UC Berkeley

This dissertation analyzes the interaction of firm level heterogeneity and selection. This selection can encompass either the survival of firms in the market, or their participation in international trade. In the first part I concentrate on the question of how large the aggregate productivity losses from the misallocation of resources across firms are. If firm exit is endogenous, micro-frictions can induce extensive margin misallocation: inefficient firms continue to survive (Zombies) and efficient firms are forced to exit (Shadows). I develop and estimate a fully specified model with plant-level micro-data to quantify extensive margin misallocation. This method allows me to identify Zombie firms, estimate the efficiency distribution of Shadow firms and calculate aggregate Total Factor Productivity (TFP) gains when Zombies are replaced by Shadows. Estimates for Indonesia show that extensive margin misallocation can increase aggregate TFP losses from micro-distortions by over 40%. Compared to existing estimates aggregate TFP losses from micro-distortions are 50-100% higher.

In the second part of the dissertation I trace out the implications of a simplified version of the framework I developed in the first part to questions of international trade. The cross-country comparisions in measured micro-distortions suggest that differences in firm heterogeneity could be potentially important to explain aggregate TFP and therefore also trade patterns. I consequently develop a tractable multi-country general equilibrium model of such differences in firm level heterogeneity across countries. I show how the model naturally links measured trade frictions to national firm-efficiency distributions and endogenously generates asymmetries in trade flows in the absence of asymmetric trade frictions. The model is able to generate key stylized facts, specifically the absence of a strong negative relationsship of firm-size dispersions and internal trade shares as predicted by the standard heterogeneous firm trade model with identical efficiency dispersions across countries.