Under its new leadership,
is making loans that are going bad faster than credits granted by its former, now discredited, management team.
Carmel, Ind.-based Conseco, whose finance arm focuses heavily on mobile home lending, is now headed by CEO Gary Wendt, a former
executive who joined the company in June last year with the aim of repairing the damage done under its former permanent chief exec, Stephen Hilbert.
But mobile home loans made since Wendt took over are showing an early and alarming spike in delinquencies, according to data contained in documents filed with the
Securities and Exchange Commission
and rating agencies. Some $1.16 billion of Wendt loans, made in the second half of 2000 by Conseco's lending business,
, are deteriorating faster than loans made in early 2000, when Hilbert's lending policies appear to have been at their most reckless.
Investors have shown a lot of confidence in Wendt, bidding Conseco stock up over 100% since his appointment was announced. But if the deterioration of credit quality proceeds at the current pace, questions will be raised about the CEO's turnaround strategy. According to the data, Wendt's strategy consists of charging more interest on mobile home loans to boost interest income. At the same time, Conseco is now repossessing more mobile homes and then writing new loans on many of these properties.
While tougher, this strategy carries significant risks: A higher interest rate may not, over time, compensate for the higher delinquencies it causes, and borrowers buying repossessed homes are considered by industry experts to be significantly less creditworthy. In addition, the slowing economy could make this more aggressive strategy all the more difficult to execute, as people will be finding it harder to pay back loans, especially those made at higher interest rates.
Conseco representative Mark Lubbers declined to make Wendt or Conseco Finance CEO Bruce Crittenden available for comment. When
Detox sent an email outlining the issues and numbers raised in this article, Lubber replied that the material was "both selectively misleading and selectively negative." Despite several requests,
wasn't able to get the company to comment beyond that statement.
The company hasn't commented publicly elsewhere on the quick deterioration of Wendt's loans, but the CEO has made remarks about aggregate delinquency measures.
In a March 16 fourth-quarter earnings release, Wendt said that total mobile-home loans delinquent for more than 60 days -- a narrow measure of delinquencies that excludes a large number of past-due loans, as well as credits on repossessed properties -- should fall to 1.65% of overall mobile-home loans at the end of 2001, from 2.16% at the end of 2000. In 1999, the figure was 1.55%.
At the end of 2000, Conseco mobile-home loans totaled $26.3 billion, or about 55% of the firm's $46.6 billion in financial receivables, according to its annual report, filed Monday.
It's hard to find more extensive aggregate data on delinquencies, but a March 27 Conseco prospectus states that all loan contracts more than 30 days past due (but again excluding loans on repossessed properties) rose to 4% of loan contracts at the end of 2000 from 3% a year earlier. That number fell to 3.33% at the end of February, according to the filing. (For an explanation of how to get a more complete picture of Conseco's delinquency trends, please click
Wendt said in the March memo that the upheaval at Conseco in the middle of last year is to blame for the rise in 2000 delinquencies, as "employees took their eye off the ball." And he stressed that 2001 credit quality indicators wouldn't affect his profits guidance. In the memo, Wendt added that Conseco's projections for pretax 2001 Conseco Finance earnings of $350 million to $390 million take into account predicted delinquencies.
Wendt's recovery plan, outlined in December, sees Conseco Finance as the chief source of profits growth, with him forecasting earnings of $800 million from the unit in 2004.
However, for that target to be achieved, Wendt's aggressive new lending policies can't result in too many bad loans. But that may not be possible if credits made under his tenure keep going bad at current rates.
Conseco makes mobile home loans and then, in a process known as securitization, it bundles them up into intricately structured bonds, most of which are sold to institutional investors. Conseco makes money if the sum of money paid out on the pool of underlying loans packaged in the bond is greater than the interest Conseco has to pay out to the bondholders, after factoring in bad loan losses and the cost of servicing the loans in the pool.
On Oct. 5,
Conseco Finance Securitizations Corporation
, the entity that carries out Conseco's securitizations, issued a bond, called Series 2000-5, that is backed by $720 million of mobile home loans. On Dec. 28, Wendt's team made another issue, Series 2000-6, secured by around $440 million of loans. Late last month, CFSC issued Series 2001-1, with loans totaling $580 million. The vast majority of loans in these bonds will have been made under Wendt, since lenders that use securitizations rarely hold the underlying loans for more than three to four months.
In a very short time, many loans backing the 2000-5 and 2000-6 bonds have started to go bad. By the end of February, or in just under five months, loans more than 30 days past due (including credits on repossessed properties) totaled $21 million, or 2.93% of loans in the pool, according to data from a
securitization Web site, as well as data from
Web site (as of Tuesday, the CFSC filings of much January and February data had yet to appear on Web sites posting SEC documents).
And after just two months, the 2000-6 bond showed total 30-plus day delinquencies equivalent to 1.39% of total loans at the end of February.
In and of themselves, these might seem like small percentages. But they are much higher than losses that were notched up at equivalent time periods on securitizations carried out by Hilbert in 2000. And his securitizations from last year had lending standards that were considerably worse than others carried out after Conseco acquired Conseco Finance, then called
Green Tree Financial
, in 1998.
Hilbert's 2000-1 securitization, from February 2000, was experiencing a massive 6.95% 30-day-plus delinquency ratio at the end of February 2001, after only one year in existence. Three months from its issue date, that ratio was 0.77%, which is lower than the 1.39% after only two months on Wendt's 2000-6. And after five months, Hilbert's now-toxic 2000-1 bond had a delinquency ratio of 2.4%, lower than the 2.93% seen on Wendt's 2000-5.
It normally takes three to four years for mobile-home delinquencies to reach a peak, so it's particularly worrying to see this level of deterioration taking place in under a year.
Clearly, Wendt can't let credit quality continue to worsen at this pace. Annualized, the two-month delinquency rate on the 2000-6 comes to a bruising 8.34%, while the annualized total for the 2000-5 is 7%.
A Wall Street consumer finance analyst who requested anonymity speculates that losses may be trending up so sharply in the 2000-6 deal because it contains a lot of loans granted to customers who have bought mobile homes that have been repossessed (the analyst's firm rates Conseco a hold and it has done underwriting for the company).
These loans, termed repo refis, made up 22% of the 2000-6 deal, according to
Moody's Investors Service
. The share for the 2000-5 issue is 9%, according to Moody's. Repo refis "generally go to a weaker borrower," says Pramila Gupta, a senior analyst at Moody's. Hilbert-era securitizations generally contained very small amounts of repo refis.
"Securitizations with a large share of repo refis are going to fare worse than others," says the consumer finance analyst. This assertion is supported by data on Conseco's 2000-4 securitization, from Aug. 11 last year. In this issue, repo refis make up 16% of the loan pool. Here, 30-plus-day delinquencies stood at 5.32% at the end of February, or 9.82% annualized. (The 2000-4 securitization can't safely be attributed solely to Hilbert or to Wendt, as its loans may have been made under both CEOs.)
If the high level of repo refis is to blame for the delinquency spike, investors also need to closely track Conseco's most recent 2001-1 deal, in which these high-risk loans have a 19% share.
It's not hard to see how the repo refi game can backfire. Conseco may be finding itself repossessing homes that it took back only a few months earlier, and the value of a twice-repossessed property can't be all that great.
Other indicators show that Wendt's underwriting standards haven't improved on Hilbert's. Wendt's three securitizations since October last year show an average loan-to-value ratio of 87.6%, with 42% of loans having a loan to value over 90%. (The loan-to-value ratio measures the loan size as a percentage of the cost of the property bought.) The three 2000 securitizations that can associated with Hilbert had an average LTV of 87, but "only" 37.6% of loans had an LTV over 90% on average (you can see other comparative data in the tables within).
Of course, Wendt's get-tough strategy may pay off. As word gets around that Conseco is willing to repossess properties, other borderline borrowers could make more efforts to pay, and delinquencies may come down -- as Wendt is predicting. In fact, 31-to-59-day delinquencies dropped by 3,741 contracts to 10,569 in the two months ended Feb. 28, according to a Conseco filing.
If that happens, and actual losses come down, Conseco gets to keep a good chunk of the interest rate spread it takes in from its securitizations. And Wendt has been managing to obtain a slightly higher interest rate on his securitizations. The average coupon on his three deals is 12.4%, compared with 11.3% on Hilbert's first three in 2000.
The 2000-6 securitization had an excess spread of around 5.25 percentage points at the end of January, according to a bond analyst at a large insurance company that owns Conseco asset-backed bonds. He takes the weighted average interest rate on the loan collateral (12.52%) and subtracts the interest that Conseco must pay (7.27%). "This appears to be how Gary Wendt aims to make Conseco Finance more profitable," says the analyst who requested anonymity.
By contrast, Hilbert's 2000-1 deal has an excess spread, before losses and operating costs, of 3.11 percentage points. But if half of its January 30-plus-day delinquencies were to eventually become actual losses, and total around 3.4% of the portfolio, that excess spread would be wiped out, even before accounting for operating costs. That would eat into the company's overall profitability.
If losses pile up, the current profitability of even the high-coupon 2000-6 deal could be much reduced and endanger Wendt's goal of making Conseco Finance the powerhouse of the new Conseco, says a financial services analyst at a non-U.S. investment fund that is
short Conseco stock.
The analyst notes that the low-interest-rate segments of securitizations mature quickly, which could result in a higher payout coupon for Conseco in two years, thus shrinking the excess spread. He says that the excess spread could fall to around 5 percentage points in two years. Subtract possible losses of 3.5%, as well as operating costs of half a percentage point, and Conseco ends up making only 1% on the deal, he surmises.
"That's nowhere near enough for Gary Wendt to achieve his aims for Conseco Finance, especially if he has to cover big losses on old Conseco deals," says the fund analyst.
Know any companies that the market may be misvaluing? Detox would like to hear about them. Please send all feedback to
In keeping with TSC's editorial policy, Peter Eavis doesn't own or short individual stocks. He also doesn't invest in hedge funds or other private investment partnerships.