Beyond Mechanical Markets

Asset Price Swings, Risk, and the Role of the State

Roman Frydman, Michael D. Goldberg

• Publisher: Princeton University Press
• Year: 2011
• Print ISBN: 9780691145778
• eBook ISBN: 9781400838189
• Topic: Economics
• Subtopic: Finance
• Index: Economics

A powerful challenge to contemporary economics and a new agenda for global finance

In the wake of the global financial crisis that began in 2007, faith in the rationality of markets has lost ground to a new faith in their irrationality. The problem, Roman Frydman and Michael Goldberg argue, is that both the rational and behavioral theories of the market rest on the same fatal assumption—that markets act mechanically and economic change is fully predictable. In Beyond Mechanical Markets, Frydman and Goldberg show how the failure to abandon this assumption hinders our understanding of how markets work, why price swings help allocate capital to worthy companies, and what role government can and can't play.

The financial crisis, Frydman and Goldberg argue, was made more likely, if not inevitable, by contemporary economic theory, yet its core tenets remain unchanged today. In response, the authors show how imperfect knowledge economics, an approach they pioneered, provides a better understanding of markets and the financial crisis. Frydman and Goldberg deliver a withering critique of the widely accepted view that the boom in equity prices that ended in 2007 was a bubble fueled by herd psychology. They argue, instead, that price swings are driven by individuals' ever-imperfect interpretations of the significance of economic fundamentals for future prices and risk. Because swings are at the heart of a dynamic economy, reforms should aim only to curb their excesses.

Showing why we are being dangerously led astray by thinking of markets as predictably rational or irrational, Beyond Mechanical Markets presents a powerful challenge to conventional economic wisdom that we can't afford to ignore.

PART I

The Critique

Indeed, the chief point was already seen by those remarkable anticipators of modern economics, the Spanish schoolmen of the sixteenth century, who emphasized that what they called the mathematical price depended on so many particular circumstances that it could never be known to man but was known only to God.

—Friedrich A. Hayek,
“The Pretence of Knowledge,” Nobel Lecture

1

The Invention of Mechanical Markets

ALTHOUGH THE raison d'être for financial markets implies that they cannot assess asset values perfectly, over the last four decades of the twentieth century, economists developed an approach to macroeconomics and finance that implied that financial markets allocate society's capital almost perfectly. To reach this conclusion, economists constructed probabilistic models that portray an imaginary world in which nonroutine change ceases to be important; indeed, it becomes irrelevant.

An economic theory of the world that starts from the premise that nothing genuinely new ever happens has a particularly simple—and thus attractive—mathematical structure: its models are made up of fully specified mechanical rules that are supposed to capture individual decisionmaking and market outcomes at all times: past, present, and future. As one of the pioneers of contemporary macroeconomics put it, “I prefer to use the term ‘theory'…[as] something that can be put on a computer and run…the construction of a mechanical artificial world populated by interacting robots that economics typically studies” (Lucas, 2002, p. 21).

To portray individuals as robots and markets as machines, contemporary economists must select one overarching rule that relates asset prices and risk to a set of fundamental factors, such as corporate earnings, interest rates, and overall economic activity, in all time periods. Only then can participants' decisionmaking process “be put on a computer and run.” But this portrayal grossly distorts our understanding of financial markets. After all, participants' forecasts drive the movements of prices and risk in these markets, and market participants revise their forecasting strategies at times and in ways that they themselves cannot ascertain in advance.

To be sure, with insightful selection of the causal variables and a bit of luck, a fully predetermined model might adequately describe—according to statistical or other, less stringent, criteria—the past relationship between causal variables and aggregate outcomes in a selected historical period. As time passes, however, market participants eventually revise their forecasting strategies, and the social context changes in ways that cannot be fully foreseen by anyone. The collapse of the hedge fund Long Term Capital Management in 1998, and the failure of ratings agencies to provide adequate risk assessments in the run-up to the financial crisis that began in 2007, shows that models assuming that the future follows mechanically from the past eventually become inadequate. Trading in financial markets cannot, in the end, be reduced to mere financial engineering.

ECONOMISTS' RATIONALITY OR MARKETS?

Ignoring such considerations, contemporary macroeconomic and finance theory developed models of asset prices and risk as if asset markets, and the broader economy, could be adequately portrayed as a fully predetermined mechanical system. And, recognizing that reducing economics and finance to engineering requires some justification, contemporary economists developed a mechanistic notion of rationality that they then claimed provided plausible individual foundations for their mechanical models.

Lucas hypothesized that the predictions produced by an economist's own fully predetermined model of market outcomes adequately characterizes the forecasts of rational market participants. The normative label of rationality has fostered the belief, among economists and noneconomists alike, that this so-called Rational Expectations Hypothesis (REH) really does capture the way reasonable people think about the future.

Of course, hypothesizing that any fully predetermined model can adequately characterize reasonable decisionmaking in markets is fundamentally bogus. Assuming away nonroutine change cannot magically eliminate its importance in real-world markets. Profit-seeking participants simply cannot afford to ignore such change and steadfastly adhere to any overarching forecasting strategy, even if economists refer to it as “rational.”¹

The Soviet experiment in central planning clearly shows that even the vast and brutal powers of the state cannot compel history to follow a fully predetermined path.² Change that cannot be fully foreseen, whether political, economic, institutional, or cultural, is the essence of any society's historical development.³

Such arguments were, however, completely ignored. Once an economist hypothesizes that his model generates an exact account of how an asset's prospects are related to available information about fundamental factors and adopts the Rational Expectations Hypothesis as the basis for “rational” trading decisions, it is only a short step to a model of the “rational market.”

Such a model implies that prices reflect the “true” prospects of the underlying assets nearly perfectly. An economist merely needs to assume that the market is populated solely by this sort of rational individuals, who all have equal access to information when making trading decisions. In the context of such a model, “competition…among the many [rational] intelligent participants [would result in an] efficient market at any point in time.” In such a market, “the actual price of a security will be a good estimate of its…[‘true'] value” (Fama, 1965, p. 56).

Economists and many others thought that the theory of the rational market provided the scientific underpinning for their belief that markets populated by rational individuals set asset prices correctly on average. In fact, this theory is the proverbial castle in the air: it rests on the demonstrably false premises that the future unfolds mechanically from the past, and that market participants believe this as well.

WAS MILTON FRIEDMAN REALLY UNCONCERNED
ABOUT ASSUMPTIONS?

Fully predetermined models presume that nonroutine change can be completely ignored when searching for adequate accounts of outcomes. These models offer an extreme response to the daunting challenge that such change poses for economic analysis. In contrast, relying on largely narrative analysis, Hayek, Knight, Keynes, and their contemporaries focused on the inextricable connection between nonroutine change, imperfect knowledge, and the pursuit of profit in capitalist economies.

As insightful and rich as these narrative accounts were, most contemporary economists probably felt that jettisoning them in favor of the clarity and transparent logic of mathematical models was a move in the right direction. After all, any explanation of the real world, let alone of the highly complex interdependence between individuals and the market, must necessarily abstract radically from its numerous characteristics. Even the detailed narrative accounts of Hayek, Keynes, Knight, and others left out many features of this interdependence.

But to instruct economists to embrace models that are constrained to portray capitalist economies as a world populated by interacting robots was not merely to call for more clarity and transparent logic in economic analysis. Economists were being asked to adopt an approach that went well beyond useful abstraction, because what it left out was actually the essential feature of capitalist economies.

Of course, most economists would readily agree that assuming away the importance of nonroutine change is not realistic. But they nonetheless cling to this core assumption in the belief that it transforms economics into an exact science. They also would agree that the Rational Expectations Hypothesis is not realistic, often describing it as a convenient assumption.⁴ When confronted with criticism that their assumptions are unrealistic, contemporary economists brush it off by invoking the dictum put forth by Milton Friedman (1953, p. 23) in his well-known essay on economic methodology: “theory cannot be tested by the ‘realism' of its assumptions.”

Our criticism, however, is not that the core assumptions of the contemporary approach are unrealistic. Useful scientific models are those that abstract from features of reality. The hope is that the omitted considerations really are relatively unimportant to understanding the phenomenon. The fatal flaw of contemporary economic models is their omission of considerations that play a crucial role in driving the outcomes that they seek to explain.

In fact, at no point did Friedman suggest that economists should not be concerned about the inadequacy of their models' assumptions.⁵ Indeed, at the time that he wrote his essay, examining assumptions was an important aspect of the discourse among economists. Friedman (1953, p. 23) recounted a “strong tendency that we all have to speak of the assumptions of a theory and to compare assumptions of alternative theories. There is too much smoke for there to be no fire”

Consequently, Friedman devoted substantial parts of his essay to an effort aimed at reconciling a “strong tendency” among economists of his time to discuss assumptions with his main conclusion that a theory cannot be tested by the realism of its assumptions. Using a variety of arguments and examples, he reiterated the essential point: an economist's success in devising a model that is likely to predict reasonably well crucially depends on the assumptions that are selected to construct it. As Friedman (1953, p. 26) put it, “The particular assumptions termed ‘crucial' are selected [at least in part] on the grounds of…intuitive plausibility, or capacity to suggest, if only by implication, some of the considerations that are relevant in judging or applying the model.”

The need to exclude many potentially relevant considerations is particularly acute if one aims to account for outcomes with mathematical models, which ipso facto make use of a few assumptions to explain a complex phenomenon. So the bolder an abstraction that one seeks, the more important it is to scrutinize the assumptions that one selects “on the grounds [of their] intuitive plausibility.”

Even more pertinent is Friedman's emphasis on the importance of understanding when the theory applies and when it does not. As he emphasizes throughout his essay, economists should scrutinize whether the assumptions they select are relevant in judging or applying the model.

Viewed from this perspective, the contemporary approach's core assumptions—that nonroutine change and imperfect knowledge on the part of market participants and economists are unimportant for understanding outcomes—strongly “suggest, if only by implication” that models based on them cannot adequately account for asset prices and risk. As we have seen, these assumptions imply that financial markets do not play an essential role in capitalist economies. Moreover, reliance on fully predetermined models and the Rational Expectations Hypothesis to specify decision-making assumes obviously irrational behavior in real-world markets (see Chapters 3 and 4). Friedman would surely consider assumptions that yield such implications unsuitable to serve as the crucial foundation of a theory that purports to account for how profit-seeking individuals make decisions and how modern financial markets set asset prices and allocate capital.⁶

Re-reading Friedman's essay makes it clear why he wanted to alert economists to the dangers of interpreting his methodological position as a lack of concern about the assumptions underpinning economic models. As a superb empirical economist, Friedman understood that basing economic models, as abstract as they must be, on bogus premises is a recipe for predictive failure.

For various reasons, contemporary economists typically do not mention the parts of Friedman's essay in which he acknowledges that careful selection and scrutiny of assumptions is the crucial aspect of building successful scientific models. Instead, they invoke a selective reading of his essay to justify their steadfast refusal to consider that the core assumptions underpinning their approach could be irreparably flawed.

It is not surprising that, as Friedman cautioned, models that place such assumptions at their core repeatedly failed what he considered the ultimate test of a good theory: its ability to provide adequate predictions. Nowhere is this failure more apparent than in asset markets. Indeed, after considering many empirical studies, Maurice Obstfeld and Kenneth Rogoff (1996, p. 625) concluded in their magisterial book on international macroeconomics that “the undeniable difficulties that international economists encounter in empirically explaining nominal exchange rate movements are an embarrassment, but one shared with virtually any other field that attempts to explain asset price data.”

Interestingly, however, this sober assessment—and the hundreds of studies that underpinned it—has not shaken economists' belief in the foundational principles of contemporary macroeconomics and finance. Despite their abject predictive record (indeed, despite the global financial crisis), fully predetermined accounts of economic outcomes are still held up as the only scientific kind, and the Rational Expectations Hypothesis remains the only widely accepted method for portraying how rational individuals forecast these outcomes.⁷

THE POST-CRISIS LIFE OF INTERACTING ROBOTS

One might have thought that the crisis that began in 2007 would have precipitated widespread questioning of the core assumptions of the contemporary approach to economic analysis. On the contrary, nearly all arguments about the causes of the crisis, and about the reforms needed to guard against the recurrence of such a catastrophe, take for granted the relevance of Lucas's conception of macroeconomics as a science of interacting robots whose rationality is based on the Rational Expectations Hypothesis.

Imperfectly Informed Robots

To be sure, insistence on fully predetermined models and the use of the Rational Expectations Hypothesis does not necessarily imply that the market sets asset prices at their supposedly true values—the key claim behind the so-called “Efficient Market Hypothesis”⁸ This implication of the Rational Expectations Hypothesis applies only if one assumes that all market participants have the proper incentives to search for the relevant information and are not somehow denied access to it. To preserve “rational expectations” and yet conclude that the Efficient Market Hypothesis is false, economists exploited the key distinction between information on fundamental variables, which serves as an input to the forecasting process, and the formal and informal knowledge that participants rely on to interpret that information and arrive at their forecasts.

In a seminal contribution to the Rational Expectations approach to the analysis of markets with informational imperfections, Grossman and Stiglitz (1980) pointed out that if all participants understood the supposedly true process driving asset prices, they would not devote the necessary resources to gathering information about the prospects of underlying assets.⁹ Thus, perfectly efficient markets (which, according to the Efficient Market Hypothesis, incorporate all available information into prices) are impossible.

Moreover, for various reasons, participants may not have equal access to information that is relevant to their decisionmaking. Formalizations of this idea in Rational Expectations models with so-called “asymmetric information” produce market prices that substantially deviate from the supposedly “true” values generated by their perfect-information analogs.¹⁰ These conclusions were generally viewed as providing the scientific underpinning needed to interpret important aspects of the crisis as stemming from distorted and asymmetric information, inadequate incentives, and imperfect market competition.

This helps explain why Rational Expectations models with asymmetric information have become ascendant in the public debate about the crisis that began in 2007. After all, woeful lack of transparency, glaring informational asymmetries, and distorted incentives for key market participants played significant roles in bringing the financial system to the brink of collapse.

Undoubtedly, Rational Expectations models with asymmetric information achieved “their aim [in showing] that the standard [Ratio...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Table of Contents
• Acknowledgmentss
• What Went Wrong and What We Can Do about It
• Part I - The Critique
• Part II - An Alternative
• Epilogue
• References
• Index