This course follows a detailed step-by-step process designed to insure state of the art credit model accuracy and testing, using actual data from the Kamakura Risk Information Services public firm default models as a basis for course materials.
Objectives of the Credit Risk Process: Macroeconomic Factors and Correlations in Default Correlated Default: The Evidence A Review of Credit Models Testing and Calibration of Credit Models
- In sample versus out of sample testing
- Is each of the N competing models better than random chance?
- If so, which is the best?
- Is a combination of 2 or more models better than any one alone? If so, which 2 or 3?
- Periodicity of the data: monthly versus annual
- Creating the term structure of default probabilities and testing effectively
- Point in time versus "through the cycle" default probabilities and ratings
- ROC accuracy ratio and Cumulative Accuracy Profile
- Van Deventer and Wang test for comparing actual and expected defaults over time
- Falkenstein and Boral test for detecting bias in default probability levels
- Out of sample test regime
- Usage of the Credit Models and Model Testing in Practice
- Naïve Models vs. Reduced Form and Structural Credit Models
- Case Studies in Default
- Implications of Cyclicality in Default Probabilities and Recovery Rates
- Implications for Credit Spread Modeling
- Implications for CDOs and Credit Portfolio Management Financial Reporting of Test Results for Risk Management: A Regulatory and Managerial Perspective
Dr. Donald R. van Deventer, Kamakura founder and member of the RISK Hall of Fame
Ms. Li Li, Managing Director for Credit Advisory Services at Kamakura Corporation.
For more information, please contact firstname.lastname@example.org.