Are bad modelling practices creeping into some retail banks under the auspices of Basel II compliance?
I built my first credit scorecard nearly 20 years ago, and these days I regularly work with retail banks on their approach to the predictive model lifecycle, from undertaking health checks to improving efficiencies and insight. Recently for a few banks, a concern struck me: their interpretation of Basel II regulations seems to influence their credit scorecard development methodology in ways that could increase their bad debt levels.
There are a number of aspects to this, but in this blog, I’ll focus on:
The use of Basel II defined PD models as the primary credit risk assessment
The primary purpose of Basel II PD (probability of default) models is to input into IRB Regulatory Capital Calculations, which in turn provide for comparative measures across organisations. Consequently, a standard default performance definition is required (90 days delinquent, 12 months performance). However, there is a misconception in some areas that credit risk models can only use the Basel II performance definition.
From a retail credit risk perspective, this performance definition may not provide the best assessment for some decision areas and portfolios. For example, in the origination of unsecured loans, default performance may not mature until 15, 18 or even 24 months after account opening.
At one bank, a vintage analysis undertaken on unsecured loans showed a 12-month bad rate of 2.5%, whilst at 18 months it was 4.5% and at 24 months 5.5%. More than half of the accounts that went bad did so in months 13 to 24. Under the PD model design they were using with a 12 month outcome, these accounts were being classed as “good” because they were good at month 12. This affected the ability of this model to assess an applicant’s credit risk, compared to one built on a longer outcome period.
For one bank, we conservatively estimated that this issue was costing them around €5 million per year in additional bad debt. That’s doesn’t include the negative impact on profitability from reduced application acceptance rates.
So what’s a better approach? It is best practice to develop an initial PD model on the most appropriate performance definition for that portfolio or decision area, which is then aligned to the Basel II performance definition for capital calculations. This ensures that the model used for the primary credit risk assessment better reflects the probability of default leading to more accurate lending decisions.
In the coming weeks, I’ll continue to share lessons learned when it comes to modelling and Basel II compliance. For additional best practices, I encourage you to download these two FICO ebooks: High-Performance Predictive Analytics and Comply and Compete: Model Management Best Practices (registration required).
