How to Resolve the CDS Basis Trade Blowup

Tyler at Zero Hedge has a wonderful post on the CDS basis trade today, which is a must-read for anybody who's interested in what happened to the CDS basis in the fourth quarter of last year or how Merrill Lynch could have lost so much money in December.

Tyler explains that while the CDS basis was positive during most of the Great Moderation -- it cost slightly more to insure a bond against default than you were receiving in coupon payments -- the basis on many names has become very large and negative over the months since Lehman Brothers collapsed.

Now a negative basis violates, in theory, the no-arbitrage rule: it means you can buy a bond, fully insure it, and lock in risk-free profits just by holding both the bond and the CDS to maturity.

But after Lehman and AIG blew up, that arbitrage trade became hard to put on, because prime brokers started asking for collateral from the buyers of credit protection:

Traditionally the margin requirements on CDS would be in the sub 1% range; after Lehman some counterparties raised the margin requirements as high as 20%, and others even asked for the whole margin to be paid up front. On a $10 million CDS position, which traditionally would only have cash outflows every quarter to fund the quarterly insurance payment, all of a sudden accounts would have to pony up to $2 million in margin just to put a trade on (or keep it on, leading to many basis forced unwinds).

This is truly a strange world: it's rare to ask people to put cash up front for the privilege of being able to pay an insurance premium every quarter. But as a result, the negative basis hasn't been arbitraged away, and the mark-to-market losses on anybody playing the CDS basis trade can be enormous -- yes, bigger even than the losses that Jerome Kerviel managed to rack up at SocGen.

A $15 billion loss could have been created as simply as experiencing a blow up on $25 billion on basis trades. And this assumes no leverage which is naive for the prop desk model: if ML had leveraged its pre-existing basis trades even 10x, the total basis trade notional needed to create this loss would have been only $2.5 billion. Is it inconceivable that ML had $25 billion in basis trades? Not at all - after all they were a preeminent CDS trading powerhouse and had one of the most active basis trade prop desks.

It seems to me that this is one area where government funds could be put to very good use -- and be guaranteed to make a profit. The government starts buying up lots of corporate bonds where there's a negative CDS basis -- Tyler cites CIT, Marriott Hotels, Home Depot, Temple-Inland and Omnicom as examples of credits where the basis is 300bp or more -- and then buys CDS protection on those bonds; it then promises to hold both the bonds and the CDS to maturity. Naturally, the government would buy the CDS protection on the new CDS exchange which it's trying to get the market to set up.

The result would be good for credit spreads, as bond prices would rise. It would be good for price discovery, as the confusion generated by the huge difference between bond and CDS prices would largely go away. It would lock in significant profits for the government. And it would get the new CDS exchange off to a flying start. What's not to love?