Now that the furor surrounding the initial incursion of rational choice methods of inquiry into the field of comparative politics has died down and a second generation of research has matured, the approach is showing its full analytic power. Economic Origins of Dictatorship and Democracy uses the technical apparatus provided by game theory to deliver empirically important insights about the origins of modern political regimes at both the micro level and, even more intriguingly and perhaps somewhat surprisingly, at the macrohistorical level. As its title suggests, with its allusion to Barrington Moore's still widely read Social Origins of Dictatorship and Democracy (1966), this is a big, theoretically ambitious book of the type that one feared had been permanently eclipsed in political science with the adoption of ever more technically demanding methods of investigation. Daron Acemoglu and James Robinson offer new insights into an old question of major theoretical significance—namely, how to understand the origins of modern political democracy—and they do so using a remarkably elegant and unified game-theoretic framework buttressed by historical case studies.

This book aims to explain why some countries develop stable democratic political institutions whereas others do not, and yet others alternate between democratic and autocratic rule. The gist of Acemoglu and Robinson's class-based interpretation of institutional change is that democratic institutions are established because it is not credible for predemocratic political elites, in the face of a revolutionary threat by the masses, to promise citizens that, once adopted, redistributive policies will remain in force even in the future. Representative political institutions, on the other hand, once established provide the assurance that future policies will reflect the interests of the majority. The creation of democratic institutions is, in other words, designed to lock in a commitment that guarantees citizens majority control over the formulation of economic (redistributive) policies in the future. The authors work out this explanation formally, and they illustrate it with historical materials drawn from Britain, Argentina, Singapore, and South Africa.

In this view, democratic political institutions are fundamentally a device to ease social conflict and prevent revolution by the lower classes. As the authors write: “democratization … [is] a commitment to future pro-majority policies by the elites in the face of a revolutionary threat and, perhaps more important, … a commitment made credible by changing the future distribution of political power” (p. 27). This account is deeply informed by British history, where chronic unrest and the fear of revolutionary upheaval similar to that taking place across the English Channel drove British elites over the course of the nineteenth century to adopt legislation that gradually enlarged suffrage and introduced parliamentary democracy.

Underlying Acemoglu and Robinson's formulation is the distinction between de jure and de facto political power. De jure political power arises from the structure of institutions regulating access to political power. Examples that the authors provide include suffrage and electoral system laws, because these set the rules for who has the right to vote and for how votes will be counted. De facto political power, by contrast, arises from the immediate relations of force between social groups, typically though not necessarily understood empirically to be social classes. When de facto power relations shift, an actor may seek to establish de jure political power by establishing institutions that guarantee the ability to shape policy outcomes in the future. Citizens, in other words, knowing that a situation of de facto political power will not endure indefinitely, press elites to establish democratic political institutions that will give the majority de jure political power for the future.

How, in this formulation, does variation arise in the adoption of democratic institutions? Why do some countries establish democratic institutions whereas others do not? The differences in observed outcomes—classified by the authors as nondemocracy, democracy, unconsolidated democracy, and repression—derive from differences in strategic interactions between elites and citizens that depend, in turn, on the preexisting level of economic inequality in society. In the Acemoglu-Robinson world, the preferences of elites and masses for redistribution arise from the extent of existing inequality, which can be neatly conceptualized as the share of total income accruing to the rich. These preferences lead to different strategic interactions between elites and citizens. In their theory, democracy emerges as an equilibrium outcome only when an intermediate level of inequality exists. In situations of low inequality, citizens have less incentive to press for democracy since there is not much to redistribute from elites; in contexts of high inequality, elites are more likely to resist democratization with repression because redistribution is so threatening to their economic well-being. The aggregate result of these competing interactions is, thus, that the relationship between intergroup inequality and democracy is parabolic. Democratization occurs at intermediate levels of inequality, rather than when inequality is very low (when citizens do not have strong revolutionary incentives) or when it is very high (when elites are likely to use repression to retain political power). Universal suffrage is established when the lower classes favor redistribution and instigate revolutionary upheaval to obtain it and when, in addition, elites decide to concede rather than repress.

The core formalization shaping the analysis in Economic Origins is the median voter model, laid out in detail in Chapter 4. In this model, the tax schedule adopted by government is that preferred by the median voter. In almost every society for which data are available, the median income is in fact lower than the average income; as a result, the median voter prefers to redistribute income. This model informs a wide range of theoretical and empirical studies of redistribution, among other things. Acemoglu and Robinson extend this framework in new directions to offer predictions about when elites will concede democratic institutions in order to stave off revolution and when they will not. Of equal importance, they present their models in such a way as to make the book accessible even to political scientists with relatively little formal training. Their modeling is, therefore, at once interesting and readable. It successfully unifies what might otherwise appear to be a collection of distantly related empirical observations. In fact, the use of formal modeling makes the book straightforward and easily interpretable.

It was not until 1973, when Theda Skocpol published what we might think of as the Cliff Notes to Social Origins of Dictatorship and Democracy, that Barrington Moore's long and initially bewildering tome became easily accessible to a wide professional audience. Acemoglu and Robinson's study will not need similar secondary treatment to be widely adopted in graduate seminars and to influence future generations of scholarship. Throughout, their argument is easy to follow. The value of the parsimony and elegance offered by formal models is well represented here.

What issues does the book raise without fully resolving? I see two main sets of concerns. First, the distinction between policies and institutions is neither clear-cut nor unambiguous. Acemoglu and Robinson contend that this distinction is precisely where the commitment problem bites: If the legislature is not elected on the basis of universal suffrage but rather by a restricted franchise, then the law is not credible for the future because if the ruling elite decides it is inconvenient or too expensive, the law can be arbitrarily overturned unless the majority of citizens is in possession of sufficient de facto power at the time. This is why, they contend, the majority threatens revolution: because redistributive policies are not credible unless locked in institutionally.

Some policies lock in specific distributional schedules for the future, however. Perhaps the foremost example of a policy almost impossible to reverse is land reform. Once land is seized, carved up, and distributed to peasants, it is hard to imagine how it could be reconsolidated and handed back to its original owners. The adoption of suffrage reform, conversely, though classed by Acemoglu and Robinson as a purportedly irreversible institutional commitment, has historically been reversed frequently. Consider France, which adopted universal male suffrage in the constitution of 1793—a constitution that was never implanted— and adopted it again in 1848, after which it was reversed; adopted it a third time in 1852; and finally adopted universal male suffrage definitively only in 1884.

The French example illustrates that if the elite controls de facto power even under democratic institutions, there is little to prevent that elite from arbitrarily altering the constitution. This in turn may make institutions more fragile than the authors would like to admit. Acemoglu and Robinson's answer to this is that “policies, even though they can sometimes be difficult to reverse, are generally easier to reverse than institutions” (p. 174). But this, though perhaps nominally true, is analytically unsatisfactory since it concedes that institutions are reversible. It asserts by definition what needs to be demonstrated. If we consider as well that all rules are subject to strategic manipulation by rational agents, the problem becomes even more severe. It is a good deal more difficult to distinguish policies from institutions than the authors admit. Since this distinction underlies the entire argument of the book, clarifying it will be an important contribution for future work.

Second, the empirical evidence provided in the book in support of the hypothesis that democracy arises in contexts of middling levels of economic inequality is suggestive but not systematic enough to be entirely convincing. Acemoglu and Robinson offer scattershot historical examples that, they contend, corroborate their theory. In Latin America, they argue, inequality between citizens and a small elite was very high, leading the elite to use force to resist democratization for much longer than elsewhere. In Singapore and some other Asian nations, by contrast, inequality is less pervasive than elsewhere, leading, they argue, to less revolutionary pressure from below. But just what is meant by low or high is not clearly operationalized or used in an entirely consistent way throughout the book. An additional if inevitable problem for long-term historical studies is determining the point at which to measure the outcome: When was democracy consolidated in those European countries that succumbed to fascism or to communism, for instance?

Acemoglu and Robinson justify their use of historical examples rather than quantitative analysis with the claim that available measures of inequality, such as the Gini coefficient, do not accurately tap the concept of intergroup inequality that they believe theoretically relevant. This presents a genuine conundrum. Without systematic evidence, we cannot assess if their argument is true, but the authors have ruled out the use of available systematic data. They therefore rule out systematic tests of their main claims. This is doubly disturbing since at times, the case-based evidence presented by Acemoglu and Robinson is incomplete or inconsistent with their theory. On occasion, they play a little fast and loose with the evidence, such as their unsatisfactory discussion of the data on British income inequality in the nineteenth century (p. 70).

Of course, this very weakness will provide generations of future scholars a prepackaged research agenda. Perhaps if extended back far enough historically, the newly available data on the incomes of the highest 1% in selected countries that has been assembled by Anthony Barnes Atkinson and Thomas Piketty Top Incomes Ove the Twentieth Century (2007) will prove relevant to Acemoglu and Robinson's claims and can be used to assess them. Arguably, the income of the top 1% is a good measure of elite income, so that data on it combined with data on average income might provide empirical leverage for investigating the political impact of intergroup inequality. Although I do not believe that this volume presents sufficient evidence to nail their case, they have offered a compelling theory supported by enough evidence to make it plausible. Arguing about whether they are right or wrong will be good fun in seminars and classrooms for years to come.

Does Economic Origins therefore successful meld theory with empirics so as to render obsolete the contention made in 1994 by Donald Green and Ian Shapiro that rational choice offers nothing of value for our empirical understanding of politics? Despite the extraordinary value provided by a book that has Barrington Moore and Theda Skocpol speaking directly to William Riker and Anthony Downs, I am not sure it will convince anyone as skeptical as Green and Shapiro. Even so, it is a creative and agenda-setting book with more interesting and nuanced theoretical implications than this review has had space to mention. It qualifies as required reading for students of comparative politics.