Friday, May 23, 2008

Friday epic: last of the Romans

(FT Alphaville) Bloomberg have an interesting story:

Citigroup Inc. created a $2.5 billion mortgage-backed security called Bonifacius Ltd. in August as capital markets seized up and panic swept Wall Street.

The issue took the name of a general, called by historian Edward Gibbon the “last of the Romans,'’ who fought and died for a fading empire. The bonds were created from subprime home loans as demand evaporated. Within six months, Bonifacius collapsed as homeowners fell behind on their payments in record numbers.

There are surely more etymylogical ironies out there when it comes to CDOs. Bonifacius means “good fate”, which is a prediction the investors in that particular CDO might take umbrage at.

The broader point of Bloomberg’s story though, concerns accounting rules, and the various ways banks - in this case Citi - handled exposures and structuring of CDOs off balance sheet.

Bloomberg somewhat gets the wrong end of the stick though, making a distinction between CDOs and VIEs (variable interest entities) like they’re apples and oranges. VIE is just a catch all term for vehicles, including CDOs. A CDO can be a VIE, or can become a VIE, depending on various factors, all of which, are explained in the gripping FASB Interpretation of Rule 46(R).

The key thing with VIEs is whether they are consolidated or not, and the FASB’s rules - introduced in the wake of the Enron scandal - deal with how VIEs must be consolidated onto the balance sheets of corporations/institutions/individuals other than the ones which nominally “own” the structure as traditional equity holders.

In short, VIEs are “owned” (and consolidated) by people other than the equity “owners” if…

1) The equity investment at risk is not sufficient to permit the entity to finance its activities without additional subordinated financial support provided by any parties, including the equity holders.

2) The equity investors lack one or more of the following essential characteristics of a controlling financial interest:
a. The direct or indirect ability to make decisions about the entity’s activities through voting rights or similar rights
b. The obligation to absorb the expected losses of the entity
c. The right to receive the expected residual returns of the entity.

3) The equity investors have voting rights that are not proportionate to their economic interests, and the activities of the entity involve or are conducted on behalf of an investor with a disproportionately small voting interest.

Take then, a CDO, structured in early 2007 by a generic Wall Street bank. The bank offloads all the equity and mezz tranches but can’t shift all the senior ones. So it creates a senior and a super senior set, allowing it to offload the (junior) senior triple A at a higher spread. With just the super-seniors unsold, it enters into a negative basis trade or puts them in an LSS (more on this one shortly).

The CDO, though it might still be managed by the bank, or one of its subsidiaries, doesn’t register on balance sheet though, since the equity noteholders are still the controlling noteholders.

As with most CDO documentation though, there’s a clause which gives the most senior class of noteholder - in this case the bank - control to decide what to do with the CDO (liquidate, accelerate, nothing) in the event that the CDO breaches a default trigger. When that happens, of course, under FAS 46R, the equity investors lose “the direct or indirect ability to make decisions about the entity’s activities through voting rights or similar rights”. Which is, of course, exactly what happened with so many CDOs. In the case of Bonifacius then, it’s not quite right for Bloomberg to be asking whether it was “put into a VIE”, since it was more as case of it just transforming into one.

Unless that is, we’re talking LSS conduits. Apart from through negative basis trades, one way banks could make use of and money out of, those dull, poorly paying super-senior tranches was to use… even more leverage.

It’s at this point that we start digging into the area where this whole financial crisis started. An LSS is a piece of synthetic wizardry: it’s a swap contract, written on the super senior. (Or else, the swap contract is the super senior - clarification on this point gratefully received.)

As a swap contract, of course, the up front cash cost is only a fraction of the contract’s notional value. Thus an “investor’s” participation in the contract might cost $10m, while the contract’s actual notional value (against the actual value of the super-senior tranche) is $100m.

How though, do you get investors to participate in those derivative contracts? Not easily. Far easier if you put those contracts up as collateral in a new structure altogether which issued some kind of fixed income note: voila, you have an LSS conduit.

You - the bank - want, of course, to find as broad a range of investors as possible, so you go to the broadest, deepest structured debt market there is: the asset backed commercial paper market. LSS conduits, like all those other ABCP conduits (SIVs etc), issued a load of commercial paper over and above a small equity tranche. As is common in ABCP conduits, you also provided a liquidity backstop to the CP roll to reassure investors.

LSS conduits got hugely popular in, of all places, Canada. As observed by Pimco:

At one point so many of LSS notes were arranged by Wall Street and Bay Street on behalf of Canadian conduits (thanks to the cheap financing provided by the ABCP investors) that they became the marginal price setter in the global market for credit risk!

All fine. But the cumulative leverage in this house of cards is incredible. And there’s a market value risk. For one thing, LSS conduits were laden with protective unwind triggers. Then there’s the nature of the LSS swaps themselves: the “notional” amount which the conduits stump up, like any collateral haircut, will increase dramatically relative to a small movement in price of the reference asset: the super senior tranche.

In other words, if spreads on super-senior tranches widened just a little, then there was going to be a lot of trouble. As rating agencies downgraded CDOs, of course, that’s exactly what happened: magnifying the impact of the crisis hugely:

Super senior tranches started to slip in mark-to-market value, thus causing all those LSS conduits to suffer net asset value losses. Some LSS conduits hit their unwind triggers. Extremely conservative commercial paper buyers suddenly pulled back from the markets, refusing to touch asset-backed paper issued by opaque structures. From Canada, the crisis spread to Germany, and from Germany, through the UK and the US, causing the SIV meltdown.

In the US, with banks like Citi, there was a knock on effect too, since they had provided liquidity backstops against the revolving CP issuance of those LSS conduits. As FT Alphaville reported at the time, Citi suddenly found itself forced into buying $25bn of commercial paper from LSS conduits it had dumped super-senior CDO debt into.

Back then to the original point: were banks deliberately gaming accounting rules? Possibly, but if they were, they failed. All this stuff has come straight back onto balance sheet.

If the VIE rules instituted by the FASB were indeed intended to stop institutions hiding exposures or losses, then Bloomberg is quite right: by their own yardstick, the accountants have failed. Bloomberg quotes Lynn Turner, who was the SEC’s chief accountant during Enron:

They never got the real problem fixed after Enron… When people find out how little FASB did, they’re going to be shocked. FASB needs to be taken out behind the woodshed and given a good whoopin’.

No comments: