Critics of conventional macroeconomic models -- critics of the Rational Expectations and Efficient Markets Hypotheses in particular -- are often accused of simply rehashing old complaints (e.g. see the second paragraph of The Economist article mentioned below). This paper from the INET Conference attempts to depart from “familiar complaints” about REH and EMH. The paper argues that RE requires a mechanistic model of expectations that cannot possibly capture "the diversity of 'reasons' upon which individuals in real-world contexts might base their decisions," and then it offers an alternative approach to modeling expectations:
Life After "Rational Expectations"?, by Roman Frydman and Michael D. Goldberg [presentation slides]: Many people regard the recent financial crisis as a painful addition to an already massive body of evidence that demonstrates the inadequacy of today’s economic models of “rational” markets. ...
But very few have interpreted the inability of “rational” market models to account for such swings as a potentially decisive indication that economists’ approach to modeling rational decision-making is irreparably flawed. The debate triggered by the crisis, summarized by The Economist in two articles addressing “[w]hat went wrong with economics [a]nd how the discipline should change to avoid the mistakes of the past,” has largely overlooked the key problem: the impossibility of establishing a standard approach to modeling how a rational individual makes decisions in every situation.
Precisely the presumption that economists’ have found such a standard has come to underpin models of rational decision-making in a wide variety of contexts – diverse economies, markets, and even fields of inquiry, such as political science and law. In order to arrive at such a universal approach, economists’ standard of rationality must abstract as much as possible from differences in individuals’ interpretations of the social context, including the process driving market outcomes, history, norms and conventions, and public policies and institutions. For the last three decades, the vast majority of economists, including those following the behavioral approach, have considered the “Rational Expectations Hypothesis” (REH) to be the cornerstone of this standard.
In this paper we sketch the emergence of REH and how it evolved to become the centerpiece of contemporary macroeconomics and finance. We focus on major arguments advanced by the promoters of the hypothesis that seemed to have contributed to its rapid and broad acceptance. We argue that REH models are fundamentally flawed on epistemological and empirical grounds and thus cannot serve as a foundation for thinking about markets and public policy.
Consequently, we urge economists to jettison REH. We have recently proposed an alternative approach, called Imperfect Knowledge Economics (IKE)... In contrast to contemporary models, IKE recognizes the inherent limits to economists’ knowledge, as well as the imperfection of knowledge on the part of market participants and policy officials. ...[P]sychological findings, as well as observations concerning the context within which participants make decisions – including historical market outcomes, past policies, norms, and conventions – play a key role in formalizing the foundations of IKE models.
Rationality and the Social Context
...For economists, “a decision-maker is rational if [she] makes decisions consistently in pursuit of [her] own objectives” (Myerson, 1991, p. 2). Economists typically suppose that individuals are motivated by self-interest, and thus that in making decisions they attempt to maximize their own well-being. Because selfishness is widely considered to be an innate trait, using self-interest to stand for decision-makers’ objectives is compatible with economists’ belief that their approach to rational decision-making is universally applicable.
The problem is that what constitutes self-interested decision-making depends on the context within which it occurs. ... As Sen (1993, p. 501) has argued,
“[S]uppose the person faces a choice at a dinner table between having the last remaining apple in the fruit basket (option B) or leaving the apple for someone else to take and forgoing the opportunity of eating the nice-looking apple (option A). She decides to behave decently and picks nothing (option A), rather than one apple (option B). If, instead, the basket had contained two apples, and she had encountered the choice between having nothing (A), having one nice apple (B), [or having two nice apples] (option C), she could reasonably enough choose one (B), without violating any rule of good behavior.
In checking whether these choices are internally consistent, economists would consider them on their own, without any reference to an individual’s values or the context within which she makes decisions. To be sure, if her sense of decency or some other reason were not behind her apparent preference for A in the first case and for B in the second, such choices would undeniably be inconsistent on purely logical grounds. But, as Sen (1993, p. 501) emphasizes, although this combination of choices would violate the standard consistency conditions, “[t]he presence of another apple (C) makes one of the two apples decently choosable. [T]here is nothing particularly ‘inconsistent’ in this pair of choices (given her values and scruples).” Indeed, as this example shows, “There is no such thing as internal consistency of choice” (Sen, 1993, p. 499).
Of course, were an economist to take into account the fact that, beyond the anticipated pleasure from eating an apple, the individual is also concerned about the “decency” of her choice, he could have rationalized her choices as being consistent after all. Consequently, once we enlarge the set of factors that an individual considers important for her well-being to include “decency,” her choices can be seen to be consistent with self-interest, broadly understood.
Inventing Rational Expectations
In simple decision problems, as in Sen’s apple example, the individual is fairly confident about the consequences of her choices for her own well-being. In such cases, an economist might be able to use his own understanding of the individual’s social context to interpret whether she makes consistent choices. However, matters are much more complicated in financial markets, in which the consequences of individuals’ choices lie in the future.
Here, an economist must not only model the preferences with which an individual ranks alternative options – for example, in terms of their returns and riskiness – but also how she forecasts these outcomes. Even assuming that all individuals are motivated by self interest, there is simply no general procedure or model that would pick the precise set of reasons – and the forecasting strategy based on those reasons – that would adequately capture how a generic rational individual thinks about the future. But an economist requires precisely such a standard of rational forecasting in order to formulate a set of conditions that would enable him (or anyone else, for that matter) to interpret whether in most situations an individual makes decisions “consistently in pursuit of [her] own objectives.” ... Consequently, in modeling “rational” decision-making, economists largely ignore the diversity of “reasons” upon which individuals in real-world contexts might base their decisions.
Behavioral economists and non-academic commentators often criticize economists’ standard of rationality as implying that rational individuals make decisions as if they had superhuman abilities to understand the future – that they can compute correctly the consequences for their well-being of alternative options available to them. The raison d’être of behavioral economics has been that most people lack these abilities which supposedly explains why they do not make decisions consistent with economists’ standard of rationality.
But REH presumes no such thing. ... Instead, REH supposes that individuals adhere steadfastly to a single mechanical forecasting strategy. Indeed, economists’ characterizations of rational forecasting would appear to any reasonable person, let alone a profit-seeking participant in financial markets, to be obviously irrational.
After all, a profit-seeking individual understands that the world around her will change in non-routine ways. She simply cannot afford to believe that she has found an overarching forecasting strategy and, thus, she will look for new ways to forecast, which cannot be fully foreseen. Thus, REH imagines a place devoid of the crucial features that characterize, even in the most rudimentary abstract way, how individuals forecast in real-world markets.
Economists, particularly macroeconomists and finance theorists, often view REH as a “useful abstraction.” But the necessity to abstract, intrinsic to all science, cannot render coherent, let alone justify, imputing to individuals demonstrably unreasonable beliefs and then claiming that these individuals are rational.
Unsurprisingly, REH models turned out to be woefully inadequate as characterizations of how profit-seeking individuals make decisions, particularly when it comes to selecting and revising forecasting strategies. In real-world markets – in which there is diversity and an ever-present possibility that non-routine change will alter the process driving market outcomes – REH models have no connection to what participants would consider “logical and reasonable.” ...