How vertically specialized is Chinese trade?

Two recent phenomena have transformed the nature of world trade: the explosive growth of Chinese trade, and the growth of vertically specialized trade due to international production fragmentation. While vertical specialization may explain much of the growth and unique features of Chinese trade, few papers have quantitatively assessed these two phenomena together. In part, this is because it is difficult to measure just how vertically specialized Chinese trade is. The unique features of China's extensive processing trade cause both the identification of imported intermediate goods, and their allocation across sectors, to depend upon the Chinese trade regime. In this paper, we estimate the vertical specialization of Chinese exports, addressing these two challenges. Using two Chinese benchmark input-output tables, and a detailed Chinese trade dataset which distinguishes processing trade from other forms of trade, we develop a new method of identifying intermediate goods imported into China. Vertical specialization is then estimated using two methods. The first method uses the Hummels, Ishii and Yi (2001) measure, the official benchmark IO tables, and incorporates our identification correction. The second method follows the first, but also incorporates the Koopman, Wang and Wei (2008) method of splitting the benchmark IO tables into separate tables for processing and normal exports, in order to address the allocation problem. Results show strong evidence of an Asian network of intermediate suppliers to China, and the two methods provide a range of estimates for the foreign content of Chinese exports. In 2002 aggregate exports ranges between 25% and 46%, with some individual sectors are as high as 52! %-95%. Across destinations, under both methods, the vertical specialization of Chinese exports declines with the level of development of the trading partner.