Picking on Pyxis

By Felix Salmon
August 10, 2010
Louise Story continues to talk to the SEC, and it seems that Merrill Lynch is now in the spotlight, and specifically a vehicle it underwrote named Pyxis. I wonder whether Merrill/BofA, like Goldman before it, has received a Wells notice and hasn't bothered to disclose the fact.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google,mail" data-share-count="false">

Louise Story continues to talk to the SEC, and it seems that Merrill Lynch is now in the spotlight, and specifically a vehicle it underwrote named Pyxis. I wonder whether Merrill/BofA, like Goldman before it, has received a Wells notice and hasn’t bothered to disclose the fact.

Just like in the Goldman and Citigroup cases, the SEC seems to be concentrating on disclosure problems here: specifically, that Merrill failed to disclose to its own shareholders the extent of its subprime exposure. (Yves Smith thinks this doesn’t go nearly far enough, but it’s unclear what other kind of case against these banks would be legally colorable.) Failure to disclose to shareholders is bad, but it’s Citigroup bad, not Goldman bad. The fraud which the SEC alleged against Goldman was failure to disclose to direct investors in securities that the bank was underwriting, which is worse.

I would love to see a bit more reporting on the Pyxis case, though, especially now that it seems to have caught the attention of the SEC. Specifically, Jesse Eisinger and/or Jake Bernstein should write something on this: they have covered the Magnetar story in great detail, and Pyxis was a Magnetar deal.* They, I think, are probably better qualified than Story to explain clearly exactly what’s going on here. Because I, for one, am very confused.

Story writes of Pyxis, and similar entities named Steers and Parcs:

These programs generally issued short-term I.O.U.’s to investors and then used that money to buy various assets, including the leftover C.D.O. pieces.

But there was a catch. In forming Pyxis and the other programs, Merrill guaranteed the notes they issued by agreeing to take back any securities put in the programs that turned out to be of poor quality…

To further complicate the matter, Merrill traders sometimes used the cash inside new C.D.O.’s to buy the Pyxis notes, meaning that the C.D.O.’s were investing in Pyxis, even as Pyxis was investing in C.D.O.’s…

To provide the guarantee that made all of this work, Merrill entered into a derivatives contract known as a total return swap, obliging it to cover any losses at Pyxis. Citigroup used similar arrangements that the S.E.C. now says should have been disclosed to shareholders in the summer of 2007.

This makes it sound very much as though Pyxis was an SIV, much like the notorious Citigroup entities which had embedded “liquidity puts” and which forced Citi to take enormous losses even though they were nominally off Citi’s balance sheet.

SIVs, after all, were exercises in maturity transformation: they would issue short-term IOUs and then buy longer-term securities like those issued by CDOs. That’s exactly what Story says was going on here. And in some cases, like Citi’s, the parent bank would be on the hook for losses at the SIV, which is also what Story says happened at Merrill.

But as far as I can make out, Pyxis was not a Citi-style SIV, but rather an old-fashioned CDO. And after Story’s single reference to its “short-term IOUs”, she never once mentions them again, so it’s far from clear what she’s talking about. I’ve certainly never heard of a CDO issuing short-term paper.

I suspect — but I’m not sure — that the big picture is relatively simple. Merrill created Pyxis, and in doing so managed to move a lot of subprime exposure off its balance sheet. But somewhere in the deal was a mechanism for those exposures to bounce back onto Merrill if things went very bad. And that’s what happened. Most likely, Merrill used the Pyxis deal as a way of getting its subprime exposure insured by MBIA and/or ACA. But that insurance became worthless when both companies imploded.

If that’s the case, however, then I reckon the SEC case is pretty weak. If Merrill had a subprime exposure which was insured by MBIA or ACA, I think they could make a strong case that the exposure did not need to be disclosed to shareholders. But if it turns out that the Pyxis affair is more than just a bad choice of insurers on the part of Merrill, then it would be great to get some much clearer detail on what went on than I can find in Story’s article.

*Update: Pyxis was not a Magnetar deal after all, I was confused by Yves Smith’s post. There were a couple of Magnetar trades with that name, but they had nothing at all to do with Merrill. The Pyxis that Story is talking about was a Merrill SIV, and not a CDO at all. Which makes much more sense.


Comments are closed.