Predikce pravděpodobnosti selhání českých bank

Abstract

One of the most important tasks of credit risk control is prediction of borrower’s probability of default. The wrong estimation of probability of default could lead to many problems in financial sectors. That is the reason why great attention is just pursued credit-scoring models. The aim of this dissertation work is estimation of probability distribution of future value of probability of default of five Czech banks. We use a logit model, exactly model GaG3, to determine the probability of default. The future values of the probability of default are modelled by plain simulation Monte Carlo (PMC). The theoretical part of this dissertation work is composed by the second and the third chapter. First of all, the second chapter pursues mainly approaches to evaluation of efficiency and competitiveness of banks and description of bank financial indicators, whereas the third chapter engages in credit-scoring models and definition of credit risk. The third chapter also contains a procedure of estimation of distribution of probability of default, stochastic processes and simulation Monte Carlo. The practical oriented forth part represents application of methodology on chosen banks. In this chapter the model GaG3 is used for calculation of probability of default and finally a simulation of future values of probability of default is carried out.