Structural VAR Estimation with Exogeneity Restrictions

Exogenous variables arise quite naturally in macroeconomic models of small open economies. In these models, overidentification is also a common feature. In the presence of exogeneity restrictions and overidentification, the usual two-steps approach to the estimation of structural vector autoregressions is not equivalent to maximum likelihood. The authors propose a simple modification of that usual approach which produces maximum likelihood estimators. Copyright 1996 by Blackwell Publishing Ltd