Credit gaps as banking crisis predictors: A different tune for middle- and low-income countries

The aim of this paper is to assess the quality of credit-based variables as early warning indicators of systemic banking crises. The existing literature focuses mainly on developed economies and shows that the best performing indicator is the credit-to-GDP gap computed via one-sided HP filter (the so-called Basel credit gap). The empirical evidence legitimates the use of the credit-to-GDP gap as a key indicator in macro-prudential banking regulation, i.e., in the determination of the countercyclical capital buffer. We take advantage of a new database on bank credit series and credit gaps covering more than 160 countries (Bouvatier, Delatte and Rehault, 2022) to focus specifically on middle- and low-income countries. Our findings suggest that the credit-to-GDP gap remains the single best performing indicator regarding the high-income group while the same does not hold for middle- and low-income countries. This result highlights that a one-size-fits-all approach is not relevant in the design of the operational framework of the countercyclical capital buffer. Further, we show that the credit gap turns to be a fair banking crises predictor when the financial development, captured by the trend's value in credit-to-GDP ratio, exceeds 20%.