The Roots of Democracy

Hailed as the key to the solution of poverty, corruption, bad governance and, last but not least, terrorism, spreading democracy around the globe has become the centerpiece of U.S. foreign policy since 9/11. Unfortunately, however, this enterprise is at risk because most of our policymakers have a poor understanding of the economic and institutional landscape that is most favorable to the extension of political liberties and free elections.

The experience of the past two years in Iraq shows that simply removing the old governing elite and holding elections is unlikely to suffice to establish a peaceful democracy. On the other hand, realizing that political freedom will not happen at a snap of our fingers should not draw us into holding a fatalistic, and equally mistaken, view that democracy is impossible in that region of the world. There does not seem to be anything inherent in Islam or even Arab culture that blocks the introduction of free elections in the Middle East. It is worth remembering that not many years ago Catholicism was (wrongly) believed to thwart liberal institutions in Southern Europe and Latin America.

To make sense of the plight of Iraq, the Middle East and, for that matter, broad swaths of the developing world, we must understand first the nature of the democratic game. Democracy is a mechanism of decision in which, to a large extent, everything is up for grabs at each electoral contest. The majority of the day may choose to redraw property rights or alter the institutional and taxation landscape, thus dramatically reorganizing the entire social and economic fabric of the country. Hence, democracies survive only when all sides show restraint in their demands and accept the possibility of losing the election. In the Middle East, where inequality is rampant and wealth often derives from well-defined and easy-to-expropriate assets such as oil wells and other mineral resources, democracy poses an undeniable threat to those who profit from the authoritarian status quo. Not surprisingly, the minority in control of the state will be relentless in opposing the introduction of free elections. Thus, it is not nationalism or even religious animosities that explain the current violence in Iraq — but rather oil, its geographical distribution, and the loss of its political control by the Sunni minority that monopolized the state until two years ago.

This diagnosis has very straightforward implications for any democratization strategy. Since the absence of democracy in the Arab world and, for that matter, in regions such as Africa and Central Asia derives from a particular distribution of wealth and power, this distribution must first change (or be changed) for democracy to flourish. This in turn means that democratization is possible everywhere — that is, there are no inherent cultural, psychological, or national-character reasons that block the attainment of political freedom anywhere. But it also means that its success is much harder than many wish to believe.

Idealists versus realists

Broadly speaking, there are today two competing schools of thought on the underlying forces that have pushed for and delivered democracy around the world — and which, for the sake of brevity, we may choose to label as “idealist” and “realist.” Both are, however, mistaken.

Idealists, who currently seem to enjoy the media’s ear, explain today’s democratic momentum in a way that is strikingly similar to how past democratization waves were portrayed by their contemporary publicists. The triumph of democracy, the argument goes, is rooted in a universal yearning, intimately connected with the best part of human nature, that should lead to blossoming liberal institutions once we knock down the stifling cliques and institutions of the past. To support their claim, they point to the impressive strides made by democracy in recent decades. By 2000 there were around 100 democracies — almost twice the number in 1989 and about three times as many as there were just after World War ii (see Figure 1). More recently, even after the effects of the fall of the Berlin Wall seemed to have run their course, democracies have kept cropping up at a steady rate. A few African countries have held democratic elections. Georgia and Ukraine have recently joined the democratic pack. Iraq has had free elections. And the stirrings of the Iraqi election have been recently felt in Lebanon.

FIGURE 1
Number and Proportion of Democracies in the World, 1800–2000

The majority of mankind may indeed wish to live under free institutions. Recent surveys put at two-thirds the proportion of those that claim to prefer democracy to any other regime in each and every continent. But a look at history tells us a more cautionary tale about the chances of democratic progress. Figure 1 shows that still today, only slightly more than 50 percent of all sovereign states have a democratically constituted government. That proportion is not very different from the share of democracies in the peak years of 1920 and 1955. Those two bright moments, which came along with the spread of Wilsonian ideas in Europe and then again with the first lights of decolonization, faded quickly. Moreover, a substantial number of today’s nominal democracies are not performing as well as they should. There is little doubt that democracy has triumphed in most of the Western world. Yet democratic practices appear fragile in the core of the former Soviet Union and in Latin America. And they remain elusive in most of sub-Saharan Africa and in the Middle East. In short, we are still far away from having reached the end of history, crowned with liberal politics.

Rejecting the idealist account, realists insist that democracies do not arrive in deus ex machina fashion — and certainly not as a mechanical outcome of the introduction and spread of new ideas around the world. As scholars have pointed out repeatedly, democratic stability is well correlated with economic development. At least since World War ii, as per capita income has gone up, so has the chance a country will be democratic. Whereas only 20 percent of countries with a per capita income of $1,000 have been democratic since 1950, about 40 percent with $4,000 and 90 percent with $11,000 or more have. Furthermore, no democracy has collapsed in any country with a per capita income over $7,000.¹

Yet even these seemingly robust results are, upon further reflection, weak and unconvincing. First, the threshold of development at which democracy becomes likely has varied over time. Before 1945, 90 percent of all countries with a per capita income of $4,000 were democracies (versus 40 percent afterwards). Or, to put it differently, nineteenth-century Norway and Switzerland were democratic with a per capita income not dissimilar to what many less developed countries have today. Second, there are glaring exceptions to the relationship between development and democratic regimes. Germany and England had similar incomes but experienced very different political fortunes in the 1930s. India is a democracy in spite of its economic performance. Most oil countries are rich but impervious to liberal institutions. Last but not least, there are still no good explanations for how economic development boosts the chances of democracy. In short, we need to push further our inquiry about the roots of democracy.

Realists are right in claiming that democratic life is possible only when certain, mostly material conditions are in place. But, in acting like the drunkard that searches for his lost keys onlyunder the lamppost — that is, by looking at the easy-to-measure variable of income — they have missed the true nature of those conditions. It is, rather, excessive economic inequality, particularly in agrarian countries and in nations rich in oil and other minerals, that exacerbates the extent of social and political conflict to the point of making democracy impossible. In an unequal society, the majority resents its diminished status. It harbors the expectation of employing elections to drastically overturn its condition. In turn, the wealthy minority fears the outcome that may follow from free elections and the assertion of majority rule. As a result, it resorts to authoritarian institutions to guarantee its social and economic advantage. By contrast, in societies endowed with some relative social and economic equality, inhabitants are willing to accept the inherently uncertain results of free elections — that is, they are willing to agree to be temporarily reduced to the status of minority and to be governed by the party they oppose.

The democratic game

Truly free elections are like fair card games. Their consequences may be clear-cut and often rather divisive. But they begin as very uncertain events. Before the cards have been shuffled and players have played their hands, there is no winner at the table. Similarly, before voters cast their ballots, the politicians who run for office cannot predict whether they will be elected or not. But once the electoral campaign is over and the winning candidates have assumed the offices or seats under dispute, losers must give up all their claims to power. They must simply wait for new elections (to be held at some point in the future) to have any chance to attain power. In the meantime, they have to accept the decisions and comply with the policies of their electoral adversaries.

The electoral process carries no guarantees, in itself, that any of those politicians, who in most instances opposed each other rather ferociously during the campaign, will respect the terms and continuity of the democratic procedure. The loser, who is now governed by the winner, may abide by the election, accept the defeat. and wait until the next electoral contest takes place. But she may be inclined to denounce the result, mobilize her supporters, appeal to courts and international observers, and perhaps even stage a coup to grab by nonelectoral means the office she lost at the ballot box. In turn, the winner may have an incentive to use his position to shift public resources to fatten his campaign chest and boost his electoral chances, to twist the rules that govern elections and, last but not least, to delay or cancel any future electoral contests.

Hence, a stable or successful democracy — that is, the uninterrupted use of a free and fair voting mechanism to make political decisions and select public officials — is possible only if both the winner and the loser (or, more generally, the majority and the minority) comply with the outcomes of the periodic elections they have set up to govern themselves. The minority, which is now at the mercy of the majority, must accept its defeat. And the majority, which now controls the levers of the state, must resist the temptation of permanently shutting out the losers from power. In short, liberal institutions exist only when all parties consent to the possibility that in the future they may have to exchange their roles in the political theater, with today’s winners becoming tomorrow’s losers and the current losers turning into the future winners.

Equality of conditions

When will voters and politicians abide by the rules of the democratic game? Think again of elections as a game of cards. If too much is at stake — that is, if bets are too large — the incentives to cheat become irresistible. Similarly, the participants in elections will assent only to the rules of the democratic game if the effects of the electoral outcome do not fall on any of them too heavily. The losers will submit to the electoral result if what they forsake at the moment of defeat is not too excessive, that is, if it does not threaten their living standards or political survival. Likewise, the winners will not exploit their preeminence to redraw the electoral mechanisms (to diminish the uncertainty of future elections) only if the value of the offices they hold and the political decisions they make is not too large.

Generally speaking, democracy will be possible only if both winners and losers — that is, if all voters and their representatives — live under some relative equality of conditions. When voters do not differ excessively in wealth among themselves, not much is up for grabs in elections. Democracy is then a quiet business, feared by few and welcomed by most. By contrast, if social and economic inequality is rampant — that is, if a few control most wealth — the majority will look forward to an election as an event whose outcome will enable them to redistribute heavily to themselves. Facing such strong pressure for redistribution, the wealthy will prefer an authoritarian regime that would exclude the majority of the population and hence block the introduction of high, quasi-confiscatory taxes.

The insight that equality of conditions is a precondition for democracy has a long and often forgotten tradition in the study of politics. It was apparent to most classical political thinkers that democracy could not survive without some equality among its citizens. Aristotle, who spent a substantial amount of time collecting all the constitutions of the Greek cities, concluded that to be successful, a city “ought to be composed, as far as possible, of equals and similars.” By contrast, he noticed, a state could not be well-governed where there were only very rich and very poor people because the former “could only rule despotically” and the latter “know not how to command and must be ruled like slaves.” They would simply lead “to a city, not of free persons but of slaves and masters, the ones consumed by envy, the others by contempt.”² Two thousand years later Machiavelli would observe in his Discourses that a republic — that is, a regime where citizens could govern themselves — could only be constituted “where there exists, or can be brought into being, notable equality; and a regime of the opposite type, i.e. a principality, where there is notable inequality. Otherwise what is done will lack proportion and will be of but short duration.”³

More contemporary empirical evidence will be brought to bear later in a separate section. But a quick look at the history of the past two centuries shows that equality loomed large in the choice of political institutions. Big landowners have always opposed democracy, whether in Prussia, Russia, the American South of the nineteenth century, or Central America in the twentieth. By contrast, for democratic institutions to prevail, at least before industrialization, there had to be a radical equality of conditions. The Alpine cantons of Switzerland in the Middle and Modern Ages or the Northeastern states of the United States in the eighteenth and nineteenth centuries are cases in point.

Land, minerals, oil

Faced with the risk of suffering high taxes at the hand of a democratic majority, the wealthy minority may choose an alternative path to authoritarian rule as a form of protection. Or, instead of investing in repression and violence, it may simply decide to take its assets somewhere else. Or it may even hide (or underreport) its wealth.

But which course the wealthy pursue depends on the type of property they own. If the sources of their income are not mobile — that is, if they own lands, mines, or oil wells — they cannot transfer their assets abroad. And so, whenever they are threatened by a majority hungry for redistribution, they have only one solution: controlling the state, repressing the opposition, and raising all types of barriers to the exercise of freedom and to elections. By contrast, if wealth is mobile and therefore can actually escape the brunt of excessive taxation by fleeing abroad, its owners will tolerate democracy. As a matter of fact, voters, knowing that asset holders may exercise their exit option, will moderate their fiscal demands. And with some of their bluntness gone, democracies will be more likely to survive, even if economic inequality was not low. The recent transition to democracy in South Africa is a good example of this logic. While opposition to democracy ran high among the Afrikaner farming communities, it hardly existed among the English-speaking financial and industrial elites, who could easily (and actually did) move their capital abroad. In fact, over the postwar period, South African prospects for democratization improved as part of the Afrikaner community gradually moved from farming to industrial and financial activities — that is, from holding fixed assets to investing in mobile capital.⁴

Economic development and industrialization go hand in hand with the expansion of education, the formation of a skilled labor force, and hence with growing equality of wages and conditions across the population. But economic development is also the story of a shift from highly immobile fixed assets to progressively more mobile capital — that is, from societies that rely on the exploitation of mines and agricultural land to economies based on manufacturing industries and human-capital-intensive businesses.

The transformation in both the distribution and the nature of wealth that is at the heart of economic growth explains why many recent empirical studies have found democracy to be well correlated with the level of per capita income. But it also accounts for the well-known fact that that correlation breaks down for most oil and mineral-rich countries. Since 1950, 80 percent of all non-oil-exporting countries with a per capita income over $8,000 have been democracies. The proportion is roughly reversed among high per capita income countries whose export revenues from oil amounted to 25 per cent or more of total trade revenues.

In those economies where industrialization did not take off and natural resources remained or became the sole or main source of wealth, authoritarianism was likely to prevail. Democracy is too threatening to the riches that accrue to the owners of oil wells and mineral fields. Since they cannot flee their country to avoid the potential nationalization of their assets, they have no other option than suppressing demands for liberalization.

Islam has been much brandished as the cause of authoritarian attitudes and institutions in the Middle East and North Africa. But as Freedom House recently pointed out, if we take into account the large Muslim populations of countries such as India, Indonesia, Bangladesh, and Turkey, the majority of the world’s Muslims live now under democratic regimes.⁵ In turn, some scholars have noted that, even if Islam is compatible with free elections, the Arab world is not. Indeed, all Arab states remain undemocratic as of today — and do so by employing substantially repressive policies. The problem with this claim, however, is that it never specifies the ways in which Arab culture and behavior may be at odds with the principle of mutual toleration among winners and losers that makes democracy possible. Moreover, the few surveys we do have seem to show that Middle Eastern populations favor democracy by margins similar to those found in Latin American or Asian publics.⁶ The truth is that the politics surrounding the control of natural resources, rather than any religious or cultural factor, is what explains the preponderance of authoritarianism in the Middle East (and much of sub-Saharan Africa as well).

At this point it is worth stressing that oil or mineral wealth does not lead to authoritarianism on its own. In other words, nothing said so far implies that a democratic country that discovers oil in its seas will revert to dictatorship. It will certainly not do so if it can sustain a set of institutions that spread the newly acquired wealth in a fair manner across the population and that block any attempt by a minority to monopolize the newly discovered assets. Norway is a good example. Before oil was discovered in the North Sea and its flow transformed that country into one of the richest economies of the world, Norway had a well-established and well-governed democracy. Its government managed to craft a set of institutional devices to make sure that oil would not corrupt the political and electoral accountability mechanisms already in place.

By contrast, the discovery of natural resources is fatal in already authoritarian or weakly democratic societies. Unconstrained by any domestic mechanism of control, the existing political and economic elites alone benefit from those natural riches. The new wealth reinforces their political muscle. It reduces, too, any incentives to follow economic policies that may foster growth and industrialization. Were the latter to succeed, new social and economic strata might arise that could in turn challenge the existing political status. In short, illiberal countries that suddenly strike oil or any other mineral wealth become trapped in authoritarianism. This is the standard story of many oil- and gas-rich Middle Eastern and North African countries. It is the tale of all the former Soviet Union states in Central Asia. It partially explains the fragile nature of liberal institutions in Russia. It fits the pattern of violence and dictatorship that devastates many African nations. And, in combination with appalling levels of inequality, it is behind the turmoil of several Latin American economies, such as Venezuela, Ecuador, and Bolivia.

Political violence

If authoritarianism is well entrenched in a country — that is, if the governing elite is not willing to give up any of the levers of power — then the opposition may have only one tool to take over the state: political violence.

The use of openly violent means to achieve power will be most likely in countries that are highly unequal and where wealth is mostly immobile. As already stressed above, the higher the income inequality, the more resistant the well-off will be to the introduction of democratic, peaceful means to set government policy. The losses they would incur from majority rule would be too substantial. By the same token, resorting to violence to effect political change will also become more attractive to those excluded from the state. Since violence implies an enormous loss of human life (without any certainty of victory), opposition groups will risk a civil war only if inequality is substantial — that is, if the existing elite owns a considerable fraction of the economy. Moreover, political violence will become particularly acute in unequal economies in which assets are fixed. In countries with abundant natural resources, the potential rebels can apply violence to overturn the existing regime relatively certain that if they win, major assets will be unable to flee the country.

FIGURE 2
Type and Distribution of Wealth and Political Regimes

A convenient way to summarize our discussion so far is presented in Figure 2. The vertical axis of Figure 2 depicts the level of inequality, rising as one moves upward in the graph. The horizontal axis captures the level of asset immobility, growing to the right. Each point in the graph indicates a particular combination in the nature and distribution of wealth. We can associate each point in the graph with the most likely type of political regime and level of political violence we would expect based on the theory presented here.

As shown in Figure 2, democracy would be expected to prevail either when equality is high or when assets are relatively mobile — that is, all the area to the left of the first curve. By contrast, authoritarianism should be more common as both equality and capital mobility decline (and the incentives of the holders of assets to accept a democratic outcome fall) — all the area to the right of the first curve. Finally, political violence should flare up in unequal countries rich in natural resources — that is, the upper right corner of the graph.

Contemporary evidence

With figure 2 in hand, we can now turn to see how well theory matches our contemporary evidence. I do so in three steps. First, I offer data on types of political regimes worldwide since the middle of the nineteenth century. I then consider the American experience and the differences across its states. Finally, I examine the location of political violence around the world and its causes.

Political regimes across the world. Following the same format of Figure 2, Table 1 presents the distribution of countries (or, more exactly, the proportion of country-years) that have been democratic in the world since 1850, classified by type of wealth and its distribution. Since data for such a long period of time are hard to come by, here I employ two relatively good proxies. The type of wealth is measured through an average of the percentage of the economy in manufacturing and the percentage of the population living in cities — the higher the percentage, the more mobile wealth is assumed to be. The distribution of wealth is measured through the distribution of land or, more precisely, through the percentage of arable land in the hands of family farms (as opposed to big landowners); the higher the proportion of family farms, the more equal the economy. The two variables decrease as we move upwards and to the right — in this way Table 1 can be read in parallel to the predictions summarized in Figure 2.

TABLE 1
Proportion of Democracies in the World, 1850–1995

		80–100	60–80	40–60	20–40	0–20
	0–20	No data	60	42	9	1
	20–40	100	91	43	27	5
	40–60	100	100	71	48	10
	60–80	100	100	84	72	0
	80–100	No data	100	91	14	No data
Note: The percentage in each cell refers to the proportion of country-years that are democratic over the total number of country-years in each category.

Table 1 fits Figure 2 well. In highly urbanized and industrialized areas, a very high proportion of countries are democratic. The proportion of democracies is also very high in equal societies, even if they are hardly developed. But for underdeveloped and unequal countries, authoritarianism prevails. In the upper right corner, only 1 percent of the cases are democratic.

Cultural or religious variables play a substantially secondary role in the democratization process compared to the impact of the nature and distribution of wealth. Employing a statistically more systematic analysis of the evidence we have for all the countries of the world since the early nineteenth century, it is possible to show that the religious composition of the population does not have any systematic impact on the type of political regime in place.⁷ Refuting any Weberian assumptions about religion and politics, Catholic countries are not any more likely to be authoritarian. A growing proportion of Muslims does not jeopardize democracy either, once we take into account the structure of the economy and the weight of fixed natural resources. Democratic instability grows only with more ethnic fragmentation, but this result takes place only in middle income economies.

Democracy in America. A second piece of empirical evidence comes from the political history of the United States. The forces that shaped the level of democracy across its states are similar to those that operate at the world level.

Before the Civil War, the quality of democracy in the different states of the United States was strongly related to the equality of economic conditions across the country. The Northern states were fairly equal societies, at least in comparison to Europe. Whereas the top 1 percent of men held 29 percent of gross assets in the United States in 1860, the proportion was 61 percent in the United Kingdom in 1875. Between 1820 and 1850, the ratio between the wages of skilled and unskilled workers remained stable at around 1.5 on the northeastern seaboard. In England the ratio peaked at 2.6. Not surprisingly, about two-thirds of all adult white men had the right to vote in America in 1790. That proportion climbed to 90 percent before the Civil War in most cases. By contrast, in those northeastern areas experiencing an industrial revolution and the rapid formation of an urban proletariat, such as Massachusetts and Rhode Island, taxpaying requirements were maintained or reinforced and resulted in the exclusion of about a third of adult white men. And in the South, the very restrictive franchise in place was tightly linked to slavery.

FIGURE 3
Franchise Restrictions in the United States in 1910

For a few years after the Civil War, the imposition of military rule in the South led to the democratization of the entire country. But the removal of the last federal troops in the 1870s opened the way to the exclusion of blacks and a portion of poor whites in the South. Figure 3 shows the degree of formal restrictions on the right to vote (such as poll taxes and literacy and residency requirements) across states in 1910. Economic conditions explain a good deal of the variation in place. In the highly rural and highly unequal Southern states, franchise conditions were very exclusionary. Only the combination of economic growth, considerable emigration, and federal intervention brought democracy back in the 1960s. In the Midwest and especially in the Plains, electoral requirements were scarce even though their economies were agrarian because economic equality was high. Finally, in spite of relatively unequal conditions, electoral restrictions were either very mild or completely absent in the Northeast. The ratio of the managers’ salary to the wage of domestic servants was similar to that of the South. Inequality, moreover, increased at the turn of the twentieth century, driven by the arrival of about 25 million immigrants. An agitated public opinion approved the exclusion of paupers in half a dozen states of the Northeast coast and favored the introduction of new systematic registration procedures, which purportedly led to a substantial fall in turnout among unskilled workers in the first decades of the twentieth century. Still, partly due to the political opposition of urban machines and partly due to a flourishing urban economy with very mobile assets, both the threat posed by the immigrants and the ensuing political backlash were much more subdued than in the agrarian South. As a result, even if imperfectly, democracy persisted in the North.

Violence. The use of violence to effect political change is a generalized phenomenon around the world. Some estimates put at 137 the number of civil wars during the period from 1820 to 1990 and at 16.2 million the death toll from civil wars fought after World War ii. Between 1919 and 1997, there were over 500 spells of guerrilla warfare around the world. In the same period, close to 1,500 politically motivated assassinations or attempted assassinations of high government officials or politicians were committed — an average of one every three weeks.⁸

In a way similar to Table 1, Figure 4 displays the average level of industrialization and urbanization (on the horizontal axis) as a proxy for the type of assets and the percentage of family farms (on the vertical axis) to capture the degree of concentration and therefore inequality in the ownership of land. Both axes are drawn in the reverse order (decreasing in value as one moves away from the origin) so that the high inequality/high specificity area is in the upper right corner. Figure 4 then plots as small black dots all the country-year observations from 1850 to 1992 for all sovereign states. It marks the onset of civil wars with black squares.

The graph shows considerable variation in the combination of the level of industrialization and on the inequality of the agrarian sector across countries. Yet most civil wars have occurred in countries where the agrarian sector was still dominant and land was distributed unequally (basically within the triangle to the right of a diagonal going from no industrialization and less than 50 percent of the land to middle levels of industrialization with no family farms at all). Several cases that are closer to the middle (that is, farther away from the upper right corner) have considerable oil resources, and so conflict there may be related to asset immobility. All in all, the distribution of observed civil war onsets matches quite well the predictions expressed in Figure 2.

Figure 5 does the same for the distribution of guerrilla warfare from 1919 to 1997. The occurrence of guerrilla war is more widespread than systematic civil war, but the pattern is still similar: Violence is heavily concentrated in unequal agrarian economies.

Preparing the soil for democracy

Given that democracy flourishes only once certain social conditions are in place, what can be done? Can we actively shape them to foster democratization? In other words, can we reshape social conditions in a country to satisfy the underlying economic requirements for a successful political transition to democracy?

The answer cannot be and is not a simple one. The door to liberal democracies undoubtedly exists. But it is narrower and its opening harder than is often granted. Or, to put it differently, policymakers need to understand that they are confronted with sharp trade-offs: between short-run versus long-run solutions, between violent and not-so-violent strategies of intervention, between betting on economic development to change political institutions over the course of one or more generations and toppling the elite rule of the ancien régime through war and occupation. In a way, the very acrimony of the current debate about the democratization of the Middle East is the best demonstration of how hard it may be to adopt clear-cut policies and follow them through.

Historically, democracies have replaced authoritarian regimes through two paths. On the one hand, democratic institutions have emerged after a long process of economic development spreads material wealth across society, equalizes economic conditions, and erodes the strength of the old authoritarian elites. On the other hand, absent economic modernization, social and political change has happened only after enormous violence — generally through military intervention of a foreign power.

FIGURE 4
Economic Structure and Civil War Onsets, 1850–1992

FIGURE 5
Economic Structure and Guerrilla Onsets, 1919–1997

Before the irruption of commercial and industrial capitalism in modern Europe, most wealth was fixed in the form of farmland and mines. A few agrarian communities (mountainous Switzerland, Norway, or Iceland) were equal and democratic. But most pre-industrial societies were (and are) characterized by the combination of inequality, authoritarianism and underdevelopment.

Authoritarianism is pervasive in an agrarian economy for a simple reason. In a Hobbesian world infested by bandits and generalized war, autocrats are a standard, reasonable mechanism to enforce peace and to protect the peasant population against plunder and death. Still, the price of authoritarianism is inequality. In exchange for protection against bandits like themselves, rulers such as the Bourbons, the Tudors, or the Sauds seize an important part of their subjects’ assets. For example, at the death of Augustus (14 A.D.), the top 1/10,000 of the Roman Empire’s households received 1 percent of all income. In Mughal India around 1600 A.D., the top 1/10,000th received 5 percent of all income. In fact, the annual income of the Indian emperor was the equivalent of the wage of about 650,000 unskilled workers.

The formation of the state and the pacification of its territory made possible agriculture and the extension of some mild forms of commerce and industry. But, overall, growth occurred at a snail’s pace. Worried about the emergence of economically independent strata that may eventually challenge their political preeminence, authoritarian rulers favored the maintenance of those noncommercial, pro-land policies that were the basis of their wealth and power. Moreover, the king’s vassals had no legal mechanism to resist any of his potentially arbitrary actions. With property rights insecure, very few individuals had any incentive to invest in new businesses and create new forms of wealth.

Although coming in sundry forms and with different degrees of intensity, this political and economic landscape of stagnation dominated the whole world until the modern period. Its transformation and the progressive democratization of previously illiberal societies took place through two different paths. The first one developed in the long haul, caused by economic modernization. The second path was short and abrupt, triggered by war and occupation.

Democratization resulted, on the one hand, from modern development. Commercial capitalism, then followed by an industrial take-off, led to the spread of wealth, the erosion of the relative value of immobile assets and natural resources, and more economic equality. These new conditions then made the transition to liberal democracy possible. This economic and political transformation proceeded in waves. It first happened in an almost self-generating fashion in a few places located in the North Atlantic area — Britain, Belgium and the Netherlands, the Rhine area of Germany, Switzerland, and the Northern states of the United States — where no monarch was able to suffocate pre-existing medieval and pluralistic institutions in the name of modern absolutism. The parliamentary institutions of those nonabsolutist states protected the interests of merchants and investors and hence allowed the latter to take advantage of the scientific revolution of the seventeenth and eighteenth centuries. As capital accumulated in the already developed core, it gradually spilled over to the near periphery — particularly when the latter had either stable political institutions or foreign military pacts (generally with the United States) that credibly protected capital against the threat of expropriation. This is the story behind the boom of Southern Europe and, to some extent, of East Asia in the postwar period. Once those countries grew in the 1960s and 1970s, they went through very peaceful transitions to democracy in the last quarter of the twentieth century.

In those countries that had neither an equal agrarian economy, like Norway or some Swiss cantons, nor equalization through economic development, democratization rarely came peacefully from within. Even enlightened tyrants do not pass economic and institutional reforms to equalize conditions, since doing so would jeopardize their grip on power. It is true that authoritarian states sometimes push for economic reforms to industrialize their countries, as Meiji Japan did in the late nineteenth century. But their reforms, mostly implemented in response to foreign military competition, rely on the heavy intervention of the state and the creation of big industrial conglomerates tightly linked to the governing elite, hence avoiding a distribution of assets conducive to democracy.

Without society-centered economic development, the destruction of the old authoritarian elite (and of the institutions that blocked growth) comes about only as a result of war, defeat, and foreign occupation. This is the case of Central and Eastern Europe and of East Asia. It took World War ii and the Allies’ victory to destroy the ancien régime’s social coalitions and political institutions hindering democracy and economic development. The story of political instability and authoritarian governments that burdened Germany and Italy in the first half of the twentieth century ended only with American occupation. Similarly, the United States democratized Japan and imposed key agrarian reforms in Korea and Taiwan that would then sow the seeds for growth and liberal institutions. Although its consequences were otherwise catastrophic, the Soviet occupation of Eastern Europe made tabula rasa of the past quasi-feudal structures of that area. Once the ussr collapsed, Eastern Europeans could easily transit to democracy in a way they were unable to before World War ii.

Spreading democracy?

In taking stock of how democracy and, more generally, political modernity came about across the globe, we can derive some lessons for political action in our time. First, we need to act as “realists” in thinking about the true foundations of liberal regimes. The unprecedented expansion of democracy in the past 20 years might not be the final act of world history. The past century alone has already witnessed democratization waves that then quickly ebbed in dramatic and unexpected ways. To put it differently, it is not enough to hold free elections to inaugurate a full-grown democracy. Democratic constitutions take root only in fertile ground, and fertile ground requires in turn much thorough and patient work.

Second, this work cannot consist of half measures. Tying foreign aid to domestic political change is unlikely to transform the governing institutions of the recipient country for two reasons: There is normally too much at stake in the control of the state to convince the ruling elites to establish a truly free electoral arena in exchange for cash flows. More important, foreign aid has little effect without exacting control of its use and of the recipient’s compliance with its attached conditions. But monitoring by democratic donors is notoriously imperfect. In fact, to be effective enough requires a level of intervention close to occupation and re-colonization. And this brings us back into the war-and-occupation path discussed earlier.⁹

Third, democracy happens after growth and economic modernization happen. But even if this is in a way a peaceful path to democracy (and sometimes it is not, since countries may need a Glorious Revolution to get the right set of economic institutions), this process takes considerable time. It took East Asia and Southern Europe a few decades to join the democratic club.

Two additional points must be made on the possibilities of promoting economic change. The debate on what policies yield growth is still much alive and mostly unresolved among economists and policymakers. The question of how to implement pro-growth strategies is even more disputed. Short again of direct intervention, democratic countries have two tools at hand to foster growth: (1)foreign aid in exchange for institutions that protect property rights and foster investment and (2) free trade and liberal immigration policies to give Third World producers and laborers access to Northern markets. Both instruments have a somewhat diminished political status around the world today. Conditional aid has been much depreciated by a generalized backlash among the Latin American public in recent years. Open markets are in turn increasingly contested among substantial parts of the European and North American electorates. Moreover, even if economic growth may be quite desirable to developing populations, the ruling elites of nations rich in natural resources have very little incentive to establish a modern economy for a very particular reason: The sizeable revenues they extract from oil and mines shelter them from having to create the set of institutions that would generate growth but then would certainly empower their populations and shake the authoritarian status quo.

Finally, short of the path of economic modernization, democracy can be established only through radical and violent change imposed from abroad. The liberal left has been wrong to dismiss the neoconservative logic of war and foreign-led transformation applied to democracy. Quick democracy can only come in this way because authoritarian elites have no reason to give up their control of the state. It is true that, in some dictatorships, internal turmoil, in the form of civil war and revolutionary action, can turn the tables upside down. Historically, however, revolutions have hardly led to democracy. They have generally resulted in a continuous succession of authoritarian regimes and outbursts of violence. This is the story of both Latin America in the nineteenth century and a good chunk of the twentieth century and Africa today. But the fact that foreign imposition of democracy has happened in the past does not mean that we should rush to embrace the solution of war and occupation. This strategy involves extremely heavy-handed and painful measures. To be sustainable, it needs considerable resources, exceptional resoluteness, and, therefore, broad consensus at home. Without generalized support, it ends up in failure — as shown by the partial reconstruction of the American South after the Civil War and its abandonment in the 1880s. And even if it is carried through, no one assures us that the occupier will behave as a benevolent planner keen on leaving the country fully democratized and reformed.