Investigating the GIC

Came across this article on Creditflux this morning. Think you’ll find it (and our comments below) interesting:

Single-name spread widening biggest factor in CSO valuation declines, says Lehman Brothers
News Digest, 3 September 2008

In a recent research report, Dissecting the losses in synthetic CDO investments and their implications for the credit markets, Lehman Brothers researchers point out that marks on corporate synthetic CDOs continue to deteriorate. They say a valuation below 50% of par is quite possible even if the reference credit portfolio is relatively high in quality.

They take the hypothetical example of an August 2006 10-year double A rated CSO issued at par paying 100bp over Libor at issuance and with collateral invested in a monoline GIC. The deal is linked to a 4-5% tranche of a 100-name portfolio with an average rating of strong triple B and an average spread of 42bp.

They estimate the current value of that deal as 42 cents in the dollar. Most of that fall in value comes from the general widening in credit spreads, which are on average 2.5 times wider than they were when the transaction was issued. The researchers calculate that this has contributed a 34% decrease in the value of the deal after taking account of the gain from convexity and moves in correlation. The decline in the value of the cash collateral has added a 20% loss in value with the balance from transaction costs.

The paper concludes that unwinds or restructurings of CSOs have the potential to send spreads wider and curves steeper and to cause credits which are popular in synthetic CDOs to underperform. But Lehman says that significant unwinds are unlikely in the near term unless there are unforeseen accounting or regulatory changes or large-scale changes of rating methodology.

Let’s just agree upfront that we’re not too surprised by these valuation numbers, but we’ll focus on the carefully hidden sentence: The decline in the value of the cash collateral has added a 20% loss in value with the balance from transaction costs.

We’re assuming this “cash collateral” is the GIC. Essentially, the AA tranche they’re describing suffers -- in addition to the pains of this liquidity crisis -- from the depletion of the eligible investments (i.e., the GIC), which has lost 20%. GICs are typically comprised of assets of the highest credit quality, usually with an emphasis on their being short-term assets, typically at least P-1 rated (short-term rating) and/or Aa3-rated or higher (long-term rating).

The GIC’s loss -- especially in a fully-funded transaction -- may itself be enough to wipe out the AA tranche, which detaches at 5%, even if the synthetic portfolio suffers no losses.

Perhaps it’s not too surprising that the AA CDO tranche is down 50 plus percent in mark-to-market losses, given the GIC – the supposedly most reliable portion of its investments is itself down 20%, and the circular relationship entailed. (Let’s only hope this 20% is a recoverable MTM loss, and hasn’t been realized.)