Auto loan fraud losses rose to $6 billion last year, doubling from 2016, according to fraud detection specialist PointPredictive. The company is launching products to help lenders catch fake identities and learn from the fraud experiences of their peers.
The Fraud Data Exchange, which PointPredictive plans to roll out next week, gives members of PointPredictive's auto lending consortium access to a pool of data from confirmed fraud cases, while the Income Validation Alert, which launched this week, helps verify the listed income on consumers' credit applications.
With the Fraud Data Exchange, lenders can cross-reference loan applications with confirmed fraud cases. Incoming credit applications will be scanned against data points -- such as phone numbers or addresses -- historically associated with fraud. The software then reports a fraud score on a scale of 1, indicating a low risk, to 9.9, the highest risk, along with three reason codes indicating the type of fraud suspected.
A fraud score is more difficult to compile than a traditional credit score, given the volume of fraud entry points into the process, said Frank McKenna, the San Diego company's chief fraud strategist.
The system flags suspicious or potentially implausible information, such as an income too high relative to the borrower's occupation or a far distance between the borrower and the dealership, McKenna said. Still, it takes a combination of factors to produce a high risk score. Data in the exchange comes from repeat offenders, and not one-off cases, McKenna told Automotive News.
"The score is based on lots of interrelationships between a lot of the data," McKenna said. For instance, if the borrower's address is 2,000 miles from the dealership, "that might be the icing on the cake."