Revisiting The CDS Menace

As readers of this blog know, I have written extensively about credit default swaps (CDS) and the risks involved. I explored several aspects, from the theoretical (CDS Factors in Equity Valuation, a four part series), to the more practical (CDS: Phantom Menace).

Today, I am revisiting the CDS subject because of a reader's comment to last Saturday's post. He/she said: "I am unwinding every ITM CDS* on Monday with all my counterparties". There is a ton of trouble brewing within that one simple, declarative sentence.

I assume the commenter is a CDS dealer/market maker, likely running a square book, i.e. not exposed to credit risk on a net basis. But he/she is exposed to counterparty risk and by unwinding positions he/she wants to minimize that as much as possible. After Bear Stearns' implosion, I totally understand and I bet dozens of other institutions are thinking along the same lines.

As with all insurance contracts, the ultimate risk in CDS is counterparty risk, i.e. when time comes to collect on the insurance (or the premiums, for that matter) the other side won't pay up. Since these derivative instruments trade 100% over the counter, there is no central clearer in between to guarantee payment. And this is where notional amounts come into the picture.

The conventional view on CDS is that enormous notional amounts should not concern us because they mostly cancel out between counterparties, greatly reducing the net market exposure. The Bank for International Settlements puts CDS notional amounts at $42 trillion, but with a gross market value of $721 billion (BIS data as of June 2007 - pdf). The International Swaps and Derivatives Association provides a slightly higher figure for notional amounts, at $45 trillion (ISDA data as of June 2007 -pdf).

But if one or more counterparties were to fail, they would set off a series of dominoes which could easily result in a daisy chain. Netting would then become a matter for the bankruptcy courts, taking years to resolve. Not exactly a favorable outcome when market volatility is measured in minutes.

One more aspect to explore is CDS pricing. Amounts outstanding exploded upwards during the last three years, i.e. when credit spreads were at historical lows. The chart below shows the notional amounts of CDS outstanding versus the yield spread between BAA corporate bonds and the 10-year treasury bond (click to enlarge).

CDS Amounts Outstanding and Corporate Yield Spreads (Data: ISDA, FRB)

We see that while credit spreads stayed at a low 150-170 bp, CDS amounts boomed from $8 trillion to $45 trillion. In other words, the vast majority of credit insurance was sold at very low prices. Credit spreads have now spiked to 350 bp and the CDX index for North American corporate CDS is at 185 bp, coming from a low in the 20's last June.

This means that marking all those CDS to market is currently resulting in very large valuation losses for those that sold the insurance. This further exacerbates the counterparty risk problem, particularly with those that made a habit of collecting such pennies in front of steamrollers, e.g. hedge funds. And this, even before credit events have begun in earnest, during what many are already describing as the worst crisis since WWII.

We may be in for very unpleasant surprises when those who sold insurance on the cheap are called upon to pay. Let me put it this way: if the large, professional credit monolines (MBIA, AMBAC, etc.) were not adequately reserved, what are the chances that Joe's Greasy Spoon and Hedge is?
______________________________________________

(*) ITM = in the money.