1 Introduction

I like to categorize theories of the West’s wealth according to the hypothesized time period of the proximate cause. Arguments like Diamond’s (1997) Guns, Germs, and Steel or the fractured land hypothesis of Fernández-Villaverde, Koyama, Lin, and Sng (2023) locate the roots of Europe’s rise in the late Precambrian. Another set of theories identifies something about modernity as the driver of the liberal economy: McCloskey (2010) and Mokyr (2016) offer distinct but complementary cultural explanations; North and Weingast (1989) famously find an institutional explanation in the rise of parliamentary government.

In between a billion BC and 1500 AD lie all the other explanations. Some of these are cultural, such as Holland (2019), or Schulz (2022) and Henrich (2020) on European exogamy, in each of which Christianity takes center stage as a force for reform. Others are institutional: accounts of the emergence of a distinct body of Western law which could alternately be traced to common law (Benson, 1990; La Porta et al., 2008) or the Church (Berman, 1983). What Salter and Young do in The Medieval Constitution of Liberty (2023) in one sense falls into this latter camp. They offer an account of economic liberalism that is firmly grounded in medieval, feudal institutions. They also, however, offer explanations of medieval institutions grounded in geography, and explanations of entrepreneurial culture grounded in medieval institutions. In short, they offer a theory that bridges existing work on the Great Enrichment while remaining a novel contribution in its own right.

What’s more, they largely succeed. I will have much to say by way of criticism, but my criticism should not overshadow my praise. The Medieval Constitution of Liberty is an excellent book. It works as an introduction to the Middle Ages, as an introduction to the economic way of thinking, as an interdisciplinary account of the origins of economic liberalism, and as a survey, synthesis, and pruning of the existing literature on institutions and development. It maintains a high level of accessibility without sacrificing much in the way of sophistication. Its few infelicities of syntax do not much detract from its readability. Its theoretical contributions are fraught with interesting implications for conversations far beyond the realm of medieval and early modern history. In short: buy it, read it, assign it, and make sure your university library has a copy. In what follows, I offer a critical précis of the argument of The Medieval Constitution of Liberty. Then I articulate two challenges. I conclude by highlighting what I think are some of the book’s best features.

2 The argument of the medieval constitution of Liberty

After an introductory chapter, Chaps. 2 to 4 articulate the relevant “Historical Backdrop” for the emergence of the medieval constitution. The authors provide a wealth of anecdotes and (what strikes me as a supremely competent) survey of the historical literature. Since they pass up the opportunity to distill the political developments of the early Middle Ages into a convenient list of those features which were essential to the ultimate constitution, I take the liberty of doing so now. The early Middle Ages saw the emergence of:

  1. 1.

    A landed and militarized nobility (2023, p. 33);

  2. 2.

    Vassals with land benefices whose loyalty to their lords had to be monitored (2023, p. 47);

  3. 3.

    The Church as a legitimating agent (2023, pp. 47–51);

  4. 4.

    The Church as a pan-European legislator through the Peace of God movement (2023, see Chap. 4).

That is, around, roughly, the turn of the millennium, Western Europe became governed by feudal lords whose power depended on their capacity to command loyalty from their landed vassal; vassals in turn held their land at the behest of their lords. Vassals tended to reserve their allegiance for lords whose authority was recognized as legitimate by the Church, and the Church offered neutral arbitration services and interpolations of God’s law that restricted – on pain of excommunication – warfare among the nobility. In short, Western Europe became simultaneously governed by multiple elites who held partially overlapping and partially divergent interests.

Salter and Young do not, in the main, offer theories or models as to why this state of affairs obtains. They take the medieval backdrop as a set of facts – exogenous parameters, if you will – which inform the development of the medieval constitution during the High Middle Ages. But their historical presentation is nonetheless compelling, and they frequently lead their readers to an understanding of the political backdrop through narratives that trace institutions to their origins. As I mentioned above, it is unfortunate that they do not always clarify which developments remain relevant to their project. For example, they describe in detail the role of missi under the Carolingians, but it is not clear that missi are important, or matter at all, for the subsequent medieval constitution. The history is interesting, but I think the section might have been omitted without great loss, and there is nothing to indicate its comparative unimportance. An attentive reader should have no trouble eliciting something like the list I have given, but the fact that the list of relevant developments has to be elicited at all is unfortunate.

With such political conditions in mind, Salter and Young turn to the task of describing how the medieval constitution becomes a constitution of liberty: “The primary thesis we advance in this book is that the structure of political property rights in medieval Western Europe was conducive to the evolution of good governance” (2023, p. 75). They begin with a series of definitions, which I reproduce in full because the technical vocabulary is useful, and incurring the fixed costs of definition here will profitably lower the marginal costs of exposition:

  • Political property right: “First, [a political property right] details who is entitled to wield political power. Second, it specifies the costs and benefits that arise from exercising that power. Third, it determines if those costs and benefits are borne by those who primarily make political decisions, or if those costs are passed on to others (… whether the costs and benefits are primarily internalized or externalized)” (2023, p. 75).

  • Residual claimancy: “To the extent that a holder of political property rights internalizes the costs and benefits arising from their exercise, that holder has residual claimancy” (2023, p. 75).

  • Constitution: “The set of rules that constrain those in authority in terms of how they wield that authority” (2023, p. 76).

  • Sovereignty: An agent is sovereign if and only if he is able to “defend his own political property rights” (2023, p. 97). Note that “there is no relationship between sovereignty and territorial monopoly” (2023, p. 101).

The constitution emerges from a series of bargains. Salter and Young explain what they mean by illustrating constitutional bargaining with a stylized example, familiar to any who have read North and Weingast (1989). A king confronts a credible commitment problem: he can pledge not to tax very much at time t, so that his vassals produce a high volume of output. At time t + 1, however, the king has no incentive not to tax all of the wealth of his vassals. The vassals backwards-induct and produce a low volume of output; they distrust the king’s stated tax plan. The king can thus generate more wealth for himself by ceding some power to his vassals – for instance, by giving them a representative council that has veto authority over new taxes. Then, his vassals know that the king’s tax plan is credible, and they will consequently produce a higher volume of output. By ceding power to his nobles, the king becomes richer. The creation of the council is a constitutional bargain that represents a Pareto improvement for all political elites.

2.1 Ideal types of governance

Salter and Young explicitly characterize the medieval world as composed of “shareholders states,” (2023, pp. 89–90) a third ideal-type government distinct from both autocracy and democracy. They follow Olson (1993) in their analysis of autocracy: the autocrat who has an encompassing interest limits his predation, but his revenue maximizing-tax rate is higher than the social optimum. They challenge, however, Salter and Young analysis of democracy: “Olson assumes the interests of voters are adequately aligned with social wealth maximization, but beyond a very small proportion of the polity’s population, this is almost certainly not true” (2023, p. 88). Democracy and autocracy both suffer from different versions of the same problem. Political property rights in both are such that there is imperfect residual claimancy; the costs and benefits of political decisions are largely borne by people other than the decision-maker. In the former case, the revenue maximizing tax rate for the autocrat diverges from the social optimum. In the latter, the benefit/cost of a productivity increase/decrease is only very fractionally internalized by each decision-maker (voter), rather like a joint stock company with hundreds of millions outstanding shares, where each shareholder holds exactly one share. A shareholders state, conversely, eliminates political externalities by making the structure of government resemble that of a corporation. Political decision-makers largely internalize changes in productivity (as in an autocracy), but power is fragmented among distinct shareholders (as in a democracy).

Bargaining among “shareholders” yields an emergent pattern of sovereignty. The nobility and the Church bargain their way to a self-enforcing distribution of sovereignty. Salter and Young characterize these agents as players in iterated Prisoner’s Dilemmas, Ultimatum, and Hawk-Dove games.Footnote 1 Each player is clearly sovereign on some margins and lacks clear sovereignty on others. Where sovereignty is weak, it becomes contested by other players. The conflict results in a more precise clarification of political property rights, where the weaker party accepts the sucker payoff and resigns himself to a more limited sphere of sovereignty and the stronger party extends his sphere to include the previously contested space. Importantly, sovereignty is not only or even primarily geographic in nature. As one example of the process of contestation, Salter and Young discuss the investiture controversy, where the battle was over who – kings or the Church – had the right to make (invest) new bishops. Ultimately, “in constitutional equilibrium, some political property rights holders will be sovereigns. Those who are not will come to hold only an attenuated set of political properties” (Salter & Young, 2023, p. 109).

2.2 Polycentric sovereignty

The process of bargains led to the emergence of polycentric sovereignty; that is, sovereigns held distinct but partially overlapping spheres of sovereignty. The beneficial result is obvious to economists: Tiebout competition (competition to make one’s jurisdiction attractive to potential members who can sort themselves into a set of jurisdiction) and yardstick competition (competition to emulate the success-making features of highly successful jurisdictions). Medieval governance was also hierarchical in ways that parallel contemporary federalist states. For example, kings governed many noble vassals hierarchically, while the Church, provided overarching rules governing the interaction of kings across all of Western Europe. The sovereignty of political property rights holders ensured the stability of the hierarchy; it was extremely difficult for king, church, lord, or vassal to overstep his bounds with respect to any other king, church, lord, or vassal.

Broadly speaking, the medieval world was ruled by three “estates”: the Church (first estate), militarized, landed nobles (second), and autonomous merchant cities (third). The Church was sovereign over the ecclesiastical affairs of all of Europe; it behaved, then, in many respects, like the federal government of a federalist republic. It was prevented from overstepping its bounds by the fact that its weapons were largely reducible to excommunication and the suspension of religious services in specified regions. In other words, the Church governed all of Europe through the selective granting and withholding of legitimacy from nobles. The Church also maintained a monopoly on important kinds of human capital (e.g. literacy) that it could leverage to great effect.

Each noble, in turn, was sovereign in his own fiefdom and demesnes. Nobles have residual claimancy, in the sense that they receive revenue streams from the land that they govern. Thus, lords run their lands like shareholders run firms. Each lord is made better off by increases in output. The constitutional bargaining process – the creation of credible commitment mechanisms that constrain the discretionary power of elites – is progressively emulated at every rung of the lord-vassal ladder. Cities do not arrive at the bargaining table until later, and I will hold off, for now, on describing their function until below.

In short, Church and noble check each other, and the fragmented military authority of nobles ensures that no one becomes sovereign over all. Instead, separate spheres of sovereignty emerge, and within each sphere of sovereignty, agents’ political property rights are such that they internalize the benefits and costs of their decisions. Each agent, then, when he successfully bargains for powers that improve his own revenues, improves the living standards of those he governs (the governance externalities are minimal). The result is a gradual emergence of limitations on the predatory capacities of every agent.

“Why Western Europe?” ask Salter and Young, “Why not elsewhere?” (2023, p. 137). The answer that they give comes down to geographic fractionalization. Western Europe’s geography makes governance of large regions uniquely costly. Thus, it makes sense that Europe would evolve this system of polycentric sovereignty – the costs to monopolization and thus monocentricity are just too high.

2.3 Liberal fruits of the medieval constitution

In the most compelling section of the book, Salter and Young go on to illustrate two institutional fruits of the medieval constitution: representative assemblies and self-governing cities. There is apparently an open debate among historians about the function of medieval assemblies. They may serve something like the function of Parliament in North and Weingast’s (1989) famous story: the king can raise more revenue by tying his own hands and submitting to the will of an independent legislative body. They may, conversely, function as coordinating mechanisms for making the nobility aware of the king’s agenda. Or, instead of coordinating, there might simply be occasions for kings to strongarm nobles into obedience. Salter and Young paint the history of the assembly as a gradual movement from serving the latter to the former function. Salter and Young present some evidence that pre-Carolingian assemblies tended to create consensus by pretending that more consensus existed than really did; this pretense of agreement can create genuine agreement when each party who might otherwise object to the findings of the assembly expects a scant number of allies. Thus, the initial assembly was a cross between a coordinating device and a vehicle for the imposition of royal will. But as the medieval constitution solidified, assemblies began to cut both ways. The emergence of well-defined spheres of sovereignty meant that nobles had real power at the bargaining table, and thus, by the High Middle Ages, kings would submit themselves to assemblies in order to secure their own ends.

Self-governing cities emerged from a similar bargaining process. Wealthy merchants leveraged their wealth to secure increasing levels of political autonomy. For example, merchants occasionally married serfs, and since the legal status of children was traced matrilineally, “burghers… were inconvenienced by the non-free status” (Salter & Young, 2023, p. 201) of other city dwellers. The fact that a freeman might have serf children was an obvious injustice demanding reform. Thus, the legal status of city-dwellers became principally territorial: one had rights as a member of the city qua his or her membership; rights were no longer defined by relationship to a lord. And since the cities were governed by and for merchant interests, they tended to evolve relatively liberal market institutions with sophisticated legal apparati for resolving property disputes. Cities thus became the first bastions of individual rights.

Having established that free cities developed the first truly modern market-friendly legal institutions, Salter and Young advance their most provocative claim: McCloskey (2010) gets the timing wrong on her Bourgeois Dignity thesis. Instead, Salter and Young (2023, pp. 202–206) argue that self-governing medieval cities cultivated bourgeois virtues and lent dignity to merchants in the High Middle Ages. McCloskey locates the source of bourgeois dignity in modernity because that is where the paper trail takes her: large volumes of extant literature which see merchant activity as something honorable fail to appear until the late 15th century at the earliest. Salter and Young report that merchants did not produce large volumes of literature at all, but nonetheless were accorded dignity in city life. Thus, bourgeois dignity is a product of the Middle Ages, not the Renaissance or Enlightenment, and McCloskey mistakes the timing because literature is a bad method for tracking culture in a pre-literate world, where most writing was produced by a decidedly anti-commercial clergy.

Salter and Young conclude with a good-natured takedown of the state capacity literature. State capacity is only good insofar as state capacity is employed for protective and productive ends. High-capacity states are engines of terror if not appropriately constrained. The interesting question for economic development is thus not, how did European states become so effective at raising revenue, but rather, how did European states become so effective at raising revenue and then primarily use that revenue to fund public goods? The answer, posit Salter and Young, lies in the medieval constitution of liberty. Early Modern European states were importantly shackled on account of the rules developed under centuries of polycentric sovereignty: “modern states and political liberalism are complements under the right circumstances. Those circumstances included the High Middle Ages constitution” (2023, p. 246).

In short, then, Salter and Young give us three ideal types of government: democracy; autocracy; and shareholder state. Of these three types, only shareholders states internalize the benefits and costs of political decisions to the decision makers. Thus, shareholders states should dynamically produce the best governance (defined as the governance with the fewest externalities suffered by the governed). The medieval system of polycentric sovereignty approximated the shareholder state ideal type, and thus ended up producing modern economic and political liberalism. They supplement their account narratively and connect their story to a vast range of empirical literature. Their theory makes sense of more than that which they set out to explain.

3 Criticisms

There are two serious problems with the argument. First, Salter and Young seem to take it for granted that political property rights are tradeable. But alienability is often the last right to be added to the bundle of property rights; for instance, in the Middle Ages, lords had use and exclusion rights to their land, but frequently lacked the right to sell their land. This fact has important implications for the theory of polycentric sovereignty that they advance. Their account is incomplete without a discussion of how certain rights became tradeable in the first place (and why others did not – a king cannot sell his kingship even if he can sell his land). Second, the role of the Church is left deeply ambiguous. It is clear from their account that they see the Church as a critical player in the bargaining process, but it remains unclear how that is the case. While many of their examples deal with religious controversy, it is also unclear how the results of religious controversy in the Middle Ages helped to make the constitution one of liberty. In both (excellent) examples of how polycentric sovereignty created explicitly liberal governance institutions – assemblies and free cities – the Church does not seem to play an integral role. In what follows, I articulate both these criticisms in more detail and sketch a picture of what I think a version of The Medieval Constitution of Liberty that accounted for them would look like.

3.1 Alienability and property rights

Salter and Young underspecify their analysis of residual claimancy. Alchian (1965) – in whose tradition Salter and Young claim to work – argues that the distinguishing feature of private property rights is alienability. Public property is property which one is entitled to use while others are excluded, but the user cannot alienate his right of use.Footnote 2 If the right is alienable, as Alchian tells us, “The two derivative implications are: (1) concentration of rewards and costs more directly on each person responsible for them, and (2) comparative advantage effects of specialized applications of (a) knowledge in control and (b) of risk bearing” (1965, pp. 823–824). (1) has important implications for Salter and Young’s treatment of residual claimancy. If you define, as they do, residual claimancy as the extent to which “a holder of political property rights internalizes the costs and benefits arising from their exercise” (2023 p. 75), and you agree, with Alchian, that making a right alienable importantly increases the level of internalization, then it follows that the holder of a political property right which is not alienable has only an attenuated degree of residual claimancy.Footnote 3

Apply Alchian’s point to Salter and Young’s tripartite division of ideal types. Salter and Young themselves characterize democracies as being relevantly like joint-stock companies where each citizen has a single share and there are a few million shares outstanding (2023, p. 90). What Alchian would add is that, importantly, each voter’s share is inalienable. If the shares were alienable, then the state would no longer be a democracy; it would become, quite literally, a shareholders state, where political decisions are made by people who have the highest willingness-to-pay for decision-making power and where, consequently, benefits and costs of decisions would be fully capitalized in the market price of the alienable share of government. Autocracy could, presumably, be characterized as the case where exactly one person or organization holds the single inalienable share of government.

To use Olson’s language, the stationary bandit has a suboptimal revenue maximizing tax rate because he fails to internalize all of the distortionary effects of taxation. If the bandit could, however, alienate his right of banditry, then the market price of the right would capture those externalities – at least as long as the victims of banditry can purchase the right from him (i.e. at least as long as transaction costs are negligible). What prevents this Coasean bargain from taking place is precisely the fact that the bandit typically cannot alienate his right of banditry. The credible commitment problems involved in selling a stationary-banditship are why, as Acemoglu (2003) explains, there is not a “political Coase theorem.”Footnote 4 It strikes me, then, that the relevant difference between autocracy, democracy, and the shareholders state is that in the former two, political property rights are generally inalienable, and thus holders lack (or have only very attenuated) residual claimancy, whereas in the latter, political property rights are alienable, and holders have residual claimancy.

The implications for political economy are enormous. The process of constitutional bargaining can take place only with bargainable rights. Insofar as rights are inalienable, they cannot be brought to the bargaining table. And since it is costly to make rights alienable (if Acemoglu (2003) and Nye (1997) are to be believed, sometimes infinitely so), then some rights will be left off the bargaining table. What is more, insofar as political property rights remain inalienable, there are political externalities.Footnote 5

I want to suggest, then, that to answer the question of whether medieval states could be thought to approximate shareholders states requires us to answer the question of whether the holders of political property rights could alienate those rights. Leeson (2011) examines the evolution of the medieval English institution of trial by battle: land disputes were resolved by the disputants’ representatives bludgeoning one another (sometimes to the death). He argues that the institution was an efficient mechanism for land allocation given the legal difficulty of land alienation. Substitution (replacing a given landholder on the feudal chain with person) and subinfeudation (creating a new link in the feudal chain) imposed obvious costs on heirs, and less obvious (but no less real) costs on lords. Thus, potential alienators were frequently required to secure the consent of their lord(s) and heirs. This system resolved the externalities of land alienation by dramatically raising transaction costs, crippling land markets. Similarly, as Salter and Young themselves note, “Since bishops were not allowed to alienate Church property, once land came into church hands it stayed there” (2023, p. 129). Land was not the only inalienable right of the medieval lord. The canon law’s prohibitions on simony raised the costs for the sale of ecclesiastical offices (even if enforcement was very imperfect). Slaves were saleable human capital; serfs were not. The reader can generate his own examples at his leisure.

Do not misunderstand me: none of the above suggests that important political property rights were de facto inalienable. Indeed, Leeson’s entire argument is that trial by battle approximated an all-pay auction, and thus that the institution functionally facilitated land markets where such markets were prohibited by norms. But the point is that land markets emerged where they were not present before. Land became more alienable as a result of trial by battle. I am thus not arguing that Salter and Young’s characterization of the Middle Ages as polycentric sovereignty under shareholders states is wrong. In fact, I think it is basically right (for whatever my opinion is worth). Rather, I am arguing that they have not fully shown that it was so – because they have not shown how some important political property rights could have been alienated – and that they have not shown how such rights came to be alienable. But the alienability of political property rights is essential to their claims about the residual claimancy of sovereigns and the constitutional bargaining process. My objection, in other words, is not that the model is wrong, but that it is underspecified, and takes alienability partially exogenous when it demands explanation.

This is really a missed opportunity for Salter and Young. The shareholders state is conducive to long-run liberty in a way that autocracy is not. But to the extent that claims to capital (and labor) are inalienable, individual lords are more like autocrats than shareholders. This observation gives Salter and Young the chance to really articulate the function of polycentricity in their model. Where landholders are imperfect residual claimants on account of the fact that they cannot sell their land, they may nonetheless approximate shareholders on account of polycentricity, manifested in the form of federalism and Tiebout competition, in a broader context of North-and-Weingast style constitutional bargains. This is what I take Salter and Young to ultimately argue, but by not clearly specifying the margins on which the medieval constitution is characterized by alienation-based residual claimancy and where, in the absence of means of alienation, polycentricity nonetheless constrains medieval political property rights-holders, they leave too much work to their readers.

The difference between alienability and the emulation of its effects might seem trivial, but it is not. It is the difference between privatizing a commons and governing the commons. In the former case, political externalities are eliminated by making the political decision-maker the owner of the fiscal commons – by giving him an alienable residual claim. In the latter case, externalities are managed by limiting use of the fiscal commons to members and by creating clear boundaries between membership and nonmembership. The point by Nye (1997) and Acemoglu (2003) is that the former case is basically impossible: credible commitment problems prevent effective privatization of the fiscal commons. What Salter and Young show is that this is not strictly the case: the medieval world developed mechanisms for resolving those credible commitment problems. But there are other margins where it seems like polycentric governance of the fiscal commons does the work of approximating true ownership.

The authors’ failure to distinguish between ownership and its approximation prevents their book’s profound generality from becoming immediately obvious. If they had spelled out more clearly, more technically, the relationships between ownership, governance, polycentricity, externality, etc., then it would be apparent to everyone that The Medieval Constitution of Liberty is really about the constitution of liberty simpliciter, and that the modifier just indicates their very compelling case study. As I have suggested above, I think that the way they should have done so is by introducing the concept of alienability; the concept would have enabled them to draw distinctions which remain blurry throughout the text.

3.2 The ambiguous role of the Church

My other major worry is that the Church very obviously takes center stage in the development of the medieval constitution – it is a frequent player in examples throughout the book – but Salter and Young leave its role theoretically underspecified, and, more troublingly, the Church seems to do very little work in explaining either of the two triumphs of the medieval constitution. Representative assemblies and free cities were the result of constitutional bargains between merchants, nobles, and kings, and while bishops acted in the constitutional bargains, it seems that they primarily acted in their capacities as landholding elites and not as representatives of Rome.

For example, Salter and Young devote an entire chapter to the Peace of God movement, where they show how the medieval Church became a kind of pan-European arbitrator that was capable of leveraging its legitimating agency to establish basic international law (mainly laws of war). But Salter and Young fail to make use of this fact in their theory of polycentric sovereignty. Their theory devotes roughly four pages (2023, pp. 128–131) to the Church directly, and mainly sketches historical episodes to evidence the following: “The Church’s spiritual authority… was by no means absolute, but neither was it something to be dismissed lightly” (2023, p. 130). There is a great deal of space between absolute and negligible authority, over which constitutional bargaining takes place. I wanted much more precise answers to much more precise questions: how does the Church’s role as an international arbitrator affect bargains between nobles? And what about bargains with the Church directly – how do they affect the subsequent distribution of political property rights? Salter and Young mention the investiture controversy in connection to these questions, as an example of a constitutional bargain. But what is the upshot of investiture? Why does controlling the investiture of bishops matter, or, in other words, what is the marginal benefit to kings of having authority over the investiture of bishops, and what is the marginal benefit to the Pope? And for that matter, what is the marginal benefit to the Pope, or to agents of the Church, of having the capacity to arbitrate disputes, or of declaring days of peace where no war can take place? How does the Church leverage its legitimating power to secure benefits for itself?

All of these questions are raised and partially answered in the first few chapters, when Salter and Young bring the readers up to the starting line, so to speak. They recount interactions between Church and state, describe the emergence of a distinct sphere of sovereignty for the Church, and suggest that the Church’s international governance played an important part in medieval Europe’s quasi-federalist polycentricity, from the fall of Rome to, say, 1000. But in their main theoretical section, the Church never receives a systematic treatment.Footnote 6 As it is, their model seems to treat the Church merely as one player among many, though their historical work (correctly) suggests that the Church actually had a uniquely important role. As I mentioned above, the problem is compounded when they apply their theory to history. Given the medieval constitution, they argue, we can expect the emergence of good governance. They give us two (excellent) examples thereof, but in each case, the Church seems to play only an incidental role.

What makes this all so worrisome to me is that it seems like any story of the Middle Ages in which the Church is not the main character is probably going to be incomplete, at best, or mistaken. Salter and Young give a brief nod to Berman’s magisterial Law and Revolution (1983), the central argument of which is that Western law developed as a result of the Gregorian reforms. If Berman is right, then the medieval constitution of Salter and Young’s model owes much to radical transformations within the Church in the 11th century. But Berman – and indeed, the Gregorian reforms – barely receive mention. A more recent example of scholarship on the medieval roots of economic growth is Henrich’s The WEIRDest People in the World (2020); Henrich argues, along with Schulz (2022) that the medieval Church forced a policy of extreme exogamy on the people of Western Europe, with long-run cultural ramifications that made liberalism possible. I may have missed something, but I did not find any engagement at all with either Schulz or Henrich.

Again, this is not to suggest that Salter and Young’s characterization of the Middle Ages is wrong. It is underspecified. Do the Gregorian reforms matter for defining the spheres of sovereignty? Are the reforms made possible by the medieval constitution, or is the constitution changed by the reforms? If so, why? If not, why not? Is exogamy a competing or complementary explanation? Does the medieval constitution suggest something about the roots of Church-mandated exogamy? Or vice versa: does exogamy affect the structure of the medieval constitution?Footnote 7 What about other uniquely Christian features of Western medieval civilization: monogamy; monasticism; and the Church’s control over divorce? And what about theological doctrine? So much of the conflict of the Middle Ages is over theological controversy, including investiture, but also various heresies, papal supremacy, the filioque, and the requirements for conciliar validity. Salter and Young’s account is made incomplete by its failure to trace the origins and consequences of the Pope’s unique power: the Papacy and the Church play a uniquely important role at the constitutional bargaining table in the High Middle Ages, but Salter and Young only allude to this fact, and keep their most detailed discussions of the Church in the “Historical Backdrop.”

4 A wholehearted recommendation

After all that, I still recommend The Medieval Constitution of Liberty to every library, for four reasons. First, its breadth of history is impressive. Salter and Young make sweeping claims about European governance that they then evidence by furnishing the reader with interesting and often surprising anecdotes. For instance, to illustrate political decentralization in the High Middle Ages, they offer the following example (2023, p. 73): Louis VI raised an army to attack the Count of Lauvergne, who was a vassal of William X of Aquitaine. William, in turn, was Louis’ vassal, but William demanded that Louis back off. The Count of Lauvergne had been harassing the bishop of Clermont, and so needed handling, but William argued that, though his own actions were subject to Louis’ judgment, the Count of Lauvergne’s actions could not be directly constrained by Louis. The Count was in William’s jurisdiction only. Louis conceded to his vassal, and the Count was tried in court instead. Examples like the forgoing abound; Salter and Young give an excellent sense of the sovereignty enjoyed even by lesser nobles. There were rights that no king could abridge.

Second, the basic models of constitutional bargaining, of ideal-type governments, and of polycentric sovereignty are extremely useful heuristics for thinking about governance everywhere. For all my nitpicking about alienability and residual claimancy, the fact remains that framing governance quality as a function of the size of political externalities is brilliant. While that idea itself is not novel to Salter and Young, it is too rarely deployed in applied work. The Medieval Constitution of Liberty successfully argues that medieval governance institutions were characterized by a high level of internalization at the outset, which lowered transaction costs for subsequent bargains that internalized further political externalities. Moreover, they develop a concept of de facto sovereignty as a phenomenon emerging from controversies that can be characterized as iterated games. I expect scholars with interests far beyond the Middle Ages to profitably appropriate variants of this model in applied work. I look forward to reading – and using – the term sovereignty equilibrium elsewhere.

Third, Salter and Young are generally convincing. The examples of representative assemblies and free cities are compelling evidence for their thesis; medieval governance institutions must have birthed institutions of modern market liberalism in at least these two clear ways. Their identification of bourgeois dignity with the cities of the High Middle Ages is an interesting idea that merits further research. The point is not one that they can heavily substantiateFootnote 8; nonetheless, the idea seems plausible and, if true, certainly strengthens their thesis. Their analysis of the state capacity literature is exactly right: “What constrains and profitably directs state capacity cannot be state capacity itself” (2023, p. 226). High-capacity European states must have had additional institutional features that made governance good. Plausible candidates for these features include delimited, local sovereignty (observed most clearly in cities) and parliamentary constraints on executive power, both of which arise from the medieval constitutional bargaining process.

Fourth, the entire project of reframing the feudalism of the High Middle Ages as a constitutional equilibrium is novel and admirable. It is for this last reason that I mainly love the book. Salter and Young have inaugurated a research program. They set out to explain the medieval roots of liberalism; in the process they have developed and applied the tools of economic analysis to political history. The Medieval Constitution of Liberty should become the ur-text for students of the political economy of the Middle Ages, not because it is flawless, but precisely because its flaws show us what we should talk about next. It is also an admirable stab at a theory of medieval institutions, and it really does succeed on many margins. The development of the concept of polycentric sovereignty is an important contribution in its own right, its application to the Middle Ages apt, and the illustrations of liberal governance emerging precisely because polycentric sovereignty lent itself to liberal constitutional bargains are compelling.

Perhaps it is entirely appropriate, then, that Salter and Young leave it to future scholars to trace the interaction of alienability, canon law, exogamy, monasticism, etc., with polycentric sovereignty. The Kuhnian in me appreciates The Medieval Constitution of Liberty for (what I hope is) the invention of a paradigm – and the economist and amateur medievalist in me appreciates this book for a potentially appropriable stream of research rents. Salter and Young ask a novel question, develop novel tools for answering it, and if their answer does not satisfy its readers entirely, that is all to the good: we can pick up where they leave off.