(from Appendix Dollar Of
Table 9) 5
($) Sales
Local Industries1
E1ec., Gas, Water 8
Sanitary Services 18,009 .0873
Printing 8 Publishing 16,362 .0512
Transp. G Warehousing 89,478 .0135
Whlse. 8 Retail Trade 94,419 .0146
Communications 6,939 .0943
Finance, Ins. 6 Real
Estate 27,297 ---
Government 8 Private
Services 97,038 .0219
Construction 5,400 .0160
Sub-Total
National Industries1
Agriculture, For. 8
Fisheries 825,363 .0024
Mining 22,680 .0403
Non—durable Goods 539,622 .0371
Lbr. 8 Wood Prod. 3,510 .0402
Chemicals 8 Allied Prod. 20,439 .0615
Primary 8 Fabricated Mtls. 7,236 .0500
Paper 6 Allied Products 1,323 .0528
Stone, Clay 8 Glass Prod. 837 .0464
Machinery 8 Equipment 6,831 .0477
Misc. Durable Goods 46,953 .0327

Sub-Total

Total

305

APPENDIX TABLE 12 (cont'd)

 

 

 

3 4 5
Induced Localization Local
Profit . 1 Profit
Gains Ratio Gains

($) ($)
1,572 1,572
838 838
1,208 1,208
1,379 1,379
654 654
2,125 2,125
86 86
:3862 79862
1,981 1.0 1,981

914 .0078 7
20,020 .0093 186

141 .0361 5

1,257 .0059 7
362 .0055 2
70 .0111 1

39 .0065 *
326 .0036 1
1,535 .0069 11

9 2,261
34,507 2,20I

306
APPENDIX TABLE 13

Poteau Project-Sales and Local Net Wage Gains
Stemming-from Increased Poultry Processing

 

 

 

 

 

l 2 3
Poultry Distribution Sales
Plant of Poultry Stemming
Sales 6 From
(s'OOO) Sales ($)
Local Industries1

“Eiec., Gas, Water 6

Sanitary Services .0000 0
Printing 8 Publishing .0001 270
Transp. G Warehousing .0017 4,590
Whlse. 8 Retail Trade .0098 26,460
Communications 0 0
Finance, Ins. 8 Real

Estate .0014 3,780
Government a Private

Services .0069 18,630
Construction .0003 810

National Industries1

Agriculture, For. 6

Fisheries .0489 132,030
Mining 0 0
Non-durable Goods .1778 480,060
Lbr. 8 Wood Products 0 0
Chemicals 6 Allied Prod. .0078 21,060
Paper 6 Allied Products .0016 4,320
Primary 6 Fabricated Mtls. 0 0
Stone, Clay 6 Glass Prod. 0 0
Machinery 6 Equipment .0002 540
Misc. Durable Goods .0304 82,080
Final Demand .6909
Net Export .0221

Total 2,700

307

APPENDIX 13 (cont'd)

 

 

 

4 5 6
Induced Net Localization
Sales Production . 1
(from Appendix Ratios
Table 10)
18,009 (-) 1.6
16,362 (-) 1.0
89,478 (-) 1.0
94,419 (-) 1.0
6,939 (-) 1.0
27,297 (-) 1.0
97,038 (-) 1.0
5,400 (~) 1.0
825,363 (-)
22,680 (-) .0078
539,622 (-) .0093
3,510 (-) .0361
20,439 621 .0059
1,823 997 .0111
7,236 (-) .0055
6,831 (-) .0036
46,953 127 .0069

308

APPENDIX TABLE 13 (cont'd)

 

 

Local Wage Net Wage
Exportable Rent 7 Gains
Surplus Coefficients ($)

($)

 

.4624
.1979
.5181
.6152
.4315

.3850

.4333
.3049

.1771
.0358
.0497
4 .0810 *
33 .1168 4
.1169
.0669
.1429
242 .1187 29

309
APPENDIX TABLE 14

Poteau Project-Local Profit Gains Stemming-from
Increased Poultry Processing

 

Exportable After-Tax Local
Surplus Profits Per Profit
(From Appendix Dollar of Gains
Tab1§)1l) Sales5 ($)

 

Local Industries1
Elec., Gas, Water 6
Sanitary Services
Printing 8 Publishing
Trans. 8 Warehousing
Wholesale 6 Retail Trade
Communications
Finance, Ins. & Real
Estate
Government 8 Private
Services
Construction

Sub-Total 0

National Industries1
Agriculture, For. 8

 

 

Fisheries
Mining
Non-durable Goods
Lumber 8 Wood Prod. 4 .0615 *
Primary 8 Fabricated
Metals
Paper 8 Allied Prod. 33 .0528 2

Stone,C1ay 8 Glass Prod.

Machinery 8 Equipment

Misc. Durable Goods 242 .0327 8
Sub-Total 10

Total 10

310
APPENDIX TABLE 15

Poteau Project-Sales and Local Net Wages Induced-by
Project Construction

 

 

 

1 2
Project Direct2
Construction Requirements

Cost Per Dollar Of

($‘000) Sales
Agriculture, For. 8 Fisheries 0
Mining .041
Non-durable Goods .035
Lumber 8 Wood Products .012
Stone, Clay 8 Glass Prod. .069
Primary Metals .005
Fabricated Metals .126
Non-electrical Machinery .003
Construction Equipment .041
Electrical Machinery .001
Transportation Equipment .021
Misc. Durable Goods ---
Transp. & Warehousing .038
Wholesale 8 Retail Trade .082
Services .027

Total 5,500

311

APPENDIX TABLE 15 (cont'd)

 

 

 

3 4 5 6 7
Induced Wage Net Localization Local
Sales Rent. Wage Ratiol Net Wage
($) Coeff1c1ents Ga1ns Ga1ns
($) ($)

0 .4217 0 - 0
225,500 .4405 99,333 .0078 775
192,500 .1800 34,650 .0093 332

66,000 .3165 20,889 .0361 754
379,500 .3454 131,079 .0065 852
27,500 .2714 7,464 .0028 21
693,000 .3490 242,134 .0048 1,162
16,500 .3902 6,438 .0024 15
225,500 .3868 87,223 .0027 236
5,500 .3877 2,132 .0066 14
115,500 .3868 44,675 .0027 121
--- .3477 --- .0069 ---
209,000 .5181 108,283 .0 108,283
451,000 .6152 277,455 .0 277,455
148,500 .4333 64,345 .0 64,345

454,365

i i | (I'll! 'lll. Al I I. I 'l‘ll Ill

312
APPENDIX TABLE 16

Poteau Project-Local Profit Gains Induced-by
Project Construction

 

 

 

1 2
Induced Localization
(froiaAggendix Ratiol
Table 15)
($)
Agriculture, For. 5 Fisheries 0 ---
Mining 225,500 .0078
Non—durable Goods 192,500 .0093
Lumber & Wood Products 66,000 .0361
Stone, Clay 6 Glass Prod. 379,500 .0065
Primary Metals 27,500 .0028
Fabricated Metals 693,000 .0048
Non-electrical Machinery 16,500 .0024
Construction Equipment 225,500 .0027
Electrical Machinery 5,500 0066
Transportation Equipment 115,500 .0027
Miscellaneous Durable Goods --- .0069
Transp. & Warehousing 209,000 .0
Wholesale 6 Retail Trade 451,000 .0
Services 148,500 .0

Total

313
APPENDIX TABLE 16 (cont'd)

 

 

 

3 4 5

Local After-Tax Local
Induced Profits per Profit

Sales Dollar Of Gains
($) Sales ($)

0 .0024 0

1,759 .0403 71
1,790 .0371 66
2,383 .0402 96
2,467 .0464 114
77 .0547 4
3,326 .0438 146
40 .0533 2

609 .0319 19

36 .0415 2

312 .0319 10

--- .0327 ---
209,000 .0135 2,822
451,000 .0146 6,586
148,500 .0219 3,252

13,189

314
APPENDIX TABLE 17
Mountain Home Project-Sales, Wages and Net Wage Gains

Induced-by Increase in Agricultural Production:
National and Local

 

 

 

 

 

1 2
Increase Direct
in Requirements
Agricultural Per Dollar
Sales 2
($ million) Of Sales
Local Industries1
EIec., Gas, Wdter 8
Sanitary Services .00291
Printing 8 Publishing .00018
Transp. 8 Warehousing .01841
Wholesale 8 Retail Trade .02987
Communications .00125
Finance, Insurance 8
Real Estate .03265
Government 8 Private
Services .01438
Construction .00524
Sub-Total
National Industries1
Agriculture, For. 8
Fisheries .23992
Mining .00117
Textile Products .01576
Lumber 8 Wood Prod. .00290
Chemicals 8 Allied Products .01472
Paper 8 Allied Products .01242
Misc. Non-durable Goods .02624
Metal Products .02471
Stone, Clay 8 Glass Prod. .00049
Machinery 8 Equipment .01147
Misc. Durable Goods .00006

Sub-Total

Total 16.0

315

APPENDIX TABLE 17 (cont'd)

 

 

 

3 4 5 6 7
Induced Wage Net Localization Local
Sales Rent Wage Ratio1 Net Wage
($) Coefficients Gains Gains
($) ($)
46,560 .0782 3,645
2,880 .0336 97
294,560 .2312 68,157
477,920 .0161 7,773
20,000 .0817 1,635
522,400 .1142 59,673
230,080 .0000 0
83,840 .1434 12,030
9 IS29g:U
3,838,720 .0000 0 .0084 0
18,720 .1771 3,316 .0053 18
252,160 .0556 14,023 0 0
46,720 .0497 2,323 .0010 2
235,520 .0810 19,090 .0013 25
198,720 .1168 23,229 .0014 33
419,840 .0358 15,069 .0022 33
395,360 .1169 46,223 .0004 18
7,840 .0669 525 .0014 1
183,520 .1429 26,243 .0016 42
960 .1187 114 .0016 0
1569133 1:2
36391:25 I539Izz

APPENDIX TABLE 18

316

Mountain Home Project-National and Local Profits
Induced-by Increased Agricultural Production

 

 

 

 

 

Sub-Total

Total

1 2
Induced After-Tax
Sales Profits per
(from Appendix Dollar of
Table 15) 5
($) Sales
Local Industries1

Elec., Gas, Water 8

Sanitary Services 46,560 .0873
Printing 8 Publishing 2,880 .0512
Transp. 8 Warehousing 294,560 .0135
Wholesale 8 Retail Trade 477,920 .0146
Communications 20,000 .0943
Finance, Insurance 8

Real Estate 522,400 --
Government 8 Private

Services 230,080 .0219
Construction 83,840 .0160

Sub-Total

National Industries1

Agriculture, For. 8

Fisheries 3,838,720 .0024
Mining 18,720 .0403
Textile Products 252,160 .0290
Lumber 8 Wood Prod. 46,720 .0402
Chemicals 8 Allied Prod. 235,520 .0615
Paper 8 Allied Products 198,720 .0528
Misc. Non—durable Goods 419,840 .0371
Metal Products 395,360 .0500
Stone, Clay 8 Glass Prod. 7,840 .0464
Machinery 8 Equipment 183,520 .0477
Misc. Durable Goods 960 .0327

317
APPENDIX TABLE 18 (cont'd)

 

 

 

3 4 5
Profit Localization Local
Gains . 1 Profit

($) Ratlo Gains
($)
4,065 1.0 4,065

147 1.0 147
3,977 1.0 3,977
6,979 1.0 6,978
1,886 1.0 1,886

--- 1.0 ---
5,039 1.0 5,039
1 341 1.0 1 341

W: 3 7m.
9,213 .0084 77

754 .0053 4
7,313 0 0
1,878 .0010 2

14,484 .0013 19
10,492 .0014 15
15,576 .0022 34
19,768 .0004 8

364 .0014 1

8,754 .0016 14
31 .0016 *

318
APPENDIX TABLE 19
Mountain Home Project-Sales, Exportable Surplus,

and Local Net Wage Gains Stemming-from Increased
Agricultural Production

 

 

 

 

 

1 2
Increase in Distribution
Agricultural of
Sales Agricultural
($ million) Sales
Local Industriesl

Elec., Gas, Water 8

Sanitary Services 0
Printing 8 Publishing 0
Transp. 8 Warehousing .0015
Wholesale 8 Retail Trade .0029
Communications 0
Finance, Insurance 8

Real Estate .0442
Government 8 Private

Services .0121
Construction .0056

Sub-Total

National Industries1

AgriEulture, For. 8

Fisheries .3100
Mining 0
Textile Products .0295
Lumber 8 Wood Products .0188
Chemicals 8 Allied Prod. .0009
Paper 8 Allied Products 0
Food 8 Kindred Products .3950
Misc. Non-durable Goods .0205
Metal Products 0
Stone, Clay 8 Glass Prod. .0001
Machinery 8 Equipment 0
Misc. Durable Goods .0002

Sub-Total

Total 16.0

319
APPENDIX TABLE 19 (cont'd)

 

 

 

3 4 5 6
Sales Sales Exportable Localization
Stemming- Induced-by Surplus
From (from Appendix
($) Table 15)
($)
0 46,560 ---
0 2,880 ---
24,000 294,560 ---
46,400 477,920 ---
0 20,000 ---
707,200 522,400 184,800 1.0
193,600 230,080 ---
89,600 83,840 5,760 1.0
4,960,000 3,838,720 1,121,280 .0084
0 18,720 ---
472,000 252,160 219,840 0
300,800 46,720 254,080 .0010
14,400 235,520 ---
0 198,720 ---
6,320,000 711,520 5,608,480 .0078
328,000 419,840 ---
0 395,360 ---
1,600 7,840 ---
0 183,520 ---
3,200 960 2,240 .0016

320

APPENDIX TABLE 19 (cont'd)

 

 

 

7 8 9
Local Wage Local
Exportable Rent Net Wage
Surplus Coefficients Gains
184,800 .1142 21,110
5,760 .1434 826

21,936

9,419 .0000 0
0

254 .0497 13

43,746 .0674 2,948

4 .1186 0

29961

321
APPENDIX TABLE 20

Mountain Home Project-Local Profit Gains Stemming-from
Increased Agricultural Production

 

 

 

 

 

l 2
Local After-Tax
Exportable Profits per
Surplus Dollar Of
(from Appendix 5
Table 17) Sales
Local Industries1
’Elec., Gas, Water 8
Sanitary Services
Printing 8 Publishing
Transp. 8 Warehousing
Wholesale 8 Retail Trade
Communications
Finance, Insurance 8
Real Estate 184,800 ---
Government 8 Private
Services
Construction 5,760 .0160
Sub-Total
National Industries1
Agriculture,’FOrestry
8 Fisheries 1,121,280 .0024
Mining
Textile Products 219,840 .0290
Lumber 8 Wood Prod. 254,080

Chemicals 8 Allied Prod.

Paper 8 Allied Prod.

Food 8 Kindred Prod. 5,608,480 .0209

Misc. Non-durable Goods

Metal Products

Stone, Clay 8 Glass Prod.

Machinery 8 Equipment

Misc. Durable Goods 2,240 .0327
Sub-Total

Total

322

APPENDIX TABLE 20 (cont'd)

 

 

Local
Profit
Gains

($)

 

2,691

6,875
10,214

117,217

323
APPENDIX TABLE 21

Mountain Home Project-Local Wage Gains Induced-by
Project Construction

 

 

 

1 2
Construction Direct
Cost Requirements
($1000)

Agriculture, For. 8 Fisheries 0
Mining .041
Non-Durable Goods .026
Lumber 8 Wood Products .006
Stone, Clay 8 Glass Prod. .060
Primary Metals .035
Fabricated Metals .077
Non-electrical Machinery .043
Construction Equipment .044
Electrical Machinery .046
Transportation Equipment .006
Miscellaneous Durable Goods .002
Transp. 8 Warehousing .039
Wholesale 8 Retail Trade .061
Services .027

Total 153,550

324

APPENDIX TABLE 21 (cont'd)

 

 

 

3 4 5 6 7
Induced Wage Net Localization Local

5.1.. Coe‘f‘i‘i‘iientg 328i. Ratiol ”65.1%?

($) ($)

0 0 0 .0084

6,295,550 .1771 1,114.942 .0053 5,909
3,992,300 .0358 142,924 .0022 314
921,300 .0497 45,789 .0010 46
9,213,000 .0669 616,350 .0014 863
5,374,250 .1056 567,521 .0008 454
11,823,350 .1199 1,417,620 .0001 142
6,602,650 .1505 993,699 .0003 298
6,756,200 .1509 1,109,511 .0000 0
7,063,300 .1239 875,143 .0016 1,400
921,300 .0807 74,349 .0000 0
307,100 .1187 36,453 .0016 58
5,988,450 .2313 1,385,128 1.0 1,385,128
9,366,550 .0161 150,801 1.0 150,801
4,145,850 0 0 1.0 0

 

1,545,413

325
APPENDIX TABLE 22

Mountain Home Project -- Local Profit Gains
Induced-by Project Construction

 

 

 

1 2
Induced Localization
(froiaAggendix Ratiol

Table 19)
Agriculture, For. 8 Fisheries --- .0084
Mining 6,295,550 .0053
Non-durable Goods 3,992,300 .0022
Lumber 8 Wood Products 921,300 .0010
Stone, Clay 8 Glass Prod. 9,213,000 .0014
Primary Metals 5,374,250 .0008
Fabricated Metals 11,823,350 .0001
Non-electrical Machinery 6,602,650 .0003
Construction Equipment 6,756,200 0
Electrical Machinery 7,063,300 .0016
Transportation Equipment 921,300 0
Misc. Durable Goods 307,100 .0016
Transp. 8 Warehousing 5,988,450 1.0
Wholesale 8 Retail Trade 9,366,550 1.0

Services 4,145,850 1.0

326

APPENDIX TABLE 22 (cont'd)

 

 

 

3 4 5
Local After-Tax Local
Sales Profits Per Profit
($) Dollar of Gains
Sales5 ($)
0 .0024 0
33,366 .0403 1,345
8,783 .0371 326
921 .0402 37
12,898 .0464 598
4,299 .0547 235
1,182 .0438 52
1,981 .0533 106
0 .0319 0
11,301 .0415 469
0 .0319 0
491 .0327 16
5,988,450 .0135 80,844
9,366,550 .0146 136,752
4,145,850 .0219 90,794

311,574

327

Footnotes

 

*Less than $0.50.

**Set at zero to reflect importing of limestone used
in steel production.

1. Local Industries sell primarily to local markets,
while National Industries sell to National Markets. See
Leontief, et. al, "The Economic Impact--Industria1 and
Regional of an Arms Cut," Review Of Economics and Statis-
tics (August 1965). Locaiization ratios derived by'
d1v1ding local employment in each industry by National
employment in that industries (set at 1.0 for local in-
dustries).

 

2. From 1963, Input-Output Study, "Direct Require-
ments Per Dollar of Gross Output, 1963", Survey Of Current
Business (November 1969), Vol. 49, NO. 11, p. 38.

 

3. From Leontief, et. al, "The Economic Impact . . .,"

o cit. Also "Income and Employment by Industry," Surve
0% Current Business (July 1969, Vol. 49, NO. 7, pp. 39-44

 

4. Local wage rates used for Local Industries;
National wage rates used for National Industries. Source:
Employment and Earnin 5: States and Areas 1939-68.
BfiIIetin NO. 1370-6, . S. Dept. Of Labor, Bureau Of
Labor Statistics (August 1969); Agricultural Statistics
1969, U. S. Department Of Agriculture. GOv‘t Printing
Office (Washington: 1969).

 

 

5. Survey of Current Business (July 1969).

6. 1963 Input-Output Study, "Table 1. Inter-industry
Transactions, 1963," Survey Of Current Business (November

1969) Op cit.

7. Defines as

 

Wi - We

° w
1 1 C1, here

Wi - average production worker wage in industry i.

We = average production worker wage in lowest
average wage industry.

Ci

labor compensation in industry 1, and
Si - sales in industry i.
8. Opportunity Cost set at zero. Wage Rent Coef-

ficient reduces to "labor compensation per dollar Of
sales."

328

9. Haveman and Krutilla, opycit, p. 20.

10. As estimated by Kalter, et a1, Op cit.

 

11. Marion Clawson, Reports to the Outdoor Recreation
Resources Review Commission, NO. 24. Economic Studies Of
Outdoor RecreatiOn (Washington: 1962), 85.

 

178 0293

mm