A key feature of our theoretical framework is that collective action is intrinsically transitory. Even with the use of ideology or incentives, solving the collective-action problem is difficult to begin with and very hard to sustain. The empirical literature also emphasizes that the difficulty of solving the collective action problem leads collective action to typically be transitory. Lichbach (1995, p. 17) notes “collective action, if undertaken on a short-term basis, may indeed occur; collective action that requires long periods of time does not.... Given that most people’s commitments to particular causes face inevitable decline, most dissident groups are ephemeral, most dissident campaigns brief.” This transitory nature of collective action is echoed by Tarrow (1991, p. 15), who notes “the exhaustion of mass political involvement,” and Ross and Gurr (1989, p. 414) discuss political “burnout.” Similarly, Hardin (1995) argues that
... the extensive political participation of civil society receives enthusiastic expression only in moments of state collapse or great crisis. It cannot be maintained at a perpetually high level. (p. 18)
3. Modeling Preferences and Constraints in Nondemocracies
Let us now put the collective-action problem aside and start investigating the implications of the revolution constraint (5.4) binding on nondemocratic politics. To do so, consider the following game depicted in Figure 5.1. In writing about this game and others in the remainder of the book, we treat the elite and the poor as single players. In general, to specify what an equilibrium is in such a game, we would have to describe the payoff functions and strategies for all the elites and all the citizens. A Nash equilibrium would then entail a specification of strategies, one for each player, such that no member of the elite and no citizen could increase their payoff by changing their strategy. Nevertheless, this level of generality is redundant. All members of the elite are the same, as are all citizens. Moreover, as discussed previously, we assume that both groups have solved their collective-action problems. This justifies us in treating both groups collectively and talking about “the elite” and “the citizens” and examining an equilibrium stemming from interactions between these two groups. Nevertheless, in specifying payoffs, we do so at the individual level because even when the collective-action problem has been solved, behavior has to be individually rational.
In Figure 5.1, the elite move first and set the tax rate, τN. We use the notationto refer to a specific value of τN set to avoid a revolution. After observing this tax rate, the citizens decide whether to undertake a revolution. If they do not, the game ends with payoffs:
(5.8)
where= ( — C()). These payoffs follow from redistribution in nondemocracy at the tax rate The second equality in these equations rearranges the expression for V ( yiτN =) in a way particularly instructive for the remainder of the book. In particular,(— yi) — C() is the net amount of redistribution for i = p, r so that(- yP) - () > 0 while( - yr) - C()< 0; that is, the elite loses from income redistribution.
Alternatively, the citizens might choose to attempt a revolution, in which case we assume that the revolution always succeeds and they receive the payoffs:
where the payoff to the citizens comes from the way we specified the revolution technology, and the elite receive nothing because all income is expropriated from them. What matters is not that the elite receives nothing but simply that what they receive is sufficiently low that they want to avoid revolution.
How do we solve a game like this? The answer is “backward induction,” starting at the end of the game tree. This technique, which we appealed to in Chapter 4, is useful because it characterizes the subgame perfect Nash equilibria of the game. Subgame perfection is a refinement of the original Nash equilibrium concept, useful in games with sequential moves and in dynamic games. The key feature of such an equilibrium, noted originally by Selten (1975), is that it rules out Nash equilibria supported by noncredible threats “off the equilibrium path.” By “off the equilibrium path,” we mean that the equilibrium strategies are such that the threat will not be carried out - it remains just a threat. A noncredible threat is a threat that the player making it would not find optimal to actually undertake if called upon to do so.
To consider an extreme example, imagine that the citizens demand all the money of the elite or they will blow up the world, including themselves. Faced with this threat, it is optimal for the elite to give the citizens all of their money. This is one Nash equilibrium. However, it rests on the threat that if the elite refuses, the citizens will blow up the world. This threat is off the equilibrium path because the elite hand over their money and the citizens, therefore, do not have to carry out their threat. Imagine, however, that the elite refuses. Now, the citizens must decide whether to blow up the world. Faced with this situation, the citizens renege on their threat because, plausibly, it is better to get nothing from the elite than to kill themselves. Therefore, their threat is not credible, and the Nash equilibrium supported by this noncredible threat is not appealing. Fortunately, there is another more plausible Nash equilibrium in which the elite refuses to give the citizens anything and the citizens do not blow up the world. This second Nash equilibrium is indeed subgame perfect, whereas the first is not because it rests on noncredible threats. Given the importance in this book of the credibility of threats and promises, we make heavy use of the restriction that equilibria be subgame perfect.
We need to distinguish two cases. In the first, the revolution constraint (5.4) does not bind. This implies that even if the elite sets the tax rate most preferable for themselves, τN = τr, undertaking the revolution is not in the interests of the citizens. Then, in the subgame perfect equilibrium of the game, the elite anticipates that the revolution will never occur and, therefore, set their most preferred tax rateτN=τr =0.
The more interesting case for our exposition is the one in which (5.4) binds. Now, if the elite were to set τNτN = τrit would be in the interest of the citizens to undertake a revolution. Anticipating this, the elite would try to make a concession-for example, change policy closer to that preferred by the citizens. In this context, this implies that they will set a tax rate sufficient to prevent the revolution. The first question to ask is, therefore, whether such a tax rate exists. The best tax rate from the point of view of the citizens is τN = τp, as given by (4.1 I ) - after all, τ P is the tax rate that the citizens would have set themselves, so the elite can never do better than setting this tax rate in trying to maximize the utility of the citizens. Thus, the question is whether:
holds or, using the definitions in (4.7), whether:
(5.9)
holds. We use a weak inequality because, as noted previously, we assume that if the citizens are indifferent between the status quo and revolution, then they do not revolt.
If (5.9) does not hold, then even the best tax rate for the citizens is not enough to prevent revolution. This might be because the citizens are well organized and have managed to fully solve the collective-action problem or because they can use the economy’s productive resources quite productively after a revolution. Both of these scenarios translate into a low value of it. Alternatively, (5.9) may fail to hold because taxation is costly, so even the best tax rate for the citizens is not sufficiently redistributive. In this case, the unique equilibrium involves the citizens undertaking a revolution.
The other case, which is arguably more interesting from the point of view of our analysis, is when (5.9) holds. In this case, a unique tax rateexists such that V(yPτN= τ) = Vr(R,µ) given by:
(5.10)
It follows from (5.9) that this tax rate is such that. Therefore, in this case, the unique equilibrium involves the elite setting the tax i to prevent revolution.
The interesting feature of this simple game is that, despite the fact that the elite has complete control of formal political power in nondemocracy, they may have to deviate from their most preferred tax rate, τr, because there are other sources of political power in nondemocracy constraining their actions - i
n our formulation captured by the revolution constraint. This type of political power is de facto; the citizens are excluded from the political system, but they can pose an effective challenge from the outside. Fearing a revolution coming from this de facto political power of the citizens, the elite makes concessions and sets a tax rate that redistributes some of their resources toward the citizens.
Before stating the main result, we need to introduce a more formal definition of strategies. Let σr = τN) be the actions taken by the elite, which consists of a tax rate τN[0, 1], in which the superscript N refers to nondemocracy. Similarly, σp= {p(.)} are the actions of the citizens that consist of a decision to initiate a revolution, p(τN) (ρ = 1 representing a revolution) where this decision is conditioned on the current actions of the elite who move before the citizens in the game according to the timing of events depicted in Figure 5.1. Hence, p is a function, p : [0, 1]{0, 1 }. Then, a subgame perfect equilibrium is a strategy combination, {such that Q and Q’ are best responses to each other in all proper subgames. We always use the tildes to represent a particular equilibrium.
Various strategy profiles can be in equilibrium, depending on the parameters. Nevertheless, for any specification of parameters, the equilibrium is unique. When θ µ, the revolution constraint does not bind and the following strategies constitute an equilibrium: τN = 0 and p(τN)= 0 for all τN. According to these strategies, the elite sets the tax rate at zero and the citizens never revolt, whatever the tax rate is. Here, it does not matter what the elite does (i.e., p= 0 irrespective of τN) because the poor have a dominant strategy. Note the important property that strategies must specify behavior both on and off the equilibrium path. Even though the elite’s strategy stipulates a zero tax rate, the citizens’ strategy specifies what action to take for all tax rates, not just zero.
When θ > µand (5.9) does not hold, then the following strategy profile is the unique equilibrium: p(τN) = 1 for all τN. In this case, even setting the tax rate τ will not stop a revolution so, whatever the elite does, the citizens revolt. The citizens again have a dominant strategy, this time to revolt irrespective of τ N.
Finally and most interesting, when θ > µand (5.9) does hold, the following strategy profile is the unique equilibrium: τN =and p(τN) = 0, for all τN ; also off the equilibrium path, p(τN) = 1 for all τN <. Here, revolution is attractive if the elite makes no concessions, but because (5.9) holds, the citizens can be dissuaded from revolution by concessions, specifically by setting the tax ratesuch that (5.10) holds. Note again the specification of behavior off the equilibrium path. The elite set the tax rateand the citizens do not revolt if offered a tax rate τN >. Nevertheless, the strategy of the citizen says that if offered a tax rate τ N< , they will revolt. It is this “threat” off the equilibrium path that induces the elite to give redistribution. This threat is credible because if the elite deviated and tried to get away with less redistribution, it would be optimal for the citizens to undertake revolution. The concept of subgame perfect Nash equilibrium explicitly imposes that such threats have to be credible.
Summarizing this analysis, we have the following:
Proposition 5.1: There is a unique subgame perfect equilibrium {, }in the game described in Figure 5.1, and it is such that:
• If (5.4) does not bind, then τ N = 0 and p(τ N) = 0 for all N.
• · If (5.4) binds and (5.9) does not hold, then p(τN= 1 for all τ N.
• If (5.4) binds and (5.9) does hold, then τN= where is given by (5.10), and p(τN) = 0, for all r N i, and P(τN) = for all τN<
This discussion and Proposition 5.1, therefore, highlight how in nondemocracy equilibrium policies are determined by a combination of the preferences of the elite and the constraints that they face. When these constraints are absent or very loose, as in the case in which (5.4) does not bind, what matters is the preferences of the elite. When the constraints are tight (e.g., when (5.4) binds), the elite are constrained in the choices they can make.
Our model builds in a natural way on existing models of revolutions. This research-for example, Roemer (1985), Grossman (1991, 1994), Wintrobe 1998), and Bueno de Mesquita et al. (2003) - examines simple games where authoritarian regimes can be overthrown by the citizens and then make various types of responses, concessions such as cutting taxes and redistributing assets, or repression. Like our analysis, these papers abstract from the collective-action problem. Our main innovation comes later when we show how democratization can emerge when concessions are infeasible and when repression is too costly. To understand when concessions are or are not feasible, we need to examine their credibility.
4. Commitment Problems
4.1 Basic Issues
An important issue throughout this book is the inability of those controlling political power to commit not to use it. In other words, the problem is that when those with political power make promises to those without, the promises may sometimes be noncredible. This is important, in turn, because without such credible promises, those in power have fewer options open to them and, in particular, they may sometimes be unable to deal satisfactorily with crises, such as an immanent threat of revolution discussed in the previous section.
The issue of commitment is intimately linked to that of political power. To see this, consider a nondemocracy in which political power lies with the elites. For one reason or another - but, as we will see, most probably to avoid revolution - the elites would like to promise to choose policies in the future that are more to the liking of the citizens - for example, they might want to promise to redistribute income to the citizens. However, the elites hold political power in nondemocracy and, therefore, have the right to determine the level of taxes and transfers in the future. They can promise to make transfers in the future, but these promises may be noncredible. Tomorrow, they get to decide these transfers and, if it is not in their interest to be making them tomorrow, they will not make them. They get to decide whether to make the transfers tomorrow because they hold political power.
It is important to emphasize that the commitment problem arises from the potential de-coupling between the beneficiaries of the decisions and the identity of those holding political power. The transfers benefit the citizens; but, they are made by the elites, who are not the beneficiaries. On the contrary, they are the ones who bear the burden of any transfers. Therefore, typically it is not in their interest to make these transfers in the future and their promises of future transfers and redistribution are not credible. Contrast this with a situation in which political power is in the hands of the citizens. There is a congruence between the identity of those holding political power and those benefiting from the transfers. The citizens would certainly like to implement the transfers from the elites to themselves. This highlights that commitment problems arise when political power is not in the hands of the beneficiaries of the promised policies. In essence, those with political power cannot commit not to use it to renege on the promises made in the past.15
Commitment problems are not only present in politics but also in all areas of social life. Almost all economic transactions have a temporal dimension. Traders typically deliver goods today but receive payment tomorrow. A commitment problem arises if customers promise to make a payment tomorrow but, when tomorrow comes, it is not in their interest to make the payment. In this case, they renege on their promises and fail to make the payment. Therefore, there is ample room for commitment problems in social and economic relations. However, in most instances, society has relatively low-cost ways of dealing with the most major potential commitment problems. To remove potential problems, we need to remove the freedom of customers to decide whether to make a payment tomorrow without facing any repercussions if they renege on their promise. As we saw, the problem is that whenever customers get to make such a decision in an unconstrained manner, they prefer not to make a payment (and thus keep the money in their pocket). There have to be some “constraints” on their actions or some potential repercussions (i.e., punishments) if
they decide not to make the payment. There are three potential way to deal with these commitment problems: contracts, repeated transactions, and changing the identity of who gets to make the decision.
The most common way of dealing with potential commitment problems is to write enforceable contracts. For example, the trader could get the customer to sign a contract at the time of delivery stipulating that in a number of days, the customer will make a payment to the trader. What happens if the customer fails to make the payment? If the contract is in fact enforceable, there is an outside agency, typically a court of law, where the trader files a complaint that the customer broke the terms of the contract. This agency, after determining the truth of the claim, punishes the customer and forces him to make the payment, if possible. Contracts solve most potential commitment problems in an ideal world. However, even in the realm of purely economic transactions, we are far from this ideal world, and there are many problems with these types of contracts in economic transactions, including those stemming from asymmetries of information. They also include those related to the fact that certain important characteristics that one would like to contract upon, such as the quality of the good that the trader delivers to the customer, may not be “contractible” because the outside agency is unable to observe the true quality (the implications of this type of contracting problem is the topic of a large literature in organizational economics; for example, Williamson 1985 and Grossman and Hart 1986). However, potential problems with contracts are much more severe, even unsurpassable, when we come to the political arena.
Economic Origins of Dictatorship and Democracy Page 20
