Tsebelis: Nested games

Tsebelis. 1990. Nested games: Rational choice in comparative politics. Berkeley: University of California Press.

Chapter 6

Objective

Models consociationalism as a nested game, in an effort to answer three questions.

1. Under what strategies can elites pursue accommodating strategies despite their followers' demands for intransigence?
2. Why would followers continue to reelect elites who continually pursue such strategies?
3. How do political institutions promote accommodating strategies?

Model

• MASSES: Within the electoral arena, societal groups are sharply divided. They would rather have mutual intransigence than get suckered by the other group (i.e. cooperate when the other guy defects in a PD). Thus, the masses would play "prisoner's dilemma" or "deadlock"; intransigence is always a dominant strategy.
• ELITES: If it weren't for reelection concerns, elites play chicken. They would prefer to get their way and have the other guys lose (like defection against a cooperator in PD), but mutual intransigence (i.e. mutual defection in a PD) is even worse, since they prefer a continuation of the current government to governmental collapse. Thus, elites play "chicken" within the parliamentary arena; if the other player is intransigent, they will give in rather than have mutual intransigent. Thus, my dominant strategy depends on what you do.
• NESTED GAME: In reality, elites face reelection. As such, they must balance between appeasing their followers (being intransigent) and playing chicken. They weight these two games (PD and Chicken) with k and (1-k). K is influenced by two things.
  1. Information. If masses lack information about what elites are doing, then elites can play chicken.
  2. Monopoly on representation. If members of a single group have only one party trying to represent them, then the elites can ignore some electoral pressures and thus play "chicken" instead of "PD" in their relationships with other elites. See Fig 6.1, p 169.

Equilibria

Iterated Game

On unimportant issues, it is an iterated game, since favors today can be traded for favors tomorrow. There are two possible equilbria. First, we might decide to always cooperate on everything. Or, we might decide that we're better off if I always let you win on some issues and you always let me win on other issues.

Under the second equilibrium, INSTITUTIONS might matter. For example, federalism guarantees regional groups complete control over some aspects of policy. Though Tsebelis doesn't discuss it, presumably you could divide up cabinet portfolios to meet a similar goal.

Even without institutions, strategic elites might recognize that their followers care more about an issue thatn another elite's followers do. As such, the first elite might rile up her followers, thus allowing her to credibly commit to intransigence. Since the second elite's followers don't really care about the issue, the second elite gives in.

One-Shot Game

For important issues, it's a one-shot game; neither elite is willing to make a concession in exchange for future concessions. Still, accommodation is possible. For example, the Belgian Egmont Pact was negotiated behind closed doors in strict secrecy. Since the public had no information, elites could play "chicken" (although the masses did have information during the implementation stage).