A Probit Model for Insolvency Risk Among Insurance Companies

We estimate a probit model of insolvency risk, using a dataset of about 400 Dutch insurance companies during the period 1995-2005. The results suggest that surplus capital, company size, profitability, long-tailed business and being a mutual insurer reduce the risk of insolvency. The model can be used to identify insurers with high insolvency risk one year ahead. It is shown that the choice of the threshold above which an insurer is classified as having high insolvency risk, is an important determinant of the relative occurrence of type I and type II prediction errors. We use a loss function to find the optimal threshold given the supervisor's relative preference with respect to missing insolvencies and false alarms