Traditional analysis of the determination of the current account balance of a country is based on static Keynesian models of saving and investment. In the early 1980s, several authors challenged the appropriateness of the traditional model by arguing that the observed movements in the current account balance of a country are the outcome of saving and investment decisions by economic agents. Saving and investment decisions are inherently dynamic in nature in the sense that they involve intertemporal choice which is affected by current as well as expected future movements in economic variables. Therefore, static models in general, and Keynesian models in particular, are incapable of accommodating the dynamic nature of the decisions involved in saving and investment. These authors present explicit dynamic optimizing framework in which they distinguish between the effects of transitory and permanent changes in income and relative price on the current account balance of a country;Despite their elegance, empirical testing of the intertemporal models of current account determination has been limited by our inability to identify the transitory and permanent components in observed economic time series. However, recent developments in time series econometrics provide ways in which one may attempt to decompose an observed nonstationary time series into a transitory and a permanent component. Such a decomposition opens the opportunity to empirically test whether real world data support the predictions of the intertemporal models of current account determination. In this study, two different methods have been used to obtain such decomposition of nonstationary economic variables. Cointegration analysis has been used to examine the long-run relationship among the variables. Then Vector Autoregression (VAR) technique is used to investigate the short-run dynamic behavior of current account balance in response to shocks to transitory and permanent components in income and real exchange rate. The analysis is performed for two countries: the United States vis-a-vis the rest of the world, and Japan vis-a-vis the rest of the world. The results of the empirical analysis are inconclusive. Results for Japanese data are more supportive of the intertemporal models than those for U.S. data.