About this Issue

Today’s westerners are likely to be highly tolerant of other religions, but this was not always the case. Indeed, repression was the norm in many European societies during the early modern era. So what changed? 

The standard story holds that after the Reformation, Europe simply fought itself to exhaustion. Whole nations were decimated in religious wars between Catholics and Protestants. Eventually, rulers and common people alike began to look for a way to stop the bleeding, and from this necessity, toleration emerged. 

This month’s lead essayists beg to differ. Mark Koyama and Noel D. Johnson argue that stronger states had less need of religious unity, because stronger states could rely on contracts and the rule of law to establish interpersonal trust and national solidarity. As a result, they didn’t need the type of solidarity that comes from religious unity. Liberalism was therefore a luxury they could afford. 

This is bound to be a controversial idea, though Koyama and Johnson stress its origins in classical liberal historiography. Joining them to discuss religious freedom, liberalism, and how we can further religious toleration in the modern world, we have invited Prof. James A. Robinson of Harvard University, Dalibor Rohac of the American Enterprise Institute, and Dr. Hans Eicholz of Liberty Fund. We also welcome readers’ comments through the end of the month.

Lead Essay

The Trouble in Getting to Denmark

How did modern liberal states emerge? Francis Fukuyama calls this the problem of “getting to Denmark.”Economists have traditionally focused on the question of how sustained economic growth got started. Political scientists, on the other hand, have focused on questions of democratization and political stability. Both of these are critical components of liberal states. But there is one factor that social scientists have neglected: religion.[1]

This neglect is easy to understand. In the developed world religious liberty is so ingrained in formal and informal institutions that it rarely emerges as a newsworthy issue. This was not always the case. Religion was so central to premodern societies that it is difficult to fully understand the transformations associated with modernity without attending to it. Religion was used to justify the categories in which government and society more broadly used to structure everyday life. Women versus men, nobles versus commoners, guild members versus non-guild members, Muslims versus Christians, Christians versus Jews. All of these categories—as well as the different statuses associated with them in law and in culture—relied to a varying degree on religion to legitimize their use.

If we are interested in the rise of liberal societies and liberal principles such as freedom of expression, then we have to understand how society came to adopt new political and economic institutions that were largely separated from religious doctrines. That is, we need to explain the rise of religious freedom.

The emergence of religious freedom in the West was not the product of ideas alone. The mid-20th century historian of religious freedom Roland H. Bainton observed that “the best things on religious liberty were said in the sixteenth century but not practiced until the nineteenth.”[2] What he meant was that during the 16th century there were many statements in favor of religious liberalism despite an intensification of religious persecution and violence. Writers including Sebastian Castellio (1515–1563)—who condemned the burning of Michael Severtus by Jean Calvin—and Anabaptists such as Felix Manz (d. 1527) and Balthasar Hubmaier (1480–1528), both of whom suffered death for their beliefs—articulated arguments in favor of religious pluralism. But they were largely ignored by their peers and had no influence on policy. It was only in the 19th century that some measure of religious freedom was attained in Europe.

Religious freedom should be sharply distinguished from the practice of toleration. Today what the word “tolerance” refers to is better called permissive individualism, that is,the idea that we have no right to judge or condemn individuals for their lifestyle choices so long as those choices don’t impose harms on others. But the original Latin meaning of tolere, which persisted all the way through the 17th century, is to bear something that one strongly disagrees with. It was a practical rather than a moral principle, and it was contingent and subject to revision. In our book Persecution and Toleration, we use the term “conditional toleration” to emphasize the difference between toleration and genuine religious freedom.

Why does this distinction matter? Too often historians praise societies that didn’t actively persecute religious minorities as “tolerant.” But the absence of persecution does not imply that individuals were free to worship as they liked or, indeed, to pursue economic or social betterment. The assumption that dissent could invite repression was sufficient to create a climate in which potentially subversive thoughts were not freely expressed. In our reading of the historical evidence, neither ancient Rome nor the Islamic or Mongol Empires had religious freedom. They often refrained from actively persecuting religious minorities, but they were also ruthless in suppressing dissent when it suited their political goals.[3] Religious freedom is a uniquely liberal achievement, and liberalism is an achievement of post-1700 modernity. What explains it?

Personal Rules versus Identity Rules

In Violence and Social Orders, Douglass North, Barry Weingast, and John Wallis distinguish between personal rules, identity rules, and general rules. Personal rules involve treating people differently on the basis of their individual characteristics. This works when social cooperation is limited to a small number of individuals, say a band of ten or twenty. Members can trust one another because the chances of future interactions are high. Consistent with game theoretic reasoning, this very often leads to sustained cooperation.

Personal rules have a serious shortcoming, however. They are of limited use in dealing with strangers. As the scale of society increases and the complexity of social interactions grows, personal rules cease to be sufficient. With the shift to settled agriculture after 8,000 BC, political organizations became larger and states oversaw the introduction of more sophisticated legal systems to prevent theft, fraud, and uncontrolled violence. For most of history, and in much of the developing world today, these laws have taken the form of identity rules.

Identity rules depend on the social identity of the parties involved. This could refer to an individual’s clan, caste, class, religious affiliation, or ethnicity. Examples from historical legal systems abound. Aristocrats faced different rules from commoners. Slaves faced different rules from freemen. The Code of Hammurabi, for example, prescribed punishment based on the relative status of the perpetrator and the victim. Identity rules were common historically because governing individuals on the basis of their legible social characteristics was cheap. As religious identity was particularly salient, many identity rules treated individuals differently on the basis of their religion.

Religion was an especially important component of identity in the large agrarian civilizations of Europe and the Near East in a time before nationalism and nation states. Shared religious beliefs and religious identities were seen as crucial to maintaining social order. Religious differences were extremely destabilizing because they were associated with a host of deep societal cleavages.

In an environment where a common religious identity undergirded not only the institutions of the church, but also those of the state and civil society, both religious freedom specifically, and liberalism more generally, were unthinkable.

For instance, in medieval and early modern Europe oaths sworn before God played an important role in upholding the social order. These were thought so important that atheists were seen as outside the political community, since as John Locke put it, “promises, covenants, and oaths, which are the bonds of human society, can have no hold upon an atheist.”[4]

A shared religious identity was also crucial for guild membership. Guilds in Christian Spain excluded Muslims. Guilds in 14th century Tallinn excluded Orthodox Christians. Jews were excluded almost everywhere. In parts of Europe converts from Judaism and even their descendants or remote relations could not be guild members.[5] In a world governed by identity rules, an individual’s religious identity determined what economic activities were open to them. The question is: How did we move from a world of identity rules to one of general rules?

General Rules and State Capacity

The eventual abandonment of identity rules in favor of general rules in Western Europe is intricately bound up with the development of state capacity. State capacity refers to the ability of a government to enforce rules and raise taxes. Identity rules are a cheap, but ultimately limited, way to do both these things. In contrast, it is expensive to implement general rules that are uniformly applied to individuals regardless of their personal beliefs or family background.

As an illustration, consider how early modern governments often used Jewish communities as a source of tax revenue. Usury restrictions made lending by Christians very costly. However, rulers could grant monopoly rights to Jews to lend without violating their religious principles. In turn, the rates of interest charged by Jewish lenders were high, and the profits were taxed away by the very rulers who granted these rights. Finally, the specialization of Jews as moneylenders exacerbated preexisting antisemitism among the Christian population. This in turn made it relatively easy for rulers to threaten Jews in case they didn’t intend to pay up.

So long as rulers relied on Jewish moneylending as a source of revenue, Jews were trapped in this vulnerable situation. Their position could improve only when states developed more sophisticated systems of taxation and credit.

As suggested by the above example, low state capacity and a reliance on identity rules are self-reinforcing. States that rely on identity rules face less incentive to invest in the fiscal and legal institutions that would increase state capacity. This, in turn, makes them more reliant on identity rules and less able to enforce general rules.

Furthermore, religion is a convenient source of legitimacy for weak states. This gives rise to a partnership between state and church. While there are many forms that this partnership can take, a common one involved the state enforcing religious conformity in return for support from the religious authority.

In Persecution and Toleration we provide a detailed historical narrative that discusses how events like the Black Death and the Reformation disrupted the medieval pact between the church and state. This narrative focuses on the political and institutional consequences of these shocks.

Following the Reformation, we document how religious violence was most extreme in states like England and France where the state was relatively strong. It was in these states that these conflicts first undermined the old church-state partnership. As this partnership weakened, new sources of legitimacy had to be sought. In contrast, Poland, where state capacity was low, saw little religious persecution despite the spread of Protestantism.

It took repeated failures to reimpose the old partnership between church and state before rulers gave up trying to eliminate religious diversity. In France, for instance, Louis XIV’s attempt to eradicate Protestantism by expulsion failed, and during the 18th century French Protestants gradually secured a degree of recognition, culminating in the Edict of Toleration of 1787. In England, while formal laws against Catholics and nonconformists remained on the books until the 19th century, the 18th century also saw a marked decline in religious tension and violence.

In our argument it was not that the Wars of Religion simply exhausted confessional and doctrinal disputes. Rather there was a transformation at the institutional level. The leading European states shifted away from identity rules towards more general rules. This shift was related to 19th-century historian Henry Sumner Maine’s discussion of the passage from status to contract: Status was imposed and ascriptive. Contracts, in contrast, are the outcome of voluntary choices. Status-based rules are invariably identity rules. Contracts provide the foundation for a system of general rules.

Moving from a fixed status to a contractual society helped set in motion a range of developments, including the growth of markets and a more extensive division of labor. But it had the unintended consequence of diminishing the political importance of religion, and this made liberalism feasible for the first time in history.

Throughout our book we stress that there is no necessary relationship between state capacity and liberalism. Powerful states have been and still are capable of repressing minorities, of course. The horrors of the 20th century were conducted by all-encompassing Leviathan states. But liberalism is fragile and highly vulnerable in the absence of a powerful state.

Readers of Cato Unbound might find this argument uncomfortable. After all, it places a lot of explanatory weight on the rise of state institutions and stands in contrast to much scholarship in the libertarian tradition. However, we see it as building directly on the insights of the most important classical liberals, including Adam Smith and F.A. Hayek, who emphasized the importance of strong but limited states in providing the institutional framework in which markets and liberalism can flourish.

Modern States and Liberalism

What does our argument imply about the fate of liberal societies today? While the far left has never accepted liberal values such as freedom of expression and freedom of religion, antipathy towards liberal values is now evident in mainstream progressive publications as well. Liberalism is indicted because it is perceived as legitimating inequality and failing to endorse social justice.

Similarly, writers from the right such as Patrick Deneen and Yoram Hazony critique liberalism for destabilizing the family and the foundations of social order.[6] They construct a strawman version of liberalism based almost entirely on John Locke’s account of the social contract. As these writers would have it, liberalism is based on the idea that society is justified because rational self-interested individuals came together to form a society for mutual self-betterment. This legitimates a society where individuals become atomized and dislocated, unmoored from tradition, religion, and family.

This essay is not the place to fully engage these misconceptions. But it is important to note that we think the core characteristics of a liberal society are the rule of law and reliance on general rules. It does not require belief in the realism (or even usefulness) of social contract theory. Nor does it require us to believe that human beings correspond to the model of Homo economicus.

Liberalism is valuable because it is the only form of social order we know of that is consistent with a high degree of autonomy and human dignity. By documenting the vital role played by the strong states in adopting liberal institutions, Persecution and Toleration sheds new light on how we think of the prospects of the liberal order in the future.

In parts of the developing world, where states are weak and religious and ethnic identities remain strong, the future does not look bright for religious freedom or liberalism more generally. Globalization and the spread of Enlightenment values were once thought likely to lead to liberal democracy becoming the default mode of government. This no longer seems so. There is simply little evidence that liberal ideas alone will sway governments to adopt general rules and the kinds of institutions required to upload a liberal social order.

By contrast, in the developed world where states are relatively strong and liberal ideas already have legitimacy, there remains room for optimism. Though liberal values appear in retreat today, the costs of reverting to a world of identity rules would be very costly. Understanding the origins of liberalism allows us to take some measure of those costs.


[1] Historians, on the other hand, have rediscovered the importance of religion in recent decades. Conflicts like the English Civil War (1642-1649) that were once studied through the lens of class analysis are now viewed as religious struggles.

[2] Roland Bainton, The Travils of Religious Liberty. Westminster. 1951, page 253.

[3] See for instance, Maria Rosa Menocal, The Ornament of the World: How Muslims, Jews and Christians Created a Culture of Tolerance in Medieval Spain. Little Brown, 2002 for the claim that Islamic Spain was pluralist and tolerant. Similar claims are made for the Mongols by George Lane, Daily Life in the Mongol Empire, Westport, Conn.: Greenwood, 2006. A recent book that promotes the notion that the Romans were comparatively tolerant despite their persecution of Christian is Catherine Nixey, The Darkening Age: The Christian Destruction of the Classical World. MacMillan, 2017.

[4] John Locke, Letter on Toleration, 1689; 52.

[5] Sheilagh Ogilvie. The European Guilds: An Economic Analysis. Princeton University Press. pages 103-104. We discuss this in Johnson and Koyama, pages 56-57.

[6] See Patrick J. Deenen. Why Liberalism Failed. Yale University Press and Yoram Hazony, The Virtue of Nationalism. Basic Books. 2018.

Response Essays

The Perks and Perils of Having a State-Run Church

Johnson and Koyama make an ambitious and original claim about the origins of religious toleration. In brief, certain big shocks to western European society made it difficult for states to base their legitimacy and social control on a pact with the church. In response they had to develop capacity and new instruments which freed them to tolerate religious freedom.


There is a lot to like about the argument. For one, it moves us beyond Charles Tilly’s politics-free notion that “war made states and states made war” and instead asks us to focus on the internal politics of control and the development of capacity. This move is long overdue, and it is exciting.

I have three big issues with the argument, however. First, imagine that it is correct that something sunders the pact between state and church and forces states to rule in different ways. Why does this mean they wish to allow, or promote, religious toleration? What is the mechanism exactly? Is it that once the state exits from its ruling pact with the church, it wants to foment religious competition to undermine the church and make sure that it is, if not a useful ally, at least submissive? I can see how the argument might allow for religious toleration to emerge, but I didn’t see why it happens. It seems possibly necessary, but not sufficient. At the moment the thesis seems to be more a new argument about state formation than one about religious toleration.

My second main issue is whether this argument really makes sense empirically, even in the European case. Take Britain. Henry VIII broke with the Catholic Church in the 1530s, expropriated the monasteries, and created the Church of England, with him as the head. That doesn’t sound like a weakening of the state-church pact, it sounds like a nationalization of it. In fact, one would have thought that a Protestant church with the British king at the head of it would have been even more useful as a tool of legitimation and social control than a distant, and not so cooperative, Catholic Church. This moment did indeed coincide with a period of state formation, what Geoffrey Elton called the “Tudor Revolution in Government,” where something like state offices separate from Henry’s personal retainers begin to emerge. But was this due to the separation of state and church, or the latter’s indigenization? Roll on to the next century. If you wanted to pick an even more pivotal moment in British state formation it would be the Glorious Revolution. This led to the emergence of a bureaucratized fiscal system, including the creation of the Bank of England to manage the public debt and other measures. At the heart of 1688 was the overthrow of James II. How was that justified? The Bill of Rights put it like this:

it hath been found by experience that it is inconsistent with the safety and welfare of this protestant kingdom to be governed by a popish prince

No Catholics on the throne! Moreover, one clause of the Bill stated, “That subjects which are Protestants may have arms for their defence suitable for their conditions and as allowed by law.” No arms for Catholics. The various Test Acts meant that no public servant in Britain, including members of Parliament, could be a Catholic. They were finally repealed in 1828. So the heyday of British state formation was accompanied by systematic and institutionalized discrimination against Catholics. Rather than vanish as a tool for legitimation, the British state created its own more pliable religion to accompany its rise.

Third, let’s think outside the European case and ask about the relationship between religion, religious toleration, and state formation elsewhere. There are cases which certainly resonate with the thesis of the book. Think of the Saudi state, the first version of which formed in the 1740s with the famous pact between Muhammad ibn Saud and Muhammad ibn Abd al-Wahhab. It was a fusion of a nascent state with an expansionary jihadist religious ideology. The Saudi state remains very weak as a consequence of this pact; for example, the necessary commitment to Sharia and religious judges creates a huge amount of discretion and uncertainty in the legal system, with huge adverse economic consequences. If for some reason this pact was broken, the Sauds might indeed have to build a very different sort of state.

But perhaps the part of the world where powerful states have been most in evident, at least for their economic consequences in the past 70 years, is in East Asia. A plausible hypothesis about why the East Asian Miracle was possible was the legacy of bureaucracy which ultimately seems to come from Imperial China. But what has that got to do with religion? Not much I think. The usual story is that it is linked to Confucianism. Confucius said in the Analects, “promote those who are worthy and talented.” Confucianism also promoted a very different model of state-society relations which wasn’t primarily based on some religious sanctioning—even if the ruler had the “mandate of heaven.” Of more practical importance seems to be the notion that rulers maintain legitimacy by “sowing wealth amongst people,” or as President Xi recently emphasized, quoting Confucius,

He who rules by virtue is like the North Star. It maintains its place, and the multitude of stars pay homage.

This was not a model based on participation. Confucius also observed, “When the Way prevails in the world, commoners do not debate matters of government.” So meritocracy, no accountability, all legitimized by benevolent rulers. Where is the church? Is there religious toleration in China? Certainly not, because this would promote a different set of loyalties other than those espoused by the ruling ideology. But the lack of religious toleration does not stem from a pact between the rulers, whether imperial or communist, and some organized church, for there is none.

What do I conclude from this? It’s a brilliant argument, and it might be right, though I have doubts as discussed above, but is it, if not Euro-centric, since Medina isn’t in Europe, at least maybe restricted to the domain which generated the great monotheistic religions? There is something about the way the relationship between Catholicism, Islam, and the state played out that seems to generate a particular set of interactions between religion and state. But it is not clear to me that this is true elsewhere in the world. Yet if that is so, it is difficult to disentangle this argument from a whole host of other things that makes this part of the world distinct.

Realizing Religious Liberty: “Stateness” or Competition?

Initially Johnson and Koyama appear to be presenting a succinct, even compact, theory of historical development. The opening question leads one to believe that their theory is about the emergence of liberal states. To my mind, however, that framework does not quite fit.

In point of fact, they seem also to want to make a more specific claim: That the early modern state is the very source of the principle of religious liberty itself.But this claim too, is a bit confusing. Do they mean as an idea, or as a principle of law? Much of my confusion stems from the kind of explanation they have chosen to give.

What our authors have presented is a repackaging of what in German was called eine Stufenlehre,or a stadial theory, with all of the attendant problems these theories entail for historical causation. Under the mistaken impression that they have identified certain specific causal relations through the identification of the distinctions of specific kinds of rules, they have actually provided only descriptions of large aggregate events. And each step in this process of change has unfortunately encouraged insufficient attention to the complexity of the relationship of thought to action. That has, in turn, churned up all the classic chicken and egg problems that such ideas involve, whether we are speaking of the systems of G.W.F. Hegel, Friedrich List, Karl Marx, or the early ideas of Douglass North.

But make no mistake. I like Johnson and Koyama’s descriptions. Their typology of rules is vital for making important qualitative distinctions in the interpretation of human actions. I very much agree with their aim to caution all liberals of either classical or libertarian inclination not to treat the importance of a powerful state so cavalierly as to think they can make do without one. “Liberalism,” they want us to recognize, “is fragile and highly vulnerable in the absence of powerful states.” I accept this up to a point.

But from their typology of rules, we do not obtain an explanation for change, and from the need for government, we do not get a prescription for policy. Now, it may well be that in my rather quick overview of their book length treatment, I have missed where they offered a theory of historical change, or a means for evaluating public policy choices. They certainly have not done so in this essay, and I remain dubious that it can be done in any definitive sort of way while keeping to their framework.

For starters, what exactly does powerful mean? Many, I suppose, might be tempted to assert that it entails a claim of arbitrary will; that is to say, the authority to do whatever the sovereign wants. Such power over the lives of others, however, would not be the sort of state any liberal or libertarian would endorse.Might they mean what the American Founders meant when they spoke of “energy in government”? By this, they would stress efficiency in the exercise of specifically enumerated and limited powers. It is then a question of the kind of government, and not so much about its being strong or weak. If some liberals confuse these terms, they should not.

The real problem, however, is to be found in the treatment of status and contract as legal indications of trust. Henry Sumner Maine’s typology of status versus contract is a very handy way of classifying the differences among ideas about what is and is not acceptable behavior. What explains this shift in attitudes?

Koyama and Johnson’s answer amounts to a restatement of the question: “In our argument it was not that the Wars of Religion simply exhausted confessional and doctrinal disputes. Rather there was a transformation at the institutional level. The leading European states shifted away from identity rules towards more general rules.”

Ok. But what prompted this shift? All throughout their more extensive treatment, it is in fact the processes of competition, among princes, prelates, cities and, yes, eventually, what we now call states and, ultimately, nation states. Is it more accurate to say then that it is “The State” that gives us the full articulation of the principle of religious liberty in question, or is it the process of competition that drove states to do so? I would make a case for the process in this instance.

History is, of course, always a dynamic relation between thoughts and actions. Institutions are themselves the products of the articulation of various ends and means, what Ludwig Lachmann (1971) aptly called the “nodal points” of coordination. To understand either thoughts or the forms of action, you need to slice into the hermeneutical circle at some point in order to interpret the relationships involved. This is the nature of all feedback loops. Our authors recognize this in part, observing that “The emergence of religious freedom in the West was not the product of ideas alone.” I am not aware of any serious historian who would assert such an extreme claim, but they then go on to favor just one part of the circle: the primacy of the institution of the modern nation state: “It was only in the 19th century,” they assert, “that some measure of religious freedom was attained in Europe.”

Some measure? Do they not mean, some degree? And what could this be?

Tolere we are told is not really freedom. It is putting up with something with which you disagree. Freedom, on the other hand, is the positive affirmation of your right to practice what you wish; or in other words, “the idea that we have no right to judge or condemn individuals for their lifestyle choices.” Really?

If condemning means use of the state, then yes I agree, but if it is the right to judge, then no, I do not. We still tolerate those with whom we disagree, we just do not attack them. Something of the positive has crept into their notion of rights. Real religious freedom always has, and always will, retain the right to judge for oneself and one’s religious community.

Consequently, if it is true that the earlier forms of governance, from ancient Rome forward, sometimes observed more, and sometimes observed less tolerance, then it is not clear why the modern state must be privileged in this regard. And if the idea of religious freedom can be shown, as they have noted, to precede the early modern state (“there were many statements in favor of religious liberalism despite an intensification of religious persecution and violence”), then the origins of the principle must in fact be rooted in even earlier processes. And this is borne out by their own evidence.

That some states (e.g. Spain, and England under the Stuarts) did not observe the principle, and others like France under the premiership of Louis Henri, Duke of Bourbon, even reversed themselves, indicates that it was not the characteristic of “stateness” per se, but the processes of constitutional thinking that were more especially involved—processes of both thought and action.

Douglass North’s early works, which appear to predominate in our authors’ longer treatment, do indeed, at times, seem to favor institutional structures over mere thought. His later work, however, did not (See especially Understanding the Process of Economic Change 2005). Above all, North emphasized the important dynamic of competitive processes in the shaping of ideas in the West. This move was not taken out of thin air. He rather expanded on some older ideas.

Among those taking up the competitive argument very early on were Hume (1742) and Kant (1784). James Madison also recognized the importance of a multiplicity of competing faiths among the colonies in early America as the basis for American recognition of political constitutional freedom. Likewise Max Weber, in his more developed reflections published posthumously as the General Economic History (1920), observed the critical importance of the competitive element in the economic and political relations of Europe. And it was this that Randall Collins expanded upon in his Weberian Sociological Theory (36, 1986). It was also a central theme in the histories of E. L. Jones (1981), Harold Berman (1983, 2003) and Hendrik Spruyt. The latter’s book, The Sovereign State and Its Competitors (1994), is particularly relevant here, because Spruyt drew specific attention to the importance of the league cities.

Koyama and Johnson themselves have noted, in their longer treatment, the interesting case of the city of Hamburg, one of the great Hansa cities. It was especially “favorable to Jewish settlement,” they noted. Being open, as the merchant senators of the city were, “to international trade” (300, 2018), Hamburg welcomed moral strangers. Granted, the city-states were ultimately replaced by the nation-state, but they were responding to the same competitive process to which the later nation-states were also subject.

To conclude then, I would contend that it was the processes involved in the context of decentralized authorities, and the ensuing competition among them, that paved the way for the widening of toleration and eventually the formation of constitutional republican forms of modern governance.

The point is not to disagree with the importance of well-constituted political regimes. It is however misleading to say that it was the nation-state, full stop, which gave us religious liberty. It was rather the processes of competition that forced open the various monopolies of power.

Federalism Is a Condition for Better States

The libertarian distrust of government rests on solid empirical foundations. According to work by the late R.J. Rummel, after all, around 262 million unarmed people died at the hands of governments over the course of the 20th century – six times as many as those who died in combat over the same period. Although Mark Koyama and Noel Johnson’s essay and the body of research on which it draws do not show that such distrust is completely unwarranted, they do qualify it in important ways.

Mark and Noel’s findings establish links between state capacity and institutions on the one hand, and conflict and prosperity on the other. Many libertarians will find those counterintuitive and perhaps even a little disturbing. Before dismissing them as flukes or historically contingent artifacts, it is worth noting that they are entirely consistent with the thrust of the literature on the drivers of long-term prosperity, including popular expositions in Daron Acemoglu’s and James Robinson’s Why Nations Fail and in Tim Besley’s and Torsten Persson’s Pillars of Prosperity. More importantly, taking these findings seriously opens an avenue to a more nuanced, and hopefully more viable, version of the classical liberal project.

If the list of state-sponsored horrors is long, violence can be inflicted by a variety of nonstate actors as well. What Mark and Noel’s research demonstrates is that the decline of such decentralized forms of violence in Medieval and Early Modern Europe—against Jews and witches—went hand in hand with the rise of state capacity and formal mechanisms of law enforcement. Accordingly:

[L]ow state capacity and a reliance on identity rules are self-reinforcing. States that rely on identity rules face less incentive to invest in the fiscal and legal institutions that would increase state capacity. This, in turn, makes them more reliant on identity rules and less able to enforce general rules.

These findings naturally connect with a larger political economy on violence, which finds that civil wars are strongly associated with poorer societies. Moreover, a strong correlation exists between fiscal capacity of government and income. The link is also seen in historic data from Europe, conceivably driven by the necessity to wage war. As an aside, it has also been hypothesized that the economic divergence between China and Japan can be explained partly by the gap in state capacity that opened after 1750. In the context of developing economies of Sub-Saharan Africa, better tax collection appears associated with improved quality of public administration.

This does not necessarily show that state capacity is itself the cause of high incomes and domestic peace. Rather all three—state capacity, growth, and lack of civil strife—seem driven by partly overlapping sets of institutional drivers. What that means is that in practice it might not be easy or desirable to seek to disentangle “good,” or “inclusive,” institutions and state capacity. As development practitioners learned the hard way in places ranging from Somalia to Afghanistan, where there is no state capacity, “good” formal institutions will not sustain themselves, giving way to rules and organizational forms built around tribal identities.

If the machinery of the state can be put to good uses as well as to bad ones, perhaps the orthodox libertarian view of government, especially “big” government, as a nemesis needs rethinking. Historically, the flourishing of dynamic private economies has clearly relied on the existence of state-financed public goods. Furthermore, as Will Wilkinson and others have argued, building a constituency for a dynamic economy involved a lot of creative destruction and is much easier to accomplish with a robust social safety for those left behind. That might also provide an explanation for the significant improvements in economic freedom in recent decades seen in countries with sizeable welfare states, such as Canada or Denmark.

The observation that states are not the only source of coercion and violence, but only one of its many sources, has also deeper implications for the classical liberal project. Coercion, violence, and repression can be inflicted by ethnic and religious communities, families, and the variety of intermediating institutions of civil society, which are normally seen as benign and even necessary hallmarks of a free, self-governing society. As a result, freedom might well require a constant balancing and renegotiation between these different sources of authority. Singling out just one of these organizational forms as inherently suspect is misguided. In reality, state power can perform an important balancing act against locally grown sources of coercion—think of French secularism responding to the aggressive anti-republican Catholicism of the late 19th century, or of the civil rights movement in the United States using levers of the central government to respond to injustice inflicted by U.S. states.

In other words, the success of the classical liberal project requires a plurality of sources of authority keeping each other in check: federalism. In its original, broad meaning (in Latin, foedus means a covenant) federalism means more than just a division of government into different levels tasked with the provision of different types of public goods. Rather, “principles of federalism permit people to function through self-governing institutions among local, regional, and national communities in the interest of organizing collective endeavors,” as Vincent Ostrom put it. Its appeal lies in the simultaneous presence of different forms of governance, all of them available to the individual simultaneously.

Some of the turbulences observed around the world today seem linked to the fact that this process of balancing and renegotiation has been disrupted by technology and globalization. As Martin Gurri argues, information has become radically democratized. Globalization has reduced the ability of governments to get things done—yet much of the political class appears to be eager to continue with business as usual. The result has been a decline of trust in political and nonpolitical authorities and a reversal towards tribalism, as Mark and Noel hint at towards the end of their opening essay.

The federalist reasoning extends to the world of international affairs. Many on the political right—both conservatives and libertarians—harbor a deep distrust of international institutions and international organizations. For some, the distrust springs from the veneration of national sovereignty as a supposedly self-evident good. Others, especially those steeped in crude realist theory of international relations, see such institutions as an irrelevant veil to power competition by states.

Yet both criticisms miss the mark. As the experience of the post–World War II world suggests, institutions can restrain national governments—and that is, furthermore, a good thing. International institutions can provide a backstop against the rogue behavior of states, including in extreme scenarios of state collapse or their degeneration into vehicles of mass murder. The global GATT/WTO framework, with its principle of non-discrimination, is a dramatic improvement over the reality of warring trade blocs that characterized the first half of the 20th century. Legitimate criticisms can be directed at the European Union, but the process of economic integration has imposed useful constraints on the ability of European governments to engage in protectionism and economic nationalism. I worry that many of my British Euroskeptic friends will understand the full meaning of this only after Jeremy Corbyn becomes Prime Minister of the post-Brexit United Kingdom.

Do any of these mechanisms, at home or internationally, work perfectly? Of course not—but perfection is not attainable in human affairs. Governance of complex, open, and pluralist societies is hard work, with no panaceas in sight. There is no guarantee that free societies will successfully weather the current crisis of trust in institutions and authorities. Mark and Noel’s work is important because it shows the way in which the classical liberal project can help societies navigate the current storm: by embracing the diversity of institutional and organizational forms, and by encouraging contestation and competition between them to ensure that they serve as effective checks on each other. The sooner classical liberals do that, instead of insisting on their fixation on government as an intrinsic enemy of freedom, the better our odds of getting through the current situation unscathed.

The Conversation

The Fitful March of Religious Liberty

James Robinson raises three concerns with our thesis. First, he argues that we provide necessary, but not sufficient, conditions for religious freedom to emerge. Second, he notes that the two major shocks in English history—Henry VIII’s declaration of independence from Rome and the Glorious Revolution—were not, in fact, characterized by greater secularism or religious freedom. Third, Jim wonders about cases outside of Europe, particularly China, where religion was less salient than in Europe but which, nonetheless, never developed in a liberal direction.

James’s first point concerns mechanisms. He is correct to note that the mechanisms we identify provide necessary, rather than sufficient, conditions for the emergence of religious freedom. What we argue is that absent a set of institutions based around general norms, it is difficult to see how religious freedom could have emerged. There was no religious freedom, or liberalism more generally, in the feudal states or despotisms that frequent premodern history. An institutional transformation was required. But this institutional transformation did not make religious freedom inevitable. Individual rulers like Louis XIV could change the course of policy. This fragility is an important part of the story; after all, the 20th century saw much of Europe embrace totalitarianism.

What we point to are systematic factors that can explain why religious unfreedom was once ubiquitous and why the West evolved towards liberalism. Early modern France provides an interesting example. Following the Wars of Religion, an uneasy religious peace emerged under Henri IV and the Edict of Nantes. This peace exemplifies the kind of conditional toleration that we discuss at length in our book. There was no commitment to religious freedom. French statesmen like Cardinal Richelieu and Colbert sought to reduce the independent political power of the Protestant nobility while preserving religious toleration for Protestants. They also encouraged Jews to settle in France. These developments are consistent with the theoretical mechanism highlighted in our book.

This gradual move towards greater religious liberty was interrupted under Louis XIV. Led on by ministers who exaggerated the number of converts to Catholicism, and convinced that it was possible to reunify France religiously, Louis XIV inaugurated a policy of persecution that would cumulate with the revocation of the Edict of Nantes. Autocratic rulers can always reverse policy. What is less appreciated is that the revocation ultimately failed. As documented by historians such as David Garrioch in The Huguenots of Paris and the Coming of Religious Freedom, 1685–1789, persecution proved counterproductive. It was impossible to convert all Protestants; some of them fled, while many who converted under threat of coercion soon reverted to their old religion. By the 1720s, many anti-Protestant laws were no longer being enforced. Garrioch quotes a Lieutenant-General commenting on a family known to practice Protestantism:

[They were ] all new converts who accomplish badly their Catholic duty. They have, however, a very good reputation among the merchants and I am assured that they conduct their business as wood merchants with great probity.

French Protestants finally achieved religious toleration by 1787, under the auspices of the ancien régime. The French case points to the mechanisms that we believe pushed European states toward liberalism. But it also shows that these forces certainly did not make liberalism inevitable.

Robinson’s second point is also valuable. Whereas formerly historians saw the Reformation as marking a decisive point on the path to the secular state (and the privatization of religion), in recent decades scholarship has emphasized the religious character of the Reformation. So it is important to note that that while the Reformation did shatter the medieval equilibrium between church and state, it did not break the relationship between religious and secular authority. Indeed, Henry VIII and his successors sought to re-order this relationship, not to abolish it.

What matters is that, whatever they attempted, none of Henry’s successors succeeded in establishing stable relations between church and state. Acknowledging this failure is required to understand the crisis that the English state underwent in the 17th century. A Church of England that was broad enough to encompass both Catholic sympathizers and radical Protestants was bound to satisfy neither. It struggled to provide legitimate authority in the same way that Catholicism had. Rather than a source of stability, it generated instability. Civil War was sparked by Archbishop William Laud’s failed attempt to impose a common prayer book on Scotland. Political disagreements were impossible to separate from religious beliefs. Even after Charles I was deposed, there seemed no way to bridge the divisions that had arisen among Anglicans, Presbyterians, and Independents.

The Glorious Revolution was the solution to this crisis. And, as Robinson correctly notes, it also did not bring about religious freedom. Indeed, the toleration that James II was prepared to offer in 1687 went beyond that guaranteed by the Bill of Toleration of 1689! As the title of our book suggests, the road to religious freedom was a long one, and it was only in the 19th century that the British state can be viewed as liberal with regard to religion (and even then imperfectly).

The decline in the political role of religion was gradual. Consider the institution known as the “royal touch.” For centuries it was believed that monarchs could cure the disease known as scrofula by laying their hands on the afflicted. This practice did not disappear with the Reformation. In England it came to an end with the Glorious Revolution and the accession of William and Mary (Marc Bloch, 1973).[1] But in France it continued through the 18th century, although belief in its efficacy declined.[2]

Turning to Jim’s third point. Our opening essay indeed focused on Europe. And while we believe our argument has traction for other continents, we have not proposed a universal account that fully explains the path of state development across the entire world.

That said, we discuss China and Japan at length in Persecution & Toleration. Briefly, the argument is that the unique history of early state formation in East Asia meant that the partnership between religious and secular authorities was always tilted in favor of secular authority. Because of this, Chinese identity rules were not tied to religious beliefs as they were in the West and, as the Chinese state built capacity, the tensions between religion and governance did not emerge in the same way. The result was a powerful illiberal state which relied comparatively less on religion for political legitimacy. China has never had religious freedom and from the Tang period onward the Chinese state closely regulated all religion. This is the path that China is on today.

Having a powerful state per se is not certainly sufficient for religious freedom or liberalism. We are not providing a blueprint for how to become liberal (e.g. “build a strong state”). Rather our interest lies in uncovering the factors responsible for the emergence of liberalism in Western Europe. There it was the process of state formation and increased religious diversity after 1500 that made it incentive compatible for European states to establish liberal institutions.


[1] It was briefly revived by Queen Anne.

[2] The last instance was in fact in 1825, during the reign of Charles X, when it was greeted skeptically.

Competition Among States Wasn’t Sufficient for Religious Liberty

We are grateful for the insightful comments of Hans Eicholz. They provide us with an opportunity to clarify some misunderstanding and to further elucidate our argument.

Eicholz is dissatisfied with our framework, or at least our description of it in our opening essay. Eicholz’s first comment is that our theory is a stadial theory like that of the Scottish Enlightenment thinkers, or G.W.F. Hegel or Karl Marx. It is true that in Persecution and Toleration we aspire to write social scientific history. This means we are explicit in proposing a theoretical framework. But we don’t see our theory as stadial—at least not as strictly defined. Certainly we don’t find it useful to think in terms of rigid historical stages with their own or laws or dynamics. What is true is that, as economic historians, we are conscious of the vastly different scale of economic activity and living standards that differentiate modern societies from pre-modern societies. And we see important differences in the way past societies were ordered. In particular, we think societies that are governed by identity rules function quite differently from those where more general rules obtain.

In Persecution and Toleration we acknowledge the importance of historical contingency in shaping the paths taken by different countries. For example, not all regions of Europe, when confronted with religious heterogeneity, chose to relax the bounds of toleration. Notably, Spain expelled its Jewish population in 1492—choosing to eliminate a source of tension and maintain its relatively decentralized and low-capacity institutions. This represented what Acemoglu and Robinson term a “critical juncture” for Spain. It could have continued towards building state capacity and abandoning identity rules. This is not how history played out, however, for reasons as difficult to predict as silver coming in from the Americas, and the personality of Spain’s rulers.

Eicholz suggests that what he calls stadial theories, whether of Hegel, Friedrich List, Marx, or the early Douglass North, suffer from the same problem. But he doesn’t elaborate on what this problem is. In any case, rather than the early North, we are explicit in building on the ideas outlined in the last book North co-wrote (with Barry Weingast and John Wallis), Violence and Social Orders. But this is by-the-by. Our work is structured around a theoretical framework because the alternative to an explicit framework is an implicit theory and often an incoherent one, as Karl Popper observed.

Eicholz also finds our definition of state power problematic. The concept we rely on is that of state capacity. How does state capacity differ from simple power? The Ottoman Empire serves as an example. From Machiavelli onward, European commentators noted the absolute power of the sultan. Served by slaves, the Ottoman sultan could expropriate whomever he wished.[1] But did the Ottoman Empire have high state capacity? While it was a powerful state at its height, by the late 17th century it had less fiscal capacity than any of its major European rivals. While the sultan could seize the property of any of his subjects, he collected far less taxation on a per capita basis than the King of France, let alone the States-General of the Dutch Republic.[2] Furthermore, as we elucidate in the first chapter of the book, taxes collected per capita are just an imperfect, albeit useful, proxy for true state capacity. For example, in the Ottoman Empire, tax collection continued to be farmed out to private individuals well into the modern period—resulting in poor oversight of fiscal institutions as well as frequent violations of fiscal contracts by the sultan.[3] Investment in state capacity and reliance on general rules are thus often strategic complements.[4] Perhaps in our essay we should have been clearer (as we are in the book) that the types of modern states we envision as promoting religious freedom were constrained and limited.

Eicholz also suggests that we clarify our use of the term freedom. Here there seems to be a simple misunderstanding. We use freedom in a political sense and the sense that is widely accepted within liberal theory as freedom from restraint. Religious freedom simply requires the state not to coerce people on the basis of their religious beliefs and not to favor or penalize one religion over another. And it goes without saying that religious freedom is a continuous rather than a binary measure.

Finally, Eicholz emphasizes the role of inter-state competition. He sees polycentric competition among divided states as critical to the emergence of religious freedom. Here there is some agreement between Eicholz and ourselves. Joel Mokyr’s recent book A Culture of Growth restates the view that repression in early modern Europe was rendered ineffective because thinkers and their books could cross borders.[5] One of us has written a paper on the beneficial consequences of political fragmentation in Europe.[6]

Political fragmentation and Inter-state competition superficially seem like an attractive alternative explanation. A deeper look however reveals that interstate competition is not an alternative explanation to investments in state capacity; they are part of the same story.

Political fragmentation alone doesn’t get you very far. Scholars who wish to argue that Europe’s political fragmentation was crucial to its dynamism and growth have to explain why fragmentation among Indian polities after the decline of the Mughal Empire had no such effect. Not all fragmented political systems have the same properties. What distinguishes the European state system after 1500 is not simply that it was fragmented but that it was characterized by competitive investment in state capacity and state-building.

And one reason why political fragmentation prompted religious liberalism was that it was one way competitive states sought to gain an advantage over their rivals—this was why Prussia embraced Huguenot immigrants and one reason why Oliver Cromwell and Cardinal Richelieu were favorable to Jewish settlement. We don’t believe it coincidental that these were the same states making the greatest investment in state-building. Interstate competition is what prompted European states to raise larger armies, increase taxes, and run up debt. In other words, we agree with Eicholz that competition matters, but it is part of the larger story we wish to tell and not an alternative to it.

Why does this distinction matter? Well, Eicholz concludes that it was “the processes of competition that forced open the various monopolies of power”. We don’t disagree with the sentiment, and we favor both competition and breaking down monopolies. However, while the process of interstate competition and state building that occurred in early modern Europe did prevent the emergence of a single hegemonic state in the continent, it was characterized by the establishment of monopolies of violence within each European state.


[1] At the Battle of Lepanto (1571), the Ottoman commander Ali Pasha is said to have taken with him his entire treasure because he feared expropriation.

[2] In per capita terms by 1700, taxes in the Dutch Republic were 26 times higher (in silver prices) than in the Ottoman empire. Data from K. Kivanc Karaman and Sevket Pamuk, (2010) Ottoman State Finances in European Perspective, 1500–1914, The Journal of Economic History 70(3), pp 593-620.

[3] Eliana Balla and Noel D. Johnson, (2009) Fiscal Crisis and Institutional Change in the Ottoman Empire and France, The Journal of Economic History 69(3), pp 809-845.

[4] In the sense of J. Bulow, J. Geanakoplos, and P. Klemperer (1985), ‘Multimarket oligopoly: strategic substitutes and strategic complements’. Journal of Political Economy 93, pp. 488-511.

[5] Reviewed here by one of the authors http://www.independent.org/publications/tir/article.asp?id=1267.

[6] “Unified China; Divided Europe” Chiu Yu Ko, Mark Koyama, and Tuan-Hwee Sng International Economic Review, February 2018, Volume 59, Issue 1, pp 285-327.

State Capacity and General Rules

I thank Mark Koyama and Noel D. Johnson for their clarifications, and I think that we are not so far apart.

A couple of issues though could still be developed further: What is a state’s “capacity,” and what is its relationship to general rules?

The reason I took their theory of history to be stadial was due to the degree of emphasis they placed on the transitions from personal to identity to general rules. This emphasis seemed to be doing the heavy lifting in the definition of a strong versus a weak state, and also to be rendering the principal framework of their historical process.

If “capacity” is defined as a power to effect some desired change, then it is not clear why a state that is serving rules of identity, if it is in fact effectively preserving such, is not displaying “capacity.” It must rather be that they mean that the state is displaying a capacity of a particular sort.

Is capacity simply in the marshaling of resources? But every state calls upon resources, and if those resources are sufficient to the ends sought, then again, it begs the question of what it means to speak of capacity outside of specific purposes. If we accept that competition of various sorts, among various types of polities and institutions, drove the development of modern states to embrace a general rule approach to certain aspects of their domestic policy, it won’t help simply to define it as such. Rather, we ought to probe more deeply into why they did that.

And our authors have in fact done this, noting that fiscal policy was driven by the needs of war, or as they stated in their response to me, to “gain advantage over their rivals.” Thus Prussia worries about what France is doing, and it becomes more effective policy for Prussia to stop engaging in domestic persecution so as to gain access to the resources (i.e. men and money) to repel rival states. So capacity is capacity for war, and the competition for such is what drove “statesmen” to seek ways of unlocking the coffers of all their inhabitants. Where did they get the idea then for a general truce among the adherents of different faiths within their territories?

Freeing commercial relations, improving currency, regularizing rules of finance and banking by states were also products of long and arduous processes for very much the same purposes that entailed the application of general rules. The general rules that are particular to modern commerce, however, grew up initially in the interstices of all these various powers. Merchants were able to form quite extensive networks simply by observing certain rules of property and contract, while obtaining sufficient compliance through a hodgepodge of custom, reputation, and piecemeal enforcement by sometimes this lord over here, and sometimes that lord over there. And the various potentates noted the fortunes being so accumulated from the late Middle Ages onward. Again, the Hanseatic league cities are key to understanding the processes involved at their earliest and most basic level.

I would even venture to say, that it was here, among the cities, that we find the very source of the idea that treating all men equally with regard to certain objects was first realized, the old notion that one is set free by breathing the air of the city. In this instance, the advantages of regular and equal enforcement of the rules of property provided an analogy for so treating individuals with regard to other aspects of their persons, including their religious beliefs. And this certainly can be seen in the case of the League of the Rhine when the cities of “Mainz, Cologne, Worms, Speyer, Strasbourg, Basel and other cities” sought to assure the peace by uniting for defense and observing that “This promise shall be kept in such a way that not only the greater ones among us shall have the advantage of this common protection, but all, the small with the great, the secular clergy, monks of every order, laymen, and Jews shall enjoy this protection and live in the tranquility of holy peace.” It was a beginning, and it was 1254.

Why didn’t all decentralized areas of the world move along these same lines? Different ideas, different degrees of urgency, and even serendipity all played their parts. And it is not that competition was not having an effect elsewhere. I see that our authors cite an essay by Roland Vaubel, but not the collection he edited with Peter Bernholz, Political Competition, Innovation and Growth in the History of Asian Civilizations (2004). Sometimes competition did play an important role in other parts of the world, but then other factors would intervene. It is almost never just linear causality in history.

So, I would simply encourage them to highlight further the issue of competition, and in conjunction with Dalibor Rohac, would emphasize that federalism as well as separation of powers all express varieties of rivalrous institutional forms within polities that help explain why certain types of rules are maintained despite the ease with which identity groups can form.

Rules of Community, Economy, and Society

As I think further on Koyama and Johnson’s interesting use of the typology of rules, it strikes me as particularly relevant to the question of how certain conceptual and institutional forms relate to each other.

Specifically, it is critically important to understand that different rules will predominate in different forms of human association. This is an old point that goes back to both Ferdinand Tönnies and Max Weber. In this instance, however, Koyama and Johnson’s narrative affords a very useful way to see the process by which individuals move from the application of rules of status and identity, in one sphere of their lives, to rules of more general application in another sphere. It is always going to be a process of learning.

Initial rote learning by imitation is the the foundation of our first conceptual orientation to ourselves and those around us. This is the role that family and community play, and it is why concepts of who we are are so powerful. They fill up the categories of meaning that give us our most basic notions of what to pursue and what to avoid, what is good and bad, what is right and wrong, defining substantive ends to be sought. This is Gemeinschaft, or community.

General rules are almost always means to an end, and have use for all individuals, regardless of specific substantive notions of the good. They are in fact useful for the pursuit of substantively different ends, so long as they are not invariably in conflict. Rules of economic life tend in this direction. This is Wirtschaft, or economy.

And finally there is a realm where both types of rules meet each other, and are coordinated through specific institutional forms as varied as customary practices and organized associations. This is Gesellschaft, or society.

Whether we speak of the theoretical understandings of Douglass North, Joel Mokyr, or Deirdre McCloskey, we are speaking of competition as processes of learning. And this is ultimately why I think the specific conditions of pluralism of various political and civil forms in the Middle Ages must have given rise to the formation of an earlier orientation to a general rule of personal liberty in faith than our authors’ explicit answer would admit.

It was not for the sake of freedom that modern states enforced general rules. It was because they recognized the usefulness of such rules in the amassing of wealth, as shown previously in cites.[1] Their capacity for war was not first, but came after states secured the allegiance of their inhabitants. The idea of such a rule, however, must have drawn initially from the economic realm in which modern civil society had had its first inklings. The processes of learning, the processes of competition, precede the state, and most certainly precede the modern nation state. Exhaustion in the religious wars was just one way competition of a very brutal sort promoted learning, but the tools in their toolkit must have already been there, at least in the form of useful analogies from other civil associational forms.

Again, we must look to the cities.


[1] The references to the League of the Rhine in my previous comments come from a collection of tracts that were among the favorites of a scholar who was both a mentor and a friend to so many of us, the late Professor Leonard Liggio. The documents dealing with medieval city life can be found in the last section of this volume.

Religious and Other Freedoms

Johnson and Koyama make a bold argument about the connection between state formation and religious freedom. In my previous response I argued that I didn’t understand why states that had freed themselves with pacts from the Church had any interest in actually promoting religious freedom. But more importantly, I wondered how this argument applied elsewhere in the world. Their response was that China, for example, developed a different historical relationship between state and society, with a secular state dominant and no religious or other type of liberty.

OK, but let me now question the association between religious liberty and liberty more generally. Is it true that there is a necessary connection between religious liberty and liberty more broadly? Europe developed both; China developed neither. What about the “off diagonals”? Johnson and Koyama’s account might seem to imply that these are empty, but I don’t think that is right. Could you have a situation where there is religious liberty, but not liberty? I think the answer is yes, and the obvious case for me is pre-colonial Africa.

There was little general liberty in historical Africa. Robert Rattray, the first head of the first Anthropological Department of Asante, and part of the British colony of the Gold Coast, now Ghana, transcribed an Asante proverb in 1929,

When a chicken separates itself from the rest, a hawk will get it.

As Rattray put it,

[A] condition of voluntary servitude was, in a very literal sense, the heritage of every Ashanti; it formed indeed the essential basis of his social system. In West Africa it was the masterless man and woman who ran the imminent danger of having what we should term ‘their freedom’ turned into involuntary bondage of a much more drastic nature.

By involuntary bondage of a “much more drastic nature,” Rattray of course meant slavery. The Hobbesian nature of West African society in the era of the slave trade and its aftermath meant that to be safe, you had to surrender your freedom and become a client, tied to someone more powerful. Another proverb he wrote down said “If you have not a master, a beast will catch you.”

People were not free, but they were free to practice whatever religion they wanted. In fact Africans seem to have had a very pragmatic attitude toward religion. There were no Bibles and texts defining what was orthodoxy or heterodoxy and religious ideas and practices seem to have mostly been viewed on the basis of whether they worked or not. One of the myths of the LoDagaa people on Northern Ghana, transcribed by the anthropologist Jack Goody, belonged to the Bagre initiation society. It recognizes that the practices of the society were not “God’s way” but then continues

In God’s country What is there to surpass this?

If it worked, it was worshipped. Notions of one god (for example Onyame among the Akan Rattray studied) intersected with ancestral cults, secret societies, and rain shrines. The Kuba kingdom of the Congo rainforest had three different creator spirits, Mboom, who they likely brought with them on their migration, Ngaan, adopted from the local Kete peopleb and Ncyeem apoong, probably brought by King Shyaam, in the 17th century when he founded the state. After that, the Kete were not forced to worship Mboom.

This the African experience shows that there is no necessary relationship between religious toleration or liberty and liberty more broadly construed.

Were European Cities Responsible for Liberalism?

We appreciate Hans Eicholz’s further comments on our essay. His conclusion is that we should look to European cities as the source of liberal tendencies in medieval and early modern Europe.

This is a venerable argument that goes back to Max Weber and Henri Pirenne, and indeed to Adam Smith. Cities were incubators of civil republicanism, market liberalism, and greater religious freedom.

There is much to this argument. Certainly, we agree that the Italian city-states in particular provided governance models that later states imitated. They developed republican political institutions and built fiscal capacity. Sovereign debt markets first developed in late medieval Italian cities were only adopted by larger states in the 17th and 18th centuries. Cities also offered freedom to serfs.[1] But it is important to point out that recent research cautions against viewing Europe’s independent cities as forebears of modern liberalism.

Stephan R. Epstein’s important work Freedom and Growth pointed out that city-states such as Florence were far from liberal in how they treated their rural hinterlands.[2] Contrary to their reputation, princely states were often less rapacious or extractive.

In the late Middle Ages, Florence imposed protective tariffs and restrictions on the products of its local rivals, Pistoia and Prato. Epstein documents how powerful cities like Florence had a detrimental impact on rural proto-industrialization. In Epstein’s view,

Up to the 1450s Florentine policy towards the contado combined paternalistic and authoritarian norms established in the early fourteenth century, including rural export bans and forcible supplies at fixed prices to the city, the forced distribution of grain imports and excess stocks among the peasantry, and loans to stimulate agricultural recovery (Epstein, 2000, 161).

Recent work by Sheilagh Ogilvie also casts shade on the pro-market reputation of medieval city-states. Medieval cities were dominated by craft and merchant guilds that established barriers to entry and regulated trade. Ogilvie notes that medieval cities were less open than is often supposed:

Few towns admitted outsiders freely to the full economic rights of ‘citizens’ status.’ They placed conditions on citizenship and, as a consequence, excluded people who could not fulfill the requirements. Towns often charged citizenship admission fees to raise revenue, something they were able to do because citizenship meant access to a pool of profitable privileges which only urban citizens enjoyed, among them the right to apply for guild membership (Ogilvie, 2019, 97).

This research cautions against a reading of European history which posits “liberal” cities against “autocratic” states. Medieval city states did cultivate Republican notions of civic virtue and autonomy and were important centers of long-distance trade and manufacturing. But sustained economic growth required access to larger markets than any single city-state could provide. The economic rise of Western Europe certainly cannot be understood without recognizing the legacy of the independent cities of the Middle Ages, but it also can’t be fully appreciated without recognizing the important contribution of nation-states and larger federations, such as 19th-century Zollverein.


[1] It is less clear whether medieval cities offered meaningful bastions against religious persecution.

[2] S.R. Epstein, (2000). Freedom and Growth. Routledge: London.

Is Federalism the Solution?

We thank Dalibor Rohac for his insightful commentary. He is correct in noting that mass killings can result both from the actions of large and powerful states like Maoist China or Stalin’s Soviet Union, and in the absence of a state.

Dalibor’s main substantive point concerns the importance of federalism. The promise of federalism—as outlined in Barry Weingast’s important work on market-preserving federalism—is that a federal state can obtain the best features of both centralization and decentralization. A federal state benefits from economies of scale and a large internal market while benefiting from decentralization at a local level.

We do not disagree with Dalibor on the prospective benefits of federalism. One message of Persecution and Toleration is that the historical process through which state formation takes place matters. In Britain and France the state building process resulted in highly centralized nation states. In Germany this process did not take place. When Germany was finally unified in the 19th century it was as a federal state. As a large part of our argument is focused on the rise of modern states in Western Europe, we did not dwell on federalism at any great length.

Moreover, the extent to which federalism will safeguard personal, social, and economic liberties depends on the character of the federal state. The Holy Roman Empire was a federal structure, but it did not safeguard individual liberties. In particular, our reading of the evidence suggests that it is critical that the federal state is market-preserving and that it enforces general rules.