Real Exchange Rate Movements in Developed and Developing Economies: an Interpretation of the Balassa-Samuelson's Framework

This paper investigates empirically the Balassa-Samuelson hypothesis (BSH) using annual data over 1970-2008 from 33 countries grouped into developed and developing countries. The innovative feature of our study is that we introduce a new approach for classifying traded and nontraded industries. Our proposed approach allows for country specific heterogeneity over each industry, and changes in classifications of industries across periods. We apply panel cointegration tests with group-mean panel dynamic ordinary least squares (DOLS) estimators suggested in Pedroni (2000, 2001) to examine the BSH. We also use persistence profiles in order to investigate the speed of adjustment to equilibrium in response to shocks on cointegrating relations, and employ them as a complementary tool for checking whether the long run relationships obtained from the tests are actually valid. Our main finding is that there is not enough evidence in favour of the BSH nor in favour of purchasing power parity (PPP) in either developed or developing countries. In developed countries, there is strong evidence that higher growth countries will experience real exchange rate appreciation. However, this relation does not provide evidence in favour of the BSH. We actually find opposing evidence that as productivity growth in traded goods relative to that in nontraded goods increases, the real exchange rated tends to depreciate. For developing countries, our results support the BSH. However, the speed of reversion to equilibrium is very slow. We also find moderate evidence that comparatively rapidly growing developing countries will experience real exchange rate depreciation.