US sub-prime: ABS CDO pipeline dries up as sub-prime fears start to bite


Alex Chambers, Louise Bowman
Published on:

Hysterical headlines and wild index swings have disguised what is really going on in the world of home equity loan securitization.

There has only been one topic of conversation in the structured finance market in recent weeks – and for once it was the same subject that everyone else in the financial markets is talking about – US sub-prime mortgages. Problems at originators such as HSBC, Fremont, Accredited Home Lenders and New Century Financial have generated negative headlines around the world. As far as the structured finance market is concerned, the key issues are what will be the impact on existing and future ABS and CDO transactions?

For most banks the whole rationale for origination of sub-prime mortgage loans is predicated on the ability to shift much of the credit risk to the capital markets – indeed, HSBC’s problems partly stem from the fact that Household Financial Corp had not securitized a large portion of its exposure. But even the houses that have been big securitizers and originators of this collateral – such as Lehman, Bear Stearns and Merrill Lynch – were originally thought to be facing serious problems. These concerns stemmed from the belief that they might have retained residual interest in those securitizations, and also have large mortgage warehouse lines that would need to be seasoned before they could be repackaged into a securitization.

But first-quarter results from the brokers showed few ill-effects from sub-prime exposure. It appears that much residual risk has been sold on to third-party first-loss investors and that much of the warehousing risk has been hedged in the CDS market.

"Thank god we had a synthetic market – before the introduction of synthetics, dealers had no way of expressing short positions. The introduction of single-name CDS and ABX allowed dealers to run client-facing inventory and manage credit risk," says David Martin, global head of MBS and ABS at UBS.

Bankers say that many investment banks were shorting sub-prime risk even before ABX took its sharp downward turn this year. The ABS synthetic market’s liquidity grew substantially in the US during 2006. So while everyone was expecting banks to take a hit from warehouse lines (average arrears in unseasoned US sub-prime loans are between 5% and 10% but average 90-plus day arrears have now reached 25%), prudent use of warehouse insurance has seemingly cut risk exposure.

Having successfully negotiated this dislocation, the leading mortgage players are confident of the newer ABS vintage. "Newly issued securities in sub-prime will trade up substantially. Since the start of 2007 new sub-prime loans stopped being granted without proper verification of incomes and without people putting down deposits. So the underlying loans are much better from 2007 than they were in 2006," Warren Spector, president and co-chief operating officer at Bear Stearns, told Euromoney.

Spector argues that this year, as well as the quality of loans being better, the rating agencies are much stricter about how they rate deals. What is a triple-B today is a very different security to what was a triple-B a year ago.

Unknown, unknowns

But that is not the end of the story because many market participants fear there is more bad news to come. Furthermore, future events including the path of interest rates and house price appreciation (HPA) will determine whether the bad news becomes atrocious.

"There are problems in the US mortgage market – but how deep they go nobody knows," reckons Alexander Batchvarov, international structured finance strategist at Merrill Lynch. "Historically problems arise when unemployment rises and HPA falls. We now have stable employment and an economy that is still growing. We do not know how many sub-prime borrowers will be unable to refinance, we don’t know how far HPA will fall and we don’t know whether the US is going into recession or stabilizing. And even if people default this year we won’t know what the impact will be as losses will not start to crystallize until mid to late next year." US HPA fell from 13.2% to 3.4% in the third quarter of 2006, its steepest deceleration in 30 years.

"Much depends on house prices. If they start falling, the market could be in trouble," says a US banker. "Equally, if it becomes much harder to get sub-prime mortgages, that will impact the market as there’s no one to buy your house and the market could begin to spiral downwards. But the market needs to make the distinction."

Originations of single-family mortgages

Conventional and government-backed mortgages, 1990–2006 Q2, by loan type

Source: OFHEO

Or, as Spector puts it: "Right now, everything is being pushed down together, and I don’t think that’s right."

The US sub-prime sector now accounts for 12% of all US ABS, so huge volumes of sub-prime loans have found their way into the market, particularly as collateral for mezzanine CDOs of ABS (a sector where quarterly issuance topped $20 billion in the second half of last year). Indeed, according to Credit Suisse, home equity or sub-prime accounts for some 80% of ABS CDOs. CDOs backed by 2005 and 2006 vintage loans – which suffer from weaker underwriting standards and a higher percentage of adjustable-rate loans and second-lien loans – are under increasing scrutiny.

According to analysts at JPMorgan, ABS CDOs can withstand 2% to 4% of losses before overcollateralization (OC) tests are triggered and 8% to 10% of losses before there is any principal loss to the triple-B tranche. Forced sales would not kick in until a bond had not paid interest or had been triple-C rated for a year or two. Thus, the likelihood of actual losses is remote at present. Credit Suisse analysts point out that $6.8 billion of mezz ABS CDO tranches (typically rated single A to double B) are PIK-able and will be seriously affected if they fail to meet OC and interest coverage (IC) tests. This is because cashflows are then diverted to pay senior note holders but missing interest is added to the balance of the PIK-able tranche without even trigging default.

The sector has already experienced a sharp (and perhaps long overdue) spread widening that will have had severe implications for the growing numbers of ABS investors that mark to market. The first-loss investors in ABS CDOs are now hedge funds, proprietary trading desks and pension funds, all of which will suffer mark-to-market losses due to spread widening. Many credit hedge funds have typically gone long CDO equity and short the higher tranches of CDOs and they will suffer losses if not completely hedged. Many funds have also set up permanent capital vehicles to invest in RMBS and CDO equity and these vehicles will be particularly exposed to the downturn because of their concentration of exposure to the bottom end of the capital structure.

The ability to hedge out exposure to sub-prime assets has been made much easier over the past 18 months by the development of the ABX indices and single-name CDS. But the behaviour of the index – particularly the triple-B and triple-B minus ABX 06-2 and 07-1 (which have exposure to the largest sub-prime deals from 2006) – has attracted some criticism. Indeed, the index was described as an "emotive indicator" by one structured finance banker, who thought that the wild swings in the ABX were a diversion from much of the real action.

Markit admits that the ABX is clearly overstating the risks. "The ABX has widened further than fundamentals might suggest it should," concedes Ben Logan, managing director, structured finance, at the firm. "It is a young product and will have more liquidity as time goes on." There has been remarkable tiering between issuers, with as much as 100bp separating different names at the same rating.

"Having a benchmark like this to convey information to the market is a good thing – many within the HEL market believe that it has outstripped fundamentals," Logan maintains. But others argue that the index – particularly because of its composition of the 2006 vintage – is misleading and at worst contributes to market panic. "The ABX is a one-way market and TABX is even worse," says Batchvarov at Merrill Lynch. "It is an illiquid, phantom product that is divorced from the fundamentals of the market. Look at the cash market – triple-B minus risk is 300bp, not 1,000bp. The people who really look at this risk (the buy-and-hold investors) understand the fundamentals. The ABX is just hedge funds taking a view and creating huge volatility in the market." Real questions remain as to the suitability of the mechanisms dealers use to price the ABX, but for anyone long sub-prime risk in any way the index offers the easiest and most liquid way to hedge positions – and that is why it has been so hammered.

Adjustable rate mortgage arrears as a percentage of original balance

60 days

Source: Moody’s, SG Credit Research

But it is still unclear what has happened to an estimated $60 billion to $90 billion of unsecuritized mortgages from the 2006 vintage. Only the bankers at the exposed institutions really know how they have managed these positions – it could be that some of this has already been repackaged and will be sold down if and when the market has firmed up.

Not surprisingly, negative headlines and a collapsing index have cut CDO appetite for HEL. And reduced appetite at the bottom of the capital structure is bad news for the top as well. The first couple of months of 2007 were unusually busy in the ABS CDO market, with 17 mezz ABS CDOs launched in February alone. Originators knew that some form of spread widening was coming and wanted to get deals out of the door as quickly as they could because of the warehousing risk. Now that the correction has come, many outstanding deals have been shelved (as many as 12 according to market participants) and the ramped assets are being sold.

In theory, the widening of spreads in the underlying mezz collateral should have made the arbitrage potential of ABS CDOs more attractive, but as the liability spreads have also sharply widened (triple-B ABS CDO tranches have gone from 325/375bp over in December last year to 550bp to 600bp over now) this potential has dissipated.

Analysts do not expect the hiatus in the market to be long term, but issuance of ABS CDOs during 2007 will probably be a long way short of 2006.