Managing the impact of COVID-19 on credit risk models
The COVID-19 pandemic is creating many challenges for credit risk models. At the start of the COVID-19 pandemic, government(s) immediately started introducing relief programs and financial institutions began implementing payment arrangements, such as payment holidays, and introducing moratoria on loan agreements. Defaults were prevented or at least delayed. This resulted in adverse effects on the modeling side of credit risk.
On the one hand, existing models forecasted an increasing level of credit risk because risk drivers in the models deteriorated to, sometimes, extremely adverse values. On the other hand, the observed default rate decreased because of relief programs. This ‘disturbed’ the historical relationship between risk drivers and defaults that are captured by the credit risk models currently in place.
This white paper discusses the potential impact of the COVID-19 pandemic on credit risk models. We explain some of the possible modeling methodologies to enhance credit risk techniques to capture the COVID-19 crisis scenarios. We have identified three techniques that can be applied: Quantile Regression (QR), Markov Switching model, and Panel Smooth Transition Regression (PSTR).
To download the full articles ‘Managing the impact of COVID-19 on credit risk models‘, click on the button below.