International Trade and Industrial Evolution: An Empirical Framework for Policy Analysis

Pennsylvania State Univ University Park, University Park PA

Investigators

Abstract

Industrial sectors are continually evolving in terms of the variety of products they offer, the mix of their active producers, the prices these producers charge, and their productive efficiency. Using plant-level panel data from developing countries, this project characterizes the way that these evolutionary processes respond to trade policy reforms and to changes in the exchange rate regime. The analysis is based on structural models of industrial evolution that highlight producer heterogeneity, firm-specific productivity shocks, new firms' start-up costs, and uncertainty about the future. By moving away from representative plant analysis, the project captures many micro features of industrial sector responses that the existing trade literature misses. The basic model used in this project quantifies an industry's dynamic response to changes in the intensity of import competition. It characterizes the time paths of firm-level capital gains and losses, entry and exit patterns, price-cost mark-ups, job creation and destruction, average productivity, productivity dispersion, and the menu of product varieties available to consumers. Generalizations of the basic model are also analyzed. These allow firms to export some of their output when it is profitable for them to do so, and/or to make investments that expand their capacity or imperfectly control their productivity shocks. The generalized models address additional issues, including the question of how import competition or exporting opportunities might induce or discourage investment and productivity growth. All versions of the model quantify the role of expectations and policy credibility in shaping responses to policy reforms. The models are econometrically fit to large data sets that the principal investigators have acquired from Colombia, Morocco, and Taiwan. For each country, these multi-year data sets contain detailed plant- or firm-level information on sales revenues and input costs. Augmented by data on tariffs, the exchange rate, and the total value of imports, this cost and revenue information supports estimation of the basic model. For the extended models, data on export revenues and on productivity-enhancing expenditures (like fixed investment, worker training and technician salaries) are also used. The methodological contribution of the project is to adapt applied industrial evolution models to an open economy setting, and to implement these models using available micro data. In so doing the project breaks new ground concerning the measurement of sunk start-up costs and firm-specific productivity trajectories. The project also facilitates policy debates by quantifying many dimensions of industrial responses to policy reforms in a single integrated framework. The consequences of commercial policy reforms that are of interest to business and labor (changes in profitability, capital gains or losses, and job creation or job destruction) are measured and compared to the consequences of interest to consumers (changes in prices and the available menu of product varieties).