The quality of predictive algorithms plays a crucial role in Creditinfo operations. We strive to help our Clients perform efficient credit decisions through smart and innovative use of data.
Our operations are spread across numerous territories in five continents and many of our countries have lived through challenging periods before. This has allowed Creditinfo to aggregate an invaluable experience of risk management in a crisis. Each crisis poses unique challenges and sees different government responses. However, amidst all uncertainties we have numerous times come to the following conclusions:
- Credit Bureau Scorecards are behavioural in nature and thus they promptly adjust to the changing economic environment.
- Average Credit Bureau Score is higher in an expansion phase than it is in a recession.
- Scorecards developed by Creditinfo continue to rank order credit risk during a recession.
- Understandably, observed default rates are likely to increase across the full spectrum of scores. In other words, we are typically seeing that all customers become riskier.
- Scorecards continue to rank from the highest risk at the lowest scores to the lowest risk at the highest scores. This is because Creditinfo’s model development methodology pays a significant attention to the interpretability of trends utilized in the predictive algorithms. Furthermore, we validate each developed scorecard against models that we have created historically in similar projects. Such qualitative measures help to capture causal relationships and thus ensure that our scorecards perform well through all parts of the economic cycle.
- It is necessary for lenders to change their cut-off scores to higher values and make other score-driven policies more conservative to retain the quality of the portfolio and compensate for the increased risks.
- This can be done based on expert judgement or economic impact analysis and affect either all customers or just specific industries that are expected to be hit the most by the crisis. Lenders should not neglect increased correlation between Probability of Default and Loss Given Default that is observed in a recession. Furthermore, financial institutions should review limit assignment policies. This is critical in an environment where incomes are vulnerable to reductions.
- Where regulators and governments are providing assistance to borrowers or asking the lenders to do so, we are working closely with the data providers to ensure that these cases are identified and have a minimum impact on the scores.
- The data provided to the Creditinfo database will incorporate the updated agreements with payment holidays and other short-term measures, which in turn, will be reflected in the scores. For more details on the specifics of each market please contact your local Credit Bureau.
We are confident that Credit Bureau Scorecards remain an extremely valuable tool at your disposal. It is necessary to review score-driven policies to ensure that this tool continues being optimally used.
Together we will get through the storm!
More questions on we can to support you in managing your limits and cut-offs during these challenging times?