The SEC settled a little crisis-era CDO fraud case with UBS today and the fraud is pretty entertainingly shitty. Basically UBS provided the warehouse for a synthetic CDO where the notorious ACA was the collateral manager, and the disclosed deal was that, when the CDO closed, it would enter into (as protection seller) any CDS contracts that UBS had entered as part of the warehouse at (1) the market price of those CDS or (2) the price UBS had received for them as initial counterparty, whichever was more favorable to UBS.1 Now right there you've got some optionality and room for fuzziness, and you could imagine various unpleasant schemes where, for instance, UBS cherry-picks some contracts to transfer at market and some at historic price, or where UBS mis-marks some contracts to get a better deal when it transfers them.
But the actual scheme was simpler and dumber: the initial auctioning of the contracts was not - as was the norm for ABS CDS in those heady days of 2007 - just "bid the highest spread you're willing to pay for these contracts." Instead, ACA and UBS ran an auction where the running spread was fixed, at say 300bps, and the auction was just over the up-front points that the bidder would pay to enter into the deal. And so UBS ended up getting $23.6 million in up-front points. And then when it sold the CDS to the CDO, it did so for just the 300bps running, or 298bps really since UBS clipped a 2bps fee. The 2bps was disclosed; the up-front points - which amounted to around 971bps2 - were not. UBS just kept them, I guess on the theory of "well the counterparties are paying us 300bps for protection and we're paying you 298bps for protection, what is the problem?," which is a pretty dumb theory. You can see why UBS is exiting other derivatives businesses; their work here demonstrates the appropriate shadiness but an insufficient level of smarts.
Everyone involved in structuring the deal knew about this, especially ACA, especially Laura Schwartz, whom you may remember as the ACA executive who claimed not to know that John Paulson was shorting Fab Tourre's Abacus CDO, so you might wonder about her. Nobody saw fit to tell the CDO's investors, and when those investors asked everyone sort of fumbled and lied about it:
ACA employees were aware that UBS would not transfer the upfront points to ACA 07-2. During the ramp of this CDO, employees from ACA and UBS discussed in telephone conversations recorded at ACA whether UBS would transfer all, some, or none of the upfront points to the CDO. In one conversation, an ACA employee said that UBS would transfer the CDS to the CDO at a “mid-market” price and keep the rest of the upfront points (even though the price was not “mid-market” at the time). In another conversation about the upfront points, an ACA employee asked: “Is there, uh, 20 million dollars lying around?” The UBS employee responded: “There’s no 20 million. . . . We spent it already.” Finally, after a prospective investor learned of the existence of the upfront points, he was told by an ACA employee that UBS was keeping those upfront points as a “hedge” for itself.
Hahaha that's a bullshit answer but also not false! Because the inevitable occurred:
UBS ultimately was able to sell only $186 million face value of the CDO’s securities to 9 investors, for which UBS actually received $153 million because of discounts offered on such securities. UBS retained approximately $598 million of the CDO’s securities, including a $375 million super senior note, approximately $188 million in junior notes, and $35 million of the CDO’s equity securities. Only four months after it closed, ACA 07-2 issued a notice of default, as a result of the deterioration of the subprime mortgage-backed securities market, and ratings agency downgrades of thousands of RMBS bonds, including bonds referenced in ACA 07-2. At the time the CDO was liquidated in June 2008, outside investors lost approximately $130 million on their investments in this CDO.
So if investors lost 85% of their investment you gotta figure UBS's $223mm of equity + junior notes was pretty much toast. And I suppose a pile of cash is a good hedge to a pile of garbage, though on the other hand I believe them when they said "we spent it already." On other garbage I mean. It probably wasn't much of a hedge, in the event.
These CDO cases, man. It's hard not to be struck by the disconnect between the naughtiness involved, which is varied both in species and in unpleasantness, and the actual losses that befell their investors. Was Abacus "designed to fail," or was it a "darn fine portfolio"? Meh, the distinction between best and worst subprime CDO portfolios collapsed pretty quick in the summer of 2007. Did UBS and ACA screw investors out of $23 million with this stupid fee shenanigan? Oh absolutely, but the investors are probably more concerned about the $130 million they lost by buying subprime CDO notes just before the subprime crash. And, y'know. UBS seems to have lost more on this deal than all its outside investors combined - even before the fine! - while ACA, um, ACA is pretty much no more because it spent too much time hanging out with subprime CDOs.
Still! The thing is: if you found out about how UBS was surreptitiously eating all these fees, wouldn't you stay the hell away from this deal? That would have been a good idea. For UBS too, as it turns out.
1.From the SEC's order:
One requirement in the CDO’s indenture was that transactions be conducted “on an arm’s-length basis for fair market value.” The CDO’s offering circular also stated that CDS could only be acquired by the CDO if such CDS satisfied this arm’s-length, fair-market-value requirement, but it allowed the transfer of collateral into the CDO at the same price that UBS paid during the warehouse period. ACA could fulfill its fiduciary duty to the CDO by determining the fair market value of the collateral at close and transferring the collateral into the CDO at such price, or transferring the collateral into the CDO at the price at which it was acquired for the warehouse. This second option was the industry standard for this type of CDO.
2.That math is:
Through the first three BWICs, by early April 2007 UBS acquired $297.5 million notional amount of CDS with running spreads of 300, 350, or 375 basis points and with upfront payments totaling approximately $28.9 million. The remainder of the CDS referencing RMBS for ACA 07-2 were ramped by May 14, 2007, but no other positions were acquired for ACA 07-2 using upfront points. Because two of the CDS needed to be unwound before the CDO closed, the total amount of upfront points collected during the warehouse dropped to approximately $23.6 million. The total notional value of the CDO’s collateral when the CDO launched was $750 million.
So 28.9 / 297.5 = 9.71%, and I assume it all scaled down proportionally. 9.71% on just the relevant CDS with up-front points; it's a hybrid CDO so the rest of the collateral is either fairly bid CDS or physical bonds, I dunno.