Large data sets, nonlinearity and the speed of adjustment to real exchange rate shocks

A well known puzzle in international finance concerns the very slow speeds of adjustment of real exchange rates observed in response to shocks. In this article, we explore whether allowing for a wide range of influences on the real exchange rate in a nonlinear framework can help resolve this puzzle. Using, recently proposed econometric methods for summarizing very large macroeconomic data sets into a small number of observable factors, we find that there is a long run relationship between these factors and real exchange rates. When put into a nonlinear framework, we find that allowing for the effects of macroeconomic factors dramatically increases the measured speed of adjustment of the real exchange rate.