FERNBACH supports the consideration of credit risk throughout all stages of the life cycle of loans. From the deal origination date, on we provide solutions to track the development of credit risks.
For the Loan Application Process, the APP “Credit Spread (AI)” provides a credit spread that can be used for calculating the interest in the offer. This credit spread takes into account the Expected Credit Loss for the individual deal.
Existing approaches that use scoring, rating or Basel II PD for pricing have generally been determined using static procedures for portfolios.
A specialised neuronal net is trained with the following historical data (coming from a performance database (PDB) that contains defaulted and non-defaulted deals in the past):
- Customer data (age, income, provenience, job title, …), rating (if available),…
- Deal data (principal, maturity, deal type, …)
- Macroeconomic data, Market Data
- ECL, defaulted or not
The trained network is able to predict the ECL for new deals directly.
Pricing: Based on the predicted ECL the Credit Spread can be recalculated directly.
For early warning systems the ECL is predicted in case a running deal changed its contractual data or the customer changes his repayment behaviour. If the ECL change exceeds a defined threshold, a notification is generated in order to trigger optimisations of the contractual relationship.