Upcoming current expected credit losses (CECL) requirements will call for data collection of both loan-level and industry-wide information to create accurate models, experts from Wilary Winn advised during a breakout session Monday at the CUNA CFO Council Conference in Orlando.
Focusing on vehicle loans, Douglas Winn, president of the financial advisory firm, recommends grouping loans into four major categories due to differing expected credit performance: new vehicle direct and indirect, and used vehicle direct and indirect.
These categories should be further divided into FICO bands as historical credit losses vary significantly by credit score range, Winn says.
The more granular and, therefore, more predictive a financial institution makes a pool of loans, the fewer historical loan defaults it will contain, Winn says.
He notes that few credit unions have sufficient numbers of defaulted loans to be statistically valid.
“The more granular you are the more the predictive it is, but the less likely you are to be able to infer credit losses from the data,” Winn says.
As result, he advises credit unions to combine their results with industry-wide data.
For example, to be statistically accurate, a financial institution would need 3,700 loans in the 660 to 719 FICO score group, for which there likely would be 24 defaults.
►Click here for more conference coverage from CUNA News, and get live updates on Twitter via @AdamMertzCUNA, @cumagazine, and @CUNACouncils, and by using the #CFOCouncil hashtag. Learn more about the CUNA CFO Council, a member-led professional society for credit union executives, at cunacouncils.org.