Defending against speculative attacks: The policy maker's reputation

This paper studies how speculative attacks and regime change arise from the interaction between a policy maker's reputation for defending against speculative attacks and speculators' learning of the policy maker's type. If speculators receive conditionally independent and identically distributed private signals about the policy maker's type in every period, then no matter how precise their per-period private signals are, there can be no equilibrium with attacks. If the speculators receive private signals whose precision increases as a power function of the time with the power strictly greater than two, there exist infinitely many equilibria with attacks, but this equilibrium multiplicity arises from the timing of the attacks only. In such a case, the set of status quos surviving in any equilibrium is strictly larger than in the case of a short-lived policy maker.