Health risks

The ailing American subprime market is languishing on its sickbed and some experts fear it will contaminate international credit markets – but is the alarm about contagion exaggerated? By Vanessa Drucker in New York

The financial media have been issuing dire and daily warnings: the sky is falling and the foundations are crumbling under the American subprime mortgage market. There is no question that some investors will suffer large losses. The bigger riddle is, will America’s subprime woes spill over into the international credit markets at large, resulting in more serious tremors – or even a meltdown.

The jury is still out and no one can accurately predict the extent of damage. Yet while much of the media predicts a possible Armageddon, most securitisation professionals sound less alarmist. While he acknowledges that valid arguments can be made on both sides, “it will probably turn out a storm in a teacup,” says Paul Forrester, a Chicago attorney with Mayer, Brown, Rowe & Maw.

Several factors have led to the erosion in American subprime performance. Most salient has been the deterioration in underwriting standards among lenders, for loans originated from about the fourth quarter of 2005 through to the end of 2006. For example, lenders tolerated much higher loan-to-value ratios, debt-toincome ratios and borrowers’ leverage when they extended financing, and accepted lower credit scores. They also allowed unprecedented levels of loans with “stated” documentation, which rely on the borrowers’ own statements for proof of income. “Most tellingly, the number of subprime loans originated to first time buyers was historically high in 2006,” says Michael Youngblood, managing director asset-backed securities research at Friedman Billings Ramsey.

Meanwhile, the subprime sector has been stealing a larger share of the mortgage pie. It represented 10% to 11% of the American market for the previous decade, but suddenly swelled to 25% of originations in the past couple of years.

In part, slick advertising and easy terms account for that increase. Some borrowers, who could have qualified for prime rates, chose subprime to obtain a larger house or investment properties to flip. “Other people were not ready, emotionally or financially, to become first time home buyers. But mortgage brokers, estate agents and Wall Street institutions were all prepared to accommodate them, at least until the market turned,” says Amy Crews Cutt, deputy chief economist at Freddie Mac, a government agency.

Poor or liberal underwriting standards have not been unique to subprime. Between the prime and subprime categories, the class of Alt-A borrowers, who are frequently selfemployed, may lack verifiable income documentation, and therefore pay higher rates. Traditionally prime loan credit (FICO) scores range from 850-720, Alt-A from 720-680, with subprime representing 680 and below. In the face of rising interest rates, when the Federal Reserve began to tighten in June, 2004, lenders became anxious to increase market share and profitability, and preserve origination volumes. “So they reached down into the lower regions of their respective loans types,” Youngblood explains. “And Alt-A lenders tried to expand share, by reaching into the upper reaches of subprime.”

The main credit rating agencies, Moody’s, Standard & Poor’s and Fitch, have taken criticism for paying insufficient regard to the loose underwriting criteria. Quincy Tang, senior vice president RMBS at Dominion Bond Rating Service, responds that her firm has been careful. “DBRS doesn’t need to rate 100% of the market, so it is easier for us to say no to certain issuers. Since we can be more selective, we were cautious about rating some mortgage companies that lacked internal financial strength and that were relying on Wall Street conduits to fund the loans they were originating.”

“Rating agency models tacked on extra points for diversity for mortgages that incorporated second liens,” adds Glenn Schultz, managing director consumer ABS and mortgage research at Wachovia Capital Markets. “Now second lien loans are defaulting at a higher rate.” Homebuyers, who previously put down 20% of the purchase price, began to obtain second loans for that 20% instead, without spending a penny.

Those liens have turned out to be one of the reasons the 2006 vintage mortgage pools have performed badly, and they could affect about half the subprimes from 2006, versus merely 10% in 2001. The foreclosure process for first lien loans takes time, typically 12 to 18 months. Second lien servicers, on the other hand, tend to write them off immediately, realising there will be nothing left to recover.

Contributing to the perfect storm, the American housing bubble burst, as prices began their descent in autumn 2005. Until then, the steady gains in property prices had seemed to provide a safety net for homebuyers – there was always a profit on the horizon.

Arizona, California, Florida and Nevada have witnessed the highest home appreciation. These are now experiencing the largest increases in subprime delinquencies and foreclosure starts, which reflect the downturn in housing prices. “We are still some way off from seeing the market bottom,” says Jay Brinkman, vice president research and economics at the Mortgage Bankers Association. Speculators need to get out since prices are falling; others, who overreached to own a house, must now sell. “They expected they would always be able to sell at a profit, while the market was rising,” Brinkman reports.

At the same time, economic conditions and employment are closely linked to foreclosure rates, on a regional basis. The hardest hit areas have been in the Midwest, where manufacturing industries have shed jobs. Those states reveal about 20% of loans in foreclosure.

The next headache due will become resets. Many subprime mortgages are structured to reset rates after the initial 24 months, so those of 2005 are entering the critical period. As their rates jump, the pain will particularly strike those with poor credit, or whose homes have fallen in value. “We’ll see the main action in 2008, with a tail end in 2009 on 36-month resets from 2006. We won’t be able to stop gritting our teeth and holding the arms of our chairs until well into 2008,” predicts Mark Adelson, head of structured finance research at Nomura Securities International.

Since we are still gritting teeth at this stage, what risks can we already discern? The problem is that it may still be too early to ascertain. “Our dilemma is that we are facing unquantifiable uncertainty, with a capital ‘U’”, says Adelson. “That is more challenging than a comparatively ‘nice’ uncertainty like risk, which can be measured.”

We do know, however, that Wall Street repackaged many triple-B and triple-B-minus tranches into a fresh generation of collaterialised debt obligations (CDOs). We still do not know how to value the BBB pieces of deals, where CDOs have taken on exposure beyond the dollar amounts of paper issued.

For the CDOs craved them so much that they had to create synthetics to meet their voracious demand. To compound the complexity, issuers have also structured CDO-squared products from the bottom tranches of existing CDOs, which are now even further removed from the original subprime collateral. It is some comfort, though, that they do provide investors with double protection, from their own credit enhancement as well as the credit enhancement on the underlying securities.

Few CDOs are ever composed entirely of subprime mortgage backed securities (MBS). Most are a blend that reflects a manager’s particular expertise or sense of market opportunity, and consist of a mixture of prime, Alt-A, commercial mortgage backed (CMBS) and non-mortgage asset-backed securities (ABS). CDOs originated in 2006 did, however, contain higher proportions of subprime loans than in prior periods, so we should expect to see a higher proportion of default rates. “CDOs created in the second quarter of 2006 held up to 80% in subprime,” says Youngblood.
“Already in 2007, the composition of the deals is differing sharply, with new CDOs containing only 10 to 20% in subprime.”

We simply have no idea who is holding the riskiest tranches of the CDOs, or how concentrated the holdings are. Indeed, it worries the rating agencies – who are accustomed to rating one security at a time – that the risk is piled up at some point in the system, but it is unclear just where. David Wyss, chief economist at Standard & Poor’s, highlights two logical suspects: hedge fund and non-US investors. “We have anecdotal evidence that CDOs have been sold heavily overseas, although we do not know who bought them”, he says. “Many were sold in London and in Zurich, and might well have been purchased with recycled petrodollars.”

Celia Chen, director of housing economics at Moody’s Economy.com voices concern about the plethora of foreign investors. “These pools of buyers from all over the world have exploded over the past decade. They are pouring money into the US housing market, without realising how cyclical those housing markets are. Most have had no direct experience of a US property downturn.”

The other suspect is hedge funds, which have been keen purchasers of CDOs. The danger is that a huge hedge fund might send a wide ripple effect into the credit markets, particularly if it is overlevered. That situation is distinct from normal trading losses. Investors gamble, on emerging markets for example, where they win and lose each day.

Resecuritisation can further cloud the issues. Although the Securities and Exchange Commission prohibits the practice in public transactions, private markets are not constrained, and entities like hedge funds can reissue new shares, using “tainted” securities, says Jim Croke, a structured finance attorney at Orrick, Herrington & Sutcliffe in New York. A usual construct would be to buy triple As, with a yield of Libor (the London inter-bank offered rate) plus 25, borrow funds for leverage, and put the securities into a new structure that can fund itself at Libor minus five. Croke says: “If the performance of the bought securities starts to hurt, then the securities of the new vehicle could be adversely affected.”

Determining the value of the CDOs can be difficult. Because these instruments are fairly illiquid, and do not trade daily, valuation can be complicated, and is often based on credit quality. Outside the scope of markets like common stocks or government bonds, dealing spreads increase sharply too. The difference of opinion on values can depend on bid and offer prices.

“Still the fact that no over-thecounter or exchange traded markets exists does not mean values cannot be assigned,” Youngblood says. Take the recent $1.6 billion (£780m) Bear Sterns bailout in June. Merrill Lynch seized $825m in assets it held as collateral for its loans to the Bear funds. Clearly, Merrill Lynch had a reasonable basis for valuing the CDO equity it held, or it would not have entered that business. Likewise, Bear must have taken comfort in Merrill’s valuation, or it would not have engaged Merrill as a repo counterparty.

Under American Generally Accepted Accounting Principles (GAAP) rules, most significant investors must mark-to-market once it becomes clear their interests are imperiled. Those losses are taken through the income statement, which is a “one-way rule”, says Forrester. “Once you take a write-down, you can’t write it back up.”

So even “buy-and-hold” CDOs will be affected, if they are subject to US GAAP. Subprime troubles do impact valuations as the rating agencies begin to downgrade the investment pieces. Investors face exposure as they mark their holdings to market. Typically the rating agencies start to downgrade at the lowest tranches, which are the pieces hit first, as massive losses eat into the over-collateralisation level of extra support and safety. “We are already seeing almost weekly downgrades in the lower tranches, although not among the triple As, which account for the highest volume,” Brinkman reports.

CDO managers must take action when they acknowledge the problem. The legal structures of CDOs include triggers for particular rights. For instance, substantial losses in subprime loans may affect the calculation of coverage ratios. “Once a security defaults, it is normally kicked out or haircut to a pessimistic recovery rate,” Forrester explains. Therefore, a manager must sell collateral to restore compliance with mandated collateral quality, and use the principal proceeds to pay the senior classes. The subprimes will flood the market, which further depresses value, leading to a spiral or selling cascade.

Lenders also must protect their interests. Remember that even while assets may still be yielding an income stream, nervous lenders may put in redemption calls, when they would rather not take a chance on default. (Hedging strategies are expensive right now, too.) “If a fund blocks the redemption calls, lenders may threaten to sell collateral, preferring to take a small loss and move on to the next investment,” says Croke. Again, the market value dips, putting pressure on liquidity.

These bleak scenarios have prompted concerns that the subprime component will drag down the lowest rated, highest yielding tranches of the securitised pools, which could then expose the higher rated tranches to declines in value. But many experts disagree. “There just aren’t enough subprime loans out there to create such a problem,” responds Adelson. In the $10 trillion American mortgage market, about $1 trillion consist of subprime. Out of that $1 trillion, about 10% are going to become delinquent, and then only a fraction of those will foreclose. How many? Estimates vary from about 6% to 15%.

The analysis is not a binary one, whereby delinquencies automatically lead to default. “People pay in arrears all the time. As soon as we introduce the possibility of outcomes others than default, the numbers begin to drop,” says Shultz. Moreover, all the losses do not kick in simultaneously, as the process of foreclosure is a long one.

It is an overlooked piece of good news that we are still not seeing a rise in delinquencies in credit cards, which are “the canary in the mineshaft”, according to Schultz. One school of thought holds that consumers are withholding their mortgage payments for the very reason that they do not want to fall behind on their credit card payments and risk losing that access to unsecured debt. What that analysis does not take into account is that the credit card companies themselves are scrutinising an applicant’s mortgage history before they issue new cards.

To keep it in perspective, it helps to remember that a similar cycle in the American housing markets repeats itself every decade or so. After a massive collapse in the early 1990s, the pattern played out yet again in 1998. At that time, rumors swirled that Lehman Brothers was insolvent, and the value of MBS plummeted. Many lenders, who went out of business, no longer exist. Croke recalls, “Even then, the events only affected that specific market, rather than the whole system.”

Having experienced so many cycles, America has established various public and quasi public entities to offer protection, as buffers. Although UK housing markets are also volatile, no comparable institutions, like Freddie Mac, Fannie Mae or the Federal Finance Housing Board, exist in Britain. For example, after the Long-Term Capital Management (LTCM) crisis in 1998, Freddie Mac stepped up, to buy volumes of agency securities, thus fulfilling its responsibility to help stabilise the mortgage market. “What’s more, MBS constitutes eligible collateral at the Fed discount window,” Youngblood points out. “If needed, Bear Stearns could have even taken its triple As to the Fed and borrowed against them.”

Fears of a full-scale financial meltdown appear to have been exaggerated. Yet investors, too, need to renew their own balance of risk and reward. In the flurry of the past years, many lost sight of the reality that triple Bs are not bullet proof bonds. They latched onto the hope that right down to single Bs, there would be no danger of default. But even triple Bs still carry some risk. Wyss concludes,

“There is nothing wrong with buying risky securities, but the rewards must be sufficient to justify them. Recognise when you are gambling. We didn’t name these loans ‘subprime’ because we thought they were totally safe.”

Dodging the bullet

In late June, Bear Stearns, an investment bank, agreed to infuse $3.23 billion (later reduced to $1.6 billion), to bail out one of its two collaterialised debt obligation (CDO) hedge funds, the High-Grade Structured Credit Fund. As the largest scale rescue since 1998, the money is to supply a credit line, and replace loans from banks like Citigroup and Lehman Brothers. The hedge fund lost about 5% of its value in the first four months of 2007, while a related vehicle, the High-Grade Structured Credit Strategies Enhanced Leverage Fund was under water by 23%. By mid-July, both funds had become worthless.

“We are still uncertain about exactly what transpired at Bear,” says Michael Youngblood, managing director of asset-backed securities research at Friedman Billings Ramsey. The firm has not published a list of its holdings, and nor is it obliged to. At first, the fact that Bear did not move rapidly to internalise the issue rattled markets, by suggesting that the magnitude of losses might exceed its capacity to handle them. “It raised prospects of another LCTM,” Youngblood describes. For that reason, it was critical for Bear to make its pledge, to forestall a firesale that might have depressed the entire sector.

Yet within two weeks of the auction, the spreads on the securities had already begun to narrow. That shows how the market continues to function. When Bear finally rode to the rescue, it demonstrated that it had more than ample capacity to deal with it.

Britain has its own troubles

Britain’s non-conforming markets are likely to face higher levels of arrears during the next six months. However, the drivers are somewhat different from those across the Atlantic. For instance, Britain has not witnessed the same growth in second lien types of products.

“Both markets do share loosened underwriting checks, when it comes to low documentation (America) or self-certification (Britain),” says Stuart Jennings, head of EMEA RMBS, Fitch Ratings, in London. But the main culprit for British arrears will probably be the recent rise to 5% in interest rates, which will expose variable rate non-conforming borrowers to payment stress. “It will particularly affect those coming off discounted or fixed teaser rate products originated prior to the rate rises,” says Jennings. “This stress will be compounded should house price inflation slow, or even reverse, which would limit these borrowers’ capacity to refinance onto new deals.”