Skip to main content

Court Finds That UBS Exploited Clients And Tricked Them Into Overpaying For Toxic Securities, But In A Legal Way

  • Author:
  • Updated:

This is kind of exciting: a bank won a CDO case! Or: something nice happened for UBS!

So the story goes like this (from the court opinion): in March 2002, UBS did a synthetic CDO deal called North Street 2002-4 with Landesbank Schleswig-Holstein, a German landesbank, where LSH (now called HSH, due to mergers probably caused in part by this unpleasantness) sold $500mm of protection on a $3bn portfolio "comprised predominantly of assets linked to the United States real estate market (for example, mortgage-backed securities and instruments issued by real estate investment trusts)." The protection attached after $74mm of losses (i.e. UBS bought the most subordinated $74mm of notes) and UBS could manage the stuff in the pool of reference assets:

Under the credit default swap at issue here, [North Street], as protection seller, in exchange for UBS's agreement to pay premiums, agreed to make certain payments to UBS, as protection buyer, upon the occurrence of defined adverse "credit events" affecting securities in the aforementioned reference pool. While the securities in the reference pool were required to meet certain ratings specifications, UBS selected the initial securities for the pool, and also had the right to substitute assets in and out of the pool during the life of the credit default swap, within defined parameters and through the use of internal procedures specified in a reference pool side agreement between UBS and HSH. The governing documents required that by March 2004, 70% of the reference pool would be comprised of asset-backed securities, real estate investment trust assets, and commercial mortgage-backed securities.

The landesbanks were kind of famous for being muppets back before being a muppet was cool, and this was a pretty muppety deal. Which they've now figured out, and sued UBS in New York state court, claiming that UBS was selling them the deal based on the ratings of the underlying securities but actually stuffing North Street with the worst available securities with those ratings:

The amended complaint alleges that UBS harbored the undisclosed intent to engage in "ratings arbitrage" in managing the reference pool. HSH explains "ratings arbitrage" as "[t]he systematic selection and substitution of credits which had the requisite credit rating, but traded at wide spreads (i.e., were higher risk) for that rating. This ratings lag' reflects the market's understanding, evidenced by the lower value of the security, of a deterioration in credit quality in advance of ratings agency downgrades" (emphasis added).

This week, though, HSH's fraud claims were dismissed by a state appellate court. The court took seriously the disclaimers in the offering memo saying things like "UBS is not your adviser" and "consult your own lawyers" and "this is super risky and you'll probably lose all your money so don't invest," and couldn't believe that LSH was so dumb as to rely on UBS salespeople saying "no, trust us, this deal is great, would we steer you wrong?" when all the badness was right there in the prospectus. And it wasn't troubled that UBS was selling LSH this deal based on credit ratings while it was internally modeling it based on reality:

Here, the core subject of the complained-of representations was the reliability of the credit ratings used to define the permissible composition of the reference pool. The reliability of those ratings was the premise on which the entire deal was sold to HSH. Far from being peculiarly within UBS's knowledge, the reliability of the credit ratings could be tested against the public market's valuation of rated securities. According to the allegations of the amended complaint itself, a study of the market for the relevant kinds of securities would have revealed that the credit rating conferred on a security by a rating agency did not necessarily correspond to the security's risk level as perceived by the market. In particular, the amended complaint alleges that UBS's internal analyses of its own risk, profit and loss from the transaction was based on models using "market-implied default probabilities" — in other words, probabilities of default that were derived from publicly available market information. Nowhere in the amended complaint does HSH identify any kind of factual data UBS used in its internal analyses of the NS4 deal that was not readily accessible to finance professionals worldwide. Indeed, the amended complaint does not even allege that the models used in UBS's internal analyses — the economic assumptions and mathematical methods employed — were so unique to UBS that an independent appraisal of the risks of the transaction would not have revealed them.

Anyway the opinion is really really of the school of "you get what your OM says you get; you set up a structure that allowed UBS to stuff you with shit so don't complain that they stuffed you with shit."* There's a throwaway "By no means do we suggest that UBS, if it engaged in the sharp dealing alleged by HSH, is to be commended; such practices are indeed troubling. Still, however much UBS's alleged conduct leaves to be desired as a matter of business ethics," etc. etc. you can fill in the rest.

You can if you like get pretty Greg Smith about this case. The court makes a lot of noise about how this landesbank is a big sophisticated financial entity with top advisers and so forth and they went into this with their eyes open and - maybe? Like, I suppose it's possible that a sophisticated party could have done its own analysis on the US real estate market and decided that all the securities that were trading out of line with their ratings should have been trading in line with their ratings and so determined that buying this thing from UBS was the most efficient way to express that view and make money on that market anomaly correcting itself. I mean, I guess.

But it didn't happen here. There was category arb all over the place. UBS took the riskiest securities in their ratings category and sold them to the least sophisticated bank in the "sophisticated financial institution" category. And I bet that the UBS salesman covering LSH was all "trust me, don't I give you good advice?" and that carried more weight than the doorstop offering memo saying "DON'T TRUST US, WE DO NOT GIVE YOU ADVICE."

One exercise here is to think about whether the SEC would have won its case against Citi or Goldman or whoever if it had brought those cases in front of this court, or conversely whether LSH would have won its case against UBS if it had brought it in front of Judge Rakoff in federal court. There are distinguishing facts in the Citi/Abacus/etc. cases where the SEC got massive settlements; for the most part, in those cases the conflicts of interest were sort of "undisclosed" whereas here they were stamped in really big letters all over the OM. (That is, in the SEC's cases, there was a pool manager who was supposedly looking out for the interests of protection sellers/noteholders but was actually maybe comatose or in the pocket of the protection buyer; in this case, the pool manager was UBS and was pretty clearly looking out for the interests of the protection buyer/itself.)

But the important facts are the same. Because the important facts are, highly rated securities tanked. The sellers/protection buyers (UBS, Citi, Paulson/Goldman) thought that this would happen because they were smart, or lucky, or had models that took a sophisticated view of publicly available data rather than just relying on ratings. The note buyers/protection sellers did not think that this would happen because they were dumb, or unlucky, or relied blindly on ratings. Everyone knew that the sellers expected the securities to decline in value and that the buyers expected them to go up, and everyone had more or less the same public data though some of them had more sophisticated means of analyzing it.

The SEC settlements seem to come from a place of suspicion about those imbalances: even though everyone had to know that (1) someone was short the credit in a synthetic CDO and (2) that someone would probably have some hand in picking the underlying assets (because they'd have to agree to go short them), the SEC sort of pretends to be shocked by those facts on behalf of investors. The state court's ruling in the UBS case comes from a different place, a place where everyone in financial markets is trying to take advantage of their counterparties by exploiting informational or analytic advantages. That place seems to me to be ... well, right.

But you can also see why the big banks mostly haven't been fighting the SEC in court over it's faux-shock approach. Because saying "well of course we take advantage of our counterparties by exploiting informational or analytic advantages, we just disclose that to them and it's what they expect," while probably true, sounds uncomfortably close to "we enjoy ripping the eyes out of muppets."

UBS wins reversal of fraud finding in NY CDS case [Reuters]
HSH Nordbank AG v. UBS AG (pdf)

* It - I mean, I guess this is obvious, but I was sort of surprised to see just exactly how easily UBS was allowed to stuff LSH with shit. Like, they get to control what goes into the pool, change assets out when they feel like it, it doesn't have to hedge any actually existing UBS exposures. There were some limits, including advance notice of changes of reference securities and some veto rights for LSH, which UBS maybe breached - they're also being sued for that, and that wasn't addressed in this opinion. But basically their only incentive is to put the worst securities in their ratings category into the pool.


Appellate Court Willing to Entertain the Possibility that Citi Was Not Committing Fraud

I've had some fun these last few days proposing counterintuitive theories for why Citi might not suck as much as you probably think it does and it's nice to see others joining in the pastime, even if this sounds a little far-fetched: The district court’s logic appears to overlook the possibilities (i) that Citigroup might well not consent to settle on a basis that requires it to admit liability, (ii) that the S.E.C. might fail to win a judgment at trial, and (iii) that Citigroup perhaps did not mislead investors. That piece of rank conjecture is from the Second Circuit's opinion on an appeal* of Judge Rakoff's rejection of the settlement between the SEC and Citi over some mortgage-backed securities. Here's DealBook:

Today In Swiss Banks With Creepy But Defensible Structured Products

I don't really understand it but the TVIX thing is creepy fun. If you haven't followed it, Credit Suisse issued this exchange-traded note called TVIX that was a 2x levered bet on the VIX. They suspended new issuance about a month ago due to position limits, and people were just so damn excited to own the thing that its price crept up to 189% of its fair value, where "fair value" is a reasonably easily measurable thing based on the formula in the TVIX prospectus. Then last week Credit Suisse announced that they would be creating more units, and the price plummeted to and then through fair value, which is what you'd expect to happen. Except that it started plummeting a few hours before that announcement, which is Suspicious. So of course people are sad and so there's a Bloomberg Brief with sort of sad-funny quotes like: “When it started to fall, I bought more because I couldn’t believe how low it was going. I didn’t realize I was playing with a hand grenade.” – Michael Gamble [heh! - ed.], 67, who doubled down on his TVIX investment before the price collapsed. Investors “all think: ‘Oh, I’ll just buy these things, I’ll be hedged against volatility and everything will be wonderful.’ And now they’ve seen the market goes down and their volatility protection goes down too, and they’re going ‘Hmm, what happened here?’ These people are going to have to pay a really expensive lesson.” – Larry McMillan, who manages $30 million as president of McMillan Analysis Corp. So, yes, Larry, they are going to pay a really expensive lesson. But what is it? Stephen Lubben has a little thing in DealBook today where he frets: