5, Welfare Economics and Water Allocation
In an attempt to develop criteria for public policy, economists have de-
veloped a branch of normative economics called "welfare economics." More re-
cently, this branch has become known as the "new" welfare economics to empha-
size its development in England and the United States since the 1930's. Its
essential problems were recognized and its relevant theorems developed in the
1890's by Pareto.1/
In formulating policy criteria, welfare economics takes explicit account
of differences in individual preferences and incomes and of the resulting prob-
lems in aggregating individual utilities. It is an economic axiom that the
marginal utility of individual income decreases with increasing income. There
is no agreement among economists on whether and in what sense--ordinally or
cardinally--individual utilities can be compared; but welfare criteria that
avoid interpersonal comparisons are generally preferred.
Classical and neoclassical economists were well aware of these problems.
They, however, focused on an increase of (real) aggregate national income as the
main criterion of economic welfare.W Pareto's views were not in conflict with this
1/Pareto, Vilfredo, Cours d'Economique Politique (Lausanne: F. Rouge,
2/The first edition of Alfred Marshall's Principles of Economics appeared
in 1890, seven years before publication of Paretols main work in French.
Marshall mentions Pareto only in passing and in a different connection.
.__________________________________ _______ _____------- W ^^ ^ ^^ ^ ^^ ^ -.........--- V^ -
emphasis because he believed--supported by historical experience as he saw
it--that an increase of aggregate national income and greater equality of
income distribution tended to be associated. In this case, an increase of
aggregate national income means also an increase of economic welfare according
to Pareto's criterion at least under some generally accepted assumptions.
The positive correlation between changes of aggregate national income
and of equality of income distribution--sometimes called "Pareto's law"--was
challenged by Pigou/and others; but Pareto's welfare criterion is indepen-
dent of his "law." This does not imply that the correlation noted by Pareto
does not exist nor that Pareto's criterion is of greater significance for
economic theory than his "law." Quite the contrary, one may wonder whether
the great intellectual effort of the last 20 years which has been invested in
developing Pareto's criterion might not have yielded greater dividends, in
terms of knowledge as well as welfare, if it had been employed for further
investigation of Pareto's law and of the problems associated with the increase
of aggregate national income.
-------------------------rr ---- ---------------
l/Pigou, A. C., The Economics of Welfare (London: Macmillan and Company,
The Pareto criterion says that a change that makes at least one individual
better off and leaves no individual worse off represents an increase of welfare.
This criterion is usually interpreted to mean that welfare is increased by a
change rendering it "possible" to make at least one individual better off and
leave no individual worse off by compensating the losers. Most of the discussion
in the new welfare economies deals with this compensation principle.
The Pareto criterion "without" compensation is so restrictive that it has
little relevance for an appraisal of public policies--even if it could be
practically applied. There are scarcely any policies which make nobody worse
off. Furthermore, if there were such policies, the criterion would be ineffec-
tive for choosing between more than one alternative to the status quo. The
Pareto criterion "with" compensation is not so restrictive, but its application
is even less practical.
The Pareto "with" criterion is conceptually not identical with the criterion
"increase of aggregate national income." But the latter criterion may be regarded
as a practical first approximation to the former, provided that the policy under
consideration does not appreciably increase inequality of income distribution;
and provided further that there are other policies in operation which work
independently and continuously in the direction of greater equality of income
distribution. Such policies are, for example, progression in income and property
taxes, high inheritance taxes, and "social welfare" legislation in the narrower
sense (relating to old age, invalidity, unemployment, minimum wages, and public
health, education, and so on). These two conditions can be regarded as fulfilled
when considering resource policies in modern western societies.
Accepting an increase of aggregate national income as an economic criterion
for public water policy does not imply that application of this criterion faces
no theoretical and practical difficulties or that it is the most useful cri-
terion under conditions where economic change and uncertainty are the central
problems. We shall return to these problems in the concluding section.(Section 6).
The contribution of welfare economics has been a clarification of the theo-
retical meaning (or absence of it) of a social welfare function and social indif-
ference curves and of the difficulties (or impossibility) of applying the Pareto
criterion in actuality.2The disservice of welfare economics has been that its
terminology is used by economists and others without pointing out these theoretical
and practical difficulties. The false impression is created that a simple cri-
terion is available that can be used for legislation, court decisions, administra-
tive regulation, and social planning in general.
In the field of water allocation policy, such use of welfare economics can
best be shown by an example.
In a recent paper, optimum water allocation in social planning is analyzed
by superimposing smoothly convex social satisfaction indifference curves on a
single production possibilities curve equidistant from the point of origin.-
Anyone with a little knowledge of high school geometry has no difficulty in
locating accurately a point of maximum social satisfaction in water allocation
between two uses.
^ ~~ -, - - - - - - - - - -
l/For the last (but probably not final) word in this clarification, see:
Samuelson, Paul A., "Social Indifference Curves," Quarterly Journal of Economics,
vol. LXX, February, 1956, pp. 1-22. This article cites the significant previous
He Earl O. and John F. Tiamons "Ec Frework for P a
y Zone LPrw Center 0ce .ed
iversi 1Of Iow."ii c7-" s cn=Icul U"e, I_ e College, Ames,
Iowa&, M Zc'h' :, i 2, .p Processed,
__________________________________________--------- -- --- n ______
No information is given as to how the social satisfaction indifference
curves can be determined theoretically and computed in actuality. A higher
indifference curve does not become "Pareto-better" by word magic. The sugges-
tion that such curves could be used for water allocation in legislating and
planning must be regarded as not warranted by the state of welfare economics.
For some time, cost and revenue indifference systems have been in use to
analyze decision making in firm economics.2 But applying a single equidistant
possibilities curve to policy decisions raises questions no less serious than
the use of social satisfaction indifference curves.
The ancestry of a single production possibilities curve, the apparent sim-
plicity of which has made it rather popular recently, can be traced to two
basic assumptions of programing.W These assumptions are, first, existence of
limitational factors--especially capital--and, second, independence of decisions
regarding the intensity of each process and decisions regarding the combination
of processes. No information is given as to what limitational factors are
assumed. From the standpoint of policy, capital and other factors are limita-
tional only under narrowly defined short-run static assumptions. Under such
assumptions, water allocation through legislation, court decisions, and adminis-
trative regulation has little meaning.
/ Ciriacy-Wantrup, S. V., "Eeonomics of Joint Costs in Agriculture," Journal
of Farm Economics, vol. XXIII, no. 4, December, 1941.
/ The reader with no technical training in economics may want to consult an
easily understandable explanation of the assumptions and techniques of programing,
for example: Soles, James N., "Linear Programing and Farm Management Analysis,"
Journal of Farm Economics, vol. XXXVII, no. 1, February, 1955.
Under long-run static and under dynamic assumptions, policy decisions
regarding water allocation are not independent of policy decisions regarding
water development (Section 4). A single production possibilities curve "assumes
away" the essential problem of the meaning and the determination of marginal
costs of water development. Water allocation policy deals with a whole system
of cost indifference curves. Which one is relevant can be ascertained only
after comparing them with a system of "revenue" indifference curves. The
optimum point of water allocation then becomes a curve of "optimum direction"--
using the terminology of vector analysis.- This curve is not monotonic, and
under assumptions approaching reality, it is a space curve.
The equidistant feature of the single possibilities curve implies that
changes in the water use vector have no influence upon production possibilities.
By this implication, one "assumes away" another essential problem of water
allocation. This problem is created by great differences in water quality
requirements between water uses. This problem presents itself in economic terms
through important relations between quality, quantity, and costs. Some of these
relations were indicated above (Section 2).
One must conclude, therefore, that a single equidistant production possibil-
ities curve has no meaning for water allocation policy--whatever its use may be
in firm economics.
The foregoing example is of some interest because it combines the approaches
and techniques of the two most important branches of normative economics--welfare
economics and firm economics--and attempts to derive from such a combination
criteria for public water policy. The questions raised so far with respect to
this attempt are overshadowed in their implications for water policy by another:
Is it conceptually useful to make the maximization principle the basis of economic
criteria for pursuing the public interest?
I/ Ci acy-Wantrup, S. V., "rconoios of Joint Costs in Agriculture," ~. ci.,