Josef Bajzík, Jan Janků, Simona Malovaná, Klára Moravcová, Ngoc Anh Ngo
We synthesized 3,175 estimates (454 impulse responses) of the semi-elasticity of credit with respect to changes in the monetary policy rate from 91 vector autoregression studies. We found that monetary policy tightening consistently yields a negative and long-lasting response in both credit volume and credit growth. Several factors contribute to the substantial heterogeneity of the effect sizes in this literature. First, publication selectivity significantly exaggerates the mean reported estimate, because insignificant results are under-reported. Second, researchers’ choice of estimation design has a significant impact on the estimated response. Studies using Bayesian methods and including house prices report a smaller decline in credit, while studies with sign restrictions show a larger drop than those using recursive identification.
JEL codes: C83, E52, R21
Keywords: Bayesian model averaging, credit, interest rates, meta-analysis, monetary policy transmission, publication bias
Issued: December 2023
Download: CNB WP No. 19/2023 (pdf, 3.8 MB)