FICO @ AFSA: Auto Lenders Losing $2-3B to Fraud

One of my favorite movie scenes is from Blade Runner. A policeman is interrogating his detainee, asking a series of questions to determine whether the person is a real human-being or a synthetic.

Funnily enough, this scene parallels the problems discussed at a FICO roundtable on ‘How to Fight Auto Loan Fraud’ held this week at the AFSA Vehicle Finance Conference & Expo.

While none of the 25 attendees resembled Harrison Ford, they agreed they were getting busier looking for synthetics, just a different kind of synthetic. In this case, we were talking about synthetic identities, a type of fraud in which a criminal combines real and fake information to create a new fake identity that is used to open fraudulent accounts and make fraudulent purchases.

The use of synthetic IDs in the auto loan industry has been a growing problem in the last 18 months. Attendees at the roundtable told us that, amongst other drivers pushing fraud their way, they felt part of the problem was the use of originations systems that require fields such as social security numbers (SSN) in early stage and initial runs, when dealers often enter random numbers rather than inconvenience a customer. The misuse of SSNs from children and the elderly was also seen as a growing loophole for fraud and abuse. Attendees felt first party fraud was further compounded the by the multiple paths for material misrepresentation such as the use of:

• Digital image and document editing to falsify paystubs
• Non-existent addresses that aid in skip trace fraud
• Falsified data critical to underwriting – such as vehicle information - on official documents.

In the discussion, some lenders told us they have started using alternative data in the form of social media to confirm identities. An example would be scanning a loan applicant’s Facebook profile to check they look like a ‘normal’ person with a ‘normal’ social circle and then giving them a trustworthiness rating. A few people in the group discussion felt this technique may eventually fall foul of privacy laws, as well as pose a reputational risk to firms with consumers. FICO believes there is already enough rich lending data available to scrutinize application fraud to avoid this approach.

Another point raised in the discussion was concern around internal and dealer fraud. In fact, one lender felt that half of their fraud was dealer fraud, and that data matching or analytics would be useful to spot collusion.

Lastly, the majority of attendees agreed they felt they still catch the majority of auto loan fraud through manual verification and common sense processes, but they don’t always have a good grasp on how much fraud is out there or what technologies can help them automate their manual processes. Recent estimates peg the industry losses at between $2-3 billion annually, so the problem is considerable and still growing.

Luckily, FICO has some experience in solving ‘big data’ problems. The company’s Falcon® Fraud Manager has found global success with banks using its market-leading predictive analytics to stop card fraud by looking for out-of-pattern transactions.

Similarly, FICO has been developing its network analytics capability for identity and entity resolution as a killer technology against application fraudsters using synthetic IDs. The solution works by reviewing the relationships between people, entities and data to reveal discrepancies in a matter of minutes and it can uncover collusion and fraud rings within seconds.

To learn more, check out our webinar on: Why Auto Loan Fraud is a Growing Problem – And How to Fight It