Risk measurement for a global credit portfolio

This thesis examines different methods of measuring risk in a global credit portfolio. Because the portfolio includes exposures across multiple markets, it is necessary to capture the dependence structure between them. To achieve this, several statistical techniques are explored, namely different copulas for modeling correlation. The core risk measures evaluated are Value at Risk (VaR), Expected Shortfall (ES), and Component Value at Risk (CVaR), all within the context of credit risk modeling. These measures are analyzed and compared to assess their effectiveness and reliability under varying conditions. In addition, sensitivity analyses are conducted to understand how changes in key parameters within the loss model affect the overall risk profile of the portfolio. The goal is to better understand the implications of model choice and parameter sensitivity on risk estimates for diversified credit portfolios.