This paper traces the evolution experimented by the concept of ’’the rational consumer’’ in the published writings of J.R.Hicks on demand theory
His closely related work on welfare economics will...
. These extend, from the 1934 Economica paper co-written with R.G.D.Allen (Hicks and Allen 1934), where they established today’s standard interpretation of consumer rationality as behaviour consistent with a given ordering of alternatives, to the 1986 short paper in a volume honouring the neo-Austrian economist Ludwig Lachmann (Hicks 1986), where he made sceptical comments on the applicability of the assumption of rational consumer behaviour. This change of position seems coherent with Hicks’s general trajectory. It is a common appreciation that, following his retirement in 1965, Hicks walked on a divergent path from the mainstream of economic theory, notwithstanding the direction followed by the mainstream derived in a good part from his earlier works. In fact, the manifested dissatisfaction with some of his own achievements, like the IS-LM model or the ’’dynamic’’ part of Value and Capital (Hicks 1939), was so sharp that many commentators have recurred to the rhetorical device of attributing to ’’different’’ fictional authors the writings included in different stages of the Hicksian work
See among others, (Leijonhufvud 1979), (Solow 1984)...
In this paper, a different interpretation is put forward, stressing the elements of continuity in the development of Hicks’s thoughts in order to provide a picture of gradual changes that can be related to the general evolution of mainstream economic theory. I consider this emphasis is particularly justified to deal with the contributions of Hicks to demand theory, a field in which he seemed distinctly more satisfied with his early work
For example, note that A Revision of Demand Theory...
Two fundamental features of the Hicksian approach to rationality in consumption decisions will be highlighted. The first is Hicks’s insistence in asserting that the proper concern of demand theory (of economic theory, indeed) is the explanation of the behaviour of groups (’’classes’’) of people, and not the behaviour of single individuals. This led him to consistently reject the need of a psychological basis for assuming rational behaviour, and therefore to criticize both utilitarian and behaviourist interpretations of it.
The second feature is the justification of the assumption of rationality in consumption based on two arguments, one of obviousness and another of convenience. In most of his works Hicks tended to identify rational behaviour as ’’economic’’ behaviour, the subject of economic analysis. He considered the procedure of assuming rationality as characteristic of the economic discipline, and provided several examples of its use in the history of economic analysis. Regarding convenience, he defended the rationality assumption as the simplest hypothesis consistent with the requirement to arrive at empirically verifiable implications.
For expository purposes, in this paper I divide Hicks’s contributions to demand theory on three stages, which are developed in corresponding sections. Section 2 deals with the first stage, which includes (Hicks and Allen 1934) and Part I of (Hicks 1939). These were the works that established the approach to demand theory based on the assumption of rational consumer behaviour according to a given scale of preferences (hereafter ’’the preference approach’’) as a replacement to the traditional Marshallian approach based on the assumption of rational consumer behaviour aimed at maximization of a given function which measures consumer satisfaction or ’’utility’’ (hereafter ’’the cardinal utility approach’’). The superseding of the traditional approach was presented as a generalization of the scope of demand theory, allowing it to account for the implications of changes in income on consumption decisions (’’income effects’’), which in Marshall’s theory were assumed away. The preference approach explicitly rejected utilitarianism, offering as justification for its assumptions their simplicity and concordance with empirical evidence.
The subject of Section 3 is the second and central stage, represented by (Hicks 1956), (Hicks 1958) and Chapters 4 and 5 of (Hicks 1979). In these works, Hicks confronted the development of a behaviourist interpretation of the preference approach in the form of Samuelson’s ’’revealed preference hypothesis’’, which forced him to explain more carefully his own position. He regarded the aim of explaining individual behaviour with a simple hypothesis like revealed preference as too ambitious, and prescribed the explanation of aggregate market behaviour as the proper subject of demand theory. In this respect, rationality in consumption decisions could be accepted as a knowingly unreal assumption, useful in the construction of simple models. This procedure was defended with two lines of reasoning. On one hand, Hicks claimed support from economic tradition, presenting several historical examples of what he considered as similar successful applications of the same principle. On the other hand, he insisted again on the econometric applicability of his approach.
The third and final stage, considered in Section 4, corresponds to the last years in the life of Hicks, and deals exclusively with the contents of (Hicks 1986). Even if this short paper is not comparable, in scope or influence, to the main Hicksian contributions to demand theory, it should not be overlooked that it contains an important qualification of the views Hicks maintained in his whole career.
While sketching a brief history of the use of rationality assumptions in economic analysis, Hicks came to the conclusion that its justification is mainly empirical, based on the observation of actual behaviour by different classes of agents. But he found the empirical evidence about the existence of rational consumers not so strong as in other cases. This finding led him to alert about the limited range of situations in which rational consumer models are applicable.
Although the reasons for this important qualification are nowhere explicit on the paper, there are hints that he was worried about the legitimacy of the extensions of economic analysis to questions traditionally considered as unrelated to the discipline.
Finally, Section 5 ends the paper with a summary of the main issues.
2 - An individual with a given scale of preferences
The first contribution by Hicks to demand theory was a paper co-written with R.G.D. Allen, A Reconsideration of the Theory of Value (Hicks and Allen 1934). It was published in two parts on two consecutive numbers of Economica. The first part, signed by Hicks, was a non-mathematical exposition, while the second part, signed by Allen and subtitled A Mathematical Theory of Individual Demand Functions, contained all the mathematical formulation. Despite this division of tasks, they warned the reader that:
Our co-operation has been so close that (…) this division does not (…) correspond in any way to the actual process by which the theory was constructed. Mathematics and economics went hand in hand; nor would the reader find it easy to identify our respective shares by a consideration of the technique necessary to reach particular points..
(Hicks and Allen 1934, p.55, fn.2)
In fact, as Hicks himself pointed at in several occasions
See, for example, the prefatory note to the reprint...
, the extent of collaboration in the development of that paper went even further, in the favourable environment provided by the London School of Economics (LSE) under the direction of Lionel Robbins. Hicks’s involvement in demand theory was a by-product of his research into a reformulation of general equilibrium theory starting from the works of Walras and Pareto, which he started around 1930 following Robbins’s suggestions. Moreover, the basic results of the paper are derived from an adaptation to indifference curves of the idea that the concept of elasticity of substitution -as defined in (Robinson, 1933)- provided a measure of the curvature of an isoquant, idea advanced by A.P. Lerner, then also at LSE (Lerner 1933). Hicks claimed he had derived the relationship between the price-elasticity of demand for a commodity and the elasticity of substitution between that commodity and one another in a simple two-good case before his collaboration with Allen, another LSE fellow, began. The use of the income-elasticity of demand in that relationship interested Allen at a time in which he and A.L.Bowley (yet another LSE man) were working on family budget data (Allen and Bowley 1935)
The marks of the LSE are also evident in the cosmopolitan...
It seems a reasonable guess that this LSE background was determinant in the initial conceptions that Hicks maintained about the rational consumer. The subject of Hicks and Allen’s paper was the reconstruction of demand theory using a Paretian scale of preferences to define the consumer’s ’’system of wants’’ instead of a Marshallian utility function. This was a task entirely consistent with Robbins’s emphasis in the construction of a Positive Economics which rejected utilitarianism, famously expressed in (Robbins 1932).
Rational behaviour in (Hicks and Allen 1934) was embedded in the concept of individual equilibrium. The main character in the paper was called an ’’individual’’ or ’’consumer’’ without any qualification, but it was assumed that his actions always could be interpreted as equilibrium positions. The meaning of equilibrium was that the individual would take all the advantage opportunities he identified on the basis of his given ’’system of wants’’. For Hicks and Allen’s purposes, the hypothesis of rational consumer behaviour was the obvious assumption to use, so there was no necessity to offer explicit justifications for it. In fact, they were proposing a particular change in the exact meaning of rationality, claiming that the traditional concept, based on cardinal utility, was unsatisfactory.
To begin with, they quoted Pareto’s demonstration that the utility function associated to any scale of preferences is not unique and provides only an index, not a measure of utility. This introduced a degree of arbitrariness in the cardinal approach, and made the theory unnecessarily complicated. The main charge against Marshallian demand theory, however, was its inadequacy as a guide for the econometric estimation of demand functions. In fact, as Hicks would recall later (Hicks 1981), this was the direct motivation behind the paper. They were trying to find an explanation to the empirical results of Henry Schultz (Schultz 1933), who failed to estimate symmetric cross-elasticities for two goods, as Marshallian theory predicted.
In Hicks and Allen’s exposition, the relative prices between any two goods, X and Y, have to reflect ’’the quantity of good Y which would just compensate him for the loss of a marginal unit of X’’ (Hicks and Allen 1934, p.55). Otherwise the consumer could exchange equal values of the goods to achieve a more preferred combination. This ratio, defined as the ’’marginal rate of substitution’’ (MRS) of Y for X, would correspond, in the diagrammatic Paretian representation of the simple two-good case, to the slope of the indifference curve. One additional assumption is needed to warrant ’’stable equilibrium’’ (i.e. rational behaviour): MRS of Y for X has to increase when the consumer substitutes Y for X, or the consumer would be choosing his least-preferred combination instead of the most-preferred one.
This theoretical structure allowed a detailed description of the consumer’s scale of preferences. The rate of increase of MRS can be measured by the Robinson-Lerner elasticity of substitution. In the simple two-good case, the elasticity of substitution is then equivalent to the degree of curvature of the indifference curve, it is symmetric, and it can be interpreted as an indicator of substitutability. As variations in MRS are measured on the same indifference curve, another indicator was needed to express the relationship between diverse indifference curves in the consumer’s scale of preferences. Consideration of hypothetical changes in income allowed them to define income-elasticity of demand for this task. Therefore, alterations in the demand of a particular good X as a result of changes in its price, measured by the price-elasticity of demand could be related to the individual’s scale of preferences through what Hicks called the Fundamental Formula:
where Epx denoted the price-elasticity of demand of good X, Eμ(x) the incomeelasticity of demand, σ the elasticity of substitution, and κx the proportion of total income spent on good X. Theoretically, σ cannot be negative, thus imposing a verifiable restriction on econometric demand equations. Besides, this expression provided a decomposition of the price-elasticity of demand into ’’income’’ and ’’substitution’’ effects, which was used in the paper to define with precision the conditions under which Marshall’s ’’law of demand’’ holds.
The generalization to the case of more than two goods allowed Hicks and Allen to propose a new definition of complementarity. The old one, derived from Pareto and Edgeworth, was based on the impact that marginal changes in the supply of one good would have on the marginal utility of other. When scales of preference replace cardinal utility functions, there is no two-good relationship that corresponds to the concept of complementarity, because marginal utilities are indeterminate. Instead, the new definition was based on the impact that marginal changes in the relative consumption of two goods would have on the MRS of the other goods for one of the first two. They called the measure of this impact the ’’elasticity of complementarity’’ between any two goods. Assuming the satisfaction of additional ’’integrability’’ conditions, elasticities of complementarity would be symmetric and would help to describe a complete ordering of preferences.
The generalization of the Fundamental Formula provided a decomposition of the cross-elasticity of demand into complementarity and income effects. That decomposition opened the door to explain Schultz’s results as the consequence of different income effects for each good, neglected by Marshall as a result of his simplifying assumption of a ’’constant marginal utility of money’’. Therefore, the preference approach was shown to be more appropriate for the interpretation of empirical evidence.
Some of the new concepts introduced in Hicks and Allen, like the income-elasticities of demand could be based on available empirical evidence. But there were others, like the assumption of ’’increasing MRS’’ or the satisfaction of the integrability conditions, which could not be dealt that way. To adduce introspective experience in their favour would mean reintroducing utilitarian concepts through the back door.
Hicks did not consider that integrability conditions were really crucial. While Allen’s part of the paper showed great care to derive the main results without having recourse to integrability, at the cost of increased complexity, in Hicks’s part of the paper the conditions are disposed of very cursorily
As (Samuelson 1974) has noted, subsequent works by...
This proposition, which exercises a great fascination over the minds of mathematical economists, remains of doubtful economic significance. (Hicks and Allen 1934, p.56, fn.1).
The implications of the assumption of increasing MRS were much more serious. All the results obtained in the paper rested on its satisfaction. The only justification provided about it was the following:
To assume it true universally is a serious assumption, but one which seems justifiable until significant facts are adduced which make it necessary for us to pay careful attention to exceptions. (Hicks and Allen 1934, p.58).
In other words, the preference approach based on increasing MRS should be accepted until its implications were falsified by empirical data. Otherwise, equilibrium analysis of consumer behaviour would not be possible.
Hicks’s research on general equilibrium continued after his move from LSE to Cambridge in 1935, and came to be published as a book with the title of Value and Capital (Hicks 1939). There, the substance of his previous paper with Allen was restated in Part I, with some minor differences. In the first place, Hicks wanted the book directed to a wider audience, and so chose to expose his ideas on demand theory in a form more closely related to the Marshallian one, because he considered English-language readers were more acquainted with it. Substitution of the preference approach for the cardinal utility one was presented, in a less controversial way than in (Hicks and Allen 1934), in the form of a generalization. The Marshallian assumption of a ’’constant marginal utility of money’’ was considered as a simplification that could be valid if applied to those situations where income effects were empirically irrelevant. Hicks considered that the particular problems Marshall wanted to address were precisely of that kind, but that for other purposes it was better to have more general hypotheses at the economist’s disposal.
These changes were also facilitated by the rediscovery of (Slutsky 1915), that provided a much more elegant mathematical formulation of the whole theory. Slutsky’s formulation made use of partial derivatives of utility functions without depending on the existence of any determined utility function. Hicks adopted this formulation, instead of that used in Hicks and Allen’s paper, in the Mathematical Appendix to Value and Capital, with only minor variations. He seemed to be convinced at the time that the concept of the utility function could be useful as an auxiliary construction that greatly simplified the mathematical version of the theory
Later, he would regret ’’the disappearance, from most...
Other modifications were dictated by the purely theoretical nature of the book. The econometric implications of (Hicks and Allen 1934) had no place in it
But were not neglected, as it is shown by the following...
. This altered the type of criticisms raised against cardinal utility. Pareto’s demonstration of the indeterminateness of utility functions was repeated, but it alone was not a very persuasive argumentation. Hicks was led to put more emphasis on the rejection of the introspective basis of the traditional assumptions about measurable utility, like that of diminishing marginal utility. He refused explicitly to accept subjective experience as empirical evidence, arguing that it cannot be tested
Statements based on the concern of demand theory with...
. But this left the correspondent assumptions in the preference approach, like diminishing MRS (redefined this way to resemble more closely Marshall’s own terminology), in need of a justification. The explanation offered for them was rather obscure, relying on the need to assume some regularity (i.e. the absence of ’’kinks’’ in indifference curves) to warrant the possibility of stable equilibria. In the end, this amounted to repeat again that the assumptions were better accepted in absence of empirical falsification, adding to it that it was the simplest available and that it was of a similar nature to the usual assumptions supporting the whole building of economic theory
(Hicks 1939, pp.23-24). That this justification was...
3 - An ideal representative consumer
The wide influence exerted by (Hicks 1939) helped to establish the preference approach as the dominant one in demand theory. The book soon became a classic, and defined the direction of economic research in the subsequent years. The utilitarian basis of the Marshallian approach was definitively discarded, and marginal utilities remained only as a mathematical convenience. A new generation of economists devoted its time to problems linked to the development and correction of the main themes advanced in Value and Capital. Regarding demand theory, efforts concentrated in two main directions
Both originated from the work of Samuelson (Samuelson...
. On one hand, the logical structure of the preference approach was thoroughly developed, leading to a fully axiomatized general theory of choice, based on the mathematical theory of ordering, of which consumer decisions were a particular case. On the other hand, dissatisfaction with what was perceived as an insufficient justification for the hypothesis of diminishing MRS, gave rise to an entirely new interpretation of consumer choice. The ’’revealed preference hypothesis’’ claimed to avoid any reference to non-observable concepts and substituted behaviourism for utilitarianism as the foundation of demand theory.
In the meantime, Hicks directed his research to welfare economics and the rehabilitation of Marshall’s concept of the consumer’s surplus using the preference approach. His move from Cambridge to Manchester and the Second World War kept him distanced from the new developments until 1946, when he travelled to the United States and had discussions with Samuelson, Arrow, and others. The reception he then gave to the new directions taken by demand theory was a mixed one: the generalization of the preference approach was congenial to him
The potential for this kind of generalization was present...
, but revealed preference was not.
He wrote an entirely new version of his views on demand theory, which was published as a book with the title A Revision of Demand Theory (Hicks 1956). The exposition of consumer theory that Hicks presented in this book marks the central stage in the evolution of his ideas. Revealed preference forced him to revise the foundations of the preference approach from the bottom up, because the new approach could be defended with the same reasoning he had advanced before in support of his own views. Not only it was at least as simple as the principle of diminishing MRS, but also their empirical implications seemed equivalent. On top of that, the new approach claimed the advantage of being more clearly distanced of utilitarian implications. In fact, revealed preference was proposed as an interpretation that thrived on the whole logical structure of the preference approach, just as this approach thrived on that of the cardinal utility approach.
What Hicks found unacceptable with revealed preference was its purpose of explaining actual individual behaviour in rational terms. This was for him too much ambitious:
I retain a considerable scepticism about the ’’Revealed Preference’’ approach. It is a long-standing trouble of demand theory that it is always tempting us to overplay our hand..
(Hicks 1956, p.vi)
He doubted that any simple assumption, like that of consistency with respect to given preferences on the face of changes in prices and income, could provide a precise account of so complex a matter like actual human behaviour. In consequence he thought that, in case consistency tests based on the axioms of revealed preference could be applied to individuals, they would generally show inconsistent behaviour. In such situation, inconsistent behaviour could be attributed to diverse causes: changes in preferences, non-transitive preferences, change in other variables, etc. The impossibility of assigning priority to any one of them undermined the usefulness of the approach. He expressed the opinion that for economic purposes, the more modest task of explaining aggregate group behaviour was enough.
Furthermore, Hicks warned about the difficulty in inferring consumer’s preferences from the statistical data usually available, i.e. market data reflecting aggregate behaviour. He argued that to apply consistency tests to such data it was necessary to assume that consumers were homogenous, in preferences as well as in income. Otherwise, apparent inconsistencies could be attributed to diversity in any of these two characteristics.
On the constructive side, Hicks offered detailed argumentations to support his own preference approach as the best alternative for the explanation of aggregate consumer behaviour
The explanations were couched in the new fully axiomatized...
. These expanded the views presented in his previous works. In the first place, he identified rational behaviour as a common assumption in Economics. Hicks considered its use was so extended that it could be claimed to define the whole discipline. He advanced the view that utility theory was simply an extension of production theory based on profit maximization. Traditional production theory took producer behaviour as motivated by the pursue of profits:
So far as man the producer (…) is concerned, it has commonly been thought that adequate rules could be got from the principle of maximizing money gain -the economic man. (…) (I)t was natural to approach the behaviour of the consumer in something like the same way, and to look about for something which he might be considered to maximize. Thus we got the Utility theory (…)..
(Hicks 1956, p.5)
Therefore, the pretended psychological basis of cardinal utility was really an auxiliary construction to persuade doubters about the verisimilitude of maximizing consumers. Demand theory could be included as another part in the program of explaining economic facts as the result of rational behaviour by self-interested agents.
Economic facts are usually of statistical nature. Therefore, the theory has to take into account the needs imposed by econometric methods. For econometric purposes, the rationality assumption is needed as a classification device. Consumption data can be explained from a wide number of different variables, and the statistical methods are designed to identify the effects of each one. To estimate the effects attributable to economic variables it is necessary to isolate economic from non-economic explanations of consumption decisions:
The kind of theory which is needed for this purpose is one which will tell us something about the ways in which consumers would be likely to react if variations in current prices and incomes were the only causes of changes in consumption. This is precisely what the theory of demand, considered from the econometric point of view, has to do. It proceeds by postulating an ideal consumer, who by definition is only affected by current market conditions, and asks how we should expect such a consumer to behave..
(Hicks 1956, p.17) (italics in the original)
It was at this point where Hicks defended again the hypothesis of rationality on the grounds of simplicity and correspondence with the procedures followed in the rest of economic theory.
Hicks considered that this justification was easily extensible from Econometrics to ’’Plain’’ or ’’Pure’’ Economics -the object of (Hicks 1939)- because, in fact, the econometric procedure above described is based on pure economic theory.
One strong implication of the necessity to establish artificial classifications to isolate economic explanations is that the consumers of demand theory cannot be considered as actual individuals. The explanation given to their behaviour is not a full explanation of individual behaviour. The conscious adoption of this point of view, which Hicks later would call ’’the cool line’’ (Hicks 1981, p.xiv), permitted him to justify the particular form of rationality he preferred to assume, following the criterion of convenience. Each time Hicks had to decide among alternative assumptions in the building of his theory, he acted on the premise that ’’(s)ince the preference hypothesis is only a hypothesis, we are at liberty to assume it in any form that we choose(…)’’ (Hicks 1956, p.23). Therefore, he defined preference orderings as transitive instead of non-transitive, to avoid problems of circularity and unconnectedness
This is the assumption equivalent to the mathematical...
(Hicks 1956, p.23), he assumed weak instead of strong orderings, to allow for continuity (Hicks 1956, p.44), and so on. The abstract nature of these hypotheses allowed him to build a relatively simple theory with sharp empirical implications.
As stated above, Hicks defended his approach on the grounds of the possibilities it opened for empirical applications. At first sight, those possibilities seem limited. Empirical market data are the product of the behaviour of a multiplicity of actual consumers that do not generally satisfy the ideal properties of rational behaviour, defined as the consistency of the decisions taken with a fully integrable ordering of preferences that satisfies additional properties like local non-satiation.
The procedure to deal with such data was sketched in (Hicks 1958). It is based on the consideration of the ideal consumer described by the theory as ’’representative’’ of the group of consumers actually taking decisions in the market
Hicks uses sometimes the term ’’average’’ as synonymous...
. Under the assumption that the differences between actual individual preferences and those of the representative consumer are due to random causes (i.e. are not related to non-economic explanations), a sufficiently large population of actual consumers would present a similar aggregate behaviour to a population of representative consumers, each of them identical with any other in preferences and income. If this condition is not fulfilled, then the difference in behaviour must be attributed to some non-economic explanation
It is easy to recognize this methodology in the representative...
. One clear defect of this construction is that it neglects distributive issues. For situations in which distribution is considered too important to be set aside, Hicks proposed to construct models with several representative agents, each one representing one social group or class
An idea not very dissimilar to the models with heterogeneous...
Hicks planned initially to complement his revision of demand theory with a parallel work on welfare economics, something that he would only accomplish partially in his aforementioned 1958 paper. The revision of Hicksian welfare economics would never be completed, and Hicks changed his research interests to another subjects. He returned to write about demand theory from time to time, usually to maintain the positions expressed in (Hicks 1956). However, in his little book about methodological questions, Causality in Economics (Hicks 1979), he extended his interpretation of the role of rationality in the history of Economics.
Hicks insisted there on his consideration of the rational behaviour assumption as one of the very foundations of the discipline of Economics. He called it ’’the Economic Principle’’. According to his definition, a rational agent (the ’’Economic Man’’) is the one who, if it is free to choose, acts in order to take advantage of all the opportunities to promote its self-interest. He illustrated the presence of the assumption from the very beginning of modern Economics by advancing an interpretation of Adam Smith’s analysis of locational advantages as an early example of use of the Economic Principle. He attributed to Smith the definition of self-interested action as economic action to demarcate the limits of the analysis and justify the use of the principle, discarding the alternative interpretation that the principle was empirically based on the observation of ’’capitalist’’ behaviour. On this account, Smith was also a pioneer in the use of what Hicks called the ’’equilibrium method’’, i.e. the interpretation of economic data such as prices and income as the result of a situation where the economic men have taken all the opportunities of advantage to them presented
This is precisely the concept of ’’equilibrium’’ applied...
4 - A working-class housewife
After publishing the first volume of his collected essays, Wealth and Welfare (Hicks 1981)
Which contained reprints of (Hicks and Allen 1934)...
, it seemed that Hicks had closed his involvement with the issues of demand theory and consumer rationality. The parts especially written for the occasion (introduction and prefatory notes), together with some papers previously unpublished, expressed substantially the same views that he had repeated in various occasions since 1956. But when we came back, yet again, to explain his views on rationality on a volume of ’’essays in honour of Ludwig M. Lachmann on his eightieth birthday’’ (Hicks 1986), he gave an unexpected turn to some of those views.
The paper poses the question of the interpretation of rational behaviour in a very wide context, spilling out of the borders of economics and into other related social sciences. Hicks used this detachment from the economic field in order to illustrate better his often-repeated claim that Economics is concerned with classes of decisions, not with particular ones. This meant again that the behaviour that has to be explained is that of a representative agent. Thus he arrived again to the point where it is necessary to make an assumption about that behaviour, noting that rationality is the usually adopted one in Economics. To explain why this is so, he almost reversed his exposition of (Hicks 1979) about Adam Smith’s use of the Economic Principle.
One can evade this issue by defining economic behaviour as being such that it is susceptible of being analyzed in that manner. I gave some countenance to that approach in my Causality in Economics (…), but I have come to feel that it is not good enough..
(Hicks 1986, p.297, fn. 4)
In place of his former explanation, he embraced the alternative there rejected. Rational behaviour was the product of observation, after all. Drawing on his Theory of Economic History (Hicks 1969), he considered it was the behaviour of the merchant which served as a model for the Economic Man, because the business of the merchant is no other than to make profits, i.e. to be rational.
The main part of the paper is devoted to pursue the implications of this idea. Hicks observed that the drive for profit was the basic motivation behind the invention of accounting, so that the practice of keeping accounts could be taken as external evidence of the profit motive. The general extension of that practice in Smith’s time could then be adduced in favour of the interpretation that Smith assumed rational behaviour because he actually observed it.
The extension of the rationality assumption from the merchant to the businessman throws in some complications, which were revised in this paper. The weight of the evidence considered allowed Hicks to claim plausibility for the assumption of profit-making rationality, although he recognized the existence of particular situations in which alternative approaches, like that of ’’satisficing’’ behaviour, would be preferable. Labourers seem to be subject to similar considerations.
The judgement about the rational consumer is not so clear. He acknowledged that the main justification for the assumption is ’’convenience’’: the mathematical appeal of the possibility to consider consumer behaviour as a problem of maximization against constraints and the suitability of the assumption for the econometric interpretation of data. But in this occasion he did not considered convenience to be enough. To provide a full justification of the use of the assumption, additional evidence would be necessary. External signs of rationality are needed in order to identify the situations that can be analyzed by the preference approach. At the end of the paper, as an illustration, Hicks identified some of the features of the kind of evidence required.
One of the limits to the application of the rationality assumption in consumer behaviour clearly identifiable is the need of knowledge about the relevant alternatives. The consumer has to be in the position to know all the relevant alternatives between which to choose, or at least those that are also known to the observer. This means that the applicability of the rationality assumption is limited to those situations in which that knowledge is clearly possessed by the consumers, or at least to those where it is obvious that they could get it without a great waste of resources in comparison to the gain in satisfaction which they obtain from consumption. An empirical counterpart of this situation would be one in which the consumers were people with reduced incomes (so their marginal gain from spending would be large) and ample spare time (so their cost of opportunity would be low). Hicks provided an example of that situation which he obtained by personal experience when he was in Manchester and could observe the shopping of working-class housewives in a time when there were little employment opportunities. That let them time to search for information about goods and prices, as a rational consumer would do.
This example is a very restrictive one and contrast vividly with the optimism expressed years before about the scope of representative consumer models. For a better appreciation of its intended meaning it would be necessary to know the reasons that led Hicks to this sudden change in the views he had held in print for more than fifty years. Although these reasons are not explicit in the paper, some clues can be taken from the conversation he held in 1986 with Arjo Klamer, which was published as (Klamer 1989). There, after briefly describing the content of the paper, he expressed his doubts about the propriety of assuming rationality for certain classes of human behaviour. Klamer then asked him to precise to which classes of behaviour he was referring, and he answered that writing ’’about population, and population behaviour, all in strictly maximizing terms (…) seems to me pretty ridiculous (…) that was something I wanted to avoid’’ (Klamer 1989, p.172). This seems to suggest that Hicks was very concerned about the extension of economic analysis, in the form of representative agent models, to a wide variety of problems traditionally considered as out of the discipline.