Research

Working Papers


Abstract: A large and growing literature focuses on product quality differentiation and its positive and normative importance for international trade. In this literature, there are a number of approaches that indirectly infer product quality based on product price and quantity data. I propose a simple test for the robustness of commonly used approaches to quality inference: when purchasing goods from a common set of exporters, do two importers agree on which goods are high quality? This paper examines three most applied approaches (unit values, demand residuals derived from a CES model, and demand residuals derived from an augmented nested logit framework) and calculates the inferred quality of each exporter-product from the perspective of multiple importers. Exporters' unit values and market shares exhibit positive rank correlations across importers. Counter-intuitively, adding additional variables suggested by the demand residual approaches lowers rank correlations, i.e. worsens the agreement among importers. The inconsistency in quality rankings across importers is very robust, occurring across all industries and regardless of the extent of product differentiation, and corresponds to quite large differences in the price-equivalent variation in quality levels. Further analyses show that such inconsistency is not well explained by changes in the destination-specific quality mix induced by variation in trade costs as in Melitz (2003) or the Alchian-Allen effect (Hummels and Skiba, 2004), or by time invariant differences in consumers' preferences and tastes across importers.

Work in Progress


Impact of Labor Market Conditions on College Major Choices: Aim Low? (Draft coming soon)
(with Xiaoxiao Li and Paul Thomas) 

Abstract: A growing literature emphasizes the importance of human capital investment through college education. College major choice is a dynamic decision associated with future earnings, students' abilities and economic stability. This paper studies the effect of labor market conditions on major switching behavior. We explore a rich administrative data set that covers the complete educational path of each Purdue college student. We find that students start in difficult majors with high expected mid-career earnings. Before entering the market, they are likely to switch into easier majors with high expected early-career earnings. Results from propensity score matching show that switchers stay longer in college, while there is little improvement in GPA compared to non-switchers with similar SAT scores and first-year GPA. We quantify the cost of switching and investigate the welfare effects of major switching by college.