Set-Identified Structural Vector Autoregressions and the Effects of a 100 Basis Point Monetary Policy Shock

Abstract

Abstract I estimate impulse responses to a 100 basis point US monetary policy shock using setidentified structural vector autoregressions. Identified sets for these responses may be unbounded when the identifying restrictions admit zero impact response of the federal funds rate following a standard-deviation shock. Such a zero response is always admissible when there are fewer sign and zero restrictions than endogenous variables. This is the case under existing restrictions on the systematic component of monetary policy, which consequently yield uninformative identified sets. Additional sign and narrative restrictions yield informative identified sets and imply smaller output responses than some previous estimates.