Current Accounts and Exchange Rates: A New Look at the Evidence

Abstract

This paper 'goes back to basics' in empirical analysis of the J-Curve. First, we document strong violations in the distributional assumptions that underlie nearly all previous work on this issue. Second, we employ distribution-free, non-parametric statistical tests to characterize the data and summarize the key relationships between real exchange rates, the current account, and real GDP. We find some (weak) evidence of a J-Curve in the data. Interestingly, however, we document that this evidence is not consistent with the standard theoretical explanation of the J-Curve. Consequently, our empirical results pose a strong challenge for international economic theory.