Tvorba a vybrané aplikace scoringového modelu defaultu komerčních bank pro období finanční krize

Abstract

One of the most important problems in risk management is the correctly determining the probabilities of default (PDs) of particular financial subjects. In this thesis a possible determination of a commercial bank’s PD based on credit-scoring models is discussed. The aim of this doctoral thesis is to model the probability distributions of PD of selected U.S. and Czech commercial banks using Lévy processes, based on the estimated prediction model. The thesis is divided into two parts. The first part is devoted to the estimation of the three different models (based on linear discriminant analysis, logit regression and probit regression) from a sample of almost three hundred US commercial banks. These models are then compared and verified with the control sample to determine the best model. The second part of the thesis aims to apply the chosen model to a portfolio of chosen US and Czech banks to estimate their present financial stability. However, it is also important to be able to estimate the evolution of PD in the future. So, in the next step of application part, the values of particular indicators are obtained through random sampling and used to estimate the PD distributions. It’s assumed that the indicators are distributed according to a multidimensional subordinated Lévy model (Variance Gamma model and Normal Inverse Gaussian model, particularly). A chance that “a financial crisis” will occur, at least in terms of probability, is indicated by estimation of the various quantiles in the estimated distributions. It is verified in the thesis that Lévy processes can capture the characteristics of the empirical distribution better than traditional Geometric Brownian Motion a therefore, that use of Geometric Brownian Motion for simulation of future probability distribution of PD can lead to undervaluation of risk. Finally, it should be noted that the applicability of the estimated model (with respect to the used data) is limited to the recessionary phase of the financial market.