Measuring Credit-To-GDP Gaps. The Hodrick-Prescott Filter Revisited

32 Pages Posted: 9 May 2019

Date Written: May 8, 2019


The credit-to-GDP gap computed under the methodology recommended by Basel Committee for Banking Supervision (BCBS) suffers of important limitations mainly regarding the great inertia of the estimated long-run trend, which does not allow capturing properly structural changes or sudden changes in the trend. As a result, the estimated gap currently yields large negative values which do not reflect properly the position in the financial cycle and the cyclical risk environment in many countries. Certainly, most countries that have activated the Countercyclical Capital Buffer (CCyB) in recent years appear not to be following the signals provided by this indicator. The main underlying reason for this might not be only related to the properties of statistical filtering methods, but to the particular adaptation made by the BCBS for the computation of the gap. In particular, the proposed one-sided Hodrick-Prescott filter (HP) only accounts for past observations and the value of the smoothing parameter assumes a much longer length of the credit cycle that those empirically evidenced in most countries, leading the trend to have very long memory. This study assesses whether relaxing this assumption improves the performance of the filter and would still allow this statistical method to be useful in providing accurate signals of cyclical systemic risk and thereby inform macroprudential policy decisions. Findings suggest that adaptations of the filter that assume a lower length of the credit cycle, more consistent with empirical evidence, help improve the early warning performance and correct the downward bias compared to the original gap proposed by the BCBS. This is not only evidenced in the case of Spain but also in several other EU countries. Finally, the results of the proposed adaptations of the HP filter are also found to perform fairly well when compared to other statistical filters and model-based indicators.

Keywords: credit-to-GDP gap, cyclical systemic risk, early-warning performance, macroprudential policy, statistical filters

JEL Classification: C18, E32, E58, G01, G28

Suggested Citation

Galán, Jorge, Measuring Credit-To-GDP Gaps. The Hodrick-Prescott Filter Revisited (May 8, 2019). Banco de Espana Occasional Paper No. 1906 (2019), Available at SSRN: or

Jorge Galán (Contact Author)

Banco de España ( email )

Alcala 48
Madrid 28014

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics