Search with Partially Informed Stopping Decisions

Abstract

Search models commonly make the simplification that all the relevant information about a sampled object is revealed once it is sampled. In this paper we present and analyze a search model in which searchers have only partial information about the sampled objects at the point of decision. We show that the combination of this imperfection with searchers discretion over which imperfect signal to choose may have a qualitative effect on the equilibrium outcomes. In particular, it generates large multiplicity of equilibria in an environment that would have a unique equilibrium otherwise, and gives rise to equilibria in which stopping decisions might be non-montonic in the strength of the signal that the searcher obtains. These insights owe to the negative search externality that agents exert on each other and which in some form is familiar from other search contexts. But the way in which it plays out here is novel and not immediately obvious.