This is the working paper version of “Review: Selected Essays by Frank H. Knight: Volumes 1 and 2,” published in Review of Political Economy 14(2), April 2002, pp.269–275. Please cite the published version.
Knight, Frank H.,
“What Is Truth” in Economics?
Selected Essays by Frank H. Knight: Volume 1.
Laissez-Fair: Pro and Con.
Selected Essays by Frank H. Knight: Volume 2.
Edited by Ross B. Emmett.
(Chicago, University of Chicago Press, 1999),
pp.406+xxiv, $58.00 hardcover, ISBN 0-226-44695-6 (v.1)
pp.406+xxiv, $58.00 hardcover, ISBN 0-226-44697-2 (v.2)
These two volumes contain an excellent collection of twenty-nine essays, carefully chosen to give a cross-section of Knight's interests and contributions. Knight's writing is thoughtful and careful, while being at the same time wide ranging and wonderfully engaging. The essays are offered roughly in the order they were published, rather than being grouped by theme. This is convenient for the purposes of reference, and since there are so many recurring themes in Knight's writings, it presents few difficulties for one who simply reads through the volumes.
For those of us who are not historians of thought, the prospect of wading through two sizable volumes raises a natural question: What do Frank Knight's writings have to offer contemporary economists? Given Knight's lack of currency, we may be surprised by the answer: much indeed. Anyone interested in the methodological issues involved in characterizing and developing an economic science should read these books.
Most economists associate Knight with a single important book: Risk, Uncertainty and Profit (1921). Few have a sense of his role a as founder of the interwar Chicago School of economics. We tend to have a stronger sense of the views of students of Knight--- such as Patinkin, Friedman, Stigler, and Buchanan---than of Knight himself, despite their open acknowledgement of strong intellectual debts to Knight. These two volumes offer us a glimpse of the range of interests and the insightful criticisms that characterized this unique, important economist.
Both volumes emphasize essays with large methodological and ethical themes, with volume 1 running more toward methodology and volume 2 running more toward ethics. The essays on economic methodology emphasize the limited scope of economic science, the reasons for methodological dualism, and the relationship between economics and ethics. I recommend beginning with the last essay in the first volume---the eponymous “ `What Is Truth' in Economics?” (1946 JPE)---which is both a serious but accessible meditation on the methodology of social science and an hilarious send-up of T.W. Hutchison's The Significance and Basic Postulates of Economic Theory. (“It is perhaps the chief merit of the work that … the author ends up by virtually abandoning the `criteria' on which at first he lays so much emphasis.”) This taste of Knight's style and concern will provide useful context for the other papers.
Knight does not suffer fools gladly, and his targets range from the inconsistencies and hidden agenda of churchmen to the silly excesses of fellow liberals. Much of the writing in these volumes is critical, and most of Knight's critiques have aged very well. He reminds us to temper our social science ambitions and to confront the limitations of our profession. Knight shows these limitations to be extremely constricting, yet the severity of his criticism is tempered by its friendliness. This is true in both his methodological and ethical writing, which this review will emphasize, and in his explorations of economic theory, which we will now touch upon.
Most readers can expect to be irritated by Knight's quick dismissal of arguments in their favorite tradition. For example, his quick dismissal of Austrian capital theory in “Capitalistic Production, Time, and the Rate of Return” (1933) and his laconic yet incisive recitation of the reasons why marginalism completely supercedes classical economics in “The Ricardian Theory of Production and Distribution” (1935) may gall some economists, and his attacks on the pretensions of mainstream theory and methods will irritate many more. Consider his impatience with those who try to dispense with marginal utility, especially the diminishing marginal utility of income, as expressed in “Realism and Relevance in the Theory of Demand” (1944) or in “Laissez-Faire: Pro and Con” (1967). In “Historical and Theoretical Issues in the Problem of Modern Capitalism” (1928), Knight trains his sights on another target: “In the popular mind, and of course especially in the Marxian and most other socialistic literature, it is viewed as an axiom that the owner of `capital,' or `the means to labor,' has the worker in his power. But the proposition is false or meaningless if employers are in competition with each other and act in accordance with economic motives. … The essential thing, on the organization side, is not at all … the dominance of property over propertyless labor. The subjection of workers to owners is certainly less extreme now than it was in the precapitalistic age, due to the fact of impersonal competition.” Even if Knight moves a bit too quickly through such arguments, they are not simply empty rhetoric, and it is an important context for his arguments that Knight grants capitalism no ethical primacy over other economic systems. One of the wonderful things about Knight is that he never hesitates to prod any school of economic thought precisely at its most vulnerable spots.
“The Quantity of Capital and the Rate of Interest” (1936 JPE) is a fascinating paper which, in the guise of capital theory, shows Knight wrestling with some basic difficulties in moving from static theory to growth theory. Knight wishes to accomplish several things in this paper: clarify the nature of capital and its role in a capitalist economy by uncovering difficulties in the views of Ricardo, Jevons, and Böhm-Bawerk, show the irrelevance of the stationary state to a actual capitalist economies, to privilege production possibilities over consumption trade-offs in the determination of interest rates, and to dismiss the loanable funds theory of interest rate determination and replace it with a theory that emphasizes asset stocks rather than flows. It is an interesting aspect of the article that it emphasizes the urge to accumulate, perhaps indicating a Marxian influence: “the creation and maintenance of a supply of capital in a society depends upon men having or acting as if they had interests which look beyond their own lives”. (Indeed, in “The Ideal of Freedom” (1943), Knight argues that this urge to accumulate leads over generations to increasing inequalities of wealth and power, which exposes the greatest contradictions within our ideals of freedom and equality. He reiterates this in “The Sickness of Liberal Society” (1947), and in “Laissez-Faire: Pro and Con” (1967) he excoriates Hayek for ignoring it.)
In contrast with Keynes's liquidity preference view, Knight wishes to emphasize the stock of capital rather than the stock of money as determinative of the rate of interest. Indeed, one of the interesting and puzzling things about Knight's view of interest rates, seen through Keynesian eyes, is that Knight manages to discuss interest rate determination without reference to money markets. In his 1936 paper “Unemployment: And Mr. Keynes's Revolution in Economic Theory,” Knight considers Keynes's view of interest rate determination, but not with great friendliness. “The theory of interest is the most difficult part of the whole construction to take seriously.” (p.356)
The skepticism with which Knight greets the liquidity preference theory is typical of his reception of Keynes's General Theory, which he considers merely to suggest “modifications of conventional equilibrium analysis to account for temporary, possibly more or less chronic, disequilibrium conditions” (p.346). However one reacts to Knight's barrage of detailed, perceptive criticisms of Keynes's work---by finding it utterly damning, or by finding Knight to cut down dozens of trees without damaging the Keynesian forest---there is one criticism that must receive serious attention. Knight takes himself to be trained in and to work in the “Classical” tradition that Keynes purports to supplant, and Knight claims to find little that is recognizable in Keynes's portrait of the Classics. He finds that Keynes has characterized the whole of “classical economics” by a particularly simple subset of the classical analysis, more sophisticated versions of which he believes to contain most of what is valuable in Keynes's work. The title of Knight's essay is therefore ironic.
Knight clearly loves and respects economics, but his core interests are in ethics and economic methodology. These interests are well represented by the first three essays of the first volume: “The Limitations of Scientific Method in Economics” (1924), “Ethics and Economic Interpretation” (1922), and “The Ethics of Competition” (1923). (The second and third essay were published in the Quarterly Journal of Economics. I try to imagine such articles appearing in the QJE today, and my imagination fails me.) The first essay attempts to characterize science, to draw distinctions between the requirements of science and the limitations faced by those who study human behavior, and to show that there is nevertheless a science of economics. The second and third essays focus more on the limitations of economic science and the inability of economics to provide ethical justifications for any particular economic system. These themes pervade Knight's writings.
Knight quickly characterizes science as the formulation of empirical laws suitable for making predictions about events we wish to influence, and just as quickly he explains why human behavior resists such formulations. Like Veblen, Knight recognizes that science attempts to discard teleological explanation, which is however necessary for the understanding of human behavior. Knight argues (against the behaviorist program) that we are unable to understand human behavior without reference to values, and furthermore we are unable to characterize the values and perceptions in individual choice situations in ways that yield useful predictions about the specific actions individuals will take. One severe problem lies in the difficulties of objective measurement of the constituents of any choice situation, and a possibly more severe problem is that individuals bring their entire history to bear on each choice situation (so that choice exhibits radical path dependence).
Knight sets a very modest goal for economic science: to improve upon the performance of common sense. He is impatient with attempts to render economics more “scientific” by simply adopting obfuscating vocabularies. His pointed critiques of behaviorism, discussed below, remain relevant to modern discussions of revealed preference. His criticisms in “Statics and Dynamics: Some Queries regarding the Mechanical Analogy in Economics” (1930) are as perceptive as any modern deconstruction of the economic use of terms such as `statics', `dynamics', `equilibrium', `inertia', or `friction'.
Given Knight's criticisms and explicitly limited ambition, his conclusions about the cognitive status of economic theory may seem startling. “In spite of all the foregoing, there is a science of economics, a true, and even exact, science, which reaches laws as universal as those of mathematics and mechanics.” The reason will appeal to Austrian economists: Knight holds that while economics cannot make predictions about the content of economic behavior, it can make precise predictions about its abstract form.
Knight is kind enough to offer specific examples. Unfortunately, if we hope for startling revelations, we will be disappointed at the unveiling. Indeed, it is unclear that his examples go beyond common sense. Here are some of them. A person will want more of a good rather than less, up to a point. Wealth is desired. Goods have diminishing (marginal) utility, so consumer expenditures will comprise multiple goods. More production requires more inputs. These examples have for Knight the character of “axioms”, but he also offers a more complex example: he proposes that the conditions of supply and demand are largely responsible for the price of goods. It seems that we are not getting much past common sense here, and Knight agrees. Early on he comments that economics “can tell us little in the concrete, and its chief function is negative---to offset as far as possible the stupid theorizing of the man in the street.”
In contrast with many classical liberals, Knight does not presume the ethical superiority of laissez-faire. In part, this is because he puts great stress on the endogeneity of tastes. Individuals not only produce in an economic system but are produced by it. This makes comparison of economic systems very difficult, for they must be assessed not only on how well they satisfy the desires expressed within the system but also by the desires that the system produces. His recognition of this crucial point allows him to situate economic actors socially. For example, “[t]he competitive economic order must be partly responsible for making emulation and rivalry the outstanding quality in the character of the Western peoples who have adopted and developed it.” This allows him to offer a much more sophisticated characterization of the motivations of individuals in consumption and production than those offered by mainstream economics. Satisfactions are socially produced, and do not simply inhere in objects of consumption or technical job characteristics. This also forms one of Knight's fundamental critiques of economic science, for the things that mainstream economics treats as data (e.g., tastes) are endogenously produced. Against the notion that there is no point in disputing about tastes, Knight argues that we dispute about tastes constantly and that the development of “good” values is a fundamental human goal. Since a competitive system engenders both good and bad values, a general defense against an absolute standard is impossible. The primary defense of the competitive system, such as it either has been or can be realized, is other systems pose a danger of “vastly worse.” For Knight does assert, like Karl Marx and Friedrich Hayek, that capitalism has at least engendered a remarkable material progress.
In this context, it might be surprising that Knight launches such vigorous defenses of the competitive process. For example, “Some Fallacies in the Interpretation of Social Costs” (1924), attacks two cases for government intervention: Pigou's example of road congestion externalities, and Graham's example of a decreasing cost industry in the context of international trade. In both cases Knight is determined to show that a fully competitive process in a private ownership economy yields the efficient outcome.
Another of Knight's concerns in these essays is to dismiss the behaviorist impulses that seduced economists of his time and continue to seduce economists today. Behaviorism attracts economists who want desperately to render their studies more scientific but find that they keep referring to the unobservable desires and beliefs of individuals. At some point the idea occurs to anyone who thinks about this that perhaps one could base economics on observables outcomes rather than on hidden processes of the psyche. Knight saw clearly that this is an impossible hope. For example, in “Facts and Metaphysics in Economic Psychology” (1925), he observes that “we do know that men have interests, and that they cannot be left out of account without vitiating the observation.” Knight places great stress on a methodological contradiction in behaviorism: scientific observation is inherently intersubjective, so a behaviorist can dismiss the mental in his subjects of study only to insist upon it when publicizing his research. However, while Knight is right about this contradiction, I doubt this criticism will prove persuasive. People are used to adopting contradictory methods in different domains of action. I feel Knight would have pressed his case more firmly by elaborating on the previous point: it is impossible to make observations of economic behavior without acknowledging the subjective interpretations of the actors. To take a famous example, it is not useful to describe monetary exchange in terms of the movements of colored paper. Fifteen years later, in “ `What is Truth' in Economics?”, he makes this point explicitly with a very similar example: the example of a check system.
Although Knight in some sense disavows pragmatism (which he seems to associate with a Jamesian will-to-believe), a remarkably modern pragmatism pervades Knight's writings. For example, in “Fact and Metaphysics in Economic Psychology” (1925), Knight comments that, “The principles of logic themselves rest upon common sense, are derived from it by a process of abstraction and generalization, and have to be abandoned as soon as their deliveries come to run counter to common sense itself, that is, as soon as normal minds agree in repudiating them.” In 1925 Knight is saying, “Truth is established by consensus as much as beauty.” Fifteen year later, in “God, Professor Adler, and Logic,” he is still saying, “Truth is a matter of judgment, ultimately, and accurately speaking, a matter of taste.” And to give just one more example, in “The Rights of Man and Natural Law” (1944) he comments, “To the extent that any truth is subject to a supposed objective test, as in natural science, the issue is merely carried back to the validity of the test, which depends on … social criteria. … even the truths of science are finally judgments of value.”
Related to his pragmatist leanings is Knight's clear methodological dualism. For example, once again in “Fact and Metaphysics in Economic Psychology” (1925), Knight comments that, “subjective idealism and scientific realism finally split on the same rock, the common-sense impossibility of disbelieving in the reality of other minds. … we cannot talk about human beings in monistic terms and talk sense.” How much turns on what he means by “cannot”: cannot now, or cannot ever? He seems to mean cannot ever, which may appear a strong statement. Note however that some of the best modern philosophy of action (Davidson 1980) lends support to the strong version of this statement.
The bulk of Knight's methodological concerns are epistemological. However he also consider metaphysical questions occasionally relevant, most obviously in his discussions of freedom. At times he even broaches the question of free will, which may interest Austrian and Post Keynesian economists---the latter have especially tended to stress free will as a source of economic uncertainty. Knight claims that free will is obvious. For example, in “The Meaning of Freedom” (1943) he states, “To deny freedom is to deny the reality of denying and to assert that error itself, as well as all effort, is an error, an illusion, and illusion itself an illusion. Beyond this, self-contradiction cannot go. It is literally impossible to assert that one is not asserting, or that there is no difference between making a statement and making a noise. Freedom is an ultimate datum of experience and a condition prerequisite to all discourse, even to all thinking.” Similarly, in “The Rights of Man and the Natural Law” (1944) he states, “[Freedom] is an intellectual mystery or surd, yet the most certainly known of all facts, since it is a presupposition of all thinking and cannot be denied without asserting it.” In such statements, Knight asserts that truth must be attributed freely to propositions if the notion of truth (and therefore even the notion of a proposition) is to bear any weight. However it is worth noting that (at least in these articles) Knight has not left us was a clear metaphysical statement on the reality of free will. (Indeed, in “The Sickness of a Liberal Society” (1947) he claims that “[a] philosophical discussion of freedom would merely lead into endless metaphysical speculation.”) He has argued that we cannot coherently deny free will, but this is subtly different: it may simply be a statement about our capabilities. In addition, it seems to me that Knight goes too far here. He should have stopped with the pragmatic observation offered by Mises (1966) that we experience ourselves as choosing. Whether the conscious experience of intentionality and choice is possible in a purely physicalist universe is a question that is viewed as very much open by many researchers on consciousness. I would hazard a guess that in most early 21st century intellectuals, a vague hope that the answer is no cohabits with a forlorn conviction that the answer must be yes.
Freedom is the most essential of Knight's concerns. While his two 1943 articles, “The Meaning of Freedom” and “The Ideal of Freedom,” lay out his concerns in great detail, they are pervasive in his writings. For Knight the great problems of society are not purely economic, they are ethical. The core ethical problems concern the reconciliation of the conflicting objectives of individuals who must live together. “The essential fact about human society … is that human beings have both conflicting and common interests.” Common interests are the reasons we associate. Law, in the broadest sense, serves to prevent a “violent or destructive clash in consequence of conflicting interests.” The essence of freedom is the absence of coercion, and institutionalizing the proper limitations on coercion is the fundamental ethical task facing society. Taking a longer view, the need is for institutions that produce individuals “fit for membership in a free society.” (Recall that Knight emphasizes that individuals are products of their social environment.) The romantic temperament, which inadequately tempers the ideal of action with rational thought and knowledge of what is feasible, is the largest threat to society.
Knight's writing is scholarly, witty, acerbic, and methodologically sophisticated. I cannot recall the last time I so enjoyed the writing of an economist. The contrast with contemporary economic writing, which tends toward methodologically ungrounded empirics and dry formalisms, could hardly be greater. I find myself eager and wryly amused when I am reading Knight, and I encourage economists to read these essays. We are reminded that current econometric explorations usually remain methodologically ungrounded (estimate the “true” parameter values!), and that popular formalisms hide much of what should interest any serious economist (“assume” markets clear!). Knight offers important prophylaxis against the many unreflective excesses of contemporary economics, but his critiques are fundamentally friendly to anyone who earnestly seeks to discover what is truth in economics.
Davidson, D. (1980) Essays on Actions and Events (Oxford: Oxford University Press)
von Mises, L. (1966) Human Action: A Treatise on Economics (Chicago: Contemporary Books, Inc.) Third Revised Edition.
Alan G. Isaac
American University