Theoretical Welfare Economics

A Critique of Welfare Economics was the major contribution from Ian Little in his early career as an economist. It can also take its place beyond doubt as one as the most important publications on economics from the decades of the early post-war years.

To appreciate this work, it must be seen in the context of its time. These were the years of the “New Welfare Economics”. This was founded in the rejection of the “Old Welfare Economics” of Mill, Bentham and Marshall, which depended upon measurable utility. To these writers, it made sense to discuss whether it is a good idea to take £10 from a rich man to give the pro­ceeds to a poor man, even if leakage created by this transfer reduced the poor man's gain to £3. A comparison of the marginal utility of money of the two parties provides a precise numerical answer. This kind of reasoning was a vic­tim of the revolution in philosophical thinking that was logical positivism, and the ideas of the Vienna School. Taken to extremes, as it sometimes was, this new philosophy reached such bizarre conclusions as the refusal by the Oxford philosopher A.J. Ayer to admit to being an atheist on the ground that the proposition “There is no God” is untestable, and hence without meaning.

If arguments are valid only if they discuss exclusively the observable and the measurable, there is no room for cardinal utility.

It is in response to this difficulty that John Hicks and Nicholas Kaldor came up with the concept that came to be called the Kaldor-Hicks criterion (the K-H test). According to this test, a change can be recommended if the gainers are able to compensate (bribe) the losers and still be better off. That looks appealing, but what exactly does it mean? Are we asked to accept that a change that passes the K-H test, plus the required compensation, is to be rec­ommended? That is no more than a particular case of the Pareto test, and is similarly limited in scope. Instead of such a narrow application, the K-H test did not require that the compensation be paid.

Then Tibor Scitovsky showed that the reversal of a K-H improvement could also pass the K-H test. With inefficient states, well inside the produc­tion possibility frontier, there is plenty of surplus to pay compensation, so Scitovsky's finding is not unexpected. It is to this confusing tangle of ideas that Ian brought his sharp and precise intellect. In place of the K-H test, he proposes a two-item checklist for a change to count as a welfare improvement. First, it would produce a not-unfavourable redistribution of income; and sec­ondly, the losers from the change could not bribe the gainers to vote against it. These two tests together define the “Little criterion”. The second test takes care of Scitovsky's point.

The first three chapters of A Critique develop carefully and thoroughly the theory of welfare comparisons based on the choices made by individuals in market situations. It is shown how consistent choices can generate indiffer­ence curves (or behaviour curves) that provide a behavioural definition of “better off” for an individual consumer. The many difficulties that this approach encounters are noted at every step. Ian eventually relies on the pos­sibility that the theory might work better for an average individual than for a particular genuine individual. One of the striking features of A Critique is its focus on the central field of a basic welfare economics. Its author refuses to be diverted towards extensions, such as dynamics, or the cardinal utility mea­sures of von Neumann and Frank Ramsey. He is aware of this material, but chooses not to go down those side roads. As the reader will learn from A Critique, there is plenty to be done with the most elementary welfare econom­ics; and the author does just that.

The balance between rigorous scepticism, and a determination to achieve what can be achieved, is perfectly captured in the short paragraph that closes Chapter III of the volume:

We must not pretend that our analysis is anything but rough and ready. As we have already implied...it is particularly inapplicable in respect of choices between jobs, and different hours and kinds of work. Nevertheless, enough has, I think, been said to show that it would be foolish to dismiss the whole of wel­fare economics solely on the ground that the analysis of “individual” behaviour, on which it rests, is hopelessly at variance with the facts (Little 1950: 50).

Chapters IV and V of A Critique move on from the behaviour of individu­als and the evaluation of individual welfare to the difficult fields of the distri­bution of welfare, interpersonal comparisons, and value judgements. This is economics, yes; but truly it is high-standard philosophy. Central to Ian's case is a head-on assault on the clear fact-value distinction of David Hume and G.E.

Central to his argument is the observation that terms such as “happy” or “better off” do not refer to the entirely subjective and personal, as it might be maintained does “tastes good”. Even this last term cannot be completely sub­jective. A man who says that raw sewage tastes good is not truthful. Also, some terminology that appears to be no more than a value judgement reflects commonly understood criteria for its application. So, while the description “a good man” may be less precise than “a tall man”, it is not available for anyone to use as he likes. To say that a mass murderer is a good man is simply to reveal linguistic incompetence. Now the sentence ‘John would be happier if he gave up drinking' can be considered a positivist statement. One who insists on a rigid fact-value distinction cannot claim that this last sentence does not entail a value-loaded recommendation that John gives up drinking. Clearly, the positivist statement does imply a recommendation in favour of abstinence in John's case. A crucial conclusion of Ian's detailed analysis heads a list at the end of chapter IV: ‘Interpersonal comparisons of satisfaction are empirical judgements about the real world, and are not, in any normal context, value judgements' (ibid.: 68).

Chapter VII is a short chapter devoted to the social welfare function, such as is proposed by Bergson and Samuelson.

Chapters VIII and IX of A Critique examine the optimal conditions of production and exchange: equal marginal rates of substitution for different individuals or producers. Yet the important point delivered by these chapters is that the satisfaction of one of these conditions is not sufficient for an opti­mum, however defined, if other marginal conditions are not satisfied. For example, direct taxation is not necessarily superior to indirect taxation when direct taxation destroys the equality between the rates of transformation and substitution of leisure and goods. This type of argument is now always called the theory of the second best. Little is perhaps its originator, although a few would realise that. As he himself put it: ‘Unfortunately for me, I did not name the theory!' (CaR 1999: 8).

Marginal conditions do not work when there are indivisibilities. A bridge across a river is either built or not built; one cannot have a little less bridge. Ever since Marshall, this problem has been treated by applying the theory of consumer surplus to focus on the difference in total utility that the bridge delivers. This approach was obviously undermined when cardinal utility was abandoned. Hicks applied much energy to rehabilitating the concept without cardinal utility, while Little took a different route, preferring direct ordinal assessments of lumpy changes.

The remaining chapters of A Critique (XI-XIV) examine output and price policy for public enterprises; the valuation of national income; welfare theory and international trade; and welfare theory and politics. Chapter XV con­cludes. There are numerous sharp insights in these discussions, and also some intriguing surprises. Take, as an example of cutting analysis, the question of marginal-cost pricing for public enterprises. It is evident that the theory of the second best will take issue with a simplistic argument in favour of marginal­cost pricing. This is because with average costs far higher than marginal costs, as is typically the case with public enterprises, such as the railways, strict marginal-cost pricing leaves a large revenue gap to be filled. There is no non­distorting way of raising that revenue, so the case in favour of marginal-cost pricing collapses. Little goes further by showing marginal cost to be a slippery concept. In the short run, marginal cost oscillates wildly, as when the mar­ginal cost of a rail journey varies according to how crowded are the carriages. In the extremely long run, marginal cost is much the same as average cost.

Given his espousal of the second best, one might expect Ian to reject the case for free trade. His actual position is more subtle and interesting. In the preface to the 2002 edition of A Critique he writes:

The basic fallacy is that the free trade dogma neglects the distribution of income. Fifty years later I can find no fault with this. However, I fear that the cursory reader might think that I believe that free trade generally worsens the distribu­tion of wealth both between and also within countries. On the contrary, I believe that for most developing countries, especially the poorest, trade benefits the poor: this is because exports are relatively labour intensive, and raising the demand for labour reduces poverty (Little 2002a: xii).

A good way of assessing the weight of the contribution that is provided in A Critique is to ask what a contemporary undergraduate studying welfare economics would lose if told to read nothing but that one volume. The answer must be that this imagined student would not be badly disadvantaged. Of course, there are numerous other references that would benefit that individ­ual. Ideally, he or she should certainly study some social choice theory, which does have relevance for classical welfare economics. Also, the welfare econom­ics of risk and uncertainty, and intergenerational welfare, should not be neglected. Analysis using welfare weights, rejected by Hicks and only adopted later by Little, is moreover hugely valuable. Yet a must-have tool kit of welfare economics, with the correct emphasis on the distribution of welfare, is all to be found in the pages of A Critique.

4