Do Bank-Specific Indicators Matter for Explaining and Predicting Systemic Banking Crises?

Journal of Youth and Adolescence · Springer Nature · 2026-09-24 · DOI: 10.1007/s41549-026-00145-2

Abstract

This paper aims to contribute to the literature on early warning systems by assessing the role of bank-specific variables in explaining and predicting systemic banking crises. We employ a multinomial logit regression to address bias in crisis identification and apply the Bayesian Model Averaging (BMA) technique to select potential crisis indicators, thereby reducing uncertainty in developing a credible and robust model. The study uses a sample of 48 developed and developing countries from 1980 to 2015. Our findings highlight the significance of microeconomic variables in explaining financial fragility and crisis duration. These variables are also relevant for predicting the onset of crises, both in-sample and out-of-sample, although they are less important for correctly identifying crises when using BMA to select the optimal combination of early warning indicators. Additionally, the results demonstrate that aggregated bank-specific characteristics are crucial for explaining and predicting banking defaults.

Read at publisher →

Similar articles