Predicting the Risk of Insurance Company's Bankruptcy

Nowadays many participants leave the insurance market. In the period from 2003 to the present, more than 1,300 licenses were withdrawn from the Russian insurance market. One of the main reasons for this phenomenon is the bankruptcy of insurance companies. Consequently, insurance companies, their contractors, and the state are interested in the existence of economic models that allow predicting the bankruptcy of insurers. The purpose of this study is to predict the risk of bankruptcy of Russian non-life insurance companies by building and analyzing survival models (the Cox model, the Cox model with time-varying covariates), and the logit model. The investigated interval: from 2003 to 2016. During the study, two models predicting bankruptcy of insurance companies were obtained: with an accuracy of 72.8% for the Cox model with time-varying covariates and 68.5% for a logit model using company data 1 year before bankruptcy. It was proved that differentiation by groups of authorized capital and types of insurance affect the risk of bankruptcy of non-life insurance companies. Also, there is an impact on the risk of bankruptcy by financial indicators such as the growth of insurance premiums; profitability of assets, equity, sales; current solvency; level of debt ; coefficient of profitability; the ratio of net premiums written to the surplus of the insurer.