What do Exporters Know?

The decision of firms to participate in export markets drives much of the variation in the volume of trade. To understand this decision, and in particular to predict how firms will react to currency devaluations or changes in trade policies, policymakers need a measure of the costs firms incur when entering foreign markets. Prior estimates of these costs are often large relative to most exporters' observed revenues. We show that these estimates depend heavily on how the researcher specifies firms' expectations over the potential revenue they would earn upon entry. In response, we develop a novel moment inequality approach that allows us to (1) recover entry costs placing weaker assumptions on firms' expectations and (2) quantify the effects of counterfactual policies. Our approach both introduces a new set of moment inequalities, odds-based inequalities, and generalizes the revealed-preference inequalities introduced in Pakes (2010). We use data from Chilean exporters to show that, relative to methods that must specify firms' information sets, our approach generates estimates of entry costs that are approximately 70% smaller than previous measures. We predict gains in export volume from reductions in entry costs and currency devaluations that are between 30% and 60% larger, respectively, than those predicted by existing approaches.