Guillermo Noguera (with Robert Johnson):
How Large IS That Trade Deficit?

Trade in intermediate inputs accounts for as much as two thirds of international trade, direct evidence of the fragmentation of production processes across borders. While a large volume of theoretical work has explored the causes and consequences of this type of trade, most empirical work is either country or sector specific and therefore fails to provide a broad perspective on the integrated global production structure. In our paper, we quantify cross-border production linkages for the world as a whole. We use input-output tables and bilateral trade data from GTAP 6 Database, the purchase of which was made possible by an IBER Dissertation Award.

By combining data into a new dataset, we showed that production sharing significantly distorts bilateral trade patterns and openness measures. This occurs for several reasons. First, bilateral production sharing implies that exports and imports are scaled down in value added terms relative to gross terms. For example, US trade with Canada is about 40% smaller measured in value added terms than in gross terms.

Second, multilateral production sharing gives rise to indirect trade that occurs via countries that process intermediate goods. For example, Apple’s iPod is produced with a blueprint from California and electronic components from Japan, China, and other countries, which are assembled in China into the final product that is then shipped to the US and other markets for distribution. In US trade statistics, the iPod appears (roughly speaking) as a $150 import from China. But the value of the Chinese content added to the iPod is only a fraction of that value; rather, a substantial portion is indirectly attributable to Japan. This type of indirect trade in turn implies that bilateral trade imbalances can be quite different when measured in value added terms. For example, while the US-China gross trade balance was about $25 billion larger than that US-Japan balance in 2001, we show that they were approximately equal in value added terms.

Finally, input-output linkages also distort sector-level measures of exports. Many sectors that appear to export a small amount of their production (e.g. raw milk sector) are in fact exposed to international markets because their output is embodied in goods that are directly exported (e.g. cheese). We show in particular that the Services sector appears much more tradable relative to Manufactures under our value added export metric.

I am currently working on a second paper that uses our dataset of trade in value added to quantify the gains from global production sharing in a multi-country multi-sector general equilibrium model of Ricardian trade. Stay tuned!

Guillermo Noguera is a PhD Candidate in Economics at UC Berkeley.
http://www.ocf.berkeley.edu/~gnoguera/