About a month ago, I wrote a pessimistic post about Credit Acceptance (CACC). Glenn Chan, who had written about CACC previously, replied in the comments. William Knecht, who had written CACC up at Corner of Berkshire and Fairfax, also commented.
In contrast to my pessimism, Chan and Knecht are bullish and own the stock. They both make good points that are worth responding to. I'd actually meant to respond much sooner, but life got in the way... better late than never.
I think the difference of opinion is partly philosophical. Chan and Knecht look at the company from bottom-up and see that it's historically been a superior lender. I care less about a business's quality and focus on the macroeconomic picture. I think macro is particularly important for CACC because lending is a commodity business and depends on access to the financial markets.
I also focus on the big picture because CACC doesn't disclose much information about its business. This reply to a shareholder inquiry, although a decade old, is typical:
We have not found portfolio delinquency rates to be useful in running our business, primarily because portfolio delinquency rates are influenced to a significant degree by the rate of growth in loan originations. Since we do not find the metrics you requested useful, we do not believe adding these metrics to our disclosures is appropriate at this time.
Chan and Knecht make a persuasive case that CACC is better than many of its competitors and that there are structural reasons why it's better. They're probably right about this. When I'm bearish about a sector for macro reasons, I like to avoid the good companies and short the bad ones. Since CACC seems to be a good company, I'm not short the stock, but I wouldn't buy it at anywhere near today's price.
An anonymous commenter asked what i thought about America's Car-Mart (CRMT). I'm less familiar with CRMT, so I don't have a strong opinion about it, but from what I've read it seems to be very similar to CACC: a strong player in a sector I dislike for macro reasons.
Returing to CACC, here are some of the points on which I disagree with Chan and Knecht.
Is subprime auto lending a bubble?
In my original post, I drew a parallel between the housing bubble and the current boom in subprime auto lending. Knecht writes, "I don't think comparisons to sub prime housing super appropriate" and gives several reasons why.
I didn't mean to make a strict comparison between subprime auto loans and subprime mortgages. Certainly, many of the details are different. The big similarity, in my opinion, is that historical performance has given lenders and ABS buyers a false sense of security and encouraging them to take excessive risk.
When the tech bubble burst, house prices performed very well, and many investors assumed that was an inherent feature of the housing market ("housing has never gone down on a nationwide basis"). When the housing bubble burst, auto loans performed very well, and now many investors assume that's an inherent feature of auto lending ("people need to drive to work, so they pay off their auto loans first").
One sign of a bubble is that subprime auto borrowers' average FICO score is at a record low, although the average loan-to-value ratio is still a few percent below its record high.
Another sign is that subprime auto lenders have been able to raise money on very favorable terms. 70% of one recent Americredit ABS deal is rated AAA and has interest rates below .7%. Similar deals from other lenders have 60-80% rated AAA with rates below 1%. Borrowing at a ~2% blended rate and lending at 20%+ has inflated the ROEs of CACC and its competitors, but it comes close to "return-free risk" for ABS buyers.
Is CACC participating in the bubble?
CACC's annual loan volume and outstanding loan balance are multiples of their 2006-07 levels. Part of this may be a superior operator taking share, but it's also a result of the company leveraging itself: CACC's loans net of reserves were 150% of tangible equity at YE2006 and 300% at YE2013.
All of this growth has come from new dealers, however: per-dealer volumes are lower than during '06-07. This seems to indicate underwriting discipline on CACC's part. As Knecht writes, "CACC has exhibited strong discipline, walking away from business without a margin of safety. Witness the dealer attrition they experience and infrequent pricing changes." On the other hand, the new dealers present a risk of adverse selection.
Chan writes that "Loans with longer terms would be a sign of potentially dangerous lending." CACC has extended its average loan term from ~3 years in 2006 to ~4 years today. In the company's defense, most of its competitors offer longer terms. As the industry has loosened standards, so has CACC, but it's stayed more disciplined than most other lenders.
To what extent does CACC share credit risk with dealers?
Chan writes, "If the lender is like Credit Acceptance or Westlake, then the lender gives an advance for only part of the loan. That's not so problematic."
Knecht writes, "CACC sweeps all the money until it's advance is paid off. In other words, the dealer is in the first loss position."
I don't think that dealers take the first loss. Dealers must pay to join CACC's lending program and then pay monthly fees to stay in, and I doubt they would do this if the program also forced them to retain credit risk.
As I understand it, CACC measures its anticipated collections and dealer advances against the total expected payments from the loan. For a car with a purchase price of $10,000, the loan may call for the borrower to pay $20,000 in principal and interest. For 2013, the advance rate was 47.6%, which would amount to $9520 on this hypothetical $20,000 loan. If the loan was made at an LTV greater than 100%, as subprime auto loans typically are, then the dealer should be able to lock in a profit.
For comparison, here are CACC's loan statistics for 2006:
Average consumer loan (principal + interest) $15,016
Average purchase price $ 7,374
Purchase price 49.1%
Forecasted collection rate 70.9%
On the other hand, dealers receive significant profit-sharing payments if losses on the loans they've originated stay below a certain level. They can also get kicked out of CACC's lending program if they originate too many bad loans. So they do have incentives to keep credit quality high.
Can subprime auto borrowers refinance their loans?
In my initial post, I suggested that subprime auto borrowers can refinance their loans and that may be suppressing loan losses. Chan writes that he "[doesn't] believe it's easy for deep subprime borrowers to effectively refinance."
After doing further research, I think that Chan is correct: subprime auto refinancings seem to be relatively uncommon, although Westlake, one of CACC's direct competitors, offers them.