Man, Economy, and Liberty

7

Welfare Economics:
A Modern Austrian Perspective

Israel M. Kirzner

Among the most notable of Murray Rothbard’s many contributions to the literature of modern Austrian economics, is surely the major paper on utility and welfare theory that he wrote for the 1956 Mises Festschrift.1 This writer can personally attest to the excitement engendered by the lucid manner in which this paper deployed Austrian insights to illuminate fundamental theoretical issues (concerning which contemporary economics was floundering), and by the characteristic erudition which Rothbard poured into that single essay. Whether or not one fully accepted Rothbard’s conclusions, it was impossible not to glimpse the power of consistent Misesian thinking which that paper so excellently exemplified. The present paper, written thirty years later, seeks to reexamine a small part of the terrain covered by Rothbard’s essay. In offering a modern Austrian perspective on welfare economics we shall be emphasizing some of the same basic Austrian tenets that Rothbard so rightly insisted on thirty years ago. While our perspective may not entirely dovetail with some of Rothbard’s conclusions, we venture to hope that our observations concerning welfare economics be judged to be in the same subjectivist, methodologically individualistic tradition that Rothbard’s work has so valuably carried forward for so many years.

Some Observations Concerning Welfare Economics

Welfare economics has, in its numerous incarnations, sought to offer criteria by which it might be possible scientifically to evaluate the economic merits of specific institutions, pieces of legislation or events. Such evaluation would have to transcend the narrow economic concerns of specific individuals whose interests might be involved, and to express, somehow, a perspective flowing from the economic interests of all individuals in society. As we shall see, Austrian economists have been particularly sensitive to the difficulties that must beset such an undertaking. Indeed, many of the difficulties have been recognized again and again by the economics profession at large, and it is for these reasons, of course, that welfare economics has undergone so many attempted reconstructions “from the ground up.”

We shall briefly survey the more important of these attempts from a perspective that seeks consistently to apply the following (related) Austrian concerns: (a) methodological individualism: we shall refuse to recognize meaning in statements concerning the “welfare of society” that cannot, in principle, be unambiguously translated into statements concerning the individuals in society (in a manner which does not do violence to their individuality); (b) subjectivism: we shall not be satisfied with statements that perceive the economic well-being of society as expressible in terms (such as physical output) that are unrelated to the valuations and choices made by individuals; (c) an emphasis on process: we shall be interested in the economic well-being of society not merely in terms of its level of economic well-being (however defined), but also in regard to the ability of its institutions to stimulate and support those economic processes upon which the attainment of economic well-being depends.

Welfare Economics—Some Highlights of its Past

a) During the period of classical economics it was, of course, taken for granted that a society was economically successful strictly insofar as it succeeded in achieving increased wealth. Adam Smith’s Inquiry into the Nature and Causes of the Wealth of Nations expressed this approach to the economics of welfare simply and typically. It was taken for granted that a given percentage increase in a nation’s physical wealth (with wealth often seen as consisting of bushels of “corn”) meant a similar percentage increase in the nation’s well-being. From this perspective a physical measure of a nations’ wealth provides an index of that nation’s economic success, regardless of its distribution. A bushel of wheat is a bushel of wheat. Clearly this notion of welfare offends the principles of methodological individualism and subjectivism; it was swept away by the marginalist (subjectivist) revolution of the late nineteenth century.

b) Marshall and Pigou sought to preserve certain central elements of the classical approach, while avoiding the trap which sees well-being as identified with (or directly proportional to) physical wealth itself. They focused attention not on goods themselves, but on the utility of those goods. In principle a nation’s physical wealth, given its pattern of distribution, corresponded to a given level of aggregate utility. Moreover they believed this aggregate to be measurable, in principle, by the “measuring rod of money.” They sensed no problem in conceiving of “aggregate utility”; they thought of utility as something that could be compared and aggregated across individuals. They certainly did not see utility as associated uniquely with an individual act of choice; rather they saw it as a kind of psychological shadow that closely followed physical wealth. Its central advantage over wealth, as an index of well-being, was that it incorporated the refinement of diminishing marginal utility. It was no longer acceptable to consider a bushel of wheat to be identical, welfare-wise, with each other bushel of wheat; the margin of consumption by the individual must be considered. But it was still considered valid to treat one dollar’s worth of utility as entirely equivalent to a second dollar’s worth of utility.

This approach to welfare economics is clearly unacceptable to economists who have absorbed the Misesian (and Rothbardian) lessons concerning the true meaning of utility in economic analysis. Utility is, for Austrians, not a quantity of psychological experience, it is merely an index of preferability as expressed in acts of choice. To attempt to aggregate utility is not merely to violate the tenets of methodological individualism and subjectivism (by treating the sensations of different individuals as being able to be added up); it is to engage in an entirely meaningless exercise: economic analysis has nothing to say about sensations, it deals strictly with choices and their interpersonal implications.

c) The approach to welfare economics that has, of course, been central to economics for the past half-century, is that which revolves round the notion of Pareto-optimality. A change is seen as enhancing the economic well-being of society if it renders some of its members better-off (in their estimation) without rendering any others worse off. This approach certainly avoids the problems of interpersonal comparisons of utility, and would thus seem to be consistent both with the methodological individualism and with the subjectivism that Austrians insist upon. Several points need, however, to be noticed.

While the notion of Pareto-optimality is indeed concerned with the individual members of society it nonetheless reflects a supra-individual conception of society and its well-being. After all, a Pareto-optimal move is considered to advance the well-being of society—considered as a whole. Otherwise it is not at all clear what is added (to the bald observation that the change is preferred by some and objected to by none) by the judgement that the move is “good for society.” Indeed the Pareto-criterion turned out to become an integral element in the development of the idea that society faces an “economic problem”—that of allocating its resources among its competing goals, in the most efficient manner. Societal inefficiency in resource allocation came to be identified with sub-optimality according to the Pareto criterion. Now this notion of society facing its economic problem in the resource-allocation sense arose, as is well-known, as an extension of the concept of individual economizing behavior that was articulated so definitively by Lionel Robbins in 1932.2 But, as has before been recognized, this extension is in fact an illegitimate extension, not at all faithful to the spirit of Robbins’s formulation. Robbins was concerned to identify the economic problem facing the individual. It is the individual who has goals and who deliberately deploys his perceived resources in order to most efficiently achieve his goals, as far as is possible. To transfer this important concept of individual allocative choice, to society as a whole is, at best, to engage in metaphor. Society, as such, neither possesses goals of its own nor deliberately engages in allocative choice. Insofar as the idea of Pareto-optimality came to reinforce the faulty and misleading notion of society’s “economic problem,” it was part of an approach to the analysis of economic welfare that fell grievously short of consistent adherence to the principle of methodological individualism.

Hayek and the Critique of Welfare Economics

It was against this mainstream notion of society and its purported allocative problem that Hayek’s famous 1945 paper3 was directed. Hayek’s attack might, it is true, be seen as not being primarily against the welfare notion that was embedded in the idea of society’s economic problem. Hayek focused on the circumstance of dispersed knowledge. The relevant information that “society” would have to possess in order to solve its economic problem is widely dispersed. Society is thus simply not in a position to address its supposed economic problem (even if, for the sake of discussion, this societal allocative task could be held to be meaningful). Hayek’s critique might thus be seen as emphasizing the problems obstructing the practical solution of a nation’s economic problem, rather than as a critique of the standard conception of that problem itself. But Hayek’s paper constituted, nonetheless, a profound—if indirect—critique of the very meaningfulness of societal efficiency as developed, for example, in the Paretian context.

For once it is recognized that the relevant information is inevitably and definitely dispersed among many minds, it is impossible to avoid the conclusion that the notion of social efficiency is correspondingly devoid of meaning. Social efficiency must refer to the extent to which the allocation of social resources corresponds to the priorities implied in the relative urgencies of social goals. But in order for the notion of “social resource” to be meaningful, and in order for the notion of “relative urgencies of social goals” to be meaningful, it must, at least in principle, be possible to imagine a single mind to which the relevant arrays of social resources and social objectives are simultaneously given. Hayek’s insight concerning dispersed knowledge was, in effect, to deny such a possibility. Thus dispersed knowledge turns out to be not merely a phenomenon that constitutes a practical difficulty with which would-be planners must grapple; it turns out to be a phenomenon (not necessarily the only one) that robs the very concept of social efficiency of its meaningfulness, even in principle. To choose, presupposes an integrated framework of ends and means; without such a presumed framework allocative choice is hardly a coherent notion at all.4 Hayek’s insight into the subjectivism of knowledge and information has thus decisively dislodged the foundations of Paretian welfare economics, at least insofar as those foundations have been held to support the concept of social choice and social efficiency. (More recent extensions by Hayek and others of this subjectivism of information to encompass also Polanyi’s idea of “tacit knowledge”—knowledge incapable of being deliberately communicated to others—have rendered these damaging implications for standard welfare economics even more destructive.)5

Coordination as a Hayekian Welfare Criterion

Several writers have, pursuing the implications of these Hayekian insights, seen the concept of “coordination” as offering a normative yardstick consistent with these subjectivist and methodologically individualistic insights.6 As discussed, the notion of social choice (and thus of the efficiency of such choice) has been fatally undermined (except at the level of metaphor). If Jones (who prefers Smith’s food to his own enjoyment of a day’s leisure) fails to trade with Smith (who prefers the labor of Jones over his own food), we may not be able to say that society has failed to efficiently allocate the food and labor time among Jones and Smith—but we could surely still say that Jones and Smith have failed to coordinate their activities and their decisions. It seems plausible and intuitively appealing to perceive coordination—permitting each agent to achieve his goals through the simultaneous satisfaction of the goals of the other agent—as constituting a desideratum transcending the individual goals of the respective agents. Failure to achieve coordination might thus be seen as a failure of the social apparatus to achieve a supra-individual result—but such a judgment relies not at all on any notions inconsistent with subjectivism or with methodological individualism.

It is of course true that the fulfillment of the coordination norm appears to be formally equivalent to the fulfillment of the Paretian welfare criterion. Any sub-optimal situation (in the Paretian sense) clearly corresponds to the failure of a pair of potential market participants to trade with one another on feasible, mutually attractive terms—in other words, it corresponds to a failure to achieve coordination. But, unlike the Paretian norm, the coordination norm escapes interpretation as a yardstick for social efficiency in social-allocative choice. Coordination does not refer to the well-being achieved through its successful attainment; it refers only to the dovetailing character of the activities that make it up.

Thus Hayek’s emphasis on the dispersed character of knowledge appears to provide not merely the definitive critique of standard Paretian welfare economics, but also the basis for an alternative normative yardstick, one thoroughly consistent with the tenet of methodological individualism. Scope for this new normative yardstick is provided precisely by the circumstance of dispersed knowledge. Fragmented knowledge is responsible for activities that are not mutually coordinated. The “social” problem faced by Hayek’s economic society is precisely that of overcoming the discoordination to be expected to flow out of such fragmentation. There is a deeper issue here. If one abstracts from the fragmented character of information, if one treats all existing information as if it were known to all market participants, one is, of course, abstracting from the possibility of discoordinated activities. With the Hayekian “economic problem” assumed to be out of the way, in this fashion, it might seem that the standard (Paretian) economic problem comes back into its own, invulnerable to Hayekian strictures. The problem facing society would, on such assumptions, appear to reduce to that of achieving Paretian optimality in respect of the relevant social objectives, in the face of its limited resources. But, surely, if we assume away the dispersed character of information, the standard economic problem facing society presents no challenge at all. If we can assume that what is known to one is known to all, then (averting our gaze from the remaining quibbles which the methodological individualist might have against the concept of social efficiency) it seems difficult to imagine the possibility of any social allocation of resources that might be pronounced socially inefficient. Given perfect mutual knowledge it appears obvious that all possible Pareto-optimal moves must have already been implemented. To imagine otherwise would be to imagine that agents deliberately refrain from taking advantage of available opportunities known by them to exist. Knowledge of all such opportunities, and knowledge of all relevant transaction costs, must appear inevitably to lead to Pareto-optimality (given these transaction costs)—achieved either through market activity or through centralized organization (with this later choice itself determined by comparison of the respective transaction costs). Thus Hayek’s insights concerning fragmented knowledge might appear to provide not merely a critique of standard welfare criteria, and also a substitute yardstick (in terms of the coordination norm)—they might appear at the same time to salvage welfare economics from the extinction to which it would be doomed by the inevitability of perpetual optimality. But the situation is not quite so simple.

Hayek in the Panglossian World

The truth is that many of the observations made in the preceding sections of this paper might seem to be vulnerable to serious challenge. Such challenges, it would seem, can be launched at several distinct levels—with the challenges stemming precisely from the paralysis arising from the inevitability optimality. On the one hand it might appear that the circumstance of fragmented knowledge does not salvage welfare economics from the extinction spelled by perpetual optimality. Further it might be argued that Hayek’s insights in fact deepen the perplexities created by such Panglossian concerns. We shall in the present section develop these challenges. In subsequent sections we shall rebut these challenges, showing how the observations made in the preceding sections in regard to Hayekian welfare economics can be defended (despite the challenges developed in the present section). Moreover we shall use our discussion to point out a novel sense in which “coordination” offers a normative criterion that escapes Panglossian paralysis. (It will be in the context of this latter discussion that we shall deploy the third Austrian tenet referred to at the outset of this paper, that of maintaining a concern with processes rather than exclusively with states of affairs.) We turn now to develop the apparent challenges to Hayekian welfare economics referred to at the outset of this section.

The difficulties that we must face up to, in considering the Hayekian thesis of dispersed knowledge and information, consist in the fact that, from a mainstream perspective, the Hayekian “knowledge problem” might appear not to be a problem at all, in the relevant sense.7 To point out that knowledge is scattered in society is, it might be argued (contrary to our earlier assertions), not necessarily to note that standard welfare analysis is inapplicable—it is merely to point out that such standard welfare analysis is to be carried on in the context of a hitherto unsuspected cost, the cost of ascertaining and of communicating information. Dispersal of knowledge and information indeed introduces new costs for the acquisition of the knowledge necessary for economic choice. But surely the presence of a novel class of costs does not, in principle, render inapplicable the standard criteria for the evaluation of social efficiency.

Moreover, once it is recognized that the fragmentation of information complicates standard welfare analysis without vitiating it, it seems appropriate to point out that the Panglossian paralysis referred to earlier offers as serious threats to a “coordination”-based approach to welfare analysis, as it does to the mainstream approach. After all any discussion of Jones and Smith “coordinating” their activities must refer to a potential for coordination in the context of the relevant resource constraints confronting the respective parties. Surely then, the availability and costliness of information-acquisition must be counted in as part of these “relevant resource constraints.” If engineer Jones, Sr., and farmer Smith can exchange engineering services for food, with mutual gain, it may seem that only a coordination failure could prevent such exchange from taking place. But it will not constitute a coordination failure if Jones, Jr., schoolboy, refrains from enrolling in an engineering program on his graduation from high school if the costs of the training program are too high. Similarly, it might appear, all coordination “failures” attributable to Hayekian knowledge fragmentation, turn out not to be failures at all once one properly considers the cost of searching for the information needed to bridge the dispersed knowledge gaps. If Jones, Sr., and Smith fail to engage in mutually gainful exchange, as a result of knowledge dispersal, they are not, it might be contended, acting sub-optimally, from a social point of view; they are fully taking advantage of each other’s availability in the context of their limited knowledge of each other’s situation. To pronounce this state of affairs to be socially inefficient or “uncoordinated” might seem to be succumbing to a temptation warned against in elementary economics, viz. that of pronouncing welfare judgments without regard to resource scarcities. Participants in an economy can be counted upon to engage in mutually gainful exchange transactions, insofar as their knowledge permits. Moreover, insofar as participants are aware of worthwhile possibilities for learning useful information that may reveal as yet unexploited opportunities for mutual gain, they can surely be counted upon to engage in such useful learning. It does seem, then, that in a world of dispersed information as in a world of omniscience, sub-optimality or states of discoordinatedness cannot be postulated to exist (if one properly includes the costs of information-acquisition).

Indeed it might be contended that it is precisely Hayek’s dispersed information insights that are capable of focusing needed attention on the costs of learning and of knowledge-communication. Once the paralyzing assumption of perfect knowledge has been dropped it becomes impossible to avoid grappling with the economics of learning and communication. Our contention thus far is that, once such economics of learning and communication has been taken into account, Panglossian perpetual optimality paralysis sets in once again. At all times agents will be engaging in the optimal mix of decisions (including decisions to learn and to communicate). No pair of decisions can be pronounced uncoordinated, given the costs of learning.

Dispersed Knowledge, Optimal Ignorance and Genuine Error

We shall discover, however, that these contentions are invalid. The Panglossian paralysis we have found to afflict mainstream welfare economics is not a threat to the Hayekian coordination approach. It is not the case, we shall see, that Hayek’s fragmentation of information does nothing more than to complicate matters through the introduction of a new cost. Rather the dispersal of knowledge creates scope for a genuinely fresh approach to normative analysis. This is so because such dispersal of knowledge necessarily involves not merely new costs (of learning and communication) but also the very real possibility of what we may call “genuine error.” This writer has elsewhere argued8 that genuine error, so often exorcised form economic analysis, in fact deserves a central place in that analysis. Genuine error occurs where a decision-maker’s ignorance is not attributable to the costs of search, or of learning or of communication. In such cases the decision-maker’s ignorance is utter ignorance—i.e., it is a result of his ignorance of available, cost-worthy, avenues to needed information (which includes, of course, the possibility of his being altogether ignorant of the very existence of valuable information). At the level of the individual decision-maker we may describe his activity as having been sub-optimal when he subsequently discovers himself to have inexplicably overlooked available opportunities that were in fact worthwhile. He cannot “condone” his faulty decision-making on the grounds of the cost of acquiring information, since the information was in fact costlessly available to him. He can account for his failure only by acknowledging his utter ignorance of the true circumstances (i.e., of his ignorance of the availability of relevant information at worthwhile low cost). Such utter ignorance cannot be explained in cost-benefit terms; it is simply a given.

Two implications of the phenomenon of utter ignorance, of genuine error, may be noticed. First the injection into economic reasoning of the possibility of genuine error, introduces a degree of “looseness” into our understanding of economic processes that is of great importance. It is no longer true that the configuration of exogenous variables, tastes, resource availabilities and technological possibilities, unambiguously marks out the course of individual activities. This is so because while these data do mark out the optimal opportunities, we cannot be confident that such optimal opportunities will be known to the relevant decision-makers—even if we make provision for deliberate processes of search and learning. We cannot be sure that available processes of search and learning are known to those who might benefit therefrom. The second implication (flowing from recognition of the phenomenon of genuine error) is that we must now recognize the possibility of corrective actions within an economy, that are not to be traced to shifting cost patterns. Corrective action may be set off by the sudden (“entrepreneurial”) discovery by a market participant of a hitherto unperceived opportunity for pure profit. Let us now return to examine Hayek’s dispersed information.

We objected that the introduction of the need for costly search, learning and communication (forced upon us by Hayek’s insight) does not really threaten the mainstream economizing view. The fragmentation of knowledge, we pointed out, merely introduced an additional cost-dimension—that of mobilizing and centralizing scattered bits of information. We now see that the fragmentation of knowledge is likely to affect matters far more seriously and fundamentally. The fragmentation of knowledge injects into the picture scope for genuine error, resulting from utter ignorance. Pursuing once again the line of reasoning introduced earlier in this paper (and subsequently challenged in the preceding section) the circumstance of dispersed and fragmented knowledge compels us not merely to recognize a practical difficulty to be encountered in seeking to address society’s allocative-efficiency problem—this circumstance undermines the very meaningfulness of such a social “economic problem.” Given the scope for genuine error which we see to be implicit in the circumstance of dispersed information, we now see that this circumstance indeed erodes the meaningfulness of the concept of social allocative efficiency. Before we can even begin to contemplate what we may mean by social allocative efficiency we must somehow confront the problem of overcoming that utter ignorance which obstructs the relevancy of the efficiency concept for social policy. It is here that the norm of “coordination” is to be perceived in a fresh light, rather different from that coordination-norm discussed earlier in this paper.

Coordination and Coordination9

We must distinguish carefully between (a) a possible norm of coordination in the sense of a coordinated state of affairs, and (b) a possible norm of coordination in the sense of the ability to detect and to move towards correcting situations in which activities have until now been discoordinated. The distinction between these two possibilities corresponds to the two different meanings of the word “coordination”: the word may refer to the activities being carried out when these activities are indeed dovetailing with one another; alternatively it may refer to the process through which initially clashing, discoordinated activities, are somehow being hammered out in a manner such as to approach a more smoothly dovetailing pattern of activities. The discussion earlier in this paper implicitly referred to coordination only in the first of these two senses. (It is for this reason that we were able to note formal equivalency between the coordination norm and the norm of Pareto-efficiency.) We wish now to draw attention to the possible relevance of the second coordination norm for a modern Austrian approach to welfare economics. Once we have identified genuine error as a culprit responsible for a failure of a society’s economic system to successfully fulfill its functions, we have placed ourselves in a position to appreciate the meaning of this second coordination norm. Absent the phenomenon of utter ignorance, we have seen, our first coordination concept (like its Paretian counterpart) turned out to be of little normative interest. After all, we noted, given the absence of utter ignorance, all activities must be carried on in optimal fashion. Even if some activities are being carried out “erroneously,” because of incomplete information, we saw, we could hardly describe these activities as being sub-optimal or “wrong”—after all, they took advantage of every scrap of information it was judged worthwhile to lay one’s hands on. In this sense the world is, at all times, at a Pareto-optimum, in a state of full coordination—the best of all possible worlds, given the costs of change. But injection of the possibility of genuine error arising out of simple utter ignorance introduces us to the possibility of genuine discoordination—and to the possibility of evaluating the institutional environment in terms of its potential to inspire genuine discovery (of opportunities previously overlooked as a result of utter ignorance). Thus a norm of coordination looms into center stage in the sense of permitting us to ask what potential a society’s economy possesses to inspire such pure discovery of its earlier genuine errors. Such an approach to welfare economics is made possible by our escape from the Panglossian world; that escape was, in turn, made possible by our emphasis on genuine error (arising out of utter ignorance); we have seen in this paper that scope for genuine error is widened most considerably by the circumstance of dispersed and fragmented information identified by Hayek. It is for this reason that we see Hayek’s criticisms of standard approaches to welfare analysis as opening the door, at the same time, towards the possible reconstruction of normative economics along truly Austrian lines, that is, in a manner fully consistent with (a) subjectivism, (b) methodological individualism, and (c) an emphasis on dynamic processes.

Notes

1. Murray N. Rothbard, “Toward a Reconstruction of Utility and Welfare Economics,” in Mary Sennholz, ed., On Freedom and Free Enterprise (Princeton: Van Nostrand, 1956), pp. 224-62.

2. Lionel Robbins, An Essay on the Nature and Significance of Economic Science (London: Macmillan, 1932).

3. Friedrich A. Hayek, “The Use of Knowledge in Society,” American Economic Review 35 (September 1945).

4. Cf. J. M. Buchanan, “What Should Economists Do?” Southern Economic Journal 30 (January 1964).

5. See Friedrich A. Hayek, Law, Legislation and Liberty, vol. 3, The Political Order of a Free People (Chicago: University of Chicago Press, 1979), p. 190.

6. Israel M. Kirzner, Competition and Entrepreneurship (Chicago: University of Chicago Press, 1973), chap. 6; Gerald P. O’Driscoll, Jr., Economics as a Coordination Problem (Kansas City, Kans.: Sheed Andrews and McMeel, 1977).

7. See also Israel M. Kirzner, “Economic Planning and the Knowledge Problem,” Cato Journal 4 (Fall 1984): 407-18.

8. Israel M. Kirzner, Perception, Opportunity and Profit, (Chicago: University of Chicago Press, 1979), chap. 8; idem, Discovery and the Capitalist Process (Chicago: University of Chicago Press, 1985).

9. See also Israel M. Kirzner, “Prices, the Communication of Knowledge, and the Discovery Process,” in Kurt R. Leube and Albert H. Zlabinger, eds., The Political Economy of Freedom: Essays in Honor of F. A Hayek (Munich: Philosophia Verlag, 1984).