Louise Story has penned what presents itself as an important story at the New York Times, one that charges Merrill Lynch with misrepresenting the size of its subprime, specifically, collateralized debt obligation exposures, in the runup to the global financial crisis. The ruse the article depicts is a CDO called Pyxis., which purportedly served as a dumping ground for exposures Merrill could not unload. Initially, Merrill was able to escape reporting these positions because it claimed to have hedged the risk. In fact, the hedges failed, the bank was ultimately on the hook and was later forced to 'fess up to the magnitude of its holdings. This revelation sounds juicy in that Citigroup and some of its recent senior executives paid fines to the SEC for similar, albeit less convoluted-sounding, misconduct.
But in fact, the story is astonishingly incomplete, to the point of being misleading. While Merrill's probable accounting improprieties are noteworthy and merit investigation by the authorities, they are not the most important element of this episode. CDO abuses amounted to accounting fraud to enable employees and executives to loot their companies. Moreover, they were not perpetrated by isolated actors, but were part of what Bill Black calls a criminogenic environment.
To put it more simply, if you think Merrill's misrepresentations to investors are a big deal, they are only a small aspect of the bigger, and frustratingly largely untold, tale of the role of CDOs in the crisis. CDOs were the epicenter of the upheaval, the device that magnified a what otherwise would have been contained subprime bubble into an economy-wrecking meltdown. When the music stopped, it was the dealers themselves that wound up holding much of the toxic paper they'd created. AAA rated CDOs went from haircuts of 2-4% in early 2006 to 95% in later 2007. The collapse in CDO valuations and the resulting inability to use CDOs as collateral for repo was a major, if not the major, cause of dealer illiquidity and insolvency which resulted in massive bailouts and backdoor subsidies.
Accounts like Ms. Story's are blind man and the elephant affairs: at best, they do a good enough job of depicting, say, the trunk, but leave the beast undimensioned.
The New York Times account muffs how the deal works, making it sound as if the only important actors were the bank itself and three Merrill traders (and it isn't even clear whether they acted as employees of Merrill, or were operating through affiliated vehicles).
Ms. Story somehow completely misses that the critical actor in these transactions was the sponsor of the CDOs, the hedge fund Magnetar, whose name appears nowhere in this article. A simple Google search of "Pyxis" and "CDO" turns up not only the Magnetar connection, provides links to deal lists (ours, recapping and adding to our discussion in ECONNED and ProPublica's), which show that there was not one Pyxis deal, as the story suggests, but two, a Pyxis 2006-1, issued in October, 2006 with Calyon as the underwriter (that deal is not the subject of this article), and the deal in which Merrill provided assets, Pyxis 2007-1, issued in March 2007, with Lehman as the underwriter. She also fails to mention Putnam, the CDO manager for the Pyxis deals, anywhere in the story or indicate that Magnetar's CDO deals were being put together by an analyst they hired from Putnam.
Why did Ms. Story not contact either of the two groups that own the Magnetar story, or at least review their work? The bizarre failure to mention Magnetar, which is integral to understanding this deal, suggests Ms. Story was going to some lengths to present the story as a new account, at the expense of reader understanding.
A key illustration of the shortcomings of the New York Times reporting: the description of the deal is so muddled as to be incomprehensible, which suggests not onlydid she not understand it, she did not consult anyone with expertise in the CDO market.
Nevertheless, the article does serve to provide some pieces of a mystery of sorts: how and why did Merrill manage to wind up with so much CDO dreck on its balance sheet when the bubble burst? But as you will see, it missed the context and the implications.
The conventional account is that Merrill's head of its CDO business, Chris Ricciardi, who had a history of trading out to a better paid job at the peak of a market (and his timing was again accurate), quit in early 2006. Merrill, determined to keep its leadership position, set a very aggressive transaction volume target, which resulted in it underwriting three times as many CDOs in 2006 as 2005. The amount Merrill was stuck with looked, superficially, like a pipeline problem: Merrill was making CDOs faster than it could sell them.
While the subprime market had gotten wobbly at the end of 2006, it went serious distress in February 2007. Everyone in the market knew that Merrill was long a ton of subprime paper, at least $10, perhaps as much as $20 billion. Thus the critical parties - investors and bond guarantors - knew that Merrill needed badly to lighten up on its exposures, and the fastest route would be CDOs, but since Merrill was expected to be putting a lot of supply into the market, their deals were certain to be bid wide (meaning at prices unfavorable to Merrill).
To make matters worse, it is likely that Merrill was even more long subprime exposure as of February 2007 that the Street surmised. Merrill had an active relationship with Magnetar, whose massive CDO program (named after constellations like Pyxis) has taken Wall Street by storm (industry sources and our own analysis suggest Magnetar's CDOs drove the demand for at least 35%, but more likely 50%-60% of subprime demand in 2006). The role dealers like Merrill played for Magnetar, among other things, meant that they served as the long side of the CDOs Magnetar created. So Merrill, as of February 2007, desperately needed to unload a ton of CDOs, both to offload residential mortgage backed securities exposures, and to bundle up tranches of unsold CDOs on its balance sheet (even first gen mezzanine CDOs contained up to 10% of other CDOs; so called "high grade" were as much as 30% CDO by early 2007).
Between March 1 and September 1 of 2007, Merrill issued just under $23 billion of CDOs - a record setting level. But in early March, the market was still wobbly, everyone knew Merrill was long, and there was no assurance it would be able to dump as much dreck as it wanted to (even this $23 billion was short of what Merrill needed to offload). So using another outlet, in this case, pushing some of its exposures out through a Magnetar deal underwritten by another firm, would have been a very appealing idea.
During this timeframe, even more remarkably, MBIA insured $10.8 billion of the Merrill CDOs, a phenomenal amount of new CDO exposures in a short window. This, plus the 2007 Pyxis deal (insured by FGIC), would explain why Merrill thought they had hedges in place which ultimately turned out to be worth less than they thought. Merrill was an investor in ACA. It appears likely that Merrill used ACA as an additional insurer for CDOs it was dumping in 2007 to get out of its massive pipeline. ACA blew up by the 4th quarter of 2007, which would be another reason Merrill's "hedges" on its exposures had failed.
So what is troubling about this transaction? Two issues stand out.
First is the way some critical parties are unlikely to have operated on an arm's length basis. One of the critical assumptions of a bond guarantor (and Pyxis 2007-1 was insured) is that the manager who is selecting the bonds, in this case Putnam, in an independent party. During this frenzied stretch of activity, Merrill underwrote one deal in March for Magnetar, the notorious Norma, which a Wall Street Journal makes clear the manager was not even remotely independent. Just like ACA in the Goldman 2007 Abacus trade, the manager of a Magnetar-sponsored trade took its marching orders from the supposed sponsor (in the Abacus case, the famed subprime short John Paulson, although curiously he never stepped up to actually play that role, but participated solely on the short side). It is even more likely that Magnetar had considerable influence over what exposures went into the CDO if the objective of Pyxis 2007-1 was, as it appears, to allow Merrill to dump CDO and RMBS exposures through a hidden channel.
How is it possible that Merrill's MBS exposure ended up in a Lehman led, Putnam managed Pyxis CDO if not for Merrill's connection to Magnetar? And if this is so, how plausible is Magnetar's previous denial that they played any role in influencing the bond selection process in deals on which they were the equity sponsor?
Second is the motivation for Merrill to get itself into this mess, which is bonus fraud. For the most part, the firms that wound up with a lot of CDOs on their balance sheet were Eurobanks (Citibank is a special case, in that it was the biggest player in SIVs and had to take those transactions back). The Eurobanks, thanks to an unfortunate interaction of Basel II rules and internal metrics that rewarded managers and traders who found ways to use less capital in their businesses, incentivized dealers to engage in the so-called "negative basis trade." If a trader bought or retained an AAA tranche of a CDO and hedged it (at some banks in full, in others, merely in part), the difference between the income on the CDO and the cost of the hedge over the life of the trade was discounted and included in the trader's P&L. In what other business is it legit to be paid bonuses on yet-to-be received profits?
Ms. Story alludes to bonuses as a driving factor in the alleged disclosure fraud, but this is merely an aside, when it should be a central issue. No one, in particular regulators, seems willing to look at how firms paid particular actors and executives for deals that went bad, particularly ones that were questionable even at their inception.
With this as background, the supposed revelation is framed, per the title, as "How Regulators Unearthed Merrill's Dodging of Risk Disclosure" (later revised to "Merrill's Risk Disclosure Dodges Are Unearthed"). This is laughable. The regulators certainly did not unearth this story. But given Ms. Story (and Gretchen Morgenson's) close connection with the SEC, the agency is probably a significant source for this story. The SEC appears to be eager to show progress on pursuing bad behavior during the crisis and has latched onto the failure of various dealers to make proper disclosure of their subprime exposures.
But the real story is about the widespread corruption in CDOs and about Magnetar. The SEC appears reluctant to pursue this angle and is using Ms. Story to focus on a different, narrow aspect. It is embarrassing that the Times cites the widely derided Citi settlement as a strong precedent. But this is consistent with SEC patterns - extolling their success with Citi (such as it is) to show the advances they are making. The SEC's strategy is evidently to limit its effort to the narrow issue of adequacy of disclosure, which is the common thread among the Goldman, Citi, and Merrill cases. Appallingly, that blinkered approach means the agency is likely to give Magnetar a free pass, since the SEC apparently regards Magnetar to have made adequate disclosure of conflicts of interest. But presenting CDO managers as working on behalf of all the investors when they were in fact serving the interests of the sponsor, whose true interests were opposed to all the other investors, is a separate basis for investigation and action, but one the SEC has evidently decided not to pursue.
Thus what the SEC tries to publicize as progress is grossly inadequate and misses the drivers of crisis- of widespread fraud throughout the CDO market, which involves price manipulation, inadequate disclosure of parties and intentions, and tying among transactions, in addition to corporate level misleading statements regarding mortgage exposure.
The SEC, via Ms. Story, appears to display a desire to get credit for investigating financial crisis issues without addressing the problem with the CDO market head on and almost seems afraid to capitalize on the progress they made with Goldman on the Abacus deal. Are they afraid what they will uncover (might it implicate Deutsche Bank, which would undermine the credibility of SEC enforcement chief Ray Khuzami, who was General Counsel for Deutsche Bank in the US, which along with Goldman, was a leader in synthetic CDOs)? Or is it simply that they felt these CDO cases are too much work and they'd rather go back to the usual SEC low hanging fruit of insider traders? Given how central abuses in the CDO market were to the crisis, neither of these is good enough reason for the SEC to punt.
Update and correction 4:45 PM:
We owe a bit of an apology to readers and to Louise Story of the New York Times, for an apparent error in our analysis. We have been informed that, remarkably, there were two separate Pyxis vehicles which were issued in the 2007 time frame one of which was a CDO and a second a structure which some sources classify as a CDO and others describe differently.
The Pyxis 2007-1 transaction which we discussed in the post was a CDO but did not have any Merrill involvement. The Pyxis deal referenced in the NY Times article that was issued by Merrill, Pyxis Master Trust, Class 2007, was an entirely different vehicle and was underwritten by Merrill
The brief description of the Merrill-devised transaction in the NY Times article, which was frustratingly vague, included references to "notes" being issued, which we took as an indication that it was an external transaction with investors and corresponded to the Magnetar-sponsored CDO, since there was no evidence of any other deal named "Pyxis" being marketed during this time period. It appears that the Merrill transaction was actually an internal transaction.