CREWS : A CAMELS-based Early Warning System of Systemic Risk in the Banking Sector

This document proposes an aggregate early-warning indicator of systemic risk in the banking sector. The indicator is derived from a logistic model based on the variables in the CAMELS rating system, originally developed for the US, and complemented with macroeconomic aggregate variables. The model is applied to the Spanish banking sector using bank-level data for a complete financial cycle, from 1999 to 2021. The performance of the model is assessed not only during the last global financial crisis and the subsequent sovereign crisis, but also during the recent Covid-19 shock. The proposed indicator has a macroprudential orientation, which differs from most of previous studies predicting individual bank defaults. The indicator is found to provide accurate early-warning signals of systemic risk in the banking sector within a two-year horizon. In this context, the indicator provides mid-term signals of systemic risk that complement those derived from macrofinancial indicators and from measures of the materialization of risk