Jan Janků, Simona Malovaná, Josef Bajzík, Klára Moravcová, Ngoc Anh Ngo
Credit cycles have become longer and more pronounced since the mid-1980s, often amplifying business cycle downturns. While many studies examine the interplay between credit and output, they disagree on the strength and persistence of these effects. We conduct a meta-analysis of over 2,600 point estimates extracted from impulse response functions reported in 68 VAR-based studies across 63 countries. We find that output reacts quickly but briefly to credit shocks, whereas credit responses to output shocks are larger and more persistent, especially in advanced economies. Publication bias inflates reported effects, yet adjusted estimates remain economically significant. We also document substantial heterogeneity, with stronger responses in European samples, studies of corporate credit, and models using Bayesian methods or sign restrictions. These findings help clarify the typical dynamics between credit and output, informing monetary and macroprudential policy design.
JEL codes: C32, C83, E32, G21
Keywords: Credit cycles, impulse response functions, meta-analysis, output, publication bias
Issued: September 2025
Download: CNB WP No. 12/2025 (pdf, 4.3 MB)