Credit risk mangement in finance - a review of various approaches

Classification of customers of banks and financial institutions is an important task in today’s business world. Reducing the number of loans granted to companies of questionable credibility can positively influence banks’ performance. The appropriate measurement of potential bankruptcy or probability of default is another step in credit risk management. Among the most commonly used methods, we can enumerate discriminant analysis models, scoring methods, decision trees, logit and probit regression, neural networks, probability of default models, standard models, reduced models, etc. This paper investigates the use of various methods used in the initial step of credit risk management and corresponding decision process. Their potential advantages and drawbacks from the point of view of the principles for the management of credit risk are presented. A comparison of their usability and accuracy is also made.