Augmented credit-to-GDP gap as a more reliable indicator for macroprudential policy decision-making
|E32, G01, G21, C22
credit-to-GDP gap, out of sample forecasts, augmented credit gap, countercyclical capital buffer, estimation uncertainty
This paper aims to evaluate the possibilities of augmenting the credit-to-GDP gap series with out-of-sample forecasts to obtain a more stable indicator of excessive credit growth. The credit-to-GDP gap is a standardized and harmonized indicator of the Basel III regulatory framework used to calibrate the Countercyclical Capital Buffer (CCyB). Thus, a good indicator should be valid, stable, and represent future financial cycle movements. This research focuses on reducing the end-point bias problem of the Hodrick-Prescott (HP) filter approach to estimating this indicator. This is appropriate for those authorities whose analysis shows that the HP approach best predicts the financial crisis. Several popular models of out-of-sample forecasting are tested on Croatian data to extend the filtered original series, and the results are compared based on multiple criteria. These include the stability of the indicator, not just the usual model forecasting capabilities. The autoregressive approach, alongside the random walk model, was the best-performing one. The results of this study can be used in real-time decision-making, as they are relatively simple to estimate and communicate. Such augmented gaps reduce the bias in the series after the financial cycle turns. Moreover, the paper suggests possible corrections to the credit-to-GDP gap so that the resulting indicators are less volatile over time with stable signals for the policy decision-maker.