Welfare economics is that branch of economics which concerns itself with the principles by which alternative economic arrangements may be ranked in terms of social welfare. Although commonly regarded as a normative study, preliminary propositions of welfare economics that have reference to the welfare of the individual only need not be normative. If, for instance, welfare is used as a synonym for happiness, the statement that a person’s welfare increases with an increase in his range of choice becomes a judgment of fact rather than of value. However, unless the welfare of each person in a community were increased, a value judgment about the distribution of welfare would be necessary to any conclusion that the welfare of the community as a whole, or social welfare, had increased. One might try to avoid all value judgments by defining some measurable magnitude as an index of social welfare. It might then seem possible to regard welfare economics as a positive study concerned with testing hypotheses about the factors that influence this chosen index. But unless the criteria of social welfare implied by the definition, particularly the distributional aspect, command broad assent, the results of this arbitrary procedure will have limited interest.
History. Until the turn of the present century it was not customary to draw a sharp distinction between positive or descriptive economics, on the one hand, and welfare or prescriptive economics, on the other. Indeed, economic doctrine was such that statements of economic policy were hard to separate from propositions of pure theory, policy often seeming to follow ineluctably from theory. With the publication of Pigou’s Wealth and Welfare in 1912 (revised as The Economics of Welfare in 1920), it became generally realized that welfare economics could be usefully developed as a separate study and regarded as a body of principles to which policy decisions might usefully be referred. However, one thing remained to be made explicit: no policy propositions whatsoever could be deduced from the axioms of economic theory. This task was admirably discharged in Lionel Robbins’ classic Nature and Significance of Economic Science (1932). Although the distinction between positive and normative economics is now commonplace among the academic fraternity, and the text for a sermon to all first-year students, it has to be admitted that much of the writing in popular financial journals and in the reports of government and international bodies continues to convey the impression that policy proposals flow inevitably from well-worn economic principles.
Several of the key concepts in welfare economics were elucidated in the nineteenth century. The concept of consumer’s surplus, for instance, was used by Dupuit (1844), a French engineer, to justify the building of bridges in those cases where even the maximum revenue derivable from the sale of their services fell short of their current costs. In the first edition of his Principles of Economics (1890), Alfred Marshall, father of the Cambridge school of economists, defined consumer’s surplus as “the excess of the price which he would be willing to pay rather than go without the thing, over that which he actually does pay.” Assuming the marginal utility of money to be unchanged for small price variations, he attempted to justify what was in effect Dupuit’s method of measuring this surplus as the area under the demand curve and above the price line. This concept and the analogous concept of economic rent are useful tools in partial equilibrium analysis, in which prices in all markets, other than those immediately under survey, are taken as remaining constant.[See CONSUMER’S SURPLUS;RENT.]
The progenitor of general equilibrium analysis, Léon Walras, introduced the idea of a position of maximum welfare for society, a position he identified with the market solution of a purely competitive economy (1874-1877). However, the concept was more successfully established before the turn of the century by Vilfredo Pareto (1896-1897). Having discarded marginal utility in favor of the notion of ordered preference fields for individuals, Pareto was able to define a social optimum as a position from which no change could be made that would make everybody better off. A few years later Enrico Barone (1908), in an article entitled “The Ministry of Production in the Collectivist State,” used some fairly simple mathematics in exploiting the implications of such an optimal position. However, the neatest demonstration of the implications of an optimal welfare position for society remained to be given by Abram Bergson (1938). By maximizing a welfare function for society containing all the relevant economic variables subject to the constraints of techniques and resources, most of the first-order conditions for a maximum, familiar to economists as the marginal equalities, were seen to be common to various schools of thought which differed mainly on the question of an ideal distribution of welfare.
The “New Welfare Economics” and the old . The so-called New Welfare Economics is dated, somewhat arbitrarily, from the appearance of a short note by Nicholas Kaldor (1939) in response to a paper by Roy Harrod (1938). Using as an illustration the repeal of the Corn Laws in Britain in 1846, Harrod argued that the gain to the community as a whole might be said to exceed the losses to the landlords only if all the people affected were treated as equal in some sense, a view held to be unwarranted by Robbins. Kaldor, however, denied the relevance of interpersonal comparisons of utility to the problem by attributing to the classical economists a more “objective” test of economic efficiency: a new economic arrangement is an improvement if the losers thereby could be more than compensated by the gainers. Whether compensation should be paid in any instance was a political question on which the economist had no special authority to pronounce.
This test, or principle of hypothetical compensation (sometimes abbreviated to principle, or test, of compensation), was hailed by John Hicks (1939a) as a more suitable foundation for welfare economics than was the utility foundation provided by Marshall and at the time associated with the Cambridge school—in particular with the welfare economics of Pigou. Allocative problems had frequently been expressed in value terms: the attainment of an “ideal” output—an optimum position, in effect, requiring that for any class of factor the value of marginal products be the same in all lines of employment. Yet it was generally believed that one could not proceed to improve allocation with a clear conscience unless the marginal utility of money was in fact the same for everyone concerned. For so long as differences in the marginal utility of money existed between people, the redistribution of income associated with any reallocation might well result, on balance, in a reduction of total utility. With the Kaldor-Hicks test, however, this precondition was regarded as superfluous: one could proceed without inhibition to recommend as allocative improvements all changes which met the compensation test, leaving distribution as a separate consideration.
Although 1939 is a useful watershed in welfare economics, the perspective allowed by the passage of time since then reveals the New Welfare Economics to be less of a novelty and more of an adaptation of the existing Pareto approach as developed by Barone and Bergson, among others. On the one hand, it obviously involved a shift from cardinal utility to ordinal utility, or to preference fields in general—although the habit of expressing welfare propositions in utility terms has lingered into the present. On the other hand, the New Welfare Economics appears as a straightforward extension— to some extent anticipated by Pigou (1920) and Hotelling (1938)—of Pareto’s definition of an optimum position (1896-1897) to nonoptimal positions generally. If an optimum was defined as a position from which no movement could make everybody better off, nonoptimal positions could be ranked on a similar principle: thus a movement from one nonoptimal position, I , to a better, or “more efficient,” nonoptimal position, II , could be defined as a movement which could make everyone better off than he was in the I position—in effect, a statement that gainers could overcompensate losers in the movement to II. And one should be reminded in passing that a requirement that everyone in fact be made better off in the change from I to II —which is sometimes referred to as “the principle of compensated adjustment” and which, although cautious, might appear an inoffensive rule of procedure— was explicitly repudiated by Kaldor, for whom this sort of hypothetical compensation was in itself a test of economic efficiency divorced from distributional considerations.
Several of the criticisms of the Kaldor-Hicks test serve to highlight some of its logical and ethical implications. Tibor Scitovsky (1941) has pointed out that a new batch of goods, II , might by this test be shown to be more efficient than an existing batch, I , notwithstanding that if the n batch were adopted, application of the same test would sanction a return to I . The key to this paradox is to be found in the interconnection between any distribution of a given batch of goods and the corresponding common set of relative prices. The Kaldor-Hicks test is rather like comparing the aggregate values of the two batches of goods, using as relative prices those generated by the existing distribution of the I batch. Scitovsky’s proposed reversal test, on the other hand, is akin to comparing their aggregate values with the prices resulting from the II distribution. And it is a well-known index-number phenomenon that certain changes in the weights—prices, in this instance—attributed to the magnitudes in alternative situations might alter the ranking of the two indices. In any event, the Scitovsky criterion for an economic improvement required, as one should expect, not only that in a movement to II gainers be able to overcompensate losers but that, in addition, gainers in a return to I should not be able to compensate losers.
It might be thought that the compensation principle is such as to sanction blackmail, since anyone contemplating socially damaging behavior ought, on this principle, to be dissuaded by adequate compensation. In fact, since hypothetical compensation is the criterion, a movement to the new position of no social damage is countenanced only because the losers (those who agree to refrain from mischief) may be more than compensated by the gainers (the rest of society). It is not required, however, as Kaldor emphasized, that the losers actually be compensated. Indeed, this objection, voiced by George Stigler (1943), might more legitimately be urged against the principle of compensated adjustment, which does require of the efficiency test that everyone actually be made better off. According to this principle, all existing, or even potential, criminals ought to be pensioned off by society. Nonetheless, as understood today, the compensation principle, actual or hypothetical, is not a universal principle but one applicable in cases in which the law, as the expression of public opinion, is neutral as between contending interests. Obviously this is not always so. If, for example, it were universally held that all men had a right to peace and quiet, a small airline company would be legally prevented from disturbing the quiet of a wealthy residential area even though, in the absence of legal restraint, the rich could have improved their situation by compensating the airline for the losses it would sustain by abandoning the route in question.
A more telling objection to the compensation tests as criteria of economic efficiency was voiced by Radomysler (1946): that their constant application might conceivably act to reduce the purchasing power of the poorer sections of the community. This consideration was uppermost in Little’s mind when he argued (1950) against the adoption of tests of hypothetical compensation as definitions of an increase in real income, in welfare, or in economic efficiency. He proposed to use such tests only as one part of a dual criterion which would require that an approved economic change also realize a satisfactory distribution of welfare.
It is less accurate to speak of different schools of welfare economics than of different approaches which, were they all realizable in practice, would tend to the same solution. We can divide the present theoretical literature into three broad streams.
Optimality of resource allocation . The traditional interest in rules of resource allocation, itself closely associated with the development of the theory of value since 1870, is still foremost in the theoretical literature. It has resulted in continued emphasis on those rules regarded as necessary conditions for an optimal position, the simplest rule being in fact the most comprehensive in requiring that resources be so distributed that—granted complete divisibility—the marginal products of each factor class are the same in alternative uses. (If this condition is not met, there must be room for improvement: If, for example, the marginal value of labor’s product in good X is $2 but is $3 in good Y, then, provided always that the marginal product of labor was diminishing in X and Y, continued increments in total value are created by switching labor from the production of X to that of Y until its marginal value in each is the same.) We can usefully break this over-all condition down into those preliminary conditions most frequently used in examining popular welfare propositions by expressing the value of the marginal product for factor A in all lines of product as
Similarly, for factor B,
(where X and Y are goods, px and pv their respective prices, and A and B factor units; px(dX/dA) is, therefore, the price of the good X times the marginal physical product of factor A in terms of X; that is, the value of A’s marginal product).
From equations (1) and (2) three sets of optimal conditions are derived without difficulty.
Exchange optimum requires that the product rate of substitution be the same for each person. This condition is met by having a common set of goods prices, px, pv, …, to which each person adjusts by setting equal to it the ratio of his marginal utilities of X, Y, …, or, what comes to the same thing, his goods rate of substitution. Thus, for each person:
All conceivable distributions of a given batch of goods meeting this exchange optimum condition trace out a locus known as the contract curve.
Production optimum, which requires that the ratio of marginal physical products of pairs of factors be the same for all goods or, put another way, that the factor rate of substitution be the same for all goods, is derived simply by dividing equation (1) by (2). Thus,
and so on. This condition is met along a boundary containing all combinations of goods producible with a given factor endowment.
Top level optimum builds on the two preceding optimal conditions and requires that the product rate of substitution in consumption (the same for each person) be equal to the product rate of substitution in production. It is obtained by dividing (1) by pv(∂X/∂A) and (2) by pv(∂X/∂B), which gives
The price ratio on the left-hand side faces each individual and results in a common rate of substitution ∂Y/∂X, this being equal to the technical rate of transformation (or product rate of substitution) ∂Y/∂X on the right-hand side. This condition is fulfilled by any tangency between the production possibility boundary and a community indifference surface; in the construction of the latter the exchange optimum condition is always met.
The allocative rule, or optimum conditions, has frequently been employed in the attempt to establish (among other things) general propositions concerning the welfare superiority of perfect competition or “equiproportional” imperfect competition (in which the ratio of price to marginal cost is the same in the production of all goods and services) over other forms of market organization, the welfare superiority of poll taxes over other forms of taxation, of free trade over no foreign trade. It is also used for deriving the conditions necessary to the achievement of optimal tariffs by a country which is immune from tariff retaliation but can sell more abroad only by lowering its price and can buy more only by paying a higher foreign price. The conditions can easily be reached by a little algebraic manipulation guided by the idea of maximizing a country’s net benefit from foreign trade, although they are illustrated most simply in the two-goods, two-country model represented in Figure 1. The curve OB is country B’s offer curve and traces the amounts of Y it is willing to export in exchange for X imports. Similarly, OG is country G’s offer curve and indicates the amounts of X it is willing to supply against its imports of Y. The free trade equilibrium is at D, at which terms of trade, OD, country G’s imports of Y, or OY, is equal to B’s
exports of Y, and G’s exports of X, OX, is equal to B’s imports of X. Under free trade, country B enjoys a level of welfare indicated by the social indifference curve I3 passing through D. Any tariff on the imports of X reduces the amount purchased by citizens of B at the existing terms of trade and, indeed, at any conceivable terms of trade. Any given tariff, in effect, contracts B’s offer curve. The optimal tariff is chosen to generate an offer curve OB’ that will intersect OG at D’, this being the point at which B’s social indifference curve I’s, is tangent to OG. The conditions are now readily stated: the rate of transformation of X for Y through international trade (the slope of G’s offer curve at D’) equals B’s subjective rate of substitution (the slope of B’s social indifference curve, I’s, at D’), which, given perfect competition in B, is also equal to B’s domestic rate of transformation (since everywhere in B prices are equal to domestic marginal costs).
Finally, it should be noticed that an optimal position requires the simultaneous fulfillment of the optimal conditions, which are deemed realized under perfect competition insofar as price is equal to marginal cost in every activity. If, owing to some constraint in the economy, this condition is not met in some sectors, the best that can be done in the circumstances may not be to have price equal to marginal cost in all the unconstrained sectors. In very simple cases there may be simple adaptations of the rules. If there were only one constrained sector, which sold at a price 30 per cent, say, above marginal cost, the best thing to do would be to adopt a 30 per cent rule in the rest of the economy. In other cases there may not even be a best solution, or if there is one in principle, collecting the information required to determine it would be too vast an enterprise to be seriously considered. If the number of constrained sectors were few and their aggregate product small in relation to that of the total economy, much the best thing would be to ignore these few and impose a marginal cost pricing rule on the others, with the assurance that, though obviously not meeting the conditions for an optimum, the economy as a whole would not be far from an optimal position.
Consumer’s surplus. The notion of economic rent as the surplus accruing to a resource owner, dating from Ricardo, and the parallel notion of consumer’s surplus enjoyed by a purchaser of finished goods are common in rough computational estimates of net benefits in specific projects, although they are less popular in the investigation of general welfare propositions. The consumer’s surplus concept was easily purged of its utility content by Hicks (1939b) and further elaborated in a series of papers in the next few years. The crucial income effect discovered by Slutsky (1915), and independently by Hicks and Allen (1934), was central in drawing a distinction between the compensating variation and the equivalent variation of a price change. The compensating variation may be defined as the compensation which would make an individual as well off as he was before the change if he is constrained to accept the change. If, for example, the price of oranges fell from six cents to four cents each, the compensating variation would be measured by the maximum sum per period the individual would be prepared to pay for a license enabling him to buy all the oranges he wished at the new price of four cents rather than be constrained to pay six cents. If, on the other hand, the price of oranges rose from six cents to nine cents, the compensating variation could be measured by the minimum sum per period he would accept in exchange for an agreement from him to buy at nine cents instead of six cents.
The equivalent variation of a price change is definable as the compensation which makes a person as well off as he would be after the change if he were constrained to forgo the change. Thus, if oranges fall from six cents to four cents the equivalent variation would be measured by the minimum sum the individual would accept in order to forgo buying at the new price. If oranges rose from six cents to nine cents, the equivalent variation would be measured by the maximum sum he would be willing to pay in order to have the privilege of buying at the old price of six cents. It follows from these definitions, and can be made apparent on an indifference diagram, that the compensating variation for a fall in price from, say, six cents to four cents is identical with the equivalent variation of a rise in price from four cents to six cents, the symmetry holding for the other possibility.
Market data. In this broad stream of development we can conveniently place the investigations into the appropriateness of market data in applying the optimal rules: how far does marginal cost in industry reflect the “true” cost to society, or the market price of a good its true marginal valuation to society? The pioneering work in directing attention to this enormously important problem was Pigou’s Economics of Welfare (1920). Even if the market were perfectly competitive, with outputs such that price everywhere equaled marginal cost, the resulting allocation might be far from optimal as a result of divergencies between what Pigou called marginal social net benefits of any factor class in different occupations. Today it is more common to talk of external economies and diseconomies, or of external or neighborhood effects. [Seeexternal economies and diseconomies.] Broadly speaking, such effects arise whenever a relevant variable in the economy is unpriced or inadequately priced. A familiar example chosen from Pigou’s great work is that of smoke from a factory chimney, which imposes costs on the inhabitants but which, in the absence of effective antismoke legislation, the factory owners do not include in their cost of production. Were they obliged to install antismoke devices, or to compensate for the damage inflicted on the neighboring inhabitants—or, in the event of inhabitants organizing in order to compensate the factory owners for reducing output, to reckon as costs all offers of compensation they will forgo by continuing production—these additional costs of production would generally be expected to curtail output.
The theoretical discussion of the diversity and implications for welfare of these external effects has been growing with the years, along with skepticism about the virtues of the competitive market as an allocating mechanism. Even such relatively unmeasurable effects as the response of a person’s welfare to changes in the welfare of other people or to changes in their patterns of expenditure— the so-called demonstration effect, sometimes facetiously referred to as the Joneses’ effect—has been the source of considerable theoretical speculation, in particular by Baumol (1952), Duesenberry (1949), and Graaff (1957).
The concepts mentioned in this approach to welfare economics are shaped for use in the applied part of the subject known as cost-benefit analysis, which is gradually superseding ordinary commercial and accounting procedures in determining whether or not to undertake large-scale investments with public funds.
Optimality of distribution . In the preceding category we discussed the literature on optimal conditions. We must remind ourselves that there is an indefinite number of such optimal positions open to society, each differing, among other things, in the pattern of distribution. We place in this category the approach of those who seek a conceptual solution to this problem, although one at a high level of abstraction.
A simple representation of what is being sought may be illustrated by a Samuelson “utility possibility curve,” drawn as VV for a two-person community in Figure 2. The level of total utility of individual A is measured, on any scale, along the UAaxis, that of individual B along the UB axis. The VV curve may here represent a boundary of points attainable with a given endowment of resources. Therefore, any point such as p is a combination of Ob utility for B and Oa utility for A. Point p, being on the boundary, is by definition a Pareto optimum, and in this figure it differs, in respect of distribution, from any other optimal position along the VV boundary. In order to determine formally which of these boundary points is best, the optimum optimorum, as it is sometimes called, we impose on the boundary a simple representation of the social welfare function introduced by Bergson (1938). In a manner analogous to the individual’s indifference map, which orders all combinations of goods, the community may be deemed to rank all combinations of A’s and B’s welfares. W1, W2, and W3 indicate successively higher curves of a given social welfare, although along any one of such curves all combinations of the welfares of A and B are of equal social value. Clearly, q on W2 at a point tangent to the VV boundary indicates the position of the highest social value attainable with the given resources. But on what principle can we construct these W curves? It must be admitted that they are very nebulous things, and if we had some notion of the boundary VV, at least for certain ranges, the choice of some point q would perforce be a political decision.
A far more ambitious, and more abstract, approach begins with the notion of each person in the community having his own ideas of a complete welfare ordering of all opportunities or “social states” open to the community, each such social state including some distribution of all goods and services among the members of the community and some assortment of collective goods and other things which use up scarce resources. A liberal society would require that the social welfare function should be derived in some way from these individual constructions, and from them only. What rules will enable us to generate a “satisfactory” welfare function for society from the variety of individual blueprints? In fact, Kenneth Arrow demonstrated (1951) that if certain “reasonable” conditions are to be met—in particular, that the social welfare function should not be imposed either by custom or by dictatorship, that it be positively associated with the orderings of each of the individuals, and that the removal of any possible economic arrangement be not allowed to disturb the order of the remaining possible arrangements in the social welfare function—no such rules can be found. Although this result excited a good deal of controversy at the time, its bearing on the other aspects of welfare economics has not been very marked.
The Little approach. The problems of distribution and allocation, however, have been approached in a far less abstract form and, indeed, have been divorced from the notion of optimality by some further development of the New Welfare Economics. Little and others turned to a more cautious and partial approach, seeking to establish sufficient criteria based on widely accepted value judgments to enable the community to choose between alternative economic arrangements. The criterion proposed in Little’s A Critique of Welfare Economics (1950) regarded a position II as superior to I if (a) a test of hypothetical compensation showed II to be superior whether the test is based on the I distribution or on the II distribution and (b) the II distribution is held to be at least as good as the I distribution.
There has been intense controversy recently in the pages of the Economic Journal (Robertson et al. 1962) on the meaning of the alternatives I and II, about what is held constant under procedures purporting to entail hypothetical compensation, about the meaning of a better distribution, and about the consistency and transitivity of the two parts of the dual criterion. Agreement has not yet been reached, although it has been suggested that Little’s criterion, when amended to remove the possibility of yielding contradictory results, can be reduced to the rather austere proposition that a movement to a new position is justified if the distribution is made better and not everyone is made worse off. However, this is only a conceivable result and comes from considering the worst that might happen. Its application to practical problems might reveal that it was more usual to find a better distribution and most people made better off.
Some weaknesses in welfare economics . Other than welfare criteria and the constructions used in their representation, the theoretical scaffolding used in constructing welfare propositions is much the same as that used in positive economic theory. In consequence, many of the simplifications used in positive theory have been adopted as a matter of course. The movement of goods and factors is usually treated as costless except, of course, for problems in which transport costs are the central issue. Tastes are taken to remain unchanged, and all relevant information is supposed to be available without cost. Again, since almost all welfare theorizing is conducted within a framework of comparative statics, the time taken, and the costs involved, in adjusting from one economic arrangement to another is left out of the formal analysis. Such simplifications as these may be more readily admitted into the axiomatic structure of positive theory, where their appropriateness is, in the last resort, determined by reference to the predictive performance of the relevant hypotheses. In welfare economics, by contrast, the degree of truth or falsehood of these simplifications is directly pertinent in deciding whether or not the welfare criteria in question have been met or whether an optimal position is reached by adopting certain economic measures. For we cannot legitimately talk of a movement to a better, or to a best, position unless all the costs associated with the act of moving have been allowed for. We are therefore justified in looking more closely at these simplifications.
Taste. As for constancy of tastes, which— despite some foredoomed attempts to make provision for possible changes—is the sine qua non of all welfare propositions, there are considerable difficulties of interpretation to contend with. It may be urged that inasmuch as a person’s pattern of choice is influenced by season, climate, age, environment, and other distinguishable circumstances, no formal change in that pattern need be recognized: such influences enter into his preference function in much the same way that prices and income do. An autonomous change of this preference function may, on the other hand, arise in response to some novel experience, including persuasion and new information. If the period of time over which a given economic arrangement continued to exist were long enough, and people had the foresight to make allowance for those changes in their choice patterns which inevitably arise from seasonal fluctuations and the passage of time, welfare economics could afford to ignore the autonomous changes in taste. But if instead the autonomous changes are rapid and conspicuous relative to the time required to move to a better economic position, especially in cases where the tastes that are changing are of immediate relevance to the contemplated alternative arrangements, little benefit can be expected to society from the study of welfare economics.
Information. The increasing variety and complexity of modern goods, both finished and intermediate, is acting to increase the costs of assembling and distributing the information among those who can make use of it. Whether, and how far, to extend the existing information services is obviously an applied problem in welfare economics that in itself calls for a vast amount of statistics in the endeavor to meet the condition that the dissemination of information be expanded to the point at which marginal benefit equals marginal cost.
Moreover, information about the future, whether it be about the weather, patterns of demand, sources of supply, or technological innovation, would be to some extent uncertain, no matter what resources were devoted to forecasting. There has been much informal discussion, by those in broad agreement about the meaning and desirability of a good allocation of resources, concerning the advantages or disadvantages of more centralization of planning decisions in adjusting the economy to continually changing optimal positions. Most of this literature —works of Hayek (1944), Friedman (1962), and Dobb (1924-1954) come to mind—is of a reflective nature, drawing heavily on historical interpretation and inspired by strong political conviction.
Uncertainty. Uncertainty is almost always discussed in connection with entrepreneurs’ decisions to invest in one thing or another, yet this sort of uncertainty, much of which can be reduced by better information and, perhaps, by a reduction in competitive advertising, is of negligible importance for welfare economics, compared with the inability of the consumer to foresee the longer-term consequences of introducing goods which, in the absence of such foresight, would be sanctioned under any welfare criteria. Many years are required for significant and manifest external diseconomies to be experienced by society, and by then vast material interests are entrenched and all the circumstances of daily living are so conditioned as to make drastic change politically difficult. Mechanized transport, for instance, speeds the pace of travel and initially saves time. What could hardly be foreseen is that the greater facility would be more than offset by the rapid increase of daily distances, so that in the second half of the twentieth century people spend a far larger portion of their lives traveling to and from work than at any other period in history, and have come to accept the more blatant external diseconomies—foul air, unabating engine noise, loss of life and limb—as inevitable features of modern life. Indeed, we might conjecture that the more important consequences of the introduction of new goods, chiefly durables, are likely to be the least measurable. The effects of television on emotional well-being—indeed, the effects of increasing mechanization in isolating men from one another—are beginning to be realized, though more in the field of fiction and biography than in sociological studies. The economist who tried to bring them to bear on the welfare calculus would be regarded as eccentric.
For these and other reasons it would be optimistic to expect that continued study in welfare economics will contribute toward making the world a happier place. The way in which we live—the material environment and social institutions, and above all, the relation between man and his fellows upon which happiness ultimately depends—has become largely a by-product of technological innovation wherever it advances. Nevertheless, appreciation of the methods of welfare economics can do much to mitigate some of the more blatant ills of the affluent society by combating conservative presumption in favor of commercial criteria and by revealing manifest injustices in any price system that has not been corrected to make allowance for visible and widespread external diseconomies.
E. J. Mishan
[See alsoConsumer sovereignty; Economic equilibrium; External economies and diseconomies; Game theory; Public expenditures; Utility; and the biographies ofBarone; Dupuit; Marshall; Pareto; Pigou; Walras.]
Arrow, Kenneth J. 1951 Social Choice and Individual Values. New York: Wiley.
Barone, Enrico (1908) 1935 The Ministry of Production in the Collectivist State. Pages 245-290 in Friedrich A. von Hayek (editor), Collectivist Economic Planning: Critical Studies on the Possibilities of Socialism by N. G. Pierson, Ludwig von Mises, Georg Hahn, and Enrico Barone. London: Routledge. → First published as “II ministro della produzione nello stato collettivista.”
Baumol, William J. 1952 Welfare Economics and the Theory of the State. Cambridge, Mass.: Harvard Univ. Press.
Bergson, Abram 1938 A Reformulation of Certain Aspects of Welfare Economics [by Abram Burk]. Quarterly Journal of Economics 52:310-334.
Dobb, Maurice (1924-1954) 1955 On Economic Theory and Socialism: Collected Papers. New York: International Publishers.
Duesenberry, James S. 1949 Income, Saving and the Theory of Consumer Behavior. Harvard Economic Studies, Vol. 87. Cambridge, Mass.: Harvard Univ. Press.
Dupuit, Jules (1844) 1952 On the Measurement of the Utility of Public Works. International Economic Papers 2:83-110. → First published in French. Mainly of historical interest.
Friedman, Milton 1962 Capitalism and Freedom. Univ. of Chicago Press.
Graaff, J. De V. 1957 Theoretical Welfare Economics. Cambridge Univ. Press.
Harrod, R. F. 1938 Scope and Method of Economics. Economic Journal 48:383-412.
Hayek, Friedrich A. Von 1944 The Road to Serfdom. London: Routledge; Univ. of Chicago Press.
Hicks, J. R. 1939a The Foundations of Welfare Economics. Economic Journal 49:696-712.
Hicks, J. R. (1939b) 1946 Value and Capital: An Inquiry Into Some Fundamental Principles of Economic Theory. 2d ed. Oxford: Clarendon.
Hicks, J. R.; and Allen, R. G. D. 1934 A Reconsideration of the Theory of Value. Economica New Series 1:52-76.
Hotelling, Harold 1938 The General Welfare in Relation to Problems of Taxation and of Railway and Utility Rates. Econometrica 6:242-269.
Kaldor, Nicholas 1939 Welfare Propositions of Economics and Inter-personal Comparisons of Utility. Economic Journal 49:549-552.
Little, I. M. D. (1950) 1957 A Critique of Welfare Economics. 2d ed. Oxford: Clarendon.
Marshall, Alfred (1890) 1961 Principles of Economics. 9th ed. 2 vols. New York and London: Macmillan. → A variorum edition. The 8th edition is preferable for normal use.
Mishan, E. J. 1957 A Re-appraisal of the Principles of Resource Allocation. Economica New Series 24:324-342.
Mishan, E. J. 1967 The Costs of Economic Growth. London: Staples.
Pareto, Vilfredo (1896-1897) 1964 Cours d’économie politique. New ed. Oeuvres completes, Vol. 1. Geneva: Droz. → Mainly of historical interest.
Pigou, A. C. (1920) 1960 The Economics of Welfare. 4th ed. London: Macmillan.
Radomysler, A. 1946 Welfare Economics and Economic Policy. Economica New Series 13:190-204.
Reder, Melvin W. 1947 Studies in the Theory of Welfare Economics. New York: Columbia Univ. Press.
Robbins, Lionel (1932) 1937 An Essay on the Nature and Significance of Economic Science. 2d ed., rev. & enl. London: Macmillan.
Robertson, D. H. et al. 1962 Welfare Criteria: An Exchange of Notes. Economic Journal 72:226-244.
Samuelson, Paul A. (1947) 1958 Foundations of Economic Analysis. Harvard Economic Studies, Vol. 80. Cambridge, Mass.: Harvard Univ. Press.
Scitovsky, Tibor 1941 A Note on Welfare Propositions in Economics. Review of Economic Studies 9:77-88.
Slutsky, Eugen 1915 Sulla teoria del bilancio del consumatore. Giornale degli economisti 3d series 51:1-26. → Mainly of historical interest.
Stigler, George J. 1943 Communications: The New Welfare Economics. American Economic Review 33: 355-359.
Walras, Léon (1874-1877) 1954 Elements of Pure Economics: Or, the Theory of Social Wealth. Translated by William JaffÉ. Homewood, 111.: Irwin; London: Allen & Unwin. → First published as Éléments d’économie politique pure. Mainly of historical interest.
"Welfare Economics." International Encyclopedia of the Social Sciences. . Encyclopedia.com. (May 19, 2018). http://www.encyclopedia.com/social-sciences/applied-and-social-sciences-magazines/welfare-economics-0
"Welfare Economics." International Encyclopedia of the Social Sciences. . Retrieved May 19, 2018 from Encyclopedia.com: http://www.encyclopedia.com/social-sciences/applied-and-social-sciences-magazines/welfare-economics-0
Modern Language Association
The Chicago Manual of Style
American Psychological Association
Welfare economics is a normative branch of economic theory that attempts to assess the implications of laws and institutions, including market outcomes, for human well-being. Welfare economics begins with John Stuart Mill’s “canons of taxation,” in which he applies rule-utilitarian ethics to suggest guidelines for taxation that might reduce its bad impacts. Nevertheless, A. C. Pigou’s Economics of Welfare (1920) can be thought of as the founding book of welfare economics. Among the propositions of welfare economics there would be a broad consensus, for example, that (1) in the absence of externalities, competitive equilibria are efficient, and (2) with few exceptions, taxes, monopoly power, and externalities tend to move the economy predictably away from an efficient allocation of resources.
The ideas of Mill and Pigou are utilitarian in the narrow sense that they assume the following:
- Acts, laws, rules, and institutions should be evaluated on the basis of their consequences rather than on some intrinsic rightness or wrongness; that is, nothing is good unless it does somebody some good.
- Good and bad subjective states of mind are the consequences that should be considered in the assessment.
- For a particular individual, the degree to which good states of mind are attained, and bad states avoided, can be expressed by a number (called utility).
- Moreover, this number can be compared between individuals and cumulated over the population; and so this cumulative number can be made the objective of public policy.
This narrow utilitarianism would imply, among other things, that efficiency depends on the distribution of income as well as the allocation of resources. Many economists found this undesirable, on the grounds that one ought to be able to assess the efficiency of resource allocation apart from the distribution of income, as indeed Mill had suggested. This led them to reject the fourth assumption. For this purpose, Vilfredo Pareto had proposed the criterion that bears his name: An allocation of resources is said to be “Pareto optimal” if no one person can be made better off without making another person worse off. Without interpersonal comparability of utility, however, no discrimination can be made among Pareto optimal allocations as to which is better or worse.
Moreover, to some, the third assumption also seemed implausible. In place of numerical or “cardinal” utility, they held that individual decisions and well-being are based on a system of preferences. The “good states of mind” are the ones that the person prefers (which may not correspond to pleasure and pain), and therefore the vectors of consumption goods and services that produce the states of mind can be placed in an order from better to worse, from that individual’s point of view, but the ordering does not correspond to any unique numerical measure. As Paul A. Samuelson (1948) observes, however, the preferences could in principle be reconstructed from (“revealed” by) the observed choices of the individuals. The revision of welfare economics without the last two assumptions is known as the “new welfare economics.”
A preference system can be visualized by a map of “indifference curves.” Beginning from a particular vector of consumer goods, such as “one coffee and two doughnuts,” the indifference curve corresponding to that vector is the boundary between all of the vectors preferred to “one coffee and two doughnuts” and those to which “one coffee and two doughnuts” is preferred. If one assumes that this boundary is well-defined, it forms a curve with the property that, taking any two vectors along the curve, the individual feels no preference for one over the other. Thus, for example, if “one coffee and two doughnuts” and “two coffees and one doughnut” are on the same curve, then the individual can be said to be indifferent between the two, and accordingly the curve is called an indifference curve. The preference map can also be represented by any one of an infinite array of “utility indices,” provided that, comparing two consumption vectors, the higher number is assigned to the one that is preferred. In this case, however, the only valid conclusions of the analysis are those that do not depend on the specific utility index numbers used.
Francis Ysidro Edgeworth addressed the efficiency of allocation of goods in a pure exchange economy using the indifference curve approach (1995). First, assume that the total quantities available of two goods, good x and good y, are X and Y as shown in Figure 1. If the coordinates of a point in the interior of the diagram are x 0, y 0, then those are the quantities of the two goods allocated to individual j, and the quantities allocated to individual k are X – x 0, Y – y 0. Thus, the indifference curves for individual j are oriented to the x, y axis and are shown by curves 1, 2, 3, while the indifference curves for individual k are inverted and are illustrated by curves i, ii, iii. The points of tangency of two indifference curves, shown by the contract curve LM, are all Pareto optimal allocations of the two goods between the two individuals. In general, then, there will be infinitely many Pareto optima. A shift from one Pareto optimum to another trades off one person’s preferences against those of the other person.
To visualize this trade-off, one assigns a numerical index uj to each indifference curve 1, 2, 3 and a numerical index uk to each indifference curve i, ii, iii. One must keep in mind that these arbitrary indices of utility cannot be compared as between the two individuals nor added. Figure 2 shows a diagram with uj on the horizontal axis and uk on the vertical axis. Any combination of utility
indices (indifference curves) on or beneath this curve are attainable when X and Y are produced, while none of the combinations beyond the curve are attainable. This is called the constrained utility possibility frontier corresponding to the production of X and Y.
Suppose one carries out the same exercise for each technically efficient, feasible set of outputs X, Y. One would then trace the outer limit of all of the constrained utility possibility frontiers obtained in this way. This outer limit is the grand utility possibility frontier or simply the utility possibility frontier. With an appropriate change in the definitions of u 1, u 2, … uiii, one may take Figure 2 as illustrative of the utility possibility frontier. Every point on this frontier has the properties that (1) the two goods produced are allocated between the two persons in a Pareto optimal way; and (2) production is technically efficient—that is, no reallocation of resources can increase the production of one good without reducing the production of the other; and (3) therefore, production and allocation is Pareto optimal over all possible allocations of resources and consumer goods. With appropriate mathematical notation, this conception can be extended to very large numbers of distinct goods and services and a very large population of agents.
Two issues remain in this analysis. First, how does one choose the “best” among the points on the frontier? Second, for many practical problems, one must choose between two allocations, at least one of which is not Pareto optimal. How may one do that?
Taking these questions in reverse order, the new welfare economics answers the second question with a cost–benefit analysis based on the “Kaldor-Hicks compensation test,” named after Nicholas Kaldor (1939) and John R. Hicks (1939). One may illustrate this compensation test with an example of property on the banks of a river. John Doe owns land on the banks of Flowing River, while Richard Roe owns downstream property, including both banks. Roe builds a dam, entirely on his own property. The impoundment of Flowing River floods Doe’s property. In this sequence, Roe is the gainer and Doe the loser in clear senses. But can one say that Roe’s benefits exceed Doe’s costs?
To answer this question, one considers whether Roe could allocate some of his benefits to compensate Doe and still be a gainer on net? This is the Kaldor-Hicks test, and if the answer is yes, then building the dam is a potential Pareto improvement. That is, if the dam is built, there is a distribution of income that would leave everyone in society better off than they were without the dam. But another, equally reasonable, compensation test asks: Could Doe compensate Roe for ceasing and desisting from dam building, so that Roe would be better off than he would be if the dam were built? This is the Scitovszky test—named after Tibor de Scitovszky (1941)—and if the answer is yes, the world without a dam is potentially Pareto superior to the world with the dam. The two tests would be equivalent if benefits and costs were independent of the distribution of wealth, but, as Hicks noted, they are not. If the dam is built, and compensation is not paid, then wealth is redistributed and a new situation created. From the new situation, it might be a potential Pareto improvement to return to the original position. In practice, there is likely to be little difference between the two, because the impacts of the shifting income distribution is likely to be small relative to the impact of a project, and the Kaldor-Hicks test is the one usually used.
Is it possible, then, to choose one among the allocations on the utility possibility frontier that is the optimum of optima? For example, some might prefer an allocation with a more equal distribution of income or one in which certain “basic needs” are more thoroughly met, or that satisfy other conditions. Ideally, one might hope to express some consensus of such conditions in a social welfare function that would indicate which allocations are socially preferable to other allocations independently of their technical feasibility. Visualizing this social welfare function as a set of social indifference curves, and superimposing it on the grand utility possibility frontier, results in Figure 3. In this figure, the grand optimum allocation of resources could be identified with point V. Further, following Samuelson (1956), the utility index could be made an index of real income. Remember that any numbers that correspond to indifference curves and assign larger numbers to the more highly preferred curves can be used as utility index numbers. In general, an index of real income will have this property. Therefore, suppose the axes measure
the real incomes, Rj and Rk, of the two individuals. With this interpretation, the curvature of the social indifference curves could express one’s attitudes toward income equality, so that, for example, social indifference curves tightly curved around the 45-degree line would express a strong preference for equal incomes.
This ideal was set back by the Nobel laureate research of Kenneth J. Arrow (1951), in his general (im)possibility theorem. Arrow showed that, if we require some reasonable-seeming conditions on the social welfare function, such as the requirement that it generate transitive preferences, there could be no social welfare function that would satisfy them all. For a system of majority rule, for example, Arrow adapted a paradox attributed to the marquis de Condorcet (1972) to show that (imposing his other conditions) voting outcomes could always be cyclical rather than transitive. But Arrow also extended this to any (nonmajoritarian) social welfare rule, and in the context of his theorem, it is equally impossible to say that a market outcome defines a social welfare optimum.
Arrow’s theorem has spawned a large literature of reconsideration. Sometimes the conditions he proposed for a social welfare function can seem less self-evidently appropriate on further consideration. For example, one of the conditions is “nondictatorship,” mathematically expressed; but in some simple models of majority rule, the median person is a “dictator” in Arrow’s mathematical sense, although not on a conventional understanding of dictatorship. Thus, much of the controversy has centered on the refinement of Arrow’s conditions, and there is a large variety of possibility and impossibility theorems. All the same, Arrow’s discussion raised enough doubts about the concept of a social welfare function that it has been little used since the mid-twentieth century.
The new welfare economics is utilitarian only in the broader sense in that it accepts assumptions (1) and (2) above. But the narrower utilitarian approach is not dead. Game and decision theorists reformulated utility theory so that an objective numerical measure of utility can be based on observations of risk-averse behavior. In social philosophy, John Rawls (1971) renewed the social-contract theory, arguing that public policy should promote the interests of the least-favored individual. One interpretation of this is that it demands interpersonal comparison of utility and supplies a social welfare function based on the minimax principle. Drawing on the reformulation of utility theory and Rawls’s social-contract approach, John C. Harsanyi (1975) argued that social welfare could after all be based on a summation of individual utilities. Like Harsanyi, Amartya Sen (1985) admits interpersonally comparable utilities, at least as a logical possibility, and he has proposed conditions less limiting than Arrow’s that allow the possibility of a consistent majoritarian social welfare function. Sen, however, rejects what he describes as the welfarism of both the old and the new welfare economics, by which he means the supposition that the goodness of a social system depends only on the welfares of individuals in those social systems. In addition, Sen would have data on the capacities and perhaps freedoms of individuals reflected in the normative evaluation of economic society.
For many welfare economists, inequality in the income distribution is something to be avoided, but that cannot be completely avoided because of the sacrifice of production that would result. But inequality is not a simple thing in itself. Anthony Atkinson has made important contributions to the measurement of inequality and incorporation of inequality in discussions of policy issues such as taxation and fiscal policy. In 2001, Atkinson addressed “The Strange Disappearance of Welfare Economics.” Atkinson does not suggest that welfare judgments are disappearing from economics. Quite the contrary. For Atkinson, “welfare economics” comprises systematic and critical thinking in normative economics, and he notes with regret the de-emphasis of welfare economics from graduate programs and from the current research literature. The result, he suggests, is a proliferation of ill-considered value judgments in economic research, including a good deal of prejudice and confusion. Atkinson illustrates this with examples from recent macroeconomics, a field which perhaps has never been rich in careful thinking on value issues. For Atkinson, then, welfare economics has never been more needed, although this field of study receives far less attention than it did two generations ago. On the whole, nevertheless, welfare economics is a highly developed branch of economic theory that supplies tools for applications in cost–benefit analysis, but that also raises unsettled questions for future research.
SEE ALSO Arrow Possibility Theorem; Arrow, Kenneth J.; Paradox of Voting; Pareto, Vilfredo; Pareto Optimum; Samuelson, Paul A.; Social Welfare Functions; Utilitarianism
Arrow, Kenneth J. 1951. Social Choice and Individual Values. New York: Wiley.
Atkinson, Anthony B. 2001. The Strange Disappearance of Welfare Economics. Kyklos 54 (2/3): 193–206.
Condorcet, Marie-Jean-Antoine-Nicolas de Caritat, Marquis de. 1972. Essai sur l’application de l’analyse à la probabilité des décisions rendues à la pluralité des voix [Essay on the application of analysis to the probability of majority decisions]. New York: Chelsea Publishing. (Orig. pub. 1785.)
Edgeworth, Francis Ysidro. 1995. Mathematical Psychics. Mountain Center, CA: James and Gordon. (Orig. pub. 1881.)
Harsanyi, John C. 1975. Can the Maxim in Principle Serve as a Basis for Morality? A Critique of John Rawls’s Theory. American Political Science Review 69 (2): 594–606.
Hicks, John R. 1939. Foundations of Welfare Economics. Economic Journal 49 (196): 696–712.
Kaldor, Nicholas. 1939. Welfare Propositions of Economics and Interpersonal Comparisons of Utility. Economic Journal 49 (195): 549–552.
Mill, John Stuart. 1987. Principles of Political Economy. Ed. William Ashley. Fairfield, NJ: A. M. Kelley. (Orig. pub. 1909.)
Pareto, Vilfredo. 1971. Manual of Political Economy. Trans. Ann S. Schwier. Eds. Ann S. Schwier and Alfred N. Page. New York: A. M. Kelley. (Orig. pub. 1906.)
Pigou, A. C. 1920. The Economics of Welfare. London: Macmillan.
Rawls, John. 1971. A Theory of Justice. Cambridge, MA: Harvard University Press, Belknap Press.
Samuelson, Paul A. 1948. Consumption Theory in Terms of Revealed Preferences. Economica 15 (60): 243–253.
Samuelson, Paul A. 1956. Social Indifference Curves. Quarterly Journal of Economics 70 (1): 1–22.
Scitovszky, Tibor de. 1941. A Note on Welfare Propositions in Economics. Review of Economic Studies 9 (1): 77–88.
Sen, Amartya. 1985. Commodities and Capabilities. Amsterdam: North-Holland.
Roger A. McCain
"Welfare Economics." International Encyclopedia of the Social Sciences. . Encyclopedia.com. (May 19, 2018). http://www.encyclopedia.com/social-sciences/applied-and-social-sciences-magazines/welfare-economics
"Welfare Economics." International Encyclopedia of the Social Sciences. . Retrieved May 19, 2018 from Encyclopedia.com: http://www.encyclopedia.com/social-sciences/applied-and-social-sciences-magazines/welfare-economics
Modern Language Association
The Chicago Manual of Style
American Psychological Association
WELFARE CAPITALISM. Welfare capitalism is a system of private, employer-based social welfare provisions that first gained prominence in the United States from the 1880s through the 1920s. Promoted by business leaders during a period marked by widespread economic insecurity, social reform activism, and labor unrest, it was based on the idea that Americans should look not to the government or to labor unions but to the workplace benefits provided by private-sector employers for protection against the fluctuations of the market economy. Welfare capitalism, according to its proponents, was a new, more enlightened kind of capitalism, based on the ideals of corporate social responsibility and business-labor cooperation rather than unfettered individualism and class conflict. It was also a way to resist government regulation of markets, independent labor union organizing, and the emergence of a welfare state. For all its promise of industrial harmony, welfare capitalism was a way to keep private employers firmly in control of labor relations.
U.S. businesses began to adopt a variety of what were initially known as "welfare work" practices in the 1880s. From the beginning, the benefits employers offered were inconsistent and varied widely from firm to firm. "Welfare work" encompassed minimal benefits such as cafeteria plans and company-sponsored sports teams as well as more extensive plans providing retirement benefits, health care, and employee profit-sharing. By far the most elaborate and ambitious of the early plans were the company towns, such as the one established by and named for railroad car manufacturer George Pullman in 1881, just outside of Chicago, Illinois. In Pullman, as in the company towns established by textile mill owners in the South, workers lived in company-built houses, shopped at company-established stores, played at company-provided recreational facilities, went to company-hired doctors, and were often expected to worship at company-sanctioned churches.
Portraying themselves as benevolent father figures, many employers sought to exert parental authority and control over their workers as well. Thus, workers drawn to car manufacturer Henry Ford's promise of high ($5.00 a day) wages were subject to home inspections and a strict moral code as conditions of employment. Other employers offered cooking, hygiene, and language classes in efforts to regulate and "Americanize" their immigrant workers. Welfare capitalists went to greatest lengths, however, in efforts to quash independent union organizing, strikes, and other expressions of labor collectivism—through a combination of violent suppression, worker sanctions, and, as welfare capitalism became more widespread, benefits in exchange for loyalty.
By the 1910s and 1920s, welfare capitalism had become an organized movement with a diversifying base of business, social, scientific, and political support. It had also become the leading edge of the quest for corporate competitiveness and efficiency: benefit packages, employers reasoned, would attract a higher skilled, more productive, and stable workforce. Even at its height, however, welfare capitalism left the vast majority of workers without adequate social welfare protection and actively discriminated against low-skilled, non-white, and female wage-earners. Since employer benefits remained unregulated, companies could—and did—abandon their obligations during hard times.
The Great Depression of the 1930s brought the inadequacies of welfare capitalism into sharp relief, as New Deal policymakers joined labor leaders and reform activists to establish the basis of the modern U.S. welfare state. Far from retreating, welfare capitalists subsequently adapted to the era of public provision and stronger labor unions. Private employer benefits, subsidized by tax incentives, became an essential supplement to the basic government safety net and a key bargaining chip in negotiations with organized labor. There is considerable cause for concern, then, that recent decades have seen a dramatic decline in the percentage of the U.S. workforce covered by employer-provided health, pension, and other benefits—especially as these declines have been accompanied by significant reductions in the public provisions of the welfare state.
Gordon, Colin. New Deals: Business, Labor, and Politics in America, 1920–1935. Cambridge, U.K.: Cambridge University Press, 1994.
Jacoby, Sanford M. Modern Manors: Welfare Capitalism Since the New Deal. Princeton, N.J.: Princeton University Press, 1997.
Tone, Andrea. The Business of Benevolence: Industrial Paternalism in Progressive America. Ithaca, N.Y.: Cornell University Press, 1997.
"Welfare Capitalism." Dictionary of American History. . Encyclopedia.com. (May 19, 2018). http://www.encyclopedia.com/history/dictionaries-thesauruses-pictures-and-press-releases/welfare-capitalism
"Welfare Capitalism." Dictionary of American History. . Retrieved May 19, 2018 from Encyclopedia.com: http://www.encyclopedia.com/history/dictionaries-thesauruses-pictures-and-press-releases/welfare-capitalism