Peter Norman (UNC)

“Sequential Persuasion”

Abstract
This paper studies sequential move persuasion games with multiple senders. We use convex analysis to transform a problem with innite action spaces to a nite action model. This way we prove the existence of equilibria by the Zermelo-Kuhn backward induction algorithm, show that equilibrium outcomes are generically unique, and obtain a simple algorithm for nding equilibrium outcomes. We also obtain a simple condition for when full revelation is the unique equilibrium outcome and some comparative statics results. Adding a sender who moves rst cannot reduce informativeness in equilibrium, and will result in a more informative equilibrium in the case with two states. Sequential persuasion cannot generate a more informative equilibrium than simultaneous persuasion and is always less informative when there are only two states.