Estimation of Probability of Default

Default Probability: Regulatory Aspects

Institutions implementing IRB have to establish estimates of Probability of Default (PD). Estimates of PD have to be based on internal ratings assessments. PD is the probability of default in 12 month time of the respective internal borrower. Estimating 12 month PD is a challenging task: The financial position of the customer has to be measured, estimated and forecasted. A structural model is needed to forecast the customer financial status and position 12 months ahead

The model has to reflect impact of revenues, costs and changes of values of assets and liabilities on the equity position of borrower. Alternatively, such models can be based on market capitalization of customers or on NPV of future cash flows of customers. The former type of models are only applicable in case of listed customers, the latter in principle in all cases.

Default Probability: Economic Capital Perspective

From economic capital viewpoint an institution has to be able to evaluate contributions of all types of risks, including default risks, of an individual customer’s business transactions. Accurate estimation of risk contributions helps an institution in pricing new business so that good business is not lost due to too tight pricing and that bad business is not attracted by too easy pricing.

The approach developed by CDFT is a completely transparent and repeatable, and a truly institution-specific approach. It can also be run on a reliable basis afterwards, which supports elaborate back-testing and audit trail requirements.