Dynamic Seemingly Unrelated Cointegrating Regression

Abstract

Multiple cointegrating regressions are frequently encountered in empirical work as, for example, in the analysis of panel data. When the equilibrium errors are correlated across equations, the seemingly unrelated regression estimation strategy can be applied to cointegrating regressions to obtain asymptotically ecient estimators. While non-parametric methods for seemingly unrelated cointegrating regressions have been proposed in the literature, in practice, specification of the estimation problem is not always straightforward. We propose Dynamic Seemingly Unrelated Regression (DSUR) estimators which can be made fully parametric and are computationally straightforward to use. We study the asymptotic and small sample properties of the DSUR estimators both for heterogeneous and homogenous cointegrating vectors. The estimation techniques are then applied to analyze two long-standing problems in international economics. Our first application revisits the issue of whether the forward exchange rate is an unbiased predictor of the future spot rate. Our second application revisits the problem of estimating long-run correlations between national investment and national saving.