Concessions and Information Cascades in Autocracies

• In Event: Authoritarian Politics

Abstract

This paper offers an explanation for Tocqueville’s long-held insight that ”The most dangerous moment for a bad government is when it begins to reform.” (1856) with a formal model featuring information cascade. In the authoritarian context, this Tocquevillian proposition builds on an issue every dictator of a declining regime faces. On one hand, making concession to challengers raises the legitimacy of the regime among the beneficiaries. On the other hand, it may also reveal the weakness of the regime and thus cause challengers to demand more concessions (e.g. populist policies such as franchise extensions). This poses an interesting puzzle since, historically, we do see quite a few examples of partial concessions---e.g., franchise in England was extended progressively to a wider population in 1832, 1867, and 1884---that might undermine the regime according to Tocqueville.
This paper examines, despite this issue, when it is optimal for a dictator to make a concession in order to avoid revolutions. We develop a game theoretical model in which a dictator decides whether to make a costly concession, and then citizens sequentially decide whether to revolt. Each citizen receives a private signal about the level of grievance against the regime among the citizens, and then decides whether to revolt. Each citizen revolts if she believes that the level of grievance is sufficiently high. If a citizen revolts, a revolution is organized. If all the citizens revolt, the revolution is initiated. Conditional on being initiated, the revolution is successful if and only if the grievance level is sufficiently high. Since the dictator does not receive any private signal about the grievance level, the dictator’s action does not reveal any information about it.

We examine the trade-offs faced by the dictator. On one hand, if he makes a concession, each citizen is treated better, which makes a revolution less attractive to them. On the other hand, the concession affects strategic inference among the citizens, which may trigger the revolution to be initiated (information cascade may be more likely to happen). An intuition is as follows. First, if the dictator makes a concession, the citizens are treated better and a revolution becomes less attractive. Thus, each citizen is less likely to revolt. Second, however, once a citizen revolts, then a revolution is more likely to be initiated after a concession (information cascade is more likely to happen). To see why, suppose that the first citizen revolts, even though the dictator has made a concession and a revolution is less attractive. Then, the following citizens would infer that the first rebel must have received a strong signal that the grievance level is high. This inference may trigger an information cascade. Thus, after a concession, a revolution is less likely to be organized but more likely to be initiated conditional on being organized.

The comparative statics of the equilibrium show that the dictator makes a concession if the information possessed by each citizen is sufficiently accurate. This suggests that dictators are more likely to make a concession if more advanced information technology becomes available to the citizens. This paper provides an explanation for when the elite implement populist policies.

Authors

• Yuki Takagi, Department of Political Science, Stanford University

• Hans Hanpu Tung, National Taiwan University