If economics is the study of how to make the best, or optimal, use of limited resources, welfare economics is concerned with the meaning of the term ‘optimal’ and with the formulation of statements that permit us to say that a given policy or event has improved or reduced social welfare.
Optimality is defined in terms of maximizing social welfare, so that the focus of concern is on what comprises the latter concept. Typically, it is taken to be the sum of the welfares of all members of a defined society. By adopting the value judgement that it is individuals’ own judgements of their welfare that is to count in the formulation of a measure of social welfare, we have the basis for Paretian welfare economics (after Vilfredo Pareto). In this case, to say that individual A’s welfare has improved is to say no more than A prefers one situation to another. To say that social welfare has improved requires a further definitional statement, namely, that the improvement in A’s welfare has occurred without any other individual being worse off. Thus social welfare has improved if, and only if, at least one individual’s welfare has improved and no one’s has decreased. It may be noted that while the first requirement is a value judgement, the second is a matter of definition. It is not an additional value judgement.
Paretian welfare economics is almost self-evidently sterile, since we can envision few situations in which no one is harmed by a policy. Some individuals gain and some lose. The sterility of the pure Paretian principle arises because of the alleged difficulty of comparing one person’s gain in welfare and another’s loss: the so-called fallacy of interpersonal comparisons of utility. If this is accepted, there are obvious difficulties for the formulation of criteria for a gain in social welfare. The principle emerging from the work of Kaldor (1939) and Hicks (1939) declares that there is a net gain in social welfare if those who gain can use part of their gains to compensate the losers and still have something left over. In other words, if compensation occurred, those who stand to lose would be fully compensated and their welfare would accordingly be the same before and after the policy in question. Gainers would still be better off provided the required compensation is less than their gross gains. This is the Kaldor-Hicks compensation principle.
Scitovsky (1941) pointed out that a further condition is required, since a policy may alter the distribution of income in such a way that those who lose may be able to pay those who gain sufficient to induce them back to the initial situation. The requirement that this should not be the case defines the Scitovsky reversal test for a state of affairs to be defined as a (modified) Pareto-improvement. Since actual compensation mechanisms are complex, all compensation criteria are typically formulated in terms of the potential for compensation. There is no requirement for the compensation to occur. This provides the complete separation from the Pareto principle: the compensation principle may sanction a policy that leads to a (strict) Pareto deterioration in social welfare. Scitovsky’s work opened the way for an explicit treatment of the distribution of income. Little (1957) defined various alternatives whereby social welfare can be said to increase according to the fulfilment of the compensation criterion (the efficiency test) and an improvement in the distribution of income (an equity test). The seminal work of Rawls (1971), however, best defines the turning-point in welfare economics, whereby there is explicit and simultaneous attention paid to both efficiency and equity through the adoption of Rawls’s ‘maximin’ principle of benefiting the least well off in society.
The historical oddity of welfare economics remains that it has survived as an elaborate framework in itself, and as the foundation of practical techniques such as cost-benefit analysis, despite severe and arguably fatal criticism in the 1950s—notably in the work of de Graaf (1957). Arrow’s (1963) famous ‘impossibility theorem’ also indicates the problems of defining any social welfare function based on the fundamental Paretian value judgement about consumer sovereignty.
The basic analysis of welfare economics has remained largely unchanged since the mid-1970s. It has formed the foundation for environmental economics as well as cost-benefit analysis. Recent work has integrated psychological concepts, notably in work on the divergence found in practice between the Hicksian concepts of compensating and equivalent variations (CV and EV), two measures of consumer surplus. Theoretically, the two measures should be nearly the same. In practice, EV appears to be substantially above CV, or, in plain language, ‘willingness to accept compensation’ to tolerate a welfare loss greatly exceeds ‘willingness to pay’ for an equivalent environmental improvement. Some work speculates that these differences amount to non-reversible indifference curves, that is, consumers’ valuations are heavily dependent on where they are to begin with. If this is correct, not just welfare economics, but demand theory generally, rests on a fallacy.
David W Pearce
University of London
References
Arrow, K. (1963) Social Choice and Individual Values, 2nd edn, New York.
Hicks, J. (1939) ‘Foundations of welfare economies’, Economic Journal 49.
Kaldor, N. (1939) ‘Welfare propositions of economics and interpersonal comparisons of utility’, Economic Journal 49.
Little, I.M.D. (1957) A Critique of Welfare Economics, 2nd edn, Oxford.
Rawls, J. (1971) A Theory of Justice, Oxford.
Scitovsky, T. (1941) ‘A note on welfare propositions in economies’, Review of Economic Studies.
Further reading
Just, R.E., Hueth, D.H. and Schmitz, A. (1982) Applied Welfare Economics and Public Policy, Englewood Cliffs, NJ.
Sudgen, R. (1981) The Political Economy of Public Choice: An Introduction to Welfare Economics, Oxford.
