Solving the “blind eye” problem of loan fraud

0

Almost every consumer knows the pain of filling out applications for auto loans, mortgages, and other types of loans, having to produce pay stubs or W-2s to prove that your income is high enough to earn. the purchase in question. Additionally, virtually all lenders and finance providers experience the frustration of verifying that the information is correct, while hoping that a fraudster has failed to get around the defenses.

However, the best criminals are nothing if they are not determined and resourceful, and like two Predictive Point Executives chatted with Karen Webster in a PYMNTS conversation on Thursday (September 12), pay stubs are easy to forge, and scammers can make a good living from vehicle theft or advanced money theft for buying a house. That said, digital technology and so-called big data promise to make the jobs of these criminals harder in the future, at least according to the story told by Tim Grace, CEO of PointPredictive and Frank McKenna, the chief strategist. of the company in matters of fraud.

Blind eye problem

As Grace described, much of the problem stems from what he called a blind eye. With lenders trying to authenticate so many documents – and do so within a reasonable timeframe so as not to cause friction for potential borrowers – it is pretty much inevitable that some fraudsters will go through the doors. Indeed, according to Grace, about 1 in 12 car loan applications involve some form of income fraud, that is, the applicant exaggerates the amount of income in order to access a larger loan, which can in turn result in a vehicle that can be shipped overseas and sold for a healthy profit, or a vehicle that may be difficult for repo professionals to track down, given that the fraudster also likely used a fake ID.

“For $ 5 in four minutes, you can create a pay stub to say you’re doing anything,” McKenna told Webster. Indeed, a Google search reveals several sites offering such services. Sometimes scammers are incredibly brazen, at least from an honest consumer – McKenna’s company validates loan applications, and he has talked about checking bank statements of applications and seeing clear fees for these sites. fraudulent pay stubs. “Just checking documents doesn’t work,” he said.

So, how exactly do you protect yourself against such fraud, and this blindness problem? It’s all about data, combined with machine learning, said the two leaders of PointPredictive.

“We have a data consortium,” Grace said, adding that the company’s data comes from eight main sources – including auto lenders and others, and payroll databases and other companies. This data collection, in fact, includes over 35 million auto loan applications, which contain vital information about work history and income levels. “We also know the borrowers’ zip codes and can link the income to the zip codes,” he added, giving an example of how the company’s fraud detection model works.

Detect lies

The idea – using all this data, these mathematical models, and machine learning – is to get a clear idea of ​​when a claimant is lying about their earnings and flag those claims while reducing documentation and verification for them. honest and potential borrowers. Specifically, the technology is designed to identify when an applicant’s assumed income is 15% or more than it should be in that particular situation, according to predictive modeling, Grace said. He told Webster that the model can erase 75% of applications with an accuracy rate of 90-97%. This leads to reduced friction for all parties involved – lenders, sellers and consumers. The remaining 25% is what lenders need to take a closer look at, as fraud can occur in this segment.

Those rates could improve over time as more data is added to the fraud detection system, McKenna said. “We can be more predictive on capturing fraud,” he said.

Additionally, as more data becomes available, Grace said the system takes into account the economy of odd jobs, i.e. relatively irregular sources of income for potential borrowers. “Twenty-five percent of our data in our consortium comes from independent borrowers,” he said.

There are differences in fraud between various industries – income fraud with respect to mortgages, for example, tends to be more subtle and sophisticated, as mortgage lenders have 30 days to approve or reject. loans, compared to around 15 seconds for auto finance providers. , said Grace. However, in general, he said, the broader tactics are quite similar.

More data, of course, means better predictions. Additionally, better forecasting can mean less paper and documentation for honest consumers – much less friction when it comes to securing vital loans.

——————————

NEW PYMNTS DATA: MAKING LOYALTY WORK FOR SMALL BUSINESSES – UNITED KINGDOM EDITION

About the study: UK consumers see local purchases as essential for both supporting the economy and preserving the environment, but many local High Street businesses are struggling to get them in. In the new Making Loyalty Work For Small Businesses study, PYMNTS surveys 1,115 UK consumers to find out how offering personalized loyalty programs can help engage new High Street shoppers.

Leave A Reply

Your email address will not be published.