Saturday, August 30, 2014
Adam Levitin at Credit Slips has an interesting breakdown of MBS litigation settlements. He points out that of the $94.6 billion in settlement funds, only 2% has gone to private investors alleging securities-fraud-type claims.
First, it shows that legislative reforms and court rulings have seriously impeded the effectiveness of securities class action litigation. If ever there were an area ripe for private securities litigation, private-label RMBS is it, yet almost all of the recoveries are from six settlements. This should be no surprise, but it's rare to see numbers put on the effect. This is what securities issuers and underwriters have long wanted, and the opposition has mainly been the plaintiffs' bar, but perhaps investors will take note of the effect too.
Second, the distribution shows how badly non-GSE investors got shafted. Remember, that private-label securitization was over 60% of the market in 2006. Yet investors have recovered only 38% of that which the GSEs/FHFA have recovered, and most of that is from the trustee settlements or proposed settlements (I'm not sure that any have actually closed). Private securities litigation has recovered a mere 4% of what the GSEs/FHFA have recovered.
The real question is whether investors have learned that they cannot rely on either trustees or the securities laws to protect them from fraud, and if they have, what they plan to do about it. One sensible thing would be simply to invest in other asset classes. The other would be to try and reform the trustee system and/or the securities laws.
I'm sure there are many reasons for the disparity, but I think one major contributor is a series of rulings narrowing the definition of standing in the class action context.
(okay, that was my attempt to jazz up a procedural post)
Anyway, these standing issues are now pending - sort of - before the Supreme Court, as I previously posted. What's interesting is that these standing rulings have had a dramatic effect on private investors' ability to bring claims, but they aren't usually mentioned in the same breath as other, more obvious, limitations on securities class actions.
[More under the jump]
Public Employees’ Retirement System of Mississippi v. IndyMac MBS, Inc., No. 13-640, is about the statute of repose, but more fundamentally, the issue is class action standing.
As I previously posted, courts have reached a number of different, and conflicting, rulings regarding the requirements for serving as a named plaintiffs/class representative in mortgage crisis litigation. The more that courts have tightened the screws, the harder it has been for plaintiffs to bring class cases.
As we know, during the height of the housing bubble, banks pumped out billions of dollars with of mortgage-backed securities. The numbers were staggering – for example, in the IndyMac case pending before the Supreme Court, the original complaint alleged claims in connection with 106 different offerings that took place over a 2 year period. The sizes of the offerings varied, but they tended to run about $300 to $800 million each, with a few outliers going over $1 billion or even $2 billion – and that was just IndyMac RMBS. Another case involving JP Morgan alleged just one year of RMBS, with a total face value of over $36 billion spread across 33 offerings - and there were many more cases exactly like it.
Structurally, the way it worked was that the banks would set up a trust for each RMBS offering, and the trust would issue securities that would be sold to the public. The securities were backed by mortgages owned by the trust, and would be divided into different tranches representing different payment priorities (and thus different risk levels). Each trust, of course, held a different set of mortgages.
These RMBS were registered with the SEC, and while many were purchased by investors directly in the initial offering, in some cases, purchasers bought in the secondary market.
Now, there are a few things to note. First, RMBS and other kinds of mortgage-backed instruments are not like stock – they don’t trade in an open and well developed market. That means that there’s no fraud-on-the-market presumption of reliance, which generally means that there’s no chance of bringing a class action for any kind of claim where reliance is an element – there are too many individualized issues. (Some investors have managed to do it – see Dandong v. Pinnacle Performance Ltd., 2013 U.S. Dist. LEXIS 150259 (S.D.N.Y. Oct. 17, 2013) – but others haven’t, see Abu Dhabi Commer. Bank v. Morgan Stanley & Co., 269 F.R.D. 252, 253 (S.D.N.Y. 2010)).
That means that the only claims available to investors in the class context are going to be claims that don’t require reliance. And that, for the most part, means Section 11 of the Securities Act. There are other kinds of claims, but most won’t allow recovery for secondary market purchases, and others involve state law and thus require that all members of the class have enough of a connection to the state to justify application of that state’s law.
Section 11 of the Securities Act allows recovery for false statements made in a registration statement, so it’s only available for registered securities (CDOs are out; RMBS are in). It allows for claims against the issuer of the security, and all of the underwriters.
And when the crisis first hit, lots of Section 11 cases were filed – just like the IndyMac case – involving 20, or 50, or 100 MBS put together by a single issuer. The allegations were generally that the registration statements falsely represented the quality of the underlying loans – they claimed that loans adhered to certain underwriting criteria when they did not, that the properties had a certain appraisal value when they did not, and so forth. Each one of these cases was, predictably, absolutely massive – billions of dollars worth of securities issued, and it didn’t matter if the issuer itself was bankrupt, because the underwriters were still solvent. And best of all, from the plaintiffs’ perspective, Section 11 has no scienter requirement – so once the false statements were proved, liability was almost certain to follow (not that defendants didn't have certain affirmative defenses, but they'd have been tough slogging). The news was filled with disclosures about all of the wrongdoing that led to the crisis - making pleading an easy task - and Section 11 does not even contain the same kind of heightened, detailed pleading requirements for which securities cases are famous. Even causation does not have to be proved - it's defendants' burden to disprove. From the perspective of a plaintiffs’ attorney, at least at the pleading stage, these cases were like shooting fish in a barrel.
But obviously, you can’t bring a claim with a new named plaintiff for every new MBS – it would be impossible, you’d have hundreds of plaintiffs, and for the smaller issues, it might be hard to find someone willing to do it. So instead, the cases would have one or two plaintiffs who purchased a sample of the MBS in a particular case, and they’d purport to bring claims on behalf of all other purchasers.
Beginning in roughly 2009, however, courts started dramatically chopping down the size of these cases, holding that named plaintiffs could only represent purchasers in the exact same issue. So if a plaintiff bought a security from the 2006-WF1 trust, the plaintiff could only serve as a class representative for other purchasers from that same trust. For purchasers of securities from the 2006-WF2 trust, there would need to be a new named plaintiffs, or all claims in connection with that issue would be dismissed. In some cases, courts went even further, and held that named plaintiffs could only represent purchasers in the same tranche.
The theories varied. Sometimes, courts claimed this was a requirement of Section 11 itself- which can't be right, because Section 11 was passed in 1933, long before the federal rules were amended to allow for modern class action procedures. It is unlikely that the 1933 Congress expressed any view on procedural requirements for class actions.
Other times, courts saw this as a constitutional requirement – if you didn’t purchase in the “same” issue, you weren’t injured by that particular issue, and therefore you didn’t have the requisite Article III injury to bring claims based on that issue.
But that was odd, too – in the class action context, named plaintiffs always represent people with “different” injuries than the ones they themselves suffered – that’s what it means to sue in a representative capacity. As the Supreme Court put it, “The class-action device was designed as an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.” Gen. Tel. Co. of the Southwest v. Falcon, 457 U.S. 147 (1982). We allow it because sometimes, the injuries suffered by a group of plaintiffs are so similar that it makes sense to allow one plaintiff to represent the interests of the others. Judge Swain explained it when rejecting a tranche-level standing argument:
Every time a lead plaintiff prosecutes an action on behalf of a class, she brings claims based on injuries she did not personally suffer — in other words, claims she could not have advanced individually. For example, the cancer-stricken lead plaintiff in an asbestos case brings claims based on other peoples' cancers; a lead plaintiff, paralyzed from the waist down due to a car brake malfunction, can bring product liability claims on behalf of other people who were paralyzed from the neck down due to the same faulty brake design. To say, as Defendants do, that named Plaintiffs cannot bring claims on behalf of purchasers of other securities because they cannot trace an injury to securities they did not hold is analogous to asserting that a driver who suffered injury when her brakes malfunctioned cannot sue on behalf of purchasers of cars with the same defective brake design because a plaintiff cannot claim to have been harmed by a car she never drove.
In re Bear Stearns Mortgage Pass-Through Certificates Litigation, 851 F. Supp. 2d 746 (S.D.N.Y. 2012).
Finally, some courts held that plaintiffs who purchased in one issue would not have sufficient personal motivation to represent purchasers in other issues, because each issue was backed by a different set of mortgages. The logic here is a bit more nuanced. The substantive claim in each case was that the registration statements misdescribed the quality of the underlying mortgages. But each issue had a different registration statement. And each issue was backed by a different set of mortgages. So the theory was, a plaintiff who purchased from the 2006-WF1 issue might be personally motivated to show that the mortgages backing the 2006-WF1 issue were misdescribed in the registration statement, but has no personal motivation to examine the mortgages backing the 2006-WF2 issue, or the 2007-CTY5 issue. And because proof that the 2006-WF1 mortgages were defective doesn’t tell you anything about whether the 2007-CTY5 mortgages were defective, the 2006-WF1 purchaser would not be able to represent 2007-CTY5 purchasers.
That’s a more complex argument, to which I will return in a moment.
But this is what courts were holding – on the pleadings, prior to any discovery, in the context of a motion to dismiss under Rule 12(b)(6).
Now, to be fair, there has been some confusion on the issue from the Supreme Court, which openly admitted in Gratz v. Bollinger, 539 U.S. 244 (2003), that there had been “tension in our prior cases” regarding the concept of class action standing. But there had been nothing like what the district courts were holding now – in fact, these rulings proliferated so quickly as a result of the MBS litigation that Newberg on Class Actions actually rewrote its chapter on standing in the class action context to account for the changes.
(This narrow concept of class action standing also infected other areas of law - for example, courts examining false advertising claims began to follow reasoning similar to the securities cases, holding that named plaintiffs could only represent classes of plaintiffs who purchased the exact same product as the named plaintiff - even if similar false statements were made across multiple similar products. Rebecca Tushnet has blogged about this issue on several occasaions.)
And when cases were chopped down to maybe less than 10% of their original size, new plaintiffs would emerge – plaintiffs who had purchased in the dismissed issues. They’d try to intervene to revive the dismissed claims, and that’s when courts would hold that they were time-barred, because the original action had not “tolled” the statute of limitations or the statute of repose – which is the issue pending before the Supreme Court in IndyMac.
Now, eventually, the Second Circuit stepped in. The Second Circuit agreed that it’s common for class action plaintiffs to represent people with “different,” but similar, injuries. But it was persuaded by the problem of different mortgages backing the different trusts. And it came up with a new rule in NECA-IBEW Health & Welfare Fund v. Goldman Sachs & Co., 693 F.3d 145 (2d Cir. 2012): If any two issues are backed by mortgages issued by the same originator, then that’s sufficient similarity for a purchaser in one issue to represent purchasers from the other issue. In the Second Circuit's words, if the two issues share at least one mortgage originator, then claims based on the two issues "implicate the same set of concerns." (quoting Gratz). The Second Circuit's rule, as far as I can tell, does not require that there be any minimum amount of overlap – so if, for example, 2006-WF1 and 2007-CTY5 are each 1% backed by Countrywide mortgages (no matter when they were originated or the types of mortgages involved), a purchaser of 2006-WF1 could represent purchasers in both issues.
But this ruling came too late for a lot of cases, which had already settled (with dramatically trimmed claims). And outside the Second Circuit, NECA wasn’t binding – so Judge Pfaelzer of the Central District of California, who controls Countrywide MBS litigation, has continued to require tranche-level standing. See Fed. Deposit Ins. Coroporation v. Countrywide Fin. Corp., 2012 U.S. Dist. LEXIS 167696 (C.D. Cal. Nov. 21, 2012).
Now, as is probably clear from my post, I think all of these standing holdings are wrong, and the NECA attempt to split the baby makes no sense. Because if NECA is trying to find some kind of practical rule – the point at which a named plaintiff has a real interest in proving falsity across a number of different issues – the line it drew is random.
See, the registration statements described the mortgages in general. They didn’t say “We have absolutely no defective mortgages here, everything’s perfect!” They said “our originators adhere to certain criteria.” And those statements are not going to be rendered false just because of one bad loan here or there – they’re only going to be rendered false if a very large number of the loans in a given issue are bad. And if there’s only a tiny overlap of originators between two issues, that won’t help plaintiffs prove that both sets of issues were populated by defective mortgages in general.
Or, to put it another way, overlap in originators is not, by itself, sufficient to show that the "same set of concerns" is implicated across issues.
The problem, of course, is that if you want to really do a practical analysis – are the issues sufficiently similar that a plaintiff who purchased in one issue can use one set of proofs to establish falsity across multiple issues? – well, we have a procedure for that. It’s called Rule 23, the class certification hearing, and it comes with discovery, and evidence. You simply can’t determine, on the pleadings alone, whether two issues are sufficiently similar that a purchaser from one issue has an interest in bringing claims based on both.
To make things more concrete: imagine that in the NECA case, plaintiffs conducted discovery and they found out that one single person was in charge of constructing Goldman’s RMBS, and that person was actually possessed by the devil and decided to intentionally populate Goldman’s RMBS with defective mortgages. If plaintiffs uncovered that evidence, they’d be able to prove all claims across every issue with a single deposition – done and done, in and out in 7 hours.
But let’s say, instead, that Goldman had a completely different team working on every single new RMBS issue. And those teams had completely different systems, completely different methods of quality control, got completely different sets of reports, and actually operated in offices in separate countries without ever speaking to each other.
Well, in that case, there’s a very real chance that proof of problems in one issue wouldn’t do very much to advance claims based on a different issue.
The truth, of course, is somewhere between these poles. But the important point is this: You can’t make this determination on the pleadings. It requires discovery. It requires a theory of how the plaintiffs’ claims are going to be proved, how the case is going to be tried. It requires, in other words, a Rule 23 hearing.
Ultimately, by definition, a class action involves one injured person - the named plaintiff - suing for redress of the injuries that happened to other people. This is justified on the ground that the named plaintiff's injuries are close enough to those of the rest of the class to allow the named plaintiff to sue as a representative for the others. What's "close enough"? Well, you can either look for some Platonic concept of "closeness" based on first principles - like, whether all the securities were bought from a single trust, or tranche, or shared an originator - or you can look to practical realities: what will the trial look like? And if it's the latter, it'll be a very rare case where decision on the pleadings is warranted.
Now, with that said, I have my suspicions about why courts adopted this narrow view of standing, particularly in the context of MBS. As I said, these cases were massive. And they all had the same set of defendants – because no matter who the issuer was, the same set of large Wall Street banks served as underwriters, and could be held liable on that basis. So Goldman, Citi, JP Morgan, Barclays, RBS, Morgan Stanley, Merrill Lynch, Deutsche Bank, UBS, Credit Suisse – all of them defendants, over and over, with billions of dollars in what were essentially strict liability claims. If these cases were litigated to judgment and the defendants held liable, the damages could have remade Wall Street – again.
And my legal realist view of the matter is, courts were very, very uncomfortable with the implications. Not just because of the size of these cases – so large that even some plaintiffs’ attorneys feared they couldn’t litigate them, because they wouldn’t even know how to go about administering damages funds of that size – but because these cases started to look like regulation, not compensation for injury. If Wall Street was going to be remade as a result of the financial crisis, that was Congress’s job – not the job of a handful of plaintiffs’ law firms bringing private claims.
Anyway, IndyMac’s pending, and the Supreme Court could go in any direction, but by now, so many cases have been settled and claims dropped that even the most expansive plaintiff-friendly ruling would not bring back a lot of the cases that were dismissed. Particularly persistent purchasers have turned to state law, bringing individualized fraud claims, which have longer statutes of limitations but place more onerous burdens on plaintiffs. Those cases have had mixed success, but no matter how much each individual investor recovers, these actions will never replicate the kinds of damages that class actions can bring.
So I don't know that this explains all of the disparity that Adam Levitin identifies, but it's sure responsible for some of it.