BEHAVIORAL ECONOMICS To appear in Elsevier’s Handbook of the Philosophy of Science, Vol. 5 Methods: Theory. Simple models of psychological phenomena Erik Angner, University of Alabama at Birmingham [email protected] George Loewenstein, Carnegie Mellon University [email protected] Date: 20 November 2006 Word Count: 15622 TABLE OF CONTENTS 1 INTRODUCTION.......................................................................................................................................................1 2 THE INTELLECTUAL BACKDROP......................................................................................................................2 2.1 Classical and early neoclassical economics.....................................................................................................4 2.2 Postwar neoclassical theory..............................................................................................................................8 2.3 Discussion........................................................................................................................................................13 3 PSYCHOLOGICAL APPROACHES DURING THE LATE NEOCLASSICAL PERIOD...............................15 3.1 The institutionalists: Veblen, Mitchell and Clark.........................................................................................16 3.2 The macroeconomists: Fisher and Keynes....................................................................................................17 3.3 The outlier: Scitovsky.....................................................................................................................................19 3.4 The “old” behavioral economics: Simon and Katona...................................................................................20 3.5 Discussion........................................................................................................................................................24 4 THE “NEW” BEHAVIORAL ECONOMICS........................................................................................................26 4.1 Behavioral decision research..........................................................................................................................27 4.2 Tversky and Kahneman’s biases and heuristics, prospect theory................................................................30 4.3 Thaler’s anomalies..........................................................................................................................................32 4.4 Examples of current research.........................................................................................................................33 4.5 Discussion........................................................................................................................................................36 5 THE METHODS OF BEHAVIORAL ECONOMICS...........................................................................................37 5.1 Hypothetical choices.......................................................................................................................................39 5.2 Experiments with actual outcomes.................................................................................................................41 5.3 Field research...................................................................................................................................................43 5.4 Process measures, including fMRI.................................................................................................................46 5.5 Discussion........................................................................................................................................................47 6 CURRENT DIRECTIONS.......................................................................................................................................48 6.1 Neuroeconomics..............................................................................................................................................48 6.2 Affect...............................................................................................................................................................50 6.3 Light paternalism.............................................................................................................................................52 6.4 Behavioral economic imperialism..................................................................................................................54 6.5 Discussion........................................................................................................................................................55 7 CONCLUSION.........................................................................................................................................................56 BIBLIOGRAPHY.........................................................................................ERROR! BOOKMARK NOT DEFINED. ii 1 INTRODUCTION In recent years, behavioral economics has emerged as a bona fide subdiscipline of economics.i Because behavioral economics in certain ways represents a sharp departure from mainstream – that is, neoclassical – economics, it raises a number of questions of a philosophical, methodological and historical nature. Yet, to date, it has not received the attention it deserves from historians and philosophers of science.ii In this chapter, we take some initial steps to address this deficiency. Our purpose is to shed light on (a) the nature and historical origins of behavioral economics as a field, (b) its main results and their interpretation, (c) the methods used by its practitioners, (d) its relationship to traditional economics as well as to other emerging subdisciplines such as neuroeconomics, and (e) some of its philosophical and methodological underpinnings. We make no claim, of course, to settle or even raise all issues raised by the emergence of behavioral economics; we do want to go some way toward figuring out what those issues are. The term “behavioral economics” was in use as early as 1958 (cf. Johnson 1958; Boulding [1958] 1961, 21). These days, as it is typically employed, “behavioral economics” refers to the attempt to increase the explanatory and predictive power of economic theory by providing it with more psychologically plausible foundations (Camerer and Loewenstein 2003, 3; cf. Weber and Dawes 2005, 91). Notice that behavioral economics so defined has little to do with behaviorism; in fact the historical roots of behavioral economics can be traced to cognitive psychology, which emerged in direct opposition to behaviorism (see section 2). The modifier “behavioral” – which is sometimes criticized for being redundant on the grounds that all economics is or should be about behavior – stems from the origins of behavioral economics in 1 behavioral decision research (see section 4.1). Behavioral economists do not deny that there may be much to learn from sociology, anthropology and other neighboring fields. However, most of the work characterized as behavioral economics these days – and virtually all the work reviewed here – is, in fact, inspired by psychology. A separate subfield that draws on sociology, and which is sometimes referred to as “socioeconomics,” has coalesced around a different set of researchers and journals. Our main thesis is that behavioral economics should be seen as a branch of cognitive science. Thus, we agree with Russell Sage Foundation president Eric Wanner, who has helped fund research in behavioral economics since the mid-1980s, and who has been instrumental in the establishment of behavioral economics as an independent subdiscipline. Wanner describes behavioral economics as an application of cognitive science to the realm of economic decision- making. “The field is misnamed – it should have been called cognitive economics,” he says. “We weren’t brave enough” (quoted in Lambert 2006: 52, italics in original). Drawing on Howard Gardner’s work on the origins and nature of cognitive science – especially his classic The Mind’s New Science: A history of the cognitive revolution ([1985] 1987)iii – we will argue that seeing behavioral economics as a branch of cognitive science is eminently useful for understanding both the historical origins, nature, strengths and weaknesses of behavioral economics. 2 THE INTELLECTUAL BACKDROP When cognitive science emerged in the 1940s and 50s, it did so in opposition to behaviorism and associated doctrines, including logical positivism and verificationism (Gardner 1987, 10 ff). In Gardner’ words, scientists of this era had developed “a growing awareness ... that adherence to behaviorist canons was making a scientific study of the mind impossible” (Gardner 1987, 12). 2 This was so in large part because behaviorism and associated doctrines “eschewed entities (like concepts and ideas) that could not be readily observed and reliably measured” (Gardner 1987, 15). Here, we will argue that something very similar is true for behavioral economics. Behavioral economics emerged in opposition to neoclassical economics, which was heavily inspired by behaviorism and associated doctrines, including verificationism and operationalism. In particular, behavioral economics emerged in reaction to the notion, held by many neoclassical economists, that social and behavioral science should avoid reference to entities (like cognitive and affective states) that cannot be directly observed. Our examination of the origins and development of neoclassical economics serves two main purposes. First, because behavioral economics largely emerged in reaction to neoclassical economics, a historical excursion allows us to paint a fuller picture of the views against which behavioral economics reacted. Second, because most of the critics of behavioral economics have a neoclassical background, it allows us to achieve a better understanding of their criticism. In passing, this section is also intended to illustrate that the project to rid economics of its ties to psychology (a project described in section 2.2) is relatively modern; in fact, we will argue, both classical and early neoclassical economists were deeply interested in the psychological underpinnings of economic behavior. Our exposition largely follows that of Michael Mandler (1999), who divided the history of modern economics into three main periods: classical, early neoclassical, and postwar neoclassical (Mandler 1999, 3). Like all divisions of this sort, Mandler’s is imperfect – for one thing, authors representative of e.g. postwar theory may have published before the war – but for present purposes it is good enough. 3 2.1 Classical and early neoclassical economics Before the emergence of behaviorism during the first decades of the twentieth century, psychologists were largely comfortable with talking about to mental states and other unobservables. “At the turn of the century,” Gardner writes, “investigators had been addressing the key issues of mental life: thinking, problem solving, the nature of consciousness” (Gardner 1987, 11). Similarly, as we will see, classical and early neoclassical economists made frequent reference to cognitive and affective states. Their conception of human nature – and therefore human decision making – was often relatively sophisticated, and in many cases inspired by developments in psychology. These facts are important, because there are many misconceptions about the views of human nature implicit (or explicit) in the writings especially of the classical economists of the 18th century (including Adam Smith). Perhaps because they are interpreted in light of modern neoclassical economics, these economists are often falsely attributed a particularly simple psychology, according to which people consistently, everywhere and always, pursue their own self-interest narrowly construed. The actual views of the classical economists could hardly be more different. Regarding the nature of human ends, for example, Smith wrote: “How selfish soever man may be supposed, there are evidently some principles in his nature, which interest him in the fortune of others, and render their happiness necessary to him, though he derives nothing from it except the pleasure of seeing it” (Smith [1759] 2002, 11). Similarly, regarding people’s rationality, or utility-promoting behavior, Smith wrote: How many people ruin themselves by laying out money on trinkets of frivolous utility? What pleases these lovers of toys is not so much the utility, as the aptness of the machines which are fitted to promote it. All their pockets are stuffed with little conveniences … of which the whole utility is certainly not worth the fatigue of bearing the burden (Smith [1759] 2002, 211). 4 Whatever the exact implications of these quotes, they show rather clearly that Smith did not have the simple-minded view of human nature that some would have it. Smith does not, of course, have a theory of decision in the modern sense, but he did express a vision of human nature (and therefore human action) which is remarkably multi- faceted. As Mark Perlman and Charles McCann (1998) describe it: Smith’s homo economicus ... was a man with a temporal sense, a man with loyalties, a man who clearly understood that he was part of a larger social collective. What Smith’s man wanted and needed was the responsibility for making his own decisions and accepting the consequences of those decisions. This responsibility had to be understood as existing in concert with the twin principles of self-love and sympathy, for all were combined in the Smithian calculus. In brief, in modern parlance what was to be maximized by Smith’s man was the right of self-determination, while still allowing a place for both moral and social sensibilities and even expressions of altruism (Perlman and McCann 1998, 239). The exact details of Smith’s conception of human nature are contested (cf. Otteson 2002; Schliesser 2005). It should be clear, however, that Smith – like his contemporary David Hume – was deeply interested in the psychological underpinnings of human behavior. Moreover, arguably, Smith's views about human psychology were not incidental to his more purely economic work, and may have had an important impact on them (cf. Davis 2003, 270). Reading the classical economists’ philosophical and economic psychology, several contemporary authors have gone so far as suggesting that Hume and Smith in fact identified and discussed some of the phenomena that now occupy behavioral economists. Thus, Ignacio Palacios-Huerta argues that both Hume and Smith analyzed dynamically inconsistent behavior, and that “their analyses of this behavior remain novel” (Palacios-Huerta 2003, 243). Similarly, Nava Ashraf, Colin F. Camerer and George Loewenstein (2005, 140) argue that Smith’s work “is not only packed with insights that presage developments in contemporary behavioral economics, 5 but also with promising leads that have yet to be pursued.”.iv These insights include phenomena now referred to as loss aversion, overconfidence, social preferences, and more. Early neoclassical economics is best characterized by the work of William Stanley Jevons. These economists, including Jevons, explicitly built their economics on the foundation of hedonic psychology, that is, an account of individual behavior according to which individuals seek to maximize pleasure and minimize pain. In Jevons’ words: “Pleasure and pain are undoubtedly the ultimate objects of the Calculus of Economics. To satisfy our wants to the utmost with the least effort ... in other words, to maximize pleasure, is the problem of Economics” (Jevons [1871] 1965, 37). The early neoclassical economists were inspired by Bentham, who wrote: “Nature has placed mankind under the governance of two sovereign masters, pain and pleasure.... They govern us in all we do, in all we say, in all we think” (Bentham [1823] 1996, 11, italics in original). These economists understood utility in terms of conscious experience like pleasure or happiness. As Jevons puts it: “Utility [arising from any commodity] must be considered as measured by, or even as actually identical with, the addition made to a person’s happiness” (Jevons [1871] 1965, 45). When it came to welfare economics, early neoclassical economists were unabashed utilitarians. A. C. Pigou, author of The Economics of Welfare ([1920] 1952) and commonly considered the father of welfare economics, went to great lengths exploring and measuring “total welfare” ([1920] 1952). Early neoclassical economists like Pigou believed that welfare or utility could meaningfully be aggregated across individuals, and that one state was superior to another if total welfare was greater in the former than in the latter (Mandler 1999, 4). Of course, welfare economists of the time shared the focus on conscious experience. As Pigou put it, “the elements of welfare are states of consciousness and, perhaps, their relations” (Pigou [1920] 1952, 10). 6 Mandler argues that the hedonic foundations of economics – and especially the assumption that people maximize pleasure – conferred multiple advantages. First, hedonics came with an account of deliberation, according to which individuals weigh the pleasure and pain that would result from various actions and choose the one they perceive as leading to the greatest balance of pleasure over pain (Mandler 1999, 76). Second, hedonics provided a rationale for several critical assumptions, such as the completeness and transitivity of the preference relation, and (given the further assumptions of separability and diminishing marginal utility) the convexity of indifference curves (Mandler 1999, 76-77). Third, “the early neoclassical account of rational deliberation allowed for a rich description of irrational (‘incorrect’) behavior” (Mandler 1999, 77). Hedonic psychology permits people to act irrationally because, for example, they fail to properly anticipate the pleasure resulting from certain actions, or because (in the intertemporal context) they fail to properly take future pleasure into account in their deliberations (see, e.g., Loewenstein, O'Donoghue & Rabin, 2003). In sum, the assumption that people maximize pleasure could explain both why preferences in general are transitive, etc., and why people sometimes act irrationally. The identification of utility with conscious experience had important methodological implications. Because it was assumed that individuals have direct access to their conscious experience, many economists defended the principles of hedonic psychology on the basis of their introspective self-evidence alone. Thus, John E. Cairnes wrote: “The economist starts with a knowledge of ultimate causes. He is already, at the outset of his enterprise, in the position which the physicist only attains after ages of laborious research” (Cairnes [1888] 1965, 87, italics in original). The reason, Cairnes continues, is that “we have, or may have if we choose to turn our attention to the subject, direct knowledge of these causes in our consciousness of what passes in 7 our own minds” (Cairnes 1965, 88). Because of their commitment to introspection, in conjunction with the belief that introspection supported the principles of hedonic psychology, early neoclassical economists like Cairnes saw little reason to explore alternative methods to confirm the adequacy of the foundations of their economics. The heavy reliance on introspection was not unique for the economists. As Gardner put it: “Unfortunately the scientific method favored by most researchers at the time was introspection” (Gardner 1987, 11). 2.2 Postwar neoclassical theory The emergence of behaviorism – marked by the appearance of John B. Watson’s article ‘Psychology as the Behaviorist Views it’ (1913) – included an attack on both the heavy reliance on introspection and the references to mental states. In Gardner’s words: The behaviorists put forth two related propositions. First of all, those researchers interested in a science of behavior ought to restrict themselves strictly to public methods of observations, which any scientist could apply and quantify... Second, those interested in a science of behavior ought to focus exclusively on behavior: researchers ought assiduously to eschew such topics as mind, thinking, or imagination and such concepts as plans, desires, or intentions (Gardner 1987, 11, italics in original). The two propositions are clearly present in the writings of the postwar neoclassical economists as well. The transition from early to postwar neoclassical theory, although inspired by earlier work (e.g. Pareto [1906] 1971; see Bruni & Sugden, 2007), took place over the course of some 20 years, from the mid-1930’s to the mid-50’s (Mandler 1999, 8). As we will see, postwar neoclassical economists wanted to gain distance from psychology of all kinds, objected to the notion that economics should make reference to conscious states, and rejected the idea that introspection was a scientifically acceptable means to explore such states. 8