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

A small literature has developed exploring nonlinearities in exchange rate data. This paper studies bivariate statistical relationships between current accounts, exchange rates, and cross–country ratios of GDP. The analysis differs from previous empirical work on these issues in using nonparametric methods to allow for nonlinearities, in employing a minimum of statistical assumptions, and in focusing on a fundamental characterization of the data. The authors present new evidence on the connections between exchange rates, the current account, and GDP. While the evidence loosely supports some common beliefs about the data, it conflicts with some common theoretical models. Thus the results pose new challenges for theory.