Economic Origins of Dictatorship and Democracy

by Daron Acemoglu

  Nevertheless, there is always uncertainty about what the future holds. For instance, the attempt to induce democratic consolidation through constitutional engineering in Zimbabwe has not been a great success. It is interesting that in his assessment of the future for democracy in South Africa, Thompson (1995, p. 275) notes there is “one great structural threat to democracy in South Africa: The lack of a well educated skilled labor force - the consequence of the abysmal state of education in South Africa.” Our analysis in Chapter 9 suggests that this may indeed be a problem.

  2. Extensions and Areas for Future Research

  Like any social-science theory, ours is highly simplified. To focus on mechanisms that we think are important, we abstracted from many details as well as other potentially important mechanisms. This means that there are alternative approaches to some of the basic issues we addressed and also that we excluded other forces that may be important to include for a complete theory of the creation and consolidation of democracy.

  First, our framework concentrated on social conflict as the main driving force that leads to different political institutions. Changes in political institutions occur not because of unanimity but because the side that favors change becomes more powerful and manages - at least temporarily - to impose its preferences. In Chapter 3, we briefly discussed some alternative approaches to democratization: for instance, the ideas of Bates (1991), Rogowski (1998), Herbst (2000), and Tilly (2004) that democracy emerges from the process of state formation, or the ideas of Bueno de Mesquita et al. (2003) and Lizzeri and Persico (2004) that democracy is voluntarily created by political elites because it leads to different equilibrium public policies that makes everyone better off. Although we believe that the major patterns of democratization and democratic consolidation cannot be explained only by these alternative interpretations, these are ultimately potentially complementary approaches, and empirical work must determine the relative importance of different mechanisms. As discussed in Chapter 3, empirical work on the determinants of the creation and consolidation of democracy has not progressed beyond correlations with little attention to the identification of causal relationships or isolating truly exogenous sources of variation. Thus far, there has been no serious attempt to discriminate among different mechanisms leading to democratization (Acemoglu, Johnson, Robinson, and Yared 2004).

  In addition to examining and testing alternative hypotheses, there are several important areas in which more theoretical work appears to be a high priority. Five important areas that we initially planned to discuss were omitted to keep the book length manageable.

  The first is the role of the military. In our baseline model, the only actors are different groups, and we considered that these groups could engage directly in conflict. In reality another institution, the military, plays a crucial role in revolutions, repression, and coups. Implicitly, we assumed that the military did not act as an autonomous actor but instead formed a coalition either with the elites (in the case of repression or coups) and perhaps with the citizens (in the case of a revolution). Nevertheless, there is a widespread claim in political science that the military often intervenes not on behalf of a social group but with its own interests in mind. It is also clear that in developing societies, the military is very powerful relative to other social and economic groups. An important priority for research, therefore, is to develop a theory of military politics to better grasp when the military sides with a particular group and when it may become relatively autonomous from social groups. Although there is a rich case-study literature on the military (Huntington 1964; Finer 1976; Nordlinger 1977; Rouquie 1987; Stepan 1988; Fitch 1998; Loveman 1999), there are as yet few generalizations about the objectives and behavior of the military; only Ticchi and Vindigni (2003b) have tried to use the methodological approach we adopt in this book and the tools of game theory to examine the military.

  The study of the military is related to another major research area in comparative politics. In Chapter 5, we presented a “bare-bones” model of nondemocratic politics; our analysis abstracted from differences in nondemocratic regimes. Yet, much of the political science literature precisely focuses on providing different taxonomies of nondemocratic regimes (e.g., Linz and Stepan 1996). Moreover, much research argues that the type of nondemocratic regime helps to determine the potential for the creation and consolidation of democracy. Whether this is true is ultimately an empirical question, but it is certainly a distinct possibility (Geddes 1999a,b). In this book, we chose to emphasize what we believe is the key distinction between democracy and nondemocratic regimes: the extent of political equality. Nevertheless, introducing richer models of the institutional structure of nondemocracy will undoubtedly generate many new insights.

  The second major area omitted from our analysis but clearly of central importance to understanding the dynamics of democracy is the variations in democratic institutions. A large theoretical and empirical literature emphasizes the differences between different types of democracies: for example, presidential versus parliamentarian and between those that use proportional representation as opposed to majoritarian electoral institutions (e.g., Cox 1997; Lijphart 1999; Persson and Tabellini 2000, 2003; Persson, Roland, and Tabellini 2000). Although in the appendix to Chapter 4 we provide some different microfoundations for the parameter X, the most interesting approach is to relate it to the more detailed structure of political institutions. Throughout the book, we gave examples of how the details of democratic institutions are important for the feasibility and durability of democracy. However, the formal literature is only at the beginning of a research agenda to develop models of how the types of electoral systems or whether a democracy is presidential or parliamentary influence the incentives of politicians or citizens. The choice of the equilibrium form of democratic institutions and how this influences the feasibility of democracy is an exciting area for the years ahead. Our analysis suggests that the detailed institutional structure - because of the way it influences how preferences are aggregated — will be important in determining how political conflicts take place and thus in whether democracy is created or consolidated.

  The third area in which more theoretical work is needed is within the context of what we called alternative political identities. A vast amount of political economy conceptualizes conflict along socioeconomic or class lines, but there is also a widespread understanding that this is not always the case. Although we have tried to show that our main results regarding the circumstances of when democracy arises and consolidates do not depend on the nature of political identities, having a richer model should generate many new empirical predictions. An important area for research is not just the implications of political identities but also their formation and how this depends on the institutional structure. To illustrate, historians of Africa have shown how some important current ethnic identities in Africa that are salient in political conflicts are actually an outcome of incentives created during the colonial period (e.g., Horowitz 1985 on the Ibo in Nigeria and Ranger 1991 on the Shona of Zimbabwe).

  A fourth important area for future research is collective action and revolution. In Chapter 5, we discussed the collective-action problem and argued that the available empirical evidence suggested it is circumvented by revolutionaries providing private benefits to those who take part in revolution. This inspired the model we developed and used throughout the text. Nevertheless, developing a deeper understanding of collective action is a fascinating area for future research, both theoretical and empirical. We also modeled “postrevolution societies” in the crudest way. Our justification is that revolutions (except for a brief discussion in Chapter 6) are off the equilibrium path. However, developing a better understanding of what happens in revolutions and how institutions subsequently evolve is an important topic that may generate new predictions about the creation and consolidation of democracy. As with military politics, there is a rich case-study literature on revolutions that can be the starting point for developing models and mor
e explicitly testable hypotheses.

  Finally, and perhaps most important, the future literature must provide richer models of the workings of economy and the form of economic institutions than presented in this book. A particularly exciting area for future research is the investigation of the interactions between endogenous economic and political institutions. Although in Chapter 9 we endogenized the distribution of income and discussed the important role played by economic institutions, we did not develop explicit models in which economic institutions were determined or changed over time. Moreover, we only examined situations in which income was determined by stocks of assets that were constant. In reality, capital accumulates over time and technology changes. Incorporating these dynamics of growth and accumulation into our framework is an important step (Acemoglu and Robinson 2000a, 2002; Jack and Lagunoff 2003 ). Such extensions will also help to explain why there may be path dependence in political institutions, which many scholars believe to be the case.

  3. The Future of Democracy

  The objective of this book is to develop and present a parsimonious framework to analyze democratic and nondemocratic politics and the transitions between those regimes. Our analysis is mostly aimed at understanding a relatively abstract picture of complex social phenomena. Although any simple framework makes predictions about the future at its own peril, it is useful to reflect on the future of democracy given the framework we developed herein.

  Several issues are important in thinking about whether democracies around the world will be consolidated and how they will transform themselves from what they are today. First, the world is experiencing an increased importance of human capital relative to land and physical capital for two reasons: (1) typical citizens of both developed and developing nations are more educated today than they were fifty years ago; and (2) technology throughout the twentieth century appears to have relied more on the skills and the human capital of the workers (or to have been skill-biased), thus increasing the importance of human capital in the labor market (Acemoglu 2002). Although greater returns to human capital may increase inequality in certain instances (e.g., as in the U.S. economy during the past thirty years), it generally helps to close the gap between the elites and the citizens and creates a large middle class in many less developed nations that are nondemocratic or live in unconsolidated democracies. As this gap closes and a middle class emerges, we expect less distributional conflict and more stable democracies not only in societies where political conflict has been between the rich and the poor but also where political conflict is along other lines. The recent past has witnessed many accounts of the “end of class warfare” (e.g., Fukayama 1992). We are not predicting an end to political conflict anytime soon but rather that with a greater role for human capital, the conflict will be less charged and intense.

  Second, we now live in a highly globalized world economy. For reasons already discussed, we believe that greater international economic and financial links may promote and consolidate democracy. Again, conflict between the elites and the majority of citizens will remain in the global world economy, but globalization may take the most disruptive weapons from both sides’ arsenal in this fight. The citizens do not want to pursue the most populist and redistributive policies, making the elites more secure in democracy. The elites are much more averse to coups and disruptions.

  Third, the end of the Cold War implies that the implicit economic and political support that many nondemocratic regimes received has come to an end, making the transition to democracy easier and coups against democracy more difficult (although there is a danger that the war against terrorism might offset the potential benefits of the end of the Cold War).

  These three factors imply that the future of democracy is bright. Democracy is much more likely to triumph against nondemocracy today than in the past, both in places where it has not arrived and where it has not been consolidated yet.

  Yet, given these developments, do we expect democracy to change its nature in this new era? Our argument has been that democracy is pro-majority, even possibly pro-poor. This is mainly a relative statement, comparing democracy to a typical nondemocratic regime. We also noted that there are reasons why, in democracy, the elites may be powerful even if democracy is generally more pro-majority than nondemocracy. There are two reasons to expect that, in time, the elites may become more powerful in democracy.

  First, the most important sources of extra power for the elites in democracy are their control of the party system and, thus, the political agenda and their ability to form an effective lobby against certain policies. Do we expect the elites to be able to do so more effectively in the future? There are two reasons for suspecting that the answer may be yes. With the increased bright future for democracy, the elites - especially in the current unconsolidated democracies - have to come to terms with living in democracy. In this case, they may as well do their best to influence democratic politics. Therefore, the returns to the elites for increasing their power in democracy may now be greater.

  Perhaps more important, as democracy matures, there may be a greater opportunity for organized groups, which potentially include the elites or certain segments thereof, to become more powerful. The argument that interest groups become stronger over time in democratic societies was first developed by Mancur Olson in his classic 1982 political economy treatise, The Rise and Decline of Nations. Olson pointed out that as time goes by, cooperation and trust form between different members of influential lobbies and, perhaps more important, these lobbies more effectively capture the major branches of the government and the political system. In the context of democratic politics, one of the interest groups that may become stronger and come to dominate much of politics is the elite. If so, we might expect democracies to become less pro-majority in time. The fact that new democracies appear to have been more redistributive than mature democracies throughout the twentieth century and the observation that conservative parties have become stronger in many well-established democracies during the past forty years is consistent with this notion.

  This relates to the Iron Law of Oligarchy formulated by the sociologist Robert Michels in his classic 1911 book, Political Parties. Michels claimed that all organizations, particularly political parties - even socialist ones - tended to be captured by whoever ran them; those people then came to be incorporated into the elites. He argued that this meant democracy had little chance of radically changing society because, at best, it simply replaced one elite with another. In no case would this lead to radical majoritarian social changes. If this law is true, then a natural process of elite capture reduces the radical threat of democracy.

  Second, there is also a different side to the increased importance of human capital (including skill-biased technical change) and greater globalization. By reducing distributional conflict, these economic developments are weakening many of the organizations that have played an important role in supporting the majority and policies favoring the majority. The organizations losing strength include traditional social democratic parties and labor unions. This is most visible in much of the Anglo-Saxon world, especially the United States and the United Kingdom, where labor unions today are much weaker and the traditional left parties have become generally opposed to income redistribution.

  If these changes become more widespread around the world, we may expect the elites and conservative parties to become more powerful and democracy to become less redistributive in the future, especially if new forms of representation for the majority — in both the political sphere and the workplace - do not emerge. Thus, democracy will become more consolidated; however, for those who expect democracy to transform society in the same way as British democracy did in the first half of the twentieth century, it may be a disappointing form of democracy.

  PART SIX.

  APPENDIX

  12 Appendix to Chapter 4: The Distribution of Power in Democracy

  1. Introduction

  In this appendix, we discuss the models that underp
in the analysis of distribution of political power in democracy in the last section of Chapter 4. There we argued that, under some circumstances, the equilibrium policy in democracy could be thought of as maximizing a weighted sum of the indirect utilities of the rich and the poor. We now develop a series of models that can provide microfoundations for those claims and clarify what those “circumstances” are.

  2. Probabilistic Voting Models

  2.1 Probabilistic Voting and Existence of Equilibrium

  Before we discuss the probabilistic voting model, it is useful to revisit the nonexistence of voting equilibria in models without single-peaked preferences. Recall that the MVT applies only when the policy space is single-dimensional and preferences are single-peaked. Although in this book we obtained a lot of mileage from models that satisfy these assumptions, many real-world situations - where there are cross-cutting coalitions and multidimensional differences - do not. In these situations in which the MVT does not apply, the party competition game often does not have an equilibrium in terms of pure strategies. Although in these situations mixed-strategy equilibria exist, it is often unappealing to think of parties mixing over their platforms. The probabilistic voting model first introduced by Lindeck and Weibull (1987) is useful not only as an alternative approach to policy determination but also because it provides a potential way out of the nonexistence problems that arise in the standard model.