Open Access

Kevin Jakob

• Since it is in the nature of banking that sometimes borrowers are unable to serve their loans, credit risk is one of the oldest risks financial institutions have to deal with. So-called non-performing loans where one of the most critical issues the ECB and other supervisors around the world were focused during the time of the financial crises and the following years. But still today in light of the COVID19 pandemic, the Ukraine crisis or the climate crisis credit risk and the creditworthiness of loans is one of the biggest threats within the financial system. Therefore, the individual creditworthiness of counterparties is analyzed and monitored on an ongoing basis with the help of rating models by different rating agencies and financial institutions using their own models. However, in general a single counterparty does not represent a threat to a (well diversified) financial institute or the overall systems. But risks increase when economic dependencies come into play and the defaultSince it is in the nature of banking that sometimes borrowers are unable to serve their loans, credit risk is one of the oldest risks financial institutions have to deal with. So-called non-performing loans where one of the most critical issues the ECB and other supervisors around the world were focused during the time of the financial crises and the following years. But still today in light of the COVID19 pandemic, the Ukraine crisis or the climate crisis credit risk and the creditworthiness of loans is one of the biggest threats within the financial system. Therefore, the individual creditworthiness of counterparties is analyzed and monitored on an ongoing basis with the help of rating models by different rating agencies and financial institutions using their own models. However, in general a single counterparty does not represent a threat to a (well diversified) financial institute or the overall systems. But risks increase when economic dependencies come into play and the default of one counterparty leads to the another or when multiple counterparties are affected by the same (macro)economic shock. At this stage, rating models are still useful for the analysis of individual risks but to analyze the credit risk for a whole financial institute (or system) based on its portfolio one needs credit portfolio models taking dependencies and different uncertainties into account. Quantitative credit portfolio models arise in the 1990s, were also two of the most famous credit portfolio models, namely the Credit Metrics and the Credit Risk+ model come up. However, the possibility to quantify credit risk with the help of measures like value at risk (VaR) also led to a deceptive sense of security. In the years before the financial crises, portfolio models were used for analysis of asset backed securities (ABS) which then were one of the main triggers of the crisis. Besides other issues also disadvantages of the statistical models used contributed to the crises. Today, the topic of model risk is discussed very intensively and banks are requested by supervisors to independently validate their models and to take uncertainties in the form of model risks into account. Within this thesis, we analyze different modeling techniques to enhance and improve credit portfolio risk measurement based on the two mentioned credit portfolio models. This includes the estimation and modeling of the dependencies between counterparties as well as the stochastic modeling of other risk parameters besides creditworthiness and their dependency among each other and finally we present some helpful tools to work and explore credit portfolio risk together with investigations for faster approximation techniques.…