Defining Economics: The Long Road to Acceptance of the Robbins Definition
Robbins' Essay gave economics a definition that came to dominate the professional literature. This definition laid a foundation that could be seen as justifying both the narrowing of economic theory to the theory of constrained maximization or rational choice and economists' ventures into other social science fields. Though often presented as self-evidently correct, both the definition itself and the developments that it has been used to support were keenly contested. This paper traces the reception, diffusion and contesting of the Robbins definition, arguing that this process took around three decades and that even then there was still significant dissent.
Economics is the science which studies human behaviour as a relationship between ends and scarce means which have alternative uses. (Robbins 1932, p. 15)
The sentence in which Robbins defined economics as dealing with the relationship between ends and scarce means is central to the arguments he made in An Essay on the Nature and Significance of Economic Science (1932). Though Robbins minimized the novelty of his definition, stressing its roots in continental traditions, it had radical implications. It both influenced what economists believed they could and could not say in their role as economists and figured prominently in discussions of the role of theory in economics. The reason the Robbins definition had such radical implications was that, in contrast to previous ‘classificatory’ definitions, such as that economics is the study of the production and use of wealth, or the study of what contributes to economic welfare, his definition was ‘analytical’: it identified an aspect of behaviour (Robbins 1932, p. 16). This had the implication that, insofar as it deals with the influence of scarcity, ‘any kind of human behaviour falls within the scope of Economic Generalisations. … There are no limitations on the subject-matter of Economic Science save this’ (Robbins 1932, p. 16). This laid a foundation that could be seen as justifying not only the narrowing of economic theory to the theory of constrained maximization or rational choice but also the ‘imperialism’ of economists' ventures into the other social sciences.1 Given the importance of these issues, it is not surprising that scholars have paid attention to the origin of the definition (see Howson 2004) and to the methodological conclusions that Robbins drew from it. However, the question of how Robbins' definition was received has been completely neglected. Here, we seek to fill that gap.
Though Robbins' definition is often presented as self-evidently correct, as a depiction of the economic problem faced by either individuals or societies, both the definition and the developments that it has been used to support were keenly contested. This is perhaps not surprising given the irony of the definition having been proposed when the world was at the deepest point of the worst depression ever encountered in the capitalist world. In 1932, it may have seemed counter-intuitive (to put it mildly) to argue that economics involved working out the implications of scarcity, at least at the societal level, where the pressing economic problem was a glut of capital and labour.2 It should, therefore, not be surprising to find that the notion that economists instantly recognized the Robbins definition as an appropriate summary of their discipline is a myth. We argue that it was not until the 1960s that the definition came to be accepted, and even then endorsement of it was far from universal.
Our approach is to focus on explicit discussions of the Robbins definition. We start by analysing the academic journals, considering first the initial reception of the definition and then discussions from the late 1930s to the 1950s, the period when economists began to see themselves as modellers. We next turn to textbooks, on the grounds that this is where economists typically encounter definitions of their subject. Finally, we examine the 1960s and after, the period when economists began to apply their methods to topics traditionally considered to be outside economics. Conclusions are then drawn.
I. THE DEFINITION AND ITS INITIAL RECEPTION
Most discussions of the definition of economics comprise a few paragraphs in an elementary textbook. The Robbins definition was different in that the definition was explained and its implications worked out in an argument that extended over 141 pages (158 in the second edition), drawing on material that Robbins had developed over several years (see Howson 2004). This gave his definition an importance that other definitions did not have. The definition assumed a clear demarcation between ends and means, making it clear that economic science dealt only with the latter; the policy-maker might need to consider ends, but they were not part of economics. Economics was concerned with the implications of scarcity, which had implications for the meaning of economic laws and how they should be derived. The definition therefore provided the basis for a discussion of methodology that extended to three chapters. It also enabled Robbins to examine the bearing of economic science on practice: what was and was not legitimate for economists to say in their role as economists. Thus although others had offered definitions of the subject, and there were already some classic texts on methodology (notably those of John Stuart Mill and John Neville Keynes), no one had attempted to provide such a tightly-argued analysis of the subject: all seemed to follow from this simple, one-sentence definition of economics.
Robbins (1932, p. viii) said that he made ‘no claim whatever to originality’ and that his definition did no more than sum up the way economists thought about their discipline. He claimed that his propositions were based ‘on the actual practice of the best modern works on the subject’ (1932, p. viii). That of course rested on a judgment about what those ‘best modern works’ were, for he was influenced very strongly by, among others, Ludwig von Mises and Philip Wicksteed. In an age when Anglo-American economics was still very strongly influenced by Marshallian economics and, at least in the United States, institutionalism, this was a strong assumption. Shortly after the publication of the Essay, against the background of the Great Depression and then the Second World War, economics went through the upheavals associated with the Keynesian revolution and the rise of the econometric movement. Thus while the book was welcomed, it was controversial, and acceptance of the Robbins definition was much more circumscribed than is commonly believed. This is hardly surprising, for the book came at a time when the subject was in turmoil.
The reviews in the major academic periodicals indicate some of these differences very clearly. It is hardly surprising that Edwin Cannan (1932), in the Economic Journal, should defend his own definition of economics as dealing with the production of wealth—the definition that, in one form or another, had dominated the subject since the late eighteenth century.3 A review in the American Economic Review had to wait until the second edition, when Harvey Peck (1936) observed that Robbins' view of economics (‘the individual exchange variety’) was too narrow and needed supplementing. Interestingly, the strongest praise for the book came from outside economics. George Catlin (1933), in the Political Science Quarterly, described the Essay as a ‘brilliant book’, protesting against empiricism, and he commended it to students of both economics and politics. Interestingly, in view of the ‘economics imperialism’ that came many decades later, his objection to Robbins' definition was that it included politics! To find support for the conventional view that the book contained little that was new, we have to go to an unsigned review in the Journal of the Royal Statistical Society (M 1934),4 where the reviewer explained that Robbins was presenting the view of the Austrian school and that this view was not universally accepted.
Robbins' Essay came early in a decade during which economic methodology was widely discussed. Many of these discussions took up issues raised by Robbins, of which three are relevant here: economic theory versus empirical analysis; how economic theory is to be conceived; and the role of ethics in economics.5 These themes reflect important developments in interwar economics. This was the period when institutionalism and neoclassical economics presented different ways of doing ‘scientific’ economics (see Rutherford 1999). It was also a period when mathematical theory, though still a minority activity, was rapidly developing on both sides of the Atlantic, and economics was on the verge of coming to be seen as social engineering. There was also significant disagreement, going well beyond any disagreement over interpersonal utility comparisons, about the role of ethics in economics.
Robbins was seen by many as defending economic theory against empiricism, leaving little room for the historical and statistical work favoured by the institutionalists (Catlin 1933, p. 463; Fraser 1932). Knight (1934, pp. 237–8, 225) shared this aversion to quantification, but saw Robbins as being too mechanical and argued for a more organic conception of economics. Others argued that Robbins had failed to acknowledge sufficiently the importance of empirical work, as when Hutchison (1935) criticized him for claiming that it was a virtue that the propositions of economic theory must be true: if they were tautologies, theory could have no empirical content and induction was needed.
The charge that Robbins' definition reduced economic theory to something purely formal was expressed most forcefully by Dobb:6
Professor Robbins, who has carried this contemporary fashion to its logical conclusion, explicitly emphasises the purely formal character of economic theory, without, however, seizing the full implications of this statement. Economics, as a theory of equilibrium, he points out, is unconcerned with norms and ends: it is concerned solely with constructing patterns for the appropriate adaptation of scarce means to given purposes. The corollaries of economic theory do not depend on facts or experience of history, but ‘are implicit in our definition of the subject-matter of Economic Science as a whole’. (Dobb 1933, pp. 589–90)
Thus, although some writers saw merit in formalism as defined here (M 1934), others (Souter 1933a, p. 377ff; Janes 1933; Parsons 1934, pp. 536–7) were critical. Knight, as has already been pointed out, considered that it reduced economic theory to something mechanical,7 while Harrod (1938, p. 407) argued that models hid, among other things, how little economists knew about causal sequences, implying that economic theory did not justify the claims to certainty that Robbins made.
Even stronger were the criticisms of Robbins' attempt to exclude ethics from economic science. Fraser (1932, p. 557) argued that Robbins' view implied that rationality was an end in itself. Others (e.g. Spengler 1934, p. 315) pointed out that ethical judgments were needed if policy conclusions were to be drawn, and to claim that conclusions could be derived from pure theory was to smuggle in ethical judgments in a way that was misleading. Harrod (1938, p. 396) challenged economists to accept what common sense told them—that the marginal utility of income to the rich and the poor was different, irrespective of whether such judgments were thought ‘unscientific’. Economics, for Harrod, was not so well developed that it could afford to dispense with such common-sense judgments.
In light of these controversies, it is not surprising that in 1939, Bye could observe that there was still little agreement on the definition of economics. Indeed, some economists went so far as to question the notion that there could or should be a single definition of the subject (see Fraser 1932).
II. MODELLING AND THE ACCEPTANCE OF THE ‘RECEIVED VIEW’
The debates over Robbins' definition in the 1930s took place against a background that was very different from that prevailing after the Second World War. After the war, many more economists thought of themselves as modellers. Though this shift was due in part to the wartime association of economics with operations research and economic engineering,8 there was no sudden transformation. Economics in the ‘old’ style carried on in the postwar period, and this was reflected in discussions of Robbins' Essay in the late 1940s. Joseph Spengler observed in 1948 that there was still no agreement on the subject matter of economics. He cited four definitions: that economics is what economics does;9 Robbins' definition; Fraser's definition in terms of wealth, but seen as consistent with belief in scarcity; and Parsons' definition as concerned with the ramifications of economic rationality (Spengler 1948, pp. 2–3).
The implications of the Robbins definition for pure theory versus empirical analysis were still being debated in the postwar period. Gruchy (1949) argued that Robbins' definition implied pure theory, and contended that though pure theory was needed, economics was broader than this. He found such a broader view of the subject in the Cambridge tradition represented by Marshall and Keynes: in their hands economics was both abstract and humanistic. Keynes had managed to achieve a perspective on economics that came between Robbins' formalism and Veblen's cultural perspective. The methodological narrowing at work here was called in for some particularly strong criticism by Schuller:
This ‘oyster’ view of the field of economic inquiry, in the form of a school centrism, unfortunately is not confined to the [Marxist] Soviet Empire and its representatives abroad. In the western world today there is another school of economic theory which has been so sanctified by traditional acceptance and has remained so impervious to attack by its rivals here (including the Marxists), that it is identified, in the vocabulary of most of our economists, with economic theory as such. The members of this school regard themselves not as a school but as the only practitioners of the economic profession; … they dismiss critics … for being non economists—i.e., ‘sociologists,’‘psychologists,’ or ‘historians.’ (Schuller 1949, p. 440)
The example that he selects is Robbins, who ‘solves problems and enunciates verities not from the viewpoint of himself or his school but “from the point of view of Economic Science”’. This was hardly intended as a complement, as Schuller went on to opine that ‘Such passages by orthodox economists or Marxists often sound like variations on the ancient theme: “It is not I, a human sinner, that addresses you, but God, speaking through the mouth of His prophet”’. The strength of this attack is worth noting, for it suggests that the author felt under threat.
Evidence that the Robbins definition may by this time have been gaining wider acceptance in the journals is found in a piece by Gerhard Tintner (1953), a member of the Cowles Commission, who, in the course of trying to define economics, simply gives two definitions of economics—one from Robbins and the other from Lange (1945, p. 19), whose definition was the ‘science of the administration of scarce resources in human society’.10
The theme that the choice of ends could not be ignored was carried through into the postwar period. Howard Ellis, editor of the AEA Surveys of Economics, saw economics as ‘concerned with the processes and results of free choice on the market’, deliberately defining it in such a way that choice of end fell within the subject (Ellis 1950, p. 3).11 The implication of this line of argument was that economics should be broader than the Robbins definition allowed. Thus McConnell (1955, pp. 160–1) argued that economics is about finding patterns but also about evaluating rules, institutional and ideological contexts, and the desirability of changes. As such, it includes not just ‘theoretical or analytical economics’ but also welfare economics and other branches that took it beyond what Robbins considered economic science.12
This period also saw a significant shift towards concern with prediction, testability and choice of assumptions. This change was hastened by Paul Samuelson's Foundations of Economic Analysis (1947) and became even more marked after Friedman's (1953) forceful advocacy of prediction as the ultimate test of theory.13 Samuelson's focus was on the importance of deriving ‘operationally meaningful theorems’. Not only did he not define economics (though the reader might infer that it related to maximization), he argued that ‘logically there is nothing fundamental about the traditional boundaries of economic science’ (Samuelson 1947, p. 9). His argument ran in terms of what should be taken as exogenous or endogenous; government, for example, might be in either category depending on the needs of the problem in hand. Gruchy (1949, p. 249) portrayed Samuelson as adopting the Robbins view. Others, however, saw a clear difference between them. Papandreou (1950), adopting Parsonian terminology, argued that Foundations showed that the ‘action’ framework could be used in a different way: where Robbins ended up with pure theory, Samuelson used the framework to derive empirical propositions. Robbins had confined himself to the ‘back room’ of pure theory, whereas Samuelson was started on the process of integrating the ‘front room’ with concepts from the back room. Such a view agreed with Samuelson's own remark that Robbins' claims for deductive theory were ‘exaggerated’ (Samuelson 1964, p. 736).
The way discussions of Robbins had changed by the 1950s is best illustrated in Machlup's (1955) article on verificationism. Robbins was listed as an advocate of ‘extreme a priorism’, this being contrasted with Terence Hutchison's ‘ultra-empiricism’: the former regarded economic theories as certain, whereas the latter required that assumptions as well as implications be tested. The assumptions of a theory should be understandable, but induction was useful in giving confidence in hypotheses. At the same time, it was going too far to claim that all assumptions needed to be verified, for theories and models were analytical devices used for generating testable predictions. What Machlup was doing was combining an Austrian element with what had become the so-called ‘Received View’ in philosophy of science, centred on the hypothetico-deductive model. It offered a defensible position between Robbins and his empirical critics that many economists found attractive. Broadly interpreted, it could encompass both Friedman's (1953) methodology and that of his one-time rivals at Chicago, the Cowles Commission, exemplified by Koopmans (1957).
By the early 1960s, though economists might still question whether the Robbins definition was adequate, it had come to be widely accepted. The general tone was to refer to it as being accepted and not to argue the case for it. Thus Johnson (1960, p. 552) started a very sympathetic review article on J. K. Galbraith's The Affluent Society (1958) with the concession that most economists ‘would probably accept’ the Robbins definition, ‘at least as a description of their workaday activities’.14 Its resonance with modern mathematical methods was noted by John Hicks (1960, p. 707) in his review of linear theory, where he argued that these techniques provide the tools necessary to cover economics as defined by Robbins—implying, without stating so explicitly, that this definition is the appropriate one to use. At the opposite pole, we even find institutionalist William Kapp (1968, p. 2) allowing that, in his estimation, Robbins' definition ‘characterizes very well the prevailing preoccupations of many economists’. Finally, when Albert Rees penned his entry on ‘economics’ for the International Encyclopedia of the Social Sciences (Rees 1968, p. 472), he gave a definition that was pure Robbins, calling it ‘widely accepted’. And so, a definition that very few economists had been prepared to endorse unequivocally in print had apparently come to be generally accepted.
Robbins may have denied that his ideas were novel, but there was little trace of them in the leading Principles texts of the day. The oldest, and most influential, was Marshall's Principles of Economics, renowned for defining economics simply as ‘a study of mankind in the ordinary business of life’ (Marshall 1920, p. 1).15 The next oldest, by R. T. Ely, the first President of the AEA, assisted by a series of co-authors, of whom the most prominent was Allyn Young of Harvard, offered a more traditional definition: ‘economics is the science which treats of those social phenomena that are due to the wealth-getting and wealth-using activities of Man’ (Ely et al. 1926). Sumner Schlichter offered an institutionalist definition: ‘The subject matter of economics is industry … [studied as] a complex of human practices and relationships’ (Schlichter 1931, p. 11). Other textbooks offered no definition, either providing an illustrative list of economic problems (Garver and Hansen 1928) or arguing that none was required (Taussig 1927). The closest thing to Robbins in the English-language textbook literature of the time seems to be the definition offered by Fairchild et al. (1926), who, having identified ‘the insatiability of man and the niggardliness of nature’ as ‘the foundation stones upon which rests the structure of economics’ (p. 8), define economics as ‘the science of man's activities devoted to obtaining the material means for the satisfaction of his wants’ (p. 8). Some of these texts were revised after Robbins' Essay appeared, but none adopted his definition.16
If the older generation cannot be expected to have taken to Robbins with enthusiasm, the same cannot be said of those writing new textbooks after 1932. Frederic Benham, a fellow student and later a colleague of Robbins at the LSE, defined the subject implicitly in his first edition, saying that ‘the rationale of economic activity is to satisfy human wants by producing consumers' goods’ (Benham 1938, p. 5). In his third edition, he came even closer to Robbins, referring to economic decisions as choices that involve opportunity costs (Benham 1943, p. 5), but he neither reproduced the Robbins definition nor cited the Essay. Alec Cairncross (1944, p. 8) explicitly adopted a definition in terms of scarcity, but qualified it by saying that economics was concerned only with the social aspect of scarcity: when one person's decisions impinge on other people. Furthermore, his wording—‘Economics is a social science studying how people attempt to accommodate scarcity to their wants and how these attempts interact through exchange’—restricts the subject to an exchange economy. We see a similar pairing of scarcity and exchange from Nevin (1958, p. 6), while Thomas (1952) reverted to the materialist definition in terms of wealth. The Robbins definition clearly influenced this generation, but it was heavily qualified.
If anything, North American textbooks remained further from Robbins in this period. In The Economic Organization, Frank Knight (1933, p. 4) says that economics ‘deals with the social organization of economic activity’, lately via the price system or under free enterprise. Knight found both the traditional definitions, such as Marshall's and the choice-based ones, over-broad and, as a result, ‘useless and misleading’ (Knight 1933, p. 2).17 In sharp contrast to the view that later became associated with the Chicago school, Knight insisted that ‘economizing … does not include all human interests’ and that it is both ‘error’ and ‘vice’ to ‘look upon life too exclusively under this aspect of scientific rationality’ (Knight 1933, p. 2).18Kenneth Boulding's (1941, p. 3) approach to the issue was not unlike Knight's: he told his readers that, in terms of subject matter, defining economics ‘as the study of human valuation and choice’ is probably ‘too wide’.19 Perhaps reflecting the sort of frustration hinted at by Knight and Boulding, Tarshis (1947) explained that a formal definition is not appropriate and instead listed economic problems.
There was a move towards a Robbinsian perspective in the most influential text to appear in the late 1940s, Samuelson'sEconomics (1948), in that Samuelson argued that economics deals with the relationship between means and ends, and ends are not part of the ‘science’. He also used the argument that the ‘American way of life’ requires more resources than are available (Samuelson 1948, pp. 16–17) to counter the claim that the existence of unemployed resources showed that resources were scarce. But he did not cite the Robbins definition and his focus ‘What? How? For Whom?’ resonated equally with more traditional definitions in terms of the production and distribution of wealth.
Some other texts explicitly used scarcity definitions. Most prominent was George Stigler's The Theory of Competitive Price, which was unusual in actually citing Robbins' Essay (Stigler 1942, pp. 13, 20). Stigler defined economics as ‘the study of the principles governing the allocation of scarce means among competing ends when the objective of the allocation is to maximize the attainment of the ends’ (Stigler 1942, p. 12). Stigler was not alone here. Gemmill and Blodgett (1937, p. 22) defined economics as ‘the social science that describes man's efforts to satisfy his wants by utilizing the scarce means provided by nature’, and Bowman and Bach (1946, p. 3) cited ‘economizing’, arising from scarcity, as the ‘central problem’ of economics.
Scarcity definitions of economics became more prominent in both countries from the late 1950s onward, though in the UK it was not uncommon for authors to broaden this to emphasize the social character of the subject and relate it specifically to exchange. Given the way Becker was at this time extending the boundaries of economic reasoning (not to mention later feminist critiques of economics), it is significant to note the example Nevin (1967, pp. 5–7) gave of the hazards of moving beyond the exchange context: ‘But who can assess the value of a mother's services to her family in the home? The economist, at least, is too sensible to try.’20
In the USA, on the other hand, economics came to be defined in terms of scarcity, with fewer qualifications21—to the point where Bronfenbrenner could state already in 1962 that: ‘Most of the current crop of textbook definitions of the subject stress’ the scarcity hypothesis (p. 266). McConnell, in a textbook that went through four editions during the decade, offered a definition that is worth citing in full:
Recalling that wants are unlimited and resources are scarce, economics can be defined as the social science concerned with the problem of using or administering scarce resources (the means of producing) so as to attain the greatest or maximum fulfillment of society's unlimited wants (the goal of producing). (McConnell 1969, p. 23)
Though it is tempting to ascribe this to the influence of the Robbins view, it is worth noting the emphasis on social science and the use of the word ‘administering’, echoing Lange's definition cited above.
The two most influential texts of the 1960s and 1970s were Samuelson's Economics and Richard Lipsey's An Introduction to Positive Economics. Samuelson, whose Economics continued to go through new editions, remained the most important, not least because of its use internationally. By the fifth edition (1961), the comment on the Great Depression not being a counter-example to the prevalence of scarcity was gone, and there was now a section on the ‘Law’ of scarcity. And when we come to the tenth edition (1976), written with Peter Temin, Samuelson is offering what can only be described as an expanded version of the Robbins definition, which comes after five alternatives have been considered:
Economics is the study of how people and society end up choosing, with or without the use of money, to employ scarce productive resources that could have alternative uses, to produce various commodities and distribute them for consumption, now or in the future, among various persons and groups in society. It analyzes the costs and benefits of improving patterns of resource allocation. (Samuelson and Temin 1976, p. 3)
Much of the expansion is purely expository, not changing the force of the definition at all (with or without the use of money, now or in the future), though the last sentence begs the question of how, if ends are not part of economics, one distinguishes between those changes in patterns of resource allocation that are ‘improvements’ and those that are not. If economics is broader than ‘economic science’, then this is not explicit.
The main new textbook, certainly judged in terms of sales, to emerge in the 1960s was Lipsey'sAn Introduction to Positive Economics (1963). Lipsey is an important figure because, though he came out of the LSE, he was part of a group of young economists committed to replacing the Robbinsian emphasis on deductive theory with an economics based on measurement and testing. This was the meaning of the ‘positive economics’ in his title.22 In the first three editions, he defined economics not with a single definition but by listing six economic questions, the closest he came to the Robbins definition being to refer to ‘one of the basic problems encountered in most aspects of economics, the problem of SCARCITY’ (Lipsey 1971 (3rd edition), p. 50, italics added).23 Though the use of capital letters implies that scarcity was a fundamental concept that the student must understand, the words ‘one of’ and ‘most’ make it clear that it does not encompass the whole of economics. The view here was that economics cannot be deduced, as Robbins had claimed, from one basic postulate, but has to be empirical. In the fourth edition (1975), this rejection of Robbins was even more explicit. After discussing six economic questions, Lipsey went on:
Economics today is regarded much more broadly than it was even half a century ago. Earlier definitions stressed the alternative and competing use of resources. Such definitions focused on choices between alternative points on a stationary production-possibility boundary. Important additional problems concern failure to achieve the boundary … and the outward movement of the boundary over time. (Lipsey 1975, p. 59)
Lipsey distanced himself from Robbins (without naming him) by associating him with the past, and implicitly criticizing other textbook writers who were using the Robbins definition. Unemployment and growth are signalled as important economic questions from which an exclusive focus on scarcity deflects attention.
Subsequent textbooks, too numerous to survey, offer variations on the themes encountered up to now. Scarcity is clearly considered central to economics, though the emphasis placed on it has varied;24 it is not always clear whether this is because authors see that it is limited in scope, or whether it is to make what might be a very abstract definition of the subject more digestible to newcomers to the subject. The Robbins definition appears to be in the background but to have been blended with other ideas. Significantly, it did not become universally accepted, Lipsey's textbook being the clearest example. Even today, Krugman and Wells (2004, p. 2) define economics as ‘the study of economies’, which is decidedly un-Robbinsian.
IV. ECONOMICS SPREADS ITS WINGS
By the 1960s, the Robbins definition was becoming established in the textbooks, but there was little discussion of it in the journal literature, suggesting either that it had become accepted, or that it was simply something that was irrelevant to practising economists. What changed in the 1960s and 1970s was a progressive expansion of the boundaries of the field—an expansion that was at once consistent with Robbins' definition and yet reflected a view of the discipline that was likely far beyond anything Robbins might have imagined in 1932.25 As Susan Howson (2004, p. 417) has shown, Robbins was concerned to distinguish economics from the other social sciences—to show where the boundary between them properly lay. More recent work, in contrast, has been focused on showing that such a boundary does not exist. When Milton Friedman wrote in 1962 that ‘An economic problem exists whenever scarce means are used to satisfy alternative ends’,26 he went on to note that this is ‘a very general’ conception, one that ‘goes beyond matters obviously thought of as belonging to economics’. But such is the breadth of economics today that his example of the allocation of leisure time seems rather quaint and leads one to wonder why there would ever be any fuss over excessive generality (Friedman 1962, p. 6).27 We also find Harry Johnson—like Friedman, a professor at Chicago—stating in his LSE inaugural lecture that he considered ‘scarcity’ and ‘choice’ to be the two basic conceptual ideas underlying economics (Johnson 1968, pp. 3–4)—invoking Robbins in the process—and going on to argue that economics had become more useful because of the new work being done within this framework in areas including human capital theory, the economic analysis of time, the economics of information, and the economic theory of democratic political processes.28 And already in 1968, Kenneth Boulding was referring in his AEA Presidential Address to ‘the attempt on the part of economics to take over all the other social sciences’, a movement that he labelled ‘economics imperialism’ (Boulding 1969, p. 8).29
While several scholars contributed to these early forays into areas previously considered non-economics, the process was led by Gary Becker, who in the late 1950s and 1960s had begun to advance the case that economic theory could be used to understand phenomena such as discrimination, irrational behaviour, human capital, and issues in crime and punishment. Becker did not feel compelled to justify these early excursions by appealing to any sort of definition. In fact, in his classic article on crime and punishment, Becker suggested that economists may have avoided discussing illegal activity, not because it did not fall within the boundaries of economics, but because it was ‘too immoral to merit any systematic scientific attention’ (Becker 1968, p. 170, n. 1).
When Becker defined economics as ‘the study of the allocation of scarce means to satisfy competing ends’ in his Economic Theory (1971), he immediately remarked on the breadth of this definition, noting that: ‘It includes the choice of a car, a marriage mate, and a religion; the allocation of scarce resources within a family; and political discussions about how much to spend on education or on fighting a Vietnam war’ (Becker 1971, p. 1). He later pointed out that this definition was so general that economists often found it an ‘embarrassment’, needing to be qualified ‘to exclude most nonmarket behavior’ (Becker 1976, p. 4).30 Becker acknowledged that economists tend to study the market sector, but he argued that economic principles developed for this purpose were ‘essential’ for understanding much of what had been studied by sociologists, anthropologists and other social scientists (Becker 1971, p. 2).
Becker suggested that Viner's ‘what economists do’ definition was emblematic of the problems with defining ‘a subject matter that has changed so much over time’, and, as if conceding that his definition did not seem to define a field called economics, argued that any definition of a discipline is inadequate. What is more important, he said, is ‘the economic approach’ to human behaviour:
The combined assumptions of maximizing behavior, market equilibrium, and stable preferences, used relentlessly and unflinchingly, form the heart of the economic approach as I see it. (Becker 1976, p. 5)
Although this applied to all the social sciences, it could be called the economic approach because it was economists, not other social scientists, who had adopted it (something Becker wished to change).
Interestingly, not everyone associated with the expansion of the boundaries of economics was favourably disposed toward the Robbins definition. Both James Buchanan and Ronald Coase protested against defining the field apart from its subject matter, with Coase (1977, p. 487) suggesting that the field involves the study of ‘the social institutions that bind together the economic system’,31 and Buchanan (1964, p. 220) preferring ‘the study of the whole system of exchange relationships’. Buchanan thought this subject of enough import to devote his 1963 Presidential Address to the Southern Economic Association to it, and he brought the Robbins definition in for strong criticism, calling it ‘all too pervasive’, and suggesting that it ‘served to retard … scientific progress’ (Buchanan 1964, p. 214).32 Robbins' definition, he said, made economics about ‘a problem or set of problems’ rather than ‘a characteristic form of human activity’ (Buchanan 1964, p. 214). While accepting that the Robbins definition had become standard in the profession, Buchanan argued that economics should be conceived of as the study of markets, not of resource allocation, and thus that the Robbins definition should be replaced by one that emphasized market exchange, or catallactics (Buchanan 1964, p. 214).33
Buchanan was by no means the only prominent economist to argue the case for a market exchange definition. Boulding, who came from a very different position on the political spectrum,34 offered a definition very similar to Buchanan's in his AEA Presidential Address:
Economics specializes in the study of that part of the total social system which is organized through exchange and which deals with exchangables. This to my mind is a better definition of economics than those which define it as relating to scarcity or allocation, for the allocation of scarce resources is a universal problem which applies to political decisions and political structures through coercion, threat, and even to love and community, just as it does to exchange. (Boulding 1969, p. 4)35
Unlike Boulding, of course, Buchanan (1964, p. 220) favoured the extension of economic reasoning to political structures, and in this sense did, like Becker, have an ‘outward-looking’ streak. But Buchanan did not see the Robbins definition as necessary for the expansion of economics to other fields; the exchange definition could do the same, since, for Buchanan, politics is, at its heart, a collective exchange process.
Thus, while the Robbins definition may have been widely accepted, there was still significant dissent.
Tracing the influence of a definition of economics is problematic because, most of the time, economists have little reason to cite any definition of what they are doing. There is the further problem that, though it was by no stretch of the imagination a generally accepted definition of the subject before Robbins' Essay, the idea that economics is about scarcity was, as Robbins pointed out, far from novel. Even if an economist's acceptance of the idea beneath the definition could be demonstrated (and this is often difficult), it typically would be hard to establish the definition's role in his work. This problem is exacerbated by the fact that in the main place where one would expect to find discussions of the definition of economics—in introductory textbooks—sources are rarely cited. Had the initial reaction been one of ‘Robbins has brilliantly expressed what we all know to be the essence of economics, but had never been able to articulate’, and had textbooks suddenly started adopting analytical definitions in terms of scarcity and choice between alternative uses of scarce resources soon after 1932, the circumstantial evidence would be strong indeed. But it was not like that at all. Robbins' definition of economics was challenged from the start. In the journals it was frequently attacked, and it was hardly ever accepted without qualification. Then around 1960, economists started to refer to the definition as generally accepted. What had happened was that it had gradually come to be accepted in the textbooks, though, even there, there was hardly unanimity.
If one thinks of knowledge being created in the journals and then finding its way into textbooks, this pattern is strange. It arises because one can do research in economics without being concerned with how the subject is defined, which means that it is generally not discussed in journal articles. Evidence that economists were starting to accept the Robbins definition first appears in the textbooks because that is where the definition of the subject is generally discussed. However, this does not mean that it did not have significant consequences for research. The grounds on which economists objected to the Robbins definition were that it served to narrow economics, or to constrain its methods in ways that were thought unjustifiable. This was precisely because it was an analytical definition, which appeared to define a specific way to set about doing the subject. Because of this, it simultaneously both narrowed and broadened the scope of the subject. It narrowed it through suggesting that deduction could achieve more than many economists believed it could, and it broadened it through freeing economists from being constrained to analyse a particular subject matter.36 Until around the 1970s, economists attached great importance to the latter, and many of the qualifications that were attached to the Robbins definition served effectively to confine an analytical definition to what was considered the subject matter of economics. From the 1970s, as economic methods came to be applied to social and other traditionally non-economics problems, economists became less concerned about this and the qualifications were widely dropped.
There may, however, have been another factor at work. This was that, though he did not formulate it in these terms, Robbins' definition fitted well with acceptance of a rational choice model of behaviour. As economists learned to apply rational choice to an ever widening range of problems, the Robbins definition came to be used more prominently in textbooks.37 There remained division over whether economics was defined by a method or a subject matter, but both sides in that debate could increasingly accept some version of the Robbins definition.
The authors would like to thank Gary Becker, William Baumol, David Colander, Philippe Fontaine, D. Wade Hands, Susan Howson, Fabio Masini, the editors, an anonymous referee, and the participants in the many seminars at which it has been presented for helpful comments on earlier drafts of this paper.
1. This wording is chosen to avoid commitment on whether maximizing behaviour and rational choice are or are not synonymous, or whether Robbins' definition was in fact used to justify such approaches. The latter is, in part, the subject of this paper; the former is not relevant for our current concerns.
2. But see the discussion of Samuelson's Economics, as well as note 14 below. This also raises the issue of who is the ‘agent’ facing scarcity. In introducing his definition, Robbins (1932, pp. 12–16) focused entirely on individuals, yet elsewhere (1935, p. 155) he could write about decisions facing society under conditions of scarcity. As we shall see, definitions of economics that are used in the literature after Robbins refer sometimes to society and sometimes to individuals. In the period we are considering, views on the relationship between the individual and the society were far from homogeneous and changed significantly.
3. Cannan (1914). For a survey of the evolution of the definition of economics, see Backhouse and Medema (2009).
4. It has been suggested that the reviewer was H. W. Macrosty.
5. Book-length treatments in the years following the publication of Robbins' Essay include Souter (1933b), MacFie (1936), Beveridge (1937), Fraser (1937), Wooton (1938) and Hutchison (1938). See also the articles by Fraser (1932, 1938), Souter (1933a), Spengler (1934), Knight (1934), Parsons (1934), Hutchison (1935), Machlup (1936), Leontief (1937), Ayres (1938), Durbin (1938), Harrod (1938) and Bye (1939). We leave open the extent to which this literature was a response to Robbins, though all of this did lead Robbins (1938) to seek to disentangle what he called the ‘Live and dead issues in the methodology of economics’. Note that there is a significant literature on such problems before the Essay. On the other hand, Machlup (1936, p. 39), for one, sees at least parts of this literature as a response by economists who, whatever their attitude towards laissez-faire in economic policy, want laissez-faire to apply to their own practices, and object to the ‘regimentation’ of their subject offered by Robbins.
6. Dobb is quoting from Robbins (1932, p. 75).
7. His criticism of mechanical theories came out even more forcefully later, in his review of Hutchison (Knight 1940).
8. See Backhouse (2002, ch. 11) and Morgan (2003).
9. A twist on Viner, who reputedly defined economics by saying ‘economics is what economists do’.
10. Lange's article is notable for the absence of any citations, including to Robbins. Rothbard (1957, p. 314) is the only economist outside the Cowles Commission who supported Robbins unconditionally. He welcomed him as a fellow praxeologist.
11. See also Streeten (1950), Lachmann (1951) and Shahan (1952).
12. Hicks (1941, p. 111) had also argued that because economics involved evaluating alternative institutions, welfare economics must fall within the boundaries of the subject.
13. Friedman focused exclusively on the meaning of ‘positive’ and not of ‘economics’.
14. Despite his view that scarcity permeated economic theory, Johnson went on to review Galbraith's argument that society's problems concerned its opposite—opulence—with considerable sympathy. This opulence argument was thought by some to be a legitimate critique of the Robbins definition, or to imply that it was at the very least outmoded. (See also Beckerman 1956.) Yet, as Johnson, Bronfenbrenner (1962, p. 265) and others pointed out, the issue is not the size of the physical stock of goods at some time or place, but rather the amount of the good that would be demanded at a zero price. By this definition, the problem of scarcity had been no more overcome by societal affluence than it was absent during the mass unemployment during the Great Depression.
15. Marshall was extensively cited in the literature discussed in this paper. That is true of no other textbook writer.
16. For example, Ely's text (now Ely and Hess 1937)—interesting as one of very few to provide a bibliography—dropped Mill, Cairnes and John Neville Keynes from its bibliography, and brought in Robbins' Essay. However, it offered exactly the same definition as in Fairchild et al. (1926). By the fourth edition of their Economic Problems of Modern Life, Patterson and Scholz (1948, p. 3) had expanded their ‘study of business' definition to also include ‘the social study of wealth and welfare’ and ‘the study of pecuniary values’ (supply and demand), but still had got no closer to Robbins.
17. ‘Many definitions of economics found in text books fall into this error of including virtually all intelligent behavior. One writer has actually given as his definition of economics the “science of rational activity.” Others find its subject matter is “man's activity in making a living,” or “the ordinary business of life.” Such definitions come too near to saying that economics is the science of things generally, of everything that men are for practical reasons interested in. Such a definition is useless and misleading’ (Knight 1933, pp. 1–2).
18. Knight goes on to say that ‘Life must be more than economics, or rational conduct, or the intelligent accurate manipulation of materials and use of power in achieving results. Such a view is too narrow’ (Knight 1933, p. 2).
19. Boulding also considered Marshall's definition over-broad, but ‘the study of material wealth’ and ‘the study of that part of human activity subject to the measuring rod of money’ (Pigou's definition) too narrow (Boulding 1941, p. 3). He carried these characterizations through subsequent editions of his book—see, for example, Boulding (1955).
20. We leave it to others to unpack the gendered stereotypes involved. Note that though Becker's work on the family came later, his analyses of discrimination, education and the allocation of time had appeared by this time.
21. See, for example, Snider (1962), Keiser (1965), Fels (1966), Leftwich (1969) and McConnell (1969).
22. Cf. Lipsey (2008).
23. The page number is significant. Unlike all the other texts, discussion of economics was preceded by a lengthy but non-technical explanation of the principles of modelling, including statistical methods and the philosophy of science. The latter was Popperian rather than the Received View cited by Machlup (see above).
24. It was not as though older definitions had been eradicated, however. Peach (1965, p. 15), for example, ascribes his definition, that ‘Economics is the study of, and embraces all knowledge relevant to, the production of goods and services’, to Keynes. Yet he cannot resist pointing out a few pages later that ‘Economics has long been connected in people's minds with the notion of scarcity’ (p. 21). And Hailstones, writing in 1968, defines economics as ‘a science that is concerned with the production, distribution, and consumption of goods and services’ (1968, p. 2).
25. Robbins returned to the Essay in his 1980 address to the American Economic Association, but he did not comment on the expanding boundaries of economics, though he did reiterate his views on the appropriateness of his definition of the subject (Robbins 1981, pp. 2, 9).
26. This definition is repeated verbatim in his 1976 revision of the book.
27. Friedman was providing this definition in his lectures at least as far back as January 1947; see Johnson (2008).
28. Johnson was referring to the work of economists such as T. W. Schultz, Gary Becker, George Stigler and Anthony Downs.
29. This was not a new phrase, having been used by Souter already in the 1930s: ‘The salvation of Economic Science in the twentieth century lies in an enlightened and democratic “economic imperialism”, which invades the territories of its neighbors, not to enslave them or to swallow them up, but to aid and enrich them and promote their autonomous growth in the very process of aiding and enriching itself’ (Souter 1933b, p. 94n).
30. In his chapter ‘The economic approach to human behavior’, Becker (1976, p. 3) mentioned three extant definitions of economics: ‘the study of (1) the allocation of material goods to satisfy material wants, (2) the market sector, and (3) the allocation of scarce means to satisfy competing ends’, noting that the last of these is the most general.
31. Coase included here firms, input and output markets, the banking system, etc.
32. Buchanan also suggested that this definition did not accurately describe the state of the field in 1930: ‘Only since The Nature and Significance of Economic Science have economists so exclusively devoted their energies to the problems raised by scarcity, broadly considered, and to the necessity for the making of allocative decisions’ (Buchanan 1964, p. 214).
33. ‘In so far as individuals exchange, trade, as freely-contracting units, the predominant characteristic of their behavior is “economic”’ (Buchanan 1964, p. 220).
34. Both, though, had seriously engaged with political science.
35. Peck (1936) actually said that Robbins' definition was exchange-based; see above. Kirzner (1965, p. 258) argued that Buchanan's definition can be subsumed under Robbins' definition, because the exchange process is a part of the struggle to deal with the problem of scarcity. One sees this reflected in the Cairncross (1944) and Nevin (1958) definitions, cited above.
36. In Backhouse and Medema (forthcoming) we argue that this narrowing contributed to the rise of axiomatic methods in economics, even though Robbins did not endorse such methods.
37. Support for this can be found in Stigler's essay on economic imperialism, where he suggested that the increasing abstraction of economics, particularly via the widespread adoption of ‘the machine of maximizing behavior’, brought on the ‘imperialistic age’ of economics (Stigler 1984, p. 312). Coase (1977) argued along similar lines, although he, unlike Stigler, was very critical of and pessimistic about these forays into other disciplines.