Although much has been written about modern Austrian economics in the past five to seven years, relatively few economists understand the concerns, insights and importance of this once-forgotten school of economic thought. The Austrian tradition arises, of course, out of the work of Carl Menger, Friedrich von Wieser and Eugen von Böhm-Bawerk. In more recent times, it has been self-consciously continued and developed by Friedrich A. Hayek, Israel M. Kirzner, Ludwig M. Lachmann, Ludwig von Mises and Murray N. Rothbard. Despite the rather small number of economists, at least until recently, who have explicitly worked within the Austrian tradition, some of the themes or elements of this school have worked themselves into the writings of prominent neoclassical thinkers (Hicks, 1976). Consequently, the ideas discussed in this paper may not seem entirely new. Nevertheless, the unique Austrian contribution lies in (1) the integration or unity of its themes and (2) the logical consistency with which they are pursued throughout all areas of economics.
The Austrian tradition is not a ready-made set of dogmatic answers to major theoretical questions or, still less, a set of policy prescriptions. Instead, it is a value-free way of looking at “the economic problem,” an orientation or a research program (Lakatos, 1970). Unfortunately, it is not a fully developed program, perhaps in part because of the great Keynesian diversion in the history of economic thought (Yeager, 1973). In any event, there has been a renewal of interest in this tradition due to Hayek’s 1974 Nobel Prize, the sterility of much mathematical economics and the surge of interest in the micro-foundations of macrotheory. The purpose of this paper then is to present an overall view of the Austrian research program as it can be developed through considering a number of important analytical themes.
The fundamental research policy, or “positive heuristic” (Lakatos, 1970), of Austrian economics is to render social phenomena intelligible in terms of individual purposes and plans (human action) (Lachmann, 1971, p. 31). In this spirit, the founder of the tradition, Carl Menger, saw the basic questions of the social sciences as “… how can it be that institutions which serve the common welfare and are extremely significant for its development can come into being without a common will directed toward establishing them?” (Menger, 1963, p. 146). As a statement of the purpose of economics, this is obviously quite broad. In terms of method as well, Austrians have generally considered economics to be a part of a wider social science, “praxeology” (Mises, 1966, pp. 1–3). This is the purely formal discipline that investigates the implications of the self-evident fact that individuals engage in purposeful behavior. It is a pure logic of choice and therefore applicable to all areas of decision-making, not just those traditionally considered economics. International diplomacy, war, legal institutions and the family are some of the noneconomic social phenomena that can be understood praxeologically. As a consequence, the Austrian view is clearly in step with the spirit if not the letter of the growing “imperialistic” movement of economics into many areas of social investigation (Becker, 1976, p. 5).
An institution in Menger’s framework is any coordinated pattern of individual interaction. A legal system is obviously an institution, but so are money and prices. How did it happen that many separate individuals came to agree on the use of a certain commodity as a medium of exchange? How is it possible that so many exchanges take place on mutually advantageous terms without central direction? These are just some of the questions to which Austrian economics must provide answers. Hayek has described the method by which such answers are to be constructed as “compositive” (Hayek, 1955, pp. 36–43). The object of scientific investigation is to recompose or reproduce the social institution under study by tracing its logical origins to the interaction of individual plans. This “methodological individualism” (Hayek, 1955, p. 38) is incompatible with any explanation that proceeds in terms of a solution to a social maximization problem. The relevant data are not “given” to the society as a whole; instead, small pieces of information are spread throughout the minds of a myriad of separate individuals (Hayek, 1948a). Our main question is not answered but evaded if we ignore the decentralization of knowledge and construct an explanation in terms of, for example, a constrained maximum in a social indifference system.
The Austrian concerns that we develop in the course of this essay are not a mere collection of unconnected topics. Each of them can be rigorously derived from the concept of purposeful behavior itself. Ends and means do not automatically produce behavior; it is clearly the individual perception of them that really matters. The idea of purpose thus implies the concept of subjectivity. Further, all actions must take place in time. For unless the end precedes the means all goals would have already been attained (or left forever unattained). The passage of time, as ordinarily perceived, is inextricably linked to the concept of uncertainty. This, in turn, must involve the disappointment of plans and expectations and hence the concept of disequilibrium. Disequilibrium is characterized by unexploited opportunities for mutually advantageous trade and the consequent emergence of pure profits. This, finally, provides incentives for economic processes that approach but never actually attain equilibrium. Although it is possible logically to derive the Austrian themes from a central concern, these individual categories tend inevitably to overlap. Therefore, in discussing any one of them we must necessarily refer to concepts previously developed, as well as anticipate the development of others. Dividing our investigation in the way we have just outlined reveals the internal unity and consistency of the Austrian system.
Subjectivism is a major feature of the Austrian tradition in two interrelated ways. First, as we have already seen, social phenomena are to be explained in terms of individual purposes or plans. The subjective perception of the relevant framework of ends and means is clearly indispensible to this task. Second, it must be stressed that the social phenomena themselves – the subject matter of our study – cannot be defined except in terms of individual intentions (Hayek, 1955, p. 31). A court, for example, is not just a physical building but, more importantly, an institutionalization of a pattern of individual intentions. People go to court with certain very specific purposes in mind. Similarly, money cannot be defined independently of the intentions of those who use it. A medium of exchange differs from ordinary metal or paper precisely in terms of its intended use. Therefore, not only the method but also the very questions of economics are completely intertwined with subjectivism (Morgenstern, 1972, p. 702).
Although no one is denying the existence or relevance of objective entities in economics (subjectivism is not solipsism), Marshallian “scissor” analogies are not an adequate way of portraying the subjective-objective interaction. The data of economics are not neatly divided into the subjective (tastes) and the objective (resources and technology). All of the data are subjective in the sense that only the perception of resources and technology matters. In Hayek’s (1955, p. 27) words, “So far as human actions are concerned the things are what the acting people think they are.” It is not appropriate simply to assume that subjective perceptions are always in harmony with the objective facts. Perfection of knowledge is a phenomenon to which a theoretical explanation must ultimately be given. It cannot be taken for granted. In fact, the perfect or superior knowledge of the “observer” and analyst is not directly relevant to economic explanation. It can only be relevant in the presence of a theory that says this knowledge will eventually be acquired by market participants.
In the next two subsections we elaborate further on the importance of subjectivism by developing its role in some major areas of economic theory. First, although the subjectivity of tastes has long been recognized, its corollary, the subjectivity of costs, has all too frequently been neglected – with unfortunate results. Second, both perfect and imperfect information models have generally assumed that all economic agents possess the same information, thus neglecting an important aspect of the market process. The full importance of the subjectivity of economic data, however, cannot be understood until we later undertake our investigation of uncertainty and processes.
Let us imagine a simple two-commodity world in equilibrium. Nonpecuniary elements are excluded, as are indivisibilities and allied problems. Assume that an individual must be willing to give up one beaver to acquire two deer. If the individual willingly engages in exchange at this rate, what is the cost to him of two deer? The well-trained economist would answer: the opportunity foregone, or one beaver. The answer is not entirely precise. The cost of deer is not beaver, but the want satisfaction given up when beaver is foregone. Cost accordingly belongs in utility space and not in objective space. The cost of deer is the want not fulfilled by beaver (Knight, 1928, p. 359). Cost is not a disembodied magnitude, or technological “given,” but must be related to a concrete decision. To say, therefore, that the cost of deer is beaver must be a shorthand or metaphorical way of saying that the cost is the utility of beaver foregone (Buchanan, 1979, pp. 1–15).
If this were all that there is to the subjectivity of costs, then the whole matter would be little more than a verbal exercise. Metaphors and shorthands can be useful in many contexts. In the two-commodity equilibrium world of deer and beaver, the damage done by metaphorical thinking is minimal. In more complex cases, however, the objectivist frame of mind, to which the cost metaphor gives rise, is extremely misleading.
We now take up the problems with an objectivist cost notion seriatim. Consider, for example, a multi-commodity world with money (or a numéraire), which is in general equilibrium. An individual supplier of labor finds nonpecuniary advantages or disadvantages associated with alternative occupations. His current job pays $7.00 per hour; his highest ranked alternative, however, pays $7.50 but has significant nonpecuniary disadvantages. Making the usual assumption about adjustments at infinitely small margins, the opportunity cost (in money terms) of this person’s time is $7.00 rather than the objectively measurable wage in the best alternative. This is because that occupation has a nonpecuniary disadvantage, which the individual would be willing to pay $0.50 at the margin to avoid. Compare now the slightly different case of an individual whose current job pays $7.00 per hour but also has significant nonpecuniary advantages associated with it. His best alternative pays $7.50 and is without any nonpecuniary features. Here the individual’s opportunity cost is $7.50 instead of his objectively measured wage rate. At the margin, the nonpecuniary advantages are worth $0.50. In the first example, then, where the alternative endeavor was characterized by nonpecuniary disadvantages, the opportunity cost of time was the current wage. In the second example, on the other hand, where the current job had nonpecuniary advantages, the opportunity cost of time was equal to the wage in the best alternative.
Thus even in equilibrium, whenever (as in the illustration above) there are nonpecuniary elements present, money prices do not necessarily reflect the goods foregone by the decision-maker. In order to determine what he has given up in any context, we must know his scale of values, i.e., his utility function. Out of equilibrium, however, monetary magnitudes bear no definite relation at all to subjective costs. Once we remove ourselves from the imaginary constructions of general equilibrium models, subjectivist elements become still more important. Since opportunity costs involve expectations and these expectations will differ among individuals, some agents will believe that the true opportunity costs are higher or lower than money costs observed in markets. Hence, costs will involve expectations and these expectations will differ among individuals.
In the general equilibrium world of unchanging data, the past and future merge into the ever-present. Actions are replicated period after period, so that one can no longer even distinguish between periods except by subscripts. In such a world, we could argue that the analyst could identify the relevant good foregone by an individual – the objective opportunity foregone. Outside of stationary equilibrium, in a world of change and uncertainty, the ideal observer cannot even identify the objective opportunities foregone by decision-makers, much less the subjective cost.
Some of the practical implications of these strictures are considered in the subsection (below) “Misplaced concreteness: static welfare economics.” Two points, one specific and one general, can be made here. The specific point relates to our original beaver-deer example. Only insofar as the individual is a pure medium for the flow of market forces are his costs “objective” in the metaphorical sense described above. For this to be the case, the only foregone opportunity attached to labor must be the market value of the product produced. More generally, the subjectivist cost theory points out problems with crude economic-man constructs and assumptions about perfection and uniformity of knowledge in a world of incomplete information and specialized knowledge. If our object is to explain behavior, the subjectivist framework tells us that there is more to look at than objectively measurable prices. Actions that appear to the objectivist as irrational, or misallocations, may rather reflect the fact that the decision-maker had different values or different information than the welfare theorist.
One very practical contribution of subjectivists has been in public choice and public policy discussions. Analysis has been shifted from the putative data of welfare-maximization models to the actual costs and constraints faced by the relevant individual choosers. We understand political and bureaucratic decision-making better today precisely to the degree that we have subjectified this area of economics. Our understanding of nonmarket decision-making is still in its infancy. A consistent pursuit of subjectivist economics here would add immeasurably to our economic understanding in this area. Likewise, the economic theory of the firm would be enhanced by a consistent subjectivist approach to the issues raised in this area.
In the previous section we saw the importance of subjectivity even where individuals were more or less the same. To be sure, where individuals have different utility functions the subjectivity of costs poses even more complex problems of measurement. However, the heterogeneity of individuals extends far beyond their possibly different utility functions – to the diversity of knowledge. In the standard constrained maximization problem, the equilibrium of the individual is defined relative to “given” data. Although it is not unusual to allow taste to differ among individuals,1 resources and technology are frequently assumed not to differ even in imperfect information models. In order to emphasize some fundamental conceptual issues we shall ignore taste differences in the following discussion, and instead concentrate on differential knowledge of both means and technology. There are two kinds of knowledge dispersion or diversity between which it is important to distinguish. First, individuals may have different and incompatible ideas with respect to the same objective facts. Second, they may have knowledge of different sorts of things, e.g., some people may know a great deal about conditions in the auto industry and others very little. These two types of knowledge diversity raise different theoretical issues and therefore we shall handle them separately.
Insofar as those with mistaken technological ideas improve their knowledge, their behavior will undergo change. As expectations are revealed to be incorrect, agents will revise their forecasts.2 For the Austrian, the important question is not so much what the new final equilibrium will be; instead his attention focuses on the process of adjustment toward an equilibrium. There are thus two conceptually distinct problems: (1) How do people adjust to each other and (2) how do they adjust to the underlying objective data? The divergence of expectations among individuals is one of the most important reasons for the existence of asset markets (Lachmann, 1978, p. 4). Attempts to learn the true underlying demand curve for a product manifest themselves in advertising, market research and price-cutting – all of which are incompatible with perfectly competitive equilibrium (Kirzner, 1973, pp. 151–80). In addition, price-cutting tends to create incentives for firms to improve their technological knowledge and thereby approach the best technology. Market phenomena such as these are essentially adjustment processes in a world of heterogeneous “data.”
The second type of dispersion of knowledge consists of what Hayek called “the knowledge of the particular circumstances of time and place” (Hayek, 1948a, p. 80). Different individuals will have knowledge of different things. Some people will be aware of alternate methods of producing steel while others will be experts on demand conditions in the market for rubber. Clearly, this division of knowledge is related to the division of labor in society: people will learn only what is important to their activities. Market prices tend to communicate or disseminate the relevant aspects of this differential information (Hayek, 1978, pp. 179–90). If in some uses rubber and steel are substitutes, then a technological revolution in the production of steel will be communicated to the rubber industry through a downward shift in their demand curve. No one in the rubber industry need become an expert on steel production to react properly. The divergence of knowledge of this kind may be compatible with equilibrium. Unlike the diversity of knowledge that stems from conflicting or incompatible conjectures, the division of knowledge in society may still exist in competitive equilibrium. In such circumstances, market prices convey a knowledge surrogate to all economic agents. To the extent that market processes push the system toward equilibrium, they are then the means by which the division of knowledge is continually refined and its relevant components more accurately communicated.
In recent years, Hayek’s “knowledge of the particular circumstances of time and place” has played an important role in a varied literature on incentive-compatible systems, efficient markets and rational expectations (Hurwicz, 1973; Grossman and Stiglitz, 1976). In this literature, the division of knowledge appears as “localized” or “private” knowledge. It has also arrived independently at a number of the conclusions reached in Austrian economics: the comparatively large information requirements of centralized planning (Frydman, 1980), and the inapplicability of rational expectations models to situations of divergent expectations and rapid change (Frydman, 1980; O’Driscoll, 1979). The independence of these findings confirms the conceptual power of Hayek’s original insight.
The perception of means and ends entails both the passage of time and relationship of cause and effect. Ends are usually attained after some period of acting in time (Mises, 1966, pp. 99–101). In addition, individuals must be able to perceive causal relationships in order to act: certain means must be viewed as efficacious in achieving specific goals. In a world without causality acting man would be paralyzed.
For some, Austrian economics is nothing but a particular capital theory. In this view, Böhm-Bawerk (1959) made the authoritative statement at the core of Austrian economics, while Hayek (1941) restated it in modern terms. We are now in a position to explain not only why Austrian economists have always been concerned with capital theory, but also why their capital theory does not define Austrian economics.
A consistent subjectivist approach necessarily entails an emphasis on time. All action is future oriented. There is no possibility of influencing the present (Mises, 1966, pp. 252–3). Economic agents can, however, make the decision for which time periods in the future that they will plan. Austrian economists have accordingly focused on this choice and on its implications for interest and capital theories. They have tended to separate capital from interest theory more than do most economists today. Following Böhm-Bawerk, they viewed interest as a phenomenon resulting from the exchange of goods through time. Indeed, in following this insight through to its logical conclusions, writers such as Mises, Kirzner and Rothbard removed all productivity elements from their interest theories (Mises, 1966, pp. 524–37; Kirzner, 1966; Rothbard, 1962, pp. 313–86). While their pure time preference theory of interest is still controversial among Austrians, the relatively sharp separation of capital theory from interest (or capitalization) theory is not. Interest theory, among other things, gives us the basis for a theory of capitalization or capital value. Capital theory proper deals with a highly important, though more narrowly focused, set of issues:
The central aim of this study is to make a systematic survey of the interrelations between the different parts of the material structure of the process of production, and the way in which it will adapt itself to changing conditions … Our main concern will be to discuss in general terms what type of equipment it will be most profitable to create under various conditions, and how the equipment at any moment will be used, rather than to explain the factors which determined the value of a given stock of productive equipment and of the income that will be derived from it.
(Hayek, 1941, p. 3)
While it is certainly sensible and even illuminating to talk of capital allocation for an isolated individual, the central issues of modern Austrian capital theory revolve around the problems of intertemporal plan coordination. The major question concerns the economy’s ability to produce a concrete set of capital goods (i.e., a capital structure) consistent with intertemporal plan coordination. In their applied work, Austrians have emphasized the causes and effects of intertemporal discoordination. (We take up one example of this applied work below in our discussion of money’s effects on relative prices.) Viewed from this perspective, the Austrian focus on capital theory flows from, but does not itself define, the Austrian research program. Attention to capital and time derives from attention to the process of individual decision-making. In this century, Austrians have to one degree or another moved beyond Böhm-Bawerk’s position. They have all endorsed his goal of subjectivizing capital theory by removing exclusively technical factors from capital and interest theories. Böhm-Bawerk demonstrated that one cannot establish that production produces value merely by showing that it is physically productive. Yet he relied on this same non sequitur in his own theory. The central question must be: why don’t the values of the factors of production fully absorb the value of the output? Among modern Austrians, Mises was the first to criticize Böhm-Bawerk for this error. Hayek’s Pure Theory of Capital (1941) is a link between the older and newer Austrian capital theorists. Though a Böhm-Bawerkian work, it is also an immanent criticism of parts of that system. Among other things, Hayek criticized the average period of production concept (1941, pp. 76, 144–6, 199–200). He demonstrated that one could not establish a unique functional relationship between quantitative measures of capital and the rate of interest. No measure of capital could be independent of relative prices, and thus could not be independent of interest rates. To this extent, the Cambridge capital theorists have only recently arrived at the position that the Austrians had adopted by 1941.
Having explained the traditional Austrian focus on capital theory, we can also explain the transformation in the type of contributions made by the Austrian school to capital theory. Their contribution now typically occurs in a different area (Kirzner, 1966; Lachmann, 1977). The central issue of intertemporal coordination now appears under a new guise: the coordinative function of the entrepreneur over time. While we take up this topic elsewhere, we note here that the traditional Austrian concern with intertemporal coordination has not been abandoned but extended and refined by Kirzner, Lachmann and others.
In contrast to much contemporary thought, Austrians view the framework of cause and effect as absolutely indispensable to satisfactory economic analysis. Recall that the Austrian research program rests on the potential intelligibility of social phenomena in terms of human action. From this starting point it is clear that there are two related reasons for the importance of causality. First, to understand why people undertake certain actions one cannot omit their perception of causal relations. In the absence of such perception they could not act at all. Second, to render social phenomena intelligible means to show how they are the result of, or caused by, individual action.
The atemporal framework of competitive equilibrium theory, by contrast, is incompatible with cause and effect analysis. This model is merely a set of simultaneous equations and their solutions: there is no sense of how the economy actually attains the specified equilibrium (Hayek, 1948b, p. 94). Indeed Brian Loasby tells us that “[e]quilibrium models cannot explain, because they specify no causal process.” Even comparative statics does not deal with causal explanations despite the fact we tell motivating stories “between the equations” (Loasby, 1976, p. 26, emphasis added). These stories are not a part of the formal model and may in fact be inconsistent with it (Loasby, 1976, p. 45). The comparative static methodology requires that the assumed change in a parameter or exogenous variable be completely unexpected (because expectations are assumed to be static). However, it is unlikely that after several comparative static shocks the new equilibrium will be one in which change is totally unexpected. Hence the static nature of expectations characteristic of this apparatus is inconsistent with the use to which we put it.
Recently, Sir John Hicks (1979, pp. 18–26) has explicated a notion of “contemporaneous causation” to allow for causal explanations in static equilibrium models. This idea, however, rests on an artificial view of time. Hicks’ suggestion is that if we stretch out the relevant period by an amount long enough we can accommodate both cause and effect within the same period. This notion of the relevant period is not, however, the economically pertinent one. The temporal separation of cause and effect is fundamental to any focus on individual planning and acting. Contemporaneous causation thus has nothing to do with economic causation.
The importance of uncertainty in decision-making is obvious from causal observation. However, to the extent that we are concerned with processes, there is also an analytical need to introduce uncertainty into economic models. The assumption of perfect foresight is logically incompatible with any tendency toward equilibrium.3 Attempts to construct models with equilibrating processes will dissolve into logical circularities unless the assumption is abandoned. We can easily demonstrate this proposition by use of two related examples.
More than 50 years ago, Oskar Morgenstern (1976, p. 174) showed that when one individual’s plan is dependent on that of another, perfect foresight will produce “an endless chain of reciprocally conjectural reactions and counter-reactions.” An equilibrium set of plans cannot be attained except by the sheerest of accidents. Morgenstern’s example is worth quoting in full:
Sherlock Holmes, pursued by his opponent, Moriarty, leaves London for Dover. The train stops at a station on the way, and he alights there rather than travelling on to Dover. He has seen Moriarty at the railway station, recognizes that he is very clever and expects that Moriarty will take a faster special train in order to catch him in Dover. Holmes’ anticipation turns out to be correct. But what if Moriarty had been still more clever, had estimated Holmes’ mental abilities better and had foreseen his actions accordingly? Then, obviously, he would have traveled to the intermediate station. Holmes, again, would have had to calculate that, and he himself would have decided to go on to Dover. Whereupon, Moriarty would again have “reacted” differently. Because of so much thinking they might not have been able to act at all or the intellectually weaker of the two would have surrendered to the other in the Victoria Station, since the whole flight would have become unnecessary.
(Morgenstern, 1976, pp. 173–4)
The equilibrium in this illustration is a stable set of plans that might result either in Holmes escaping or being captured. That aspect of the solution is unimportant for us; what is important is the absence of any tendency for the plans of the individuals to change. In equilibrium their plans would be coordinated (even though they are enemies) in the sense that each is optimal given what the other individual is in fact planning.4 The moral of the Holmes-Moriarty story, however, is that perfect knowledge of each other’s plans makes attaining such an equilibrium impossible. Neither individual will ever be satisfied with any tentatively adopted deterministic plan and so coordination cannot be achieved.
The second example was developed 20 years ago in a more familiar market context by G. B. Richardson (1960, pp. 30–8). Assume, initially, that in one industry there is an “objectively” expected rate of return higher than the normal rate. (There is some underlying objective structure “producing” the future.) Assume further that all market participants have perfect foresight of this and that all the other conventional assumptions of the competitive model have been satisfied. If each potential entrant must take into account the behavior of others because they can produce competitive supply, there will be either an indefinite expansion of output or no expansion at all. If everyone foresees a supernormal return then an indefinitely large number of firms will enter at the same time and the actual rate of return will be indefinitely lower than that “objectively” expected return. But if people are very clever they will foresee this and not enter. If, however, no one enters then the rate will again fail to be bid down to the equilibrium level. In neither case, then, is the assumption of perfect foresight compatible with a tendency toward equilibrium. This is not a mere cobweb-like situation but rather a problem of logical circularity. Each potential entrant must make his decisions on the basis of what others decide. Perfect knowledge of what they intend to do, however, paralyzes decision-makers. One cannot assume lagged entry to eliminate Richardson’s conundrum. Differential speeds of entry are inconsistent with perfect knowledge. If all firms have perfect knowledge, none should be slower or quicker to enter (Gordon and Hynes, 1970). As we suggest below, divergent expectations may be a necessary condition for a movement toward equilibrium.
Although it is evident that the perfect foresight assumption cannot be reconciled with a theory of economic process, it is not immediately clear how we ought to model uncertainty. The purpose of this subsection is to examine three different views of uncertainty and to determine the strengths and limitations of each.
This can be defined as an objectively measurable uncertainty that can be converted into a certain prospect by the judicious grouping of cases (Knight, 1971, pp. 231–2). The relative frequency of accidental fires, mortality rates in socioeconomic age categories, and the number of certain types of illnesses per 100,000 population are all collections of risky events. The apparently random character of these events is responsible for their stability in the aggregate. To the extent that individuals are concerned with these aggregations of events, the uncertainty is “transformable” or “eradicable” (Weston, 1950, pp. 43–4; 1954, p. 155).
The objectivity of risk is capable of at least three interpretations that are relevant for our purposes. In cases of structured games (e.g., tosses of a fair coin, throws of a fair die) the probabilities can be viewed as part of the experimental setup. The game is, in effect, programmed to yield certain relative frequencies in a long series of trials. In other circumstances (e.g., mortality tables) empirical regularities have been noted over significant periods of time. These regularities then become hypotheses about future sequences of events that are often well corroborated. Insurance companies and others used them in, for example, the computations of premiums, because they produce satisfactory results. Finally, it is possible to take a thoroughgoing Bayesian view and say that objective probabilities are merely uncontroversial probabilities. The problem with this approach, however, is that it leaves unexplained why some probabilities are uncontroversial and others are not. Furthermore, if probabilities were purely arbitrary degrees of belief it would be hard to generate any tendency toward agreement as in models of stationary stochastic equilibria. For our purposes, then, we shall adopt as a useful maintained hypothesis that some probability assignments do reflect an underlying objective structure while others, in cases of more complex phenomena, do not. This objectivity can be of a preprogrammed, a priori variety or of a statistically based, empirical kind.
“True” or Knightian uncertainty is objectively unmeasurable or, if it is so measurable, it cannot be converted into a certain prospect because the grouping of cases is economically unfeasible or irrelevant. This kind of uncertainty is clearly “nontransformable.” Although in theoretical terms the distinction between risk and Knightian uncertainty is quite sharp, in reality there is a continuum from the easily measurable to the “completely” unmeasurable. The degree of measurability or, more precisely, the usefulness of the measure adopted, clearly depends on the cleverness with which the analyst chooses the reference class for a given event. Consider, for example, the case of a “unique” investment decision. It may not be possible to find other decisions sufficiently homogeneous5 to construct a very useful class. However, “instead of taking the decisions of other men in situations more or less similar objectively, we may take decisions of the same man in all sorts of situations” (Knight, 1971, p. 228). From the point of view of a bank contemplating a loan to this entrepreneur, such a reference class may indeed be the relevant one. Consequently, the determination of the correct class depends upon the purpose of the individual economic agent. Probability measurements, then, clearly have a large subjective (but not arbitrary) component whether they are made in the context of risk or Knightian uncertainty. In the former case, once the reference class is established, the probability is objective, but in the latter, even with an accepted class, an objective probability estimate cannot be established.
Individual cases in the category of risky events are not themselves risky but instead they are uncertain (Knight, 1971, p. 234). Objective probabilities are relative frequencies and hence apply directly only to the reference class. The individual toss of a coin, for example, cannot have a relative frequency of one-half for heads: it is either heads or tails. In those cases where the economic agent is only concerned about the single toss or case, the overall objective frequencies are of no direct relevance to his decision-making (Buchanan and Di Pierro, 1980, pp. 696–8).
Even when the economic agent is dealing with large classes of events, there is a question of the relevance of the observed frequencies to the class at hand. Since no two classes are exactly the same, there is a subjective judgment involved in the decision that the two are sufficiently similar, so that the data on one are useful for predicting certain characteristics of the other. For example, to what extent are data on the frequency of tropical diseases among group X applicable to Y? In what respects must X and Y be the same in order for the inference to be reliable? Unless there is a whole sequence of such decisions to which we can refer, this judgment must take place under conditions of Knightian uncertainty.
Those who favor the personal probability approach may find the distinction between risk and uncertainty unhelpful. If probabilities are merely decision weights, then it may seem that we can apply them without regard to the nature of the uncertainty. This argument, however, confuses the behavioral and cognitive levels of analysis (Buchanan and Di Pierro, 1980, p. 695). From a behavioral perspective, decision weights may be useful under both risk and uncertainty without any distinction. From a cognitive perspective, however, the existence of uncertainty may be an important element in explaining why decision weights differ among people (Buchanan and Di Pierro, 1980, p. 696).
The heterogeneity of expectations is more likely under Knightian uncertainty than under risk. Where individuals react in the same way toward uncertainty, the divergence of expectations is necessary to ensure a tendency toward equilibrium. Without such divergence, the Morgenstern-Richardson paradoxes would reassert themselves and make process analysis an impossibility.
In the two previous forms of uncertainty, each individual was assumed to be capable of listing all of the various possible outcomes. The sample space was thus viewed as complete. However, to the extent an individual does not know what outcomes are possible he will not be confident of any probability distribution (Loasby, 1976, p. 9). In fact, he may feel quite strongly that something not part of the sample space or not on the probability distribution will happen. Cases of technological innovation, for example, are not just characterized by the inability to assign objective probabilities to each of the possible inventions. Instead, the individual may have no idea what the complete set of possible inventions is.
G. L. S. Shackle has developed an approach to uncertainty that does not require complete listability and hence is more appropriate to circumstances like those described above. He has constructed a nondistributional measure of uncertainty known as “potential surprise” (Shackle, 1969, pp. 67–113). His spectrum extends from a zero degree of potential surprise (“perfectly possible”) to an arbitrary maximum of potential surprise (“impossible”). The possibilities over which potential surprise is measured are rival, mutually exclusive outcomes. In no sense must these possibilities be exhaustive. It is not necessary to “distribute” one’s certainty over a complete set of outcomes as in the probability approach. Certainty (i.e., a probability of unity) cannot be ascribed to more than one rival outcome or hypothesis. A zero degree of potential surprise, however, can be given to any number of rival possibilities. An individual can believe that he would not be the least bit surprised if any one of an indefinite number of rival possibilities occurred. In addition, as more possibilities become known, none of the original alternatives need have a greater degree of potential surprise as a consequence. If, for example, all five initial hypotheses occurred to an individual, these additional possibilities could also have zero potential surprise. Each of the seven can be just as perfectly possible as each of the five were. The probability approach, on the other hand, requires that the new alternatives “borrow” some of the certainty from the initial possibilities and hence the degree of belief in these must fall.
The limitations of equilibrium analysis and the need for a comprehensive theory of disequilibrium adjustment are made more evident when we understand the proper function of equilibrium constructs. By abstracting from change altogether and specifying the conditions under which it will not occur, it is possible to begin to understand the forces that induce change (Mises, 1966, p. 248). Equilibrium serves as an analytic focal point toward which some forces tend and from which other forces pull away. In addition, the construct enables us to trace out the full effects of a given change under static conditions (Machlup, 1967, p. 48). While this may not be a “realistic” description of the effects, it specifies the tendencies that a certain change alone sets in motion. The essence of equilibrium theorizing is thus predictability of effects within a model (Rizzo, 1979, p. 3). Consequently, all equilibrium models are conceptual tools and not descriptions of reality (Machlup, 1967, pp. 56–8).
The foregoing interpretation of the function of the equilibrium notion sets up a critical tension within the concept itself. For example, to emphasize tracing out the full effects of an assumed change is to use causal language: A results in B. Thus time has crept into the analysis. Furthermore, it is impossible to know why the condition of plan consistency is a state of rest unless we tell ourselves some informal story about processes. If ex ante supply and demand are equal on a certain market it is not immediately obvious why people shouldn’t change their behavior slightly in an effort to do better. We have a story about how a small displacement from the equilibrium sets in motion forces to bring us back.
The unique advantages of adopting the Hayekian (Hayek, 1948c, pp. 33–44) concept of equilibrium as plan consistency6 rest on the ease with which the concept is adaptable to a theory of process. Since plans unfold in time the development of a theory of plan adjustment to unexpected changes is a natural offshoot of the Hayekian equilibrium. Furthermore, plan consistency focuses on purposes and anticipations and so gives rise to less rigid or mechanistic dynamics.
In recent years there have been attempts to reduce the need for disequilibrium models by greatly extending the meaning of the term “equilibrium” (Rizzo, 1979, pp. 2–7). Frank Hahn (1973, p. 28), for example, says that “[t]he traditional notion of an equilibrium … requires the equilibrium actions of agents to be consistent, whereas I have the weaker requirement that they not be systematically and persistently inconsistent.” Consider a world in which the prices faced by firms have a random component. In period after period the firms will try to guess what these prices will be so that they can correctly determine their level of output. If the process that generates the random price components is stable, firms will eventually settle on a constant expectations function (Stigum, 1969; Hahn, 1952, pp. 803–4; Littlechild, 1977, pp. 6–8). They will find the best way of predicting prices and never change that method. Even the best method, however, will not always be correct, but it will not be “systematically and persistently” incorrect. Hence firms will at times produce too much output and at other times too little output; equilibrium in this sense exhibits not complete consistency but only “optimal” consistency.
This approach has far more in common with traditional equilibrium models than it does with any kind of process analysis. No learning at all occurs in this world because, by assumption, people have already discovered the best way to cope with their risky environment. In fact, all that has happened is that the perfect foresight assumption of the old models has been replaced with perfect stochastic knowledge. How anyone would acquire such knowledge is not a question that can be entertained.
As equilibrium models, however, it is not clear that stochastic equilibria are preferable to old-fashioned deterministic equilibria. Perhaps the main motivation for developing such stochastic models is that they are more “realistic.” After all, it is not often that we see complete plan coordination.7 A stochastic equilibrium is more nearly compatible with the frustrations and disappointments of the real world. To the extent, however, that this idea underlies the foregoing approach, it is seriously misleading. As we have already seen, equilibrium is not meant to be a description of reality but a conceptual tool. Making such a tool more “realistic” may in fact be harmful if it now fails to isolate the proper variables. To consider imperfect coordination as the equilibrium state leaves us at a loss to describe the direction of change when, for example, the stochastic component of prices decreases. We must, at least implicitly, retain a vision of the perfectly coordinated state of affairs to understand what happens when parameters of the stochastic model change. The major problem of stochastic equilibria is that in being more “realistic” they may convince us that we needn’t bother about developing disequilibrium models that can explain the processes of change and learning. As the next section will make quite clear, this would be a terrible mistake.
The models discussed above make the artificial assumption that all agents possess the same perfect stochastic knowledge. This is at odds with not only the subjectivity and diversity of individuals but also with the costliness of gathering information. Suppose, instead, we postulate a world in which individuals’ perceptions of an underlying stochastic structure are gradually undergoing change. This could be either because the costs of information-gathering are falling or merely because more “draws from the urn” convey a better picture of the underlying distribution. In any event, such changing perceptions imply that agents are learning and, hence, the method by which they attempt to forecast relevant events will itself change. This kind of model is thus inconsistent with the constancy of expectation functions.
If we further complicate the situation and question the modeling of uncertainty purely as risk, the divergence of expectations takes on new significance. Insofar as uncertainty is objectively unmeasurable, people will develop different plans even if they have the same attitude toward uncertainty. Recall that in a Knightian world, personal probability distributions are likely to diverge and so different people will expect different things. Hence the degree of discoordination is likely to be greater than in a risky world. In addition, it is unclear that there is any meaning to the concept of a stochastic equilibrium if we jettison the objective measurability of probabilities. Under these conditions, we probably must define equilibrium in terms of the complete fulfillment of expectations.
Models of stochastic equilibrium are completely incapable of handling radical uncertainty. If an individual were to face a price that did not lie on his probability distribution, the model could not explain how he would react (Stigum, 1969, p. 549). True learning has not taken place and therefore it is unlikely that the individual would react in some constant way, i.e., his expectations function itself would change. This is because the disappointment “of the first situation must always enter as a new parameter into the second” decision (Hahn, 1952, p. 805). The agent will take account of his previous mistakes. Consequently, the stability of expectation functions and hence the maintenance of stochastic equilibrium is dependent on the complete elimination of surprises.
Is it possible to say something about how expectation functions will change as people engage in genuine learning? Obviously, some kind of metatheory is necessary to explain a change in the “theories” by which people form their expectations or make their predictions. Such a theory is not yet available (Hahn, 1973, p. 21). Nevertheless, we can say something about how it might look. In the first place, recognition of the possibility that something undreamt might happen will itself have an impact on the confidence with which people hold their expectations. Second, what people will learn from their disappointments and how they will respond to them is something that we can hope to explain only in very general terms. We might be able to exclude some possible modes of adjustment or, alternatively, specify a probably large set of possible expectations functions. Fundamentally, our inability to determine more precisely the method by which changes in these functions occur lies in the impossibility of predicting the future course of knowledge. That course is radically uncertain.
Although many theorists will admit that general equilibrium constructs are analytical tools rather than descriptions of reality, most, if not all, welfare economics proceeds on the assumption that all markets, except for the one under study, are in competitive equilibrium. In an analysis of taxes and subsidies, for example, the market prices of resources are assumed to reflect correctly their value in alternative uses. This implies, of course, equilibrium not only in those factor markets but also in the markets for the alternative output. In the static theory of monopoly, furthermore, the conclusion that monopolists underproduce is predicated on the absence of distortions in substitute and complement markets. The general theory of second best and its complex optimality conditions have been developed to deal with simple situations where distortions are known (Lipsey and Lancaster, 1957). Nevertheless, most of what welfare economists do rests on the assumption that the conceptual tool of equilibrium mirrors a good deal of the real world quite accurately. To the extent that this is not true, however, the foundations of applied welfare theory are shaky.
The focus of Austrian economics, as we have seen, is on the plans or actions of individuals rather than directly on social wholes or institutions. One of the key difficulties of some applications of static welfare theory is that focus of the individual is suppressed by use of the aggregative concept of social welfare. This is particularly true in contexts that make use of the Kaldor-Hicks potential compensation principle: the actual loss of one individual is “offset” by the gain to another. Whether the social welfare measure is based on a willingness-to-pay approach or on a social-indifference system does indeed matter. The latter, of course, is even more offensive to those who take seriously the differences among individuals. Nevertheless, both approaches tend to aggregate gains and losses in a way appropriate if society were a single individual. On the other hand, Austrians prefer to view social institutions in terms of their ability to coordinate individual plans (Kirzner, 1973, pp. 212–34). Certainty creation, knowledge dissemination, and incentive enhancement are all factors that tend to increase the opportunities for voluntary exchange and thus the possibilities for coordination. In the final analysis, however, no approach to economics can eliminate the need for ethics and social philosophy in the determination of policy. Hence the coordination approach, like all others, will never be a complete basis for policy unto itself.
Inefficiencies of all types can be traced to the inadequacy of knowledge. Since Coase’s classic article (1960), economists have tended to categorize inefficiencies by the type of transaction cost identified as their source. These costs are commonly divided into search, contracting and enforcement costs. All of these stem, however, from the absence of perfect knowledge (Dahlman, 1979). Search would, of course, be unnecessary were knowledge complete. Contracting costs consist of discovering mutually acceptable terms of exchange. And there would be no enforcement costs if everyone knew in advance who would and who would not keep agreements.8 Thus each inefficiency can be attributed to a particular transaction cost, which, in turn, is the result of certain information being absent or too costly to obtain.
In an important sense, however, there can never be inefficiencies in a static model even when information is imperfect. Inefficiency must be relative to certain constraints. A perfectly competitive world in which oil was a scarce commodity would seem inefficient when compared to one in which it was a free good. Analogously, the greater the degree of technological information possessed by the members of a society, the more “efficient” that society would seem to be. A world of complete knowledge of objectively available technological options, however, is not a relevant standard against which the efficiency of the current system can be measured (Demsetz, 1969).
Why should efficiency be defined in terms of some unattainable goal? People make the best use of the knowledge on hand in a world of limited information, just as they do in a perfectly competitive one. If some people in the former world have better information than others, then the others would have already purchased the fuller knowledge to the optimal extent. Similarly, if an institutional change could increase efficiency then that change would have already been affected to the optimal degree. In a purely static framework, all opportunities for potential gains are efficiently exploited. In that world, whatever is must also be efficient. Therefore, a static welfare framework is ultimately self-destructive since it degenerates into a crude tautology and is thus unable to provide guidance for desirable change.
Dynamic or disequilibrium elements must be introduced into a welfare framework in order to avoid the problematic reasoning discussed above. If this is done, inefficiencies can be seen as unexploited opportunities from the perspective of superior knowledge that has not yet been fully absorbed. This lack of full absorption by the economic system might, of course, be attributed to costs and then the imperfect absorption can be efficient at each point in time. This, however, would return us to a static equilibrium framework. To view such inefficiencies as a disequilibrium phenomenon, however, is essentially to predict that if agents were aware of certain facts they would exploit specified opportunities. If we have reason to believe that this superior knowledge will become more widely known, then we can begin to say something about the direction in which the economy will move.
Superior information about unexploited exchange opportunities can be possessed by the economic agent-entrepreneur or, less likely, by the economic analyst. In the first case, entrepreneurs incur profits or losses to the extent their information is correct or incorrect. Hence here is some feedback mechanism exercising at least a limited self-corrective tendency. In the latter case, however, there is no such feedback. Economists are not normally in a position to observe the realization of the potential profits or losses that are implicit in their statements about inefficiencies. Welfare economics is thus a kind of vicarious entrepreneurship,9 the full significance of which can only be recognized in a completely dynamic framework.
Static welfare theory is inextricably linked to the general equilibrium framework even when it is seemingly applied in a partial equilibrium context. Analyses of output market distortions, as we have seen, presuppose that factor prices measure “social” opportunity costs with tolerable accuracy. This, however, cannot be simply taken for granted, even abstracting from all the problems we have previously raised. Even when there is a complete absence of nonpecuniary advantages associated with factor employment decisions and when there are no externalities, market prices need not represent true social opportunity costs. Consider, for example, the type of illustration that has gained some prominence in the economic analysis of law (Posner, 1977, p. 125). Suppose that people are risk-neutral and that, unless precautions are undertaken, an accident may occur between A and B which has an expected loss of $60. If A can avoid the accident by spending $45 on avoidance resources and B by spending $50 on different resources, it may seem that liability ought to be placed on A. This would appear to create incentives for the minimization of the sum of expected accident and accident prevention costs. From a global perspective, however, it is not at all clear that this solution is the most efficient. The general equilibrium value of A’s resources might actually be $50 and B’s only $45. If this is the case then the truly efficient solution is the reverse of the one initially indicated: liability ought to be placed on B.
To recognize “socially” efficient solutions to nonmarket problems like the above liability-rule illustration requires entrepreneurial alertness. It is the function of entrepreneurs in market contexts to notice when existing prices are “wrong” and hence when unexploited profit opportunities have emerged. Similarly, in a nonmarket context we cannot automatically assume “that the available figures often approximate reasonably the information we really want” (Baumol, 1970). Here too the relevant prices are not always in equilibrium, and to assume that they are is to commit the fallacy of the misplaced concrete. Heuristic tools are not descriptions of reality (Hayek, 1955, pp. 53–63).
This is crucial to understanding why and how unexploited opportunities (inefficiencies) are eventually seized. Profit, as we define the term, is a pure entrepreneurial gain attributable to the possession of superior knowledge, but not imputable to market resources that produce knowledge. It is a residual in excess of the payments to all factors of production. Ultimately, profit must be attributed to some creative spark of insight or perhaps to what Kirzner has called “alertness” (Kirzner, 1973, p. 35). A necessary, although not sufficient, condition for the emergence of pure profit is market inconsistency, i.e., the existence of more than one price for the same good (Kirzner, 1979a, pp. 157–8). This can be in input markets where, for example, in two locations the prices of the same quality apples for apple juice differ by more than transportation costs. There can also be inconsistency between output and factor markets when the price of the output is not exactly exhausted by the prices of the inputs. From this perspective, the coordinating function of the entrepreneur is built on his role as an arbitrageur (Kirzner, 1973, p. 46). In an effort to reap arbitrage profits he reduces the inconsistencies that exist in various markets.
As our discussion in the previous section made clear, the elimination of inefficiencies depends on the existence of superior information. However, why doesn’t the cost of this superior information equal its value? More precisely, why don’t the prices of information-producing resources absorb the pure profits and leave only normal returns? The answer must be that the sellers of these resources underestimate (or overestimate in the case of losses) their value.10 Knowledge that has not yet been produced is difficult to evaluate. In fact, the only way to anticipate correctly the value of such resources is to know perfectly what they will produce (Arrow, 1971, p. 148). In the case of information, however, this is equivalent to saying that it must have already been produced. Under those circumstances, of course, there would be no demand for the resources at all. The essence of the emergence of profits (and losses), then, is the imperfect knowledge about knowledge (Kirzner, 1979b, pp. 137–53).
Until now we have somewhat awkwardly abstracted from entrepreneurial expectations, but, clearly, they are important in uncovering profit or coordinative opportunities. This is especially the case when expectations turn out to be incorrect. Under those conditions, entrepreneurship may be disequilibrating and thus increase market inconsistencies rather than reduce them. In recent years, the most important development in this respect has been the rational expectations framework. Expectations are rational if they are the predictions of the correct economic theory (Muth, 1961, p. 316). This has been usually taken to mean that the subjectively perceived stochastic structure of the economy (or the relevant portion of it) is identical to the objective stochastic structure (Barro and Fischer, 1976, p. 156). Furthermore, the psychologically expected values of the various distributions must be equal to the mathematically expected values (Poole, 1976, p. 464–5). The rational expectations approach is thus an equilibrium theory postulating the existence of perfect stochastic knowledge. As such, it says nothing about the process of acquiring such knowledge or the dissemination of it as soon as some people have acquired it (O’Driscoll, 1979, pp. 153–76). In fact, in a world of rational expecters who have the same attitude toward risk, there can be no tendency or process toward equilibrium whatsoever. Fundamentally, rational expectations theory is objectionable because it postulates what is in fact being investigated. Furthermore, any test of the rational expectations hypothesis must be a joint test of the correctness of the economic theory and of the method by which people make their predictions. Divergence between individuals’ expectations and “stochastic truth” is not even conceivable. The analysis construct itself makes error an impossibility. Far from enhancing our understanding of the formation of expectations, this theory has proven to be an unfortunate diversion from the task.
Attempts rigidly to tie the formation of expectations to some objective structure have their roots in a desire to produce (at least ultimately) fairly precise or detailed stories about processes. In contrast to this “mechanical dynamics,” a truly “expectational dynamics” (Hart, 1951, p. viii) would not produce the same kind of explanation. Instead, it would yield very general insight into the equilibrating and disequilibrating processes. Consider, for example, the perspective on explanation by von Neumann and Morgenstern (1967, p. 44, emphasis added) in another context: “[T]he complete answer to any specific problem consists not in finding a solution, but in determining the set of all solutions.” When expectations are not “objectified” (i.e., reduced to a one-to-one function of objectively specifiable variables) we shall not be able to determine them uniquely. Nevertheless, we may be able to understand in a general way how expectations are formed and hence to determine a set of possible expectations. Even if an economic model of process contained all of the proper variables in the correct relationship to each other, it would still amount to a “mere” explanation of the principle (Hayek, 1967, pp. 15–16). To the extent that the variables in the model are expectations and other subjective magnitudes (“the state of knowledge”) we will not be able to ascertain anything more than ranges of possible values for these variables. As a consequence, the most precise form of process explanation feasible would be that given certain values of the exogenous variables, a certain range of values for the endogenous variables is possible. Often, however, we may not even be able to go this far. The above formulation presupposes the existence of a correct model and the only difficulty lies in plugging in the precise values for the exogenous variables. In most cases, even the parameters or structure of the model will not be known with exactness. In those cases we may be confined to constructing a model of such generality that its only observational power is to exclude certain qualitatively specified events from happening together (Hayek, 1967, pp. 15–16). For example, we may be able to “predict”11 that given the state of expectations likely to persist over the relevant period, an increase in the money supply is incompatible with any increase in the purchasing power of money. We might also be able to predict that, under current conditions, interest rates cannot fall if the rate of increase in the money supply is raised. Other examples can no doubt be found, but the essence of the point is that, due to the subjective elements involved in the processing of objective data, dynamic theories can never be as precise as equilibrium theories in the determination of a unique or small set of outcomes. This reflects not a failing of the theories so much as the complexity of the subject matter.
Austrian insights have long been applied to monetary economics. Among the original marginalists, Menger alone made monetary analysis an integral part of his price theory. His analysis of the origins of money remains the most theoretically complete statement of the problem. Modern theorists have yet to equal much less surpass his achievement (Menger, 1892; 1950, pp. 257–85).
In this section, our concern is chiefly with a more applied problem, the analysis of the inflation process. Austrian monetary work has centered on the impact of monetary shocks on relative process and resource allocation. This is an area in which the profession has executed almost a 360-degree turn. In the late 1920s and early 1930s, the distributional and allocational effects of monetary disturbances were of great theoretical interest to economists. To employ modern terminology, monetary theorists were establishing microfoundations for monetary theory. Hayek’s four lectures (“Prices and Production”) on the subject at the University of London earned him a University chair – testimony to both their brilliance and the subject’s importance to economists (Hayek, 1935).
With the coming of the Keynesian Revolution, however, professional interest turned elsewhere. The hard work done on the microfoundations of monetary theory, distributional effects of monetary shocks, the behavior of capital stocks in monetary disequilibrium, etc., was forgotten. In their stead, the profession inherited a highly aggregated “Keynesian” macro model, which by its very construction and assumptions could not deal with any of the problems that seemed so important but a few years before.
The grand synthesis of Walrasian general equilibrium theory and Keynesian macrotheory produced even more abstract macromodels, which retained nonetheless their highly aggregative structure. Among the abstractions introduced was that of “neutral money.” This assumption by itself ensured that no analytic treatment of the issues mentioned above would be possible (Lutz, 1969; O’Driscoll, 1977, pp. 49–56). Hayek had invented the term “neutral money” and delimited the concept as part of a discussion of the goals of monetary policy. The context was a world of very nonneutral money. A neutral monetary policy was among the policy goals. Yet subsequent neoclassical literature turned the problem upside down by assuming neutrality of money, thereby abstracting from all the problems resulting from money’s nonneutrality. The Austrian (and many would say the Keynesian) message was lost in the process. Monetarist models followed in the same mold, so no real competition on this issue existed.
Rampant inflation in the 1970s brought the allocational and distributional effects of inflation back to the profession’s attention. An inflation-ridden world does not look faintly neoclassical. Thus far the work has been empirically oriented (Bordo, 1979; Parks, 1978; Vining and Elwertowski, 1976). We would argue that not only are theoretical underpinnings needed, but that the work of the Austrians can supply a good deal of the needed analysis.
Writing on money, Austrians have naturally seen the effects of monetary disturbances being mediated through entrepreneurial expectations. For instance, changes in the growth rate of the money stock alter the profitability of investments. Mises, Hayek and others focused on the distributional (“Cantillon”) effects of monetary shocks, together with an integration of capital-theoretic considerations. Entrepreneurs receiving additional credit-money as loans gain purchasing power at the expense of population segments which do not initially benefit from the net additions to credit and money.12 The monetary expansion not only transfers purchasing power but also alters relative prices by depressing real interest rates (even if nominal rates increase for expectational reasons) (O’Driscoll, 1979, pp. 163–8). Investments hitherto unprofitable become profitable, along with complementary investments. But investments in processes geared to higher interest rates will generally be rendered less profitable or even unprofitable (Hayek, 1935, pp. 32–104).
Two crucial points must be emphasized. First, the assumed monetary disturbance has altered cost-price ratios. Entrepreneurs’ expectations change because of relative price and interest rate signals. They need not suffer from any illusions. If, as rational expectations theorists claim, Keynesian macrotheory relied on unexploited profit opportunities, the Austrian analysis of economic fluctuations depends on entrepreneurs exploiting perceived profit opportunities.
The second point concerns the unsustainability of inflation. The changes in the pattern of investment (Hayek’s “material structure of the process of production”) can only be sustained so long as inflation is at least sustained, if not accelerated. (This would, of course, mean accelerating monetary growth.) In order that entrepreneurs can complete new investments, they must continue receiving net additions to purchasing power through net credit creation financed by monetary expansion. A production process involves many complementary investments, each of which is economically viable only if all are completed and maintained. Real capital considerations thus require larger and larger net additions to entrepreneurs’ purchasing power, i.e., accelerating inflation. Accelerating inflation is not, however, sustainable ad infinitum. Hence, investments dependent on continuing inflation are not sustainable. These nonsustainable investments represent malinvestment and capital market discoordination. Intertemporal tastes (e.g., time preference) and intertemporal production plans have not been coordinated (Hayek, 1970, pp. 135–56).
The movement in price averages or indices is epiphenomenal. Observed changes in real activity are a concomitant of the changes in real prices. When Hayek first presented this analysis at the University of London, he held strong empirical hypotheses about the precise nature of the relative-price changes and their impact on the structure or processes of production. He also still employed the average period of production concept. Changes in real interest rates accordingly had definite effects on the length of the production process. As we discussed above, Austrians (including Hayek) abandoned such reasoning shortly after these lectures. But the essence of Hayek’s analysis remains untouched.
Inflationary (and deflationary) shocks are exogenous forces in capital markets, being independent of notional or planned saving and investment. In placing a wedge between intertemporal consumption and investment decisions, monetary disturbances lead to intertemporal discoordination and cyclical fluctuations. The variegated nature of capital, a constantly changing financial structure, and other variations within each cycle make it difficult to make any but the most general statements about cycles in general. However disparate cycles may be, they can still be described as the product of inflation-induced discoordination. Indeed, the theme of discoordination has strong support in recent Austrian and non-Austrian literature (Leijonhufvud, 1975; O’Driscoll, 1979). Some of the recent empirical work has partly vindicated earlier Austrian work on the intensity of the 1929–33 depression (Gallaway and Vedder, 1979). More interaction between the existing theoretical literature (Austrian work being included here) and recent empirical and historical work is needed.
In the course of this survey we have developed the central Austrian themes as logical implications of a coherent research program: to make social phenomena intelligible in terms of human action. This has led us to consider a wide range of economic issues from subjectivity to uncertainty to dynamic processes. The Austrian research program, however, is not fully developed in all of these areas; in fact, in some it is only a suggestion of a way to proceed. Nevertheless, the Austrian tradition offers an alternative path of not fully explored insights to those who have become disillusioned with neoclassical orthodoxy. That alternative rejects both the excessive mechanization of individual choice characteristics of contemporary price theory and the inappropriate aggregation of many quasi-Keynesian models. The Austrian framework is consistent with genuine, nonmechanistic choice and the decomposition of social wholes into the human planning from which they arise. Surely these are still the “noteworthy, perhaps the most noteworthy, problem(s) of the social sciences” (Menger, 1963, p. 146).
1 However, in their effort to reduce significantly the importance of subjective elements, Becker and Stigler assume the homogeneity of tastes (Becker and Stigler, 1977, pp. 76–90).
2 Here we are assuming single-valued expectations.
3 It is important to understand that we are not claiming that the equilibrium state is incompatible with perfect foresight. In this situation there is no logical problem because everything, by definition, is unchanging – including the actions of individuals.
4 The “perfect” foresight assumed in this illustration is perfect in the sense that A knows what B will in fact do if A does such and such, etc. It does not include knowledge of the equilibrium plans. This is because that kind of knowledge would make process impossible since everyone would immediately proceed to the equilibrium, i.e., they would never be in disequilibrium in the first place.
5 By “homogeneous” is meant that the relative frequency of an outcome is invariant with respect to random subsequences.
6 The plans must be consistent with each other and with the objective environment.
7 Actually, we might not recognize it if we saw it. See Machlup (1967, pp. 56–9).
8 This produces some paradoxes similar to those in the Holmes-Moriarty example above.
9 We are indebted to Israel Kirzner for this term.
10 Where the entrepreneur uses only his own resources their opportunity cost must be below their value in the current use. This would seem to imply that profits are really a rent. However, rents exist in equilibrium but profits do not. The entrepreneurial “rent” must arise out of uncertainty. For the view that the entrepreneur, qua entrepreneur, uses no resources, see Kirzner (1973, p. 40).
11 The work “predict” is being used in its technical broad meaning. A model “predicts” when it yields a certain finite number of events as logical implications of that model. Hence models about past events do not forecast but they “predict” or “retrodict.” The discussion in the text presupposes that one goal of process analysis is to clarify actual, historical processes.
12 The existence of contracts is a sufficient but not necessary condition for the distributional effects to emerge.
We gratefully acknowledge the helpful criticisms of Frank Arnold, Israel Kirzner, Lyla O’Driscoll, Mark Perlman, T. K. Rymes, Andrew Schotter and Leland Yeager. We also wish to thank the Scaife Family Charitable Trusts for financial support.
Arrow, K. J. (1971) “Economic Welfare and the Allocation of Resources for Invention.” In D. M. Lamberton (ed.) Economics of Information and Knowledge Middlesex: Penguin Books.
Barro, R. J. and Fischer, S. (1976) “Recent Developments in Monetary Theory.” Journal of Monetary Economics 2: 133–67.
Baumol, W. J. (1970) “Review of Buchanan’s Cost and Choice.” Journal of Economic Literature 8: 1210–11.
Becker, G. S. (1976) The Economic Approach to Human Behavior Chicago: University of Chicago Press.
Becker, G. S. and Stigler, G. J. (1977) “De Gustibus Non Est Disputandum.” American Economic Review 67: 76–90.
Böhm-Bawerk, E. v. (1959) Positive Theory of Capital Trans. G. D. Huncke, South Holland, Ill.: Libertarian Press.
Bordo, M. D. (1979) “The Effects of Monetary Change on Relative Commodity Prices and the Role of Long-Term Contracts.” Working Paper. Ottawa: Carleton University.
Buchanan, J. M. (1979 [1969]) Cost and Choice: An Inquiry in Economic Theory Chicago: University of Chicago Press, Midway Reprints.
Buchanan, J. M. and Di Pierro, A. (1980) “Cognition, Choice, and Entrepreneurship.” Southern Economic Journal 46: 693–701.
Coase, R. H. (1960) “The Problem of Social Cost.” Journal of Law and Economics 3: 1–44.
Dahlman, C. J. (1979) “The Problem of Externality.” Journal of Law and Economics 22: 141–62.
Demsetz, H. (1969) “Information and Efficiency: Another Viewpoint.” Journal of Law and Economics 12: 1–22.
Frydman, R. (1980) “Individually Rational Expectations and Market Behavior.” C. V. Starr Center for Applied Economics Discussion Paper New York: New York University.
Gallaway, L. and Vedder, R. (1979) “The Great Depression: A Tale of Three Paradigms.” Ohio University Economic History Paper no. F-47 Athens, Ohio: Ohio University.
Gordon, D. F. and Hynes, A. (1970) “On the Theory of Price Dynamics.” In E. S. Phelps (ed.) Microeconomic Foundations of Employment and Inflation Theory New York: W. W. Norton.
Grossman, S. J. and Stiglitz, J. E. (1976) “Information and Competitive Price Systems.” American Economic Review 66: 246–53.
Hahn, F. H. (1952) “Expectations and Equilibrium.” Economic Journal 62: 802–19.
Hahn, F. H. (1973) On the Notion of Equilibrium in Economics Cambridge University Inaugural Lecture Cambridge: Cambridge University Press.
Hart, A. G. (1951) Anticipations, Uncertainty, and Dynamic Planning New York: Augustus M. Kelley.
Hayek, F. A. (1935) Prices and Production (2nd ed.) London: Routledge & Kegan Paul.
Hayek, F. A. (1941) The Pure Theory of Capital Chicago: University of Chicago Press.
Hayek, F. A. (1948a) “The Use of Knowledge in Society.” In Individualism and Economic Order Chicago: University of Chicago Press.
Hayek, F. A. (1948b) “The Meaning of Competition.” In Individualism and Economic Order Chicago: University of Chicago Press.
Hayek, F. A. (1948c) “Economics and Knowledge.” In Individualism and Economic Order Chicago: University of Chicago Press.
Hayek, F. A. (1955) The Counter-Revolution of Science Glencoe, Ill.: The Free Press.
Hayek, F. A. (1967) “Degrees of Explanation.” In Studies in Philosophy, Politics and Economics Chicago: University of Chicago Press.
Hayek, F. A. (1970) “Price Expectations, Monetary Disturbances and Malinvestments.” In Profits, Interest and Investment New York: Augustus M. Kelley.
Hayek, F. A. (1978) “Competition as a Discovery Procedure.” In New Studies in Philosophy, Politics, Economics and the History of Ideas Chicago: University of Chicago Press.
Hicks, J. R. (1976) “Some Questions of Time in Economics.” In A. M. Tan, F. M. Westfield and J. S. Worley (eds.) Evolution, Welfare and Time in Economics Lexington, Mass.: Lexington Books, D. C. Heath & Co.
Hicks, J. R. (1979) Causality in Economics New York: Basic Books.
Hurwicz, L. (1973) “The Design of Mechanisms for Resource Allocation.” American Economic Review, Papers and Proceedings 63: 1–30.
Kirzner, I. M. (1966) An Essay of Capital New York: Augustus M. Kelley.
Kirzner, I. M. (1973) Competition and Entrepreneurship Chicago: University of Chicago Press.
Kirzner, I. M. (1979a) “Alertness, Luck, and Entrepreneurial Profit.” In Perception, Opportunity, and Profit Chicago: University of Chicago Press.
Kirzner, I. M. (1979b) “Knowing about Knowledge: A Subjectivist View of the Role of Information.” In Perception, Opportunity, and Profit Chicago: University of Chicago Press.
Knight, F. H. (1928) “A Suggestion for Simplifying the Statement of the General Theory of Price.” Journal of Political Economy 36: 353–70.
Knight, F. H. (1971 [1921]) Risk, Uncertainty, and Profit Chicago: University of Chicago Press.
Lachmann, L. M. (1971) The Legacy of Max Weber Berkeley, Cal.: Glendessary Press.
Lachmann, L. M. (1977 [1956]) Capital and Its Structure Kansas City: Sheed, Andrews & McMeel.
Lachmann, L. M. (1978) “An Austrian Stocktaking: Unsettled Questions and Tentative Answers.” In L. M. Spadaro (ed.) New Directions in Austrian Economics Kansas City: Sheed, Andrews & McMeel.
Lakatos, I. (1970) “Falsification and the Methodology of Scientific Research Programmes.” In I. Lakatos and A. Musgrave (eds.) Criticism and the Growth of Knowledge Cambridge: Cambridge University Press.
Leijonhufvud, A. (1975) “Costs and Consequences of Inflation.” UCLA Department of Economics Discussion Paper no. 58, Los Angeles: UCLA.
Lipsey, R. and Lancaster, K. J. (1957) “The General Theory of Second Best.” Review of Economic Studies 24: 11–32.
Littlechild, S. C. (1977) Change Rules, O.K.? Birmingham: University of Birmingham.
Loasby, B. J. (1976) Choice, Complexity, and Ignorance Cambridge: Cambridge University Press.
Lutz, F. A. (1969) “On Neutral Money.” In Streissler, E. (ed.) Roads to Freedom: Essays in Honour of Friedrich A. von Hayek New York: Augustus M. Kelley.
Machlup, F. (1967) Essays in Economic Semantics, New York: W. W. Norton.
Menger, C. (1892) “On the Origin of Money.” Trans. C. A. Foley. Economic Journal 2: 238–55.
Menger, C. (1950 [1871]) Principles of Economics Trans. J. Dingwall and B. F. Hoselitz Glencoe, Ill.: The Free Press.
Menger, C. (1963 [1883]) Problems of Economics and Sociology (ed.) L. Schneider and trans. F. J. Nock Urbana, Ill.: University of Illinois Press.
Mises, L. v. (1966) Human Action: A Treatise on Economics (3rd ed.) New York: Henry Regnery & Co.
Morgenstern, O. (1972) “Descriptive, Predictive, and Normative Theory.” Kyklos 25: 699–714.
Morgenstern, O. (1976 [1935]) “Perfect Foresight and Economic Equilibrium.” Zeitschrift für Nationalökonomie 6 (part 3). Trans. F. H. Knight in A. Schotter (ed.) Selected Writings of Oskar Morgenstern New York: New York University Press.
Muth, J. F. (1961) “Rational Expectations and the Theory of Price Movements.” Econometrica 29: 315–35.
O’Driscoll, G. P. Jr. (1977) Economics as a Coordination Problem: The Contributions of Friedrich A. Hayek Kansas City: Sheed, Andrews & McMeel.
O’Driscoll, G. P. Jr. (1979) “Rational Expectations, Politics and Stagflation.” In M. J. Rizzo (ed.) Time, Uncertainty, and Disequilibrium: Exploration of Austrian Themes Lexington, Mass.: Lexington Books, D. C. Heath & Co.
Parks, R. W. (1978) “Inflation and Relative Price Variability.” Journal of Political Economy 86: 79–95.
Poole, W. (1976) “Rational Expectations in the Macro Model.” Brookings Papers on Economic Activity 2: 463–514.
Posner, R. A. (1977) Economic Analysis of Law (2nd ed.) Boston: Little, Brown.
Richardson, G. B. (1960) Information and Investment Oxford: Oxford University Press.
Rizzo, M. J. (1979) “Disequilibrium and All That.” In M. J. Rizzo (ed.) Time, Uncertainty, and Disequilibrium: Exploration of Austrian Themes Lexington, Mass.: Lexington Books, D. C. Heath & Co.
Rothbard, M. N. (1962) Man, Economy, and State, 2 vols. Princeton, New Jersey: D. Van Nostrand.
Shackle, G. L. S. (1969) Decision, Order, and Time in Human Affairs (2nd ed.) Cambridge: Cambridge University Press.
Stigum, Brent P. (1969) “Competitive Equilibrium under Uncertainty.” Quarterly Journal of Economics 83: 533–561.
Vining, D. R. Jr. and Elwertowski, T. C. (1976) “The Relationship between Relative Prices and the General Price Level.” American Economic Review 66: 699–708.
von Neumann, J. and Morgenstern, O. (1967 [1944]) Theory of Games and Economic Behavior New York: John Wiley & Sons.
Weston, J. F. (1950) “A Generalized Uncertainty Theory of Profit.” American Economic Review 40: 40–50.
Weston, J. F. (1954) “The Profit Concept and Theory: A Restatement.” Journal of Political Economy 62: 152–70.
Yeager, L. B. (1973) “The Keynesian Diversion.” Western Economic Journal 11: 150–63.