CarCrash

Investors willing to talk openly about their activities tend to do so in order to showcase their perceptive and analytical brilliance, either on current ideas or past investments that have proved particularly rewarding. This discussion is neither.

This is our attempt to come to terms with a significant investment error and what to do about it.

CarMax stock fell by a third in the 3rd quarter, with most of the falls occurring toward the end of September when the company posted very disappointing 2Q fiscal 2026 results (fortunately for our portfolio these declines were nearly entirely offset by large gains in our Nigerian bank investment as its listing was upgraded to the main board of the London stock exchange).

The metric that left investors most concerned was the decline in same store used unit sales of -6.3% in the three months to August – a major reversal of the 8.1% growth in same store volumes in the May quarter. The stock fell sharply again in early November as the company guided for an even weaker quarter to the end of November, with same-store volumes expected to be down 8-12% and announcing leadership change with the resignation of the CEO, Bill Nash.

CarMax is now in the midst of a major crisis.

The six-month period to the end of November now appears on track to be the worst in terms of same store sales declines outside the depths of the financial crisis in 2008 (and getting pretty close to those levels). CarMax’s business model is volume-driven, with earnings growth highly leveraged to higher like-for-like volumes, so these declines are a major problem. The difference between 2008 and today is that the overall economy doesn’t appear to be in a major crisis (although there are clearly significant pockets of consumer stress), and CarMax is facing much tougher competition from the likes of Carvana. In other words, there is a reasonably high probability that this time the problem is not with the industry or the overall economy, but with CarMax itself.

What makes this hard to judge, however, is that we appear to be in uncharted territory with respect to the used car industry given vehicle affordability is probably at its lowest level in recent decades due to high prices for both new and used cars, weak growth in real incomes, and elevated interest rates. These are industry conditions in which some of CarMax’s less conservative competitors are able to thrive. Before jettisoning our holdings of CarMax it is worth pausing to reflect on these conditions.

Rewinding just a few months, same store volumes were on an improving trajectory through the second half of last year and into May. Same store volumes grew 3.1% on average over the three quarters ended February and accelerated to 8.1% in the May quarter. Even if some of the May quarter's rise was due to tariff-related demand pull-forward, it still appeared that CarMax was turning a corner after several years of struggle – a struggle borne of the extraordinary conditions in the market for used vehicles since the pandemic. Used car prices surged in 2020 and 2021 as supply was curtailed and demand stoked by massive government stimulus. The Manheim used vehicle value index surged 67% between December 2019 and December 2021. Used prices have since fallen by 21% but remain around a third higher relative to pre-pandemic levels. CarMax’s average used retail price jumped 27% to $26,207 in fiscal 2022 (ended February) and has remained around that level ever since.

With both vehicle prices and interest rates significantly higher relative to 2019 vehicle affordability has plummeted. Monthly car repayments have reached historic highs with the September average monthly repayment on a new vehicle reaching $749 (according to Lending Tree), and $529 for a used vehicle. A record one in five new vehicle buyers are committing to a payment of over $1,000 per month and in order to make monthly repayments more digestible, loans are getting longer with a record 66% of new car loans are now being financed for six years.

Coming off a 2021 volume surge, with prices so high and consumers suffering a post- stimulus hangover, it was easy to rationalize the declines in CarMax’s volumes in 2022 (-13%) and 2023 (-5%) as a painful, albeit temporary phenomenon. As a volume-focused player the lack of affordability has been a major challenge for CarMax, which strives to offer fair and transparent pricing and has seen its gross profit per unit increase by less than 6% between 2019 and 2024, far less than would be necessary to offset the drop in volume even if other revenue lines like service contracts and interest income have grown a little faster than that. Unfortunately, in a period of higher inflation, CarMax’s costs have grown faster than its revenues. Operating expenses per unit rose by roughly a third from $2,330 in 2019 to $3,086 in 2024, which means CarMax generates less operating profit per vehicle today than it did five years ago. The biggest driver of this increase in costs has been employee compensation, which rose by nearly 48%, to $1,795 per retail used vehicle sold. Even as staffing levels per store have moderated back to 120 as of February 2025, slightly below pre-pandemic levels, cost per employee is still 24% higher.

The driver of business value for CarMax is its ability to grow used vehicle volumes faster than its growth in operating expenses (primarily compensation costs given they represent nearly 60% of total operating expenses) and considering new stores come with additional costs, generating like-for-like volume growth is absolutely key.

The industry backdrop for CarMax’s business model has been uniquely challenging. It is hard to blame just a poor environment, however, when the company’s closest competitor is growing in leaps and bounds. Carvana grew its retail used car volumes by 33% in 2024 to 416 thousand vehicles, and growth has accelerated to over 40% in the first nine months of 2025.

An obvious conclusion to draw from this data is that Carvana is eating CarMax’s lunch and is growing volumes because it simply provides a superior service and product. That is certainly the conclusion the market has drawn with Carvana's equity now worth over $95bn (and >$100bn in enterprise value, as of 10 December) vs CarMax’s $5.7bn. Game, set, match.

But something about this situation does not sit right with us. CarMax will still sell over 30% more used retail vehicles this year than Carvana and makes a consistent $2,000+ profit between the cost of each car acquired and the price at which it is sold. Carvana’s gross profit, on the other hand, is loaded with other income line items that would ordinarily be reported outside of gross profit, one of which is the profit on sale of auto loans made to customers and sold on to financing 3rd parties.

CarMax has historically financed just over 40% of the cars it retails with a focus on low credit risk customers, with some additional financing to higher risk customers provided by partners. Carvana’s financing rate is much higher, typically between 70-80% (according to its asset-backed securitization disclosures), and its lending criteria appear to be significantly more lenient. Of the $18.5bn in aggregate outstanding principal Carvana has sold to investors as of September 2025, approximately 56% is “prime”, implying around $8.3bn is less than prime. Nearly $1.2bn of that $8.3bn in “non-prime” loans were delinquent by more than 31 days in Carvana’s most recent filings – a 30 days past due delinquency rate of 14%, indicative of well below prime lending. Carvana’s most recent asset backed securities transactions also show that 40% of loans are for between six to seven years and that the annualized charge off rate for these loans was 4.2% in the first half of 2025.

In a period of record car prices and high interest rates, a willingness to make risky loans to customers who couldn’t otherwise afford to purchase a vehicle seems like a pretty obvious way to grow sales. It is not our objective to cast aspersions at Carvana, suffice to say that it pays to be wary of any financial institution growing its lending by 20%+ p.a., as other institutions appear to be pulling back.

And despite the fact that credit conditions remain extraordinarily accommodative, with BBB corporate spreads at historically tight levels over treasuries – even lower than in 2021 – there is growing evidence that the easy subprime credit chickens are heading homeward.

BBB Corporate Index Option-Adjusted Spread

Sub-prime auto retailer and lender, Tricolor Holdings, filed for Chapter 7 bankruptcy (straight to liquidation) in September with around $1.4bn in loans. In October a small provider of subprime finance to car dealerships called PrimaLend also failed. There are other signs of rising stress in private credit portfolios.

As the CEO of JPMorgan said on the bank’s 3Q earnings call, “I probably shouldn’t say this, but when you see one cockroach, there are probably more.”

With the Fed about to lower interest rates again in December, a new Fed chair starting in May, and mid-term elections less than a year away, there are plenty of reasons to believe policymakers will move heaven and earth to avoid or even just forestall an economic slowdown, tightening credit conditions or a sell-off in high-flying technology stocks. A cynical observer might suggest this can't happen until OpenAI goes public in order for it to come anywhere close to meeting its gargantuan spending commitments, as well as allowing well-connected insiders to take profit.

But given the stress for low-end consumers, how long before financing for sub-prime auto credit becomes much, much harder or more expensive?

It has been a very painful four years as a shareholder of CarMax. It is possible the business model is fundamentally broken if it cannot drive volume growth above growth in operating costs given the intensified competition. That said, it's not clear to us that the story is over, despite the recent stock price declines and upheaval in the executive ranks. We are neither a seller or a buyer of the stock until we get more clarity on where the business is heading, but the position has become substantially smaller (as of today around 4% of the portfolio).

It is worth noting that the value of the company today is a little below $6bn, less than half its worth from a decade ago, despite selling 50% more cars and generating twice as much revenue.

One simple metric is worth keeping in mind: if CarMax were simply able to return to its historic profitability levels of $1,000 in net income per retail used vehicle sold (the average of the decade to 2019) on volumes it achieved in 2021, the stock would trade on less than 7x earnings. By comparison, if Carvana achieved a similar $1,000 in net income per used unit (they are, in fact, in the same business despite one trading like a software business), assuming a P/E multiple of 20 (slightly below the currently elevated S&P500 average), an equity market valuation of $95bn would imply used retail unit sales of 4.75m, or 8x from this year's expected volume, and nearly a quarter of total used retail vehicle sales in the US.

There are clearly lessons to be learned from this experience, but we want to make sure that we learn the right ones. That will take some more time.