"Merrill and the Basis Trade
In a basis trade, an investor buys a corporate bond and simultaneously buys CDS protection on the bond. There's a "negative basis" when the price of the CDS is lower than the price of the bond (usually defined by the bond's asset-swap spread or Z-spread). In other words, the basis is negative when the coupon payments the investor receives on the bond are higher than the premiums the investor pays out on the CDS. Conversely, there's a positive basis when the price of the CDS is lower than the price of the bond. The basis trade, and trades like it, have been extremely popular on Wall Street in the past few years. Deutsche Bank's "star trader" Boaz Weinstein supposedly lost over $1bn on basis trades when the corporate bond market froze in September.
From what I hear, Merrill's biggest losses didn't come from basis trades per se, but came more generally from its massive exposure to Lehman as a counterparty on CDS trades. Now, some of Merrill's exposure to Lehman as a counterparty definitely came from CDS trades that were part of a larger basis trade. Remember, Merrill surprised the Street in October by reporting a $2bn pre-tax trading loss in Q3 that was, according to Merrill:
Primarily related to the default and spread movements of certain government sponsored entities and major U.S. broker-dealers.Any CDS contract with Lehman as counterparty needed to be replaced when Lehman collapsed. Remember the emergency "Risk Reduction Trading Session" the ISDA opened on the Sunday that Lehman was preparing to file for bankruptcy? That was so that the major dealers could start replacing their derivative contracts where Lehman was the counterparty with contracts with other counterparties. The emergency trading session was a disaster (about which more later), but it was especially disastrous for Merrill.
Remember the timing.
The emergency trading session ran from 2 pm to 6 pm on the Sunday that Lehman was preparing to file for bankruptcy. The news that BofA was in advanced talks to buy Merrill didn't break until around 5 pm. Before that announcement, everyone was looking for who would be the next to fall, and the consensus was that Merrill was next in line. So traders turned their guns on Merrill. In a standard CDS, no money is exchanged upfront. But when a firm looking to enter into a CDS contract might default before the contract matures, counterparties start to demand money upfront—known as "points upfront" or "initial margin." Since Merrill was expected to collapse if it couldn't find a buyer, and news of the BofA deal had not yet leaked, counterparties were demanding that Merrill make huge upfront payments in order to enter into a CDS contract. And because Merrill had huge exposure to Lehman as a counterparty, it had a lot of contracts it needed to replace.
I'm not saying that all of Merrill's losses, or even most of its losses, came in the emergency trading session—it took banks weeks, if not months, to replace all their trades with Lehman. The emergency trading session was just a microcosm of the next few weeks/months. Upfront payments have become relatively common in CDS trades since Lehman's collapse. Even after the BofA deal was announced, counterparties were still demanding points upfront from Merrill (though not as much as in those dark hours before the BofA deal was announced).
My feeling is that a good chunk of Merrill's losses came from the sheer number of CDS contracts it had to replace when Lehman collapsed. Merrill had a lot more CDS contracts with Lehman than anyone thought—Moody's said the number was "outside of our expectations." Merrill has taken a loss on virtually every one of those contracts, as it had to replace them with pricey off-market CDS, some of which required sizable upfront payments. And since Merrill's $2bn Q3 trading loss only reflects the CDS contracts it replaced before Sept. 30, it's a good bet that these losses extended into Q4.
While CDS that were part of basis trades no doubt account for some of the contracts Merrill had with Lehman, I don't think the basis trade is really the culprit here. The culprit is Merrill's outsized exposure to Lehman as a counterparty.
Addendum: When the ISDA announced the emergency trading session on the Sunday that Lehman collapsed, I had a good laugh, and took great pleasure in explaining to my wife all the reasons why the trading session, as well as the following few weeks in the CDS market, would be a clusterfuck of the highest order. At that point I was still a "former structured finance lawyer," so the situation was amusing to me. Of course, a few hours later my old boss called and said: "I need you to come back. This is gonna be Armaggedon." Since he's one of the nicest men on the face of the earth, I couldn't refuse. Karma is a bitch.
Here's my reply:
"Since Merrill was expected to collapse if it couldn't find a buyer, and news of the BofA deal had not yet leaked, counterparties were demanding that Merrill make huge upfront payments in order to enter into a CDS contract. And because Merrill had huge exposure to Lehman as a counterparty, it had a lot of contracts it needed to replace."
What you are describing is what I call a Calling Run. It's just my variation of Fisher's work on Debt-Deflation. From what you wrote, I am assuming that anyone familiar with the idea of a calling run could see the seeds of it when Merrill was hit as you described. From my point of view then, anyone familiar with these concepts could plainly see that allowing Lehman to fail could start a Calling Run, where investors started demanding money back in a ripple effect sort of manner, since that was what the Merrill reaction was. Assuming that you can follow my reasoning, since I'm not an expert in these matters, I am right about this?
Don the libertarian Democrat (I'm stuck with this name now, like the Goo Goo Dolls.)