An Econometric Cntribution to the Intertemporal Approach of the Current Account

This paper investigates an intertemporal optimization model to analyze the current account through Campbell & Shiller’s (1987) approach. In this setup, a Wald test is conducted to analyze a set of restrictions imposed to a VAR, used to forecast the current account for a set of countries. We focused here on three estimation procedures: OLS, SUR and the two-way error decomposition of Fuller & Battese (1974). We also propose an original note on Granger causality, which is a necessary condition to perform the Wald test. Theoretical results show that, in the presence of global shocks, OLS and SUR estimators might lead to a biased covariance matrix, with serious implications to the validation of the model. A small Monte Carlo simulation confirms these findings and indicates the Fuller & Battese procedure in the presence of global shocks. An empirical exercise for the G-7 countries is also provided, and the results of the Wald test substantially change with different estimation techniques. In addition, global shocks can account up to 40% of the total residuals of the G-7.