Using rough sets to predict insolvency of Spanish non-life insurance companies

Insolvency of insurance companies has been a concern of several parties stemmed from the perceived need to protect the general public and to try to minimize the costs associated to this problem such as the effects on state insurance guaranty funds or the responsibilities for management and auditors. Most methods applied in the past to predict business failure in insurance companies are techniques of statistical nature and use financial ratios as explicative variables. These variables do not normally satisfy statistical assumptions so we propose an approach to predict insolvency of insurance companies based on Rough Set Theory. Some of the advantages of this approach are: first, it is a useful tool to analyse information systems representing knowledge gained by experience; second, elimination of the redundant variables is got, so we can focus on minimal subsets of variables to evaluate insolvency and the cost of the decision making process and time employed by the decision maker are reduced; third, a model consisted of a set of easily understandable decision rules is produced and it is not necessary the interpretation of an expert and, fourth, these rules based on the experience are well supported by a set of real examples so this allows the argumentation of the decisions we make. This study completes previous researches for bankruptcy prediction based on Rough Set Theory developing a prediction model for Spanish non-life insurance companies and using general financial ratios as well as those that are specifically proposed for evaluating insolvency of insurance sector. The results are very encouraging in comparison with discriminant analysis and show that Rough Set Theory can be a useful tool for parties interested in evaluating insolvency of an insurance firm.