Working Paper “Measuring Credit Gaps: A Restricted Hodrick—Prescott Filter Approach”
EFSD economists have improved credit gap estimation, adopting it to emerging market credit cycles.
A newly released working paper introduces a modified methodology for assessing the credit gap. The research made by the economists of the Eurasian Fund for Stabilization and Development (EFSD) focuses on improving the robustness, interpretability, and operational relevance of this key early warning indicator for financial crises.
The credit gap serves as a key indicator for macroprudential regulation, formalised under the Basel III framework. Yet the estimation approach advocated by the Bank for International Settlements — a one-sided Hodrick—Prescott filter (HP filter) — has attracted serious critique from both scholars and practitioners.
As its main contribution, this paper proposes a modified approach to measuring the credit gap, combining the strengths of the one- and two-sided HP filters. In particular, the authors propose using the one-sided HP filter with a trend component that is subject to linear constraints based on expert assessment of the two-sided HP filter. A criterion based on bootstrap replications has been developed to help decide whether such restrictions are necessary.
Special emphasis is placed on determining the optimal value of smoothing parameter λ for the specific credit cycle dynamics of the economy in question. The results indicate substantial variation in cycle length, confirming the necessity of individual parameter calibration for each economy.
In addition, the paper proposes a solution to the ‘end-point’ problem by forecasting the credit-to-GDP ratio at least two quarters ahead. This is achieved through an ARIMA model supplemented with the procedure of bootstrapping residuals.
To enhance the practical relevance of the approach, the authors propose an algorithm for measuring the credit gap comprising six consecutive steps, from preparing the data and determining the optimal value of parameter λ to applying restricted one- and two-sided HP filters.
The approach proposed by the authors yields more balanced and robust estimates of the credit gap. This is particularly relevant for monetary and macroprudential authorities in developing economies when designing measures to prevent credit booms and mitigate systemic risks to the banking sector.