The Next Generation of Auto Buying and Borrowing: A Conversation with Kevin Filan

Credit unions have long been reliant on consumer lending, particularly consumer lending for automobiles, for much of their loan portfolio. According to Experian, credit unions account for 22% of all auto financing, after leading the market in 2022.?

Despite this market strength, auto loan growth has been challenging as of late. At year-end 2023, auto loans represented 31.1% of loan balances, a decline from 35% in 2018. Over those five years, credit unions have grown their auto loans by 7%, while other types of loans such as mortgages, equities and credit cards have in aggregate grown by 11%.?

Why the slowdown? One reason is risk management. The young consumers looking to borrow to purchase a car often fall into near-prime and non-prime credit ranges, ranges that many credit unions shy away from because of the perceived credit risk. These same credit unions then struggle to attract younger members.

Ultimately, the key to lending to the next generation of borrowers may lie in understanding how credit unions can effectively lend while managing that credit risk.

I sat down with Kevin Filan, Marketing SVP of Open Lending, to discuss how the needs of near-prime and nonprime borrowers and the needs of credit unions align so well.

The Auto Loan Cycle: Affordability, Liquidity, and Risk?

Caroline Vahrenkamp (CV): What in your view is the biggest issue for near-prime and non-prime prospective borrowers trying to get a car loan?

Kevin Filan (KF): I think there's a there's a vicious cycle at play here. If you think about it, you need two things to book an auto loan: First, you need a consumer who says “yeah, based on price of the vehicle, based on my own household budget, I can afford a car. I can afford taking on the debt.”?Well, inflation has made those calculations more difficult. New car and used car prices have always risen year over year, but what happened when we hit that shortage of inventory during the pandemic and really limited day supplies sitting on dealer lots? The prices went crazy high, and they've never come back down.

We reference the Mannheim Used Value index, which sets 2019 prices at 100.?That index was as high as 210 or 220, but it is still around 200 now. That means used car prices are around double what they were only five years ago.

CV: And these aren’t frivolous purchases, typically.

KF: That’s right. There are things that necessitated them having more reliable transportation, newer transportation, or a different capability in that transportation. “Need” tends to trump “Want” when it comes to why someone goes into the market.

CV: And the other half of the vicious cycle is?

KF: The other half of that vicious cycle is, as you know, liquidity and perceptions of risk. You know the industry; it was a wakeup call when Silicon Valley Bank and First Republic and overseas, credit Suisse all failed. Lenders were already contending with liquidity challenges, and now, because of these bank failures, the regulators putting more emphasis on CECL requirements to avoid repeats of what was happening in early 2023.

CV: All while the excess deposits that came in during the pandemic began to dwindle away.

KF: Exactly! So as liquidity started to get drawn down, lenders looking at where they were lending.?

Credit unions were starting to make decisions.?Some were pulling back from auto lending in general. Others were just pulling back from indirect lending. Many were continuing indirect but were pulling back just into prime lending because they were seeing increases in their 60-day delinquencies.

CV: And while options are decreasing, the interest rates are rising.

KF: And the net outcome of those two forces intertwining is that near prime consumers are a significantly smaller percent of buyers in the marketplace.

CV: It’s hard for those borrowers.?They need the vehicle, but the costs for both purchasing and borrowing on top of it are so high; the interest rates that they're getting now on those on those near prime loans are just astronomical, at least relative to recent experience. I always have to kind of preface this because what I consider astronomical is not what someone who lived through the 80s would consider astronomical.

Buying Cars with Gen Z and Millennials

KF: And that’s an issue for young borrowers particularly. Gen Z and millennials are opting to pay in cash where they can or opting to lease cars at a rate that is about four times the rate that Gen X and older segments lease. If you think about the millennials, almost their entire adult lives, cash was almost free. They came into adulthood as financial consumers in a time of quantitative easing, when rates were practically zero. COVID stimulus put that on hyperdrive. Now there is a rebound of tightening credit, elevating borrowing costs. And that is a bit of a shock.

CV: I find leasing fascinating, because I remember feeling like leasing was something you did if you were, for lack of better word, posh. You leased your BMW.?The thought was: I don't need equity in my car. I don't need to have value in my car. I'm just going to drive it and then I'll get a new one every time, because I can afford that.?And now, you’re saying we’ve reached the point where leasing is the affordable option??That we've reached a point where it's more equivalent to renting than buying?

KF: Like a long-term rental.

CV: It's fascinating just to see how the auto market has turned into a home market.?

KF: Another piece I was surprised to see was the extent to which they're in 48-month (about 4 year) terms, as opposed to longer-term loans. The monthly payment would be lower if they stretched out the term, so first we thought that younger Millennial and Gen Z borrowers are just more financially cautious about taking on a longer-term loan. But then, we found out that it’s the institutions that are approving loans for 48 months but not 60 or 72.

CV: From a CFO’s standpoint—the odds are that they're going to pay down more of their principal in 48 months than they will over the course of 72. So, if they default, they're less likely to be underwater. There’s a willingness to take credit risk on a shorter term because the overall dollar exposure would be less.

KF: Another finding was that more than one in five were paying in cash, and the stated reason was they were dissatisfied with what they were quoted on a loan rate and terms.

I think what we're seeing with Gen Z is different than what has happened with older cohorts when they came through this same period in their lives. What they've experienced in their financial lives until now has been very different, incredibly volatile. And so, as you would expect, they're reacting and interacting with financial services differently when it comes to an auto loan.

CV: So what can credit unions be doing then to make sure that they are able to get into this group? As your research points out, beginning, thin-file borrowers are so much more loyal than previous generations had been. In addition, they show stronger improvements in score, there's fewer charge offs; the loans perform better than the score or the loan grading would suggest.

How can credit unions move past the perceived risk and be able to make these loans, because every credit union I know is saying “We’ve got to have younger members. We’ve got to have younger members…”

KF: But you can't say on a Monday, “I want younger members” and then on a Tuesday be turning them down when they come in and apply for an auto loan.?I understand why a CEO would need to be cautious about whom we're lending to, based on what they're seeing in the macroeconomy in terms of delinquencies and in terms of requirements from regulators. But you've got to be able to use more data than just what's in a conventional credit score to really understand that consumer.

Underwriting: Balancing Accuracy and Fairness?

KF: But here's the trick: The more data you consume, the more time it takes an underwriting team to process an application. That's where lending enablement solutions that use machine learning and AI come in. I know people are almost sick of hearing “machine learning” and “AI”, but their application is what needs to happen for auto lenders to be successful. We need to be able to sift through that data more consistently and fairly to make informed decisions. And consumers want us to do it quickly—that's where a lot of consumer satisfaction comes from.

So how do you do the marry speed with accuracy and fairness? We need a platform that's able to ingest data consistently and fairly, to evaluate creditworthiness, to price and structure it appropriately, and to turn that result around to the applying consumer sitting in the dealer’s office.

CV: Let’s talk about another group of potential borrowers.?Filene has done quite a bit of research on ITIN lending for home loans specifically, and similarly you’re now looking at data that goes beyond income and credit score for auto loan borrowers. What are some things you’re seeing about ITIN auto borrowers?

KF: We looked at the studies that Filene has done on ITIN, and they speak very much the fact that ITIN borrowers actually have higher credit scores than someone would assume. They qualify for higher loan amounts. Their delinquency rates are often less.

One thing we’ve found is that if you look at the vehicles that they're buying, it speaks to a dual use vehicle, one that is used both for personal use and for work. That's a consumer that understands really well that if they don't make the payments on their vehicle, and it’s repossessed, they can't earn income. And so, I think when you start to get ITIN holders and you look at the vehicles that they're in, it is absolutely 100% tied to their ability for their livelihood.?

Obviously, that’s not true in every case, but I think ITIN borrowers would be way over indexed in terms of dual use vehicles when you start looking at those vehicle segments.

CV: Again, it's just another point of data that you could be evaluating as you're trying to determine credit worthiness.

KF: And the other thing I think that's interesting, when we looked at some of the great research that Filene Research Institute has done, was when it comes time for that ITIN holder to matriculate into holding a Social Security card number. They're no longer allowed under law to apply for lines of credit under their ITIN; it has to be under their SSN.?

So, here’s the scenario: We have a set of consumers who are ITIN holders, either resident or nonresident aliens, who are looking for a loan. They take that loan, and they successfully comply with the terms of that loan. Once they get their social security number, you would expect their credit score to go up. Well, even as their banking needs mature, your credit union now has been a significant part of helping what started off as an ITIN holder getting that line of credit for a car or a truck, for building a life. And that's exactly what the Charter is for.

CV: Right, being able to be a stronger part of the community and able to help facilitate that process. Helping the community and helping the cooperative at the same time.

Overcoming the Trust Barrier

CV: One item that really jumped out for me on the report was the trust barrier.?There is a real sentiment among near prime borrowers and non-prime borrowers that they don't trust lenders. They don't trust the people selling cars, but I think we expected to see that. But they don’t trust the lenders either. What can create unions do to improve that?

KF: I think the speed at which you get to a decision matters.?I think the longer that the decision is sitting with the institution, I think diminishes not just satisfaction but trust.

The second piece is being able to look at them as more than just a credit score. And I think that's the frustrating part to thin file consumers new to credit, whether they’re an ITIN holder, or someone who's on a younger cohort, or someone who says, “there were some issues in the past, but I'm well past those issues now, and if you look at the rest of how I conduct myself financially, I am responsible.”

Think of the frustration: they apply, it takes a long time, and then they receive an answer back where no, we're not going to approve you, or worse, I'm going to approve you, but there are conditions, particularly if you're in the near prime space and near prime applications.

These are the applications that institutions routinely tell us that their underwriting teams struggle with. So, what these teams do, lacking more information or the ability to process that information, is move them down the credit scale and price them as if they're subprime.

So, the borrower gets a term or a payment and an interest rate that was not at all in line with what they were expecting or what they feel like they deserve.

CV: Right.

KF: So how do we build trust? We need to understand creditworthiness in a way that traditional credit scores aren't doing a good job of capturing[CV1]?. Fundamentally you have to start there. Then you have to supplement that with additional information and be able to turn that decision around fairly quickly and get closer to getting it right.

CV: I think that fits with my experience with these near-prime borrowers. We would look at C loans and some of these have performed fantastically and we really ought to have given them a better rate. For some loan products, we had worse delinquency and charge offs in our B paper than we did in our C&D paper. We weren't pricing them accordingly because we didn't have this additional data to use.

KF: And if you look at the default rate on 72-month prime auto loans versus near-prime 72-month loans, defaults are often higher on the prime 72 month. It’s understanding why the loan is extended to 72 months. The prime loan is extending to lower the payment, so there are already signs of struggle for making the loan fit their budget, even with great credit. With the near-prime, it’s less about cashflow and more about the interest rate, so there’s fewer capacity issues.

CV: Showing again that there’s so much more to underwriting than looking at a credit score. This is why I’ve been a big fan of CECL, because you actually do have to pay attention to what you’re offering and why and then seeing the effects.

Now that we have tools that allow us to get that full picture of the borrower's financial situation, in theory, we should be able to say we can give you the loan and we can give you the loan at the rate that you deserve.

KF: We've got some more work that will be forthcoming. There's even more we can do to peel back the onion to understand what's happening with different generations, including Millennials and Gen Z. We’ve started to uncover additional insights with the vehicle accessibility report?that we’ve been working on.

There’s an old saying about how at the top of the mountain, you’ll find more mountains. That’s what we’ve experienced. You get to the top of this mountain of data analysis, and what you get is a line of sight to other mountains that you couldn’t have seen until you got to the top of this one. It starts to tee up other questions.

We think it’s important for credit unions to have a line of sight to insights and questions that they wouldn’t be able to see otherwise. In this case, we want to play our part in helping the industry understand the near-prime and nonprime auto borrowers better, so they can do a better job of meeting their needs.?

It’s incredible how every time we start asking questions about underserved consumers, we learn about the impact of serving those consumers. In your study of ITIN borrowers, for example, 95% of the borrowers said that ITIN loan had a positive impact on their life. That’s incredible. You buy a good or service, especially an expensive one, and you’re asked, do you regret it, or did it have a positive impact on your life, and almost everyone says it had a positive impact? That’s unique. At the same time, it isn’t entirely altruistic for credit unions to serve these consumers. You want to do it for the right reasons, but there’s also a business case. Credit unions can do well by doing good.

CV: There is absolutely something to that. And to leverage insights to bring those two goals together is significant. We have all this data sitting in these data warehouses: let’s use it to have a positive impact on both the community and the bottom line.?

Thank you so much for sitting down with me, Kevin!

— CV