The Structure and Evolution of Entry Costs in Trade.

Abstract

The first essay addresses the structure of sunk export entry costs. Heterogeneous firm models of international trade include a country specific sunk cost that firms pay to enter a foreign market. Despite the important role these costs play in determining firm participation, we understand little about their structure. Characterizing the structure of sunk entry costs relies on careful identification of state dependence conditional on heterogeneity. I employ reduced form and structural approaches and a confidential panel of U.S. manufacturers' exporting behavior to the top 50 destinations 1992-2007 to conclude that export entry costs are mainly country specific. The structural model allows me to estimate that the global entry cost is $20 thousand while country specific entry costs are $3.7 million for Canada, $4.16 million for Japan, $3.58 million for Mexico, $4.22 million for the United Kingdom, and $3.63 million for Germany.

Using confidential microdata from the U.S. Census, in the second essay we find that the fraction of manufacturing plants that export rose from 21 percent in 1987 to 39 percent in 2006. It has been suggested that similar trends in other countries may have been caused by declining costs of entering foreign markets. Our study tests this hypothesis for the first time. Both reduced form and structural estimation approaches find little evidence that entry costs facing U.S. plants declined significantly over this period. We instead argue that changes in other factors that determine export status are sufficient to explain these trends.

In the third essay, we propose a framework that embeds search and matching frictions in a general equilibrium international trade model with heterogeneous firms. Search implies that the import price lies between the domestic price and the foreign affiliate's average cost of production. With no search frictions our framework recovers the standard model. Search does not affect the quantity traded, the productivity threshold associated with exporting or the domestic price. Nevertheless, search parameters enter the standard gravity equation, and the total value of imports falls as search frictions rise. We argue for the use of disaggregated data to quantify the importance of these frictions.