Banks May Have Mispriced Things That Had No Price

Can you sell bonds for more than you think they're worth? Oh, sure, why not, but there'll probably be a federal investigation anyway.
The bonds were for sale. And they referenced houses. Which were probably for sale too, out of foreclosure. So this picture. Work with me a bit. Photographer: Jacob Kepler/Bloomberg

There's not a ton of detail in this Wall Street Journal article about federal regulators' "wide-ranging examination of mortgage-bond sales by banks in the years that followed" the financial crisis but ooh boy is it interesting. The examination seems largely metaphysical, looking not so much into the banks' e-mails as their souls. Here:

Regulators are investigating whether traders exploited the murky pricing around residential mortgage-backed securities from around 2009 through 2011 to buy or sell the investments at artificially depressed or inflated values, the people close to the inquiry said. The other parties in such transactions would typically be rival banks, hedge funds and other large investment firms.

Unlike with publicly traded stocks, where pricing is transparent, investors in mortgage-backed securities markets often rely on traders to disclose honestly the prices paid and the commissions charged. It is generally illegal for traders to misrepresent to investors aspects of a transaction that are important enough to affect the decision to buy or sell.

So what aspects of a mortgage-backed security "are important enough to affect the decision to buy or sell"? A conventional answer would be something like "the creditworthiness of the underlying borrowers, the values of the homes used as collateral, etc." And there've been lots of lawsuits about banks that misrepresented those sorts of things in the pre-crisis period and have gotten in bad trouble. 1

But in 2009 that was sort of a dumb answer. Who's to say what a thinly traded busted subprime security is "worth"? What mattered was the price at which banks were willing to buy or sell those securities. If you knew in 2009 that a bank would buy a security at 50 and sell it at 52, then that tells you more about what it was "worth" -- in 2009 -- than your deep credit analysis of the borrowers and the house prices and so forth.

Obviously, the bank controls that information. Creates that information, really. If you call the bank and say "what price would you pay for LXS 2007-15N 2A1," 2 they will just tell you a number, and that's the number. But that leads to some pretty obvious scams. Like here is the scam:

  • You have some bonds that you think are worth, I dunno, 60.
  • Someone calls you and says "hey I want some of those bonds, where will you offer them?"
  • You say "umm, 70?"
  • They say "okay, if that's the price, that's the price, I'll take a million."

How do you feel about that scam? Reasonable minds can disagree. You can be like, "that's not a scam, that sounds great!," in which case you have a bright future in clichés about used-car salesmen. You can be like, "hmm that seems okay but you'd want to be careful because they might check your offer in competition, and anyway a lot of repeat clients might eventually figure out that you're charging a lot more than you'd pay to buy them back, and if clients feel burned they may stop doing business with you," in which case, fine, you passed the test, you can be a trader.

Or you can be like, "that is obviously illegal guillotines guillotines guillotines!", in which case you're wrong. There's no law that you have to sell things at the price you think they're worth. Really, if you think they're worth that price, you should keep them. Differences of opinions about value are what make a market. Also not telling anyone else your opinion about value, that helps to make a market too.

But our little scam, or non-scam, can shade into illegality pretty quickly. For instance, instead of saying "umm, 70," you might launch into an elaborate story about how you don't actually have any of the bonds, but you know a guy who has them, and you know he wants 72 for them, but you think you can bargain him down to 69.75, and then tack on a modest profit for yourself, so you could probably see your way to selling the bonds at 70, how does that sound?

That sounds good! This story of a third party confirming your pricing makes your customer more likely to accept your pricing. But if it's entirely fictitious you might get into trouble. That's what happened to a Jefferies trader named Jesse Litvak, who was apparently a little too creative in telling stories about how he was negotiating for bonds he already owned. 3 Litvak was arrested a year ago -- he's apparently ground zero for this federal investigation -- and charged with fraud for making material misrepresentations about the bonds. The question of whether his song and dance about the lengths he went to to buy the bonds were "material" seems like an interesting one. He never lied to anyone about the credit quality of the underlying mortgages. He just created a fantasy world where his job was a lot more interesting than it was in the real world. And where his bonds were more expensive.

There are other ways to get to illegality, or semi-illegality. You can send a client fake marks after the trade to keep them from figuring out that you overcharged them -- sell a thing worth 60 for 70, then keep telling the client that it's worth 70 for, say, collateral purposes. 4 There's a popular theory that just making more than 5 percent on a trade is illegal, even without any supporting lies, though this theory is not particularly true. 5

None of these scams is all that blood-curdling: Your victims can avoid being victimized by just checking the price with multiple dealers, or by not trading complex illiquid securities in the first place. But they provide wonderful opportunities to look really scammy, because who can resist the chance to send their colleagues a high-fiving e-mail saying "hahaha we really ripped that client's face off, good work boys"? No one, is apparently the answer. I for one am excited to see the results of this federal investigation. The results will be e-mails, and they will be hilarious and ungrammatical.

They probably won't answer the metaphysical question, though, which in the Journal's words is "where does aggressive salesmanship cross the line into fraud?" I mean, the threshold question is hard enough: If you're buying at 60 and selling at 80 in a volatile and illiquid market, is that because you're being aggressive/scammy, or because you need compensation for taking serious risks?

But even the purely scammy behavior -- buying or selling "at artificially depressed or inflated values" -- isn't necessarily illegal. You don't have to be honest about where you think value is. It's not a crime to charge a price that you think is unfair (probably!). It's probably a crime to lie about the ancillary stuff around value -- your marks, your markup, etc. -- as long as that ancillary stuff would be material to your customer.

Whether the ancillary stuff actually was material is a hard question, though I guess if it wasn't, why lie about it? "Because trading is necessarily a business of misdirection and keeping information from your counterparties," is I think one good answer. Maybe the best answer! In that it is probably true, unlikely to satisfy regulators or the public, and likely to come along with a lot of really embarrassing e-mails. So we have that to look forward to.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

  1. This is distinct from whether or not those factors were actually material to investors' decisions. That is a difficult empirical question but there is not a ton of evidence that investors were, y'know, actually asking about them. Mostly it was like, "house prices will always go up, back up the truck," etc.

  2. Really the name of a real mortgage-backed security.

  3. Also though he'd sometimes really negotiate to buy bonds and then just lie to his customers about the price he negotiated, a related though somewhat less silly sin.

  4. There's a lot going on in the Goldman Timberwolf case but that's a thing that maybe happened there.

  5. Here's Chris Arnade saying "We tried to abide by an unwritten 'five-point rule': never intentionally make more than five percentage points of profit from a customer." That comes, loosely, from a written five-point rule, the Finra "excessive markup" policy that grudgingly mentions 5 percent as a rule of thumb. (Here is a law firm memo.) Sometimes brokers get in trouble for charging more than 5 points of markup.

To contact the author on this story:
Matthew S Levine at

Before it's here, it's on the Bloomberg Terminal.