Understanding regional export growth in China

In this paper, we use disaggregated data on regional trade in China to assess the channels through which the country's exports have surged. As a starting point for our analysis, we use a Bartik (1991) shift-share approach to evaluate the common component of industry-level export growth across regions in China. If regional comparative advantage or industry agglomeration patterns are roughly stable over decadal time periods, then export growth across regions will vary according to their initial patterns of industrial specialization and which industries enjoyed rapid export growth at the national level. We also consider the impact of explicit measures of policy change that the literature has identified as drivers of China's trade expansion, including reductions of tariffs final goods, tariffs on imported intermediate inputs, trade-policy uncertainty for China in the U.S. market, and MFA quotas on apparel and textile products. We find that a simple Bartik measure has substantial predictive power for China's regional export growth. Once we add the Bartik measure to the analysis, the impacts of reduced input and output tariffs or trade-policy uncertainty on China's export growth fall substantially and become statistically insignificant. These tariff-based predictors of export growth are also very sensitive to the inclusion of time trends across provinces and broad sectors, whereas the Bartik measure has considerably more success in predicting variation in export growth within provinces and sectors. There is little evidence that regions more exposed to the elimination of MFA quotas enjoyed faster export growth.