American Scientist

The Effortless Economy of Science? Philip Mirowski. viii + 463 pp. Duke University Press, 2004. $89.95 cloth, $24.95 paper.

The Effortless Economy of Science? is a compilation of Philip Mirowski's essays on economic methodology and its application to understanding the organization and output of scientific activity. Mirowski, who is Koch Professor of Economics and the History of Science at Notre Dame, is prominent within the community of "heterodox" economists, people who have challenged much of contemporary economic theory and empirical practice. He links his criticisms to the study of science by arguing that neoclassical economic reasoning—which views individuals as making purposeful decisions, based on well-specified preferences, constraints and beliefs—fails to provide a way of understanding how the enterprise of science functions. This failure in turn means that these economic models cannot address major problems facing science in modern society.

Mirowski wades into disputes among philosophers and sociologists over the extent to which the scientific enterprise leads science to evolve in ways that are consistent with such "transcendental goals" as superior explanatory power and the like. It is not surprising that he is highly disdainful of efforts to argue that the scientific research enterprise can claim a privileged status as producer of truth.

Mirowski's essays reflect an enormous breadth of knowledge and a willingness to address a wide spectrum of questions, ranging from the foundations of scientific inquiry to issues at the forefront of public policy. Unfortunately, his ambition in scope does not make up for his generally sloppy argumentation or for his frequent use of polemics instead of reasoned analysis.

Indeed, Mirowski seems incapable of engaging in debate without calling names or trying to portray prominent figures as foolish. And these characterizations are themselves silly; any author who refers to Kenneth Arrow as "politically pugnacious" is operating in a thought-space far removed from those who know the man. Frankly, Mirowski's visceral contempt for those he regards as intellectual opponents makes it hard to take him seriously.

Beneath this harsh rhetoric there is remarkably little substance. The deficiencies of Mirowski's work are exemplified in his discussion of attempts to apply chaos theory to economics. He suggests that a major reason for the failure of such attempts is that "the neoclassical orthodoxy had a vested interest in neutralizing chaotic descriptions of the economy." No clear explanation is given as to why this is so, although he seems to think that the presence of chaos in economic data would invalidate the rationality assumptions that are often used in economic theory. As a matter of economic theory, such a claim would be false.

Further, the idea that neoclassical orthodoxy (whatever that is supposed to be) would bring about a suppression of research critical of rationality has already been refuted. Mirowski makes no mention of the fact that a new area of research, behavioral economics, is based on challenging and relaxing these rationality assumptions. Economic orthodoxy has hardly rejected this work: Research on behavioral economics earned psychologist Daniel Kahneman a Nobel prize and Matthew Rabin the John Bates Clark Award (for the best American economist under age 40).

The actual reasons for the failure of chaos theory to add much are twofold: First, there is nothing in the body of economic theory that would lead an economist to expect to find chaos in economic data. Mirowski in fact notes that this lack of theoretical foundation differs from the situation in physics, but he fails to draw the obvious conclusion. Second, those economists who attempted to identify chaos in economic data generally came to the conclusion that they could not find strong evidence for it. By contrast, behavioral economics is based on ample experimental evidence from psychology and economics.

Mirowski recognizes that the conclusion of many economists that chaos is not present in economic data is a problem. He resorts to attacking the main researchers who have made this point, saying that they "have so misrepresented the question, that their crusade to eradicate chaos from the economy has done the discipline of economics a real disservice." This claim is supported by a critical discussion of the properties of the tests various authors have used to look for chaos. Mirowski's observations about the statistical analysis of chaos are incorrect and show that he misunderstands the nature of the identification problem in econometrics. But what concerns me more is his misrepresentation of the motivation of the authors. Persuasive empirical evidence of the presence of chaos could have catapulted the discoverers to a Nobel prize; call me old-fashioned, but the acknowledgment by investigators that their search failed strikes me as evidence of intellectual integrity.

Mirowski's criticisms of economic models of science are likewise unpersuasive and in many cases amount to disapproval of the metaphor of a marketplace of ideas. An underlying theme in his attacks is that economic models for understanding science are being used to justify claims that science has a special place as a producer of truth. The problem with his criticisms is that they are ahistorical (ironically, as that is what he accuses neoclassical economics of being). It is trivial to see how a marketplace of ideas can fail to weed out inferior ideas in favor of better ones. The examples in the Soviet Union of Lysenkoism in biology and Marrism in linguistics are sufficient evidence that a society can set incentives in such a way as to promote bad science. However, the relevant question is whether, across many societies over many decades, with the incentives of scholarly prestige interacting with commercial and government demand for the technological offshoots of science, the marketplace metaphor is of use. I believe that it is, and nothing in these essays represents a serious objection to this general view.

Mirowski's criticisms further suffer from the absence of theoretical context. Often he attacks the assumptions underlying an economic model in order to dismiss it as a metaphor for understanding some phenomenon, but the force of such arguments is entirely context-dependent. This is particularly pronounced in his extended discussion of philosopher of science Philip Kitcher, who has been in the forefront of work to reconcile the progressive nature of the scientific enterprise with realistic modeling of the behavior of scientists. In the weakest essay in this volume, Mirowski faults Kitcher for, among other things, using a "representative actor" model (a framework that eliminates certain types of heterogeneity when describing individual actors) and for failing to account for "neoclassical economics as reprocessed nineteenth century physics" when using economic tools. The first objection is vacuous, since there is no reason to think that any of Kitcher's conclusions hinge critically on this modeling assumption. The second, which refers to Mirowski's most famous research on the history of economics, is simply irrelevant. Wherever Kitcher's tools came from, the question is whether they elucidate the problem at hand. Similar problems exist with Mirowski's other criticisms of Kitcher.

The heterodox ideas found here and elsewhere have had no impact on economics as a whole, just as the body of science studies research has had no effect on the natural sciences. Mirowski would attribute this to the "vested interest" of neoclassical economics. An alternative explanation is that research programs such as those found here—which fail to provide either new empirical insights or criticisms of existing practice that are intellectually compelling, let alone constructive ways to proceed—do not have enough substance to warrant a claim on intellectual resources. In my judgment, in this instance the marketplace of ideas is working efficiently.—Steven N. Durlauf, Economics, University of Wisconsin–Madison