A Semi-Structural Framework for Measuring Credit Cycles in Europe

Abstract

We develop a parsimonious semi-structural framework for measuring credit cycles using an unobserved component model of credit to the private sector, which can be used for macroprudential policy activation, such as the Counter-Cyclical Capital Buffer. Credit cycle estimation is undertaken using only information available at the time policymakers have to decide on instrument activation. This contrasts with common practice in the literature, where models are estimated using the full sample rather than in real time. Our credit cycle measure's early-warning properties for predicting financial crises are compared against benchmark real-time indicators for several European countries. It performs well in shorter prediction horizons and provides more persistent signals when compared to benchmarks, especially in out-of-sample exercises.

Keywords: Credit cycle, Financial Crises, Forecasting, State space, Bayesian Methods