Banks, credit unions and captive finance companies have tended to shy away from underwriting subprime auto loans, leaving that side of the business to more specialized auto financing companies. This choice was based on the reality that the applicants are at higher risk than other borrowers, and instances of late payments and default are far more common than in conventional loans.
With the recent introduction of AI and machine learning (ML) and their ability to quickly ingest and analyze enormous swaths of data, lenders have the opportunity to obtain a much clearer picture of applicants to make better decisions. This should have enormous benefits for lenders that serve or plan to serve the burgeoning subprime markets, which translates into tens of millions of customers in the United States and Canada.
Over the past two years, we’ve seen an influx of alternative data providers that use AI and ML to redefine the identity, income and employment verification processes, giving lenders far better insights into determining creditworthiness and the propensity for making payments on time. These providers garner data from primary and secondary sources that go well beyond traditional credit bureau reports. Bureau reports were designed to serve a broad range of industries and use cases, relying on a limited set of data and intending to provide a simple readable summary of a consumer’s complex circumstances.
Alternative data and AI leverage modern computational resources to harvest billions of data points and make inferences that weren’t possible a few years ago. By tapping into public and private data sources like motor vehicle records; rent and mortgage history; credit card payments; and utility and mobile phone bills — and cross-referencing the data in real time — lenders can make decisions on hard data and not rely on broad-brush assessments based upon overly simplified data inputs.
Subprime market: Two sides of a coin
The advent of alternative data for subprime loan decisioning could not come at a better time. In fact, subprime lenders are the fastest adopters of these features. As interest rates continue to climb, so does revenue since high interest rates are attached to subprime auto loans. This data is borne out by research conducted by Ibis World. In its 2023 report on the subprime market, it found that revenue from auto loans in this sector has climbed steadily at 3.5% since 2018, with a 21.1% surge this year. But subprime lenders are finding out that more revenue does not equate to profitability, especially with today’s higher cost of capital. Margins are squeezed and subprime lenders feel the pinch.
Not surprisingly, these higher interest rates can cause havoc among consumers. Borrowers have a harder time getting approved, and those who already have subprime loans statistically have a more difficult time keeping up with their payments. S&P Global found this to be precisely the case. Current default and delinquency rates are at 6%, a rate not seen in nearly 25 years.
So, the question must be asked: Is it possible for lenders to profitably serve the subprime market, and take advantage of the higher interest rates, without subjecting themselves to runaway delinquencies and defaults that jeopardize the integrity of their financial institutions?
The answer is an emphatic “yes,” but only if lenders really take advantage of the insights available from alternative data sources to identify borrowers who are truly creditworthy, and don’t rely on the archaic, error-prone methodologies associated with traditional credit bureau reporting.
In the past, financial institutions had little choice but to review applications and supporting documents by hand, a process that all would agree is laborious, error-prone, inconsistent — and incomplete. Today, this data can be harvested from all sorts of sources and then integrated into the lender’s LOS through APIs. This enables institutions to properly score and decision the application according to their business practices and preferences. These solutions also benefit auto dealers and consumers, who want fast decisions. Speed and accuracy are the name of the game in lending, and alternative data reduces subprime decisioning from dozens of hours to minutes. The large credit bureaus recognize the appeal of alternative data, with the major providers bringing these services in-house over the past year.
Minimizing defaults, delinquencies
Along with helping evaluate creditworthiness, alternative data is ideal for notifying lenders about potential delinquencies. Powerful AI engines scour the internet for relevant information such as a change in employment status or a late payment to another vendor. These activities can trigger an alarm to an auto lender and allow the financial institution to proactively reach out to the borrower to inquire about the situation. The lender can then offer payment options or other services to help avoid delinquency.
This feature is perfect not only for subprime activity but for all lenders that must deal with a softening economy. In every credit tier there are borrowers at risk. Alternative data enables lenders to find potential issues and address them before they turn into even bigger problems.
For years, subprime auto lending was regarded as the Wild West of the industry. There were significant rewards, to be sure, but they came with even greater risks. The introduction of AI- and ML-centered alternative data has changed this paradigm. By accessing data from primary and secondary sources, lenders can make more informed decisions, reduce the incidence of defaults and delinquencies, improve cash flow and succeed in a highly profitable, expanding market segment.
Bob Metodiev is head of business development at Inovatec Systems Corp., which provides cloud-based loan origination and loan management solutions for automotive, power sports, equipment and other lenders.
Auto Finance Summit, the premier industry event for auto lending and leasing, returns October 29-31 at the Bellagio Las Vegas and features a fireside chat with Ford Credit. To learn more about the 2023 event and register, visit autofinance.live/afs/.