Improving the Bank Recovery Process: Empirical Evidence for the Italian Banking System

We develop an empirical model with which we aim to reveal the conditions of a sample of listed banks over the period 2005-2016 in terms of their ability to survive potential extreme losses and the circumstances under which the regulator should intervene. In particular, we calculate the probability of distress of each bank by applying the Merton model; then we quantify the potential losses according to the Vasicek’s approach. The probabilities of distress are then transformed into distances to default (DD), and the corresponding cumulative distribution of banks is used to identify the Type I error (not intervening to shut down operations of a bank that would subsequently fail) and Type II error (shutting down a bank that would survive on its own). Our results show that the banks analyzed have more concentrated levels of DD, with a Type II error of about 20%. This implies that for every one bank out of five, early intervention would have been triggered. The “optimal” recovery trigger should minimize the combination of the two types of error, identifying an “optimal” amount of DD as a criterion for early regulatory intervention.