A Semi-Structural Model for Credit Cycle and Policy Analysis – An Application for Luxembourg

The paper explores the nexus between the financial and business cycles in a semi-structural New Keynesian model with a financial accelerator, an active banking sector, and an endogenous macroprudential policy reaction function. We parametrize the model for Luxembourg through a mix of calibration and Bayesian estimation techniques. The model features dynamic properties that align with theoretical priors and empirical evidence and displays sensible data-matching and forecasting capabilities, especially for credit indicators. We find that the credit gap, which remained positive during COVID-19 amid continued favorable financial conditions and policy support, had been closing by mid-2022. Model-based forecasts using data up to 2022Q2 and conditional on the October 2022 WEO projections for the Euro area suggest that Luxembourg's business and credit cycles would deteriorate until late 2024. Based on these insights about the current and projected positions in the credit cycle, the model can guide policymakers on how to adjust the macroprudential policy stance. Policy simulations suggest that the weights given to measures of credit-to-GDP and asset price gaps in the macroprudential policy rule should be well-calibrated to avoid unwarranted volatility in the policy response.