Revisiting the Dynamic Impact of Asset Purchases: A Survey‐Based Identification

ABSTRACT We propose a novel instrument for identifying central bank asset purchase shocks in a proxy‐VAR. Our instrument exploits the deviations between asset purchase announcements and expectations inferred from quantitative surveys. Using euro area data, we find a positive impact of purchases on macroeconomic variables with high posterior probability. An asset purchase shock of 1% of GDP leads to median impacts on output and prices of 0.12% and 0.07%, respectively. The effects are three times as small as those in the US economy. Finally, we show that our instrument is stronger than high‐frequency instruments, both in terms of statistical strength and alignment with narrative evidence.