After Foundations
Ch. 8 of Foundations was long, and containing much more verbal explanation of the mathematics than was to be found elsewhere in the book, but Samuelson considered it “particularly condensed,” added to the book because of the “rather scandalous state of confusion and ignorance” found in the current literature (Samuelson, 1948a).
This no doubt explains why he took up several opportunities to repeat and elaborate on the ideas he had presented there. One of the first points he made was the need to go beyond the judgments implicit in individualism. In 1949, in a review of a book by the LSE economist Hla Myint, Samuelson wrote, “freed from the obscurities of geometry and Paretian French, the new welfare economics stands revealed as being merely a set of incomplete necessary conditions whose whole raison d’etre disappears if the additional ethical conditions are not adjoined” (Samuelson, 1949, p. 372). This is a very strong statement in that he was claiming that it was essential to make ethical judgments that went beyond those that economists were generally willing to make. It was a deliberate, considered remark. A year earlier he had criticized Melvin Reder for “never going beyond the weakest interpersonal ethical axioms” (those of the new welfare economics) (Samuelson, 1948b, p. 399). A year later he wrote of the new welfare economics that, not only did it not go all the way to solve problems of what policy makers should do, “taken by itself, and without supplementation, it goes virtually none of the way” (Samuelson, 1950a, p. 11). The reason was that it analyzed potential welfare improvements but if these were not translated into actual improvements, they were hardly relevant for policy. Thus Samuelson argued that a competitive equilibrium would be socially optimal only if there was an optimal distribution of income and making judgments about the distribution of income involved in making ethical judgments.A second point, crucial to understanding Samuelson’s writing, is that he considered some ethical judgments to be uncontroversial. In “Evaluation of Real National Income” (1950a) his concern was whether a change in real national income indicated a change in the welfare of a community of possibly heterogeneous individuals.7 There was an extensive literature on this problem but Samuelson believed it was confused: “most people who have seen the recent discussion between Kuznets, Hicks and Little must find their heads swimming, and must be
This was an important problem because of the recent explosion of work on creating national accounts.
in considerable doubt as to what the proper status of this vital matter is.”[80] Samuelson’s main point was that one situation could be said to be unambiguously better than another only if it was better for all possible distributions of income. Hicks and others had considered only the distribution of income prevailing before and after a policy change, but this was insufficient. Samuelson explained this using a “utility possibility frontier” (UPF), a curve showing how the utilities of two parties would change as the distribution of income between them changed. Only if one UPF is completely inside the other (i.e., if the two curves do not cross) is it possible to say that one situation is better than another: “The only consistent and ethics-free definition of an increase in potential real income of a group is that based upon a uniform shift of the utilitypossibility function for the group” (Samuelson, 1950a, pp. 19-20). This claim, resting on the Pareto criterion - that if at least one person is better off and no one is worse off, there has been an increase in social welfare - could be described as “ethics-free” only because it rested on ethical judgments that were uncontroversial. Elsewhere in the paper, he wrote with greater precision, adding qualifications, as when he wrote that the “new welfare economics” (of Kaldor, Hicks and Scitovsky) “attempts to clear the way of all issues that can be disposed of in a non-controversial (relatively) ethical-free fashion” (Samuelson, 1950a, p.
11).A third point concerns the relationship between welfare economics and the theory of the consumer. Since his first papers on the subject, Samuelson had been adamant that “utility” had no welfare implications, for it did no more than describe the choices people made. He repeated this point in “The Problem of Integrability in Utility Theory” (Samuelson, 1950b, p. 375):
A last argument might be built up against non-integrability: if people lack the consistency of behaviour that integrability implies, then that attractive branch of individualistic welfare economics which says people’s tastes should count loses most of its content; hence, we should rule out non-integrability. I am afraid that this is an illegitimate intrusion of wishful thinking by the would-be political philosopher into the facts of life.
The furthest Samuelson would go toward linking consumer theory to welfare economics was to argue that if people did not behave rationally then it weakened the case for individualist welfare economics.
If people do not behave as if it matters to them just what they consume, that is a weakness (but not necessarily a fatal one) for the Pareto-type compensa- tionist new-welfare economics. We must not bias our view of the facts to fit our wishes and prejudices, however pretty their pattern. On the other hand if integrability should turn out to be the best hypothesis to explain the empirical facts of the market-place, this makes a belief in individualistic ethics possible but still not mandatory. (Samuelson, 1950b, p. 375)
In other words, if people do not behave as if they were maximizing an ordinal utility function, there is no justification for an individualistic or welfarist welfare economics. However, this does not turn the problem of establishing a social welfare function into an empirical one because, even if behavior could be shown to be consistent with utility maximization, ethical judgments would still be needed to justify an individualistic or welfarist social welfare function.
Samuelson’s most comprehensive reflection came in a lecture apparently delivered in 1951, in which he reflected on the position taken by Lionel Robbins in The Economic Problem in Peace and War (1950), first delivered as a series of lectures in 1947.
The lecture, titled “Politics, Ethics, and Economics,” was never published, but it was not without influence. Francis Bator (1960, p. 113), a widely read author on welfare economics and a former student of Samuelson’s, acknowledged its importance for his thinking. What most interested Samuelson in these lectures was that consideration of practical problems of planning in war and peace had led Robbins to reflect on the theoretical issues of welfare economics that were involved. Robbins noted that people might not know what was best for them, raising the issue of paternalism. More serious was the problem that there was no clear division between public and private goods, for many purchases affected people other than the purchaser. An important issue in postwar Britain was saving, but the paternalism involved in government control of saving had to be balanced against arguments that a nation’s saving might bear little connection with what consumers wanted. Robbins’s conclusion was that, at least in peacetime, “it is the dispute about ends which matters most” (Robbins, 1950, p. 28).Samuelson explained that, in scrutinizing Robbins’s views, he was investigating his own self-doubts: he was re-examining the Chicago individualist faith in which he had been brought up. His starting point was a non-welfarist judgment: “[L]ike Robbins, I have an ethical preference for the use, wherever possible and efficient, of a decentralized price-enterprise system” (Samuelson, 1951, p. 1). However, although he shared Robbins’s dislike of paternalism, there were complications. Suppose that an individualist regime required “tolerance and indifference,” and people could be vaccinated against intolerance? And what about the person choosing suicide even though, if he remained alive, he would soon change his mind? The answers were not clear: “[T]he differences between childhood, maturity, and senility - between sanity, insanity, and neuroticism - between care and carelessness are all of degree.
In a thousand ways in everyday life we impose upon others and we impose upon ourselves temporary and lasting trusteeships that limit freedom” (Samuelson, 1951, pp. 6-7). Perhaps the reason why people did not go further in limiting their freedom to make mistakes was not that this was undesirable, but that the means of doing so efficiently were not available. The provision of public goods raised questions of ethical inter-personal comparisons, just as did the provision of private consumption goods to different individuals. Further problems arose from much consumption being “conventional rather than necessitous in any physiological sense.” Samuelson asked whether it was possible for even “the most refined theoretician... to identify a consistent and invariant indifference-curve field for even the most rational individual?” (Samuelson, 1951, p. 8).This was where the politics of his title entered, for he could address some of these issues only through bringing in “non-economic political ethical axioms” (Samuelson, 1951, p. 8). For example, one might say “We don't really care a bit what things we happen to consume, but we will fight for the death for the right to choose them freely on competitive markets,” a position he attributed to Stigler. To take such a stance was to see the pricing mechanism as an end, not as a means. Samuelson distinguished between the “economic well-being” of individuals (resulting from consuming goods that they want to consume) and “any subtractions, additions, or modifications to this purely economic well-being that stems from the political and economic mechanisms by which their actual consumption is brought about and determined” (Samuelson, 1951, p. 10). He claimed that the first could be summarized by “an Economic Welfare Function a la Bergson,” and he coined the phrase “Political Welfare Function” to describe the second. However, having written as though economic and political welfare could be added together, he immediately qualified this, denying that they were independent of each other:
Because of the interactions between these two aspects of welfare and for other reasons too, it would not be proper simply to add the two together in order to arrive at the larger total of welfare.
Rather it is a case where changes in either aspect of life - economic or political - will cause (ordinal) changes in welfare most broadly conceived. (Samuelson, 1951, pp. 10-11)The political problem on which he then focused was that of an ethically acceptable distribution of income. The complication this raised was that changes in distribution (political welfare) would have implications for efficiency and hence economic welfare, meaning that the two needed to be considered together.
Samuelson then took up Robbins’s discussion of the optimal level of saving. He criticized both Robbins and the Cambridge mathematician Frank Ramsey (1928) who had analyzed the problem of a planner maximizing the discounted sum of utility over an infinite time horizon. Ramsey had used some interesting methods but his theory was completely unacceptable because it involved individuals evaluating their own consumption (their own utility) in relation to that of future generations. This raises a political-ethical problem:
Just as it is a political-ethical problem to decide how much today’s Peters share with today’s Pauls, so it is a political-ethical problem to decide how much we present- day Peters must share with the Peters that will come after us. A welfare function must have been defined before normative appraisals are meaningful.... The young Peter has no more right to rob old Peter than to rob his unborn great-great grandchildren. (Samuelson, 1951, p. 18)
8.7