What Can We Learn from Sign-Restricted VARs?

Abstract

<jats:p> I use a simple business cycle model to illustrate the workings and limitations of sign restrictions in structural vector autoregressions. Three lessons emerge. First, such sign-based identification is vulnerable to “shock masquerading”: linear combinations of other shocks may be misidentified as the shock of interest. Second, since the popular Haar prior automatically overweights more volatile shocks, the implied posterior is decisively shaped by relative shock volatilities--a feature of shocks that has nothing to do with their dynamic causal effects. Third, sign restrictions on structural elasticities--rather than just the usual restrictions on impulse responses--can be highly informative. </jats:p>