Exit, Voice and Technocracy

ABSTRACT

In Power Without Knowledge, Jeffrey Friedman develops a critique of technocracy and in doing so makes an epistemic case for exit over voice. He argues that a technocracy that fails to take people’s ideational heterogeneity into account is unlikely to possess the knowledge required to solve social problems, and that the alternative of “exitocracy” may, in some cases, overcome these limits. By creating the conditions under which individuals can exit from undesirable social situations, an exitocracy may allow people to escape their social ills without knowing their society-wide causes or solutions. As Friedman recognizes, however, an exitocracy still requires technocratic knowledge, and this paper explores these requirements in further detail. First, it investigates the limits of exit as a solution to social problems, and the extent to which technocratic policies can be substituted for exit. Second, it considers the need for technocratic knowledge in the promotion of exit opportunities, and how this undermines Friedman’s defense of exitocracy.

Keywords:

• technocracy
• exit mechanism
• exitocracy
• political epistemology
• green political theory
• primary public goods
• secondary public goods
• tertiary public goods

In Power Without Knowledge (Oxford University Press, 2019), Jeffrey Friedman develops a powerful critique of technocratic politics. Unlike previous critics, such as Habermas, who aimed to steer the power of technocracy towards democratically determined ends, Friedman argues that technocracy may lack the very knowledge it requires for its legitimacy: reliable knowledge of how to solve social problems and improve the human condition. Given the epistemic complexity of social problems, there is no assurance that a technocratic state would pass the very test it sets for itself: that of creating more benefits than costs.

As the book’s preface notes, the arrival of a detailed and sustained critique of technocracy may seem untimely. The current political situation in the United States (the focus of the book) and elsewhere seems to feature a greater threat of populism than of technocracy. Friedman’s argument, however, is very far from irrelevant or outdated. It cuts to the core of a widely held belief that is common on all sides of the political spectrum, even among those who would not think of themselves as sympathetic to technocracy. This is the belief that the solutions to social problems are readily apparent. For Friedman, this belief is at the center of technocracy but also underlies much of our political discourse. Public-policy debates involve high levels of disagreement, but all sides share the assumption that they already know or can easily discover which policies will best address our social ills.

A similar assumption is also found in recent debates within political and democratic theory, which pit epistemic democrats (e.g., Landemore 2013 and 2014) against defenders of epistocracy (e.g., Caplan 2007 and Brennan 2016). While differing greatly in the form of political rule they favor, both sides in this debate share the belief that someone—whether ordinary citizens or social-scientific elites—can unproblematically know how to solve, alleviate, or prevent social ills. Friedman’s critique of this assumption has wide-ranging implications, then, for the contemporary theory of politics as well as its practice. The argument developed in Power Without Knowledge makes important contributions to both political theory and political science, and will surely come to be seen as a key work in the burgeoning field of political epistemology.

Power Without Knowledge proceeds in three parts. The first analyzes the complexity of social problems, the likelihood of unintended consequences in public policy, and therefore the underappreciated burden facing those who claim to know which policy intervention is likely to be most effective. The second investigates whether technocrats are likely to possess such knowledge, and argues that the prospects currently appear to be bleak for both epistocrats, such as trained social scientists, and “citizen-technocrats”—ordinary citizens who hold opinions about technocratic policy issues. The third part then considers “exitocracy” as a partial solution to these problems. By promoting the ability of individuals to exit from social arrangements, exitocracy allows them to escape their problems instead of waiting for a society-wide technocratic fix. Friedman argues that such a system, in some cases, can enable people to improve their condition even in the absence of judicious technocratic knowledge.

I will focus on the last of these three parts. Its claim that we can produce better outcomes by prioritizing exit in the private sphere over voice in the public sphere is not new, but such claims have previously tended to come from classical-liberal theorists (e.g., Hayek 1945 and 1960; Kirzner 1985; Pennington 2011; Tebble 2016), while being a target for critique (including epistemic critique) from those on the political left (e.g., Bennett 2016; Benson 2019a and 2019b; Gamble 1996; Greenwood 2007 and 2010; O’Neill 2012). Friedman, however, bucks this trend by advocating an egalitarian form of exitocracy that would facilitate exit by means of the redistributive state. Friedman’s proposal therefore has as much of a claim to being socialist as capitalist.

Sections I and II will give a brief overview of the epistemic challenges that Friedman sees as facing technocracy, will outline his idea of exitocracy, and make some general comments along the way. The remaining sections follow Friedman in considering the ways an exitocracy must itself rely on technocratic forms of knowledge. Section III looks at the need for such knowledge at the margins of exitocracy, where the possibility of effective exit is limited or non-existent. These limits reduce the ability of exitocracy to serve as an alternative to technocracy, as Friedman acknowledges. I explore these problems in further detail and introduce some additional limits that have implications for our ability to replace voice with exit. Recognizing these issues, however, does not challenge the general case for exit when and where it can be effectively exercised.

Sections IV and V then consider the more problematic need for technocratic knowledge at the heart of exitocracy. An exitocratic state must make technocratic judgments in order to fulfill its role of promoting and sustaining the ability of individuals to exit. While Friedman acknowledges this dilemma and admits that it limits the policy conclusions of his critique, he argues that we can still endorse a normative hierarchy that places exitocracy above ordinary technocracy. I, however, argue that this dilemma may create deeper problems for his view. Once we recognise that technocratic knowledge is needed to understand which actions increase or decrease exit opportunities, we cannot be sure whether an exitocratic state will actually create any of the benefits Friedman associates with exit. If Freidman’s critique of technocratic knowledge is correct, then, it is inconclusive whether we should prefer exitocracy to technocracy.

I. THE CRITIQUE OF TECHNOCRACY AND TECHNOCRATIC KNOWLEDGE

Friedman defines technocracy as a “polity that aims to solve, alleviate, or prevent its people’s social and economic problems” (5). On this account, technocracy is distinguished from other forms of politics not by the structure of its institutions but by its problem-solving mission. Typically, or as Friedman says “ordinarily” (Part III), technocracy attempts to accomplish its mission by implementing government policies to deal with discrete social problems, such as poverty, the need for education, and the need for health care. Technocracy’s social programs and regulations do not aim at achieving any broad conception of social justice, but rather at the consequentialist and specifically “negative utilitarian” goal of producing “the least amount of avoidable suffering for all” (Popper 2013, 602).

Given that it is defined in terms of its mission, Friedman offers a view of technocracy that differs greatly from the scholarly and popular image of it as rule by technical experts. If a technocracy is a polity that aims to solve social problems so as to minimize human suffering, then it is only once we determine who can best do this that we can say whether technocracy should involve rule by the many or the few (21).¹ Yes, technocracy may involve the elite rule of experts (“epistocrats”), but it may also take the form of direct rule by the people if it turns out they can better solve our social problems. Technocracy is, then, a “fundamentally epistemic project” (ibid.) that may take oligarchic, democratic, or mixed forms.

Negative utilitarianism, like its positive cousins, requires a calculation of the costs and benefits of any action. A negative-utilitarian technocrat is, therefore, someone who claims (explicitly or implicitly) to know the costs and benefits of public policies, and to know the scope and causes of the problems that public policies aim to solve. Without such knowledge, technocracy cannot be relied on to do more good than harm. To be more precise, technocrats need to have the following four types of technocratic knowledge, according to Friedman (46-47):

1. Knowledge of which social problems are not only real but significant, in the sense that they affect large numbers of people—or small numbers intensely.

2. Knowledge of what is causing the significant problems, and (preferably) knowledge of what might cause significant new problems in the future.

3. Knowledge of which technocratic actions can efficaciously solve, mitigate, or prevent the significant problems.

4. Knowledge of the costs of efficacious solutions, including not only

anticipated costs but those that are not intended, and thus are not anticipated.

Friedman labels his critique of technocracy “internal” because it questions whether technocrats are likely to possess these four types of knowledge, which is to say, whether they are likely to have the knowledge necessary for technocracy to be legitimate according to its own negative-utilitarian standards. While citizen-technocrats or epistocrats may possess the needed knowledge in any given case, all technocrats, according to Friedman, face epistemic problems that call the adequacy or reliability of their knowledge into question. His argument is therefore distinct from external critiques of technocracy, such as those put forward by the early Habermas (1970 and 1976), which object to technocracies’ lack of democratic credentials but do not question the likelihood of attaining reliable technocratic knowledge itself.

The Problem of Unintended Consequences

Why is it that Friedman believes that technocratic politics, whether at the elite or mass level, is inherently problematic? The main force of his argument relates to the difficulty of predicting unintended consequences, and thus the tenuousness of technocratic knowledge of the fourth type.

Friedman points out that determining the actual consequences of any technocratic policy requires making predictions of human behavior. Thus, knowing whether an increase in the minimum wage will have beneficial consequences, all in all, requires making a prediction about the likely behavioral response of employers when the policy increases their labor costs. Will they, for instance, respond by reducing their demand for labor, as neoclassical economists might argue, and if so, by how much? Without such a prediction, a technocrat would not know whether the higher wages that the policy will direct toward the lowest-paid workers is beneficial overall or is outweighed by increases in unemployment.

In general, technocracy is only able to achieve its ends to the extent to which human behavior can be reliably predicted by technocrats. Friedman, however, wonders if human behavior can be accurately predicted by anyone, at least when it comes to the “anonymous others” who are usually the objects of technocratic policy making (30 and passim).

Ideational Heterogeneity and the Problem of Prediction

At the core of Friedman’s challenge to the reliability of predictions in the social world is an ideational account of human behavior that he develops from the work of Walter Lippmann ([1922] 1997) and W.V.O. Quine (Quine and Ullian 1978). Simplifying for the purpose of brevity, Friedman’s account focuses on the importance of interpretation in the behavior of social actors, and particularly the role of heterogeneous ideas in these actors’ interpretations of how they should behave in a given case.²

Individual decisions about how to act, according to Friedman, do not involve the objective registry of the objective facts of one’s environment, but are rather based on subjective interpretations produced by individuals’ webs of belief. The ideas that constitute a given web include perceptions, beliefs, interpretations, hypotheses, theories, inspirations, and assumptions (140). Friedman’s view that individual behavior is ideationally determined does not by itself lead him to question the reliability of behavioral predictions. Rather, a problem arises because the ideas in various individuals’ webs of belief are heterogeneous and inaccessible to observation. None of us knows the content of the ideational inputs into others’ webs, but we have reason to think that, to some extent, they vary from person to person (144-48). Therefore, to some extent, we do not know how people will respond to their situations as they interpret them. To some extent, then, public policies designed to manipulate their actions so as to solve social problems may have behavioral consequences that we do not anticipate.

Friedman points out that every person’s web of beliefs is formed by tens of thousands if not millions of different inputs over the course of their lives, from their upbringing and social class, to the books they read and the films they watch. These will differ, across a given population, in somewhat unknowable ways. A degree of ideational heterogeneity among social actors means that they will interpret their environments differently to that degree (ceteris paribus) and, as a result, behave differently. In reaction to an increase in the minimum wage, for instance, one firm’s managers may interpret the situation as requiring that they reduce the number of their employees, another may find savings in other areas of production, while another may not change their behavior at all (204). The presence of ideational heterogeneity, then, makes prediction in the social sciences a much more difficult task than in the natural sciences, as it requires that the social scientist understand the subjective beliefs of those whose behavior is being predicted. One cannot simply assume an individual's motivation (e.g., self-interest) and from this derive their objective rational action (e.g., decrease demand for labor when its price increases). Instead, one must, like an intellectual historian, come to an understanding of the particular webs of belief of the actors (160-63). Although we may be able to do this with those people we are close with, or (to an extent) with “anonymous others” who have already acted, doing so prospectively for a large number of anonymous others is another matter.³ It is the latter task, however, that confronts any would-be technocrat, who must predict the behavior of “unseen men” (Lippmann [1922] 1997, 165).

The Possibility of a Judicious Technocracy

There may, however, be homogenizing forces in society that push back against ideational heterogeneity (169-77). Uniform social norms, for instance, can cause regularity and predictability in the behavior of otherwise heterogeneous individuals. The task of the “judicious” technocrat is, therefore, to determine in any given case where these regularities in behavior lie, and to distinguish them from otherwise unpredictable actions. With this challenge in hand by the end of Part I of Power Without Knowledge, Part II goes on to investigate whether real-world technocrats, be they epistocrats or average citizens, can rise to meet the challenge of judiciousness. Here Friedman asks how judicious real-world technocrats are; why they might not be very judicious, at a systemic level; and, therefore, what the prospects are for change.

There is not space here to survey the three chapters of Part II in detail, but suffice it to say that Friedman paints a pessimistic but not entirely hopeless picture. He argues that modern social scientists tend to ignore ideational heterogeneity due to a cultural “pressure to predict” (Chapter 4), and that this produces both the apriorism of neoclassical economics and the positivism of empiricist correctives to apriorism, such as the use of randomized controlled trials. In addition, he argues, the process by which one becomes an expert, and thus an epistocrat, tends to produce dogmatism, the result of which is that the most informed are also the least open minded (Chapter 5). Citizen-technocrats, on the other hand, are less dogmatic than experts, but have little technocratic knowledge and therefore tend to focus on the intentions of politicians rather than the consequences of their policies (Chapter 6).

None of this, however, rules out the possibility that a judicious technocracy may come into being. It is possible that there could be a “cultural revolution” (314) against the ideationally insensitive culture of epistocrats and citizen-technocrats. Such a revolution could lead technocrats to be interpretively charitable to those they want to predict, empirically scrupulous, open-minded, and familiar enough with their subjects to locate areas of homogenization on which to base policy interventions (although they would need to make sure these interventions did not themselves disrupt such homogenization; 172). However, Friedman argues that it is unlikely that we will achieve such a judicious technocracy in the foreseeable future, pending such a cultural revolution. On his view, as I understand it, the possibility of a judicious technocracy means that we cannot conclusively reject the legitimacy of technocracy, but pending such a revolution, the difficulty of predicting human behavior means that we should be skeptical that any would-be technocrat has gotten hold of reliable technocratic knowledge.

II. THE EPISTEMIC ADVANTAGES OF EXIT OVER VOICE

As Friedman points out at the beginning of Part III, however, it would beg the question for him to claim that technocracy has, a matter of empirical fact, done more harm than good (or vice versa). All that Part I did is establish a presumption in favor of ideational heterogeneity, and argue that it is the task of technocracy to judiciously investigate the possibility of homogenizing forces. While Part II suggests that this is not being done very much at present, it still might be done in the future if there is a cultural revolution against the injudicious neglect of ideational heterogeneity. We are left, then, with the theoretical possibility that, in empirical reality, there is a significant amount of unpredicted behavior, and thus of unintended consequences flowing from technocratic action; but it is also possible that the unintended technocratic consequences are not of sufficient magnitude to render technocracy internally illegitimate.

To resolve this uncertainty, Part III presents an ideal type of “exitocracy” as an alternative to “ordinary” technocracy. Ordinary technocracies address social and economic problems by attempting to gain the four types of technocratic knowledge for each problem that is to be solved—knowledge that is voiced and acted upon through some sort of deliberation. This deliberation is aimed at crafting the best policy response to social problems, case by case. Exitocracy serves as a point of comparison against ordinary technocracy—whether it be democratic, epistocratic, judicious, or injudicious—allowing Friedman to avoid the uncertainties created by Parts I and II. The ability of ordinary technocracy to do more good than harm depends on such factors as the degree to which ideational heterogeneity will be neutralized by ideational homogeneities in particular cases, and the degree to which judicious technocrats will become aware of this, which Friedman does not feel capable of predicting (317-319). This leaves unresolved the question of whether ordinary technocracies are legitimate, according to their own negative-utilitarian standards. Answering this question would require Friedman to make assertions about the magnitude of the unintended consequences caused by technocratic action, presupposing his access to the very type of technocratic knowledge whose reliability was undermined in Part I. By comparing technocracy to exitocracy, however, Friedman can make relative claims that do not require him to possess such knowledge.

Exitocracy as an “Extraordinary” Form of Technocracy

In contrast to ordinary technocracy, Friedman defines exitocracy as an extraordinary form of technocracy that seeks to alleviate the presumptive unreliability of technocratic knowledge, or more accurately avoid it, by prioritizing exit over voice. In Albert Hirschman’s (1970) original formulation of the exit/voice dichotomy, individuals have two options when they are dissatisfied with the goods or services of a provider. Either they can express their concerns by voicing their dissatisfaction, or they can exit from the service. Someone dissatisfied with a train service, for instance, can either complain directly to the train company or exit from it and experiment with the services of a rival company. While Hirschman was concerned that exit might undermine the potential to improve public services through voice—his example was of a government-run train service that has a monopoly of provision, such that the only alternatives to it are buses or taxis—Friedman (330) is interested in the epistemic advantages of exit in allowing people to escape from their social problems, and thus from their need to know how to solve them at a society-wide level.

The principal benefit of exit, then, is, according to Friedman, that it reduces the need for reliable technocratic knowledge in the alleviation of social problems. Compare, for instance, an individual voicing a complaint to a technocratic state with the same individual utilizing exit in a private market. In the first case, the individual expresses dissatisfaction with a social problem based on personal, experiential knowledge of the problem. For the problem to then be solved, however, the individual’s relatively reliable experiential knowledge must be coupled with relatively unreliable technocratic knowledge. The voiced complaint highlights a problem, but before it can be fixed, knowledge is also required about whether this problem is widely experienced, what its causes are, and which changes would address the problem without causing too many unintended consequences. In the second case, the individual uses experiential knowledge of the problem as the basis for exiting from it—for example, to another private firm—in search of relief. If this experiment works, then the individual will have escaped the problem without needing unreliable technocratic knowledge of the scope, causes of, or solutions to the problem not only for her but for everyone else in the society. As long as one can recognize the effects of one’s exit on one’s own well-being, one can exit without needing to be aware of the scale, the cause, or the society-wide solution to the problem. Individuals can, for instance, switch from one private provider of transportation to a better one without requiring any knowledge about why this alternative performs better.

Private firms are, of course, attempting to offer goods or services and, in doing so, are required to make technocratic judgments (in the sense of predictions about customers', employees', suppliers', and investors' behavior). Friedman believes that private firms are just as prone to error in making these judgments as are ordinary technocrats. The difference, however, is that in a private market multiple firms are attempting to provide goods and services as opposed to a single technocratic state. The presence of ideational heterogeneity among the providers of private goods and services, therefore, becomes an epistemic advantage, as it may lead firms to offer a diversity of solutions to people’s problems. An exitocracy allows for exit between these alternatives, so that people can escape their problems without needing to know how many others are experiencing the problems or even how the better experience that they might discover was brought into existence. Thus, Friedman maintains, exit-based action in the private sphere, while imperfect, should be better able to solve social problems, in some cases (and thus overall), than voice-based technocratic action in the public sphere. Exit allows people to escape (some of) their problems “without understanding the society-wide significance or cause of the problem, without knowing the efficacy of a society-wide solution, and without weighing the costs and benefits of such a solution” (333).

The contrast Friedman draws here between mechanisms of voice in the public sphere and those of exit in the private sphere may be overly stark. Although he recognizes opportunities for voice in the private sphere (331-32), he does not explore the opportunity for exit in the public sphere (and therefore in ordinary technocracy). Exit, however, comes in many more forms than Friedman, and for that matter Hirschman, seem to recognize (Dowding and John 2012). Individuals can exit from one private service to another but also from a private service to a public one (and vice versa), from a public or private service to no service at all (and vice versa), or from one public service to another public service. The latter could involve moving from a publicly provided service in one geographical location to another (such as from a public hospital in Manchester to one in Liverpool) or from different public service within the same location (such as taking the public bus service to and from work in London rather than the public underground train).⁴ The public sphere can, therefore, involve forms of exit by offering multiple kinds of services/solutions to social problems. The voice/exit distinction does not as neatly map onto the public/private distinction as Friedman sometimes seems to suggest when he runs the two distinctions together.⁵ That being said, the market and the private sphere do seem to provide a greater opportunity for exit than does the public sphere, given the number of private firms that can compete at any one time (at least subject to the level of market concentration). The private sphere may thus provide greater scope for exit-based action.

Exitocracy and Redistribution

Unlike an ordinary technocratic state, which aims to find direct solutions to society’s many discrete social problems, the exitocratic state’s “cardinal” and “distinctive” goal would be to provide a framework within which individuals (as consumers and employees) would be able to use exit to avoid the problems that afflict them (322, 341). It would provide public goods that are vital to the private sphere in order to create an environment “in which individuals using exit can flourish” (322). An exitocracy, then, is technocratic in seeking to address social problems and, as we shall see, it is technocratic in requiring an array of technocratic knowledge. But wherever possible, it aims to solve social problems indirectly, by promoting individuals’ ability to exit in the private sphere rather than by deliberate efforts in the public sphere. As a result, an exitocracy reduces our reliance on relatively untrustworthy technocratic knowledge while still helping people to improve their lot, in a negative-utilitarian sense.

An appeal to the private sphere as an epistemically superior alternative to voice and politics is not completely new. Classical-liberal theorists, particularly in the tradition of Friedrich A. Hayek (1948 and 1960), have long pointed to the epistemic advantages of markets and the private sphere more broadly (e.g., Kirzner 1985; Pennington 2011; Tebble 2016). As Friedman rightly points out, however, there are important distinctions between these arguments and those he develops in his book. His argument differs, for instance, from that of the early Hayek’s (1945) focus on the information-aggregating function of the price system. Friedman (335-36) argues that Hayek ideologically assumed that the fallible opinions of entrepreneurs constitute reliable “knowledge,” even while denying that central planners have reliable knowledge rather than fallible opinions. Hayek also placed emphasis on the dispersed, fragmented, and tacit nature of social knowledge, while Friedman’s account does not rely on knowledge taking any such form.⁶ There are also some interesting similarities, however, between Friedman’s defense of exitocracy and Hayek’s epistemic justification of a market society. In The Constitution of Liberty (1960, 377), for instance, Hayek argues that a key problem with granting a monopoly to the state to solve social problems is that it will “prevent the emergence of more effective alternatives” and deprive “individuals of any choice in the matter.”⁷ The ability of private firms to offer different solutions and to allow individuals to choose between them was important to Hayek’s epistemology of the market as well as Friedman’s. Others who have analyzed the epistemic properties of markets have also pointed to the role that market competition plays in decreasing the need for scientific knowledge, as it allows individuals to make simultaneous comparisons of alternatives (DeCanio 2014).⁸

Friedman’s account of exitocracy, however, has as much of a claim to being socialist as capitalist, and is therefore far from being classically liberal. In addition to the divergence between his and Hayek’s justifications of the private sphere, their views differ greatly over the role of the state in terms of redistribution. Friedman’s exitocratic state would engage in very significant forms of redistribution as part of its core function of providing exit opportunities to as many people as possible. Exit is not costless, and it also requires opportunities to enter alternatives. Friedman (335), therefore, suggests that the negative-utilitarian goals of technocracy require the implementation of a version of the Rawlsian Difference Principle—one not aimed at maximizing the “primary goods” of the least advantaged, in line with the demands of justice, but rather their “entry goods,” in line with the goal of helping them solve their problems. An exitocratic state, then, would aim to “maximize, for as many people as possible, the opportunity to use exit to solve problems affecting them personally,” by affording them the means “to enter into as many alternatives as possible, consistent not only with others’ roughly equal opportunity to do the same, but with the expansion of such opportunities for all” (ibid.). Friedman’s Afterword therefore provides an explicitly egalitarian defense of exit and the private sphere, which he believes could form the basis of a new left that would be both non-technocratic and non-Marxist.

It is therefore useful to distinguish what I will call classically liberal exitocracy from egalitarian exitocracy.⁹ While the former involves redistribution in order to provide a minimum level of income, the latter involves a much greater level of redistribution in line with a version of the Rawlsian Difference Principle. We will briefly return to the distinction between these two forms of exitocracy later. For now, the distinction highlights an interesting aspect of Friedman’s epistemic endorsement of exit. Unlike others on the political left who have tended to critique the emphasis classical liberals put on the private sphere (Bennett, 2016; Benson, 2019a, 2019b; Greenwood, 2007, 2010; O’Neill, 2012), Friedman offers an egalitarian defense of exit.¹⁰ This egalitarian exitocracy is said to reduce our reliance on problematic technocratic knowledge by promoting and sustaining the ability of individuals to exit, and therefore to escape their social problems rather than waiting for a technocratic fix.

III. TECHNOCRATIC KNOWLEDGE AT THE MARGINS OF EXITOCRACY

The extent to which an egalitarian exitocracy can alleviate the problems of technocracy will depend on the extent to which it reduces the need for technocratic knowledge. As Friedman recognizes, however, there is, in fact, a large need for such knowledge in an exitocracy. It is this need for technocratic knowledge on which I will focus in the rest of this paper. Sections IV and V will consider the technocratic knowledge required to promote and sustain exit opportunities, and I will argue that this requirement challenges Friedman’s view that we should prefer exitocracy to technocracy. This section will look at the need for such knowledge at the margins of exitocracy. That is, it will look at the limits facing the ability of individuals to exit from their problems, such that reliance on technocratic knowledge and voice-based solutions cannot be avoided. Some of these limitations are recognized by Friedman (332-34), and they do not negate the benefits of exitocracy when and where exit can be effectively exercised. Friedman calls exitocracy an extraordinary form of technocracy precisely because, while prioritizing exit over voice, it will often have no recourse but to “ordinary” technocratic solutions. These limits to exit do not, therefore, challenge the legitimacy of exitocracy. My aim in this section is to explore these limits further and add some additional considerations that bear on the extent to which exit can be substituted for voice.

We can see the restricted potential for replacing technocratic policies with exit when we consider the origins of ordinary technocracy. As understood by Habermas (1970, 102) and Friedman (351-52), technocracy developed to address the “dysfunctions” of free exchange and capitalist markets. Many of these dysfunctions, however, occur precisely because private exit is not an option. This is not true of all of the problems of markets; there are social ills from which people could exit if not for lack of resources, and these problems may be dealt with by the redistributive function of Friedman’s exitocracy (although as we will see in the next section, this solution does not come without its own problems). The issue, however, is that there are a host of common and pervasive problems that affect markets precisely because exit from them is not possible, and in such cases, exitocracy will offer no alternative to voice and technocratic knowledge.

The first set of these dysfunctions goes under the rubric of public goods (and, conversely, public bads), in the technical sense of goods or services that are non-excludable and non-rivalrous. The problems facing private markets when it comes to the provision of such goods and services are well known. The important point for our purposes is that they cannot always be effectively escaped by means of exit. If people experience public bads such as low air quality, there is not necessarily an exit option which allows them to escape. This is because the very publicness that makes such goods difficult to provide in private markets also restricts people’s ability to exit from their occurrence.¹¹ This is also true of some kinds of externalities. Although the costs of externalities are not necessarily public, as they can, for example, be locally confined—as in the case of effluents that run from a stream onto private land—they can also be difficult to exit from when they are of significant scope.

As has been argued, particularly by green political theorists and economists, public goods and externalities present pervasive limits to market institutions (e.g., Jacobs 1991, Daly 1992, Barry 1999, Benson 2018, and Eckersley 2004). If such problems were uncommon and isolated, then they would not present a significant limit to the substitution of exit for voice. However, in fossil-fuel economies, problems of environmental public goods and externalities are connected to almost all forms of production and market exchange, and their possible costs for human and non-human well-being range from significantly damaging to catastrophic. The profit motive can also exacerbate these problems by incentivizing individual firms to externalize costs wherever possible. These problems are not simply confined to anthropogenic climate change but involve a host of related environmental issues, from biodiversity loss to water/ocean acidification, ozone depletion, and acid rain. Problems of public goods and externalities may represent, therefore, not exceptions to efficient markets but the rule. They are a pervasive dysfunction of markets from which exit is not possible—a dysfunction that technocratic knowledge is essential to solve.

In some cases, Friedman (341) argues that public goods can be fitted with “exit-like” options that would, like exit itself, reduce the need for technocratic knowledge. While “no exit is available from the provision of any public good,” the exit mechanism “can be mimicked in some cases. Thus, an exitocracy might address global warming through something like a carbon tax rather than through something like the Green New Deal, as a carbon tax would not require technocratic knowledge of how to get people to reduce their carbon emissions: the discovery of how to do so would be left to individual agents seeking to reduce the tax they pay” (ibid.). However, as he recognizes, policy tools of this type still require significant forms of technocratic knowledge: “the size of the carbon tax, and many other details, would have to be set through ordinary technocratic deliberation about the human behavioral consequences to be expected from different policy designs” (342). Such policies therefore require technocratic knowledge to be successful. Friedman (ibid.) also acknowledges a class of public goods where exit-like policies are not even possible. While he defines “secondary” public goods as those that allow for exit-like mechanisms, “tertiary” public goods are resistant to exit options.¹² Military defense and foreign policy, for instance, do not offer any scope for exit and their provision must be wholly voice-based.

Even when exit-like policies are possible, however, they may in some cases require more problematic behavioral predictions than standard technocratic policies do. Determining which level of taxation will reduce natural gas fracking to the desired level, for instance, may be more difficult than directly limiting fracking production to that level through regulation. While the latter mandates a certain behavior to achieve a desired target, the former requires the policy maker to predict the individual’s responses to different increases in costs. Of course, calculating the desired target requires technocratic knowledge, and there may be unintended consequences, but this is true for both the exit-like policy and the more traditional technocratic policy. Similarly, if we want manufacturers to stop using environmentally damaging or inefficient technologies, and replace them with better alternatives, then simply banning the undesirable technologies may involve less significant technocratic judgments than predicting at which level of taxation manufacturers will change their behavior.¹³ It is not always clear, then, that introducing more exit options into the provision of public goods will reduce our reliance on technocratic knowledge, and it may even increase it. Public goods and externalities thus exemplify a pervasive dysfunction of markets that does not allow for exit, and for which exitocracy will be dependent on voice and technocratic knowledge as much as will an ordinary technocracy.

Another set of exit-resistant problems, which Friedman does not directly consider, stems from the inability of individuals to choose effectively among exit options (Akerlof 1970; Benson 2019a; Stiglitz 1991 and 1994). The ability to exit requires not just entry options but knowledge of which of these options are better for one’s well-being. However, there is a whole range of areas in which individuals do not possess, or are unlikely to possess, the knowledge required to make effective exit decisions. Take, for instance, decisions relating to one’s health.¹⁴ There can be significant health effects associated with the use and manufacture of various products and services, but of which consumers and workers will not always be aware. A product might, for example, affect the background risk of someone contracting a health condition and therefore be unnoticeable to the consumer or worker, or alternatively it may only be one among many possible and hard-to-distinguish causes of an illness. This lack of feedback means that the impact of different options on one’s well-being may not be apparent, and individuals will require a large amount of additional information in order to make effective exit decisions (Benson 2019a).¹⁵ They would, for instance, need to know about the different chemicals used in all the different goods and services they may use and the potential effect of these on their health. Obtaining this amount of often technical information creates a significant cognitive and epistemic burden (particularly given the many other demands on our time and energy) that we will be unlikely to overcome.

Similar limits on the ability of individuals to make effective exit decisions can be found in other areas. When it comes to financial products, for instance, the long-terms effects of alternatives are not always easily observable, and individuals often require technical information that they cannot always be expected to have. Failure to possess such information, however, can result in making decisions that are detrimental to one’s long-term well-being. Thus, even an exitocracy would have to rely on voice and technocratic knowledge to solve such dysfunctions. Of course, the epistemic limits facing individuals may also point to an additional role of the exitocratic state in helping to promote exit. The state may provide consumer information to citizens through public information campaigns or mandated forms of product labeling. However, there are significant limits to these methods of communicating information, which is often complex and difficult to understand (Benson 2019a; O’Neill 2012). Such problems can affect such a large number of goods that they place excessive burdens on individuals, who often make a number of market transactions in any one day.

One may argue that while individuals often lack information in certain areas and are therefore prone to making mistakes, these mistakes on the individual level will likely be less serious than the mistakes associated with technocratic intervention and policies. It is not clear, however, that this is necessarily the case. While it is true that any one individual mistake may be less serious than a political mistake, the aggregate effect of large numbers of individual mistakes may be just as serious if not more so. In those areas where individuals significantly lack information, exit may not be the best option when compared to technocratic intervention

Whether in the form of public goods, externalities, or the epistemic constraints facing individuals, then, there are many limitations to using exit to solve the social problems that ordinary technocracy aims to address. Given that technocracy developed to address the dysfunctions of markets and free exchange, this should not come as any great surprise. Many of these dysfunctions directly result from the inability of individuals to make effective use of exit, and it is therefore the failure of exit that often motives technocratic programs in the first place. For many of the problems that ordinary technocracy is interested in solving, exitocracy will simply not offer a viable alternative solution. This fact does not undermine the value of exit-based action when and where it is possible, but it does point to the limitations of exit as a solution to social problems and as a mechanism for avoiding a reliance of technocratic knowledge.

IV. TECHNOCRATIC KNOWLEDGE AT THE HEART OF EXITOCRACY

The previous section highlighted the need for technocratic knowledge in those cases where the ability to exit, or to make effective epistemic use of exit, is limited. It did not, however, challenge the epistemic benefits that exitocracy may have in cases where exit can be effectively exercised. This section therefore turns to the need for technocratic knowledge in achieving exitocracy’s distinctive mission of promoting and sustaining exit. The promotion of exit requires the provision of certain “primary” public goods, such as redistribution and a system of property rights, which itself require very significant predictions about the behavior of ideationally heterogeneous individuals. Indeed, “the administration of an exitocratic Difference Principle would require the administrators to make behavioral predictions that nobody—no matter how judiciously attentive to ideational heterogeneity—is well positioned to make” (340), and the same is true of “legally establishing and defending the private sphere.”

This requirement for technocratic knowledge at the very heart of exitocracy introduces a dilemma. If Friedman’s critique of technocracy is correct, then “we must accept . . . that the unintended costs of providing public goods, including those created by these essential exitocratic activities, may exceed their benefits, because technocratic reasoning will have to be deployed in the course of these activities” (341). For Friedman, however, this dilemma does not undermine the superiority of exitocracy over technocracy, or the justifiability of providing exit supporting public goods. He argues that while “technocratic reasoning will have to be deployed in providing exitocratic public goods . . .the inevitable errors that result will be offset, to some (putatively significant) extent, by the fact that these public goods will enable people to solve social problems in the private sphere” (ibid.). Thus, Friedman continues, we are still “entitled to say that the mix of benefits and costs should be more favorable in an exitocracy, despite the errors that will undoubtedly plague both exitocrats and ordinary technocrats” (ibid.).

Thus, the dilemma presented by the need for technocratic knowledge in providing exit-supporting goods does not, according to Friedman, undermine a normative hierarchy in which exitocracy is preferable to ordinary technocracy, just as a judicious technocracy is preferable to an injudicious one (319, 343). In this and the following section, however, I aim to argue that the dilemma does, in fact, undermine the primacy of exitocracy in this hierarchy. If we accept Friedman’s critique of technocracy, then we have no reason to believe that attempts to provide primary public goods will increase people’s ability to exit rather than decrease it, and therefore no reason to believe that an exitocratic state will produce any more of the benefits Friedman associates with exit. To see this, however, we first need to spell out exactly why reliable technocratic knowledge is required to determine which actions will increases the ability of individuals to exit. The next section will draw out the consequences of this for Friedman’s defense of exitocracy.

Consider first the redistributive role of an exitocratic state in the promotion of exit. Friedman is certainly right that for many individuals, limited resources and opportunities reduce and may remove their ability to escape their social problems. Knowing if redistribution will improve their ability to exit, however, and knowing which levels of redistribution to enact, requires technocratic forms of knowledge (340). Friedman thinks that redistribution by the exitocratic state should aim to “maximize, for as many people as possible, the opportunity to use exit to solve problems affecting them personally” by allowing them to enter into “as many alternatives as possible” (335), but there are many ways that redistribution at different levels could lead to no improvement or a decline in people’s exit opportunities. Consider, for instance, the following possible scenarios.

1. The exitocratic state increases the level of taxation on wealthy individuals and/or businesses in order to redistribute to those with relatively few exit options. This, however, leads many existing employers to cut back on their demand for labor while others struggle to produce profits and go out of business altogether. Would-be entrepreneurs are also deterred from setting up new businesses, which removes the employment opportunities that would have been created, as well as the potential of these new companies to solve people’s social problems through the services and goods they would have provided. The number of exit and entry opportunities open to individuals is, therefore, reduced.

2. The exitocratic state increases the level of taxation on wealthy individuals and/or businesses in order to redistribute to those without exit options. In seeing these taxes as a threat to their profits, however, businesses react not by reducing demand for labor but by increasing their prices and pushing any additional costs onto consumers. The extra money in the pockets of the worst off is then offset by inflation. Their position in terms of resources for exit is, therefore, left the same as it was before.

3. The exitocratic state decides to redistribute a certain amount of wealth to the worst off in order to significantly improve their ability to exit from undesirable jobs. This increase in income, however, leads many of the worst off to exit the employment market altogether and live off these publicly provided resources. This results in decreases to both economic growth and overall tax revenues, which together reduce the resources available for the exitocratic state to improve exit opportunities in the future.

4. As in the previous scenario, the exitocratic state decides to redistribute a certain amount of wealth to the worst off in order to significantly improve their exit options in terms of employment. However, given the costs of the alternative exit options they face, the worst off find that this increase in their income is insufficient to improve their exit opportunities.

In each of these scenarios, a redistributive intervention by an exitocratic state either fails to improve people’s exit opportunities or it reduces them, overall. I do not claim that any of these four scenarios would result from a state engaging in redistribution; they may well not. However, it is not self-evident that none of these scenarios would result from a particular level of redistribution. Discovering if they would requires technocratic knowledge about how individuals are likely to react to the specific changes in their social environment effected by various types and levels of redistribution. Will employers cut their demand for labor or increase their prices when confronted with high tax burdens? Will would-be entrepreneurs be deterred from creating new companies, jobs, and products by increasing taxation? What are the costs of exit for different groups? Are people likely to leave the job market altogether if they receive a certain level of redistributed wealth? The exitocratic state will need to have answers to such questions if it is to promote the ability of individuals to exit, but discovering such answers requires that one can understand and predict how a large number of ideationally heterogeneous individuals will react to different rates of taxation and social provision.

These questions about the effect of redistribution on exit opportunities take a very similar form to questions about the costs and benefits of technocratic policies aimed directly at solving social problems. Understanding the likelihood of scenarios 1 and 2, for instance, requires predictions about the reactions of employers to higher costs very similar to those required for understanding the effects of taxes and wage regulations. As Friedman puts it, such questions are “no less difficult to answer than the question of how [employers] might respond to various levels of a minimum wage or to other ‘ordinary’ technocratic initiatives” (340). Understanding the likelihood of scenarios 3 and 4, alternatively, requires making predictions about the effects of increased resources/incomes in similar ways to those required to understand the effects of an introduction of, or change to, a social welfare or benefit program. Determining how to promote and not diminish exit opportunities through redistribution requires knowledge about how ideationally heterogeneous individuals will change their behavior in response to a government intervention.

If Friedman’s critique of technocratic knowledge is correct, then, it is unclear whether he should support his egalitarian exitocracy over a classical liberal exitocracy, by the standards of negative utilitarianism.¹⁶ Friedman (322) writes that an egalitarian exitocracy “would undertake wealth redistribution far more ambitious than a universal basic income—redistribution along the lines of Rawls’s Difference Principle—to make the exit mechanism accessible to all.” However, determining whether a basic income or a more ambitious level of redistribution is best for people’s exit opportunities would require exitocrats to make predictions about the reactions of an ideationally heterogeneous population to different levels of taxation and resources.¹⁷ Perhaps an exitocracy with limited redistribution would guarantee everyone a minimal ability to exit while allowing the greatest possible scope for free markets to create new opportunities for entry. Alternatively, perhaps the market is less reactive to state interventions, and therefore more redistribution can significantly improve the exit options of the worst off without reducing the potential of the market to create more options. Determining which of these scenarios is likely, and therefore which form of exitocracy is preferable, requires significant technocratic judgments.¹⁸

Another exit-supporting good is the provision and enforcement of a system of property rights. As with redistribution, Friedman recognizes that the provision of such goods requires technocratic forms of knowledge (340-41). To see the problems this produces for his defense of exitocracy, it is important to understand the role of technocratic knowledge in determining whether a system of property rights will actually promote exit and its associated benefits. Alternative property systems can have varying effects on the ability of people to exit, and these effects can be negative as well as positive (Greenwood 2007). Consider, for instance, the following possible scenarios.

5. The exitocratic state provides a system of property rights that includes rights to intellectual property. Although the exitocrats predict that this will provide incentives for technological innovation, the unintended result is the creation of large monopolies in certain industries (such as those that suffer from network effects), which reduce competition and therefore the exit options of individuals within them.

6. The exitocratic state provides a system of property rights that recognizes extensive common-property regimes. Although the exitocrats predict that this will provide many people with access to property, the unintended result is to increase the cost of exit, as property rights become less flexible.

7. The exitocratic state develops a system of property rights that does not grant the state significant powers of eminent domain. Although the exitocrats predict that this will give individuals the greatest scope to use their resources to improve their condition, the unintended result is to make the provision of secondary and tertiary public goods more difficult, resulting in greater costs from which individuals cannot escape.

Again, I do not wish to claim that any of these scenarios would actually result from these alternative property systems; they may not. However, determining whether or not these problems will come about, and which rules of property will produce them, requires technocratic forms of knowledge on the part of the exitocratic state. Factors such as the balance of private and common property, the status of intellectual property, and the extent of (or need for) eminent domain can have very different effects on the efficiency of a property-rights system, and thus on the ability of individuals to exit from problematic arrangements. Determining which bundles of rights and duties will promote or decrease the ability of individuals to exit from their social problems, like judging the consequences of redistribution, requires making predictions about the behavior of heterogeneous individuals.

Different property systems can, for instance, produce different levels of market concentration and different risks of monopoly, as can be seen in scenario 5. These factors directly affect the number of exit and entry options consumers and employees face. Even if property rights are evenly distributed to begin with, in order to avoid concentration, markets can lead to large levels of inequality over time, which can threaten the ability of the worst off to exit. An exitocratic state would, therefore, need to make long-term predictions about the behavior of individuals in order to compare alternative property systems in terms of their results for exit. Of course, an exitocratic state may instead rely on competition policies (e.g., antitrust law) in the event of the unexpected development of an oligopoly or monopoly. However, the use of such policies would leave the exitocratic state in a very similar position to a technocratic state in needing to judge the consequences of policy interventions in the market economy. It would need to know how a given intervention would affect the behavior of a large number of ideationally diverse individuals, and the resulting effect of this on overall exit opportunities.

One might reply that although the choice between competing legal systems would require technocratic knowledge, we know that the exitocratic state must provide some system of private property rights in order to create an environment in which individuals can utilize exit. It does not matter which system the exitocratic state chooses to provide, as any system of property rights would promote the ability of individuals to exit in comparison to having no system at all. However, the claim that any one state-provided private property system is better for people’s exit opportunities than no state-provided system is itself a claim that requires technocratic knowledge to justify. There are examples of many ineffective state-provided legal systems, as well as examples of effective anarchist systems that manage the use of resources outside of the authority of any state, whether exitocratic or otherwise (Scott 1998 and 2009). It is possible, then, that there are some circumstances in which the best thing an exitocratic state can do to promote exit opportunities is not to define or enforce any system of private property rights at all. Knowing whether the provision of a particular property-rights system actually increases the ability of individuals to exit therefore requires significant amounts of technocratic knowledge on the part of the exitocratic state, just as does knowing how to redistribute property.

V. THE DILEMMA OF TECHNOCRATIC KNOWLEDGE

The dilemma facing exitocracy is that its distinctive mission of promoting and sustaining exit itself requires the prediction of human behavior and therefore technocratic knowledge. For Friedman, however, this dilemma should not stop us from preferring exitocracy to ordinary technocracy. He argues that although ordinary technocratic judgments are required in the provision of exit-supporting goods, “the inevitable errors that result will be offset, to some (putatively significant) extent, by the fact that these public goods will enable people to solve social problems in the private sphere,” and thus “the mix of benefits and costs should be more favorable in an exitocracy” compared to a technocracy (341). So although it is not possible to know how much better an exitocracy could be, we can know that it is likely to be somewhat better, overall. The normative hierarchy in which we should “prefer exitocracy to ordinary technocracy and judicious to injudicious technocracy” therefore remains intact (343).

The discussion of the previous sections, however, gives reason to question this claim.

If technocratic knowledge is required to understand whether the actions of an exitocracy will actually promote or undermine exit, and if we accept Friedman’s critique of such knowledge, then we can have no confidence that an exitocracy will provide any of the benefits associated with exit. In other words, we can have no confidence that the errors of an exitocracy will be offset by the resulting ability of people to exit, as Friedman claims.

Exitocracy is an extraordinary form of technocracy, which means that while it is similar to an ordinary technocracy, it is distinctive in the fact that it intervenes to increase exit opportunities whenever possible. Without reliable technocratic knowledge, however, it is not at all clear that these attempts at exit promotion will in any one case, and therefore overall, increase the ability of individuals to exit. These interventions may, in fact, lead to fewer exit opportunities overall than there would have been otherwise, as shown in scenarios 1-7 above. Thus, just as we can have no confidence that an ordinary technocracy can improve the human condition in the absence of reliable technocratic knowledge, we can have no confidence that an exitocracy can improve the ability of individuals to exit, and therefore enable these individuals to improve their own condition.

We are not therefore only ignorant of how much better an exitocracy would be, but also ignorant of whether it will be better or worse overall. In the absence of reliable technocratic knowledge, an extraordinary technocracy that attempts to promote exit may not result in a society where exit opportunities are more readily available compared to a society with an ordinary technocracy. Whether it is a change to the level of redistribution or an alteration to the property system, we cannot be sure that any exitocratic interventions will lead to more exit opportunities than there would be otherwise. We therefore have no reason to think that the errors of an exitocracy will be offset by the resulting benefits of exit, or that the mix of costs and benefits will be more favorable in an exitocracy compared to an ordinary technocracy. Without some reliable technocratic knowledge then, it is simply unclear which system is preferable. In other words, it is unclear whether the state should provide primary public goods, if the critique of technocratic knowledge given in Parts I-II of Power Without Knowledge is correct.

It could be claimed that an exitocracy would at least be preferable to ordinary technocracy when it is successful. A successful exitocracy would increase exit opportunities and therefore supplement its less reliable technocratic knowledge with the more reliable experimental knowledge of individuals in the private sphere. For exitocracy to be successful, though, it would need to have access to technocratic knowledge of some level of reliability (as established in the previous section). It would need to be able to make accurate predictions about how a population of ideationally heterogeneous individuals respond to its attempts to provide exit-supporting goods. If this level of technocratic knowledge existed, however, then it might also enable an ordinary technocracy to directly solve certain social problems. If predictions of human behavior can be made with sufficient reliability that an exitocracy can successfully promote exit options, then a technocracy may also be able to design successful technocratic policies.

A successful exitocracy may then allow more opportunities for individuals to exit and make use of their experimental knowledge. The problem is that these uses of experimental knowledge are parasitic on a prior use of reliable technocratic knowledge in the supply of exit-supporting goods. If technocratic knowledge can be used reliably in this way, then technocratic solutions may also be effective, and it is unclear whether it is best to use voice or exit in any given case. In other words, it would be unclear whether a successful exitocracy is better than a successful ordinary technocracy, given that the success of either presupposes access to reliable technocratic knowledge.

If Friedman’s critique in Parts I-II of Power Without Knowledge turns out to be correct, then, I see no reason to rank exitocracy above ordinary technocracy in the normative hierarchy. Whether we are talking about institutional practice or ideal types, or whether we have access to reliable technocratic knowledge or are ignorant of it, it is not clear that an exitocracy will produce better outcomes than an ordinary technocracy. To make any claim about the relative superiority of technocracy and exitocracy, or about which combination of exit and voice is preferable, we would need to know more about the kinds of technocratic knowledge we can access, the conditions under which we can possess it, and the kinds of social interventions such knowledge makes possible. Whether through voice or exit, some form of technocratic knowledge will be required to deal with our persistent and significant social problems, and we therefore need to further investigate when and where such knowledge can be obtained, and what it allows us to do.

One can, of course, follow Friedman in making an “Adornoesque” cultural critique (344-45). Power Without Knowledge calls out the epistemically naïve attitudes, interpretively uncharitable theorizing, blindly positivistic research, and simplistic heuristics with which technocrats “injudiciously” approach politics and public policy, and indeterminacy over whether exitocracy can help alleviate the problems of ordinary technocracy may place additional emphasis on the need for a cultural revolution in the direction of a more judicious technocratic politics. The extent to which one supports such a revolution will depend on how convincing one finds Friedman’s critique of currently exiting technocracy, and I have simply assumed this critique for the purpose of this paper. What I have argued is that if this critique turns out to be correct, then it will be inconclusive whether we should prefer exitocracy to ordinary technocracy. By elaborating on the epistemic quandaries of public good provision, this paper has questioned whether the problems that Friedman sees as facing ordinary technocracy can be reduced by having the state prioritize exit over voice. Determining the relative role of exit and voice will require further investigation into which kinds of reliable technocratic knowledge can be obtained and which kinds social interventions can be made of the basis of such knowledge.

It is possible that such an investigation will find that no form of reliable technocratic knowledge can be achieved either now or in the foreseeable future. However, an appropriate response to this ignorance would not be exitocracy, which itself requires such knowledge, but perhaps an alternative such as some form of pure procedural legitimacy. If we cannot possess any kind of technocratic knowledge, then we may have to give up hope of solving our social problems through the political or economic system, and the only remaining basis for a just politics would be its procedural fairness. Alternatively, however, if we can possess some technocratic knowledge, then we need to know what form this knowledge takes before we can rule in favor of ordinary technocracy, exitocracy, or any combination of the two.

Notes

1 Friedman (108-9) does argue that technocracy has a tendency to move in the direction of elite rule or epistocracy if people determine that specialized knowledge, not available to the general public, is required to address social problems. As a matter of definition, however, technocracy, on his view, is compatible with a diversity of different forms of rule.

2 The main focus of the essay is the extent to which an exitocracy can avoid reliance on technocratic forms of knowledge, rather than Friedman’s direct critique of the likelihood of such knowledge. A brief account of this critique will, therefore, suffice for our purposes, and I will simply take the critique as given for the rest of the paper.

3 People’s ideas are also open to change over time, which adds another layer of difficulty to the prediction of an individual’s behavior. Even if you know someone’s ideas at time T1, this does not mean you can necessarily predict their behavior at time T2, as her web of beliefs may have changed (143).

4 The first of these, which Dowding and John 2012 refers to as “geographical exit,” can be more costly, as it requires moving locations. DeCanio 2014 argues that this is also a less epistemically valuable form of exit, as it involves making comparisons of services across different jurisdictions, and therefore different conditions, rather than within one jurisdiction.

5 Throughout Chapter 7 Friedman (330-31, 322-23) tends to conflate the distinction between exit and voice with the distinction between the private and public spheres.

6 For a direct critique of these Hayekian arguments for markets, including those that focus on dispersed and tacit knowledge, see Benson 2019a and 2019b.

7 Hayek also appeals to the ability of firms to learn from and imitate each other’s successes. For many classical liberals the ability of firms to emulate each other’s successes and avoid their failures is an important part of the market’s epistemic value, as it allows good ideas to spread throughout society. Given Friedman’s pessimism about our ability to uncover the causes of and solutions to social problems, I suspect that he would question how effective any process of imitation or emulation could be. Even if some firms are more successful than others, it may be difficult, on Friedman’s view, for other firms to work out how they are successful and copy their methods.

8 For a critical comment on these claims see Benson 2019a.

9 These terms are my own and are not used by Friedman to describe his preferred system or any other.

10 It is important to note that these epistemic critiques of classical liberalism do not reject markets and their epistemic benefits outright. Rather they tend to point to their limits and the need for a greater role for the public sphere.

11 Goods are often not simply public or private but rather fall somewhere between the two. The argument of this section covers goods that tend to exhibit significant publicness characteristics.

12 The labels “secondary” and “tertiary” do not therefore speak to the importance of these goods, but to the extent to which they can utilize exit.

13 See Jacobs 1991 for examples of the problems facing these kinds of policies.

14 Friedman (333-334) does discuss medicine and pharmaceuticals, but only in relation to the problem of irreversibility or repeatability, and not in relation to the feedback problem I discuss here. In these passages Freidman rightly points out that certain products have possible fatal effects from which individuals could not possibly exit.

15 I have elsewhere referred to goods which exhibit this lack of feedback as low feedback goods Benson 2019a.

16 We may also judge these systems in terms of other normative criteria, such as standards of justice.

17 The very same problem faces classical liberals who wish to resist significant levels of redistribution.

18 Friedman’s book opens the possibility of further research comparing different kinds of exitocracy, including varieties beyond the egalitarian and the classical liberal. There could, for instance, be a libertarian exitocracy which involved no redistribution at all, or a radical egalitarian exitocracy which involved redistributions along the lines of strike equality (Carens 1981; Cohen 2009).