Front Cover
 Table of Contents
 Defining the impacts of water resource...
 The relationship between income...
 Determining the regional economic...
 Project evaluation methodology...
 Summary and conclusions
 Multiple goal account sheet for...

Group Title: Economics report - University of Florida. Agricultural Experiment Station ; no. 40
Title: Impacts of watershed projects
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00027732/00001
 Material Information
Title: Impacts of watershed projects
Series Title: Economics report
Physical Description: iii, 66 l. : illus. ; 28 cm.
Language: English
Creator: Gibbs, Kenneth C
Loehman, Edna Tusak, 1943-
Publisher: Food and Resource Economics Dept., University of Florida
Place of Publication: Gainesville
Publication Date: 1972
Subject: Water resources development   ( lcsh )
Water resources development -- Mathematical models   ( lcsh )
Watersheds   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: leaves 61-66.
Statement of Responsibility: by Kenneth C. Gibbs and Edna T. Loehman.
General Note: Cover title.
 Record Information
Bibliographic ID: UF00027732
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 001614508
oclc - 00800388
notis - AHN8934
lccn - 73622919

Table of Contents
    Front Cover
        Front Cover
        Page i
    Table of Contents
        Page ii
        Page iii
        Page 1
        Page 2
        Page 3
    Defining the impacts of water resource investment projects
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    The relationship between income and employment goals
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
    Determining the regional economic effects of a proposed investment project
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
    Project evaluation methodology and the use of accounts
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
    Summary and conclusions
        Page 53
        Page 54
        Page 55
    Multiple goal account sheet for a project proposal
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
Full Text
November 1972 Economic Report 40

Impacts of Watershed Projects

Food and Resource Economics Department
Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
University of Florida, Gainesville
in cooperation with
Natural Resource Economics Division
Economic Research Service
United States Department of Agriculture


M!A 09 1973

Kenneth C. Gibbs
Edna T. Loehman

-uuyplplr~-ur~---~-- II -yuprisllr~ps


Appreciation is expressed to departmental reviewers, J. R.
Conner and J. C. Cato, for their effort in providing valuable com-
ments and suggestions.
A major portion of the financial support for this study was
provided by the Natural Resource Economics Division, Economic Research
Service, U. S. Department of Agriculture. The authors also wish to
thank N. R., Cook and Karl Gertel of the Economic Research S-rvice, for
reviews of earlier drafts of this manuscript.


LIST OF FIGURES . . . . . .
PROJECTS . . . . . ..
Classifying Project Effects ... . .
Economic Effects . ...
Physical Effects . . . .
Social Effects . ... .
The Role of National Goals .
Defining National Goals ...... ..
Enhancing the Economic Environment .
Enhancing the Physical Environment ....
Enhancing the Social Environment ...........
Tradeoffs Among Goals . . .. .
Total Income and Income Distribution Defiled ..
Total Employment and Employment Distribution ..
Regional Income and Employment Goals ..........
Income Goals . . . . .
Employment Goals . . . .....
Delineating Regional Economic Effects . .
Construction and Operation Phases .. .
The Purpose of "Direct" and "Indirect" Effects
in Formulating Regional Models . . .
Input-Output Models and the Measurement of Project
Benefits . . .
Measuring Regional Costs .. . .......
A Dynamic Model to.Predict Regional Effects .
Definition of Variables . .
Constraints for the Linear Program . .
Choice of an Objective Function for the Linear
Program . . . . .
Evaluation of the Model and Concluding Comments ..
The Role of an Accounts Model . . .
A Comparison of Accounts Models . .








Example of Use of an Accounts Sheet .. . . 48
Problems of Setting Up an Accounts Model . . 49
APPENDIX . . . . . 55
BIBLIOGRAPHY . . . .... . .... 61


Figure Page
1. Income distribution . . . .. 17
2. Employment distribution . . . 18
3. Construction phase ....... 24
4. Operation phase . . . .. 25
5. Factors influencing net project benefits and
their allocation . . .. .. 31
6. Information inputs and outputs from the regional
model . . . 34
7. Distribution of income . . . 51


Kenneth C. Gibbs and Edna T. Loehman


The evaluation procedures used by many Federal agencies involved
with natural resource development are "guided" by recommended policies
set forth by the Federal Government. These policies are designed to pro-
vide mutually acceptable principles and procedures, so that consistency
is attained among agencies working on the same types of projects. The
first such "guideline" was presented by a subcommittee of the Federal
Interagency River Basin Committee in 1950, and revised in 1958. This
guideline, termed the "Green Book" [25], defined how economic efficiency
(referred to as the benefit-cost ratio) was to be used in agency deci-
sion making. The next major document concerned with water and related
land resources was issued in 1962 as an executive order and printed as
Sentate Document 97 of the 87th Congress, 2nd Session [18]. Senate Docu-
ment 97 encouraged agencies to consider multiple objectives in planning,
but offered few specific suggestions. Thus, justification for resource
investment projects has continued to be based primarily on the benefit-
cost ratio. There has been some mention of factors other than efficiency
in agency work plans, such as employment, income, and distribution of
benefits as indicated in a typical quote, taken from the Work Plan for
California Lake Watershed, Dixie County, Florida [26, p. 21]:
The proposed measures will reduce this unemployment
[to small timber cutting operations during the rainy season]
significantly .. The increased production of pulpwood
and other timber products as a direct result of the project

Kenneth C. Gibbs is assistant professor of food and resource economics,
and environmental engineering. Edna T. Loehman is assistant professor of
food and resource economics.

will reflect an estimated 15 percent increase in job and
income opportunities in the immediate area .
However, as indicated by the above quote, the extent of these economic
impacts has not been detailed or quantified, and emphasis has been on
other beneficial aspects.
The Water Resources Council [27], recognizing that Senate Docu-
ment 97 was not sufficiently specific, recently attempted to set out pro-
cedural details for considering national, regional, and local viewpoints.
It is now proposed that evaluation of a project be based not only on
national economic efficiency but also on a project's contribution to and
detrimental effects on other national and regional goals. Under this
structure, decision making is to take into account the interests of all
groups who may be affected, either favorably or unfavorably, by a project.
No distinction is to be made between direct or indirect effects of a
Many projects are proposed and supported by local citizens, and
the physical and economic effects of these projects are primarily local.
In these instances, the question may be raised as to why any factor other
than efficiency should be considered in making the final decision on
whether or not to construct a regional-type project. Regional projects
do, however, often affect more than the local area.
First, there are many externalities connected with the physical
effects of small watershed projects. Upstream projects affect conditions
downstream. For instance, the California Lake plan for protecting the
upper portion of the watershed resulted in flooding the lower portion.
Also, even though a project may have limited physical effects, the con-
struction of many such projects in an area could drastically alter the
hydrology of the area and affect total water use goals.
Second, it has been recognized that our political structure,
lobbying organizations, and the vested interests of an agency to expand
its own work load can promote the construction of projects despite some
local opposition. In addition, groups affected by'a project may some-
times be uninformed as to a project's true cost to them, and so may not
oppose it at the time of planning. Since the safeguarding of.all inter-
ests does not automatically occur through the institutional structure of
water resource planning, the decision makers should consider all effects,

local and non-local, physical and economic, in evaluating a project.
This is the rationale bey nd the new emphasis on evaluation of projects
in terms of multiple goals.
The first step in evaluating projects according to multiple
goals is to specify the effects to be considered in decision making.
The next step is to measure these relevant effects. Finally, the infor-
mation effects should be organized into the decision-making framework.
The role of goals is considered in a general way in Section II
of this paper. The relationship between income and employment goals is
considered in Section III. Determining and measuring regional economic
effects is the topic in Section IV. Toward this end, a model for esti-
mating such effects is presented. Finally, the steps in the decision
process, once relevant effects have been specified and measured, are
described in Section V.


In the past, work plans for watershed projects have dealt pri-
marily with economic effects from investment, and especially with the
question of economic efficiency. The economic effects considered from
a public investment such as a watershed project dealt at most with two
categories of economic effects -- those relating to economic efficiency
and those relating to income distribution. As discussed by Marglin [15,
p. 18], economic efficiency refers to the "size of the economic pie" as
it relates to total output and income; income distribution refers to
"how the pie is divided," or who will receive dollar benefits. Such
economic effects are often categorized as market effects. There may
also be non-market effects from a project, such as an effect on environ-
mental quality, which in turn may involve both monetary and non-monetary
Continuing Marglin's pie analogy, there is a question of the
method of "slicing of the pie." That is, flood control for a region
could be accomplished in any of several ways, each having different mar-
ket and non-market effects. Next is the problem of deciding how much
flood control is needed and how this control should be accomplished. A
most efficient flood control plan (highest net monetary benefits) might
not be the most desirable when its effects on the physical and social
environment are considered. The choice of project depends on the deter-
mination of relevant effects to be considered in project evaluation,
whether the effects will be beneficial or adverse, and the relative im-
portance of the different effects. In this section, we discuss the role
that national goals play in delineating and evaluating relevant project

Classifying Project Effects

In general, the effects on a project can be placed in one of
three broad categories -- the economic environment, the physical environ-
ment, and the social environment. Though these three environments are
interconnected and should not in reality be considered as separate, this
classification of effects is made for epistemic reasons.

Economic Effects

Economic effects of a resource investment project will take place
primarily in the geographic region in which the project is constructed.
Regions related through trade or physically (such as being in the same
river basin) will also be affected. The extent of national as well as
regional economic effects from a project depends, among other things, on
the size of the project and how much the project is subsidized by the
Federal Government.
National and regional effects will occur both from construction
of a project (the direct effects from investment expenditures) and from
the fulfilling of project objectives. These effects will be discussed
in more detail in Section IV.
In general, a watershed project may affect the following areas
of the economic environment:
1. prices
2. incomes
3. employment
4. structure of industrial organization
5. taxes
6. other monetary and financial variables
It would be difficult to measure a project's effects in all these eco-
nomic variables, both regionally and nationally, especially since only
some effects may be of interest to the decision maker. In addition,
economic variables may change because of conditions other than those
created by the project. In order to determine the economic effects of
a project, we need to isolate the effects on the economic variables that
are attributable to the project alone. For decision making, we should be

able to predict the relevant effects before a project is built. This can
be done by building a model of the economic system, containing the perti-
nent features to be studied, which can assess the impact of a resource
investment project on these features.
Economists generally agree that an analysis of economic variables
both with and without a proposed project is essential to the project's
evaluation. That is, the same model, assumptions, and procedures should
be used to estimate variables in both cases. A general equilibrium model
including all prices and economic structures would be needed if we were
to best estimate changes in all the variables. This would not be practi-
cal, however. Thus, we must simplify the assumptions to be made when we
build the model. For instance, since we are only interested in effects
due to the project, we make a ceteris paribus assumption that all vari-
ables except those pertaining to the project are constant. For example,
we assume that all technologies and economic structures except those con-
nected with the project are unchanging and that prices are constant.
Those assumptions could be justified for a small region and a small pro-
ject, with the reasoning that what happens in the region would not affect
the general price level and technology. One might question the validity
of predictions of economic effects based on models with such assumptions.
There are thus two problems connected with determining project effects:
not only must we identify the relevant effects of a project, but we must
also decide how good a model we need to predict these effects.
The choice of a model depends on our use of the information we
obtain: Will it be used primarily to predict or to evaluate? We need a
method of comparing alternative projects that fulfill certain objectives.
SIf we use the same assumptions and model to estimate the effects of each
plan, we at least have consistent measures of the effects in terms of
those assumptions, and so have a basis for comparison of projects under
those assumptions. If we use such models to evaluate and rank several
projects, it is not so disturbing that effects of a project may differ
somewhat from those predicted, because of changes not included in the
model. Of course, the separation between prediction and evaluation is
not so precise: in order to make good evaluations, we must be able to
make good predictions and must not underestimate important effects. The
assumptions chosen are important: the more realistic the model assumptions

are, the better the predictions, and the better the basis for evaluation.
For this reason, we discuss in Section IV the need to build a better
model for estimating regional economic effects than the input-output
model or multiplier models frequently used.

Physical Effects

Above, we discussed problems of determining and measuring a pro-
ject's effects on the economic environment. A project will also affect
our physical and social environments. Predicting such effects is not the
primary concern of this paper (and it would require biologists, physical
scientists, and other social scientists); nevertheless, we will mention
them briefly in order to discuss their relevance to the decision making
Effects on the physical environment might include changes in:
1. water characteristics (distribution, quantity, quality,
2. land characteristics (quality of soil, usage, etc.)
3. air characteristics (distribution, quantity, quality,
4. wildlife and other biological relationships
Changes in the physical environment and economic environments are inter-
related. For instance, decreases in water quality due to a project may
increase water treatment costs for the local government, hence necessi-
tating higher taxes. Conversely, changes in prices, availability of
goods, and productivity of land and water due to a project will affect
land and water use, and hence determine land and water characteristics.

Social Effects

The social environment refers to man's relation to his society
and to extra market values. Effects on the social environment may include
changes in:
1. general health and well-being
2. recreational and aesthetic opportunities
3. distribution of population

4. distribution of income
5. employment opportunities
6. educational and cultural opportunities
The social environment is related to both economic conditions and the
physical environment. For instance, changes in the distribution of in-
come and employment affect both the economic and social environments, as
does a decline in general health level of the population. Recreational
and aesthetic opportunities are affected by changes in the physical envi-
ronment. Conversely, population movements affect economic costs and the
physical environment.
Clearly, we cannot look at any of these three environments in
isolation. Changes in any one will likely produce changes in the others.
To predict the complete stream of effects from a project would require
quite an extensive theory and complicated models, including models of the
economic and physical systems and population movement. Usable predictive
socio-economic, physio-economic, and socio-physical models are needed
before we can analyze the interrelationships among the three environments.
For this reason, we must simplify our assumptions and limit the effects
we will consider in decision making.

The Role of National Goals

The question of just which effects to measure in the three areas
remains. As we discussed in the case of economic variables, it would be
impractical and costly to try to measure all possible effects. We must
restrict ourselves to those that most concern us. This leads to the
question of how we decide just which variables to include. Clearly, the
choice depends on our national goals. That is, if we have a concern (and
hence a national goal of some sort) relating to income distribution, we
need to measure the effect of an investment project on income distribu-
tion. The selection of variables to be measured is not a static matter.
As national goals change, the effects that concern us will also change.
For instance, our national goals were formerly so concerned with growth
and national income expansion that effects on environmental quality were
not considered. Now, as resources such as clean water become scarce, we
find we also need goals concerning environmental quality; hence we now

need to measure the effects of investment on such variables. Thus, there
is an inter-connection between effects and goals: our goals determine
which effects we consider important, but as some effects become of greater
concern, our goals change.
There has been some confusion as to the goal or objective of a
resource investment project. Some researchers feel that resource invest-
ment objectives should not include broad social and economic objectives.1
For instance, Back says [1, p. 1443]:
The limited scope and content of physical objectives--
for example, usually what can be obtained by a system of
reservoirs in a watershed or river basin suggests
that water projects may be inefficient or ineffective as
public instruments for achieving many of our major eco-
nomic objectives . The failure to fully and expli-
citly recognize such limitations of water projects could
deter public consideration of more relevant programs and
policies for attacking our major economic problems .
By this argument, income redistribution would not be a proper objective
for a watershed project. However, the effects of a watershed project on
income distribution must be considered in the light of national goals.
If a project has an effect of interest to us, there must be a national
goal relating to this effect, even though the goal is itself not a pro-
ject objective.
The words, goals and objectives, are sometimes used interchangea-
ably. However, Lord [14] makes distinctions between national goals and
investment objectives. Goals, he says, are
general statements of the aims of public policy .
They are the ultimate standards against which public policies
and programs in water resources, as all other areas, must be
evaluated .
On the other hand, Lord defines objectives as being more specific state-
ments relating to what specific programs are meant to accomplish. For

In remarks during his talk [23], Harry A. Steele, Associate Director,
Water Resources Council, discussed the Water Resources Council's view of
project goals. The Water Resources Council has proposed that social objec-
tives should not be part of resource investment objectives though social
effects should be displayed for a project. The Task Force Report [20], on
the other hand, does feel that social objectives are proper for resource


instance, the objectives of a watershed project might include flood con-
trol, recreation, irrigation, and so on.
Our view of the role of goals and objectives is consistent with
Lord's. The role of broad national goals is to define.and evaluate rele-
vant project effects. These goals need not be the same as project objec-
tives. To avoid the confusion between the words, "goals" and "objectives,"
we will reserve the word "objective" or "supply objective" to mean speci-
fic purposes of a project in terms of supplying resource-connected goods
to a region (flood control, recreation, irrigation, electricity, and so
on); we will use "goals" to denote more general statements of national
What then is the role of national goals in evaluating the effects
of a project? One role has already been mentioned, that of determining
relevant effects of a project. An additional role is determining whether
a given effect is adverse or beneficial (cost or benefit) as defined by
the goals. For instance, a project which brings about a redistribution
of income from the poor to the rich might be considered an adverse effect
in light of national goals. Finally, goals are used to evaluate projects,
since information on how well a project satisfies national goals, in addi-
tion to fulfilling project supply objectives, can be used in ranking pro-

Defining National Goals

We have discussed the role of national goals in project evaluation.
We should now identify these national goals. Corresponding to the three
* areas of effects defined above, we list three broad areas of goals:
1. enhancing the economic environment
2. enhancing the social environment
3. enhancing the physical environment

2This use of goals to define and evaluate relevant effects is consis-
tent with the aims of the Task Force in its proposal to evaluate project
effects in terms of accomplishing multiple national goals. It is also
consistent with the Water Resources Council proposals as described by
Steele since we do not say these broad goals should be the objectives of
resource investment.

These goals refer to goals on the regional as well as national levels.
In contrast to these three categories, the Task Force Report [20] to the
Water Resources Council lists four broad areas of national goals. These
are: enhancing national economic development, environmental quality,
social well-being, and regional development, which includes environmental
and social goals for a region. This categorization does not seem as logi-
cal as the one we present.
Of course, the ultimate goals for public policy are to promote
social well-being. As such, all economic and physical goals are also
social goals if by "social" we mean goals pertaining to the quality of
human life. For instance, we wish to enhance the quality of our physical
environment not only for its own sake but also because the lack of quality
affects our enjoyment of life, at least in the long run if not in the
short run. We wish to have a healthy economy not for its own sake but
because our continued consumption of goods and services, and hence enjoy-
ment of life, depend on it. Because of the relationship of physical and
economic goals to ultimate social goals, it is sometimes difficult to
avoid seemingly double-counting when listing effects under these three
areas of goals. The reason for dividing goals in this way is largely
epistemic, as was the categorization of effects.
The above delineation of goals was very broad; we shall now be
more specific.

Enhancing the Economic Environment

This category includes goals with respect to economic efficiency
,(that is, increasing the value of total output), income, improving the
distribution of income, employment, improving the distribution of employ-
ment, and economic growth. The inclusion of economic efficiency and growth
under economic goals is more obvious than the others. Employment is in-
cluded because unemployment is not only an undesirable social condition;
it is also an undesirable economic condition with respect to consumer
demand. The redistribution of income, also a social goal, can be included
under economic goals since redistributing income to lower income groups
can contribute to increased demand, hence to increased output. These
goals refer to both the national and regional levels.

Contributions of a given project toward regional growth and devel-
opment are difficult to measure a priori. Growth theories tell us that
certain factors such as roads, water, and sources of power are necessary
for development. However, the presence of such factors does not guarantee
growth, since growth also depends on demand for regional products. For
instance, power projects have not stimulated growth in Appalachia. Cor-
respondingly, one cannot say whether lack of flood control will retard
regional growth or whether flood control will promote regional growth.
As Back says [1, p. 1446]:
the knowledge base necessary for developing defen-
sible procedures for estimating the contributions of
water projects to regional economic development does
not exist .
In fact, Howe [9, p. 488] has concluded, from an econometric model relat-
ing water resource investments to regional economic growth, that:
water resource developments are likely to be poor
tools for accelerating regional economic growth if mar-
kets, factor availabilities, and other amenities of
living are lacking .
Therefore, we must guard against justifying a project mainly on the grounds
of enhancement of growth potential.
A given investment project may have opposing effects with respect
to goals in this category. For instance, a project which best satisfies
a regional growth goal may not satisfy national economic efficiency. Or
a project beneficial to income in one region may result in losses in in-
come to another region. The problem is complicated when considering
regional objectives, since more than one region will be involved. That
is, a regional investment project may appreciably affect not only the
Given region but also neighboring regions; therefore, the effects on these
other regions must be included in the evaluation.
Effects on economic goals are generally measurable in dollars.
Even so, they may not be directly comparable with one another, as when
comparing a loss of income to one region with a gain in another. Thus
we cannot judge a project's success in fulfilling economic goals without
judging the relative importance of the various components of economic

Enhancing the Physical Environment

This category includes enhancing and protecting wildlife habitats,
recreation, and wilderness areas and the proper management of our natural

Enhancing the Social Environment

This category includes enhancing the security of life, health,
and property; upgrading educational, cultural, and recreational opportu-
nities; and improving the distributions of income, population, and employ-
ment. As with economic goals, there may be tradeoffs between some of
these goals. That is, life could be made very secure but this might in-
volve some restrictions on the use of leisure time. For example, to secure
against flood damage, rivers could be channeled, their courses altered, and
so on; but this would drastically alter recreation and aesthetic values.
There may also be tradeoffs between employment and population goals. We
may desire to have a wide dispersion of population but employment oppor-
tunities may be greater when population is not dispersed.

Tradeoffs Among Goals

Not only are there tradeoffs between goals within a category,
there are also tradeoffs among the three categories. A project that
increases total income may adversely affect income, population, and employ-
ment distributions. A project that increases national output may decrease
environmental quality. A project that increases educational opportunities
may not be economically efficient. Many more examples of such tradeoffs
are apparent.
The problem with comparing such tradeoffs is that we lack common
units of comparison. How do we compare a recreation opportunity with an
educational opportunity; the preservation of a bird species with the pre-
vention of X dollars of flood damage; an increase in the number of jobs
for a given income class with an increase in regional incomes or with the
loss of ten miles of wild river? Such effects relating to different goals
clearly cannot be directly compared.

The first step in making comparisons is to define for each type
of effect a unit of measurement with respect to a goal. Signs could then
be attached to the effects measured, to indicate whether an effect was
positive or negative. Next, relative weights of importance would have
to be assigned to goals. These weights would depend on the units of mea-
surement: for instance, a weight of .5 on the goal of increasing the
number of jobs by one and a weight of .2 on a dollar's worth of flood
damage reduction would imply that one job is equivalent to 2.5 dollars
worth of flood damage reduction. By this process of defining measures
of goals and assigning weights to goals, we define the social welfare
function. The social welfare function thus defined can be used to rank
investment projects and choose the best one.
The use of weights will be discussed in Section V. In Section
III, we will consider in more detail the relationship of income and employ-
ment goals.


There has been some concern in recent years that public policy
should not only increase regional income and employment in general, but
more specifically should increase income and employment among certain
underprivileged groups. Thus, distribution of income and employment,
as well as their amounts, has become increasingly important. Although
increasing regional income and employment are often not incompatible
with improving their distribution, it is certainly possible to increase
regional income without increasing income to underprivileged groups.
Similarly, the total number of workers employed may be increased without
substantially increasing the number of workers in such groups. For
instance, an electric power plant may increase employment for engineers
but for relatively few manual workers. The purpose of the following
discussion is to clarify the distinction between increasing total income
and employment and changing their distributions.

Total Income and Income Distribution Defined

The total income3 of all persons residing in a region is the
sum of the total income received by employed persons, plus the total
Income received by unemployed persons:
y = Yl+ y2 (3.1)

3In an accounting sense, "total income" here refers to total personal
income and not regional income. Personal income is the sum of all income
payments received, in the region, including transfer payments and divi-
dends paid out of corporate profits. Regional income in an accounting
sense refers to all income originating from business, government, and
household sectors in the region plus income originating from outside the
region. Regional income would include corporate profits but not transfer

Total income can be further categorized into labor groups. The
categorization by groups could be made by occupation, by income class,
by industry, or by function. A categorization by occupation, such as
the U. S. Census of Populations uses, could be
1. professional (engineer, technical)
2. managerial
3. clerical
4. sales workers
5. craftsman and construction workers
6. operatives (trucking and other transport)
7. service workers
8. laborers
A categorization by industry group could be
1. agricultural, forestry, fishery
2. mining
3. construction
4. manufacturing
5. transportation, communication, utility, and other services
6. wholesale and retail sales and services
7. financial services
The choice of categorization depends on the type of policy to be studied
and the type of model to be used in studying regional economic effects.
Regardless of the categorization of labor to be used, total income
can be further defined by
SE (Yil + 2) (3.2)
where y.i is the total income of persons employed in labor sector i, and

yi2 is the total income of unemployed persons in labor sector i. Also,
by definition
yil il (1 ui)ni (3.3)
yYi2 -yi2 (ui)ni (3.4)
where Yil is the average income of employed persons in sector i, ui is
the unemployment rate in sector i, ni is the total possible labor force
in sector i, and yi2 is the average income of persons unemployed in
sector i. By substituting (3.3) and (3.4) for (3.2) we can define total
income as:

y E s [il(l-ui)ni + 12(ui)ni] (3.5)
Equation (3.5) defines the region's total income in terms of income and
employment by sectors.
The distribution of income is illustrated by a bar graph showing
the average income level for each labor group, yil and yi2, versus the
number of workers in that labor group, (l-ui)ni and uini as shown below:

Average income
of labor group

Number in ni
labor group
Figure 1. Income distribution.

The area of each bar gives the income of the corresponding labor group
and the total area of the bars is total regional income.
Under constant wage rates and welfare payments (i. e., yil and
Y92 constant for each labor sector) and no retraining programs (ni con-
stant), a change in the distribution of income could only occur by
changing the number of people employed in each class (i. e., changing ui).

Total Employment and Employment Distribution

Total regional employment is given by
E(l-u)ni = n. Suini = N u (3.6)
ii i
where N is the total labor force and u is total u employment. Expression
(3.6) relates the number of workers in each category and the rate of
employment to total employment. The distribution of employment is illus-
trated by a bar graph, showing the employment rate (1-ui) for each sector,
versus ni. A change in the distribution of employment occurs if we

change the unemployment rates, ui. The area of each bar gives the number
of people employed in that sector; the total area of the bars is total
regional employment.



Figure 2. Employment distribution.

Regional Income and Employment Goals

Income Goals

Expression (3.5) gives the relationship of employment and income
in each sector to total income. Total income can be increased by increas-
ing the wage levels, by increasing employment rates, or by shifting
labor from one sector to another through retraining. A water resource
project would probably be concerned with only the first two of these
Though the above methods increase total income, y, an increase
in y does not indicate what happened to personal income or employment.
This is due to the aggregate nature of y as a measure of personal income
and employment. A project which increases y may do nothing for and per-
haps may even harm the economic position of some group. Thus, it is
important to consider the distributional aspects of a resource investment
project as well as its effect on total income.
Funds for a resource investment could be distributed in several
ways, depending on the type of project. Haveman and Krutilla [8] have
studied the effects of several types of projects on income to industrial

groups. They thus introduce the question of what type of investment
project should be chosen, given the different implications on income
distribution from different projects. As discussed in Section II, it
is believed that income distribution implications should not be the sole
guide in project evaluation, but should certainly be considered.
If a policy Maker is concerned with income distribution, he can
approach the problem in several ways. For instance, he can select an
investment project that will increase the income of one group, say, group
k, without decreasing the incomes of other groups. (Such constraints may
be a political necessity.) This may be formulated as
max kl(-uk)n
s.t. yil(l-ui)ni Y(1-u?)ni for each i f k
0 < ui < 1
and budgetary and other institutional.constraints. On the other hand,
the policy maker may choose a project that will increase income to each
group as much as possible, without decreasing income to any group. This
can be formulated as a vector maximization problem:

max Fil(l ul)n

Y 1(I u)n
s.t. Yil(l-ui)ni yl(-u?)ni for all sectors i

and budgetary and other constraints. A solution to the above vector
maximization is also a solution to the problem of maximizing the sum
of weighted incomes,
a y = E aiil( ui)ni (3.7)
subject to constraints, for some choice of weights ai. If the policy
maker decides that each group should be weighted equally, the solution
of (3.7) corresponds to the maximization of total regional income. If
the weights are not equal, the solution will differ from regional income
maximization. If weights to groups are inverse to income (a. = 1/ i),
the problem is one of employment maximization. If only some of the ai
are positive, the problem would be to increase income of only some groups,
with possible constraints on decreasing incomes to other groups. The

expression in (3.7) illustrates the role of weights on income groups in
choosing an investment project. The greater the weight on an income group,
the more income effects from this group will be considered in choosing an
investment project. Expression (3.7) also contrasts the problems of
increasing total income, changing the income distribution, and increasing
employment since each of these problems corresponds to a different choice
of weights to groups.

Employment Goals

Above we noted that increasing employment is not equivalent to
increasing income. If a decision maker is concerned with employment,
there are several associated problems. For instance, he could be concerned
with increasing total employment and hence look for a solution to the prob-
lem of
max (1l ui)ni
s.t. o u < 1
plus budget constraints and other constraints.
Alternatively, he could be concerned with employment in one group,
say group k, subject to no decrease in employment in other groups and
would look for a solution to i
max (1 uk)nk

s.t. u < u? for each i

plus other constraints.
Or he could be concerned with increasing employment to all groups
to the point where no further increase could take place without decreasing
employment to some group. This vector maximization problem is equivalent
to attaching weights a. to each group and maximizing
Sai(l ui)ni .
If the weights chosen are all equal, this is the same as maximizing total
employment. If the weights are proportional to wage rates, it is equiva-
lent to income maximization. For other values of ae, the solution to the
problem would be different from income or employment maximization. Again,
weights on groups are important in the selection of a project which affects

It is not intended that the above maximization problems concerning
income and employment actually be solved by a decision maker in order to
choose the best investment. Solution of a continuous maximization prob-
lem would require that a continuous spectrum of investment possibilities
be available. Instead, the decision maker frequently has only a few
choices of possible investments for comparison. These problems merely
illustrate the differences and relationships and the role of weights
associated with labor groups in decision making. It is up to the deci-
sion maker to choose the income and employment goals he considers most
important and the definition of weights on relevant labor groups.
The purpose of this section is to point out that increasing
regional income is not synonymous with increasing total employment, and
that increasing regional income is not synonymous with increasing income
to certain groups. The impact of projects on the distribution of employ-
ment and income, as well as effect on total income and employment, should
be measured and considered in decision making. A model that provides
such a breakdown of income and employment effects is presented in the
next section. Labor groups in this model are classified by occupation,
but information on effects on employment by "occupation" is easily con-
verted to information relating to income class by multiplying the number
of employees by the net wage gains. Since the model includes an input-
output formula for intersectoral demands, information on income and
employment by industry groupings can also be obtained.


Effects on both the regional and national levels that should be
considered in evaluating resource investment projects were discussed in
Section II. When studying a regional project, our primary concern may
be the regional effects. However, we also have to study effects on
national goals and on goals pertaining to other regions because of the
tradeoffs which are always involved in such projects. Although this
section is primarily concerned with better measurement of regional eco-
nomic effects, the discussion is intended to fit within the same frame-
work as the discussion in Section II. By considering mainly regional
economic effects in this section, we do not mean to ignore regional non-
economic effects or national effects and effects on other regions. The
measurement of these effects is a topic for other papers.
Economic effects of concern on the regional level include changes
in output, employment, and regional income, and changes in the distribu-
tions of income, employment, and profits realized from a project. A model
to predict these effects will be presented in this section and will be
compared with input-output models.

Delineating Regional Economic Effects

Economic effects were classified in a general way in Section II.
Before we can develop a model, the nature of these effects must be more
specifically delineated.

Construction and Operation Phases

Regional economic (and non-economic) effects of a project occur
both from construction of a project and from its subsequent operation.

For example, we should distinguish between the effects of building a flood
control project and the more lasting economic effects of flood control it-
self. Therefore, we can identify project effects as occurring over at
least two phases: the first is the construction phase and the second is
the operations phase. The concepts of the two phases are illustrated by
the diagrams in Figures 3 and 4 for the example of a dam.
During the construction phase, the primary effects of the project
will be the increased demand for construction services and construction
materials and demands on related sectors. These demands will then gener-.
ate other demands in other sectors. (Such generated demands are often
called indirect effects.) The economic effects of a project on a region
during construction have been extensively studied, for example by Haveman
and Krutilla [8].
Measurement of effects from the operation phase have not received
as much attention as those from the construction phase. During the oper-
ation phase, the physical purposes of a project are being fulfilled. The
operation of a project will cause demands on some sectors of the region.
However, the demand effects will in general differ from those during the
construction phase and may not be significant. For example, after con-
struction of a canal for flood control, the only additional demand on the
construction sector is for maintenance of the canal. The operations phase
is mainly associated with increases in supplies of certain goods. For
instance, a dam on a river may increase the local supply of water, lake
recreation facilities (however, river recreation facilities may decrease),
electricity, and so on. Whether or not such increases actually raise
regional income and output depends on whether there is demand for the addi-
tional goods produced. For instance, an irrigation project may increase
the supply of water for farms and hence allow greater farm production.
Whether or not there is an actual increase in regional income from the pro-
ject then depends on whether there is demand for more of the region's agri-
cultural products.
The above comments on the importance of demand as well as supply
factors for goods supplied by a region are supported by Krutilla [11,
p. 127], who says that development projects should be judged on two counts:
S how effective were the development programs in shift-
ing the supply function of factors on which their efforts
are focused and how strategic, as indicated by their demand
characteristics, were the factors whose supply functions
were thus affected?


Labor Constr

mlld Lt

Expenditure on dar


action Services (housing,
rials transport, etc.)

n I


Non-local expenditures

Labor materials
a 1

Indirect demands generated among sections

The arrows denote demand on a sector due to the project.
The dashed lines define the area of direct local effects.

Figure 3. Construction phase.



- - -



Irrigation Recreation control

Sectors potentially
influenced by in-
creases in supply Agriculture

Effects from inter-
relationship of the
sectors (indirect


Ag. Non



- A

A- J

Other regional and non-regional sectors

The arrows denote the direction of effects.
The lines define the area of direct local effects.

Figure 4. Operation phase.

Water for
Water for coolants &
consumption waste treatment Electricity

SCommunity Comm. Industrial Indus. Comm.

Supply effects



During the operations phase, a flood control project may have the
following effects on the economic structure and technology of affected
sectors of a region:
1. production costs may decrease
2. technology may change
3. output capacity may increase
4. an import activity may be changed to a non-import
one, or even to an export industry
Of course, a decrease in costs depends on either technological change or
a decline in the prices of inputs. A project that provides electric power
at a cheaper rate would have the effect of decreasing production costs. A
technological change (change in the combination of inputs) could occur,
say in agriculture, as a result of irrigation water provided by a project.
A flood control project could increase the amount of timber harvested by
increasing the output capacity of a region's timber industry. A region
that imports some agricultural products, such as dairy products, could
become a producer of such products if a project makes the area more pro-
ductive than other areas.
Technological effects could also occur from the construction phase
if project demands encourage economies of scale and investment in new
equipment. But most technological effects would occur from the operation
phase, since the construction phase is relatively short compared to the .
project life.
Effects such as 1 through 4 above are frequently associated with
regional growth. The model presented in this paper will not have the
capability of predicting growth and technological change. However, some
immediate technological effects during the operations phase can be included,
using knowledge of the region and similar regions. Such knowledge can
also help in predicting the influx of new industries and population.

The Purpose of "Direct" and "Indirect" Effects
in Formulating Regional Models

Above we described regional economic effects occurring during con-
struction and operation phases. Economic impacts of investment projects
are also sometimes categorized according to "direct" and "indirect" effects.
The Task Force Report [20] makes no such distinctions. There is some feeling

among economists that this categorization is not helpful in evaluating
regional projects since it makes no difference to a person receiving bene-
fits or costs whether these are direct or indirect. Similarly, if envi-
ronmental quality decreases because of a project, it makes no difference
whether the cause of environmental deterioration was direct or indirect.
On the issue of distinguishing "stemming from" and "induced by"
effects for project evaluation, Long says [13, p. 18]
there is no useful purpose to be served by a sepa-
ration of changes in income to local residents, other
than direct beneficiaries, into induced-by and stemming
from benefits. The kind of local secondary effects does
not matter. What matters is who benefits, and by how
much, and who pays the cost .
However, for the purpose of formulating predictive models, such distinc-
tions can be conceptually useful. To predict complete regional effects
of an investment, a model should include both causal relationships among
industrial, resource, and household sectors such as those "stemming from"
construction as well as those "induced by" the operations phase.

Input-Output Models and the Measurement
of Project Benefits

Input-output models have been used to estimate income, output,
and employment benefits for two reasons, as discussed by Stoevener and
Castle [24]:
1. Such a model allows determination of secondary invest-
ment effects since it portrays the flow of goods and
services in the economy.
2. The distribution of effects on different labor groups
and sectors can be determined.
Much has also been written about the shortcomings of input-output
analysis in terms of its assumptions mainly with regard to those about
production functions. As Dorfman, Samuelson, and Solow [4] point out,
under the assumption of constant price ratios and technology, and homo-
genity of first degree production functions, the use of constant input-
output coefficients is justified.
It is also assumed in the input-output model that supplies of
foods produced exactly match demands, and that labor and other inputs are
available in sufficient supply. Input-output analysis thus answers the

question of the output capacities and labor amounts needed to meet a given
set of demands. On a national scale, the assumption that supplies can
always be found to meet demands, and vice versa, may be justified. How-
ever, on a local scale and in the short run, output is in fact limited by
the supply of natural resources, labor, and industrial output capacities.4
It therefore cannot be justifiably assumed on a regional basis that any
level of demands on outputs or inputs can be met within the region. Instead,
there will very likely be some excess demands that must be met from outside
the region, as indicated in Figure 1. The input-output model applied on a
regional scale would thus tend to overestimate local income and employment
benefits and output increases.
Many studies of income, output, and employment benefits are con-
cerned primarily with the construction phase of a project.5 Such one-
period models cannot include the effects of changes in technology, prices,
and so on, that might occur during the operation phase. To accurately
estimate effects during the operation phase would require a general equi-
librium model with time dimensions, including regional and national sup-
ply and demand functions for outputs produced and inputs used in the
region. Such a model would be too costly in terms of information required
for the purpose of analyzing small watershed projects.
For practical reasons, any model needs simplified assumptions
about demands and supplies. For example, we may assume that excess input
demands in a region, or excess regional demand for outputs which can be
locally produced, can be met from outside the-region at the current market
price, and that outside demand is sufficient to absorb all supply increases
in regional export goods.6 We may assume that wages and prices are constant

4By output capacity, we mean the maximum amount of output producible
in a year; this term refers both to agricultural and other industrial

5This does not imply that input-output models could not be used in
other ways. However, to study technological and other "induced by"
effects would require a model with time dimensions.

6However, we could have easily included demand limitations for some
export goods by having constraints in our model on the export of these
goods. For instance, if there is an allotment limiting production of
some agricultural crop, say cotton, this can be expressed by a constraint
on the amount of cotton which can be exported at the market price.

and that technology is constant except for industries affected by the
project. Labor supply can be assumed to be given exogenously although it
could also be considered as increasing, depending on demand for labor.
While demands for export goods and demands stemming from construction of
the project are assumed to influence the region, we may assume the region
is small enough so that economic and technological changes will not signi-
ficantly affect national economic structure and prices. That is although
we consider linkages from the nation to the region, we may ignore linkages
from the region to other regions and the nation.
These assumptions could replace the supply and demand relations
mentioned above in order to simplify the estimation of output, income, and
employment benefits from the operations phase. We will use these assump-
tions in the model presented herein. Under these assumptions, benefits
will tend to be overestimated if we ignore price effects from supply in-
creases and assume a market for all exports. Therefore, in using the
results of models such as the one presented here in evaluating a project,
keep in mind the assumptions made.

Measuring Regional Costs

We have been discussing the problems of determining regional eco-
nomic benefits. We now point out some problems in determining regional
costs. Haveman [7] has pointed out the importance of considering regional
income costs as well as regional income benefits to determine net regional
income benefits. The total expenditure on a project is determined with
the selection of the type and size of project. For a small watershed pro-
ject, part of this cost will be subsidized by the Federal Government but
some costs will have to be defrayed by the region, thus causing some
regional income losses. The regional share of the costs may be covered
by increasing general taxes in the region or by some other means such as
a bond issue or allocating costs among project beneficiaries. Just as
income gained by the region from construction of the project must be
adjusted by the region's share of the costs, income benefits to various
groups in the region must be adjusted to account for their share of the
costs. The relationship of expenditures on a project and costs alloca-
tion to net income benefits is illustrated in Figure 5. A cost allocation

scheme, though exogenous to the determination of initial income benefits,
does determine the final allocation of net income benefits and so is impor-
tant to consider.
Not only will there be direct income losses to the region as ex-
plained above; there can also be indirect income losses. If the increase
in taxes overshadows the increase in income for certain groups, these
groups will have less money for consumer items and savings, thus having
some negative effect on the local economy. In any case, the extent of
multiplier effects from project expenditure will depend on how regional
costs are allocated.
Another possible indirect cost stems from the potential shortage
in the supply of local investment capital. The supply of investment funds
in the region will be drawn upon in order to cover the monetary costs of
a project. Construction will then mean that possibly other investments,
both private and public, cannot be made or that capital will have to be
imported. The displacement of other needed investments causes opportunity
costs to be associated with a project. These opportunity costs mean that
the social cost of the project to the region may be higher than its share
of the monetary cost.
Haveman and Krutilla [8] discuss the relationship of opportunity
cost to national social cost of a project. When there is regional underem-
ployment or excess capacity, there will be no opportunity costs associated
with a project and so a project's social cost may even be lower than its
monetary cost if unemployed resources are put to use. However, investment
capital funds for public projects will in general be scarce in an under-
developed region. Haveman and Krutilla ignore opportunity costs in the
area. Opportunity costs to the public sector may be quite important because
of the unwillingness of taxpayers to assume increasing tax burdens to pay
for public expenditures. The issuance of public bonds to pay for a dam
may mean the later defeat of a school bond issue. Such opportunity costs
have not been considered in input-output studies since the government sec-
tor has'usually not even been included in models. Although the model pre-
sented here includes government sector to account for part of the demand
for labor and capital inputs and intermediate goods, it is not sufficiently
detailed to include opportunity costs in terms of foregone roads, schools,
etc. Such costs would be difficult to measure. But the model

Project selection (type and size)

Total cost of project

Costs subsidized
by government
Total project benefits
and allocation of pro-
ject benefits
Unsubsidized costs to region

Allocation of
unsubsidized -

Project net benefits and
allocation of net benefits

Figure 5. Factors influencing net project
benefits ,and their allocation.

does show that, given investment demands by industry and government, an
increase in demand for investment funds by a resource project could neces-
sitate the import of capital, at a cost to the regional trade position.
There will be other indirect regional economic costs stemming
from a project; for instance, ifwater quality decreases, water treatment
costs will probably increase and local personal expenditures for recre-
ation will decrease. If we include indirect economic benefits in pro-
ject evaluations, there is no logical reason to exclude indirect costs.
As mentioned in Section II, interrelationships between the economic, physi-
cal, and social environments generate indirect effects that ought to be
considered in evaluation. Functional relationships and data base should
be developed to include these interrelationships in model building.

A Dynamic Model for Predicting Regional Effects

This section now presents a model for predicting regional economic
effects of an investment project. The model not only includes changes
in demand during the construction phase but, because of its time dimen-
sions, the model can be formulated to include technological changes during
the operations phase. Some of the indirect costs mentioned above are
included in the model. For instance, there is a constraint on the local
supply of investment funds and a limit to the amount of investment capital
that can be imported. Consumer demand is considered as endogenously
related to disposable income, so that economic effects from tax changes
stemming from a project can be included. If growth in the local labor
supply is included, corresponding growth in the cost of local public
services should also be included so that economic costs in terms of the
scarcity of public capital can be considered.
The model is related to an input-output model since fixed input-
output coefficients are used to compute demands for output and labor. It
differs from input-output models in that it includes constraints on the
local supply of output capacity, labor, and natural resources. The eco-
nomic structure of the region can be described by the region's output
sectors, import and export sectors, labor sectors, financial sector, its
technology, and natural resources. Output sectors may include agriculture,
mining, manufacturing services, and so on. Labor sectors may be defined

both by income groups and by skills such as manual, managerial, skilled'
professional, and so on. The model is sufficiently general for use in pre-
dicting income, employment, and output, both with and without an investment.
An illustration of how the model operates, the information inputs
required, and information outputs obtained is contained in Figure 6.
According to Figure 6, the labor and resource supply and technology are
exogenous to the.model. That is, the model does not predict the technol-
ogy or labor supply; these conditions must be given and may be changed from
year to year to reflect changes due to a project. Also, the initial levels
(i. e., the level in year zero with respect to construction of the project)
of income, output capacities, and capital must be given to initiate calcu-
lations with the model. Given initial output capacity, supply of capital,
and final consumption demands (which include demands due to a project when
computing "with") under the current technology, resource and labor supplies,
the output, income, employment, and output capacity are computed for the
current year.
In the model, income from the current year determines the next
year's household demands and the savings available for the next year's
investments. Current profits by industry determine the next year's rein-
vestment by industry. The current year's trade balance (value of exports
minus value of imports) limits the capital that can be imported. These
three factors -- savings, industrial investment, and capital imports --
determine the supply of capital for the next year. The process, is repeated
to obtain the next year's outputs, income, employment, and capacities
under the technology available and resource and labor supplies; by repeat-
ing this process over several years we obtain a time stream of income,
Output, and employment predictions. Thus the model, although it uses
input-output coefficients, differs from an input-output model both in the
use of constraints and because it is dynamic rather than a one-period
analysis. To analyze the situation "with" a project, input-output coef-
ficients may be changed to reflect any expected changes in technology,
and output capacities may be expended -- though at some costs in terms of
labor and capital. Import sectors may also be changed to non-import sectors,
Since the model is recursive and provides a stream of effects, it
can be termed a dynamic model. However, the problem recurrently solved
for each year is in the form of a linear program. The programming

Information Inputs and Outputs
Initial regional economic informant
for year 5, computed from year t-l

Determined exogeneously,
ot predicted by regional Resou
economic model

information computed for year
from solving LP problem

itial information for year

Figure 6. Information inputs and outputs from the regional model.

formulation is necessary because the constraints on resources, output, and
capital make the supply and demand equations non-deterministic. For in-
stance, when demand for an input used in local production or some output
exceeds the local supply, there is a choice between importing more of the
input, or curtailing local production of the output and importing more of
it. The optimal input use, output levels, imports, and exports are deter-
mined for each year by solving the linear programming problem corresponding
to that year.7 This basically describes the operation of the model which
is given in more detail below.

Definition of Variables

In the following description of notation used in the model, the
indices refer to the various sectors. Sectors producing natural resource
products, intermediate goods, and final goods are denoted by j = 2, n.
The capital good is denoted by j = 1, and the different types of labor in-
puts by j = n + 1, n + R. The superscript n on some of the variables de-
notes exogenous effects of a water resource project. The variables denoted
by n are zero for a "without" analysis. The letter "t" below denotes a
time period, not necessarily a year.
Demand and supply factors are:

D total quantity of good j demanded in the region in year t; j = 1, 1'+,

j = household consumption demand for good j for year t; j = 2, n + A

G. = local government demand for good j in year t; j = 2, n + k

Ej amount of export of good j in year t; j = 2, n + Y

7Speigelman, Baum, and Talbert (22] have used a static linear program-
ming model including constraints on labor and resources and minimization
of capital costs. Their model is useful in computing requirements to
meet income and population targets for planning economic development in
rural South Central Kentucky, rather than predicting effects.

F. T. Moore (16] has recognized the importance of capacity limita-
tions and mentions the use of linear programming to determine whether to
add to capacity or to import, though he himself does not do this. His
paper presented a dynamic input-output model.

Mt = amount of import of good j in year t; j = 1, n + k

X = local supply of capital in year t

X" = local output of good j or use of local labor j in year t; j 2, n + A

Lt = supply of labor of type j at time t; j = n + 1, n + 2

AD't = increased demand for good j in the region in year t due to the
water resources project; j = 1, n + 2
Kt = local capacity to produce good j at time t; j = 2, n
AKj't = capacity increase at time t for good j due to a water investment
project, j = 2, n
AK = capacity increase for good j at time t, excluding increases due to
water investment project; j = 2, n

Tt = terms of trade; value of exports minus value of imports

AX't = amount of federal government expenditure on water project in year t

W t = total cost of water project for year t
Technological and consumption coefficients:
a = input-output coefficient for year t giving the amount of j used to"
produce a unit of i; j = i, n + A, i = 2, n

mji = amount of j used to expand capacity per unit i; j = 1, n + 2, i = 2, n

sj = propensity to save by labor type j; j = n + 1, n + X-

= proportion of profit reinvested by sector j; j = 2, n
aj. = propensity to consume j by labor type i; j = 2, n; i = n + 1, n + 2
(Here, only the coefficient aji is assumed to change with time due
to our ceteris paribus assumption.)
Prices and incomes:
I. = profit to output sector j; j = 2, n

= income to labor type j; j = n + 1, n + 2
j. = net profit to output sector j; j = 2, n
= net income to labor type j; j = n + 1, n + k

t. = tax rate of sector j or labor type y; j = 2, n + e

At't = change in tax rate due to water project; j = 2, n + t
p1 = interest rate for investment i' unds
pj = price of good j; j = 2, n

= wage rate for labor type j; j n +.1, n + k

Constraints for the Linear Program

The linear programming problem to be solved for each year under
consideration would be to choose the levels of output, export, import, and
capacity to maximize an objective function subject to the following types
of constraints:8

Dt < Xj + Mj j 2, n + 2 (4.1)
3 J 3
Total demand for the good j or labor type j must not exceed local
output of j plus import of j. This also includes constraints on natural

X < K-1 + AK+ + AK^t j = 2, n (4.2)
J j3 J
Local output of good j must not exceed the given initial capacity
plus any capacity increases. This also includes output of natural resources.
t t
X < L j = n + 1, n + 2 (4.3)
j -
Use of local labor type j is limited by the local supply.

D0 S X + M (4.4)
Demand for capital ii,'.etmn:i is limited by the local supply of
investment funds plus imported capital funds.

M < aTt- + a and constants given for the region (4.5)
The amount of capital which can be imported (or borrowed from non-
local banks) depends on thie previous year's trade balance.

Since positive prices are associated with capital, output goods, and
labor, the inequalities in (4.1) and (4.4) will -in effect be qualities
in the linear program.

These constraints define the possible combinations of exports Ej,
imports M1, outputs Xi, and capacity increases AK Constraints (4.1) and
(4.2) refer not only to production of manufactured goods and agricultural
and timber production but also to production and use of natural resources.
That is, good j may be the output of a mining operation. The output of j
may then be limited by the amount of ore that can be mined in a year, and
capacity expansion could refer to opening new mine shafts. Similarly, good
j could be water for irrigation whose supply is limited by rainfall. The
output capacity of irrigation water could then be expanded by a resource
The rationale behind these constraints seem obvious except for
constraint (4.5). For technical reasons, we must have at least one binding
constraint or the objective function will be unbounded. Since we allow
any level of exports and imports for goods j = 2, n + t, in order to have
a binding constraint, we limit the amount of capital that can be imported.
Capital would probably be the input hardest to obtain, and the total amount
that banks would loan depends on the region's "credit rating" as measured
by the trade balance. (Spiegelman, Baum, and Talbert [22] have binding
constraints on exports, but we find it more credible to have limitations
on capital imports.)
The variables in the constraints above may be explained in more
detail. The demand for good j is given for j = 2, n, by

Dt = Za. X + E m.. AK + E + C + G + AtDt (4.6)
3 i=2 i =2 J

That is, the demand for good j is the sum of the amount of j used in pro-
ducing other industrial and agricultural outputs, the amount of j used in
expanding capacity in various sectors, the export demand for j, the house-
hold consumption demand for j, the demand for j from the government sec-
tor, and the exogenous demand for j due to construction of a resource
investment project. If we assume any amount of export can be sold at a
positive market price, then from the linear program Ewill be the dif-
ference between X + Mt and local demand for good j. (Depending on how
j J t
finely we define the sectors, M can be assumed zero if E is greater
than zero, that is export goods will not also be imported.) Cj, house-
hold demand for j, depends on the previous year's disposable income, that
is, we assume a time lag in income and consumption changes. Formally,

t n+P t-1 (4.7)
Ct = V aji'i
j i=n+l 1
where aji is the propensity to consume j by labor group i and 7i is net
income in year t-l for group i. C. will be zero for intermediate goods.
Gj will be assumed to be exogenous to the model. Government demand for
inputs is of course related to the population and the public service out-
puts offered. (Government demand will not be predicted by this model.)
For j = n + 1, n + A, the demand for labor can be expressed in the same
form as in (4.6) except that E and are zero.
Under the definition of D. in (4.6), except for the inequality
sign, (4.1) is very similar to equations defining demand in an input-
output model except that we have also included capacity expansion demands
and government demands. Since (4.1) will hold with equality for positive
prices, the main difference between this formulation and an input-output
formulation lies in the constraints (4.2), (4.3), and (4.5). In the usual
input-output formulation, the market clearing relations would be solved
to obtain outputs and then the required labor and resources to produce
these outputs would be calculated. Here, by (4.3) the use of local labor
is limited by local labor supply. Demand.for labor of a particular type
may be less than local supply or it may exceed the local supply, causing
the need for imports. An investment which caused a need for additional
imports would exert a negative effect on regional income.
In constraint (4.2), the output of j is constrained by output
capacity, which will be the same as the previous year's capacity unless
it is expanded either by AKn1t from the water resource project or by
plant expansion AK AK"t will probably be zero for most sectors except
agriculture and timber-related in ustries. The level of AK.t is exoge-
nously determined according to the technological situation. The level of
AK" is chosen by the linear program. Since capacity expansion requires
resources as shown in (4.6), capacity will not expand unless demand is
sufficiently high and the good is sufficiently valuable to merit increased
local production rather than importation. Output of a good may also be
below capacity if the product is not sufficiently valuable. In this case,
AK = 0 and any increase AKj due to an investment would cause no in-
crease to regional income.

The variables in constraint (4.4) require a bit more explanation.
DO, the demand for investment funds, can be broken into several parts:
t ,t t + An,t (4.8)
D = 0 + D + AD
I 1,"I 1,R I
D is the demand for funds for industrial investment. D ,R is the demand
by regional government for investment funds, excluding the resource invest-
ment under consideration, and includes demand for funds for building
schools, roads, and so on. If we include population growth at a specific
rate in the model, then D should also be assumed to grow at a specific
rate. ADr't is the demand for investment funds created by the resource
investment project, defined by the regional share of the cost minus the
cost covered by taxes.
^ ^ AX nt 1 (4.9)
AD, W1Dt nt 7. Ain't
S1 j=2 3 3
D is given by an expression similar to (4.6) for j = 1, except that E.
-'i t n t t J
and C are zero, and AD1't is not included in D,..
X1, the local supply of investment funds, has two sources: local
saving and industrial investment funds, so
t n+ t-1
X = s3j (4.10)
where sJ refers to the propensity to save or reinvest. As explained above,
M1, the amount of imported investment funds, is assumed to be limited by'
the region's trade situation.
In order to use constraint (4.5) in the model, the constants a
and 0 must first be estimated for the region. (In estimating these con-
stants, one would fit the equation MiaTtl + 3.) Tt1, the terms of
trade from the previous year is given by that year's value of exports
minus value of imports minus payments on borrowed capital. Tt is defined
n n+ t
Tt PE P.M m1 (4.11)
j=2 j=2
where mt denotes the yearly payment on total regional indebtedness, m
should reflect the fact that borrowing in previous years affects indebted-
ness and hence current terms of trade. Note that imports of labor or
natural resources have a negative effect on T .

Finally, to calculate income to groups and profit to sectors, the
following formulas are used. Profit to output sector j, j = 2, n, (not
including payments on borrowed capital) is given by the value of output
minus payments for inputs and capacity expansion:

t p.X t pa X- i p.m.i AKt (4.12)
j j j Li=j it ij 2 j
Income to local labor groups j, j = n + 1, n + a is given by the amount
of local labor supplied times the wage rate, or
t t
S= p X (4.13)
and the net income and profits after taxes would be

S(1 t t Att) H, j = 2, n + A (4.14)
where At1't is the tax increase due to the local cost of a water resource
project. These values for net income and profit are used to calculate
next year's local consumption demand in (4.7) and saving and investment
in (4.10).

Choice of an Objective Function for the
Linear Program

The linear programming problem, with constraints (4.1) through (4.5)
and non-negativity constraints, is closed with the addition of an objective
function. The levels of X AK E., M. to maximize this objective function
J J J 3
can then be chosen. The selection of an objective function is rather arbi-
trary and depends on how we view the role of this objective function. We
could make the objective function the maximization of regional income, the
minimization of cost of imports, or the maximization of the terms of trade,
and so on. These would all be reasonable as regional objectives. However,
the computed values of the variables resulting from solving such a linear
programming problem would perhaps not predict very well what would happen
in the region under the given circumstances. If we want to simulate actual
choices by industries, we would have to include behavioral functions for
each industry in the region, that is, functions indicating how each indus-
try chooses its inputs and outputs, imports and exports in relation to
prices, and technology. Such a procedure would be complex and require
too much information to be suitable for analyzing resource investment

projects. We suggest use of the terms of trade as the objective function,
since it incorporates some of the behavioral features of a simulation model
while staying within the linear programming formulation. The behavioral
features of using Tt in this way are explained below.
We assume each industrial sector i solves the following problem
to find the profit maximizing level of outputs and inputs.
Max piX- Z p.D. a1 (4.15)
t i j=2 3 ...
X ,D
i j
where D. is the amount of good j demanded by sector i to produce X and ai
J 4 I
is the sector's share of yearly payments on indebtedness ml. We also
assume that the government sector is minimizing costs, i. e., solves
n+ .
Max G p.G. aGmt (4.16)
M j=2 3 J G

Labor groups can also be assumed to offer their services so as to maximize
income Z pjXj. If each sector chooses outputs and inputs to solve
(4.15) and (4.16), then the sum of the objectives in (4.15), (4.16), and
total personal income is also maximized, assuming no externalities. With
the constraint (4.1) binding the following are identities:
n i t t t t t
D + G = Xj + M Et C j = 2, n (4.17)
i=2 J J j J

n i t t t
E D + G X + M j = n + 1, n+ .
i=2 J j J J
Also, by definition
n t t t
n2 ciml + Gm-i ml (4.18)
Using the above identities, the sum of the objectives in (4.15), (4.16),
and total personal income simplifies to
n n+ n+ n+ .
E pX 2 pD. .m pjGj at + X
i=2 j=2 j=2 j=n+


n n+ n n+
= Ep X p. j + G m +Pj
i=2 j=2 i =2 j j=n+l

n n n+A

t n+ t
= Ep X.X m p. XI + M E C S p X + M
i=2 1 j=2 3 3 j=n+1 J I

m E pX

n n+ n
= j pE E.p.Mj+ p.C m
j J j=2 J J j=2 3 J

t n t
= T + E pjC

That is, the sum of regional income and revenue minus cost equals
the trade balance plus the value of current consumption. Since consumption
demand is assumed to be exogenously given from the previous year's income,
maximizing the expression in (4.19) is equivalent to maximizing T Then
Tt is maximized under the no externalities assumption if each sector maxi-
mizes profits, labor maximizes income, and the government minimizes costs
resulting in the maximization of (4.19).
Thus the use of Tt, the terms of trade, as an objective function
t t

has a behavioral basis and its use would mean that the values of Xs, E,
Mt, AKt selected by the linear program may more closely simulate actual
J 3 t
choices than other possible objective functions. The use of T as an
objective function would also serve to relate the production decisions of
each year to the next because of the constraint (4.5). This means that,
if a resource investment causes a positive or negative effect for a given
year on the local economy, this effect would also have an influence on
succeeding years.

Evaluation of the Model and
Concluding Comments

We have described the model for regional economic effects. We
believe it is an improvement over the type of input-output models which
have been used to analyze local effects of water investments. The dynamic
nature of the model allows for the consideration of various effects over
time. For instance, the model could also be formulated to include the
effects of possible technological changes over time.9 The model not only
includes more than just the construction effects of a project; it also
more closely simulates regional economic effects due to the objective
function and constraints on output, capital, resources, and labor supply,
which are not present in an input-output model. Finally, the model allows
considerations of indirect regional costs due to consumption effects and
opportunity costs to both public and private investment.
The data requirements for the model as it now stands are not much
beyond the requirements for an input-output model. In addition to input-
output and capacity expansion coefficients, propensities to save, consume,
and invest are required.10 Initial capacity limits, initial household
demands, and estimates of the labor supply are also needed. Information
is needed on government expenditure per year or government expenditure
per person, if population growth is included. For computing economic
effects "with" a project, demands due to the project and expected techno-
logical changes are needed. The model thus does not seem unreasonable
compared with input-output models such as the Spiegelman, Baum, and Talbert
model [22] which has been implemented for Kentucky.
As we pointed out in Section II, effects on the physical environ-
ment, social environment, and economic environment are interrelated. The

This is not to say that changes in technological coefficients could
not be included in input-output models also. Rather, they usually have
not been included. Spiegelman, Baum,. and Talbert'[22] include input-
output relationships for what they call conversion activities, which
correspond to technological changes in production methods. We go beyond
this since we can also incorporate capacity increases and changes in the
mix of imports and exports.
10Such information can be obtained from regression studies such as
that done by Gibbs and Stoevener [6].

model presented here is sufficiently flexible to allow interrelationships
such as increases in government spending due to increases in population, and
decreases in environmental quality due to increases in output to be included
through the use of input-output coefficients.1
One shortcoming of the model is that it does not fully express the
interrelationships between the region and other regions and the nation as a
whole. An increase in supply of an export good may result in price changes,
which have further repercussions on the region. Economic change in one
region may affect another region that trades with it,causing further reper-
cussions. Furthermore, looking at changes in all regions, there must be a
balance in total imports and exports and a balance in the total use of
primary inputs and the total national supply. To reflect these changes,
repercussions, and balances would require interlocking models for each re-
gion which would be consistent with each other and "add up" to a national
model. The national and regional input-output models presented by Haveman
and Krutilla [8] do exhibit this consistency. Further research is needed
on the means of interfacing the regional model presented here with models
for other regions and the nation.
Finally, we wish to emphasize the importance of studying regional
effects other than just the immediate income and employment effects for
resource investment projects. This importance is reflected in the new guide-
lines for evaluating water resource projects in terms of multiple goals as.
set forth by the Task Force Report [20]. Resource projects have more pro-
found effects than the temporary income and employment benefits resulting
from project construction. The distinction between construction and opera-
tion phases of a project was made with this in mind. If such distinctions
are not made, little attention will be paid to the difference between dams
and highways for regional development. The ultimate impacts of such pro-
jects on social well-being can be quite dissimilar.

11John Cumberland [3] has given an input-output model including coef-
ficients relating air and water quality to output. Using such coefficients,
considerations of environmental quality could be included in the model.


In Section IV, we presented a predictive model for the regional
economic effects of a project. The information obtained from such a model
is not directly useful for decision making; that is, the information obtained
has not yet been related to measures of performance with respect to public
goals. The information on effects obtained from a model must be organized in
such a way that it can be used for decision making. Who makes the decisions
and on what basis should the information be organized? If decisions are to
be made by direct voting or by pressure on officials by interest groups, a
display of information concerning the effectsof the project on interest
groups, such as that presented by Lord [14], would be most suitable If
decisions are to be made by agencies, the information must be displayed in
ways relevant to these agencies' decision criteria. The Task Force Report
[20] calls for effects to be displayed in an account system, showing costs
and benefits in terms of public goals. Since we are concerned with agency
decision making based on fulfilling public goals, we also adopted the accounts
framework for displaying information on effects. This section describes how
an accounts model is set up and used in decision making, and discusses the
related problems.

The Role of an Accounts Model

Perloff and Leven [17, p. 182] discuss the complimentary roles of
predictive models and accounts in policy decision making. They say that
while econometric models offer a means of predicting
future values of independent variables, b themselves they
provide no basis for deriving policy decisions from the pre-
dictions they supply. Without some kind of accounts
model, we have no basis for assessing the effects on the
economy of predicted changes in the final bill of goods.
That is, the accounts model, while it does not have any predictive power,
provides a means of organizing effects from some predictive model so that

several alternatives can be compared for decision making. Another function
of the accounts model pointed out by Leven [12, p. 149] is that, in speci-
fying the accounts framework, the "relevant phenomena" (or, in our terminol-
ogy, effects pertaining to public goals) are identified. Finally, once the
"relevant phenomrn" and their components are isolated, functional relation-
ships can be developed to explain them, and the accounts framework helps in
formulating the explanatory relationships. (For instance, an input-
output model is basically an accounting model.) Thus, an accounts model
plays an important role in developing theoretical concepts as well as
being a tool for project evaluation.

A Comparison of Accounts Models

An accounts model for decision making, corresponding to the classi-
fication of effects given in Section II, is presented.in the Appendix. We
see no need to differentiate between regional and national accounts as the
Task Force does. Regional economic effects are included in the category of
general economic effects, regional physical environmental effects are included
in the category of general physical effects, and regional social well-being
is included under general social well-being. The accounts would also exhibit
the distribution of effects in more detail than the Task Force does [21].
In its distributional aspects, the account model is similar to Lord's system
[14] of displaying impacts, since he proposes displaying impacts, monetary
and non-monetary, by interest groups. Lord's system does not display effects
with respect to goals, as this account model does, since he is primarily
concerned with a project's effects on interest groups.
Objectives relating to goods supplied by a project (irrigation,
power, water supply, flood control, and so on) are isolated from other public
goals by putting them at the top of the account sheet. The reason for this
is not to say that these objectives are more important than the other goals
but that these are the supply objectives for a project. These objectives
are isolated so that the effects of different levels of supply of goods pro-
vided by the project (in other words, alternative project plans) may be

Example of Use of an Account Sheet

An account sheet should be filled in for each of several project
proposals, for use in the decision making process. The example below illus-
trates an account sheet set up as described above, and shows how the accounts
framework is used in decision making. The example includes one supply
objective assumed -- flood control -- and three public goals. The effect
of three different levels of supply of flood control on the goals will be
considered. The three goals are: national efficiency, job creation, and
wild river preservation. The three project proposal s for flood control are
labeled II through IV. Project I denotes the null project, or the status
quo. Each column gives the information corresponding to an account sheet
for a project proposal:
Project Proposal

Project Supply Objectives Supply Level
1. flood control 0 60 50 80

Effects on Public Goals
1. national efficiency 0 -30 0 -60
2. regional employment 0 100 70 110
3. wild river preservation 0 -10 -5 -90
The account sheet shows that,for 60 percent flood protection, Proposal II
would cost $30 in terms of national efficiency (output benefits minus costs
measured in dollars), create 100 jobs in the region, and destroy 10 miles
of wild river. For 50 percent protection, Project III would have zero effi-
ciency gains, create 70 jobs in the region, and cost 5 miles of wild river;
In order to choose the best project from among the four, a decision
maker would have to choose relative weights of importance for the goals.
For instance, he could decide regional employment was 1.5 times as important
as efficiency and twice as important as wild river preservation. This would
lead to weights of 4/13, 6/13, and 3/13 respectively for the public goals.
The supply objective of flood control itself is not given a weight, since to
do so would mean double counting of benefits due to flood control. Flood
control (and other supply objectives) is carried out not for itself but
because of some public goals with respect to property, health, income, and

so on. The importance of flood control is thus included in the choice of
weights on goals pertaining to flood control.
After choosing weights on public goals, each project can be given
a score on the basis of these weights. Under the above weights, the score
for the status quo is zero; for II, 450/13; for III, 405/13; and for IV,
150/13. The ranking of the projects according to their score is then II,
III, IV, I. Thus project II, with 60 percent flood control, is the best
project in light of the tradeoffs between goals. A different choice of
weights on goals would give a different ranking of projects and a different
choice of best project.

Problems of Setting Up an Accounts Model

One of the main problems in specifying an accounts model for
effects is to determine how effects should be categorized with respect to
goals. This problem was mentioned in Section II, when we discussed the dif-
ficulty of separating economic and physical goals from social goals. For
example, although an effect on employment is an economic effect because it
has an impact on the labor market, it also pertains to social well-being.
Thus, should an effect on employment be listed under effects on economic
goals or under social well-being? A similar question could be posed for
income effects. Because of the relevance of income and employment effects
to both economic and social goals, we have listed these effects under both
areas in the account sheet in the Appendix. Although double counting seems
likely, it will not occur if weights on goals are properly chosen. For
example, the relative weights of importance for an income effe t will have
two parts, one giving the relative importance in terms of economic goals
and one giving relative importance in terms of social goals.
Another problem in setting up the accounts is the defining of
relevant regions for which effects are to be measured. The physical and
geographic characteristics of a project could be used to define the most
directly relevant regions. Using such criteria would mean that the defini-
tion of relevant regions and their size would vary for each type of invest-
ment project studies. This could present some data-gathering problems,
particularly if regional input-output coefficients are required to predict
effects. Further questions concern the number and size of regions that should
be included in evaluation of total effects.

Once the accounts are set up in terms of classifying the relevant
effects to correspond to goals and specifying relevant regions, the next
question involves the unit to be used to measure effects. In the example
given above, we were able to measure performance of the proposals with respect
to the given goals in either physical units or dollars. The net benefits
are a measure of performance with respect to efficiency, the number of jobs
created is a measure of performance with respect to employment goals, the
number of miles of river destroyed is a measure of ecological performance,
and so on. Although many effects not measurable in dollars are measurable
in some kind of physical unit, it is not so easy to measure some effects.
For instance, there is a problem of numerically expressing performance with
respect to distribution goals. This is due partly to the lack of a measure-
ment term and partly to the fact that distribution goals are not well defined.
If, for example, a definite income distribution goal could be formulated, a
performance measure could easily be devised. For instance, deviation from
an "ideal" distribution could be used as a performance measure, as indicated
in Figure 7.
Although there is a need to define better such distribution goals,
to do so is not politically feasible and definitely not the responsibility
of a water resource planner. To avoid this problem, goals relating to income
distribution could be expressed only in terms of increasing income to certain
groups, defined as relevant to a project. Employment and population distri-
bution could be similarly expressed. There remains the problem of defining
which groups are and are not relevant.
Another problem of defining performance measures relates to the
distribution of effects over time. For instance, two projects for flood con-
trol may have quite different effects over time on the distributions of
employment. One project may have quite rapid short-term effects on employ-
ment while the other may have effects occurring at a slower rate but over a
longer preiod. To use an accounts system to compare projects with differing
time distributions for effects would require that a discount rate be chosen
for each type of effects.
In benefit-cost analysis, the discount rate is used to combine a
stream of monetary benefits over time into a single benefit score for a pro-
ject. The choice of a relatively large discount rate gives present monetary
benefits relatively more weight than future monetary benefits; choice of a
relatively low discount rate gives future monetary benefits relatively more

Equal distribution of income

00 Income distribution
% .of total / / goal
Current distribution
of income

% of population

(The shaded area measures the deviation from the income distribution goal)

Figure 7. Distribution of income.

weight. Similarly, in order to compare projects having different distribu-
tions of employment effects, a discount rate for employment effects would
have to be chosen to give the relative time rate of preference for employ-
ment effects. These rates could then be combined into a single score.
Such a discount rate would have to be chosen for each type of effect that
would be distributed over time, in order to compare projects having differ-
ent time distributionsof effects.
A final problem in using the accounts model for decision making
lies in the selection of proper weights of importance for goals. There
is no easy guide for this and there will be political consequences if weights
are incorrectly chosen. To result in publically acceptable decisions,
weights should be based on the public interest, which is difficult to define
when several interest groups are involved. The proper choice of weights is
a problem in itself which belongs in the realm of political science and
social theory.
A complete decision process involves setting up an accounts model
of "relevant phenomena," defining performance measures, estimating effects
of proposals in terms of the accounts, choosing weights, and then ranking
proposals according to their scores. This section has discussed this pro-
cess and some of the problems involved.

This decision process can be applied at several levels of evaluation.
First, we can determine the best level for given supply objectives, given the
tradeoffs in public goals. Second, we can ascertain whether or not a given
supply objective is worthwhile.12 Finally, we can compare several types of
public investment -- water resource investment, schools, or sewer plants --
using a consistent accounts model and weights to rank the various types of
public investment with respect to achievement of public goals. Thus the
discussion of this paper is relevant to the whole area of regional develop-
ment planning and choice of public investment.

If scores for all project proposals except the status quo are negative,
this indicates that the status quo is best and the supply objective is not
worthwhile at any level.


This report has dealt with the evaluation of effects of resource
investment projects in terms of multiple public goals. Since a resource
investment project is likely to have positive effects in some areas and
negative effects in others, our view of the net result depends on our
view of the relative importance of the various kinds of effects. The
mixed nature of project effects is the rationale for developing decision-
making procedures in terms of multiple goals and for considering the trade-
offs in accomplishing these goals.
In Section II, we discussed the problem of determining the effects
of a resource investment project. We emphasized the role of public goals
in defining and evaluating relevant effects. Goals and effects fall into
three broad areas: the economic environment, the physical environment,
and the social environment. This categorization was made for epistemic
reasons and it was emphasized that all goals and effects ultimately relate
to social well-being.
Some aspects of multiple goal evaluation have been dealt with in.
more detail than others. In Section III we considered some of the trade-
offs involved, namely between income and employment goals. That section
emphasized the difference between changes in total income and employment
and their distributions. Other topics discussed in detail were the prob-
lems of defining relevant effects due to a project, measuring these
effects, and presenting them in a format useful for decision making.
In Section IV we presented a model to predict regional economic
effects. This model, though including input-output coefficients to
describe interindustry demands, presents a more realistic view of the
regional economy than input-output or multiplier analysis does. This is
due to the inclusion of short-run limitations on the local supplies of
labor and capital and industrial capacity. Such limitations allow us to
obtain a more accurate picture of regional benefits as well as possible
regional costs of resource investment, since an investment project will

increase the demands on limited local supplies of labor, capital, and output
and will compete for investment funds with other private and public pro-
jects. The model also includes time dimensions, so that the stream of
effects from both construction and operation of a project can be studied.
Finally, in Section V we illustrated the role of goals in evalua-
tion. Goals were used to set up an accounts model for the display of infor-
mation for decision making. The supply objectives and the level of supply
are shown at the top of the account sheet, so that the effects correspond-
ing to different levels of supply can be readily compared. In addition,
distributional aspects of project effects are emphasized in this accounts
Throughout this paper we have indicated points of weakness and
areas where more work is needed in developing better models for determin-
ing project effects. Problems for study include the interrelationship of
economic, physical, and social effects o- a project; interregional rela-
tionships; and predictions relating to technological and growth effects.
Also, goals need to be better defined, particularly with regard to distri-
bution effects. Finally, more work is needed on developing a functional
accounts model, and how such accounts can be used in the decision process.



Project Supply Objectives

Effects on Public Goals

I. Economic Environment

A. Efficiency

1. National3

2. Regional

3. Contiguous Region

B. Employment

1. National

2. Regional

3. Contiguous Region

C. Employment Distribution

1. National

2. Regional

3. Contiguous Region

D. Income

1. National

2. Regional

3. Contiguous Region

See p. 60 for footnotes.

Supply Level2

Gains Losses Net

Gains Losses Net

Gains Losses Net

Gains Losses Net

E. Income Distribution4

1. National

2. Regional

3. Contiguous Region

-Gains Losses Net

F. Economic Development (Potential for increases in output and

II. Physical Environment

A. Water Characteristics:

1. Water Quality

a. National

b. Regional

c. Contiguous Region

2. Water Quantity


Gains Losses Net

Gains Losses Net


3. Water


Gains Losses Net

B. Land Characteristics6

C. Biological Characteristics

1. Wildlife7

See p. 60 for footnotes.

Gains Losses Net

a. Wildlife Quality




b. Wildlife Distribution




c. Wildlife Quantity




2. Plant Life

Gains Losses Net

Gains Losses Net

Gains Losses Net

III. Social Environment

A. Health8

Gains Losses Net

B. Population Distribution



See p. 60 for footnotes.

Gains Losses Net

C. Recreation and Aesthetic

D. Income Distribution

E. Employment Distribution

F. Educational and Cultural

Gains Losses Net

Gains Losses Net

Gains Losses Net

Gains Losses Net

See p. 60 for footnotes.

Footnotes for Accounts Sheet

1. This account sheet is not all-inclusive, particularly in the social
and physical accounts, but does indicate how such a thing should be
set up.

2. Supply Level denotes a measure of the supply objective. For example
for flood control, it could be the percent of damage reduced; for
lake recreation, the number of acre-feet of lake available for recre-
ation; for irrigation, the number of acre-feet available for irriga-

3, In this classification of effects, the breakdown of effects into
National, Regional, and Contiguous Region is not necessarily meant
to be exhaustive. Effects in several regions might be pertinent to
a project and should then be included.

4. Income distribution changes should be displayed according to income
class for (a), (b), (c). In addition, the distribution of profits
among industrial sectors should be shown in (a), (b), and (c). In-
come effects on various interest groups should also be shown. For
instance, the destruction of a trout stream would inflict monetary
costs on fishermen and monetary benefits on farmers.

5. Effects on water distribute n would be displayed according to water
user. The monetary effects of this distribution are included under

6. Land characteristics would include land usage or distribution, land
quality with respect to erosion, fertility, arability, and land
quantity referring to land drained or land inundated. Monetary
effects by group affected are included under 1.4.

7. Monetary effects from effects on wild life should be shown by group
affected under 1.4.

8. Effects on health should also be shown in relation to income distri-
bution classes.

9. Effects on recreation should be shown according to types (lake
recreationists, river recreationists, fishermen, hunters, boaters,
hikers, etc.) and according to distribution among income classes.
Monetary effects are shown in 1.4.


I13 Back, W. B. "Estimating Contributions of Natural Resource Invest-
ments to Objectives in Regiona) Economic Development," Am. Jour.
of Agr. Econ. 51 (Dqcember 1969), pp. 1442-1448.

[2] Bromley, Daniel W., William B. Lord, et al. Procedures for Evalu-
ation of Water and Related Land Resouiir' Projects: An Analysis of
the Water Resource Council's Task Force Report. Madison: Univer-
sity of Wisconsin, April 1970.
[3] Cumberland, John H, "The Economics of Environmental Relationships."
Paper presented to the Technical Seminar Series, Department of
Agricultural Economiqs, University of Florida, May 19, 1971.

[43 Dorfman, R., P. Samuelson, and R. Solow. Linear Programming and
Economic Analysis. New York: McGraw.Mill, 1958.

[5] Eckstein, Otto. Water Resource DevelopMent. Cambridge: Harvard
University Press, 1961.
[6] Gibbs, K. C., and H. H. Stoevener. Consumer Incomes and Expendi-
tures in the Newport-Toledo Area. Ore. $t. Univ. Agr. Exp. Sta.
Sp. Rpt. 237. Corvallis, Ore:' May 1967,
[7) Haveman, Robert H. Water Resource Investment and the Public Inter-
est. Nashville: Vanderbilt UniversityPress, 1965.
[8] Haveman, Robert H., and John V. Krutilla. Unemployment, Idle Capac-
ity, and the Evaluation of PublicExpenditures. Baltimore: Johns
Hopkins Press, 1968.
[9] Howe, Charles W. "Water Resources and Regional Economic Growth in
the United States, 1950-1960," Sou. Econ. Jour. 34 (April 1968),
pp. 477-489.
[10] Hufschmidt, Maynard M., John Krutilla, and Julius Margolis. Stan-
dards and Criteria for Formulating and Evaluating Federal Wae-teF
Resources Developments. A report prepared by a Panel of Consul-
tants to the Bureau of the Budget. Washington: U. S. Bureau of
the'Budget, June 1961.

[11i Krutilla, J. V. "Regional Economics,'! Am. Econ. Rev. 45 (May 1955),
p. 127.
[12] Leven, Charles. "Regional Income and Produce Accounts: Construction
and Application," in Werner Hochwald (Ed.). Design of Regional
Accounts. Baltimore: Johns Hopkins Press, 1961
"r- f^ s/1'5 *

1133 Long, Burl F. "Concepts and Theoretical Basis for Evaluation of
Secondary Impact," in U. S. Economic Research Service. Secondary
Impacts of Public Investment in Natural Resources. U. S. epart-
ment of Agriculture Misc. Pub. 1177. Washington: U. S. Govern-
ment Printing Office, August 1970.

[14) Lord, William B. "Information Requirements for Environmental Deci-
sion Making." Paper presented to the Technical Seminar Series,
Department of Agricultural Economics, University of Florida,
March 19, 1971.

[15) Marglin, Stephen A. "Objectives of Water-Resource Development: A
General Statement," inA.. Maass et al. Design of Water-Resource
Systems. Cambridge: Harvard University Press, 1967.

[16] Moore, F. T.
(May 1955).

"Regional Economic Reaction Paths," Am. Econ. Rev. 45

[17] Perloff, Harvey S., and Charles L. Leven. "Toward an Integrated
System of Regional Accounts: Stocks, Flows and the Analysis of
S/- .the Public Sector," in Werner F. Hirsch (Ed.). Elements of Regional
'. cAccounts. Baltimore: Johns Hopkins Press, 1962.

[18] The President's Water Resources Council. Policies, Standards and
Procedures in the Formulation of Plans for Use and Development of
Water and Related Land Resources. Senate Document 97. Washington:
U. S. Government Printing Office, May 1962.

[19] Skold, Melvin D. and Arthur J. Greer. The Impact of Agricultural
Change on a Local Economy in the Great Plains. Colo. State Univ.
Agr. Exp. Sta. Tech. Bul. TB-106. Fort Collins, Colo: October

[20] Special Task Force of the Water Resources Council .
Planning Water and Land Resources. Report to the
Resources Council, Washington, D. C., July 1970.

Standards for
U. S. Water

[21] Principles for Planning Water and Land Resources.
SReport to the U. S. Water Resources Council, Washington, D. C.,
July 1970.

[22] Spiegelman, Robert, G. E. L. Baum, and L. E. Talbert. Application
of Activity Analysis to Regional Development Planning: A Case
Study of Economic Planning in Rural South Central Kentucky. U. S.
Department of Agriculture, Tech. Bul. 1339. Washington: U. S.
Government Printing Office. March 1965.

[23] Steele, Harry. "Multi-objective Planning of Water
Paper presented to the Technical Seminar Series,
cultural Economics, University of-Florida, April

and Land Resources."
Department of Agri-
26, 1971.

[24] Stoevener, H. A., and E. N. Castle. "Input-Output Models and Benefit-
Cost Analysis in Water Resources Research," Jour. of Farm Econ. 47
(December 1967), pp. 1572-1579.

[25] Subcommittee on Evaluation Standards of the Inter-agency Committee
on Water Resources. "Proposed Practices for Economic Analysis of
River Basin Projects." (Green Book) A Report to the Inter-agency
Committee on Water Resources. Washington: U. S. Government Print-
ing Office, 1958 (revised).

[26] U. S. Department of Agriculture, Soil Conservation Service, and
Forest Service. Work Plan for California Lake Watershed, Dixie
County, Florida. Report Prepared by the Dixie Soil and Conserva-
tion District, Dixie County Board of Commissioners. U. S. Depart-
ment of Agriculture-Soil Conservation Service Publication No.
4-24191, June 1967.

[27] U. S. Water Resources Council. Procedures for Evaluation of Water
and Related Land Resource Projects. Washington: U. S. Government
Printing Office, June 1969.


[28] Back, W. B. "Appraisal of Regional, Interregional, and National
Effects of Resource Development by Interindustry Analysis," in
U. S. Economic Research Service. Secondary Impacts of Public
Investment in Natural Resources. U.'S. Department of Agriculture
Misc. Pub. 1177. Washington: U. S. Government Printing Office,
August 1970.

[29] Back, W. B., and John E. Waldrop, Jr. (Eds.). Regional Studies of
Income Distribution. A report prepared by the Department of Agri-
cultural Economics and Agribusiness of Louisiana State University.
Baton Rouge, La: March 1968

[30] Bromley, D. W., and H. H. Stoevener. "The Use of Input-Output Models
in Community Growth and Development Policy." Preliminary copy of a
paper presented to the Committee on the Economics of Range Use and
Development, Western Agricultural Economics Research Council.
Tucson, Ariz: November 1968.

[31] Castle, Emery N., Maurice Kelso, and Delworth Gardner. "Water
Resource Development: A Review of the New Federal Evaluation
Procedures," Jour. Farm Econ. 45 (November 1963).

[32] Chenery, H. B., and P. G. Clark. Interindustry Economics. New York:
John Wiley and Sons, 1969.

[33] Cox, P. Thomas. "Statistical Analyses of Growth and Development Pro-
ject Areas or Potential Project Areas," in U. S. Economic Research
Service. Secondary Impacts of Public Investment in Natural Re-
sources. U. S. Department of Agriculture Misc. Pub. 1177. .7ish-
ington: U. S. Government Printing Office, August 1970.

[34] Eddleman, B. R. "Estimating the Effects of Resource Development
Programson Regional Employment," Am. Jour. Agr. Econ. 51 (December

[35] Gadsby, Dwight M. "Current Procedures Used in Evaluating Resource
Conservation and Development Projects," in U. S. Economic Research
Service. Secondary Impats of Public Investment in Natural Resources.
U. S. Department of Agriculture Misc. Pub. 1177. Washhngton: U. S.
Government Printing Office, August 1970.

[36] Gardner, B. Delworth, and Allen LeBaron. The Economics of Water
Resource Development, Vol. II--The Proceedings of a Sumrmer Institute
in Water Resources. Logan: Utah State University, April 'ib9.

[37] Gardner, B. Delworth. State Water Planning--Goals and Analytical
Approaches. Utah St. Univ. Agr. Exp. Sta. Bul. 463. LoganT Uah:
November 1966.

[38] Gibson, W. L., Jr., R. J. Bildreth, and Gene Wunderlich (Eds.).
Methods for Land Economics Research. Lincoln: University of Neb-
raska Press, 1966.

[39] Goode, Frank M., and George S. Tolley. "The Input-Oututt 1odel as a
Tool for Regional Analysis," in Gibson, W. L., Jr., R. J3 Hilri'eth,
Gene Wunderlich, (Eds.) Methods for Land Economics Research.
Lincoln: University of Nebraska Press, 1966.

[40] Haefele, Edwin T. "Social Choices and Individual Pr-,..c.- :es: Is
There a Connecting Mechanism?" A paper presented to the Technical
Seminar Series, Department of Agricultural Economics, Urnversity
of Florida, February 23, 1971.

[41] Halter, A. N., and S. F. Miller. River Basin Planning--A Simulation
Approach. Ore. St. Univ. Agr. Exp. Sta. Sp. Rpt. 224. Corvallis,
Ore: November 1966.

[42] Harrison, Robert. "The Role of Secondary Effects in Project Formu-
lation, Justification and Evaluation for the Appalachian :ear Re-
sources Survey," in U. S. Economic Research Service. Secondary
Impacts of Public Investment in Natural Resources. U. S Depart-
ment of Agriculture Misc. Pub. 1177. Washington: U. S. Govern-
ment Printing Office, August 1970.

[43] Hartman, L. M. "The Input-Output Model and Regional ''ter agment."
Jour. Farm Econ. 47 (December 1965).

[44] Heneberry, William H. "Evaluation of Secondary Benefits from the
Accelerated Watershed Program in Appalachia," in U. S. Economic
Research Service. Secondary Impacts of Public Investment in Natural
Resources. U. S. Department of Agriculture Misc. Pub. 1177.
Washington: U. S. Government Printing Office, A.ugrt 1970.

[45] Hufschmidt, M. M., and M. B. Fiering. Simulation Techniques for Design
of Water Resource Systems. Cambridge: Harvard University Press,
"i9. -- --2

[46] Isard, Walter et al. Methods of Regional Analysis: An Introduction
to Regional Science. Cambridge: The M.I.T. Press, 1967.
[47] Johnson, William K. "Assessing Environmental Impacts of Resource
Development." A paper presented to the Technical Seminar Series,
Department of Agricultural Economics, University of Florida,
April 19, 1971.
[48) Krutilla, J. V., and Otto Eckstein. Multiple Purpose River Develop-
ment. Baltimore: Johns Hopkins Press, 1958.
[49) Lacewell, Ronald D., and Vernon R. Eidman. "A Proposed Procedure
for Distributing Assessments Among Beneficiaries of Small Watershed
Projects." (Mimeographed).

[50] Maass, Arthur et al. Design of Water-Resource Systems. Cambridge:
Harvard University Press, 1962.
[51] McKean, R. N. Efficiency in Government through Systems Analysis.
New York: John Wiley and Sons, 1958.
[52) McKean, Roland N. Public Spending. New York: .icGraw-Hill, 1968.

[53] Prest, A. R., and Ralph Turvey. "Cost-Benefit Analysis: A Survey,"
in Surveys in Economic Theory, Volume III. New York: St. Martin's
Press, 1968.
[54] Sewell, W. R. D. et al. Guide to Benefit-Cost Analysis. Ottawa:
Queen's Printer7'1962.
[55] Spindletop Research Center. A Model for Sequencing Public Investment
Programs and Allocating Multi-Program Benefits, P' se I. A report
prepared for the Department of the Army Corps of Engineers, Office
of Appalachian Studies. September 1967.
[56] Spindletop Research Center. A Model for Sequencing Public Invesment
Programs and Allocating Multi-Program Benefits, PhaseT-I. A report
prepared for the Department of the Army Corps of Engineers, Office
of Appalachian Studies. September 1967.
[57] Steele, Harry A. "Multiobjective Planning of Water and Land Resources,"
Paper presented before the Technical Seminar Series, Department of
Agricultural Economics, University of Florida, April 26, 1971.
[58] Steiner, P. 0. "Choosing Among Alternative Public I:1,-estments in the
Water Resource Field," Am. Econ. Rev. 49 (December 1959).
[59] Stewart, Clyde E. "Secondary Benefit Evaluations in Federal Recla-
mation Programs," in U. S. Economic Research Service. Secondary
Impacts of Public .Investment in Natural Resources. U. S. Depart-
ment of Agriculture Misc. Pub. 1177. Washington: U. S. Govern-
ment Printing Office, August 1970.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs