By W.B. King
Calling it “a reductive approach to evaluating auto loan potential,” a recent Open Lending report found that financial institutions could be placing “too high a value on FICO scores” at the exclusion of alternative data analysis solutions.
“FICO scores should absolutely be one factor in your auto loan decisioning, but you must use debt-to-income ratios and alternative data for a more holistic picture of risk,” the report noted. “For instance, alternative data can illuminate non-traditional trade line payments, like delinquent mobile phone bills, that can help paint a future view of a borrower’s credit worthiness. If you’re not using a combination of data sources that require machine learning and artificial intelligence (AI) capabilities for auto loan decisioning, you’re putting your institution at risk.”
Loans Within Reach: Lending Enablement Benchmark 2023 surveyed 95 auto lending leaders. All respondents were based in the U.S. with a director-level title or above. Of the respondents polled in February 2023, 42 work for credit unions, with the remainder from other financial institutions, including retail/commercial banks (36 respondents), community development banks (5), online banks (3), insurance companies (7), and captive finance companies (2). Respondents included notable subsets across financial institutions of all sizes, based on annual revenue.
For the past 20 years, the Austin, Texas-based Open Lending has provided loan analytics, risk-based pricing, risk modeling and default insurance to auto lenders throughout the United States.
“The auto lending industry is facing a uniquely challenging year, with rising interest rates and the high-profile failure of several banks driving concerns about a banking crisis and continued economic instability,” the report continued. “While the landscape of consumer lending isn’t always directly impacted by a possible banking crisis, the overall financial environment is more tenuous, and that precarity is highlighting how lenders are combatting risk while increasing the volume of loans they issue.”
For banks and credit unions, and auto lenders in particular, the report noted that the use of “lending enablement solutions” can be a game changer. These solutions include AI and machine learning platforms that work in tandem with FICO scores to determine creditworthiness.
“Our most recent research reveals that lending enablement solutions are playing a clear and decisive role in improving decisioning speed, increasing ROA (return on assets), and reducing risk exposure,” the report noted. “Additionally, the use of lending enablement solutions is allowing financial institutions to better reach near-prime consumers — an audience that’s strategically critical to serve, especially in an economic downturn.”
Minimizing the Denial Pile
According to the report, the top five reasons auto loans are denied to potential borrowers are: bad credit history (71%), high debt-to-income ratio (61%), low credit score (53%), income too low for desired loan amount (48%) and unstable employment history (42%).
Respondents biggest complaints about current auto loan decisioning model are: doesn’t allow for customization (11%), current model offers low visibility into decision factors (11%), disqualifies too many applicants (14%), zero or limited after-hours availability (18%), and decisioning time is too slow (21%).
“Beyond slow decisioning time, a notable subset of auto lenders say their current models are disqualifying too many applicants, and also offering limited visibility into why these disqualifications are taking place,” the report found. “When we pair this finding with our research revealing an overemphasis on FICO scores, it’s clear the auto lending process is encumbered by an overly myopic process that’s leading potentially qualified and credit worthy individuals to end up in the denial pile.”
While 65% of respondents indicate that they are using a lending enablement platform, 96% of those surveyed said they’re only “satisfied” with their current solution. The largest portion of respondents — 52% — characterize themselves as only “somewhat satisfied” with their existing lending enablement platform, the report noted.
At 44%, respondents said improving loan-decisioning speed is their top priority. The same percentage was cited for growing ROA. Forty-two percent want to reduce risk exposure and 41% want to increase loan volume.
Is AI the Answer?
Moving forward, respondents said they will seek the following three improvements from lending providers: greater transparency and expanded features, increase use of AI/advanced analytics and better security.
“AI and predictive analytics are poised to disrupt every industry, and lending enablement is no exception to this trend. However, when financial institutions are vetting potential lending enablement solutions, they should be wary of how potential providers are defining AI,” the report stated. “ Some providers — particularly newer ones with more nascent offerings — will use the term more as a marketing device than a core and differentiating feature of the product.”
Bank and credit union leaders, the report recommended, should inquire whether the AI provider only uses the institution’s data or a broader set of federated data.
“The history of data the provider has experience with is also critical, as the algorithms that build into any AI system need to account for changes that come with how consumers, vehicle valuations and loans perform at different points in an economic cycle,” the report continued. “It’s important for financial institutions to holistically evaluate prospective lending enablement solutions and look for the specific ways AI and machine learning are being integrated.”
Comments