Let's say there are two models of how to run a financial system: make it transparent and give participants the tools they need to evaluate counterparties, or leave it opaque and trust regulators to keep things safe. There are good arguments for both sides - perfect transparency would be a competitive nightmare, and wouldn't help those too stupid to use it; perfect opacity concentrates a whole lot of risk in the hands of regulators with doubtful incentives and skills - and so we have sort of a mix, for good or ill.
This might make you ill:
When Citigroup agreed last month to pay $285 million to settle civil charges that it had defrauded customers during the housing bubble, the Securities and Exchange Commission wrested a typical pledge from the company: Citigroup would never violate one of the main antifraud provisions of the nation’s securities laws.
To an outsider, the vow may seem unusual. Citigroup, after all, was merely promising not to do something that the law already forbids. But that is the way the commission usually does business. It also was not the first time the firm was making that promise.
Citigroup’s main brokerage subsidiary, its predecessors or its parent company agreed not to violate the very same antifraud statute in July 2010. And in May 2006. Also as far as back as March 2005 and April 2000.
The incredulity continues throughout, and the article includes a handy infographic where you can see how many times each bank pulled the "I'll be good, I'll be good, I'll be good!" move in the past.
The point, of course, is that nothing much happens to the repeat violators. No one is ever held in contempt for not doing what they said they would do. Instead, it maybe sort of kind of factors into their punishment the next time in a soft and squishy way:
Robert Khuzami, the S.E.C.’s enforcement director, said never-do-it again promises were a deterrent especially when there were repeated problems. In their private discussions, commissioners weigh a firm’s history with the S.E.C. before they settle on the amount of fines and penalties. “It’s a thumb on the scale,” Mr. Khuzami said. “No one here is disregarding the fact that there were prior violations or prior misconduct,” he said.
How much are they not disregarding it? Well, they're hitting Citi with a particularly hefty $95 million penalty for this violation, because they keep breaking the law. Take that! Only that's just 1/5th of what they fined Goldman for the same conduct. And there's no real reason for the difference. Sure, Goldman is also a repeat offender (as is every big bank) - but it's got only three violations, versus Citi's eight, per the Times's chart.
The Times - and Jonathan Weil in Bloomberg last week - poke lots and lots of deserved fun at the SEC and Citi for all of this. Weil describes some of the most notable mishaps in Citi's permanent record, including the research settlement and various garden-variety frauds in auction rate securities and mutual fund costs. Those things, and its current CDO unpleasantness, all look kind of bad in hindsight and also are kind of similar basic misbehavior, which is "not telling investors everything that they really ought to have known" and usually more specifically "telling them how great things were while you were at the same time sending unfortunate emails about how terrible they were." So, yeah, it's kind of fair to say that after the fifth or sixth time they got caught selling crappy products with crappy disclosure, they should have cleaned up their act.
But it's possible to be too cynical here. The guys peddling the mortgages with dubious disclosure probably never met the guys who earlier peddled auction rate securities with dubious disclosure. Each disclosure was dubious in its own somewhat different way. And each disclosure was, in its own way, arguably plausible, plausible enough that Citi's lawyers probably signed off on it - and that other banks were doing more or less similar things much of the time. In some sense, the research settlement, say, really didn't put Citi on notice - at all - that what it was up to in CDO structuring was less than kosher.
When the SEC says "okay, now don't violate securities laws any more, or we'll getcha," what they really mean is:
(1) if you do violate securities laws again in, like, three months, we are going to shake our heads disapprovingly and maybe write something about it in a memo that might filter into your totally arbitrary fine in some loose way; and
(2) if you do the exact same thing again - like, if your research analysts keep pumping stocks to win banking business while sending around emails about how bad those stocks are - then we will fuck you up.
And that second part just has to be true. It's never really been tested - no one's been caught doing the same fraud again - but that's because banks put into place really serious compliance measures to prevent the particular problem that they got caught for, because they really are on notice about it, and they take the threat seriously. My guess is that if someone did get caught running the same fraud or quasi-fraud that was already specifically sanctioned, individuals would go to jail.
So these "I'll be good" orders are in fact a little like probation: you can get away with it once, but after that you're on notice that it's big trouble. And they do probably have their intended effect, which is to stamp out the particular misbehavior that the SEC is upset about. The Times and Bloomberg are just disappointed because they make the mistake of thinking that these orders mean what they say: that any future violations of the anti-fraud rules will be a Big Deal.
You can quite reasonably think that isn't good enough: that violations of securities laws are violations of securities laws, and that after two or three or eight strikes there should be consequences - consequences beyond the fine for the particular thing that you did. I am skeptical of that, because I think that a lot of these violations are kind of gray areas where people thought in good faith that they were following the law, but I can understand why lots of people disagree.
But the bigger problem is that the SEC's lack of transparency and, y'know, honest disclosure makes their whole project more difficult. They can't make any sort of reasoned case that they are in fact deterring specific future misconduct, because their settlements don't mean what they say, so the settlements look like they're ignored and don't deter misconduct. You're left with the SEC mumbling nice words about how they take everything into account, and telling a federal judge "we'll fine Citi what we want to fine them, don't you worry about how we got that number."
Maybe the SEC would look absurd signing a settlement with a bank that says "we won't commit securities fraud by misrepresenting how much influence the protection buyer on a synthetic CDS portfolio had in selecting that portfolio, but all bets are off on any other securities fraud!" And to avoid that, I guess they have to sign settlements where the banks say "we will never ever do any sort of bad thing ever again."
But then when nobody at the banks or the SEC takes their settlements at all seriously, the SEC ends up looking absurd anyway. And in a world where the SEC views its role as keeping the financial markets safe without being transparent to the courts or the public, that seems like a dangerous result
Citigroup Finds Obeying the Law Is Too Darn Hard [Bloomberg]