Economic Origins of Dictatorship and Democracy

Home > Other > Economic Origins of Dictatorship and Democracy > Page 18
Economic Origins of Dictatorship and Democracy Page 18

by Daron Acemoglu


  Second, equilibrium policy may be influenced not only by voting but also by campaign contributions and the activities of lobbies and special interests. In such a situation, groups that are represented by an organized special interest or who have more resources to channel through special interests tend to have more influence over policy than groups with less organization and resources. If the rich have an advantage in either of these dimensions, this allows their preferences to influence democratic policies. A model along these lines was developed initially by Becker (1983), which was greatly developed and extended by Grossman and Helpman (1996,2001).

  Third, so far, political parties have in a sense been perfect agents of the voters. In reality, however, political parties have objectives that are to some extent autonomous from those of citizens, and the policies they offer reflect them, not simply the wishes of the median voter. This is particularly true when, as first emphasized by Wittman (1983), there is uncertainty in the outcome of elections or, as shown by Alesina (1988), parties cannot commit to arbitrary policy platforms. When either of these is true, political parties’ objectives, not simply the preferences of the voters, are important in influencing political outcomes. In this case, groups that can capture the agendas of political parties can influence democratic policy to a greater extent than their numbers would indicate.

  Finally, and probably most interesting the Downsian model and many of its extensions, including models of probabilistic voting, feature a thin description of political institutions. The Downsian model introduced in this chapter is almost like a presidential election (although not in the United States because then we would have to introduce the electoral college). For example, we did not distinguish between electoral districts. If we wanted to use the model to capture the outcome of elections for the British Parliament, we would have to introduce such districts and model how the disaggregated vote share mapped into seat shares in Parliament. This may be significant because, as pointed out by Edgeworth in the nineteenth century and formalized by Kendall and Stuart (1950), small parties tend to be underrepresented in such majoritarian institutions. Thus, there is not a one-to-one relationship between vote share in aggregate and seat share in Parliament. Many other aspects of institutions might matter. For example, institutions influence voter turnout and also the abilities of minority groups to get what they want in legislatures.

  This is interesting because the institutions matter for who has power in a democracy. Consider one specific example, motivated by the attention it has received in the political science literature: the difference between presidential and parliamentary democracy. As noted previously, Linz (1978, 1994) argued that presidential regimes tend to be more prone to coups; Przeworski et al. (2000) present econometric evidence consistent with this claim. The intuitive idea is that presidents, because they are elected in a popular vote, tend to represent the preferences of the median voter in society. On the other hand, Parliament may have to reconcile more diverse interests. In this case, if we compared the same country under these two different sets of institutions, we would expect the outcome with a president to be closer to that preferred by the citizens.

  Motivated by these considerations, we use a simple reduced-form model parameterizing the political power of different groups in democracy. In the appendix to this chapter, we formally develop the first three of these ideas on modeling the distribution of political power in democracy and show how they map into the simple reduced-form model used here. Different specific models - whether they emphasize different institutional details, lobbying, relatively autonomous political parties, or the presence of swing voters - provide alternative microfoundations for our reduced form. Naturally, these details are also interesting and may be significant in specific cases; we discuss this as we proceed.

  Let us now return to our basic two-class model with a unique policy instrument, the tax rate on income, τ. Given that the citizens are the majority (i.e., 1 - δ > 1/2), Downsian political competition simply maximized the indirect utility of the citizens, Vp(τ). In this model, the preferences of the elite are irrelevant for determining the tax rate. More generally, however, the elite will have some power and the equilibrium policy will reflect this. The simplest way of capturing this idea is to think of the equilibrium policy as maximizing a weighted sum of the indirect utilities of the elites and the citizens, where the weights determine how much the equilibrium policy reflects the preferences of the different groups. We call the weight of a group the “political power” of that group. Let those weights be x and 1 - X for the elites and the citizens, respectively. Then, the equilibrium tax rate would be that which maximizes:

  which has a first-order condition (with complementary slackness).

  This yields:

  (4.16)

  where we define τ(x) to be the equilibrium tax rate when the political power parameter is X.

  It is instructive to compare Equations (4.16) and (4.11), which determined equilibrium policy in the two-class model with Downsian political competition. It is clear that the Downsian outcome is a special case of the current model for X= 0, in which case (4.16) becomes identical to (4.11) so that τ(X = 0) = τp. However, for all values of X > 0, the preferences of the elite also matter for equilibrium policies so that τ(X > 0) < τp. Moreover, the greater is X, the more political power the elites have despite the fact that they are the minority. To see the implications of this, notice that if X rises, then the left side of (4.16) increases. This implies that the right side must increase also so that C‘(τ) must fall. Because C’( τ) is increasing in τ, this implies that τ falls. In other words, dτ (X) /dX < 0. Thus, an increase in the power of the rich, or in their ability to influence the equilibrium policy in democracy through whatever channel, pulls the tax rate down and closer to their ideal point. The different models in the appendix provide different mechanisms by which the power of the elites is exerted and how the equilibrium tax rate responds as a result.

  This is important because, so far, we have emphasized that democracies generate more pro-citizen policies than nondemocracies. If, in fact, we have that as X1 and the tax rate generated by democratic politics tends to that most preferred by the elites, there will be little difference between democracies and nondemocracies. Our perspective is that there are often reasons for the elites to be powerful in democracies even when they are a minority, so X > 0 may be a good approximation of reality. Nevertheless, both the evidence discussed so far and introspection suggest that most democratic societies are far from the case where χ = 1. As a result, democracies do not simply cater to the preferences of the rich the same way as would a typical nondemocracy.

  6. Conclusion

  In this chapter, we developed some basic models of democratic politics. We also discussed in detail the workhorse models and some of their properties that we use to characterize democracy in the remainder of the book. Our analysis focuses on the two-group model in conditions where either the MVT applies or where, when the policy space is multidimensional, the equilibrium policy is that preferred by the poor. We focus, therefore, on situations in which the median voter is a poor agent and his preferences determine what happens in a democracy. We also consider extensively three substantive extensions of this model. First, a three-class model in which the middle class enters as a separate group from the rich and the poor. We defer a formal introduction of this model until the first time it is used in Chapter 8. Second, the reduced-form model of democracy in which different groups “power” can vary depending on the nature of democratic institutions, on whether they are swing voters, whether they are an organized lobby, and so forth. In the appendix to this chapter, we discuss in detail different microfoundations for the power parameter χ but, for the rest of the book, we simply work with this reduced form rather than present detailed models in which institutions, lobbying, party capture, or probabilistic voting are explicitly introduced. Finally, the simple model in which political identities differ and can be different from those based purely on socioec
onomic class or income level, and we analyze how this affects distributional conflict in society.

  5

  Nondemocratic Politics

  1. Introduction

  In this chapter, we discuss various issues that arise in thinking about policy determination in nondemocracy. For our purposes, the most basic distinction between a democracy and a nondemocracy is that the former is a situation of political equality: each citizen has one vote. As a result, in democracy, the preferences of all citizens matter in the determination of the political outcomes. In nondemocracy, this is not the case because only a subset of the people, an elite, has political rights. In principle, this could be any subset. Soviet socialism claimed to be the dictatorship of the proletariat and did not even consider “dictatorship” a word with pejorative connotations. Similarly, the dictatorship of Juvenal Habyarimana in Rwanda between 1973 and 1994 might be considered the dictatorship of a particular ethnic group, the Hutu. In Brazil between 1964 and 1985, there was a military dictatorship, with bureaucratic authoritarian and corporatist tendencies; this regime emphasized industrialization while also protecting the economic interests of the relatively rich and avoiding any radical - particularly agrarian - reforms. In contrast, the dictatorship of Mobutu Sese Seko in The Congo between 1965 and 1997 was a highly personalistic, kleptocratic regime, in which the main use of state power was to enrich Mobutu and his entourage. Despite these differences among nondemocracies (see Linz and Stepan 1996 for an influential taxonomy), our purpose is to emphasize the major difference between democracies and nondemocracies that we see as the extent of political equality.

  In general, there are two features that shape economic policies in nondemocracies : first, the preferences of the group in power; and second, the constraints faced by that group. Everything else being equal, the group in power, the elite, choose policies that maximize its utilities. However, elites typically live in fear of being replaced by different social groups or by other individuals within the same group. Therefore, an important issue in nondemocracies is to ensure that no group is unhappy enough to attempt to overthrow the regime or take other political or economic actions detrimental to the utility of the group in power.

  Our analysis in this chapter builds on the model of democracy described in Chapter 4. Thus, we think of a society composed of heterogeneous individuals. Nondemocracy is the rule of some subset of this society. In Chapter 4, we showed that democracy is the rule of the more numerous group, either the poor or, if political identities are along other lines, group X. Here, we think of nondemocracy as the rule of the less numerous group (either the rich or group Z).

  To start, we focus on models in which nondemocracy is simply the rule of the rich. In many circumstances, this is a perfectly sensible postulate. For example, formal restrictions on suffrage have typically been on the poor: those with no assets, who have low income, or who are illiterate. There have also been racial restrictions on voting - for example, in the United States before the Civil War and in South Africa before the collapse of the apartheid system - but again, the racial groups disenfranchised have always been poor. Even apparently autonomous military regimes often serve the interests of the affluent, an obvious case being the Pinochet dictatorship in Chile between 1973 and 1990. Many other Latin American dictatorships in the twentieth century were induced by the threat of radical redistributive and social policies, for example, those in Argentina after 1930, after 1955, and again between 1976 and 1983. Other examples of coups against democracy aimed at avoiding radical policies are those in Venezuela in 1948, in Guatemala in 1954, and in Brazil in 1964. Although the regimes that took power after these coups were military, the coups themselves were induced by the threat of radical redistributive policies. Drake argues in this context that

  in many ways, the anti-labor stance of these despotisms defined their raison d’être... it motivated their seizure of power, legitimized their existence, marshalled their supporters and opponents, underlay their model of economic growth, drove their social policies, and propelled their political practices. That conflict with workers also substantially affected their tenure and terminations. (1996, p. 2)

  To us, these examples suggest that there is often a close association between what nondemocratic regimes do and what the rich want. Nevertheless, as discussed in Chapter 2, our basic framework and many of the empirical results carry over to a situation in which nondemocracy is not simply the rule of the rich.

  The models we develop in this chapter emphasize the interplay between elites’ preferences and the constraints placed on them by the preferences of other groups-centrally the disenfranchised citizens - in society. Our aim is again to search for general tendencies that hold true across different types of nondemocratic regimes and to contrast those with the tendencies in a typical democracy. The dichotomous distinction we draw between democracy and nondemocracy, our desire to bring out the common elements within each regime, and our relentless reliance on Occam’s razor may appear stark, even simplistic. Nevertheless, we believe that this is the correct way to make progress, and our conviction is that this dichotomy is useful for developing intuitive ideas about the forces that lead societies to have different political institutions.

  2. Power and Constraints in Nondemocratic Politics

  2.1 The Elites in Democratic Politics

  We discussed in the previous chapter how the elites may be more powerful in democracy than their mere numbers suggest. We saw that a general model that allows the elites to have some power in democracy implies that the equilibrium tax rate is τ( x ), where χ can be thought of as a measure of the power of the elites in democratic politics. In particular, consider the limit ofχ1 in (4.16); in this case, the equilibrium policy will always be the policy preferred by the elites, which is τr.

  We think of nondemocratic politics as similar to this limit. Because the society is not a well-functioning democracy, the wishes of the majority of the population are ignored, and policies are chosen to maximize the welfare of the elites. This discussion also highlights that, despite our dichotomous distinction between democracy and nondemocracy, we can think of a more continuous distinction between the two. A society is more democratic when the wishes of the majority are incorporated substantially into major policy choices. This corresponds to a situation in which χ is close to 0 in this model. Nondemocracy, on the other hand, is a situation in which the wishes of the majority are ignored in favor of the desires of a subpopulation, the elites. Here, this corresponds to the equilibrium in which χ1.

  2.2 The Revolution Constraint

  The previous discussion highlights how we can think of nondemocracy as a situation that maximizes the utility of the elites. However, nondemocracy, especially compared to the ideal of democracy, is neither egalitarian nor fair. Therefore, the citizens would have a constant desire to change the outcome, the policies, and the regime. What prevents them is the fact that the elites control the political institutions and military power in nondemocratic societies. Because of this control, they can maximize their utility but, given that they are the minority and would like to pursue policies not in line with the interest of the majority, there can also be certain constraints on the policies they want to pursue.

  The major constraint that faces those controlling political power in nondemocracy is a danger that those excluded from political power might attempt to gain political power or to overthrow those who are in control.

  In terms of the discussion in Chapter 2, recall that we distinguish between de jure political power and de facto political power. De jure political power is that which comes from political institutions. In contrast, de facto political power comes from the ability of one group to overwhelm the other, by fighting in a battle or through other means. In democracy, de jure political power rests with the citizens. In nondemocracy, the citizens have no de jure political power; they are excluded from the political system. Nevertheless, they may have de facto political power by virtue of the fact that they are the majority, and the
y may be able to coordinate their actions to overthrow the existing regime. In the extreme, the citizens can undertake a revolution against a nondemocracy to change the political system to one that is more beneficial for them. We summarize the constraints placed on the elites by this type of de facto political power of the citizens by a revolution constraint.

  In this subsection, we discuss the origins of the revolution constraint and the restrictions it places on the actions of the elites controlling the political system in nondemocracy. As a starting point, we discuss ways of formalizing revolutions and introduce concepts related to the collective-action problems that might arise in organizing the citizens so that they can exert de facto power. Throughout, we focus on the two-class model introduced in the last chapter to make the discussion more concrete. In this model, society is divided into rich elites and poor citizens, who are more numerous.

  First, think about what happens after a revolution. By definition, a revolution in this environment corresponds to the citizens using their sheer numbers to overwhelm the elites in nondemocracy, taking control of the society and its wealth and income-generating assets. Hence, in some way, we are thinking of revolution leading to a postrevolutionary society in which the control passes from the elites to the citizens.

 

‹ Prev