Everyone expected a lot of US sub-prime mortgages to go sour and everyone knew that leverage in the LBO market was getting ridiculously high. But the liquidity crisis that hit the ABCP market in August was something that almost no one saw coming and was greeted with panic and bafflement in equal measure. "We sat down early in the month and said How far can this thing spread?" reveals an ABS veteran. "We knew that the short-term money markets would be impacted, but it seemed that the entire ABCP investor base getting up and walking on 30-day paper was something that just could never happen."
So why did it? There must be a pretty good reason for the markets to get to the stage that banks will not even lend overnight to each other. Many see the problems in ABCP (and structured investment vehicles, SIVs) as the result of a chronic dearth of transparency. "When CP investors were trying to see what was in the conduits the information was so vague that they just panicked," says an observer. "If you are simply told that the exposure to CDOs is X but you are not able to find out what type of CDOs these are or what is backing them then you will assume the worst and run."
Thus anything involving a three-letter acronym be it ABS or CDO became toxic overnight. As Matt King at Citi noted on August 16: "The main fear becomes fear itself... Of course there is no fundamental reason why the ABCP buyers should take fright the assets themselves are generally sound and most of the banks are very well capitalized but the greater the fear in the system, the greater the potential for problems."
Those problems led the ABCP market to evaporate at an alarming rate. According to UBS, US dollar ABCP outstandings were down by $48 billion for the week ending August 15, down $77 billion for the week ending August 22 and down $59 billion for the week ending August 29. The yield on 30-day CP went from hovering between 25 basis points and 75bp to an astonishing 356bp on August 20 (by August 28 it had come down to 155bp). "The market moved beyond mathematical reason and into emotion," said Douglas Connor, CP trader at UBS in a conference call on August 30.
IKBs inability to provide committed liquidity to its Rhineland Funding conduit certainly spooked CP investors across the board. And despite the fact that problems in the Canadian market were largely the result of a now-defunct structuring peculiarity, by the end of August they remained far from comfortable. "Paper is still trading but not at the size, speed or duration that is was before," says Connor. The main problem was that curves had gone completely flat, so there was little incentive to go out further than a week or two. But in this environment the structural differences between programmes has suddenly become very important and tiering will be very evident in the months to come. "The market has gone back to people rolling their sleeves up and deciding which structures are good and which are bad," says Connor.
Given that German banks hold approximately one-fifth of outstanding liquidity lines to ABCP conduits ($250 billion according to Fitch), it is no surprise that the first two vehicles to run into trouble were German (Rhineland Funding and Sachsen LBs Ormond Quay). The effect of not only a funding crunch on the liability side but having to mark to market on the asset side (when that market has disappeared) means that others will likely follow. Indeed, at the end of August there were signs of stress in other, more aggressively structured, ABCP programmes.
Providing 100% or even more in some cases liquidity to an ABCP conduit can be an expensive business. Under Basle II these lines can be risk weighted up to 20% so the market has spent the past couple of years devising ways in which to reduce liquidity requirements. One of these is to issue extendible CP that for a higher yield can be extended in a period of market disruption. The idea is that if CP funding is disrupted the extension period can be used to sell assets bringing liquidity back into line again. This line of thinking doesnt look quite so smart in the recent environment, where not only is CP squeezed but also there are no buyers for the assets that the conduit is forced to sell.
Extendible ABCP programmes have proved very popular with mortgage originators, and by the beginning of September eight such vehicles had been forced to extend their CP. They are backed by American Home Mortgage; KKR and Deutsche Bank; Luminent Mortgage Capital, Aladdin Capital; Lord Securities and Conduit Management Corp (Australia) and Thornburg Mortgage. How these programmes fare will in part depend on whether they rely on a market value swap or purely on overcollateralization. If there is a market value swap in place, investors are protected as the liquidity provider makes up any shortfall between market and asset prices. However, investors in programmes that rely on overcollateralization depend on asset sales to achieve liquidity and so could run into trouble. The KKR/Deutsche Bank, Aladdin Capital, and Thornburg Mortgage programmes are structured in this way.
The expen-SIV option
The liquidity freeze in the CP market very quickly shone a bright light into a usually dark corner of the structured finance market that of the structured investment vehicle.
The most obvious distinction between ABCP conduits and SIVs that the market has leapt upon is that conduits tend to have 100% liquidity support and SIVs as little as 10%. But the two are very different: a SIV is a very much more expensive proposition than an ABCP conduit and takes far longer to establish. There are roughly 30 SIVs in existence following a surge in activity in the sector over the past couple of years. But the sheer cost and infrastructure required to manage a SIV have checked the growth of the market.
Unlike ABCP conduits, SIVs (which operate like mini-banks and indeed are often referred to as non-bank banks) are investment companies set up to provide returns to capital note investors, not just to exploit securities arbitrage. Those capital notes are levered roughly 15 times and enable the vehicles to operate with far lower liquidity back-up. According to Lehman, the SIV market has roughly $350 billion outstanding, of which around $90 billion is non-bank sponsored.
Rolling with the Punches
European ABCP sponsors
Source: Moodys Investors Service
This is not the first time that SIVs have run into trouble. Two vehicles, ABC and AbAcAs, struggled in 1999 following the financial crisis in southeast Asia and the collapse of a film-financing investment, Hollywood Funding. In that case the vehicles survived but the capital note investors were badly hit.
In the early days of the SIV market, structures were set up with an Armageddon-style defeasance option whereby any breach of the capital adequacy trigger meant full and irreversible unwind within five days. SIVs established in the past two to three years have more staggered unwind mechanisms.
For years, one of the main problems that SIVs have faced is the tightening arbitrage available to exploit between their assets and liabilities. It is for this reason that the older vehicles (which bought assets at far wider spreads than those set up recently) are on a far stronger footing. This was brutally demonstrated by the news on August 31 that the mezzanine capital notes of Rhinebridge PLC a SIV that had only been established a couple of months earlier had already been downgraded (although given that IKB is the manager it was hardly unexpected). Cheyne Finance, which entered defeasance in August, was only set up in August 2005. These two SIVs had highly concentrated portfolios of assets in risky asset classes, which contributed to their rapid demise. But there were a string of new SIV launches in 2006 (from HSH Nordbank, Citi, MBIA and IXIS, for example), which must now be under greater spread pressure than the older vehicles. But even the older, larger vehicles (such as Beta Finance ($20 billion), which was set up in 1989, the $52.6 billion Sigma Finance (1995) and $29 billion K2 (1999), cannot operate with CP spreads where they were in early September for long. Vehicles with 10% liquidity support need to shed assets if funding comes under strain. And shedding assets in this market is just about the last thing that they will want to do. Citi, which dominates the sector, revealed that it had sold $5.3 billion of assets from its seven SIVs during August something that it will not have done by choice.
If the tough market conditions persist, the likelihood is that the majority of SIV-lites (total market $7.6 billion) and extendible ABCP programmes (total market roughly $25 billion) without market value swaps will be restructured or liquidated. Analysts at UBS estimate that $50 billion to $75 billion of ABCP could be unwound and moved into repo lines of liquidity providers. This market, which was between $1.3 trillion and $1.4 trillion before the liquidity crunch, will take time to get back on its feet. It might never return to the sub-Libor levels of the old days but liquidity should return for the better-structured vehicles (or those with 100% liquidity backstops). "A1/P1 should not be trading at 40bp to 60bp over Libor," says one exasperated observer. "Maybe at some stage people will start to realize that being 100% T-bills doesnt entirely make sense."
SIV-lites: an over-sold disaster story
When terms such as SIV-lite become common currency in the mainstream press, it is clearly time for a reality check. In late August high-profile coverage suggested that the restructuring of Cairn Capitals SIV-lite had averted a full meltdown of the entire financial system or, at the very least, Barclays Bank. These vehicles which are akin to market-value CDOs funded in the CP rather than term market swiftly assumed the role of public enemy number one.
SIV-lites have been very badly positioned in the CP funding squeeze as they do not have 100% liquidity support, were set up in the past 18 months a period of unprecedentedly tight asset spreads and operate at high and fixed leverage (unlike traditional SIVs which can dynamically adjust leverage levels). The fundamental flaw in the structure was that SIV-lites used traditional SIV liquidity support technology to back portfolios of assets that are far more risky than those that should normally found in SIVs and would more commonly be funded in the term market via a CDO. The structure relies on being able to sell assets to ease liquidity problems something that swiftly became very difficult as the bid for these types of ABS assets vanished. SIV-lites are unlikely to be used again in their current form, but the market in its entirely consists of just six yes, six deals. Even if every SIV-lite was to liquidate its entire portfolio, the impact on the structured credit market overall would be largely contained.
Aside from the small volume of SIV-lite issuance, roughly $77 billion of high-grade structured finance CDOs have also funded their triple-A tranches in the CP market (according to JPMorgan). These deals, which were most popular in 2004 and early 2005, usually have third-party liquidity backstops and fell out of favour when term funding became cheap enough to compete with the CP alternative.