Essays in International Economics

Abstract

This dissertation is a collection of three essays. In Chapter 1, I develop a theoretical framework on the sorting of heterogeneous workers across jobs to analyze the effects of tasks specialization within jobs. I adopt the view that a job consists of bundles of tasks. This view emphasizes how exogenous bundling technology limits the degree to which workers can specialize and thus exploit comparative advantage. I obtained two main analytical results: (1) Increased specialization leads to a more dispersed distribution of jobs ranked according to skill intensity, which is also known as extreme-biased technological change (EBTC); (2) There exists a sorting pattern between workers and jobs in equilibrium. The combination of these two results implies that a higher degree of task specialization gives rise to job and wage polarization.

In Chapter 2, I develop an oligopolistic model of trade with heterogeneous firms competing in a homogeneous good market. This framework aims to capture a prominent feature in reality shared by a wide range of industries--there are a few large firms along with a long tail of small ones. To this end, I incorporate the endogenous selection of heterogeneous firms for survival and exporting to the Cournot oligopoly model, and show that the general results in the heterogeneous firm literature that trade leads to improvements in aggregate productivity and as well as consumer welfare also apply to industries consisting of non-atomistic firms.

Chapter 3 investigates the quantitative importance of variation in intermediate input prices on existing techniques for estimating the trade elasticity parameter \theta

and cross-industry productivities. I find that while correcting for intermediate input price variation does not significantly alter the estimate of \theta

, such correction is important in computing measures of relative productivity. Failure to account for these variations result in productivity estimates that are biased upward when input prices are low. Building on Costinot, Donaldson and Komunjer's (2008) approach, I offer a simple method to infer industry-level productivities from only \theta

, bilateral trade flows, and intermediate input prices.