Epistemic problems in Hayek’s defence of free markets

Abstract

Friedrich von Hayek’s classical liberalism argued that free markets allow individuals the greatest opportunity to achieve their ends. This paper develops an internal critique of this claim. It argues that once externalities are introduced, the forms of economic knowledge Hayek thought to undermine government action and orthodox utilitarianism also rule out relative welfarist assessments of more or less regulated markets. Given the pervasiveness of externalities in modern economies, Hayek will frequently be unable to make comparative welfarist claims, or he must relax his epistemic assumptions and allow for greater government action than his classical liberalism would wish to accept.

1. Introduction

Friedrich von Hayek’s classical liberalism is commonly associated with one of the strongest and most influential defences of free markets. This defence has become one of the central intellectual pillars of the movement known as neoliberalism and has received increased attention within contemporary political theory and economy. ¹ A range of different and sometimes conflicting arguments can be found in Hayek’s corpus, however, leading to varying interpretations. He has been described as a utilitarian (Barry 1984; Gray 1986; Yeager 1989), a Kantian liberal (Gray 1986; Kukathas 1989), a neo-republican (Irving 2019) and a conservative (Feser 2003). I am not concerned here with reconciling these different interpretations or vindicating any one as best encapsulating Hayek’s thought. Instead, I wish to raise some problems with one influential strand of Hayekian argument for free market institutions and discuss their implications for the epistemic foundations of Hayek’s political economy and classical liberalism.

My focus will be on Hayek’s indirect utilitarian or welfarist defence of markets, which claims that free market institutions allow individuals the greatest opportunity to achieve their ends. ² While the government regulation of markets is frustrated by the dispersed nature of economic knowledge, and by the unknowability of agents’ plans, an open market spontaneously coordinates individuals in pursuit of their varied purposes. This line of argument has been highly influential, particularly among Hayekian-inspired economists who have generally taken welfarist approaches (e.g., Kirzner 1963, 1985; O’Driscoll and Rizzo 2002; Cordato 2007). Rather than providing a clear defence of free market institutions, I argue that Hayek’s assumptions about the inaccessibility of economic knowledge very often rule out comparative welfare assessments of more or less regulated markets. I therefore aim to highlight a tension between Hayek’s utilitarian arguments and the epistemic assumptions of his theoretical framework, arguing that if we take the latter seriously then we should be frequently unable to make welfarist claims either for or against the free market.

This critique is internal to Hayek’s political economy, at least on an indirect utilitarian interpretation, and is connected to the problem of externalities. ³ My argument concerning these mostly environmental externalities differs from previous debates over whether Hayek’s epistemic framework offers a purely property-based approach (Anderson and Leal 2001; Pennington 2001, 2005; Cordato 2004, 2007), or whether this underestimates the scale of externality problems and the effectiveness of government institutions (Greenwood 2007, 2015; O’Neill 2012; Benson 2018, 2019b). ⁴ Instead, I claim that the introduction of external costs forces Hayek to make assessments of welfare which are prohibited by his epistemic assumptions, with the result that he is unable to know whether free market institutions are superior in his own utilitarian terms to more regulated alternatives. Given the pervasiveness of externalities in modern economies, Hayek will either be frequently unable to make comparative welfarist claims or must relax his strict assumptions about the inaccessibility of economic knowledge. The latter, however, would likely allow for greater knowledge on the part of policymakers and therefore greater forms of government action than his classical liberalism would wish to accept.

I therefore aim to show that one of the strongest and most influential theorists of free markets very often cannot endorse the welfare advantages of generally unregulated market institutions. The paper proceeds as follows. The next section reconstructs Hayek’s indirect utilitarian argument for free market institutions and highlights the epistemic assumptions on which it is based. I do not claim that this interpretation best represents all of Hayek’s thought but will argue that a plausible and influential utilitarian argument can be identified in his work. Section 3 then introduces the problem of externalities and contemporary debates concerning Hayek’s approach to them, before the paper moves to its core argument. Section 4 explores the limits to welfare judgements in Hayek’s thought in the context of externalities, while section 5 argues that these limits undermine Hayek’s utilitarian endorsement of free markets. Section 6 then considers potential replies and section 7 concludes by discussing the broader implications of this critique.

2. Hayek’s Indirect Utilitarian Defence of Free Markets

Despite being associated with one of the strongest defences of free markets, Hayek’s (2011, 2013) classical liberalism is not naive about the need for state institutions. He recognized that the state and its legal system have an unavoidable role in establishing and maintaining the market order and he understood that this would influence the functioning of markets. Hayek (1944: 84) therefore believed that the debate over state non-interference obscured the fact that every state ‘must act and every action of the state interferes with something’. While he did not advocate for a completely ‘free’ market then, a question remains over the form and extent of state interference, and his classical liberalism provided many arguments for a system where market prices are allowed to fluctuate with minimal government intervention. This is true of his utilitarian defence of markets, which claims they can greatly increase the welfare of an individual, ‘but only if the prices he can get are determined solely by market forces and not by the coercive powers of government’ (Hayek 1978: 62, original emphasis). I will therefore use the term ‘free market’ as shorthand for the kind of lightly regulated markets favoured by Hayek’s classical liberalism. ⁵

While utilitarian and welfarist arguments can be detected in Hayek’s work, his political economy must be clearly distinguished from both orthodox utilitarianism and neoclassical welfare economics. His work cannot be seen as purely consequentialist, for instance, as he maintained that it would be a ‘misunderstanding’ to interpret him as supporting liberty only for reasons of ‘expediency’ (Hayek 2011: 52–53). However, he also does not take liberty to be an ‘ethical presupposition’ and states that any convincing defence of liberty needs to show that it is the ‘source and condition of most moral values’ (Hayek 2011: 52–53). At least for the purposes of argument, then, Hayek does not wish to take the value of liberty for granted but rather to show that it is a prerequisite for achieving other purposes. He aims to show that our ‘faith in freedom’ rests on the ‘belief that it will, on balance, release more forces for the good than for the bad’ (Hayek 2011: 83).

Hayek does not therefore deny intrinsic value to liberty. He also appeals to non-consequentialist lines of argument, such as Kantian universalization, in his work (e.g. Hayek 1966). Nevertheless, many of his arguments take a consequentialist form. For example, Hayek’s (2011: 98, 375) rejection of socialism and the welfare state often claimed that it was simply a bad means for promoting the welfare of the poorest, and that a benefit of market economies was that they would increase their position to a greater extent than any short-term redistribution of wealth. The market is also claimed to be an ‘impersonal process which brings about a greater satisfaction of human desires than any deliberate human organisation could achieve’ (Hayek 2013: 227), and a similar instrumental reasoning is applied to political institutions. Democracy, for instance, is argued to be ‘probably the best method of achieving certain ends, but not an end in itself’ (Hayek 2011: 170).

Even this consequentialist reasoning cannot be straightforwardly interpreted in orthodox utilitarian terms, given Hayek’s explicit rejection of utilitarianism as involving a constructivist fallacy. As others have argued, however, the target of his critique is the particular utilitarianism of Jeremy Bentham and starts by narrowing focus away from the general function of rules and institutions to a hedonic form of utilitarianism (Gray 1986; Kukathas 1989; Hayek 2013: 184–186). Hayek also does not aim at rejecting the goal of increasing welfare, but rather objects to this form of utilitarianism on predominantly epistemic grounds. A felicific calculus of pleasure and pain ‘presupposes that all the particular individual effects of any one action can be known by the acting person’, but our necessarily bounded rationality makes the attainment of such knowledge impossible (Hayek 2013: 186). In both its act and rule formulation, utilitarianism is argued to involve the false assumption that the acting person can have knowledge of all their effects in a large and complex society. These epistemic problems do not, however, undermine the aim of promoting welfare as such. In fact, Hayek (2013) ends his discussion of utilitarianism by concluding:

We may of course aim at the ‘greatest happiness of the greatest number’ if we do not delude ourselves that we can determine the sum of this happiness by some calculation, or that there is a known aggregate of results at any one time. What the rules, and the order they serve, can do is no more than to increase the opportunities for unknown people. If we do the best we can to increase the opportunities for any unknown person picked at random, we will achieve the most we can, but certainly not because we have any idea of the sum of utility of pleasure which we have produced. (Hayek 2013: 190)

Hayek does not therefore see his critique as requiring that we reject the goal of promoting welfare, but rather as claiming that this must be done by more indirect means than that suggested by the English utilitarians. We cannot design specific interventions which maximize utility but must rely on a more abstract order of rules which will allow individuals the opportunity to promote their welfare. It is for this reason that those who have interpreted Hayek as offering utilitarian arguments have tended to describe him as an ‘indirect utilitarian’ (Barry 1984; Gray 1986). Gray (1981, 1986), for instance, traces this indirect consequentialism back to the influence of Hume on Hayek’s work. Hume thought the laws of justice were an indispensable condition for the promotion of human welfare and their ultimate justification therefore had a utilitarian component. Such laws, for Hume as well as Hayek (2013: 183), cannot be simply broken at particular moments for reasons of expediency, but will only have beneficial effects to the extent they are followed in all cases. Whatever its origins, it is a form of indirect rather than direct utilitarianism which can be identified in Hayek’s work.

The understanding of utilitarian welfare found in Hayek is also unorthodox. His stated test is not the maximization of happiness, but rather the maximization of the opportunities or chances of any unknown person picked at random from achieving their ends (Hayek 1966: 613; 1978: 62, 183–184; 2013: 171, 190, 274, 288). ⁶ Hayek (2002: 14) argued that ‘it obviously makes sense to try to create conditions under which any randomly selected individual has prospects of pursuing his goals as effectively as possible, even if we cannot predict which particular individuals will benefit thereby and which will not’. ⁷ We can therefore judge which institutional arrangements best promote the chances of individuals, while avoiding the kinds of utilitarian calculations required not only by traditional utilitarianism but also neoclassical welfare economics. ⁸ The Kaldor–Hicks Compensation Test (or the potential Pareto standard), for instance, is commonly applied as a standard for economic efficiency and states that situation A is an improvement over B if the winners can compensate the losers and still be left better off (Hicks 1939; Kaldor 1939). This test requires similar cost-benefit calculations to those of orthodox utilitarianism, however, and is therefore also ruled out on Hayek’s view.

Hayek applies his indirect utilitarian test to general rules of conduct, as in the above quoted discussion of utilitarianism (2013: 190), but also to economic policy (1966: 613) and to spontaneous orders and the market (1978: 62–63, 183–184; 2002: 14–15; 2013: 274). The reason free market institutions better pass this test than more regulated market economies are predominantly epistemic. Rejecting the full information assumptions of neoclassical welfare theorems, Hayek argues that which institutional arrangement ‘is likely to be more efficient depends mainly on the question under which of them we can expect that fuller use will be made of the existing knowledge’ (Hayek 1937; 1945: 521). ⁹ Questions of efficiency therefore turn on the kinds of knowledge relevant to economic coordination and those groups most likely to possess it. While Hayek (1945: 521; 2011: 494) maintained that when it came to the sciences ‘a body of suitably chosen experts may be in the best position to command all the best knowledge available’, this overlooked ‘an even greater store of knowledge of special circumstances that ought to be taken into account’. It is the nature of this latter local knowledge which then frustrates the effectiveness of government institutions.

Firstly, this knowledge does not refer to general principles or laws but ‘the particular circumstances of time and space’ and is therefore only possessed by certain individuals (Hayek 1945: 521; 1978: 136). The conditions of certain resources or the costs of certain production processes, for instance, are only known to on-the-spot actors with direct access to local conditions, and it is also only individuals who know the relative importance of such things for their ends (Hayek 1945: 520). Local knowledge is therefore dispersed throughout the economy and includes information about resources and production processes, as well as the values and plans of individual actors. Secondly, the context dependence of local knowledge also means that it is contingent and open to change. A government body would therefore not only need to centralize large amounts of highly dispersed information, but continually update it. Thirdly, certain forms of local knowledge are tacit in nature (Hayek 1945). Tacit knowledge is practical knowledge embodied in experience and know-how, and cannot therefore be communicated to a centralized authority as it is only learnt ‘through social action, such as exchange or exercising a skill’ (Horwitz 1992: 198). ¹⁰ As Polanyi (1997) put it, the problem is that we ‘know more than we can tell’. It is therefore the local, dispersed and tacit nature of economic knowledge which Hayek takes to undermine the efficiency of government action.

Markets, alternatively, are said to spontaneously coordinate economic activity without ‘communicating all this knowledge to a central board’ (Hayek 1945: 525). Markets involve a system of property and exchange which allow individuals and firms to use their local knowledge in the pursuit of their own welfare. The price system then coordinates the actions of individuals as they influence the formation of market prices. By reflecting the relative supply and demand of goods, fluctuating prices act as ‘knowledge surrogates’ which allow individuals to adjust to the actions of others without needing large bodies of dispersed and tacit knowledge (Horwitz 2004: 314). An open market therefore allows individuals to utilize their local knowledge in pursuit of their goals, provides the information necessary for coordinating their plans with others, and forces them to bear the costs associated with their actions. Competition then also allows for the discovery of how best to satisfy the unknown ends of consumers, as firms try out alternatives and learn about their success through profit and loss (Hayek 2002; also see Kirzner 1985). ¹¹ In sum, market coordination provides opportunities for individuals to achieve their ends, and we cannot improve on these opportunities without forms of economic knowledge which are inaccessible to an external observer.

What is important for this paper’s purposes is that Hayek’s argument, as originally formulated, does not require him to measure or quantify the benefits of markets, in the ways required by orthodox utilitarianism and neoclassical policymaking. According to Hayek, the promise of markets is that we can judge them to improve the chances of any unknown person achieving their ends without making detailed welfare calculations and possessing the large amounts of economic knowledge these calculations entail. We can therefore expect markets to pass his indirect utilitarian test, in the long run at least, despite our ignorance. ¹² Knowing if forms of government intervention into the market will pass this test, alternatively, would require access to forms of knowledge which are inaccessible to a central authority. When evaluated from the perspective of Hayek’s indirect utilitarianism then, we should prefer a free market to a market economy with significant state interference with the price mechanism. As Hayek (1978: 62, original emphasis) put it, a market actor will ‘improve the chances of any member of his society, taken at random, as much as possible – but only if the process he can get are determined solely by market forces and not by the coercive powers of government’.

Interesting questions remain about the relationship between this theoretical utilitarian argument and the more Kantian, neo-republican or conservative elements of Hayek’s classical liberalism. ¹³ Whatever their answer, we can still consider the force of this line of argument and ask whether there is an effective utilitarian string to Hayek’s bow when it comes to his defence of the free market. This argument has also been highly influential, particularly among Hayekian inspired economists (e.g. Kirzner 1963, 1985; O’Driscoll and Rizzo 2002; Cordato 2007). Take, for instance, the standard of ‘catallactic efficiency’ within the Austrian school of economics which was first proposed by Kirzner. Rejecting the neoclassical standards mentioned above, this view states that ‘efficiency for the social system means the efficiency with which it permits its individual members to achieve their several goals’ (Kirzner 1963: 35). It therefore asks ‘what is the institutional setting that will maximize the extent to which individual members’ can achieve their varied ends?’ (Cordato 1994: 132). The influence of the utilitarian reading of Hayek can be clearly seen here, with the standard closely mirroring Hayek’s indirect utilitarian test. ¹⁴ Outside of the academy, it is also common for supporters of classical liberalism or neoliberalism to believe in the welfare-enhancing benefits of largely unregulated markets, and it is therefore interesting to examine whether one of the intellectual figureheads of these movements offered a workable argument for such a belief. ¹⁵

3. Externality Problems

Hayek makes a theoretical argument that given the nature of economic knowledge, free market institutions provide individuals with the greatest opportunity to achieve their ends. As presented in the previous section, this argument is so far silent on the issue of externalities. It claims that markets leave individuals free to further their goals, while the price system coordinates their actions and forces them to account for their effects on others. Negative externalities, alternatively, represent a failure of this coordination of individual ends, as the consensual exchanges of some impose costs on third parties which they do not take into account. When a factory owner pollutes the land of a local farmer, their pursuit of their plans interferes with the ends of another, and this imposition will not be reflected in the price of their manufactured goods. Environmental externalities, such as air, land and water pollution, as well as problems such as neighbourhood effects, offer perhaps the best-known examples of such issues. ¹⁶

Although not always discussing them in great detail, Hayek (1944: 40; 2011: 481) was well aware of such problems and accepted that they interfered with the effective functioning of the price system. While maintaining that in certain instances such problems were due to the public ownership or common use of resources, and could therefore be resolved through the establishment and enforcement of private property rights (Hayek 2011: 492), he also recognized cases where such solutions were not possible. When it came to the externalities produced by the use of large resources, such as air, water, fisheries, wildlife resources, natural gas and oil, or cases where property was too closely interconnected, such as neighbourhood effects in urban areas, he acknowledged that market prices will not function properly (Hayek 2011: 473–475, 492). In Hayek’s (2011: 481) words, these are problems where ‘the price mechanism operates only imperfectly and does not take into account many things we would wish to see taken into account’. Government policies aimed at helping to improve the performance of the price system may therefore be necessary, and possibly the use of alternative mechanisms all together. For instance, Hayek (1944) writes:

Nor can certain harmful effects of deforestation, or of some methods of farming, or of the smoke and noise of factories, be confined to the owner of the property in question or to those who are willing to submit to the damage for an agreed compensation. In such instances we must find some substitute for the regulation by the price mechanism. But the fact that we have to resort to the substitution of direct regulation by authority where the conditions for the proper working of competition cannot be created, does not prove that we should suppress competition where it can be made to function. (Hayek 1944: 40)

There has been debate over this limited endorsement by Hayek of regulation and non-market mechanisms. Certain ‘free-market environmentalists’, for instance, have argued that such measures are not consistent with Hayek’s broader political economy and that a Hayekian approach to externalities should instead limit itself to the creation and enforcement of private property rights in environmental goods (Anderson and Leal 2001; Pennington 2001, 2005; Cordato 2004, 2007). ¹⁷ Their critics, alternatively, claim that such approaches underestimate the global nature of environmental issues and that Hayek himself might have advocated for a ‘larger degree of regulation’ if he had ‘understood the potential scale of contemporary problems’ (Greenwood 2007: 81-82; 2015: 427). Others, such as diZerega (1992, 1996) and Gamble (2006), suggest that Hayek’s appreciation for spontaneous orders provides principled Hayekian grounds for protecting the complex and self-organizing ecosystems which support human society, while Benson (2018, 2019b) and O’Neill (2012) argue that political institutions are in a better epistemic position to address environmental issues than Hayekians recognize, and that market actors often lack the knowledge needed to act on environmental values. ¹⁸

While the argument I wish to make will centre around externality issues, its concerns are somewhat different from those of these previous debates. Rather than defending a property-based or more interventionist approach, I will instead focus on the problems which externalities create for Hayek’s epistemic framework. Stated briefly, the problem I see is that while Hayek wished to avoid the measurement or quantification of welfare due to its reliance on inaccessible economic knowledge, the introduction of externalities forces him to make welfare judgements in the application of his own indirect utilitarian test which require similar forms of knowledge. In the context of externalities, Hayek’s epistemic assumptions will, in a narrow set of cases, undermine his ability to make absolute welfarist assessments of markets, and in a broader set undermine his ability to make relative assessments of more or less regulated markets. In both cases, I argue, Hayek is unable to determine whether a free market is welfare superior in his own indirect utilitarian terms to more regulated alternatives, and he is therefore frequently unable to make welfarist claims either for or against free market institutions.

4. The Limits to Welfare Judgements in Hayek’s Political Economy

As we have seen, Hayek aims to provide an argument for free markets which does not require that their welfare benefits be measurable or quantifiable. As long as prices are allowed to fluctuate freely and coordinate individuals in pursuit of their ends, markets can be said to pass Hayek’s indirect utilitarian test, without access to large amounts of economic knowledge. Market externalities, however, introduce costs into this assessment. When externalities are present, a market will produce opportunities for those individuals engaged in consensual exchange to achieve their ends, but it will also disrupt the opportunities of third parties. My argument is that the same forms of economic knowledge which Hayek argued to undermine government action and orthodox utilitarianism, also tend to prohibit any precise assessments of these relative benefits and costs for people’s opportunities, and that this frequently frustrates his ability to make welfarist claims either for or against the free market. This section focuses on the first part of this argument. It shows that while a market with an externality will both create and disrupt opportunities for individuals, the amount by which it does this is unknowable from Hayek’s epistemic perspective.

Consider first the benefits which markets are said to provide for those engaged in consensual exchange. While Hayek claims that market coordination will generally provide individuals with the opportunity to achieve their ends, he is explicit in asserting that we cannot know the magnitude of these benefits because we cannot have knowledge of the specific ways markets help individuals to further their purposes. For instance, he writes:

The fact, however, is that in a Great Society in which the individuals are to be free to use their own knowledge for their own purposes, the general welfare at which a government ought to aim cannot consist of the sum of particular satisfactions of the several individuals for the simple reason that neither those nor all the circumstances determining them can be known to government or anybody else. Even in the modern welfare societies the great majority and the most important of the daily needs of the great masses are met as a result of processes whose particulars government does not and cannot know. (Hayek 2013: 170)

For Hayek, markets will have positive welfare effects, but the specific means by which they do this are unknowable and so too are the exact impacts of these means on the ends of individuals. The reason for this is Hayek’s assumptions about the inaccessibility of economic knowledge. An understanding of all the ways that markets allow individuals to further their plans would require access to dispersed and tacit knowledge relating to the workings of all producers and consumers. It would require knowledge of their access to local resources and how they make use of them to meet their or other’s needs. As we have seen, however, Hayek claims that such knowledge is only known to on-the-spot individuals, is open to change, may often involve a tacit component, and cannot therefore be communicated in its totality to any external observer. Similarly, to understand the specific impact of such processes on the chances of individuals achieving their ends would require information about these ends, and this is again said to be known only to individuals themselves. Like knowledge of resources, information about individuals’ plans and goals is dispersed to those individuals and is open to change, and cannot therefore be fully communicated to any observer. The market is therefore said to aim at achieving ‘largely unknown particular ends’ or ‘individual objectives which no one knows in their totality’ (Hayek 2002: 14–15; 2013: 274).

While Hayek therefore believes that markets will increase the opportunities for those engaged in consensual exchange, the nature of economic knowledge means that he cannot make any precise claim about the magnitude of such benefits. To know how markets increase the chances for individuals would require the very same forms of knowledge which Hayek (2013: 170) argued cannot be known ‘to government or anybody else’. It requires access to local, dispersed and tacit knowledge which are said to be inaccessible to any external observer. The same epistemic assumptions which underpin Hayek’s critique of government action and orthodox utilitarianism, therefore also limit claims he can make about the welfare benefits of markets. While he can claim that markets increase the opportunities for those involved in consensual exchange, his assumptions about the inaccessibility of economic knowledge rule out an understanding of the ways or extent by which they do this. According to the theoretical framework of Hayek’s political economy, then, we cannot know the magnitude of the opportunities markets produce.

What about the costs imposed on the opportunities of third parties by any market externality? Although Hayek did not analyse such costs in as much detail as the benefits of markets, he would seem to be in much the same position. Given that externalities involve a non-consensual imposition on third parties, they can be said to generally reduce the chances of those individuals achieving their ends. Once again, however, determining the size of such costs would require information about how a given externality affects market conditions, and information about the impacts of such effects on individuals’ ends. It would therefore require access to local, dispersed and tacit forms of knowledge which Hayek believed could not be known to any external observer. To see this, it is helpful to consider the consequences of Hayek’s epistemic assumptions for textbook neoclassical accounts of market failure which do aim to measure the magnitude of external effects. ¹⁹

On the standard neoclassical approach external costs are expressed in terms of preference satisfaction and measured through a person’s willingness to pay for a good at the margins (Pearce and Barbier 2000; Pearce 2014). For non-market goods (i.e. those involved in externality problems) such measurements are normally made through a ‘revealed preference method’ which determines willingness to pay by observing market behaviour in respect to connected goods (e.g. how much more are people willing to pay for a house in an area with better air quality). ²⁰ Determining a preference from someone’s actions in a complex market economy, and therefore outside of stylized examples, runs into conflict with Hayek’s epistemic assumptions. If someone can choose between an apple and a pear, and they choose the apple, then we may be able to say that they prefer apples over pears, but this is only because there is a known and well-defined opportunity set. For Hayek (2013: 170) complex market economies present individuals with hundreds of possibilities for satisfying their needs which cannot be completely known, and there are hundreds of possible preferences which could have motivated any given choice.

To borrow an example from Sagoff (2004), consider a person buying cookies from a girl scout. Was this action motivated by a preference for thin mints, a want to help the scouting organization, a desire to appear generous, because the girl scout was the daughter of a neighbour, or because turning the young child away may upset them? There is also the possibility of multiple determinacies as any action could be motivated by any combination of desires (e.g. a preference for thin mints plus an admiration for the scouting organization), and the opportunity set facing individuals will also not be fixed. For Hayek economic conditions are constantly in flux and the choices facing individuals will therefore alter through time. As we have seen he similarly takes people’s preferences and ends to be context dependent and therefore also open to change. Even if someone’s preference within a known opportunity set could be measured at t1, this does not therefore mean that this will be the same at t2. ²¹ On Hayek’s understanding of economic knowledge, then, people’s ends cannot easily be read from their behaviour within markets. As he put it, individuals engage in economic activity but they do so ‘for ends whose relative importance only these individuals know’ (Hayek 1945: 520).

Understanding the value which individuals attached to a good also requires controlling for the unique characteristics of the goods. Did a house receive a higher price because of better air quality, or was it closer to good schools, possessed special architectural details, had better transportation links, or any number of other factors? Even if all the goods affected by an externality could be identified, for Hayek these different features are examples of the particular circumstances which are only known to on-the-spot individuals. What is even more troubling is that Hayek (1941) argued that goods cannot be defined in purely physical terms as they are in part defined by the beliefs that people hold about them. If two watches are physically identical but one is a family heirloom, then the beliefs and values people attach to them will differ. To understand the range of factors that may influence people’s market behaviour therefore requires more than just a physical description of the good, but also an account of the subjective beliefs of the individuals involved. Both forms of knowledge are local, dispersed and often tacit, however, and cannot therefore be known to any external observer on Hayek’s view.

The same forms of economic knowledge which Hayek argued to undermine government action and orthodox utilitarianism therefore also rule out an assessment of the costs associated with market externalities. The kinds of problems just discussed will mean that while an externality non-consensually interferes with third parties and disrupts their chances of achieving their ends, understanding the ways and extent to which they do this requires knowledge which Hayek claimed to be inaccessible. According to the theoretical framework of Hayek’s political economy then, we cannot make any precise claim about the relative magnitude of the benefits of markets or the costs they impose through externalities. In terms of Hayek’s indirect utilitarian test, a market with an externality will both create and disrupt individuals’ opportunities, but the extent to which they do this is unknowable. The implications of these epistemic constraints for his defence of the free market will, however, depend on the kind of externality considered.

5. Implications for Hayek’s Defence of Free Markets

According to the previous section, Hayek’s epistemic assumptions prohibit any precise claim about the magnitude of the opportunities a market creates for individuals to achieve their ends or the magnitude of the opportunities it disrupts for third parties through any externality. To properly understand these conclusions, we must recognize that they do not rule out all welfarist or utilitarian claims. For instance, even without large amounts of local, dispersed and tacit knowledge, it would still be reasonable to claim that climate change has greater potential costs than acid rain, and acid rain greater potential costs than littering on a public pathway. Given that there are such large differences in the potential costs in these three cases, we do not require significant calculations and large amounts of economic knowledge to draw relative conclusions. Hayek’s epistemic assumptions do not result in complete ignorance, then, but they do introduce constraints when making closer comparisons as they prohibit more precise determinations of relative costs and benefits.

What does this prohibition mean for Hayek’s indirect utilitarian and welfarist defence of free market institutions? While Hayek wished to avoid the measurement or quantification of welfare due to its reliance on inaccessible economic knowledge, this section argues that the introduction of externalities forces him to make welfare judgements in the application of his own indirect utilitarian test which require similar forms of knowledge. Once externalities are introduced, Hayek’s epistemic assumptions will, in a narrow set of cases, undermine his ability to make absolute welfarist assessments of markets, and in a broader set undermine his ability to make relative assessments of more or less regulated markets. In both scenarios, Hayek will simply be unable to know whether or not free market institutions are superior in his own utilitarian terms to more regulated alternatives.

The first set of cases involves scenarios where an externality is large enough relative to the size of the associated market that it is reasonable to question whether its costs may, at least potentially, rival the benefits of that market. Under these conditions, determining if such a market passes Hayek’s indirect utilitarian test, and therefore whether it is welfare enhancing, cannot avoid assessments of costs and benefits. Applying this test would require precise judgements about the extent to which the market creates opportunities for those individuals engaged in consensual exchange to achieve their ends, and the extent to which it disrupts these opportunities for third parties. Without such judgements it will be impossible to determine whether this market increases the opportunities for individuals, all things considered, but as we saw in the previous section, these judgements require access to forms of knowledge Hayek believed to be inaccessible. Whether the benefits of such markets for individuals’ opportunities are greater than the costs it imposes on the opportunities of others is simply unknowable without some information about their relative magnitudes, but this requires access to knowledge about market conditions and individual plans which is local, dispersed and tacit. ²² In such cases, Hayek would be unable to apply his indirect utilitarian test and make even an absolute assessment of whether or not the market will, in the short or long term, be welfare enhancing.

The same can be said of more recent formulations of Hayek’s test, such as the concept of ‘catallactic efficiency’ within Austrian economics. Here efficiency is defined in terms of how well a system ‘permits its individual members to achieve their several goals’ (Kirzner 1963: 35), and asks whether a market ‘will maximize the extent to which individual members’ can achieve their ends (Cordato 1994: 132). Again, it seems we cannot apply such a standard in this first set of cases if we accept Hayek’s claims about the inaccessibility of economic knowledge. Without access to local, dispersed and tacit knowledge we cannot make precise judgements about the amount of end fulfilment promoted by a market or the amount of end unfulfilment produced by its externality. It is therefore again unclear whether such a market permits more individuals to achieve their several goals, all things considered. To clarify, the problem here is not that Hayek’s epistemic assumptions stop us from determining if such markets are optimizing welfare. Hayek (1937, 1945) explicitly rejects optimization as an unrealistic standard for institutional evaluation which would itself require full information. The problem is rather that his epistemic assumptions rule out a determination of whether a market creates more opportunities than it disrupts, and therefore whether it is even welfare enhancing in Hayek’s terms. ²³

In cases where the size of the externality can reasonably be thought to, at least potentially, rival the benefits of the market, Hayek’s epistemic assumptions stop him from making even absolute welfare assessments of markets, and determining whether or not the market is even welfare enhancing. No welfarist defence of such a market is therefore possible on Hayek’s epistemic assumptions. In many practical cases, however, external effects will not be thought to reasonably rival the size of the associated market. We are often in the position of recognizing the presence of a non-trivial externality but do not consider it to be large enough relative to the associated market to approach this market’s potential benefits. ²⁴ In these more common cases, we will not require large amounts of economic knowledge to judge that the benefits are likely to outweigh the costs. The market is simply large enough relative to the external effects that a general judgement can be made. Hayek would therefore be able to apply his indirect utilitarian test and make the absolute claim that the market is welfare enhancing, without needing to make the kinds of welfare judgements ruled out by his epistemic assumptions. While such assumptions therefore allow for an absolute welfare assessment of markets in this second set of cases, they still prohibit certain relative welfare assessments. More precisely, they rule out comparative welfarist evaluations of more or less regulated markets, and therefore any welfarist defence of free market institutions.

To see this latter problem, consider forms of government action which attempt to correct market externalities, such as the imposition of taxes or regulations. Such a policy will aim to reduce the interference of a market externality on the opportunities of third parties to achieve their ends, while it will necessarily impose on the opportunities of those engaged in consensual market exchange. Ideally, the government intervention would only impose on market exchanges to the extent that they produce external costs, eliminating the externality while allowing the market to otherwise function. Such an ideal policy would, of course, be ruled out on Hayek’s framework as it would require detailed knowledge about the impact of both the market and externality. From Hayek’s perspective, and that of many classical liberals, to compare a case of market failure to a perfect government policy is to fall afoul of what Demsetz (1969) referred to as the ‘Nirvana fallacy’. While markets are imperfect, government actions also involve forms of government failure, and these include the imposition of costs on the opportunities of third parties. Given this, government action aimed at reducing a market externality will produce benefits for those impacted by this externality, but it will also likely impose costs on others through its interference with market transactions and the price system.

What is therefore required is a comparison between the imperfect alternatives of the unregulated market with an externality and a regulated market with an imperfect government intervention. The problem for Hayek’s indirect utilitarianism is that such a comparison would again require welfare judgements which are ruled out by his epistemic assumptions. Both the unregulated market and the regulated market will create opportunities for those involved in market exchange, while also disrupting the opportunities of others through the market externality and government intervention. To know which better increases the opportunities for unknown persons, all things considered, would therefore require judgements about the magnitudes of the opportunities created and disrupted. It would require knowledge of the plans of individuals and how their opportunities to achieve these plans are impacted by market conditions, external effects and government actions. Any comparative assessment of these institutional alternatives therefore involves the very kinds of welfare judgements which the previous section showed Hayek’s epistemic assumptions to prohibit.

Understanding the extent to which a market externality disrupts the opportunities of individuals, the extent to which a government intervention reduces these disruptions, or the extent to which this intervention disrupts the opportunities for other parties, all require one to possess large amounts of knowledge which are local, dispersed and tacit. If we take Hayek’s epistemic assumptions seriously, then, we will not be able to know whether the unregulated market is better in terms of his indirect utilitarian test than the regulated market. In other words, whether a generally free market creates more opportunities for any unknown person picked at random than a more regulated market is simply unknowable on Hayek’s epistemic framework. Again, the issue here is not that Hayek cannot determine whether either alternative optimizes welfare (a claim he never made to begin with), but that he cannot determine which of the alternatives is superior in his own indirect utilitarian terms. ²⁵ Even in this second set of cases, where the externality cannot be reasonably thought to rival the benefits of the market, Hayek’s epistemic assumptions still frustrate any welfarist or utilitarian defence of free markets as they simply rule out a relative assessment of more or less regulated markets.

The problem, therefore, is that the introduction of external costs forces Hayek to make welfare judgements which his epistemic assumptions prohibit. As originally formulated, Hayek’s argument was that we could judge markets to improve the chances of unknown persons without quantifications of welfare and the inaccessible forms of economic knowledge this would entail. Determining whether government interventions into the market would improve upon these chances, alternatively, would require this knowledge. Once externalities are considered, however, costs must be introduced into the free market side of the ledger. Both a free and a more regulated market will involve the creation and disruption of individuals’ opportunities, and we cannot therefore apply Hayek’s indirect utilitarian test to these alternatives without judging the relative magnitudes of these opportunities. The trouble for Hayek is that these latter judgements also require what he thought to be inaccessible forms of economic knowledge, with the result that he is unable to know which best promotes the chances for individuals to achieve their ends, all things considered, in either the short or long run. While Hayek’s indirect utilitarian argument may therefore apply in the abstract where all external costs are assumed away, my claim is that it struggles to apply once externalities are introduced.

To clarify, I am not arguing here that Hayek either under- or over-appreciated the benefits of markets or the need for government intervention. It is rather the deeper problem that he simply does not provide a framework for evaluating these institutional alternatives in the context of externalities due to his strict assumptions about the inaccessibility of economic knowledge. Hayek is therefore just as unable to defend the kinds of government action which he at times advocated to deal with externality problems (e.g. Hayek 1944: 39–41; 2011: 473–475, 481, 492), as he is unable to advocate for free market institutions. Hayek’s indirect utilitarian and welfarist arguments simply break down in the presence of externalities.

In a narrow set of externality cases then, Hayek’s epistemic assumptions stop him from making even an absolute assessment of whether a market is welfare enhancing, and in a broader set stop him from making a relative assessment of whether more or less regulated markets are superior in terms of his indirect utilitarian test. In both scenarios, any welfarist defence of free market institutions will not be possible. From what has been argued so far, however, it may appear that this problem is only confined to a limited number of markets which are connected to large externalities, so that Hayek’s approach can be generally applied but for a few isolated exceptions. To understand the significance of such problems, however, we need to recognize the scale and ubiquity of externalities in today’s economies.

Modern industrialized and carbon-based economies involve pervasive externalities which are connected to almost all markets. Anthropogenic climate change provides the clearest example of this pervasiveness. Greenhouse gas emissions are spread throughout the whole of the contemporary economy as almost all goods and services involve some level of embodied energy as the result of their production, distribution or use. The consequence is that they all contribute to the costs of climate change and therefore an interference with the opportunities of others to pursue their ends. Such interferences will not, however, be reflected in market prices and individuals will not therefore have to take them into account. As Spash (2010: 174) argues, the fact that carbon is ‘all pervasive in the economic system’ means that its costs cannot be modelled as impacting specific products from a single sector. Instead, the price of carbon ‘affects all the prices in the economy’ with the result that failures in the price system’s ability to effectively coordinate individuals are not isolated exceptions but associated with almost all markets.

Although anthropogenic climate change is perhaps the only truly all-pervasive externality, the broader implications of this paper’s argument are not confined to this issue. Modern and highly industrialized market economies are associated with a range of significant and interconnected externality problems, including ocean acidification, the depositing of waste in oceans and river systems, acid rain, forms of air pollution, noise and light pollution, deforestation and ecosystem disruptions and large-scale biodiversity loss. While many markets may not be connected to any one of these issues, in a complex economy most markets will be in some way connected to at least some of these externality problems somewhere in their production process or supply chains. The pervasiveness of externalities in today’s economy therefore places doubt on Hayek’s (1944: 40) discussion of such problems which tends to see them as isolated issues which undermine the price mechanism in specific cases but do not impact market coordination more generally. While this underestimation of the scale of such issues may be a symptom of the time at which he was writing (also see Greenwood 2007: 81–82), modern understandings of the wide-ranging effects of economic activity on the natural environment and its associated risks to human welfare challenge such an understanding.

The problem highlighted here is not therefore specific to a small set of well-defined markets. Instead, the pervasiveness and persistence of large externalities means that Hayek will frequently be in the position of being unable to determine whether free and unregulated markets have any welfare benefits over more heavily regulated markets. This will be the case irrespective of whether Hayek assesses markets at an individual or system level. On one interpretation, Hayek wished to apply his indirect utilitarian test only to the free market system as a whole, or the ‘great society’ as he called it, rather than to any one part (Gray 1986: 57). While it is unclear that such an interpretation can account for all the ways Hayek applied his test, it does not change the conclusions drawn here. ²⁶ Whether it is a comparison between any single more or less regulated market, or a comparison between a more or less regulated market economy, relative welfare assessments will require forms of knowledge Hayek believed to be inaccessible. In modern economies where externalities are pervasive, Hayek’s epistemic assumptions frequently prohibit him from knowing whether free and generally unregulated markets will have welfare advantages over more regulated markets with greater forms of government intervention. In other words, he is frequently unable to advocate for free market institutions on welfarist or utilitarian grounds.

6. Possible Replies

Before considering further the implications of the above critique for Hayek’s classical liberalism, some potential replies need to be considered. One possible response is to reinterpret the implications of the above argument in a way more favourable to Hayek’s free market prescriptions. Rather than drawing the conclusion that we cannot evaluate the relative welfare properties of free markets on Hayek’s framework, perhaps the issues discussed merely imply that we cannot improve on the free market. If the market is producing benefits, and we cannot determine the size of any external costs or whether government intervention will reduce costs overall, then perhaps the best we can do from the normative perspective is simply stick with a generally unregulated market. The problem with this line of argument is that it is based on a status quo bias and does not provide a principled defence of free market institutions. Consider, for example, a case where a good is controlled and provided by a highly regulated market and that if a free market was introduced it would involve some non-trivial external effects. In this case, Hayek would not be able to determine if introducing the less regulated market would increase welfare compared with the status quo, and therefore the best we can do is stick with the regulated one. This reinterpretation does not therefore save Hayek’s indirect utilitarian argument as it still leaves him unable to provide a principled defence of free market institutions.

Another set of possible replies come from modern Austrian approaches to externalities. While inspired by Hayek’s political economy, this work has often combined it with Coasian insights, according to which all market externalities can be internalized through processes of bargaining and negotiation between property rights holders (Anderson and Leal 2001; Pennington 2001; Cordato 2004). Given that my argument only holds to the extent that market externalities persist, the possibility of internalizing costs through Coasian bargaining may offer a solution. As Coase (2013) himself was well aware, however, this conclusion only holds under the assumption that there are zero transaction costs, and that in a world of positive costs private bargaining will often fail. When it comes to the many large-scale externalities surveyed in the previous section, the numbers of actors involved commonly result in prohibitively high transaction costs for Coasian bargaining to take place and externalities are highly unlikely to be resolved through such means. To an extent, Hayek (e.g. 2011: 473–475, 481, 492) can be interpreted as arriving at a similar conclusion, citing a range of cases where property rights will fail to internalize costs, and these limits are also recognized in parts of the ‘free market environmentalist’ literature, at least when it comes to large scale and transboundary goods (e.g. Pennington 2001: 185). The Coasian view does not therefore provide reasons to think that large and pervasive externalities will not continue to persist in a free market, and the argument of the previous sections continues to hold.

An extension of this reply is that while many external costs cannot currently be internalized through private bargaining, we should still stick with the free market due to its problem-solving capacity. While Hayek’s epistemic assumptions frequently stop us from determining if a generally unregulated market is welfare superior to a regulated one when externalities are present, this argument would claim that the former provides the best chance of solving externality problems into the future. Firstly, it should be noted that this reply concedes a lot, as it moves Hayekians from a grounded defence of free markets based on their direct welfare benefits, to a speculative claim about their potential to solve large-scale problems in the future. It therefore concedes the central claim of this paper; that Hayek’s epistemic assumptions rule out the judgement that free market institutions better increase the chances for any unknown person to achieve their ends than a more highly regulated market.

Secondly, it is also not clear that what are normally taken to be the problem-solving capacities of free markets can be effectively applied to the kinds of externality problems discussed here. Large and pervasive externalities often do not allow for the establishment of property rights on which market experimentation can then take place, or they are at a scale that makes private bargaining prohibitively costly. Similarly, while private firms normally have incentives to solve problems by providing better goods (at least under competitive conditions) it is not clear that incentives are always appropriately aligned for externality problems where firms often benefit from externalizing costs. It is not therefore evident that we can have confidence in the problem-solving capacity of a free market system in this context or its superiority over more regulated markets. Or at least the burden would be on Hayekians to show that such capacities can be applied even in the case of large and pervasive externalities, and that they are superior to any larger role for the state.

Another potential response from more recent Austrian approaches is suggested by Cordato (2007: 17, 81–82). While I have argued that Hayek’s epistemic assumptions rule out judgements about the magnitudes of external costs, Cordato questions whether they stop us from even knowing that costs are involved. To borrow his example, storing junk cars on one’s front lawn creates costs for one’s neighbours if they find it unsightly, but benefits if they share a love for old automobiles. We cannot therefore know if an externality really disrupts the opportunities of third parties without knowledge of the preferences and plans of individuals, and Hayek claimed this cannot be known to an external observer. It is true that there are cases where the negative valence of an externality can be reasonably questioned. Someone will not necessarily know if my neighbour imposes costs on me by playing classical music in their garden, given that I may myself have a preference for classical music. To generalize this argument to the kinds of externalities discussed above, however, would be to draw too extreme a conclusion from Hayek’s epistemic assumptions. These kinds of externalities commonly involve serious damage to human health and property, and we do not therefore require detailed knowledge of people’s preferences to make the reasonable judgement that they impose costs. As already noted, Hayek (1944: 40; 2011: 473–475, 481, 492) was happy to draw such judgements and Cordato (2007: 76) himself does not seem to generalize the argument, stating that most standard externality cases can be seen as unwanted. So, while this reply may hold in certain cases, it does not for the kinds of externalities discussed here.

A final response would be to claim that the problem I have identified is really a problem for all approaches and that I have not therefore said anything specific about Hayek’s political economy. This paper’s argument, however, is internal to a Hayekian understanding of economic knowledge. While no approach can rule out the possibility of some unknown externality, I have focused on known externalities and how Hayek’s specific epistemic assumptions frequently prohibit relative welfare assessments in such cases. Approaches with alternative assumptions will not, therefore, necessarily face the same challenges. As we saw earlier, for instance, standard neoclassical economics does not take individual preferences to be unmeasurable nor does it rule out welfare assessments of even large externalities (Pearce and Barbier 2000; Pearce 2014). At least in principle, then, such an approach provides a framework for making comparative assessments of more or less regulated markets in the context of externalities. Similarly, approaches which endorse objective list accounts of wellbeing, such as basic need and capability views, allow for relative assessments of human welfare and can therefore in principle be applied to the cases I have discussed (Doyal and Gough 1991; Sen 1999). Of course, a full vindication of any alternative would require further defence of its own epistemic and welfarist assumptions, and welfare assessments on any approach will face a range of practical difficulties which should not be underestimated. The point, however, is that Hayek’s framework rules out these assessments in principle because of its particular assumptions about the inaccessibility of economic knowledge. ²⁷

7. Conclusion: A Little Less Ignorance?

This paper has argued that one of the strongest and most influential theorists of free markets very often cannot endorse the welfare advantages of unregulated market institutions. While Hayek’s classical liberalism claims that markets best promote the opportunities for any unknown person to achieve their ends, the epistemic assumptions of his political economy undermine relative welfare assessments of more or less regulated markets when there are large and persistent externalities. In modern economies where such externalities are pervasive, Hayek will frequently be unable to know whether a generally free and unregulated market is superior in welfare terms to a more highly regulated market, with the result that he cannot defend either on utilitarian or welfarist grounds. I have therefore argued that Hayek’s aim of defending free markets from his assumptions of ignorance and epistemic scepticism is in one sense self-defeating, as it ends up very often ruling out comparative welfare assessments of real-world markets.

This critique also challenges many schools of economic thought Hayek inspired. Often accepting his epistemic framework and adopting a catallactic account of efficiency, many Austrian economists will be in a similar position when it comes to relative welfare assessments of markets. Despite their strong association with free market advocacy, proponents of the Austrian school may in fact provide little basis for a welfare economic defence of unregulated markets if they hold on to Hayek’s understanding of economic knowledge. Both Hayek and modern Austrians may, of course, still make contributions to positive economics and political economy. More recent classical liberal scholarship has also moved away from a sole focus on markets to polycentric approaches involving a broader range of institutional forms (Pennington 2008; Aligica et al. 2019). This work integrates Hayekian ideas with the work of Elinor and Vincent Ostrom, and I leave open the question of whether such approaches can overcome the problems detailed here. Instead, I have focused on Hayek’s political economy and have argued that a tension exists between his indirect utilitarian argument for free markets and the epistemic assumptions on which it is based.

There are, of course, two possible ways out of this tension. One could retain Hayek’s epistemic assumptions but jettison his indirect utilitarian line of argument and the possibility of relative welfare assessments of more or less regulated markets. Given the strong belief of many classical liberals and neoliberals that increasing the scope of free and unregulated markets will bring direct benefits to human welfare, it is likely that this is a conclusion many Hayekians would find difficult to accept. It would also mean that a Hayekian classical liberalism would need to stand much less on its consequentialist claims, and much more substantially on its Kantian, neo-republican or conservative themes. It may, therefore, need to give up on Hayek’s (2011: 52–53, 83) aim of showing that liberty is not merely a ‘ethical presupposition’ and that a ‘faith in freedom’ can be grounded on its ability to realize ‘more forces for the good than for the bad’.

The alternative resolution would be to retain a welfarist and utilitarian line of argument but weaken Hayek’s assumptions about the inaccessibility of economic knowledge. While providing the possibility for a more productive Hayekian welfare economics, this second strategy is likely to challenge its free market prescriptions. On the one hand, relaxing his epistemic assumptions may broaden the possibilities for relative welfarist evaluations of more or less regulated markets, including in the context of externalities. On the other hand, relaxing such assumptions would also allow for more knowledge on the part of policymakers and therefore greater forms of government intervention into the market economy than Hayek’s classical liberalism would likely wish to accept. While these epistemic assumptions currently frustrate a clear endorsement of free market institutions, relaxing them would also likely diminish Hayekian political economy as a staunch defender of the free market.