Lending Perspectives: There’s No Such Thing as a Free (Car) Ride

raised arms while driving car at sunset
Bill Vogeney Photo
Chief Revenue Officer
Ent Credit Union

6 minutes

Assess how your future losses may vary from recent history.

Whether you believe used car values are declining due to high interest rates, an improving (or worsening, depending on your perspective) inventory of new cars, or a slowing economy, the seemingly “free ride” car lenders have enjoyed for the last two years is nearing an end, if it’s not already over.

All auto lenders have benefitted from strong car values the last few years, although the market is changing rapidly. What risk does your credit union face? Here are some thoughts on a methodology to assess how your future losses may vary from recent history.

What Free Ride?

I suppose the Consumer Financial Protection Bureau thinks that lenders have enjoyed a free ride over the last few years as used car values have been at all-time highs. Anyone with a slightly more-than-casual understanding of the lending business understands that 99.99% of lenders look at repossession as a last resort. I can count on one hand the number of cars Ent has repossessed in 20 years where we got every penny owed by the member at sale. We don’t want to repossess but sometimes we don’t have any other choice to collect a portion of the loan balance. Clearly, we haven’t had a “free” ride, but for sure we have enjoyed a reduced-price ride the last few years.

It’s very clear that lenders have benefitted from reduced car loan losses because their repossessed cars have gotten record prices at auction since late 2020 due to new car supply chain issues and the resulting increased demand for used cars. 

For example, earlier this year we had a month when we recovered more than 80% of the average balance at auction—a high water mark for Ent. (Just as a note, that metric also includes refunds from dealer-sold insurance products) Translation: That 80% recovery ratio means that on a $20,000 balance, we sold the car and had other money recovered to the tune of $16,000. Historically, that percentage tends to range in the high 50s to low 60s for our portfolio; your results may vary based on the average age of your car at the time of the loan, the term of the loan and average loan to value.

This recovery percentage is something I’ve watched closely for 15 years; it’s become my proxy for the state of the used car market. My memory is good enough to recall that after the financial crisis in late 2008, Ent experienced a low water mark of 36% in early 2009 for reasons that should be obvious. 

What Does This Really Mean for Future Losses?

I am a “back of the napkin” math guy. I may have driven one of my former peers to the verge of insanity not only because of this character flaw of mine, but I’m also pretty darn close with my quick estimates. I like back-of-the-napkin math because it often allows me to take a complex problem and state it fairly easily, allowing for my better understanding of the problem. It also helps explain the problem to others. More on that later. 

Using some approximate numbers, let’s assume that our average auto loan balance is $20,000 at the time of repossession. I already said our high-water recovery ratio was more than 80%; we’ll probably average 70% in 2022. This means we’ll lose 30% or $6,000 per loan on average once the books are closed on 2022 (we need 60-90 days to properly account for late-year repossessions). Also, half of these cars repossessed were likely financed pre-COVID-19. That’s important to keep in mind because they weren’t bought at inflated prices.

It sure seems as though what we’re experiencing in the auto market and the overall economy will cause used car values to normalize (if that hasn’t already occurred), whatever normal means these days. If I use a 55% recovery rate ($11,000 sale proceeds) as a normal recovery number for 2023, that means we’ll write off 45% of the balance on average, or $9,000, an increase of $3,000 per car. That increase represents an increase in loss of 50% per loan, meaning loan loss ratios should increase by roughly 50% in 2023.

Now, for the Big ‘But’ …

But I don’t think we’ll see 2023 return to a normal recovery level, and here’s why:

“Historically, that percentage tends to range in the high 50s to low 60s for our portfolio; your results may vary based on the age of your average car at the time of the loan, loan term and average LTV.”

I shared this with you earlier, and my words will probably ring in your ears by the end of the year. This recovery percentage can be used as an indicator of the used car market, yet it can be influenced by other factors, including the average LTV. 

While the late-2020 to late-2022 auto market may not have seen Ent’s average LTV on new loans go up (because trade-in values also increased plus buyers had more cash down), the values (or more correctly, prices) on used cars certainly went up. But how much? Based on indexes from the Manheim auction and other sources, we think used car values increased 25% to 30% from 2020 to late 2022. So perhaps a car financed at 100% LTV in 2021 isn’t really 100% in 2023.

Here’s what I mean by that: Let’s go back to that historical average of a $20,000 balance at the time of repossession and my back-of-the-napkin math. Indexes like Manheim’s are fairly complex, based on literally millions of data points. It’s a lot easier for me to go back to my own data and experience. 

For us, I’ll use an average original loan amount on repossessions, pre-COVID-19, of $25,000. That number is an approximate figure used to simplify this illustration that factors in new and used cars. That loan amount for these purposes represents an average LTV of about 100%. Let’s also assume our average car loan was backed by a 3-year-old Chevy at origination. Well, during the hot market of 2020 to 2022, based on these market experts, that Chevy was probably selling for $31,000-$32,000 based on the 25% to 30% increase in prices compared to what a 3-year-old generic Chevy sold for pre-COVID-19.

But now, that Chevy is probably back to normal values, which means we’re likely to get that $11,000 I shared before, which was an approximate historical value or recovery. But the member is going to owe more on their loan because they are financing more. Starting at a loan of $32,000 during the hot market, the member is likely to owe closer to $25,000 once they fall behind on their payments. 

Now, the loss is likely to be $14,000 ($25,000 balance minus $11,000 proceeds), not the $6,000 we “enjoyed” in 2021 and 2022, and not the $9,000 we might have lost pre-COVID-19. The recovery percentage on this typical scenario isn’t the 70% we achieved in 2022; it’s more like 44%! ($11,000 realized from sale/$25,000 balance at the time of repossession). Bottom line, I’m forecasting loss ratios for auto loans to double in 2023.

But Wait, There’s More!

All of this analysis doesn’t include the second aspect of predicting changes in loss ratios. If your credit union experiences an increase in repossessions as well, your losses will be a function of the change in magnitude and frequency of loss. Stay tuned for part two next month as I share some thoughts on frequency of loss.

Bill Vogeney is chief revenue officer and self-professed lending geek at $9.8 billion Ent Credit Union, Colorado Springs.

Compass Subscription