Dynamics of Optimal Search Intensity

Abstract

This paper develops a tractable framework for analyzing compound search problems, where a searcher chooses search intensity adaptively in each period. We fully characterize the optimal search rule and value, decomposing the inter-temporal change of search intensity into the fall-back value effect and the deadline effect. We show that the optimal search intensity (value) is submodular (supermodular) in fall-back value and time. It follows that the fall-back value effect increases when the deadline approaches, and the deadline effect decreases when a searcher’s fall-back gets higher. We further quantify the value of recall and show that Morgan (1983)’s conjecture, that a searcher with full recall searches less intensively than one with no recall, is not always true. Finally, we introduce some applications and extensions of our model.