I. INTRODUCTION
The general notion that individual preferences are influenced by the preferences of other individuals, which is the basis of the modern term “interdependent preferences,” is present in the works of a substantial number of eminent thinkers in the history of economic thought. One can even find clear signs in the theories of eighteenth-century authors such as A. Genovesi and D. Hume, and also in a large number of classical economists such as Smith, Ricardo, Marx, Senior, Mill, and many others, who discuss these issues in sympathetic terms. Furthermore, preference interdependence plays a central role in the contributions of many very well-known post-marginalist economists, such as T. Veblen, A.C. Pigou, and J. Duesenberry. There is a renewed interest in the idea recently, especially in many macroeconomic and labor papers, and also in the relatively modern research area of subjective well-being. Many theorists in these fields have found that the incorporation of interdependent preferences can help explain a variety of economic phenomena, such as wage rigidity, savings patterns, and the Easterlin paradox (see Bruni Reference Bruni2004; Drakopoulos Reference Drakopoulos2011).
In spite of the above, the idea is still ignored—to say the least—by the majority of mainstream economists, who continue to assume independent individual preferences. There are a number of reasons for this attitude, which have mainly to do with the methodological foundations of orthodox economics and with the theoretical consequences of interdependent preferences. More specifically, the notion of interdependent preferences was effectively ignored by the founders of modern mainstream economics, such as V. Pareto and A. Marshall. Pareto thought that interpersonal preferences were not part of his definition of “logical action” in economics and that their analysis belongs to sociology and not to economics. Marshall was aware of status-driven consumption but he condemned such expenditures, believing they were rooted in personal vanity and created envy in others. Thus, Marshall and Pareto did not encourage the incorporation of the idea of interpersonal preferences, and this certainly was an important contributing reason why such preferences did not become part of the corpus of subsequent mainstream analysis and of mainstream welfare economics. Orthodox reactions towards the work of J. Duesenberry on consumption theory, which is based on the idea of interdependent preferences, is a more modern representative example of this stance. Based on the historical discussion, this paper will attempt to show that the mainstream methodological preconceptions (including the prevalence of selfish economic agents) and the challenging theoretical implications of adopting interdependent preferences were two crucial reasons for the negative mainstream attitude.
This paper begins with a brief review of the presence of interdependent preferences in the history of economic ideas. It argues that the idea, in its general form, was part of the reasoning of many pre-classical and classical economists. The next section shows that this presence continued in the work of Veblen and Pigou, who explored the analytical consequences of such preferences. The next section discusses how Pareto’s and Marshall’s stance contributed to the abandonment of such preferences from subsequent mainstream economic theorizing. The following section discusses the mainstream attitude towards the work of Duesenberry as the main example of the continuation of mainstream disregard and the possible reasons for this. The next section provides a brief discussion of the more recent developments on this matter. Finally, a concluding section closes the paper.
II. INTERDEPENDENT PREFERENCES: METHODOLOGICAL POINTS AND EARLY PRESENCE
The modern term “individual preferences” was introduced as a part of the twentieth-century utility maximizing theoretical framework. As a special class of preferences, interdependent preferences are conceived, in most contemporary microeconomic theorizing, as expressing mutual and reciprocal influences among agents. However, there is a general methodological issue of using contemporary terms retrospectively when attempting to trace modern concepts in the history of economic thought.Footnote 1 Modern terms are usually part of modern theoretical frameworks with specific meanings, which need not be identical when viewed from an historical perspective. This issue is also present in the case of an historical discussion of interdependent preferences, as in most cases of the history of ideas-based investigation (see, for instance, Sugden Reference Sugden2002). One possible way to tackle this problem in our case is to adopt a general approach emphasizing the role of social influences and social comparisons on individual economic decisions. Thus, the main term used in this paper is conceived in a broad sense to reflect the general notion that individual preferences and tastes are constantly influenced by the preferences and actions of other individuals.
The general approach to the concept implies that a number of sub-categories or classes of interdependent preferences can be discerned.Footnote 2 Altruism and joint consumption are two examples of such classes. Positional goods, which are also the basis of status in the sense that they bestow status, are another main example of such preferences (see also Dugger Reference Dugger1985). Positional goods, as a class of interdependent preferences, also have a long presence in the history of economic ideas. In modern economics, the term “positional goods” appeared first in the work of Hirsch (Reference Hirsch1976). Hirsch conceived the notion as referring to goods, services, work positions, and other social relationships that are characterized by direct social scarcity, or, in other words, available only to a minority. However, since its first introduction, there is still no universal agreement concerning the exact nature of positional goods (see also Schneider Reference Schneider2007). For instance, Frank adopts a much more general definition than Hirsch’s when he states that positional goods are demanded not for any absolute property that they posses, but “because they compare favorably with others in their own class” (Frank Reference Frank1985b, p. 7). Frank’s broader definition is closer to Veblen’s “conspicuous consumption” and to Duessenbery’s “demonstration effects,” and, in this sense, it is more related to the general notion of social comparisons.
The general notion of interdependent preferences is ultimately based on the social nature of human beings. In fact, one can argue that the social nature of humans implies the behavioral importance of interdependent preferences. Clearly, the social dimension of individual decisions is not a novel idea, since it was first analyzed and emphasized by ancient thinkers and especially by Aristotle (for a review, see Schneider Reference Schneider2007). In modern times, this concept is present in the writings of many influential authors in the history of early economic thought.
The Neapolitan Antonio Genovesi (1713–1769) is one of the first authors whose social and economic thinking is characterized by an emphasis on the sociality of human nature, which is seen as the main function of society. For Genovesi, the chief advantage of society is not to be found in its production of material goods, but in the enjoyment of social relationships (Genovesi 1766, quoted in Bruni, Reference Bruni, Bruni and Porta2007, p. 31). In the same period, Sir James Steuart’s economic work can also be seen as having grasped the basis of what we call today “interdependent preferences.” As he writes:
the moment a person begins to live by his industry, let his livelihood be ever so poor, he immediately forms little objects of ambition, compares his situation with that of his fellows who are a degree above him, and considers a shade more of ease … as an advancement, not of his happiness only, but also of his rank (Steuart Reference Steuart and Skinner1767, p. 272).
D. Hume was also well aware of the importance of the social aspect of human behavior. He states: “The passions are so contagious, that they pass with the greatest facility from one person to another, and produce correspondent movements in all human breasts” (Hume Reference Hume1736, p. 605). Similar to this is Hume’s strong objection to the assumption of self-interest, which implies autonomous and thus non-interdependence of individual preferences (Hume Reference Hume1736, p. 487).
The social dimension of economic decisions can also be found in the work of many classical economists. Adam Smith recognized that individuals engage in social comparisons. In his Theory of Moral Sentiments, he also included “the love of distinction” as one of the natural human motives (1759, p. 50). The basis of this motive is the human “vanity which is always founded upon the belief of our being the object of attention and approbation” (1759, p. 50), and is “natural to man” (1759, p. 182). The idea of comparison of our income to a class or group of individuals, (what is known as “keeping up with the Joneses” in modern terms) is also present in this work (Smith Reference Smith, Raphael and Macfie1759, pp. 212–213). Similar ideas can be found in his Wealth of Nations, as the following passage indicates:
With the greater part of rich people, the chief enjoyment of riches consists in the parade of riches, which in their eye is never so compleat as when they appear to possess those decisive marks of opulence which nobody can possess but themselves (Smith Reference Smith, Campbell, Skinner and Todd1776, vol. 1, p. 190).
Adam Smith’s contemporary, John Rae, was also well aware of the power of consumption imitation. Rae’s ideas on the subject are quite similar to the ones expressed subsequently by T. Veblen, although it is not clear if Veblen was aware of Rae’s work (see Edgell and Tilman Reference Edgell and Tilman1991). Rae suggests that consumption is regulated, to a great degree, by the influence of the imitative propensity (Rae 1834, pp. 293–294). He also discusses further the idea of conspicuous consumption (influenced to a large extent by Adam Smith).Footnote 3 He attributed such behavior to the selfishness and vanity of individuals, and seems to have distinguished between mere luxuries and what were to become known subsequently as “Giffen goods” (Rae 1834, pp. 268, 270, 292).
R. Whately argued (1832, p. 51) that the level of consumption and of the consumption pattern is socially determined: “an individual man is called luxurious, in comparison with other men, of the same community and in the same walk of life with himself” (Whately, 1832, p. 53). Similarly, M. Longfield advances the argument that “the wages of the labourer depend upon the expense of his maintenance and usual style of his living, instead of his expenses and his mode of living depending pretty much upon his wage” (Longfield 1834, p. 203). Nassau Senior, in his Outline of the Science of Political Economy, drew attention to what he called “the desire for distinction.” Having discussed the desire for variety, and having attributed to it the diminishing utility to be gained from each additional unit of a particular commodity consumed, Senior unequivocally declared it to be less powerful than the desire for distinction (Senior 1836, p. 12).
Marx had also conceived the idea of relative income and consumption, and, in general, the concept of positional goods in the broad sense. Hence, the quotation “let a palace arise beside the little house, and it shrinks from a little house to a hut” (Marx 1849, p. 216). Furthermore, there are other passages where the notion of rising immiseration in spite of rising real wages is also present (see also Hollander Reference Hollander1984, p. 146). In a similar vein, J.S. Mill points out that once a basic standard of living has been achieved, the concern about social status becomes extremely strong (Mill 1874, p. 411).
Finally, by the end of the nineteenth century, Henry Cunynghame provided the first attempt to illustrate diagrammatically the effect of incorporating dependence of one individual’s demand for a good on that of others, in his 1892 article entitled “Some Improvements in Simple Geometrical Methods of Treating Exchange Value, Monopoly, and Rent.” In introducing this new diagram, Cunynghame observed:
almost the whole value of strawberries in March, to those who like this tasteless mode of ostentation, is the fact that others cannot get them. As my landlady once remarked, ‘Surely, sir, you would not like anything so common and cheap as a fresh herring?’ The demand for diamonds, rubies, and sapphires is another example of this. As the number increases, not only does the price go down, but the very pleasure of those who already have them is decreased by their becoming common (Cunynghame Reference Cunynghame1892, p. 37).
Cunynghame’s diagram showed a consumer surplus curve lying below the commodity demand curve, but shifting upwards for every reduction in the quantity supplied.
Thus, as our previous discussion indicated, reasoning akin to what is termed today “interpersonal comparisons” was present in the economic thought of the eighteenth and especially nineteenth centuries. This reasoning was part of the framework that put emphasis on consumption as a basic economic activity, including its effects on productivity and economic growth. Thus, classical economists were interested in all aspects affecting consumption and consumption patterns (see Perrotta Reference Perrotta2004, ch. 11). However, for most of the authors discussed, the above ideas were not part of a systematic theory of individual economic behavior. To a large extent, this was due to their adherence to a cost of production theory of value and to their largely macroeconomic viewpoint. This implies that a full theory of individual preferences (in which interdependent preferences belong) was not a necessary part for the construction of economic theories. A more systematic discussion of these ideas, however, became apparent with a shift towards a more subjective approach to the theory of value, which took place in the last decades of the nineteenth century. However, as we shall see, orthodox theorists effectively ignored them for a number of reasons. But before we proceed to this theme, let us see first the more detailed discussion of interdependent preferences in the works of T. Veblen and A.C. Pigou.
III. VEBLEN AND PIGOU
Interdependent preferences, and especially their manifestation in the form of positional goods, are central ideas in Thorstein Veblen’s main work. A substantial analysis of these concepts and of their impact can be found in his The Theory of the Leisure Class (1899). These concepts played such a fundamental role in his analysis that they were conceived as the foundation of a private property society. In the following statement, Veblen combines many aspects of the previous discussion. He links consumption and wealth imitation, private property, and the nature and effect of positional goods:
The motive that lies at the root of ownership is emulation; and the same motive of emulation continues active in the further development of the institution to which it has given rise and in the development of all those features of the social structure which the institution of ownership touches. The possession of wealth confers honour; it is an invidious distinction. Nothing equally cogent can be said for the consumption of goods, nor for any other conceivable incentive to acquisition, and especially not for any incentive to the accumulation of wealth.… the end sought by accumulation is to rank high in comparison with the rest of the community in point of pecuniary strength. So long as the comparison is distinctly unfavourable to himself, the normal, average individual will live in chronic dissatisfaction with his present lot; and when he has reached what may be called the normal pecuniary standard of the community, or of his class in the community, this chronic dissatisfaction will give place to a restless straining to place a wider and ever-widening pecuniary interval between himself and this average standard (Veblen 1899, pp. 25–26, 31–32).
Veblen’s famous term, “conspicuous consumption,” is obviously related to the above. In particular, conspicuous consumption is the consumption of luxuries that is observed by others, sometimes by these others participating in it. Conspicuous consumption is not confined to the “rich” or to the leisure class in Veblen’s terminology. Other social classes might also engage in such activity, even the lower classes. As he states:
No class of society, not even the most abjectly poor, foregoes all customary conspicuous consumption.… There is no class and no country that has yielded so abjectly before the pressure of physical want as to deny themselves all gratification of this higher or spiritual need (Veblen 1899, p. 85).
It is clear that in Veblen’s theoretical analysis, interpersonal comparisons of income, consumption, and wealth are extremely important factors for economic and social relationships.
Although Veblen’s ideas as a whole were not very influential for the emerging corpus of mainstream theory, the notion of interpersonal preferences was not uncommon among leading economists at the time. This can be seen in the work of A.C. Pigou, whose ideas were at the center of mainstream economics of that period. Pigou devotes considerable space to the concept of interdependent preferences. As we will see in the next section, Pigou is clearly influenced by Pareto’s usage of the term “ophelimity,” although he does not differentiate it from the term “utility,” as Pareto did (see also McLure Reference McLure2010). In Pigou’s words:
The curve of the private marginal demand prices lies above the curve of collective marginal demand prices if an addition to the consumption of one consumer diminishes the ophelimity associated with a given consumption by other consumers.… The curve of the private marginal demand prices lies below the curve of collective marginal demand prices if an addition to the consumption of one consumer increases the ophelimity associated with a given consumption by other consumers (Pigou Reference Pigou1910, p. 361).
The inequality between private and collective demand prices depends on the emergence of a social element in individuals’ states of consciousness because some or all consumers of a particular product derive utility from their direct consumption as well as consumption by third parties (see also Mclure Reference McLure2010). This is closely linked to his previous analysis in his “Some Remarks on Utility” (Pigou Reference Pigou1903), where he introduces the argument K{a,b} as an element within the individual’s utility function of a specific commodity. More specifically, U = f (...,K{a,b}), where K{a,b} is a “complex expression” in which the element a is the quantity of A possessed by the individual’s first neighbor and b is the ‘distance’ from an individual to his/her first neighbor. If the consumption of good A by neighbors has no effect on the ophelimity enjoyed by an individual from his/her consumption of good A, the ‘distance’ b is zero and thus the value of the complex expression is zero. However, when the consumption of good A by neighbors influences the ophelimity that an individual enjoys from his/her consumption of good A, the ‘distance’ in each of the elements b is non-zero and the value of the complex expression is non-zero. Pigou is clearly influenced by Cunynghame’s theory (1892) that people who like a tasteless mode of ostentation receive diminished pleasure from some commodities when they become more ‘common’ (for an extensive discussion, see Mclure Reference McLure2010).
The emergence of the ordinalist revolution in the 1930s, combined with the growing abandonment of Pigovian welfare economics, undermined two important aspects of Pigou’s theory: measurability of utility; and the theoretical possibility of interpersonal comparisons of utility. This also did not support the related notion of interdependent preferences. In spite of the criticism, however, Pigou did not convert to ordinalism, and held to his theoretical constructions (for a discussion, see Aslanbeigui Reference Aslanbeigui1992, and McLure Reference McLure2010).
IV. PARETO, MARSHALL, AND THE MARGINALIZATION OF INTERDEPENDENCE
As was seen in the previous sections, interdependent preferences were part of the economic thinking of a great number of influential theorists. However, the marginalist emphasis on the subjective theory of value required a systematic model of individual economic decision making. Thus, the issue of the nature of preferences had to be addressed. Vilfredo Pareto was the first major economist who provided the rational for the irrelevance of interdependent preferences for economic analysis, and thus contributed to their non-incorporation in mainstream economics.
Pareto regards human action as an extremely complex issue subject to many types of influences. Thus, from his early work, he introduces two distinct concepts, utility and ophelimity, which attempt to capture the multi-dimensional nature of human action. The term “utility” has a general character, while the term “ophelimity” is closer to a subjective sensation and is more specialized to the purposes of economics (Pareto 1896, pp. 3, 126). Pareto’s “economic man” is conceived as a rational being in the sense of a means–ends relationship, and this rationality can be achieved through a trial-and-error process. Moreover, Pareto’s economic man responds only to forces of economic ophelimity, and, in this sense, it has an atomistic basis, meaning that, in this framework, individual preferences are independent (Pareto 1896, p. 12). Although Pareto admits that man’s character presents other characteristics too, these are studied by other sciences. Once an individual’s preferences are influenced by interaction with the actions and ideas of third parties, action is considered with respect to utility and not ophelimity (McLure Reference McLure2010). In a subsequent article, Pareto (Reference Pareto1918) is more specific about the two scientific domains, in the sense that when the ‘economic part’ of the economic phenomena dominates, analysis should be based primarily on economic theory; and when the sociological part of the economic phenomena dominates, analysis should be based primarily on sociology.Footnote 4 It should also be pointed out, though, that in his sociological works, Pareto studied the formation of preferences and also considered the effects of social influences on preferences. For instance, in the context of his discussion on social welfare, Pareto employs the idea that each individual has a view about the relative benefits from their own consumption and from the consumption of other members of society (Pareto 1916, p. 1472; McLure Reference McLure2010).
Pareto’s overall approach regarding the nature of economic man had a strong methodological motive. Pareto was extremely influenced by the prevailing positivist scientific philosophy, a basic characteristic of which was the exclusion of all ‘metaphysical’ and ‘non-scientific’ elements from economics. Pareto’s methodological ideal of economics was that it should be a mathematical science, part of the natural sciences such as physiology and chemistry (Pareto 1896, p. 21). This implies that economics should be freed from any philosophical or psychological ideas that hamper the application of the positivist methodology (for an extensive discussion, see Seligman Reference Seligman1969; Drakopoulos Reference Drakopoulos1997). Pareto’s analysis of indifference curves, without assuming the controversial and rather psychological notion of cardinal utility, can be seen in this perspective. Interpersonal preference is a form of interpersonal utility comparisons, and cardinal measurability of utility greatly facilitates interpersonal comparisons. The subsequent dominance of the ordinalist approach to utility theory, attributed mainly to the work of Hicks and Allen (Reference Hicks and Allen1934), promoted further Pareto’s views on the model of rational economic man, which excludes social influences such as interpersonal preferences (Ranchetti Reference Ranchetti and Bianchi1998; Bruni and Guala Reference Bruni and Guala2001).
Alfred Marshall’s influence on the formation of modern mainstream microeconomics is widely recognized. Marshall was aware of consumption for the purposes of status, but he condemned such behavior basically on ethical grounds:
There is some misuse of wealth in all ranks of society. And though, speaking generally, we may say that every increase in the wealth of the working classes adds to the fullness and nobility of human life, because it is used chiefly in the satisfaction of real wants; yet even among the artisans in England, and perhaps still more in new countries, there are signs of the growth of that unwholesome desire for wealth as a means of display which has been the chief bane of the well-to-do classes in every civilized country. Laws against luxury have been futile; but it would be a gain if the moral sentiment of the community could induce people to avoid all sorts of display of individual wealth. There are indeed true and worthy pleasures to be got from wisely ordered magnificence: but they are at their best when free from any taint of personal vanity on the one side and envy on the other (Marshall 1890, book III, ch. VI, p. 137).
Thus, Marshall did acknowledge the problem, but he did not consider further the analytical consequences of status-driven consumption. The main reasons for this were methodological—no theory can be all-encompassing—and also he did not think this was a widespread phenomenon (see also Bianchi Reference Bianchi2010). In addition, it seems he was reluctant to discuss the general case of interpersonal effects on demand, mainly because that would have called into question the most convenient functional form of the standard theory: that aggregate demand could be derived from the simple addition of individual demand schedules (Mason Reference Mason1995). Although a number of authors, such as Cunynghame (Reference Cunynghame1892) and Foley (Reference Foley1893), clearly pointed out this lack concerning status-driven consumption, Marshall effectively ignored these criticisms and did not analyze the issue further in subsequent editions of his Principles (see also Mason Reference Mason1995).Footnote 5 As Bianchi (Reference Bianchi2010, p. 7) concludes, this attitude exposes “a resistance to include the topic of interrelated preferences in economic theory.”
In general, Pareto’s and Marshall’s attitude towards interdependent preferences did not promote their incorporation in mainstream economic analysis. Marshall’s explicit omission to examine their role, and Pareto’s methodological ideas against including philosophical and psychological elements in economic analysis, which started to gain wider acceptance with the emergence of the ordinalist revolution, seem to be important reasons for the mainstream neglect. Thus, by the first decades of the twentieth century, the idea of interdependent preferences had been effectively marginalized from mainstream economic analysis, although it had a constant presence in the earlier history of economic thought.
V. J. DUESENBERRY AND THE REAPPEARANCE OF INTERDEPENDENT PREFERENCES
The next major analytical use of the concept of interdependent preferences appeared in James Duesenberry’s main book, Income, Saving and the Theory of Consumer Behaviour, published in 1949. One can view Duesenberry’s work as a continuation of Veblen’s, given that there are many common points concerning income and consumption comparisons, and also concerning the role of the demonstration effect (see McCormick Reference McCormick1983). In particular, Duesenberry is known as the proponent of the relative consumption hypothesis, the basic idea of which was that “any particular consumer will be influenced by consumption of people with whom he has social contacts” (Duesenberry Reference Duesenberry1949, p. 48). This idea (labeled as the “demonstration effect”) questioned the established mainstream view that absolute levels of income only determine patterns of consumer demand (Duesenberry Reference Duesenberry1949, p. 27). Duesenberry proceeded further to analyze the basis of such behavior.
We can maintain then that the frequency and strength of impulses to increase expenditure depends on frequency of contact with goods superior to those habitually consumed. This effect need not depend at all on considerations of emulation or ‘conspicuous consumption’ (Duesenberry Reference Duesenberry1949, pp. 27–28).
Duesenberry’s consumption approach shows how levels of expenditure could be increased, not by changes in income or prices, but by following the consumption expenditures of others with whom the individual had contacts (Duesenberry Reference Duesenberry1949, p. 29). This implies serious consequences for a number of important theoretical issues, such as the pattern of savings and growth, as Duesenberry himself intended. In this framework, savings rates depend on the position of income distribution and not exclusively on the income level, as in a traditional savings function.
Duesenberry’s work was acknowledged as a very important contribution by many leading economists of the period. K. Arrow’s book review in the American Economic Review was very positive, describing it as “one of the most significant contributions of the postwar period to our understanding of economic behaviour … [and] in the best tradition of economic thought” (Arrow Reference Arrow1950, p. 906).Footnote 6 Similar positive views were expressed by G.L.S. Schackle and A.C. Pigou (see Mason Reference Mason2000). However, much in the same way as the previous approaches on interdependent preferences, Duesenberry’s ideas never gained popularity among mainstream theorists. As was discussed previously, this was partly due to the established orthodox conceptual framework, which was not susceptible to interpersonal preferences. However, there appeared some new forms of reaction, which were more specific. One type of reaction against the inclusion of such preferences was to diminish their theoretical importance. Robert Clower’s stance a few years after Duesenberry’s publication is indicative:
Interdependent preferences analysis differs but little from ordinary consumption theory; hence, while one sometimes gets different and slightly more general results in “interdependent” as compared with “independent” preference analysis, it is only in this very limited and practically unimportant sense that the two kinds of theory may be said to be at variance with one another (Clower Reference Clower1951, p. 178).
However, a number of relevant studies have indicated that the two theories are fundamentally different and that they produce different results. For instance, in an early paper incorporating interdependent preferences, Pollak (Reference Pollak1976) demonstrates that when individual preferences are influenced by other people’s consumption, the standard condition—that we can infer an individual’s preferences from his/her market behavior—is problematic. Furthermore, a pay-as-you-go pension scheme and a progressive income tax may well improve efficiency under Duesenberry’s assumptions, rather than to impair it as deduced in standard welfare economics (Holländer Reference Holländer2001; Layard Reference Layard1980; Frank Reference Frank1985a).
The position of Modigliani and Brumberg towards interdependent preferences, and more specifically towards their specific expression in Duesenberry’s analysis, was somewhat different from Clower’s. Although, initially, Modigliani (1949) embraced this idea in his own research concerning relative income effects on consumption, he was much more critical later on, arguing that it contained unnecessary social and psychological elements. Modigliani and Brumberg claimed that their new interpretation of consumption theory was sounder and much simpler (Modigliani and Brumberg Reference Modigliani, Brumberg and Kurihara1954, p. 424; Mason Reference Mason2000). In the same period, Milton Friedman attacked Duesenberry’s formulations by claiming that permanent income rather than relative income was the basis of consumer behavior. Although Friedman acknowledged some merit in Duesenberry’s concept of relative income, he argued it was basically only a biased index of relative permanent income status. Furthermore, Friedman believed his approach was superior, since it owed nothing to sociology or to psychology, in contrast to Duesenberry’s, which was full of subjective elements (Friedman 1957; Mason Reference Mason2000).
Thus, interpersonal preferences were once again excluded from the mainstream corpus of economic theory. One basic reason for this, as is inferred from Modigliani, Brumberg, and Friedman, is that Duesenberry’s analysis poses some fundamental questions regarding the cultural influences on economic decisions and the endogeneity of preferences, issues that were not popular for orthodox theorists. Thus, it seems that allegedly simpler and more “economic” explanations of consumption patterns offered by Modigliani, Brumberg, and Friedman were preferred to Duesenberry’s sophisticated approach (see also Harbaugh Reference Harbaugh1996; Cowling Reference Cowling2006). Another important reason for the lack of acceptance of his ideas was the serious problems they posed for the conventional demand theory, which assumes that individual consumption behavior is independent of the consumption of others. First, the inclusion of ‘demonstration effects’ in standard consumption theory would have made demand-and-consumption measurement far more complex.Footnote 7 Furthermore, during the same period, Samuelson’s (Reference Samuelson1947) revealed preference theory was gaining wide acceptance as a more ‘scientific’ approach based on observed behavior than the traditional utility approach.Footnote 8 However, many theorists suspected that the standard approach to individual demand functions based on revealed preference theory would be problematic if we take into account interdependent preferences (see subsequent discussion, and also Holländer Reference Holländer2001).
VI. RECENT DEVELOPMENTS
For many years after its analytical treatment in the work of Duesenberry, the concept of interdependent preferences was again effectively ignored by the mainstream corpus of economic theory. According to R. Frank, the rejection of Duesenberry’s relative income hypothesis “seems yet another testament to the power of a priori beliefs held by most economists” (Frank Reference Frank1985a, p. 111). The notion was utilized, though, by a few notable economists such as Harvey Leibenstein (Reference Leibenstein1950), where the interdependence of individual preference functions is called the “bandwagon effect”; by Tibor Scitovsky (Reference Scitovsky1962), who attacked the mainstream notion of consumer sovereignty and thus of independent preferences; by J.K. Galbraith (Reference Galbraith1969), who built on many of Veblen’s insights into consumer interdependence, conspicuous consumption, and competitive emulation; and by Fred Hirsch (Reference Hirsch1976), who introduced the concept of positional consumption and positional goods. Furthermore, non-mainstream schools, such as the institutionalist, behavioral, and post-Keynesian schools, also continued to pay attention to the idea (for an extensive review, see Drakopoulos Reference Drakopoulos2011). However, the vast majority of mainstream economists still treated individual preferences as independent.
In the last two decades, though, the research potential of interpersonal preferences has begun to be realized by an increasing number of economists, and, thus, it has begun to reappear in some economic literature. There is an increasing use of ideas such as reference income, target income, relative consumption, positional goods, and social status, which are all based on the concept of interdependent preferences (for a recent review, see Truyts Reference Truyts2010). The subfields of macro and labor economics are indicative examples. In particular, the idea that unions and workers compare income or wages with others has been expressed in a plethora of terms such as “relative wage,” “fair wage,” “aspiration wage,” “comparison or target wage” (see, for instance, Frank Reference Frank1984; Akerlof and Yellen Reference Akerlof and Yellen1990; Clark and Oswald Reference Clark and Oswald1996; Charness and Grosskopf Reference Charness and Grosskopf2001). Furthermore, in the last two decades, the notion of comparison income has also entered the ‘job satisfaction’ literature and, more recently, the quite fashionable subfield of happiness research, mainly in formulations examining the relationship between income and happiness level (for example, Clark and Oswald Reference Clark and Oswald1996; Frey and Stutzer Reference Frey and Stutzer2002; Drakopoulos Reference Drakopoulos2008). Furthermore, there are numerous studies indicating that if utility is conceived as subjective well-being, Duesenberry’s approach is empirically superior to the standard theory (for a recent survey of the evidence, see Holländer Reference Holländer2001; Blanchflower and Oswald Reference Blanchflower and Oswald2004). In spite of the above, the concept, with its specific expression of comparison income or interdependent utilities, is still not accepted by the economics orthodoxy.
It can be argued that there are some apparent reasons for the continuous mainstream resistance towards interdependent preferences. One of these reasons might have to do with their serious theoretical consequences. A number of authors have shown that that the full incorporation of interdependent preferences, and of the related concept of the comparison or relative income, in economic theory would cast considerable doubts on many well-established and important theoretical results. First of all, Pollak (Reference Pollak1976) shows that the derivation of individual short- and long-run demand functions under interdependent preferences requires a number of various conditions and assumptions. In the same tone, Clark and Oswald (Reference Clark and Oswald1996) argue that if preferences are intrinsically interdependent, the standard optimality results of the free market may fail to hold (see also Akerlof Reference Akerlof1997). Furthermore, interpersonal preferences have serious implications for optimal taxation theory. The main idea here is that the presence of income or consumption interdependence results in a negative externality, which undermines standard optimal tax theory. If consumption decisions are influenced by others’ consumption, consumption causes a negative externality by reducing the welfare of other individuals (Layard Reference Layard1980; Frank Reference Frank1985b). From a general macroeconomic point of view, the existence of interdependent utility functions might imply that economic growth is less important than income redistribution from wealthy to less wealthy citizens (see Boskin and Sheshinski Reference Boskin and Sheshinski1978; Kapteyn and Van Herwaarden Reference Kapteyn and Van Herwaarden1980). In general, optimal taxation, economic growth, and income distribution theories might need serious rethinking towards more progressive taxation and more emphasis on more equal income distribution (for reviews, see Dugger Reference Dugger1985; Postlewaite Reference Postlewaite1998; Clark, Frijters, and Shields Reference Clark, Frijters and Shields2008; and, for the implications of interdependent preferences for general equilibrium, see Ackerman Reference Ackerman1997).
Another reason for the rejection of interdependent preferences has to do with the orthodox conception that economic agents are characterized by selfish preferences. In fact, self-interest was one of the cornerstones of the traditional model of individual economic behavior (economic man). The assumption of selfish preferences fitted well with the mathematical framework of constantly maximizing agents of the Walrasian tradition. Although one can argue that selfish preferences have also a psychological dimension, most of the mainstream theorists do not conceive them as such. Self-interest has long been thought of as part of a “rational economic action” without psychological connotations (see also Nicolaides Reference Nicolaides1988; Bowles and Gintis Reference Bowles and Gintis2000). This is certainly still the case for many modern standard formulations. As Frank observes: “To many economists, the notion of consumers being strongly influenced by demonstration effects must have seemed troublingly inconsistent with the reasoned pursuit of self-interest, if not completely irrational” (Frank Reference Frank1985b, p. 146). For instance, in the well-known Stigler–Becker (Reference Stigler and Becker1977) paper, a wide range of social phenomena, such as addiction, custom, and fashion, are analyzed by using the basic utility maximizing framework in which agents are still selfish, although selfishness has a broader definition (see also Sobel Reference Sobel2005). Furthermore, mainstream resistance to questioning self-interested behavior is still very strong in spite of abundant research indicating otherwise (see, for instance, Andreoni Reference Andreoni1995). The assumption is so embedded in mainstream theorizing that, as Rabin points out, “A remarkable amount of energy has been devoted to giving self-interested explanations for laboratory behaviour that seems to be a departure from self-interest” (Rabin Reference Rabbin2002, p. 667).Footnote 9
Finally, another reason closely linked to the above is that mainstream economists are still extremely reluctant to consider psychological and sociological aspects of human behavior in their economic analysis (for a detailed account of the uneasy relationship between mainstream economics and psychology, see Lewin Reference Lewin1996). Concepts such as social status, positional goods, rank concerns, and consumer conformism, which are all based on interdependent preferences, are thought of by many mainstream theorists as not belonging to the realm of economic analysis. The anti-psychologism bias of mainstream economics is also closely linked to the irrelevance of the assumptions thesis expressed in Friedman’s (Reference Friedman1953) well-known paper. The central idea here is that the realism of behavioral assumptions in economics does not matter as long as aggregate data behaves as if these assumptions were accurate (Friedman Reference Friedman1953). This widely accepted methodological position provided an additional shield from criticism, and also support for the pure ‘economic’ approach to human behavior, which is seen as extremely successful and superior compared to other social sciences (Lewin Reference Lewin1996; Rabin Reference Rabbin2002).Footnote 10
VII. CONCLUSION
The contemporary notion of interdependent preferences, conceived in their broad sense, has had a long presence in the history of economic ideas. Our discussion indicated that they were part of the reasoning of many eighteenth- and nineteenth-century authors, such as A. Genovesi, D. Hume, J. Rae, A. Smith, K. Marx, and J.S. Mill, among others. It was also seen that the idea played an important role in the works of T. Veblen and A.C. Pigou. By the turn of the twentieth century, however, V. Pareto’s and A. Marshall’s stances towards interdependent preferences did not favor its further incorporation into mainstream economics. The main reason for this was the gradual establishment of methodological ideas that were clearly against the inclusion of sociological and psychological elements in economics. This was deemed to be essential for the drive of establishing a ‘positive’ economic science. After WWII, the concept of interpersonal preferences regained some momentum, mainly in the work of J. Duesenberry on consumption theory. However, Duesenberry’s approach was also ignored in favor of allegedly more ‘simple’ and ‘sound’ consumption theories.
In the last two decades, there seems to have been a renewed interest in the notion by some labor economists and by theorists in the relatively new subfield of the economics of subjective well-being, mainly because of its research potential for analyzing and explaining numerous contemporary economic issues. Still, however, mainstream economists are not willing to abandon the notion of agents characterized by independent preferences.
Our discussion indicated that there are some basic reasons that might account for the continuous resistance to incorporate interdependent preferences in mainstream economic theory. The traditional refusal to accept the role of psychological and sociological dimensions of economic behavior might be seen as an important reason. This is connected to the long orthodox tradition of economic agents as selfish utility maximizers, which implies independent preferences. Clearly, this is due to the methodological foundations of mainstream economics, which can be traced to the emergence of the marginalist school. The other reason, and probably a very crucial one, was the serious problems that interdependent preferences posed for many well-established theoretical results, as many contemporary mainstream theorists have realized. Thus, our discussion of the history of interdependent preferences indicates that the old methodological mainstream preconceptions concerning human behavior and ‘positive’ economics are still extremely strong.
