CACC Part 4: Will subprime auto lending end badly?

I think that the word “subprime” has an undeserved reputation.  The US subprime housing crisis was a very unusual situation caused by extreme levels of loan fraud and loans designed to blow up (e.g. ARM, negative amortization mortgages, etc.).  These excesses do not exist in other areas of subprime lending.  During the housing bubble, lenders did not verify income and allowed unemployed individuals to buy millions of dollars of housing.  In the current auto market, the lenders actually check that the information given to them is correct.  According to CACC’s training website, 75% of loan applications are initially rejected due to problems with the loan application (e.g. a utility bill to verify residence is too old).  The subprime lenders tend to perform the most loan compliance work (out of necessity) while prime lenders have fewer stipulations, looser loan compliance, and less onerous paperwork.

You don’t have to take my word for it.  The excellent website insider-car-buying-tips.com lists the documentation/stipulation requirements of major subprime lenders.  When the website talks about prime lenders, it doesn’t even mention the documentation required.  Prime lenders generally do not require 5+ references and have less stringent requirements for proof of residence.

Many lenders will have a checklist of their loan requirements.  Chase’s checklist for (prime, good credit) consumers states that verification of income, identity, residence, etc. may be requested.  It uses the word “may” instead of “will”.  It does not ask for references.  Compare this to CPSS’s guidelines:

The standard STIPS to be included in the package submitted for purchase:
1. Proof of Income (“POI”) for each Buyer
2. All pages of current telephone bill (NO PHONE, NO LOAN)
3. Proof of Residence (“POR”) for physical address
4. Proof of monthly Rent/Mortgage payment amount
5. List of 5 complete personal references
6. Evidence of Vehicle insurance in customer’s name

CPS personnel at the corporate office may stipulate additional documentation or conditions of
approval on a case-by-case basis.

Loan fraud

A New York Times / Dealbook article on subprime auto lending argues that the industry is in a bubble.  It cites the case of Rodney Durham, who obtained a loan from Wells Fargo even though he lives on Social Security, declared bankruptcy, and hasn’t worked since 1991.  The article goes on to draw parallels between loan fraud in the US housing bubble and the current subprime auto market.

I don’t believe that cases like Mr. Durham’s are that typical, at least not in the deep subprime market.  The website Insider Car Buying Tips explains common forms of loan fraud and how consumers can protect themselves against it.  While loan fraud does occur, deep subprime lenders have the most safeguards against loan fraud.  In Credit Acceptance’s case:

• They will call the employer to verify employment.
• The borrower must provide proof of income, e.g. pay stubs, bank statements, etc.  This is described on the Insider Car Buying Tips webpage for Credit Acceptance.
• The dealer holdback / pooling agreement structure.  The dealer more or less receives a large upfront payment (the advance) and smaller payments over time.  They do not receive all of the money from the bank right away.
• Fixed fees perform a function similar to the holdback / pooling structure.
• Credit Acceptance’s dealer rating system penalizes dealers for making bad loans by giving them lower advances.  When a bad dealer closes their pool of 100 loans, they will receive a smaller cheque.

As lending moves up to less risky borrowers, there are less safeguards in place.  The lenders will take on more loan fraud risk to offer a more convenient process for the customer and the dealer.  They may perform less loan compliance work because loan compliance costs money and may not be warranted for less risky customers.

The winners and the losers

I do think that some subprime auto lenders will go bankrupt or exit the business.  The industry has historically been cyclical and difficult.

I think that the main drivers of success or failure will be:

1. Reasonable levels of leverage.  A sustained level of high debt will likely doom a company eventually.  To maintain high levels of debt, a profitable company will be forced to expand rapidly.  Doing so will likely require the company to lower underwriting standards.
2. Skill in collections and loan servicing.  Being a low-cost operator is an advantage.
3. Skill in underwriting.

I would consider CACC, CRMT and NICK* to be the best managed companies in the subprime auto lending industry.  (*NICK’s founder and former CEO has left the company entirely.)

The most questionable companies would be:

1. The ones with too much debt.  A debt to equity ratio of 10:1 is probably not a good idea.
2. The ones which are trying to impress equity investors with growth.  Growth is not a good thing if the lender significantly weakened underwriting standards to achieve it.  I would be skeptical about recent IPOs and companies taken public by private equity firms.  Sellers have an incentive to push the company into doing dumb things to generate growth.

Excesses in the sector are most likely to occur with the companies trying to grow rapidly or to achieve high returns on equity (rather than returns on capital).  Excesses may also occur in the securitizations of auto loans.

Past Excesses

Auto loans

The August 2012 issue of Moody’s Structured Thinking has an excellent article on the subprime auto market titled “US Subprime Auto Lending Market Harkens Back to 1990s”.  Some of the issues that it brought up were:

• “2. Heavy use of securitization can facilitate excessive growth”
• “4. As portfolios weaken, defaults could exceed servicers’ capability”
• “6. High loan losses and servicer disruption can result in losses to investors”
• “Credit enhancement” via monoline insurers actually did not result in problems.  (In contrast, this practice played a role in the US subprime crisis.)

US housing crisis

• Rampant loan fraud on stated income loans, “NINJA” loans (no income no job), liar loans, etc.
• The owners of bad housing loans had little oversight of the origination of these loans.  Because there was a disconnect between the lenders and the origination of these loans, those originating loans were incentivized to originate as many loans as possible without thinking about loan quality.  To be fair, the truth is a little more complex.  Countrywide for example had some skin in the game because it retained residual securities.  As well, it had exposure to the representations and warranties on its subprime securitizations.  The problem seems to have been the disconnect between the CEO’s incentives and shareholders.  Countrywide’s CEO at the time made a lot of money even though the shareholders would suffer huge losses.
• Many mortgage-backed securities were insured by other companies.  When these MBS failed, insurers lost a lot of money.  Because these insurers were counterparties to many derivatives and other financial transactions, the failure of some of these insurers threatened to take down other financial institutions.  Because the US financial system is highly interconnected, this threatened to take down much of the US financial system.
• Adjustable-rate mortgages showed unusually low delinquency rates in the beginning before the mortgage reset.  This meant that it took a long time for bad loans to reveal themselves.  Negative amortization loans had similar problems.
• “A rolling loan collects no loss”.  Bad loans performed unusually well because the underlying homes were being resold and refinanced.  This led to looser lending which increased the supply of buyers, causing house prices to rise which led to more bad lending etc. etc.
• Loan losses were made more extreme because mortgage debt is non-recourse and because recoveries depend on the value of home prices (which were falling).
• There are conflicts of interest between the loan servicers and the owners of the loans.  The servicer often wasn’t being paid enough to handle delinquent loans.  (Agency loans had better structures than your typical MBS.)  The poor structure of MBS made the significant losses on them slightly worse.  Had the fees been better structured, the servicers would have put more effort into modifying mortgages and foreclosing on fewer homes.
• In some cases, servicers were not allowed to modify mortgages.
• Some of the problems in the housing crisis were due to buyers with excellent credit scores.  A credit score alone is not a good indicator of loan quality.  One can easily make bad loans to people with excellent credit.

In my opinion, the subprime auto loan market is not in a bubble and looks nothing like the US housing bubble.  The level of excess is rather low.  Yes there are some excesses in the subprime auto field, mainly in the lenders which are overleveraged and are trying to impress shareholders with growth.  But there is a part of the industry that has an excellent track record of value creation, share repurchases, and creating profits for shareholders.  Many investors may be inclined to overlook these companies because few analysts cover them, they rarely make the news, they’re non promotional, and because of the “ick factor” associated with any form of subprime lending.  Therein lies the opportunity.

*Disclosure:  Long CACC, no position in NICK or CRMT or other subprime lenders.