Research has shown that bank failures can be often be blamed on bad risk management of the entire loan portfolio. This is why credit risk management, and in particular its quantitative evaluation, is playing an increasingly important part in the process of bank management. Applying more efficient - but therefore often more complex - methods, allows a bank to get ahead of the competition and conduct a more aggressive policy on the loan market. Attempts to improve credit risk management fall into two categories: relating to the evaluation of risk of a single loan agreement and of the entire loan portfolio. The first approach has produced more literature and instances of application. Various statistical methods are utilized to assess the credit standing of potential borrowers, amongst them discrimination analysis, the probit and logit model (well familiar to econometrics), and models based on decision trees, nearest neighbors and neural networks. The pa