Identification in structural vector autoregressive models with structural changes

An increasing strand of the literature uses structural changes and different heteroskedasticity regimes found in the data constructively to improve the identification of structural parameters in Structural Vector Autoregressions (SVAR). A standard assumption in this literature is that the unconditional reduced form covariance matrix of the system varies while the structural parameters remain constant. With macroeconomic data this assumption appears untenable. This paper investigates the identification issues that arise in SVARs when structural breaks which occur at known dates affect both the reduced form unconditional covariance matrix and the structural parameters. It is shown that the combination of theory-driven restrictions with the knowledge that different heteroskedasticity regimes characterize the data generalize the necessary and sufficient identification conditions that hold for SVAR without breaks, opening interesting possibilities for practitioners. An empirical illustration shows the usef.ulness of the derived identification conditions by focusing on a small SVAR frequently used to investigate U.S. monetary policy. It is found that two heteroskedasticity regimes characterize the data before and after the 1980s, and this information is combined with economic reasoning to identify the effect of monetary policy shocks on output and inflation.