Loan Numbers Could Spell Trouble for AmeriCredit

Car buyers know to be suspicious of bargains at dealerships. The same skepticism should be exercised by investors now looking at auto-loan maker AmeriCredit (ACF Quote), which is trading at what looks like a fire-sale price after the Sept. 11 attacks.

Fearing the terrorist strikes could weaken the economy and cause more borrowers to default, investors have rushed to dump AmeriCredit, which lends to people with spotty credit histories. Now, down 25% since the attacks, AmeriCredit looks to be one of the market's most alluring bargains.

The company expects to earn as much as $3.96 a share in its fiscal year ending June 30, 2002, up more than 50% from 2001. Normally, investors would have to pay through the nose for that sort of red-hot growth rate, but the stock, at $30, is trading at a mere 7.6 times 2002 earnings.

However, the Fort Worth, Texas-based company is cheap for a reason. The low valuation reflects fears that bad loans could mount. Short-sellers have long betted against AmeriCredit, believing it couldn't sustain its earnings growth and would report increased problem loans. But until August, the stock just kept rising. Now, a maverick sell-side analyst has come out and publicly knocked the company, citing concerns about credit quality and loan accounting.

Bill Ryan, a consumer finance analyst at Ventana Capital, a newly founded New York brokerage, issued a report earlier this month that advised his clients to sell AmeriCredit, breaking with the uniformly bullish sell-side consensus on the company.

Ryan, who previously worked as an analyst at Salomon Smith Barney, thinks the stock could fall further. In particular, he contends that recent credit-quality data contradict AmeriCredit's claim that it's managing to find more creditworthy borrowers. And he wonders whether AmeriCredit's reported loss numbers are artificially reduced by allowing late-payers to defer loan payments, and by lengthening the amount of time repossessed cars are held by the company.

"If credit losses look too good to be true, they probably are," Ryan says.

A San Francisco-based hedge fund manager who's short AmeriCredit agrees. His fund's betting against the company precisely because it focuses on tainted, or subprime, borrowers. "It's real simple. We're looking for lenders with big piles of subprime loans. AmeriCredit is a one-stop shop for us."

AmeriCredit spokeswoman Kim Pulliam rejects the assertions that credit quality is a problem for the company and denies that it's using deferral and repo policies to mask a higher level of losses in its $10 billion auto-loan portfolio. She adds that the company's earnings guidance hasn't changed as a result of the terrorist actions.

Issues

Ryan's main point is that the credit quality of loans made in 2000 and 2001 is just as bad, or worse, than earlier loans. Why is this a problem? Because AmeriCredit has claimed that improving credit scoring processes allow it to make better loans now than in the past.

Investors follow credit developments by looking at monthly data for large pools of loans that AmeriCredit sells. Like many in the finance industry, the auto lender bundles its loans into bonds that are sold to investors in a process called securitization. AmeriCredit makes money if the interest rate paid to investors in these bonds is lower than the interest rate on the underlying loans, after factoring in bad-loan losses and other costs.

Some of the recently sold loan pools aren't faring that well, casting a question mark over claims that underwriting standards are getting better. Take the $1.09 billion of loans sold in an April securitization. By August, four months after the securitization was done, 2.39% of these loans were more than 60 days past due. That four-month delinquency number is much higher than on any of the company's previous 16 securitizations.

To get a fuller credit picture, Ryan adds 60-plus delinquencies to loans on cars that have been repossessed from past-due borrowers. This calculation turns up a jarring number for the $1.4 billion loan pool sold in January.

Getting Worse The numbers still raise questions about the quality of loans

* 60-plus-day delinquencies plus amount in repossession, as a percentage of total principal of securitizations done in 1997, 1998, 1999 and the three deals done in 2000 that are at least 12 months old. An explanation of the term securitization can be found in the article. Source: Ventana Capital, Detox

In this pool, 60-plus loans and repos together added up to 3.59% of the total pool in August -- just seven months after the loans were sold. This is the highest seven-month figure since a pool sold in 1997. And most other 2001 and 2000 loan pools don't compare well with earlier years.

AmeriCredit's Pulliam says this approach is misleading. Investors shouldn't get overly preoccupied with delinquencies and repos, she argues. Instead, they need to focus on actual losses caused by defaults. After all, it's losses, not delinquencies or repos, that reduce the profit margin on securitization bonds.

Pulliam explains that AmeriCredit is in the business of making loans to shaky borrowers and readily expects some of them to go delinquent from time to time. However, she adds, many troubled borrowers actually end up paying the loan back, meaning AmeriCredit avoids having to write the loan off at a loss.

And Pulliam has a point. AmeriCredit's loss rates for 2000 and 2001 are substantially lower than those for 1997 and 1998.

However, comparisons with 1999 don't look so good. A $1.3 billion loan pool sold in January 2000 shows cumulative losses totaling 5.07% of the combined loan total in August, or 19 months after the credits were sold. That's higher than three of the four securitizations done in 1999.

Then there's the question of how accurate AmeriCredit delinquency and loss numbers really are. Questions stem from AmeriCredit's practice of allowing some borrowers to defer one loan payment or more for a fee. The missed payments are added to the principal of the loan. Ryan questions whether these so-called extensions are done as part of sound and genuine efforts to reduce losses, or if they are just delaying them.

The jury's still out on that. But one thing is clear: Despite company claims that its borrowers are more creditworthy, the 2001 securitizations show the same or a greater number of extensions than loan pools sold in previous years.

Meanwhile, the extension fees are dropping as a percentage of loans extended, suggesting AmeriCredit is getting less compensation for making concessions to borrowers.

Pulliam responds that extension numbers don't shed much light on borrowers' credit quality because many extended debtors go on to pay back their loans once they've gotten an extension. And she adds that the extension fees have come down as a percentage of extensions because AmeriCredit has made a big shift in the type of loans it makes. "No amount of fees will make up for a default," says Pulliam.

The Other Way

The other way an auto lender can keep losses down, at least temporarily, is to hold cars for a longer period in repossession. Only when the lender sells the car does the loss get booked.

At AmeriCredit, the inventory of repossessed cars has risen. Repos totaled 1% of total loans at the end of June, compared with 0.6% in the year-earlier period. Pulliam strenuously disputes any claim that fewer repos are being sold to keep losses down. She says that cars can't remain in the repo stock for more than 90 days, and she says it makes little sense to look at the size of that stock at any one time. The company tries to maximize its return on repo sales by selling more cars when auction prices strengthen, she explains.

AmeriCredit's stock has taken a skid. If Ryan is to be believed, it won't be long before earnings also feel the crunch.