Real exchange rate determination and the China puzzle

Traditionally, real exchange rates are seen to be influenced in the long run by forces that return economies to purchasing power parity and by differences in productivity growth across sectors and across regions, as per the Balassa-Samuelson hypothesis (BSH). However, minor and realistic relaxations of the assumptions underlying the BSH greatly generalise the set of possible influences over real exchange rates. This paper surveys the literature on real exchange rate determination, investigates short-run behaviour, and addresses puzzles over the trends in China's real exchange rate. While China was widely expected to appreciate against the advanced economies following its first growth surge in the mid-1990s, it actually depreciated slightly until the early 2000s. Then, after 2005, its rate of appreciation was more rapid than expected. These puzzles are resolved by accounting for the effects of the trade liberalisations associated with World Trade Organization accession, China's excess saving, and the tightening of rural labour markets.