The debate on consolidation revolves around what should be added or subtracted from lists like this (for which adjectives should be added to the word “democracy,” see Collier and Levitsky 1997). Clearly, on this basis, many of the regimes that we would consider democratic are not consolidated (see Philip 2003 on Latin America, where there are probably no consolidated democracies in these terms). Although Linz and Stepan’s initial definition is consistent with our approach, the subsequent conditions they impose are not.
Our use of the word consolidation instead builds on our Schumpetarian definition of democracy. As we argued before, this seems the natural place to start in building a theory of democracy, and this view echoes that of Schedler (1998) that:
The term “democratic consolidation” should refer to expectations of regime continuity - and nothing else. Accordingly, the concept of a “consolidated democracy” should describe a democratic regime that relevant observers expect to last well into the future - and nothing else. (p. 103)
2. Incentives for Coups
We now consider a society in which democracy has been created and the preferences of the median voter determine the tax rate. We continue to use our two-group model and associate the elites with the rich and the citizens with the poor. The median voter in democracy is, therefore, a poor agent. In contrast to our previous analysis, however, we now consider the possibility that democracy may not last forever and, in fact, there may be a coup against democracy. Because of the pro-citizen policies - for example, income redistribution implied by democratic politics - in democracy, the citizens are relatively well off and the elites are worse off. This reasoning suggests that the greatest threat against democracy comes from the elites. Therefore, we model coups by focusing on the incentives of the elites to reduce redistribution by moving away from democracy to nondemocracy.
Many coups, especially in Latin America, had reducing redistribution as one of their major objectives and, in most cases, proceeded to reduce redistribution and change the income distribution significantly (see the evidence discussed in Chapter 3). Given that coups are generally undertaken by the military, our approach presumes that for various reasons, the military represents the interests of the elites more than those of the citizens. We believe this is a reasonable first pass; nevertheless, in practice, the objectives of the military are not always perfectly aligned with those of a single group and may have an important impact on the survival of democracy. Incorporating the role of the military in democratic consolidation into formal models of politics is a major area for future research, and we return to this topic briefly in the conclusion of the book.
In this chapter, we simply take as given the possibility that, at some cost, the elites can control the military and mount a coup against democracy, and we investigate the circumstances under which they would like to do so. From a modeling point of view, the interesting observation is that there is a parallel between the reasons of the citizens to want democracy and the reasons of the elites to want nondemocracy. Recall that the citizens demand a credible commitment to future pro-majority policies, and, therefore, a transition to democracy (and the elites were forced to give it to them) because they care about polices and social choices in the future as well as today and they only have temporary de facto political power. Similar reasoning applies in the case of transitions from democracy to nondemocracy. The elites want less pro-citizen policies, and they temporarily have political power to secure them. However, they care about future policies as well, and they know that once their temporary de facto power goes away, democracy will reintroduce the policies that it favors, such as higher taxes and income redistribution. Therefore, the way for the elites to secure the policies they prefer in the future as well as today is to change political institutions toward those that give them more de jure power - that is, a move from democracy toward nondemocracy.
There is much evidence that democrats would like to make concessions to the elites and the military to avoid coups, but the effectiveness of these is undermined by their lack of credibility. Nordlinger (1977) notes:
... the military have intervened despite budgetary increases designed to stave off a coup, as in the 1973 coup against President Allende of Chile. Allende was overthrown despite military salary increases which were greater than those for equivalent civilian grades, better fringe benefits, and the purchase of additional equipment.18 (p. 71 )
There is one difference between the way we are modeling the transition from nondemocracy to democracy and the transition to nondemocracy: in the first case, the citizens had the option to undertake a revolution, and the elites created a democracy to prevent it. Here, the elites actually use their political power to mount a coup and change the system. This may appear like an asymmetry, but it is not essential to our results. We adopt this particular way of modeling transitions to and from democracy because we believe it provides a good approximation to reality: in most instances, democracy resulted from the elites democratizing, whereas the move from democracy to dictatorship is almost never consensual.
3. A Static Model of Coups
To model coups against democracy, consider the basic two-class model of Chapter 4, augmented to consider the possibility that the elites can mount a costly coup. We make identical assumptions about the agents and their incomes but now allow for costs due to coups. In particular, we have:
(7.1)
where we use the convention that ζ = 0 denotes no coup and ζ = 1 denotes a coup. The notation(S) is the cost due to coup in state S. We model the costs of coups in exactly the same way as we modeled the costs of revolution and repression - a fraction of income gets destroyed. As in the static model in the previous chapter, we simply focus on the state where the coup is a threat and, hence, we suppress the notation for S. There are no costs if there is no coup; thus, if ζ = 0, then= 1. The relevant cost, therefore, is the value ofwhen ζ = 1, which we denote by 1 - ϕ where 0 < ϕ < 1.
Figure 7.1 shows the game we use to analyze coups. Initially, because we are in a democracy, the median voter sets a tax rate, τD. If there is no threat of a coup from the elites, the citizens set their most preferred tax rate, τp, as given by (4.11 ). This results in payoffs Vp(D) and Vr(D), given by (6.4). Whether the elites mount a coup depends on the continuation value in democracy and nondemocracy. We allow the tax rate initially chosen by the citizens to be different from τp because of the threat of a coup. After this, the elites decide whether to undertake the coup. If they do, the society switches to nondemocracy and the elites set the tax rate. Naturally, they choose their most preferred tax rate, τN = τr. As a result, the game ends with respective payoffs for the citizens and the elites19:
(7.2)
Figure 7.1. The Coup Game.
Alternatively, if the elites decide not to undertake a coup, the political system remains democratic. In this case, nature moves one more time and determines whether the median voter - the politically decisive agent in democracy - gets to reset the tax rate from that promised by the citizens in the previous stage. As in our simple model of democratization, this captures the notion that we model in greater detail in the next section: a regime (even a democratic regime) cannot credibly commit to future taxes. More specifically, nature determines with probability p that the tax promised, denoted, remains and the citizens and the elites receive values V (yP |τD) =) and V (yr |τD =) where, as usual:
If, on the other hand, nature allows democracy to reset the tax, the median voter chooses a new tax rate, denoted byD, leading to the values Vp(D) and Vr(D). Therefore, the values resulting from a promise of less redistribution, only at the rate, by the citizens in democracy are Vp(D, τD= ) and Vr(D, τD =), such that:
(7.3)
These expressions take into account that with probability 1 - p, the citizens get to reset the tax, in which case they are unconstrained and choose their most preferred tax rate,D = τp.
We can now characterize the subgame perfect equilibrium of this game by backward induction. Essentially, the
game has the same structure as our static democratization game in Chapter 6. The crucial issues are whether undertaking a coup is in the interest of the elites and whether the citizens can prevent a coup by promising concessions (in this case, to redistribute less toward themselves). The strategies are σr = {ζ(·), τN} and σp = {τD, D}. The actions of the citizens, who play first, consist initially of a tax rate τD ∈ [0, 1]; also, if there is no coup and nature allows the tax rate to be reset, where we again use the notation ν = 1, another tax rateD ∈ [0, 1]. Here, the superscript D again indicates democracy. The actions of the elites are a coup decision ζ : [0, 1] → {0, 1}, where ζ(τD) is the coup choice when the median voter sets the tax rate τD ∈ [0, 1], and if ζ = 1, a decision about what tax rate to set, which we denote τN ∈ [0, 1]. Then, a subgame perfect equilibrium is a strategy combination {r, p}, such thatp andr are best responses to each other in all proper subgames.
Whether a coup is attractive for the elites given the status quo depends on whether the coup constraint, Vr(C, ϕ) > Vr(D), binds. This states that a coup is more attractive than living under an unconstrained democracy. This coup constraint can be expressed as:
or
(7.4)
When this constraint does not bind, democracy is not redistributive enough, or coups are sufficiently costly that the elites never find a coup profitable. In this case, we refer to democracy as fully consolidated: there is never any effective threat against the stability of democracy. From (7.4), we can derive a critical level of the cost of a coup, denoted, such that if ϕ ≥democracy is fully consolidated. This satisfies:
(7.5)
In contrast, when this constraint binds, democracy is not fully consolidated: if the citizens do not take an action, there will be a coup along the equilibrium path. The action that the citizens can take is to reduce the tax rate. The problem, however, is that they cannot perfectly commit to doing so because of the possibility of resetting the tax once the coup threat has subsided. Considering this possibility, the value to the elites of the citizens setting a tax rate ofis). This strategy of promising less distribution prevents the coup only if this value is greater than the return to the elites following a coup; that is, Vr (D, τ D =) ≥ Vr (C, ϕ). In other words, only if:
We can now define a threshold value for the cost of a coup, ϕ*, such that when ϕ < ϕ*, the promise of limited distribution by the citizens is not sufficient to dissuade the elites from a coup. Of course, the most attractive promise that the citizens can make to the elites is to stop redistribution away from them entirely - that is,= 0 - therefore, we must have that at ϕ*, Vr (D, τ D = 0) = Vr (C, ϕ*). Solving this equality gives the threshold value ϕ* as:
(7.6)
Given this discussion, we can summarize the subgame perfect equilibrium of this game as follows:
Proposition 7.1: There is a unique subgame perfect equilibrium {r,p} in the game described in Figure 7.1. Let and ϕ* be defined by (7.5) and (7.6). Then, in this unique equilibrium, we have:
• If ϕ ≥ , then democracy is fully consolidated and the citizens set their preferred tax rate τp > 0 as given by (4.11).
• If ϕ ∈ [ϕ*,), then democracy is semiconsolidated. The citizens set a tax rate τD =where ≤ τpsuch that Vr (D, τD= ) = Vr(C, ϕ).
• If ϕ < ϕ*, then democracy is unconsolidated. There is a coup and the elites come to power and set their most preferred tax rate, τN= τr.
The analysis shows how equilibrium coups can happen as a way for the elites to limit redistribution in the future. Notably, coups happen (when ϕ < ϕ* ) precisely because democracy has a limited potential to commit to low redistribution in the future. Then, the elites use their current (and temporary) political power to change political institutions so as to reduce future redistribution. The parallel to the discussion of democratization is obvious: again, equilibrium changes in political institutions happen as a way of regulating the future allocation of political power. There is also a parallel between repression and coups - both use force to avoid democracy, but they do so starting in different political states. This is why the comparative statics of coups are similar to those for repression.
The distinction between fully and semiconsolidated democracies is useful. Democracy is fully consolidated when the coup threat is never present, democracy is not really challenged, and the citizens can set their most preferred (unconstrained) tax rate, τP. A semiconsolidated democracy, on the other hand, would fall prey to a coup if it set the tax rate τp. It can only survive by making concessions to the elites to dissuade them from mounting a coup. Empirically, this notion of semiconsolidated democracy may help us explain some otherwise puzzling behavior: Wantchekon (1999), for example, argues that in El Salvador the parties representing the majority of citizens tried in the 1990s to reduce the amount of redistribution they offered in elections for fear of inducing a coup.
It is interesting to contrast our analysis with the claim of Przeworski (1991) that consolidated democracy necessitates that all groups, even the previous elites, have a sufficiently large chance of being in power. As Przeworski (1991) put it:
... compliance depends on the probability of winning within the democratic institutions. A particular actor... will comply if the probability it attaches to being victorious in democratic competition... is greater than some minimum.... Democracy will evoke generalized compliance when all the relevant political forces have some specific minimum probability of doing well under the particular system of institutions. (pp. 30-1)
According to this argument, for democracy to be stable, all groups must have a sufficient chance of wielding power. If any group is completely excluded, they will be tempted to fight for power. This idea is widely accepted by political scientists (e.g., Weingast 1997). Colomer (2000, p. 10) reiterated this view when he wrote that “the establishment of democracy appears as a conventional agreement on new rules of the political game. Agreement is possible because democracy gives different actors reasonable expectations to gain or share power in some undetermined future.”
In contrast, in our model of democracy, the elites can never win power because policies always cater to the preference of the median voter. However, this does not mean that the elites cannot get what they want in a democracy because even when they have no de jure power, they may have de facto power. For example, in a situation in which democracy is semiconsolidated, the policies of the citizens cater to the elites despite the fact that the elites do not form the government. Indeed, this is ironic because, according to Przeworski, if the elites cannot form a government, they will try to mount a coup and, hence, democracy is not consolidated. Yet, if they can overthrow the system by force, then they must have effective de facto power; this is exactly the situation in which they will be able to get what they want from the government without having to overthrow it. When the elites do not have de facto power, they do not get what they want from democracy but neither are they able to mount a coup. Przeworski’s claim, therefore, is false in our model.
We now consider the comparative statics of coups with respect to inequality. First, we can implicitly define a critical threshold for inequality,:
such that when θ ≤, the coup constraint, (7.4), will not bind. In other words, this is the threshold level of inequality, such that when the inequality is less than this level, democracy is fully consolidated.
Next, using the definition of ϕ*, we can determinesuch that:
Democracy is semiconsolidated when θ ≤. Moreover, it is straightforward to check that> . This discussion leads to the following corollary:
Corollary 7.1: Consider a society with a fixed ϕ and p, and inequality given by θ. Then, there exist and < such that
• When θ ≤ , democracy is fully consolidated and the equilibrium tax rate is always τp.
• When θ E (, ], democracy is semiconsolidated. It sets the tax rate so as to prevent a coup in this case.
• When θ > , democracy is unconsolidated. There is a coup and the elites come to power and set the tax
rate τ N= τr.
This analysis shows that coups tend to happen in more unequal societies. In less but still fairly unequal societies, democracy is semiconsolidated and survives only by making concessions to the elites in the form of lower taxes. The intuition for why inequality matters for coups is straightforward: coups happen in this model as a way for the elites to reduce future redistribution. Democracy is more redistributive when there is more inequality and, hence, more costly for the elites. Coups, therefore, become more attractive for them in an unequal society.
These comparative statics are consistent with the evidence in Chapter 3, which discussed a cross-country relationship between measures of inequality and democracy with more democratic societies tending to have lower inequality. In the previous chapter, we suggested that this might be because in more equal societies, repression was less attractive; thus, elites were more likely to create democracy. Now we can see that once democracy is created in a more egalitarian society, it is more likely to consolidate.
Economic Origins of Dictatorship and Democracy Page 33
