An area of concern for risk managers is how “adverse selection” may impact default rates. If you’re not familiar with the term, adverse selection refers to the fact that an uncompetitive credit offer – such as one with high interest rates or a low initial credit line — will be taken up more by the riskiest people in a given population. Generally, lowering the price charged to an applicant sub-population will result in a lower average bad rate from that sub-population; while raising the price will raise your bad rate too.
One cause of adverse selection we often see is when a bank charges higher interest rates for applicants who do not deposit their regular salary into the bank than for people who do, regardless of the risk profile of the applicant. This pushes up the bad rate in the sub-population where rates are higher.
Here is why adverse selection leads to higher default rates:
- Hidden information: The borrower has private “adverse information” not possessed by the lender. For example, the customer may be showing financial stress at their regular bank, where their salary is deposited, but that delinquency is not recorded yet in the bureau. This borrower will have a lower price sensitivity — they’re keen to lay their hands on money — and so will accept terms others wouldn’t.
- Credit capacity: As the monthly payment for a loan account is an important fraction of a borrower’s payment capacity, an increase in the installment due to higher rates terms would increase the probability of default.
- Reduced population considered: The lower-risk applicants in a population do not take up the offer (because they have access to better offers), but the higher-risk applicants do, and so make up a greater percentage of the population. Say you’re making an offer to a population with historical good:bad odds of 10:1. With a strong offer, you may get the 10:1 odds from the people who take the credit. If your offer is not seen as competitive, though, people who are receiving other offers will skip it, but the riskier people in your population (who find it harder to get credit) will take it up. As a result, your good:bad odds for the booked population could be 5:1 — which will be an unpleasant surprise down the road.
We’re working with a lender who recently ran a test, launching 60,000 direct mail offers for credit. There were two offers, randomly assigned to prospects: standard interest rate and standard rate + 2.8%. Here’s the take-up on those offers, and the performance 24 months after booking:
Lenders should be mindful of wider impacts when increasing pricing for specific sub-populations above the industry standard. Best-in-class banking organizations are developing pricing strategies using decision modelling and optimization, which can ensure they make the right offer to the right customer in order to increase portfolio profitability, while mitigating (or forecasting for) adverse selection.