1 Introduction

The knowledge problem denotes the idea that for order—coordination—to emerge, there needs to be both a way to utilise the existent but dispersed and often local and tacit knowledge held by different people as well as a mechanism for generating new knowledge. It has traditionally been posed as a challenge for socialism: how can central planners meet these two challenges and produce cost-efficiently what consumers want? In his book F. A. Hayek and the Epistemology of Politics: The Curious Task of Economics, Scott Scheall argues that policy makers also face epistemic burdens when pursuing the transition to a liberal society. These epistemic burdens, he suggests, have not received sufficient attention.

I assess these epistemic burdens and compare them to the epistemic burdens involved in centrally planning an economy. I tentatively conclude that there is a difference in degree: it is more challenging to plan the economy than it is to achieve liberal transition. This is chiefly because centrally planning the economy requires targeting an endlessly shifting equilibrium, meaning that there is unexpected change, while planning for the transition to a liberal society has a fairly stable goal. Still, the adjustment to local circumstances, the collective action problems involved in getting to the desired liberal society, and especially the difficulties of creating the necessary cultural preconditions pose serious, and in the last case perhaps insurmountable, epistemic burdens.

I begin by briefly recapping the knowledge problem. Against this backdrop, I consider Scheall’s broader concept of epistemic burdens and convey what he sees as necessary for overcoming epistemic burdens and thus for successful action. In Section 3, I note the ubiquity of epistemic burdens and point to potential differences between different challenges. Moreover, I examine the epistemic burdens, or the knowledge problem, involved in central planning and then examine the epistemic burdens of liberal transition. Based on this analysis, I suggest in Section 4 that it is easier to liberalise than to centrally plan the economy—a conclusion that is motivated and illustrated by empirical findings. Section 5 concludes.

2 The knowledge problem and epistemic burdens

By clearly outlining the economic challenge for socialism, von Mises (1920, 1963) initiated the socialist calculation debate, out of which the concept of the knowledge problem emerged. Mises explained that in a market economy there is exchange. Individuals’ knowledge enters into their exchanges and thus into prices. These prices can serve as “aids to the mind” (von Mises, 1963, p. 102), allowing monetary calculation and thus helping entrepreneurs to assess past projects and forecast future projects in terms of their economic feasibility. With prices as aids, people can, albeit imperfectly, coordinate their actions. In a planned economy, however, there is one single plan and no exchange of the means of production, implying there will be no prices for them. Therefore, socialism will not exhibit economic calculation, and thus there will be no coordination and no order, Mises concluded.Footnote 1

Kiesling (2015) draws attention to two distinct elements of the knowledge problem: the “complexity knowledge problem,” which refers to the challenge of coordinating individual action in light of “dispersed private, subjective knowledge,” and the “contextual knowledge problem” which is the challenge that “some knowledge relevant to such coordination does not exist outside of the market context” (Kiesling, 2015, p. 46).Footnote 2 The second problem is decisive; solving the first problem would not be enough, as there is knowledge which does not emerge unless there is a specific institutional structure that gives rise to the market process. Only this process allows and incentivises via entrepreneurial alertness “discovering the existence of something valuable, the very existence of which was hitherto wholly unknown” (Kirzner, 1997, p. 75). Additionally, there exists tacit knowledge (Polanyi, 2009). As this is inarticulable, those who possess it need to be free to act on it if it is to be employed at all.

The knowledge problem is an obstacle to the successful coordination of people’s activities. Solving it is a means, an intermediate step, towards coordinating people’s efforts. Coordination implies there is order, which means that people’s actions are usually successful and people’s most important wants are fulfilled. It is, as Bastiat suggested, the phenomenon that Paris gets fed: “each succeeding day bring[s] what is wanted, nothing more, nothing less, to so gigantic a market” (Bastiat, 2007, p. 273). An example of disorder is the breakdown of the Soviet economy in 1920 and 1921, resulting in a horrible famine. People want coordination, as this means their attempts at cooperation succeed and they are best able to satisfy their wants—which is the ultimate end of cooperation.

In his work, Scott Scheall emphasises epistemic burdens, a wider concept than the knowledge problem, which refers solely to the epistemic challenges involved in economic coordination. Epistemic burdens arise whenever people act, and thus the concept is more comprehensive. According to Scheall, epistemic burdens refer to “the nature and extent of the ignorance that an actor in a particular context must overcome in order to use some means to deliberately realize some end” (2020, p. 3). People encounter epistemic burdens when they want to buy a new pair of sneakers or purchase a new home. Policy makers, by implication, also always face epistemic burdens, for example, when they attempt to smooth the business cycle or try to increase the supply of affordable housing.

To overcome epistemic burdens, Scheall explains, one must meet some challenges: “It is possible to deliberately realize a goal, any goal, via policymaking if and only . . . if policy makers possess both a theory and data sufficiently comprehensive to make, implement, and effect a plan for the realization of the relevant goal” (Scheall, 2020, p. 162). Consequently, it is necessary to have a functioning theory about the means-ends relationship, that is, about what means is suitable to achieve the end sought. Provided that one has such a comprehensive theory, it is necessary to possess the data to fill into the theory and specify what exactly one must do to achieve what one wants under the specific circumstances. Now, if both theory and data are present, then one needs to possess the skills, or the know-how, to successfully take the action.Footnote 3

Scheall lays much emphasis on the word “deliberately.” His elaborations imply that one may also achieve one’s goal, yet not deliberately, if “spontaneous forces intervene” (Scheall, 2020, p. 178). So, policy makers may achieve their goal of increasing the supply of affordable housing even though they lack the necessary knowledge if market forces aid them or if they are aided by pure chance. However, it is essential to remember that acting is always future oriented and thus subject to uncertainty, which simply denotes that actors must always be ignorant to some degree: “The uncertainty of the future is already implied in the very notion of action” (von Mises, 1996, p. 105). While actors’ beliefs must be adequate for their actions to be deliberately successful, it is part and parcel of acting that the basic uncertainty surrounding our actions can never be eliminated. This means that no actor can know beforehand that her action will be successful or what the future will look like—in other words, actors must always be ignorant to some degree. In a strict sense, then, we can never deliberately realise a goal, as uncertainty and ignorance are inevitable: we envision a future and act accordingly, but we cannot be sure of success. In Kirzner’s words, “for us uncertainty meant the essential freedom with which the envisaged future may diverge from the realized future. Entrepreneurial alertness means the ability to impose constraints on that freedom, so that the entrepreneur’s vision of the future may indeed overlap, to some significant extent, with that future that he is attempting to see” (1982, p. 156). Kirzner suggests that, as entrepreneurs, we can shape the future to some extent, but not absolutely. Some degree of doubt or uncertainty must always remain, and the fact that actors do not have complete knowledge, that is, are to some degree ignorant, of the means-end-relationship does not imply that it is unwise to act or that the epistemic burden is too high.

Therefore, it would be wrong to assume that an action is only advisable if we know it will lead to success—this would actually imply paralysis. Rather, it follows that what we can do is rule out certain actions as unadvisable, meaning that “the value of empirical political epistemology is likely to be more negative than positive” (Scheall, 2020, p. 163). Thus, examining epistemic burdens will primarily tell us which actions will not succeed, and not which actions will, surely, succeed. While recognising this, it is important to note that from the principle that “nothing that we cannot know enough to deliberately realize can be an obligation” (Scheall, 2020, p. 3), it does not follow that policy makers must only be obligated to pursue what they can, surely, achieve.Footnote 4 There is, instead, scope for action also in these circumstances, which is, as Schliesser highlights, crucial. For if policy makers, or other actors, only acted when they had the knowledge to achieve their goal, then—if they acted at all—this would raise “the problem that we might never discover what works” (Schliesser, 2021, p. 37).

3 The nature and magnitude of epistemic burdens

3.1 The ubiquity of epistemic burdens

Perhaps the overarching theme of Scheall’s work is the ubiquity of epistemic burdens. These burdens arise not only when centrally planning the economy but whenever people act, including when the goal is to liberalise society. One of his central messages is therefore that supporters of liberal societies must grapple with the epistemic burdens of liberal transition. He emphasises that

from a political-epistemological perspective, to aim at supplementing, rescinding, or modifying the general rules from which a social order emerges is not a different kind of thing from aiming at a particular state of affairs, say, at a specific distribution of society’s resources. Both require knowledge that may not be easily acquired by human policymakers; both may require surmounting insurmountable epistemic burdens (2020, pp. 83–84; his emphasis).

This follows from the fact that in all cases someone acts. Therefore, to succeed it is necessary to possess adequate knowledge—if spontaneous forces do not intervene. That policy makers, consequently, also need to overcome epistemic burdens when planning liberal transition is also the perspective of Douglass C. North, who mentions “the inherent difficulties involved in deliberately attempting to alter the societal framework with the very imperfect knowledge of the players” (2005, p. 8). Scheall adds critically that transitioning may be a challenge similar to that of centrally planning: “Indeed, it is not obvious that would-be liberalizing policy makers are any better epistemically equipped to deliberately realize an effective liberal order than socialist central planners are epistemically equipped to deliberately coordinate supply and demand” (2020, p. 79). This has stirred a critical response. Vanberg, in a review of Scheall’s book, objects:

A closer look at the arguments Scheall offers in support of his claim suggests that he tends to confound two obviously quite different matters: namely, on the one hand, the epistemic burdens involved in operating a liberal vs. a centrally planned system, and, on the other hand, the epistemic burden involved in implementing or setting up these systems (2021, p. 654).

Vanberg continues that these two matters “are categorically different” (2021, p. 654). However, as epistemic burdens arise in both cases, they are within the same category—and this is Scheall’s claim: “These are general facts about human action, which, in all contexts, can deliberately realize its goals, whatever they might be, only to the extent that the knowledge upon which action is based is adequate” (2020, p. 34).

It is true that epistemic burdens differ, and it is possible that Scheall underestimates the differences between these two challenges in the quote from page 79. In other places he, perhaps, is more cautious, claiming that “from a political-epistemological perspective, there is, at best, a difference in degree between these prospective goals of economic policymaking” (Scheall, 2020, p. 100). But this difference in degree can be quite large and decisive. That epistemic burdens differ is already evident in von Mises’s, 1963 paper (originally published in 1920) in which he states that some problems can be solved without prices: “As a rule, the man who knows his own mind is in a position to value goods of a lower order. Under simple conditions it is also possible for him without much ado to form some judgment of the significance to him of goods of a higher order” (1963, p. 96).

When coordination is necessary,Footnote 5 there is a problem of generating and utilising the necessary knowledge to cooperate effectively. However, not all knowledge problems are of such a magnitude that they cannot be solved centrally, by a single mind. The magnitude, or extent, of the knowledge problem, and epistemic burdens more generally, is decisive, i.e., how much knowledge is necessary.

Next to the magnitude, the nature of an epistemic burden is also essential. This refers to the kind(s) of knowledge that is necessary. I have written above that there is theoretical knowledge, empirical knowledge (or the data to fill into the theory), and the practical skills or practical knowledge. An epistemic burden, then, differs according to what kinds of knowledge—theoretical, empirical, or practical—are necessary as well as what specific theoretical, empirical, and practical knowledge is necessary. For instance, building a plane requires a lot of theoretical knowledge, but driving a bike does not—here, practical skills are needed. And the theoretical knowledge required to build a plane (e.g., knowledge in physics) differs from the theoretical knowledge that I need to set up a homepage (e.g., knowledge of C++).

The magnitudes of epistemic burdens can, and usually do, differ. But if two epistemic burdens have the same nature, comparing their surmountability is rather facile. Accordingly, it is uncontentious that it is easier to centrally plan an economy with ten thousand inhabitants than one with one million inhabitants. The comparison gets difficult once, because different kinds of knowledge are required, the nature of epistemic burdens differs as well. Any conclusion regarding the relative weight of the epistemic burdens may then need to remain tentative. As Scheall summarises this point:

We’re always thinking about both the nature and extent (magnitude) of the ignorance a policymaker needs to surmount, and the fact that there are different kinds of ignorance means that we can’t simply weigh up epistemic burdens in the balance. But, it doesn’t follow that, because epistemic burdens cannot be quantified, we cannot get an intuitive sense of whether one epistemic burden might be easier to surmount than another” (S. Scheall, personal communication, Sep 9, 2023; his emphasis).

In the remainder of this paper, I try to get an “intuitive sense” of the relative epistemic burdens of central planning versus liberal transition. I aim to show that there is a significant difference in degree between these challenges, meaning that the epistemic burdens of liberal transition are lesser than those of centrally planning an economy. To make this argument, I first consider what makes the epistemic burdens of central planning so severe and then examine the epistemic burdens involved in liberal transition. The ensuing comparison is difficult, as the epistemic burdens differ in nature. Despite this, I believe that some tentative conclusions can be drawn—informed and illustrated by empirical findings.

3.2 The planned economy

Those who want to plan the economy need to overcome immense epistemic burdens. Planners face a coordination problem in which they must use production goods to best satisfy consumer wants. In accordance with the threefold requirement for successful action, planners need firstly a theory of how to coordinate production for satisfying consumer wants, secondly the data of both consumer wants and the production goods, and finally the practical skills to implement their data-filled theory. Criticism of central planning has traditionally focused on the data problem, and not the theory and the implementation problem.Footnote 6 I will follow this path. The data challenge consists of the necessity to firstly generate new knowledge about consumer wants and production goods, and secondly of the need to employ the existent but dispersed and often tacit knowledge. Planning the economy is a question of producing together, and, as Lavoie explains, “the planners would have to grasp the whole of a complex and changing production process” (2015, p. 66). The magnitude, or extent, of the epistemic burden of planning an economy therefore largely depends on the complexity of the task, and the extent and speed of unexpected changes.

Referring to Hayek, Scheall explains that “phenomena are complex to the extent that we cannot explain, predict, and control them” (2020, p. 39). Seen from this angle, any phenomenon that is complex, by definition, poses an insurmountable epistemic burden. I understand complexity slightly differently here. Following Page, “by complexity, I mean elaborate temporal and spatial patterns and structures” (2011, p. 6), in which the usually heterogenous entities interact and are interdependent. The interaction is decisive. Only once “the parts begin to connect with one another and interact more,. . we move from the realm of complication to complexity” (Miller & Page, 2007, p. 27). These “agent interactions become highly nonlinear” (Miller & Page, 2007, p. 10) and, consequently, the ensuing “complex phenomena are hard to describe, explain, or predict” (Page, 2011, p. 6). However, they need not be impossible to describe, explain, predict, and manage. Decisive, then, is whether the phenomenon is too complex to be dealt with by deliberate action. This is in line with Rescher who notes that “the more complex something is the more difficulty we have in coming to grips with it and the greater the effort that must be expended for its cognitive and/or manipulative control and management” (1998, p. 8). But certainly, there are many phenomena which are too complex to be grasped, predicted, and managed.

The complexity in planning the economy arises first from the innumerable individuals involved who want a variety of things and who value them differently. Thus, planning needs to take into account the numerous different valuations. The second factor is that there are so many things that can be used to satisfy the different wants. These means are nonspecific and can be combined in numerous ways, and different means may be suitable for the same purposes, but different purposes may also be achieved by different means.Footnote 7 It is not technological feasibility that matters, but whether some end can be economically achieved with some means. That is, it is necessary “to separate out from among the numerous array of technologically feasible projects those projects which are economically feasible” (Boettke, 2001b, p. 242). This is further complicated because means and ends are not given: action involves not only maximising within a given means-ends framework but also “the drive and alertness needed to identify which ends to strive for and which means are available” (Kirzner, 2013, p. 27). This refers to the contextual knowledge problem—that is, the need to generate new knowledge about what consumers want, and which means are available, as distinct from the need to utilise existent knowledge about ends and means to achieve the most efficient use of resources.

Unexpected change amplifies the complexity challenges for central planning.Footnote 8 That nonspecific elements change incessantly and in unpredictable ways—people now want different things and the world changes—creates uncertainty, heavily increasing the magnitude of the epistemic burden. It may be possible to solve a complex knowledge problem if the coordination goal remains static. “With given and constant data such a state of equilibrium could indeed be approached by the method of trial and error,” affirms Hayek (1980, p. 188). But dealing with unexpected change is different, as how to coordinate the elements is not given once and for all but rather changes incessantly and unpredictably. This is amplified since the elements are not given but need to be discovered, or knowledge needs to be generated about them, each second anew.

Scheall has written that people need a workable theory, the data to fill the theory, and the know-how or capacity to actually implement their plan in order to deliberately realise their goal. From what I have said, it follows that it is the sheer complexity and massive unexpected change of the data that is so challenging for would-be economic planners. However, the epistemic burdens the data problem creates vary with different economies: “Whereas Crusoe can easily survey the whole production process, Mises argues that no single person could survey the production process of the modern economy” (Lavoie, 2015, p. 61). Only when many people cooperate and the division of labour becomes more intensive can nobody any longer assess the complete production process—what people want and what means are available and efficient—in light of not only dispersed but also contextual knowledge that must first be generated. “In none but the most simple kinds of social order it is conceivable that all activities are governed by a single mind,” elucidates Hayek (1964, p. 8). Thus, modern economies come with massively higher epistemic burdens than the small economy of a hunter-gatherer tribe, and these data-induced burdens are massive. Moreover, although this discussion focused the “data problem” (how to generate and collect the necessary knowledge about consumer wants and production goods), there is also— as mentioned before—a “theory problem” (how to predict the economy and develop a plan for action, even given the data) (Scheall, 2020, pp. 39–41), and finally a capacity problem (how to practically implement the plan).

3.3 Liberal transition

In light of epistemic burdens, Scheall suggests that “liberals need a theory of liberal transitions, that is, a theory of how more liberal contexts can be realized starting from relatively illiberal circumstances” (2020, p. 76; his emphasis). I understand liberal transition as societal change from a relatively illiberal to a relatively liberal society, while “priority is given here to institutional change” (North, 2005, p. 1). This could result in a society that constitutionally constrains policy makers to those things they can, realistically, achieve (Scheall, 2020, pp. 174–175). To gain a good understanding of the epistemic burdens involved, I now attempt to delineate what policy makers need to know in order to deliberately achieve liberal transition. They first must successfully tackle a theory problem and develop a vision of the order they want to realise, which, moreover, they need to tailor to particular societal circumstances for which they face a data problem. And second, liberalisers need to have a practical idea of how to get there and be capable of actually getting there: this again involves a theory problem, since policy makers need to come up with a theory of how they can realise their vision, but also a data problem, as policy makers need to adjust their envisioned means to societal circumstances, and finally a capacity problem, since policy makers need to actually be able to successfully implement their plan. As with the planned economy, I focus on complexity and unexpected change as decisive for the epistemic burden. In the course of the discussion, it will become clear that the main difficulty is not so much to know the desired order but rather devising a way how to get there.

3.3.1 The vision

The first challenge for liberal policy makers is to develop a vision of the desired society. They need a general, abstract understanding of the desired social order, meaning they face a theory problem. And they need knowledge of the specific societal circumstances in relation to which they need to adapt this societal blueprint—a data problem.

While the specific features of any viable liberal society are subject to discussion, there is agreement on its kernel, and this is abstract and permanent knowledge, implying that the theory problem is not so massive. To Scheall, it is an order that is liberal “in the sense of limited government under the rule of law, respectful of individual freedom and the congeries of rights associated with the ownership and transfer of private property” (2020, p. 77). There is knowledge of the rule of law, of many crucial norms and rights such as private property rights. In addition, there are societies in which these institutions predominate that can serve as models for policy makers. In general, as the desired liberal order is a specific, given goal, complexity is significantly reduced. Now, there may be disagreement about the details of the liberal order, such as disputes between minimal-state supporters and those who support moderate government interventions. But these disputes do not need be resolved by the liberal policy maker, who simply pursues her preferred design. Problems may arise only to the degree that the liberaliser faces obstacles from those who want a different liberal society; this leads to the collective action problems that I discuss below.

Nevertheless, there is complexity, created especially by the need to adapt the institutional framework to the particular society, what Scheall calls the “appropriateness in prevailing environmental conditions” (2020, p. 84), and what implies a data problem. Adapting the well-known kernel to the specific societal conditions is complicated by the fact that “the same set of rules of individual conduct may in some circumstances bring about a certain order of actions, but not do so in different external circumstances” (Hayek, 1967, p. 68). Local and temporary knowledge is necessary to ensure that the known blueprint for a liberal society is successfully adjusted to the specific societal conditions.

There is another issue. The “foundation of these flexible institutions” that allow “adaptive efficiency . . . resides in widely held beliefs embodied in the informal constraints of society” (North, 2005, p. 78). Therefore, before fine-tuning the liberal society, “the required cultural pre-conditions will first have to be realized, a potentially quite epistemically burdensome goal” (Scheall, 2020, p. 80). Boettke et al. use the Greek term metis to denote a similar concept. Metis signifies “the glue that gives institutions their stickiness” and “includes skills, culture, norms, and conventions” (2008, p. 338). Just like Scheall, who writes of necessary cultural preconditions, Boettke et al. (2008, pp. 344–345) argue that institutions only have a good chance to stick if grounded in metis. And indigenously introduced endogenous institutions, which are institutions that emerge as spontaneous orders and are not imposed by the national government or even by foreign governments, are those that harmonise with metis.Footnote 9 However, stickiness is “a necessary though not sufficient institutional attribute for creating economic growth” (Boettke et al., 2008, p. 345). Consequently, the existing metis is not necessarily the one that allows for what we typically understand to be liberal societies.

However, there is at least a rough theory of the required cultural preconditions, or metis, suggesting that the theory problem here, again, is surmountable. If we take this to be something akin to “bourgeois dignity” (McCloskey, 2016, 2017) and WEIRD—Western, educated, industrialised, rich, democratic (Henrich, 2020)—people then there is again an at least rudimentarily known destination for liberal transition. An exemplary cultural precondition may be trust (or social capital).Footnote 10 When people trust each other, they exchange peacefully, and rule of law and a liberal order are possible. Another example is general respect for the law or the norms that the law enshrines. However, given this abstract and permanent knowledge of the cultural preconditions, it is again necessary to take societal circumstances into account and adjust the cultural blueprint to local conditions, for which empirical knowledge is required.

While both cultural preconditions and the adapted institutional framework are complex, this complexity is not combined with incessant unexpected change. The desired framework and the sought-after cultural preconditions are largely stable. It is true that these two will change slightly, and, since society and its environmental conditions change, there needs to be a constant process of adaptation in which a data problem needs to be tackled. But there is a lot of stability regarding the surely complex destination which liberalisers seek to reach, massively reducing the epistemic burdens involved in liberal transition.

3.3.2 How to get there

Thus far I have focused on the objective rather than the way to get there. Given that liberalisers know the ideal society and the necessary cultural preconditions, both tailored to particular circumstances, they still need to have an idea of how to realise their ideal—and this will turn out to be the greater challenge. Policy makers thus need to first overcome a data problem by gaining knowledge about the existent culture: “Understanding the cultural heritage of a society is a necessary condition for making ‘doable’ change” (North, 2005, p. 163). Their ensuing task is simplified rather dramatically when the necessary cultural preconditions are in place.

If they are not, policy makers need to find ways to create such a culture. For example, the liberal reformer needs to foster trust between citizens. Referring back to Scheall, one needs a theory of how trust can be nurtured at all, and the data and skill to implement the plan under the particular circumstances. This also holds for other cultural preconditions. I believe this is an excessively big challenge, and one which policy makers will not be able to meet. Boettke et al. seem to think similarly when they note that “it is possible for a new and unique metis to develop over time” but add that “until that new metis is in place, reformers must realize that efforts to impose institutions, whether from within or from without, are likely to fail” (2008, p. 353). North puts it strongly when he writes that “the development of a set of informal institutional constraints” that underlies flexible societies “was more good fortune than intent; and even if we knew their source, they . . . do not appear to be replicable either deliberately or in a short time period” (2005, p. 169). Here, it is necessary to generate new theoretical knowledge of how to create the required cultural preconditions—before both the data and the capacity problem become relevant.

When cultural preconditions conducive to a liberal society are in place, liberalisers still need to change the institutional framework. Often, this will centre on “removing obstacles” (Scheall, 2020, p. 90) that inhibit people’s free cooperation. While liberal policy makers also need some knowledge of how exactly to remove obstacles and unleash spontaneous forces, it seems that this is not particularly demanding or complex. After all, what policy makers need to do is eliminate privileges, revoke measures such as price controls, and so on; as this concerns specific measures which the government had previously implemented, their abolishment does not present a particularly high epistemic burden: “Formal institutions may be altered by fiat” (North, 2005, p. 157). Removing obstacles is therefore not challenging in terms of theoretically knowing what to do as well as gathering the requisite empirical knowledge.

More challenging is the creation of new institutions which may, however, be necessary, or at least advisable. Scheall critically notes that Hayek actually intended this, giving as one example Hayek’s “proposal for a unique bicameral political system” (2020, p. 89). Designing and implementing these new institutions will, usually, prove more epistemically burdensome than solely removing obstacles since, first, there will be old institutions that need to make way for new ones, implying that removal is necessary in both cases. And second, removing should be easier than designing an institution, including adapting it to specific circumstances, as designing is a positive task for which substantial theoretical as well as empirical knowledge will be required.

There is also a capacity problem—in particular, there are issues with collective action, meaning policy makers may lack the power to implement their plans: powerful forces within society will usually tend to oppose the removal of obstacles and the creation of new institutions. “The basic organizational logic of politics conflicts with the logic of economic reform” (Boettke, 1993, p. 7) because the logic of politics is to benefit a clearly specified group in the short term, at the expense of high and long-term costs for a larger group (Olson, 1971). This applies to liberalising. After all, liberalisers intend to create long-term benefits for society at large, while, as a side-effect, usually having to impose losses on certain, often powerful, groups. In this respect, one may speak of the need to overcome path dependencies. “Once a development path is set on a particular course, the network externalities, the learning process of organizations, and the historically derived subjective modelling of the issues reinforce the course” (North, 1990, p. 99) because the existent institutional framework shapes which “kinds of knowledge, skills, and learning” (North, 1990, p. 74) pay off. There will be many organisations that developed in conformity with the institutional setup and that tend to benefit from the status quo, which is why “the status quo applies a powerful political grip” (Peltzman, 2022, p. 12), implying that many will tend to oppose bigger change. A practical example for this is the historical failure of welfare state retrenchment. Pierson notes that “welfare states have created their own constituencies” (1994, p. 2) that successfully opposed the intended societal change.

Not only vested interests but also the polity at large may create status quo bias. Two reasons why people may not want liberal transition are essential. First, people face uncertainty about who will lose and benefit from reform, so that they may oppose any change (Boettke, 1993, p. 76). Second, people may underestimate the benefits of a liberal society since they might have a flawed understanding of what means are appropriate for their ends. In this sense, Caplan and Stringham argue that “the problem with democracy is that the voters are wrong, but heeded” (2005, p. 80). The same may hold for liberal transition when citizens are wrong about its effects, but heeded. Here, policy makers may try to convince citizens, including intellectuals, of the desirability of the liberal society. This is a problem of persuasion which, again, poses an epistemic burden. Alternatively, policy makers may try to compensate those who oppose change: “Compensatory payments . . . may be required to secure general willingness to accept new laws, new rules or restrictions on behavior, or to relax or repeal existing laws” (Buchanan, 2000, p. 148). Or it may be necessary to wait for a shock: “Only a shock to the system coupled with a major reform package can break this hold that existing political structures have on the economic system” (Boettke, 1993, p. 77). One theory suggests that the reform package should be implemented overnight, as there always is a commitment problem, meaning citizens cannot trust their government to really liberalise: “the most effective way out of this impasse and to signal commitment by the regime to liberalization is to reject all notions of gradualism and embrace a radical liberalization program that is implemented overnight” (Boettke, 1993, p. 9).Footnote 11 Therefore, there exists a theory of what to do if there is scope for action, to which policy makers need to add local knowledge in order to successfully implement their plan. However, a theory problem may still exist, as there also are critics of so-called shock therapy (Weber, 2021). Moreover, if obstacles were successfully removed or new institutions implemented, undesirable effects might occur in the immediate aftermath, while desirable effects materialise only in the middle or even long term. In such a situation, policy makers may face continuing challenges in which interest groups demand policy reversal—again, a capacity problem.

Overcoming collective action problems also involves overcoming epistemic burdens. There is, however, abstract, theoretical knowledge about the problems and potential strategies; this knowledge is stable. Of course, any general strategy must be adjusted to the particular circumstances, which requires local and temporary knowledge in a situation with unexpected change. But the challenge is primarily a question of power or capacity, where liberal policy makers may simply be unable to overcome the obstacles.

3.3.3 Summary

The epistemic burdens of liberal transitions are not so much about knowing where to go as about knowing how to get there. Both cultural preconditions and the liberal order are well known and require, mainly, theoretical and permanent knowledge, and they are relatively stable destinations. Though policy makers need to tailor these blueprints to complex and unexpectedly changing societal circumstances, the difficulties involved should not be too epistemically burdensome.

Instead, the main difficulties lie in approaching the desired destination. Here, situations in which the required cultural preconditions exist differ from those in which they are absent. Policy makers need, theoretically, to be aware of the need for cultural preconditions, and then tackle the data problem of assessing society’s culture. For, as North points out, “simply putting in place the formal rules is a recipe for disappointment, not to say disaster” (2005, p. 161). However, although culture changes, it is quite stable, which is why the assessment should not be too difficult. The most difficult challenge for liberal policy makers, it seems, is creating the cultural preconditions necessary for a liberal society. Even if people wanted to be WEIRD, it does not follow that becoming WEIRD is easy, or that it is easy for policy makers to make people WEIRD. After all, the emergence of liberal democracies took a long time and witnessed major setbacks. The challenge involves both a theory and a data problem, but perhaps the theory problem is most daunting, as it is by no means clear whether there exists theoretical knowledge about how to instigate cultural change. Of course, even if adequate theories existed, policy makers would need local knowledge and the know-how to implement cultural change.

Once the correct metis is in place, things ease, but liberal policy makers still face epistemic burdens. People may not be particularly interested in a liberal society, and even if a vast majority is, strong vested interests may oppose institutional change. Dealing with these obstacles involves a power problem but also requires strategic thinking. There are some ideas what to do. For instance, Boettke (1993, pp. 129–130) provides a recipe for shock therapy. Of course, a theory problem may still persist, as it may not be clear that shock therapy works. Moreover, policy makers still need knowledge of the particular circumstances to implement the general recipe. However, if policy makers have the capacity or power to push through their plans, the epistemic burdens do not seem insurmountable, given that there is only moderate unexpected change in some areas relevant to the policy making and much of the knowledge is abstract and permanent.

To conclude this section, I want to reiterate that uncertainty is inherent in acting. Just as no action can guarantee success, no liberaliser can be certain that she will succeed. There is always an element of doubt, and that she knew enough to liberalise her society can, ultimately, only be proven by actually achieving the feat. It is important to keep this in mind when evaluating policy making.

4 Comparing the epistemic burdens

A comparison of the epistemic burdens of liberal transitions with those of central planning must rest on some intuitive evaluations, as there is no way to quantify the epistemic burdens. To evoke some intuitions, I start by examining empirical cases. Based on this, I compare the epistemic burdens, concluding that, while not always surmountable, the epistemic burdens for liberal transitions tend to be lower than those for central planning.

4.1 Empirical illustrations

Examining planned economies is difficult because, it appears, there have not been any.Footnote 12 “The only full-scale attempt [at central planning] was the one undertaken in Soviet Russia during the last six or eight months of 1920; and the results were disastrous” (Polanyi, 1957, p. 35). It was full-scale in that “this system attempted to substitute a unified plan of economic life, i.e. rational social relations of production, for the chaotic and exploititive [sic.] relations of production that existed under capitalism” (Boettke, 2001c, p. 92). However, the socialist system “would not function, and had to be supplemented by a very restricted market business,” and the attempt to “complete the development of the system . . . met with desperate resistance from all sections of the community” (Brutzkus, 1935, p. 105).

Although, in light of the resistance, the Soviet economy presumably was not really centrally planned, it already was a disaster. The Soviet Union back then was “in ruins, its national income one-third of the 1913 level, industrial production a fifth (output in some branches being virtually zero), its transportation system shattered, and agricultural production so meager that a majority of the population barely subsisted and millions of others failed even that” (Cohen, 1980 as quoted in Boettke, 2001c, p. 78). The attempt at central planning was an abysmal failure. Under mounting political pressure, the Soviets reversed course: the New Economic Policy was introduced, and with it the experiment with central economic planning was abandoned, once and for all: “Lenin not only allowed prices and profits to persist, he abandoned the cardinal goal of socialism—the substitution of a settled plan for the anarchy of the market” (Boettke, 2001c, p. 97). Roberts suggests that the Soviet economy, then, generally was “a polycentric system with signals that are irrational from the standpoint of economic efficiency” (1969, p. 175). Polycentric means that the economy was not centrally planned but rather there were numerous independent decision-makers—producing a mixed, not a planned, economy with heavy government intervention. Lavoie summarises the development of the Soviet economy: “the Bolsheviks switched from the goal of completely destroying markets to that of embracing and encouraging them, and then . . . to that of obstructing and disguising them” (1986, p. 10). So, even though the Soviet economy has usually been interpreted as centrally planned, in reality it was not. That planning, having produced abominable consequences, was swiftly abandoned speaks to the enormous epistemic burden of central planning and suggests its insurmountability. And that even the highly interventionist economy that followed the socialist experiment was struggling, even though it benefited from “the existence of world prices upon which Soviet planners could rely in formulating their plans” as well as “internal markets” (Boettke & Coyne, 2004, 81), suggests how massive the epistemic burdens are.

There are several examples, in contrast, of successful liberal transitions deliberately brought about by policy makers. One is the case of West Germany after World War II. It includes not only the economic but also the political transformation of a totalitarian, highly interventionist dictatorship into a liberal democracy that would go on to become an eminent success story. The creators of the German Grundgesetz sought to deliberately apply the lessons of history and safeguard basic rights as well as the liberal democratic institutional framework (Morsey, 2007, p. 20). Meanwhile, the German economy had been liberalised chiefly by one individual policy maker. Economic Director Ludwig Erhard simply decided to abruptly lift price controls; and from one day to the next, goods were available in the shops. Erhard, however, acted without the authority to do so, illustrating the impact a committed policy maker can have and showing that one spirited individual in the right position can make a difference (Henderson n.d.; Kleinert, 2019, pp. 54–55; Tilly, 1993, pp. 206–209). This was shock therapy. While it succeeded, illustrating that policy makers can liberalise society, the economic transition manifests the uncertainty that inevitably surrounds any attempt at liberalising. Erhard could not know, in the sense of ruling out any uncertainty, that he would, ultimately, succeed. Facing massive epistemic burdens, he took a gamble—and succeeded. It should also be noted that Erhard faced severe opposition after the institutional change when there was a massive general strike (Fuhrmann, 2017, pp. 213–217), in line with the danger that people may be dissatisfied with short-term negative effects. But Germany remained a market economy.

Another example of the success of liberal transitions is Japan after World War II. “Americans”, write Boettke et al., “played a key role in rebuilding Japan” (2008, p. 346). In this case, American policy makers were able to set the stage for an unprecedented transformation—presumably because the required cultural preconditions were already in place, as Boettke et al. add. Moreover, the work of Ostrom (2010) demonstrates the capacity of people to overcome collective action problems, although she examines primarily the issue of common pool resources, not societal change. There are also several examples in which mainly the economy was liberalised. One prominent example is Lenin’s New Economic Policy, which was a successful attempt at moderate liberalisation. Another example is the Chinese economy and the reforms implemented by Deng Xiaoping. The Republic of Singapore is also interesting, as it had a very successful, but limited, liberal transition—the country ranks near the top regarding economic freedom, but political freedom is constrained (Vásquez et al., 2022, pp. 318–319). There have also been cases in which attempts at liberalisation failed. Scheall (2020, p. 78) himself mentions two failures: Afghanistan and Iraq.

4.2 Theoretical comparison

Comparing the epistemic burdens of liberal transition and central planning is difficult because of their different nature. But it seems reasonable to distinguish, in the comparison, between the end that policy makers want to achieve and the means they need to employ to get there. In both cases the complexity of the task and the extent of unexpected change demand consideration. Lavoie offers insights for comparing the epistemic burdens. Seeing science as a polycentric order just like the economy (Polanyi, 2000), he suggests that the knowledge problem in the economy is more severe than that in science (Lavoie, 2016, pp. 76–87), as “market participants do not deliberately aim at improving the articulation of price information, their actions are largely oriented toward the uncertain future, and the information they use is normally of only temporary and local significance” (2016, p. 83). This is instructive, as the uncertain future, corresponding to unexpected change, and the temporariness and locality of economic knowledge are features setting apart the epistemic burdens of central planning and liberal transitions.

Regarding, first, the end that policy makers pursue, the epistemic burdens of central planning are heavier, given that central planning is undertaken for the people. This is because the transition to a liberal society, including its cultural preconditions, has one given, stable objective about which much theoretical and permanent knowledge exists. That liberalisers need to take into account societal conditions suggests a data problem that increases the complexity and creates some uncertainty, as these conditions change unexpectedly. But in general the difference is substantial, as central planning requires not only a theory of how to compromise between innumerable competing ends in a highly complex situation, but also faces a massive data problem: planners need to know what people want—for neither means nor ends are given (the contextual knowledge problem). And, perhaps decisively, as the sought-after equilibrium is ever-changing—since people’s wants and beliefs as well as the world change unexpectedly—the required knowledge is temporary, creating massive uncertainty and the need for central planners to constantly adjust their plans. This is far less of a problem for liberalisers whose objective is fairly stable and given.

The second dimension for comparison concerns the means that policy makers must use to get to their respective ends. In the case of central planning, this refers chiefly to the data problem of the discovery and employment of production goods to produce those consumer goods best able to satisfy consumer wants (incidentally, the challenge would be lessened, but not eliminated, if consumer wants did not matterFootnote 13). This again involves high complexity: in a world with so many things, policy makers first need to discover production goods and then go on to organise their employment in ways that are not only technologically but also economically feasible. Change drastically increases the challenge: both consumer wants and the world change unexpectedly—implying the relevant knowledge will often be temporary. For instance, new technologies create new opportunities to do things, or a catastrophe alters the available supply of some good. Moreover, planners would also need to tackle the theory problem of how to make a plan given the data.

Liberalisers must make sure the cultural preconditions that underlie a liberal society are in place. If they are not, liberalisers face an extremely complicated challenge in which they need to develop theoretical knowledge, as this appears not to exist, of how to stimulate the emergence of a culture amenable to the liberal order. While the culture they seek to realise is quite stable and well known, it seems that creating it is highly complicated—if it is even possible. If such a culture is already in place, the epistemic burdens are significantly lower. However, liberalisers still need to devise ways to implement institutional change, especially in light of collective action or capacity problems. This again involves high complexity, especially as this is a task in which local, temporary knowledge matters—think of how to manoeuvre a society in which strong interest groups exist that you need to keep satisfied, or think of the need to move quickly once a shock opens scope for action. Here, unexpected change is also relevant and increases the epistemic burden.

The biggest difference between the two epistemic burdens is the occurrence of unexpected change. Planning an economy means pursuing a coordination goal that changes constantly. What people want and what means are economical—all these factors unexpectedly change every minute. Liberal transitions also involve unexpected change. But this concerns primarily the question of how to get to a liberal society, rather than what the liberal society looks like—and, especially for the latter, mainly permanent theoretical knowledge is important. Consequently, the uncertainty-related epistemic burdens of liberal transitions should be lower than those of central planning. The complexity involved in liberal transitions should also be lower, even though things are less clear here. But that in central planning many more ends, not given in their entirety, need to be reconciled speaks to its higher complexity. Regarding liberal transitions, a lot hinges on whether the cultural preconditions are in place; if yes, the complexity-related epistemic burdens will be significantly lower.

With liberal transitions, a substantial part of the issue is a capacity problem in which policy makers need to have sufficient power to implement their plan in the face of collective action challenges. But this is also a challenge for central planners. That the focus is on the data problem that, presumably, already is insurmountable should not lead to overlooking this.

In conclusion, liberal transitions should be less epistemically burdensome than centrally planning the economy and should be actually doable. The chief factor accounting for this difference is the more stable and thus less uncertain nature of liberal transitions, combined with the fact that much more of the relevant knowledge is theoretical and permanent. Consequently, as Scheall indicates is possible, there is indeed a significant “difference in degree between these prospective goals” (2020, p. 100). This does not mean that liberalising a society is easy; but it is a challenge that can be met—although perhaps only when the cultural preconditions are in place.

As a final note, there is another difference—one that, however, favours central planners. This is that liberalisers, as liberalisers, should be much more wary of using power than central planners. There may be liberals who “would propose to achieve the liberal ideal by nonliberal means” (Buchanan, 2005, p. 3),Footnote 14 but even if liberalisers are willing to do this, they should be more hesitant in their use of coercion than central planners, who may, but need not, be (politically) liberal. In this sense, one may add, as a qualification to the above, that the reluctance which liberalisers should display restricts their available options and should thus increase the epistemic burden.

5 Conclusion

Epistemic burdens are ubiquitous, and Scott Scheall deserves recognition for bringing this to light. Specifically, Scheall has highlighted the epistemic burdens would-be liberalisers face. These burdens are substantial. They mainly consist of the difficulty of finding ways to get to a liberal society, rather than the difficulty of knowing it. The difficulties increase substantially when cultural preconditions are absent—creating a fitting culture seems very difficult, if not impossible. However, as illustrated by West Germany after World War II, the epistemic burdens of liberal transition can be surmounted. And these epistemic burdens, insofar as such a comparison can be conclusive, are lesser than those of centrally planning an economy. This latter task involves much more unexpected change and complexity in more dimensions—not only what people want is complex and ever-changing, but also what we must employ for economically satisfying consumer wants is complex, not given, and ever-changing. There are, therefore, substantial differences in degree between centrally planning an economy and liberalising society.

Policy makers cannot know that their attempt at transforming their society into a liberal one will succeed. There always is uncertainty, and this uncertainty will often be massive. There should thus be general awareness of the epistemic burdens of liberal transitions, but this should not obliterate the fact that success comes only to those who dare to try. This is also true for the challenge of maintaining a liberal society—an area for future research.