A new interpretation of the exchange rate-yield diffferential nexus

Empirical studies have had difficulty in establishing the long-run relationship between real exchange rates and real yield differentials predicted by sticky price exchange rate models. We revisit this issue in a nonstationary panel regression framework. This facilitates estimation of a long-run parameter even when the underlying relationship is subject to permanent shocks or the variables do not cointegrate. The slope coefficient estimate from a sample of 23 industrialized countries 1973M1-1998M12 has the correct sign and is statistically significant for both short and longterm yields. These results support fundamentals-based models of exchange rate behaviour while permitting real factors to play a role. Moreover, they indicate that capital markets integration is more advanced than hitherto believed.