Lenders rejected over 14% of auto loan applicants between January and June of last year, up year-over-year, according to the Federal Reserve.  -  IMAGE: Asm Arif

Lenders rejected over 14% of auto loan applicants between January and June of last year, up year-over-year, according to the Federal Reserve.

IMAGE: Asm Arif

Vehicle shoppers have been hit hard since the start of the pandemic. Low inventory levels and skyrocketing prices for new and used cars have been replaced with high interest rates, tightened credit, and lending delays.

Now it’s become harder than ever to get an auto loan and takes longer than ever for dealers to get paid.

Lenders rejected over 14% of auto loan applicants between January and June of last year, up year-over-year, according to the Federal Reserve, and although access to auto credit improved in August for all types of loans, Cox Automotive says, it was still tighter than a year earlier and, for many channels, than before the pandemic.

Meanwhile, dealers can wait up to 30 days for lenders to pay them for a vehicle they already sold, says Justin Wickett, CEO of Informed.IQ, a company that helps lenders process and fund loans more quickly.

In a climate that tough, dealers can use all the help they can get. Technology is one way to help lenders get applicants with thin credit files into automotive loans at reasonable rates, says Equifax Automotive General Manager Lena Bourgeois].

“Interest rates have gone up. Car prices have gone up. And supply continues to be a problem. So, the issue around affordability has become a hot topic,” she says.

Using new data sets in lending decisions is helping lenders navigate the chaos. Aiding the change is artificial-intelligence automation, which is also helping expedite lending decisions and payments to dealers, according to Wickett.

Equifax, for one, has created additional data sets that it says expand the ability to assess consumer risk by 16%, Bourgeois says.

Lending Considerations

Though consumer credit files tend to be very similar across the three major credit-reporting agencies Equifax, Experian and TransUnion, many consumers are under-represented in them, Bourgeois says.

“Those consumers may struggle to get a loan.”

Consumers with nonexistent to thin files lack a credit history to base a loan on, she says. They may include younger consumers who’ve never taken out a loan or consumers who aren’t active in the lending market from a traditional credit perspective.

“They also could be consumers who are trying to avoid traditional credit,” Bourgeois says. “When they make a big purchase, like an auto purchase, they want to take advantage, but their credit report might not represent their ability to meet financial obligations in the traditional sense.”

But, she says, “Lenders must look deeper than just traditional credit when making lending decisions. There is a vast consumer base that have no files or thin files compared to traditional credit bases [who are also good credit risks].”

Now credit reporting agencies offer additional data to help lenders assess applicants’ credit risk to make fair lending decisions. For example, Equifax can provide additional insights alongside the traditional credit report, including consumer payment history on telecommunications and utility accounts.

“This gives lenders a lengthy, deep history of consumers meeting their financial obligations that they can consider when accessing risk,” Bourgeois says.

She advises that the time is right to leverage the alternative data sets, especially in the automotive lending space. “It is usually the consumers in the subprime segment that are most impacted by prices and interest rates. New shoppers are also negatively impacted.”

Credit Tightening

There are three primary types of credit applicants: prime, subprime and deep subprime.

A person with prime credit has FICO credit scores that range from 670 to 790+. Those are borrowers most lenders are eager to extend credit to. A subprime consumer falls just below prime consumers at 580 to 669. They may still get a loan but will pay a higher interest rate than a prime candidate. Deep subprime is a term reserved for consumers with credit scores below where most lenders will extend credit, for example under 580.

Credit tightening is impacting borrowers in the subprime and deep subprime categories, Bourgeois says.

“Mass lenders,” she explains, “have tightened the volume of applications they approve. They have tightened credit access in the subprime or deep subprime levels because they are trying to manage their risks.”

That means lenders are examining and reviewing their portfolios more carefully to uncover ways to reduce their delinquencies.

Though a temporary situation that will improve when economic conditions brighten, more data is needed to make sound lending decisions, Bourgeois says.

“The market was performing well, interest rates were reasonable, and cars were scarce, so every deal was important. Now lenders have pulled back so they do not assume too much risk.”

Loan applicants with low scores don’t necessarily pose undue risk, Bourgeois says, nor do those with high scores come without risk. In fact, lenders realize that someone with a higher score can sometimes pose more risk than a subprime borrower.

“Just because they have a lot of credit, doesn’t mean they are great at managing their finances,” she says. “In fact, we’ve seen some subprime lenders try very hard to manage their expenses and household financials. Many of them have the financial headroom to take on a $750 to $1,000 payment. It just wasn’t necessary for them to take on that incremental payment.”

But lenders must be able to differentiate between the subprime borrower who has a lower credit score because they do not pay their bills and one who is more fiscally conservative. To cut through the stereotypes and the data, AI can help.

 “Alternative data can help differentiate consumers in this segment to allow lenders to more properly assess credit risk than they can with traditional credit reporting,” Bourgeois says.

What is Alternative Data?

Alternative data refers to financial information that hasn’t traditionally been reported to the major credit bureaus. It can include information like rent payment history, gig-economy income, utility and cellphone payments, childcare payments, subscriptions and more.

The data provides a more holistic picture of a borrower’s finances. Lenders can see spending patterns, access account information from consumer bank and credit card accounts, and use the data to evaluate and expand the view into consumer creditworthiness.

“If you have a consumer that has paid their mortgage and auto loan, but hasn’t had an auto loan in 15 years, you can look at their other obligations and see they have never missed a payment. Now you are looking at a consumer who poses little risk,” Bourgeois says.

Without such data, lenders may still approve that type of borrower, but it will be at a higher interest rate. “Now lenders can approve them at a lower interest rate,” she says. “It all depends on lender dynamics, the algorithms they use, and their own risk appetite.”

Automated Solutions

Machine learning and artificial intelligence allow lenders to easily ingest and optimize alternative data alongside traditional credit data. Combining data sets gives the algorithms greater ability to isolate nuanced risk and find pockets of opportunity. 

“The data in their consumer database is the historical data used to train the system in their algorithms,” Bourgeois says.

Equifax uses those types of automated solutions to build in a real-time approval method. Now, when a dealer sends an application to their lender, they can get an immediate response. “They can send a response to the dealer that says, ‘We can extend a 72-month term at a 4.99% interest rate, and this is what the payment will be. If alternative data must be used, this all happens seamlessly on the back end. This data helps dealers get a ‘yes’ more often.”

ML and AI are automating lending in many ways and are tools more dealers are asking for, Wickett says.

“More and more dealers are asking lenders to automate loan funding besides automating underwriting to get approvals more quickly. If a dealer cannot get paid, that’s a problem. The good news is many lenders are investing in a digital transformation to speed up their lending operations and take better care of their dealers.”

Informed.IQ speeds payments to dealers after loans are approved and the consumer drives away. The company says it works with six out of the top 10 lenders serving U.S. automotive dealers.

With or without automation in place, dealers upload auto loan documents into a portal. Some uploaded documents are fully digitized, while others are just scans of paper documents. Without AI, the documents end up in a manual review queue at the financial institution. If further documentation is needed, the process gets kicked back to the dealer to upload. Then the file returns to the lender’s manual queue. Without AI automation, that means dealers may wait 30 to 60 days for payment on a vehicle they already sold, Wickett says.

But when lenders leverage AI, the system automatically reviews documentation in real time, asks for additional documentation if needed, and gives dealers immediate confirmation that they will get paid.

“The dealers do not have to change their practices,” Wickett says. “But because their lender has adopted AI technology, they can get paid faster. Lenders can use AI to verify a driver’s license or tally up earnings on a pay stub, leaving their personnel to focus on the relationship work and give it the attention it deserves.”

The technology also helps dealers with regulatory compliance, where the rules are always changing. “Regulators now mandate that dealers present more disclosure documents to car buyers than ever before,” Wickett says. “There’s more to verify, and all these verifications take time and slow the process down. Lenders that don’t embrace the digital transformation for dealers are going to be at a disadvantage.”

As lenders dip their toes into AI technologies, Wickett offers some advice. “Lenders must be very cautious about how they deploy AI in underwriting decisions that could impact fair-lending laws or the Equal Credit Opportunity Act.”

To that end, he suggests embracing AI to address rote tasks and manual workflow and being mindful about “how to best incorporate it into your organization.”

He recommends picking partners with AI experience in the financial sector.

“These vendors know how to navigate the lending landscape, where AI can and cannot be applied, and where there are regulatory landmines to steer clear of. Don’t build this technology yourself. Lenders must be sensitive to risk management, governance and other things when using this technology. You need to be an expert.”

The right technology partners, he says, can help lenders use new data sets and AI automation to their advantage to expedite lending decisions and payments to dealers.

Ronnie Wendt is an editor at F&I and Showroom.

 

0 Comments