Essays in International Economics and Macroeconomics

Abstract

The first chapter presents a study on the employment effects of periodic policy lapses and renewals. Economic theory postulates that uncertainty and temporary disruptions in policies could lead to more permanent effects on the real economy by reducing growth and investment. The sources of policy uncertainty are manifold (Baker et al., 2014), in which political polarization is a potential explanation in the rise in policy-related economic uncertainty in the U.S.. One way in which political polarization could translate to greater policy uncertainty is the challenges involved in passing policies that are subjected to periodic lapses and retroactive renewal. This paper seeks to examine the cost of this type of policy shock on firm employment outcomes. I utilized the periodic lapses and renewals of the Generalized Systems of Preferences (GSP) in the 1990s to estimate the effects on firm employment. Depending on the amount of imports, firms with higher dependency on the GSP regime experienced slower employment growth. The difference in the employment growth rates between the GSP firms and controls persists for about four years. The second chapter discusses a key empirical challenge in measuring misallocation and proposes a method to resolve the identification problem. The challenge in the misallocation literature made popular by Hsieh and Klenow (2009) is the identification of model parameters: a standard implementation cannot separately identify the production function parameters from the measures of distortion. In this paper, my co-author and I formally characterize two biases related to this lack of identification: mismeasuring the labor output elasticity in a constant returns-to-scale world, and assuming constant returns to scale when returns to scale in production are not constant. In both cases, the econometrician’s error induces spurious correlations between productivity and distortions, leading the econometrician to mismeasure misallocation. We first show how misallocation measures in this class of models can be identified even when we cannot identify all the model parameters. We then use U.S. Census Bureau microdata and document the magnitude of the two biases. The third chapter measures the aggregate employment growth and reallocation effects of multinational firms in the U.S. over the past decade and across the manufacturing, retail, wholesale, and service sectors. At a fundamental level, understanding the contribution of multinational firms to U.S. employment growth in contrast to non-multinational firms is the first-order issue. Are MNCs a major component of U.S. manufacturing employment decline? Did they grow faster relative to their domestic counterparts? Did they create and destroy more jobs at new and existing establishments relative to controls? And, given that MNCs are often vertically integrated and operate multiple lines of business, in what sectors did they create and destroy jobs? My co-authors and I exploit a novel combination of two micro datasets: the restricted-use U.S. Census Bureau establishment-level microdata and the Bureau van Dyke Orbis firm database. The combined dataset links firm and establishment-level activity to the scope and extent of a firm’s global operations. We find that MNCs recorded higher total employment growth rates relative to the comparison group of non-MNCs. Furthermore, MNCs create more and destroy fewer jobs than non-MNCs. Moreover, MNCs are found to have created relatively more jobs across all sectors; notably in the services sector. This result suggests that within-firm reallocation across sectors may be increasingly important in the study of business dynamism.