Three Essays on the Empirical Relevance of International Trade Models: New Evidence from China.

摘要

My dissertation discusses the extensive and intensive margins of international trade. I investigate this topic using newly available Chinese Customs data and firm-level survey data from 2000 - 2006. These two datasets shed light on China, which is one of the biggest traders in the world, and allow me to test whether trade models motivated by the data of industrialized countries can explain Chinese trade.;My first paper, “Gravity or Distorted Gravity? Evidence from Chinese Data,” empirically tests the main conclusions of Chaney (2008) by using Chinese Customs data. Chaney (2008) theoretically demonstrated that in the heterogeneous firm model of trade, on top of the intensive margin adjustment as in the homogeneous firm model, trade also adjusts through the extensive margin. The elasticity of substitution dampens the impact of trade barriers on the extensive margin while magnifies their impact on the intensive margin. In my paper, to empirically test the implications of the model, first of all, I establish and collect the ad-valorem variable trade costs measures, including freight and tariff costs, for Chinese exports by merging many datasets. Moreover, by applying the measures of trade barriers and trade flows from Customs Data into the structural gravity equations which I derived based on the model, I estimate the variable cost elasticity for each industry. Combining this estimate with the elasticity of substitution data from Broda and Weinstein (2006), I find that the supporting evidences for the heterogeneous firm model are; first, 70% of trade volume growth for China can be attributed to an increase in the number of exporting firms; second, the elasticity of substitution is uncorrelated with the magnitude of the effect of variable costs on trade volume across industries. However, against the implication in Chaney(2008), the main channel of variable trade costs affecting the trade value is through the average sales per firm instead of the number of firms in Chinese data.;My second paper, “The Missing Trade of China: Balls and Bins Model,” provides a benchmark match between the model and the data in the extensive margins of trade. The balls-and-bins model in Armenter and Koren (2010) provides us a new perspective on the sparse nature of trade. Taking advantage of extremely disaggregated shipment level data, my paper finds that, similar to the U.S. data, the balls-and-bins model matches the numbers of zero-trade flows at the country-product and country-firm levels for China. However, at the firm level, with specific numbers of shipments for each firm, the balls-and-bins model cannot predict the percentages of single-product and single-destination exporters as it does for the U.S. data. Furthermore, the theoretical trade models do not out perform the atheoretical balls-and-bins model at the country-product level.;My third paper, “An Anatomy of Chinese Trade at the Extensive and Intensive Margins: Evidence from Transaction Level Data,” dissects Chinese trade into the extensive and intensive margins at the shipment, firm, and product levels. On the one hand, contrary to the U.S. data in Bernard et al. (2007), I find that distance does not dampen the extensive margin at the product or the firm level. On the other hand, the adjustment of shipment numbers accounts for 85% of total value variation. Changes in the number of firms account for 83% of the change in the number of shipments. This is consistent with the Colombia data discussed in Eaton et al. (2007) and Eaton et al. (2008).