Importing, uncertainty and the costs of trade

Intermediate goods make up a large share of world trade. However only some producers choose to use imported intermediate inputs while others do not. The decision matters for plant performance. In the data, plants that import are much larger and more productive than non-importers. Yet trade models with heterogeneous firms focus almost exclusively on firms' export decisions rather than their import decisions. Chapter 1 documents the performance advantage of importers and develops a simple, partial equilibrium model of input choice. Chapter 2 develops an analytically solvable model of a small open economy in which heterogeneous firms make endogenous import decisions. We view decisions about importing as decisions about technology adoption. In the model, firms weigh the benefit of operating a technology that uses imported intermediate goods against the fixed cost of developing trade relationships with foreign input suppliers. Only the most efficient firms choose to import. Trade liberalization leads to a reallocation of resources toward more efficient firms, just as in export-decision models, but also leads to improvements in individual firms' productive efficiency. Quantitatively, the model captures important aspects of plant-level data --- including the large performance advantage associated with using imported intermediate goods --- and generates large increases in trade from small decreases in tariffs. Chapter 3 generalizes the model to a dynamic stochastic general equilibrium setting. Costs of international trade are increasingly modeled as fixed costs paid by individual firms. In a dynamic setting, these fixed costs may take two different forms: costs of starting to trade and costs of continuing to trade. The distinction matters when firms experience idiosyncratic shocks to productivity over time. In the model, there are costs of developing relationships with foreign input suppliers and costs of continuing these relationships. The benefit of using imported inputs lies in a combination of the relative price and the technology embodied in the inputs. The model quantitatively captures many important features of plant-level manufacturing data, including the dynamics of import status, entry and exit rates, the size distribution, and the large performance advantage associated with using imported intermediate inputs.