Multiplier effect and comparative statics in global games of regime change

Abstract

This paper provides a general analysis of comparative statics results in global games. I show that the effect of a change in any parameter of a global game model of regime change can be decomposed into a direct effect, which captures the effect of a change in parameters when agents' beliefs are held constant, and a multiplier effect, which captures the role of adjustments in agents' beliefs. I characterize conditions under which the multiplier effect is strong and relate it to the strength of strategic complementarities and the publicity multiplier emphasized in earlier work. Finally, I use the above insights to identify when comparative statics can be deduced from the model's primitives when they do not depend on the information structure and when they coincide with predictions of the complete information model.