A wide definition of democracy sees it as a political system “in which those who rule have been obliged to seek consent from those they govern” through some sort of council or assembly. Rulers without strong state bureaucracies need assemblies because they face an insuperable information problem regarding what is produced and what can be taxed. They also face a tax collection problem. Moreover, pure coercive autocracy will not work if the subjects have an exit option to move out of reach of the rulers.

Early democracy / What Stasavage calls “early democracy” was characterized by councils or assemblies, a weak central state or sometimes no central state at all, and generally less than universal suffrage. Athens was “the most extensive example of early democracy” in the 5th and 4th centuries BCE. Other variants of early democracy were common in other, often primitive, societies. A sample of 186 societies studied by anthropologists — the Standard Cross-Cultural Sample — shows governing councils existing at the local level in more than half of the sample and at the central level (over many localities) in about a third. “By one estimate,” writes Stasavage, “two thirds of councils had broad-based community participation” — as opposed to elites only.

For example, early democracy existed in Mari, a kingdom that existed on what is now the border between Iraq and Syria during the first millennium BCE. The king had to negotiate with town councils over taxes.

The Huron, a native tribe confederacy in the American and Canadian Northeast, provide a more striking and recent example. At the time French explorers encountered them in the 16th and 17th centuries, a Huron village was ruled by a chief and a council of clan chiefs. Anybody in the village apparently could go and express his opinions to the council. Villages were grouped into tribes, each of which was governed by a tribal chief and a council made of all clan chiefs. Above them was the confederacy council, but its decisions did not bind individual tribes, thereby promoting consensual decision-making. Iroquois tribes had similar institutions.

Early democracy was also practiced by pre-Islamic Arabs. Sayyids (rulers) generally ruled as first among equals. The arrival of Islam in the 6th and 7th centuries CE did not immediately change that. The Koran’s governance principle of shura recommends consultation and consensus, which sometimes may have also applied to the choice of the caliph. The conquest of Iraq (and the Sasanian Empire), however, marked the beginning of the end of democracy in this part of the world as the Islamic caliphs inherited a well-functioning state that they could soon rule without assemblies. The same thing happened later to the Mongols, who abandoned their democratic habits after invading the Chinese Empire.

Early modern city-states in Europe as well as the American colonies practiced some forms of early democracy. Colonial governors, including those from London merchant companies (the Virginia Company, for example), needed to consult the colonists if only because labor was scarce and a plethora of available land offered an exit option. The rulers did not have bureaucracies to employ in their exercise of control. According to Stasavage, the promise of democratic rights attracted immigrants to America, probably (I suppose) as a symbol of individual dignity or a signal of mild governance. Governing assemblies with broad male suffrage developed with frequent elections and sometimes explicit instructions or mandates given to elected officials.

Early democratic societies were generally small-scale and, as suggested above, existed in an environment where residents had an exit option and there was no bureaucracy to help rulers tax production. A ruler therefore needed to negotiate to gain the consent of the people or the notables. In other words, democracy developed when political rulers were weak.

Conversely, early democracy failed when a ruler could easily tax production because agriculture was intensive and “legible” (that is, easy to monitor) or because the ruler employed a bureaucracy to help him. As illustrated above, a conqueror who inherited a state with an established bureaucratic hierarchy could avoid democracy simply by co-opting the bureaucrats. Hence, the decline of (early) democracy.

Early democracy is often historically associated with “the absence of many technological developments that we commonly associate with civilization” such as writing, geometry, and accounting. The advance of civilization “often acted to undermine early democracy” as these technologies were used by the state bureaucracy and “reduced the information advantage that members of society had over rulers.” Moreover, higher population densities made people “more easily monitored by bureaucrats.”

There was no state bureaucracy in Europe after the fall of the western Roman Empire in the 5th century BCE, but early democracy continued at the local level. (See “Let’s Travel That Road Again,” Spring 2020.) Kings eventually convened central assemblies composed of what Bertrand de Jouvenel in On Power (1945) called “social powers”: nobles, bishops, and later representatives of cities. The assembly or council participants had power bases independent of the state and thus “substantial blocking power.” The king needed them to govern and tax.

Modern democracy / Modern democracy gradually replaced early democracy, hence the rise of democracy after its fall. Modern democracy was a European invention. It consists of a new kind of assembly whose members are elected by extended suffrage and are not tied by mandates or instructions from those they represent. In other words, modern democracy is representative democracy with more widespread suffrage.

It developed fastest in Britain. The 11th century Norman conquerors inherited the country’s different kingdoms, which had royal councils and assemblies of local notables. At the end of the 13th century, however, the king was able to abolish the mandates in the central assembly, which eliminated much blocking power. The Tudors, who reigned from 1495 to 1603, even made progress in creating a bureaucracy.

Elsewhere in Europe, absolute monarchs (think of Louis XIV) were better able to reduce the role of councils and assemblies, but still needed them to govern. Prussia under the Hohenzollerns was an exception. Frederick William of Prussia created a permanent army that allowed him to rule without negotiating with assemblies. Except for Prussia, Stasavage tells us, European states were still “relatively weak.”

Economic historian Douglas North and political scientist Barry Weingast argue that the Glorious Revolution of 1688 created a limited state in England. But Parliament became sovereign and its executive (“the Ministry”) gained something akin to autocratic power. William Blackstone, the famous 18th century jurist, observed that Parliament has “absolute despotic power” and can “do every thing that is not naturally impossible.” Royal despotism had been replaced by parliamentary despotism. For Stasavage, though, this “state capacity” was necessary for democracy to develop, a challenging idea to which we will return.

The first fully modern democracy was born in the United States in 1787, albeit with slaves and women excluded from the suffrage. France rapidly followed, with no slavery. The formula then spread over the world and the franchise was gradually extended. Universal suffrage often came with universal conscription. Stasavage emphasizes that modern democracy recently extended to many countries in Africa, which had a pre-colonial tradition of early democracy.

Democracy and autocracy / China was an extreme case of autocracy, the opposite of the European tradition. If the Chinese ever knew early democracy, it was in the municipal assemblies of the Eastern Zhou dynasty, in a frontier area, around the beginning of the first millennium BCE. This early democracy did not last long. After the fall of the Zhou dynasty, the empire was eventually restored and went on to build a strong bureaucracy that could dispense with popular consultation.

At the apogee of the Han dynasty, at the beginning of the first millennium CE, there was one bureaucrat for every 440 subjects in the empire. It is, writes Stasavage, “an astonishing figure for a pre modern society.” We can add that it is a very low figure for a modern democracy: in the United States, which is at the low end of government bureaucracies in the rich world, public employees at all levels of government translate into one bureaucrat for 15 residents (about one for 37 at the federal level only).

The story told by Stasavage is fascinating and we might inquire, as he does, about the lessons to draw for today’s democracy and its future. We will have to tweak his ideas a bit. The distinction between early and modern democracy is useful but it hides other aspects of democracy that we should take into account. Etymologically, democracy means the power (kratos in Greek) of the people (demos). But this does not tell us what the scope of this power is nor who are the people.

Not a value per se / These many democracies do not have the same moral value and economic consequences. Stasavage is not only a dispassionate scholar but also, as he admits, “a supporter of democracy.” He believes in “the core principle that the people should have power.” The value judgment thus expressed seems consistent with the historical claim that “participative needs … are intrinsic to humans.” However, participation in essential governance, such as building a palisade, an irrigation system, or a flood control levee — is very different from exploiting minorities in a pure majoritarian government. Alexis de Tocqueville’s “tyranny of the majority” or de Jouvenel’s “totalitarian democracy” are difficult to justify morally. Socrates was condemned to death under Athenian democracy.

Stasavage identifies a “democratic anxiety” stemming from the thin participation of citizens in modern democracy and the danger of the executive power high-jacking democracy, including in the United States. He notes the “tremendous expansion of the ability of presidents to rule by executive order.” Presidential powers, he explains, “have sometimes been expanded by presidents who cannot be accused of having authoritarian tendencies, such as Barack Obama, only to have this expanded power then used by Donald Trump.” We could, or course, as well say that the new powers grabbed by Trump will likely be used by a future Democratic president “who cannot be accused of authoritarian tendencies,” or perhaps who might legitimately be so accused.

Stasavage remains optimistic for America, probably because of what he calls the issue of sequencing in the history of democracy and autocracy. The fact that, in America and in the typical Western country, democracy came sequentially before the construction of a powerful bureaucratic state gives more chance to democracy against would-be strongmen.

Yet, let us remember the many cases in which dictators were elected or plebiscited, from Napoleon III in France (with 74% of the vote) to Hugo Chávez in Venezuela, and even in a sense to Adolf Hitler in Germany. Recently, a troubling trend has shown would-be strongmen elected in the West, such as Viktor Orban in Hungary, Jair Bolsonaro in Brazil, and Trump in the United States. A would-be strongwoman (there is such a thing), Marine LePen, could continue to rise in France. Democratic majorities, including in the United States, have often repressed minorities: think about the African Americans or the interned Japanese Americans during World War II. (See “You Didn’t See It Coming,” Winter 2018–2019.)

When he leaves the field of purely positive history, Stasavage seems to consider democracy as a value per se. If, as he tends to, we give a very large extension to the concept of democracy as a regime that gives some participation in government to at least some of the people, and if we define autocracy as all other regimes, it looks difficult to oppose democracy. But this definition hides the many varieties of democracy within and across Stasavage’s early democracy or modern democracy.

Unkeepable promises / Democracy as a value per se is not only morally suspicious, but its modern form certainly cannot deliver the participation it promises. The typical individual voter remains rationally ignorant of the political stakes because, if he is not totally deluded or ignorant, he knows that his own vote has an infinitesimally small probability of changing the election outcome. Why invest time and money in information if he cannot do anything to further his own interest? Consequently, most voters vote blind, a fact demonstrated by multiple opinion polls about voters’ ignorance.

Another reason why modern democracy cannot deliver meaningful participation is the nonexistence of what economists call a “social welfare function.” Nobel economics prizewinner Kenneth Arrow has shown that, if every individual is given an equal weight, it is impossible to aggregate all individual preferences or values and obtain coherent choices. A manifestation of this impossibility theorem is the phenomenon of cycles or voting incoherence: even if no individual changes his mind, the majority could prefer A to B, B to C, and C to A. Not to mention that, as Hayek noted, “different but equally justifiable procedures for arriving at a democratic decision may produce very different results.”

A related (but different) argument on the illusion of democracy as participation is the observation that in any nontribal society made of individuals with different preferences, virtually any collective choice must violate the preferences of some individuals and is therefore discriminatory, as Anthony de Jasay noted in his 1985 book The State. The “virtually” keeps the door open to unanimously desired choices, which are necessarily abstract rules instead of specific decisions. (See James Buchanan’s 1975 book The Limits of Liberty: Between Anarchy and Leviathan.) Thus, in any nontribal society, the state has to be distant and distrusted.

These limitations of democracy are ignored in The Decline and Rise of Democracy.

Democracy and liberty / If democracy is not a value per se, it may be an instrumental value serving to achieve some other value. In the classical-liberal tradition, this ultimate value is individual liberty, or individual consent, or a social order in which individual liberty is possible. As Nobel economics prizewinner Friedrich Hayek argued, democracy is not a means of reaching collective decisions on everything, including deep philosophical issues, but merely a procedure for electing and removing governments. (See his 1979 book Law, Legislation and Liberty, vol. III: The Political Order of a Free People.) Democracy may also have the symbolic advantage of affirming the formal equality of all individuals.

In this perspective, there is a distinction more important than early and modern democracy. It is the distinction between, on the one hand, democracy as participatory power and, on the other hand, democracy as a means to individual liberty. This distinction parallels the one made by Benjamin Constant between ancient and modern liberty — that is, between collective and individual liberty. (See Constant’s 1819 lecture “The Liberty of Ancients Compared with that of Moderns.”) Democracy as power corresponds to ancient or collective liberty; democracy as individual liberty corresponds to modern or individual liberty. This distinction cuts across the early and modern categories proposed by Stasavage. The French Revolution, which oscillated between liberation and tyranny, was representative of these two faces of the democratic Janus: individual and collective.

The conception of democracy as a protection for individual liberty distinguishes democracies from autocracies much better than does the conception of democracy as participation in power. Autocrats often call themselves “democratic,” like in the “Democratic People’s Republic of Korea.” Similarly, Chinese communist leaders have often used the word “democracy” (minzhu) to describe their system. But autocrats obviously take democracy as meaning the power of the people (whom they incarnate, of course), not the liberty of the people viewed as individuals. As Stasavage himself suggests, an autocratic government — say, a Chinese emperor — ultimately needs the support of a majority or a large proportion of its subjects even if it is not expressed through formal elections. Such governments still do not qualify as democracies in the sense of individual liberty.

When Stasavage observes that “strong central state power is a core feature of modern democracy,” he is taking democracy more in the sense of the power of the people than as the liberty of the people. It is far from clear that a strong central state with much “state capacity” is beneficial to individual liberty.

One argument for a strong modern democratic state is that early democratic assemblies could block economic development by restricting entry into markets instead of encouraging experimentation and innovation. The Dutch Republic’s weak early democratic state was apparently prisoner to special mercantile interests. By contrast, Stasavage argues that the strong state being built in Great Britain in early modern times contributed to the Industrial Revolution Yet, he also seems to agree that it was mainly by creating the space for innovation that the British state was useful, not by its direct interventions. There is much theory and evidence to support the idea that economic freedom favors economic development and growth.

Democracy, autocracy, and growth / If a strong democratic state is necessary to promote development, why can’t an autocratic state do it too? Such an argument for autocracy seems to be bolstered by the case of China. If economic historians’ estimates are correct, China had a greater gross domestic product than Europe until about 1600. It is only later that the West overcame China in growth and prosperity. Why can’t innovation persist within a Chinese-style autocratic bureaucracy?

Stasavage suggests that the political instability of autocracy and “the risk of policy reversal” may be the answer. Note that the current wave of populism suggests that strong-state democracy may not be immune to these problems.

Large autocratic empires often benefit from a large internal market with transportation and communication infrastructures and relatively unimpeded trade. This advantage, however, does not require autocracy. Modern democratic countries have it too.

As Stasavage notes, considerations about economic development are very relevant to China’s place in today’s world. The millennia-old autocratic tradition of this country suggests a less optimistic future than the hopes generated by the liberalization of the economy after Mao’s death. Many analysts thought that political democratization would follow, but it does not seem to be happening now (and American trade policy does not help). Contrary to what Stasavage seems to assume, the Chinese economy will not be able to continue growing without further economic and political liberalization. (See “Getting Rich Is Glorious,” Winter 2012–2013.) One can argue that such a fate happened before, when the Ming dynasty (1368–1644) pursued policies undermining markets.

One question Stasavage does not ask is, whom is growth for? If the goods and services of which GDP is made are produced mainly for the political or bureaucratic class and aggregated with nonmarket prices as weights, growth is meaningless as an even imperfect indicator of general welfare. A strong autocratic (or collective-democratic) state may seem to promote prosperity, but it will be the sort of prosperity that the rulers and their supporting classes prefer. Only if the state is strictly limited can economic growth satisfy individual preferences as expressed on free and impersonal markets.

Limited democracy / In the end, history illustrates why the state must be distrusted, even when it pretends to be “us.” As we have seen, political rulers will try to rule as autocrats. The state will charge in taxes what the market will bear. Only institutional constraints (sometimes helped historically by exit options offered by nature) will stop it.

Democracy is valuable only as an institutional constraint — that is, if it serves to build and maintain a limited state allowing a wide margin of individual liberty. Early democracy was tyrannical against unpopular individuals and ideas, but modern democracy leaves individuals powerless before the overwhelming power of a state claiming to represent all and everybody. “Democracy,” wrote de Jouvenel, “in the centralizing, pattern-making, absolutist shape that we have given to it is, it is clear, the time of tyranny’s incubation.” “Modern democracy,” Stasavage admits, “has a somewhat autocratic feel compared to early democracy.”

The Anti-Federalists, notes Stasavage, “argued that giving the federal government the power to levy taxes risks resulting in tyranny.” He adds: “This is not what happened in the end.” But history has not ended and the federal government’s power is not decreasing. Contra Stasavage, a strong central state may be no less dangerous at the end of democracy than at the beginning. It is a sobering thought that the American republic is now roughly the age at which the Athenian democracy died.

If it is to survive its totalitarian temptations or the takeover ambitions of autocrats, the democratic state needs a weight-loss diet and humility. That is not Stasavage’s own conclusion. The Decline and Rise of Democracy is a good and instructive book, but it needs to be completed with an interrogation on why democracy is useful and how it can be relieved from its unrealistic promises.