We don't have a *lot* of pictures of Judge Rakoff, but we do have this one, which is better than any photograph. He's on the left, pondering life. Sketch: Shepard via Bloomberg News.
We don't have a *lot* of pictures of Judge Rakoff, but we do have this one, which is better than any photograph. He's on the left, pondering life. Sketch: Shepard via Bloomberg News.

Judge Jed Rakoff gave a pretty interesting speech yesterday about financial-crisis-related prosecutions and how way more people should be in prison for the financial crisis, though he's judicial enough not to say who, which makes the whole thing a bit of a strange exercise.

He starts from the premise that there's a lot of unprosecuted fraud, and then spins some theories for why that might be. They're interesting theories! One -- that prosecutors worry that anyone they charged with mortgage fraud could argue "I was just doing what the government wanted" -- practically qualifies as a conspiracy theory. But I'm going to use this speech as a tenuous excuse to give you my unified theory of financial regulation and punishment,1 which is less racy but also perhaps related to some of his questions.

The theory begins: You have some stuff you don't want banks to do. This might include selling bad things to dumb investors, selling really bad things to smart investors, taking dumb risks with taxpayers'/depositors'/whomever's money, tax evasion, all sorts of horrors really, it's a very long list. But you could probably write it down. It's probably a finite list, right?

Maybe. But in general you'll find that, for most things that (1) a bank wants to do and (2) you don't want it to do, there are a million more or less economically equivalent ways for the bank to do it. So your list can get real long real fast.2

There are two broad ways to deal with this. One is to just work really hard: Write a list of all the things you don't want banks to do, making sure to write down each of the million ways a bank could do each of the things, and constantly adding to the list as soon as anyone dreams up a new way. This approach is conventionally called "rules."

The other extreme is to just have, like, one rule, but you make it a good one. "Don't do anything that you wouldn't want to see on the front page of the New York Times," or "act only according to that maxim whereby you can, at the same time, will that it should become a universal law," or "just try not to blow yourself or anyone else up," or "come on, don't be an ass." This is called "principles."

There are obviously intermediate points on the spectrum but hopefully you get the idea.

Rules have some big problems, it turns out. For one thing, writing down all those rules takes a long time and you'll probably get it wrong. You'll forget one, or write down one that you didn't really mean to write down, or there'll be a typo, and it'll be a mess. Maybe an evil lobbyist will convince you to leave a good one off the list, or put a bad one on, knowing that no one will really notice.

Even worse though is that when people -- not all people, I guess, but the relevant people -- look at that long list of rules, they think, well hey, these look like the rules of a game, I want to play, is it my turn? And in fact anyone can play, so off they go, playing the game. The game has lots of rules, but its meta-rules, as it were, are:

  1. You can't do anything forbidden by the rules.
  2. You can do anything that's not specifically forbidden, no matter how trivially it differs from things that are.
  3. You win by doing things that are not forbidden, but that are economically equivalent to things that are.

The game is not that hard to win, because, again, it's hard to make your list of banned activities totally comprehensive, so the things you thought you banned end up happening anyway and there's not that much you can do about it. This problem is exacerbated by various other asymmetries between the rule-writers and the players of the game, including in particular that playing is more lucrative than rule-writing.

Also, everyone has so much fun playing the game. They have the excitement of the competition, and the thrill of victory. They are all, "that was fun, let's play again." And they ask other people to join their team, people whom they select for their skill at games-playing. "Oh, you're good at poker," they say to college students, "how would you like to play regulatory arbitrage?"

So you end up with a culture that treats the rules as rules of a game, and tries to win that game, and often does.

It's hard to really know what a principles-based system would look like but presumably it would be less susceptible to gaming. "Do unto others as you would have them do unto you" is actually a super boring game, and you can never really tell if you've won. You could imagine a utopian principles-based regulatory system where it never occurs to anyone to try to game the system -- because there's no real "system" to speak of -- and instead just works together, banks and regulators hand in hand, to get the "right" result. I mean, whatever, it is possible.

Now, there is one big big big problem with principles-based regulation. The problem is that nobody exactly knows what the principles mean, whatever they are. (Because the only way to know exactly would be to make the sort of list that we just got through saying doesn't work.) In a principles-based system, most of the time people will do the right thing and that'll be fine, and some of the time they'll do the obviously wrong thing and that'll be right out, but occasionally they'll do something that's in a gray area and that'll be squishy.

But that's life, life is squishy, whatever, there's nothing particularly interesting or novel or important about the fact that sometimes things are not obviously right or wrong but somewhere in the middle.3 Nor is it obvious that financial regulation, in the abstract, needs to be different from the rest of life. You could just muddle through and make reasonable judgments after the fact and occasionally say to someone, "look, maybe you thought that you weren't being an ass, but turns out that you kind of were, so knock it off."

There is a stereotype that Europe uses a lot of principles-based regulation for financial stuff, while America is much more rule-focused.4 Another thing that makes America unusual is its deep commitment to putting people in prison for long periods of time. We put many more people in prison, for much longer, and for many more nonviolent offenses, than do most European countries.

There are interesting arguments for why this might be so,5 but a particularly relevant one is found in the work of the late William Stuntz of Harvard Law School, who traced the harshness of American justice to its Civil Rights-era focus on procedural protections for defendants. We have Miranda rights and rules against illegal searches, which make it harder to convict criminals, but we make up for it by being much harsher on those we convict, and by overcriminalizing things (drug possession) that are easier to prove than the things we really want to punish (gang violence).6

That's self-reinforcing, though: The harsher your punishments, the more important it is to have clear specific rules to make sure that we only punish people who are clearly proved to have definitely broken the rules. If we're gonna send someone to jail for 20 years, we want to be really really sure that he deserves it. And since it's pretty hard to actually know if anyone deserves anything, we want the next-best thing, which is being really really sure that he violated a rule that everyone knew about.

So Judge Rakoff thinks it's perfectly obvious that lots of people committed fraud, and that prosecutors are being wusses when they say things like, "proving fraudulent intent on the part of the high level management of the banks and companies involved has proved difficult" or "because the institutions to whom mortgage backed securities were sold were themselves sophisticated investors, it might be difficult to prove reliance."7 And he has some legal arguments for why the prosecutors are wrong -- "willful blindness," "fraud on the market," etc. -- and, I mean, sure, he's the judge, but you can sympathize a little with the prosecutors, no? Rakoff argues for instance that senior bankers should have seen something in suspicious activity reports and shut down their mortgage machines, but the law probably isn't that broad. You need more specific proof of specific intent to violate the rules if you're going to put people in prison. Just "ehh you really missed a lot there" doesn't usually cut it.8

One of my favorite pieces of financial regulation is what happened to Fred Goodwin. Fred Goodwin ran the Royal Bank of Scotland for a long time, presiding over its not-so-hot-in-hindsight purchase of ABN Amro. He was knighted in 2004. In due time, RBS blew up. Goodwin naturally deserved some blame for that. It would be hard to prove that he was a giant crook, though. There were not, like, pages of e-mails from him saying "ooh let's commit crimes and blow up RBS." He just did some dumb stuff and blew up RBS is all.

So they took away his knighthood. Isn't that delightful? He probably really wanted to stay a knight, I don't know, I would have anyway. Taking away the knighthood hurt. But not as much as 20 years in jail would have. And it sent the right signal, which was to the effect of, look, regardless of the technicalities here, you look like kind of an ass, so give us back your armor or whatever. There's no hypertechnical fight over who knew what when, or what some awkward e-mail really meant. The case didn't bog down in process and technicalities, because there was no process. Goodwin did bad, and was punished, simple as that, without a lot of worrying about how exactly to describe the bad that he did and fit it within the rules. The message is not "aha, here's a terrible e-mail, we got you!" It's "don't blow up a bank, jerks."

Really everyone who messes up at a bank should be stripped of a knighthood and we should knight everyone right now so as to be prepared.

This of course is utterly unsatisfying if you want to put bank executives in jail, and the conspiracy-theory/prosecutorial-wuss/whatever model of why they're not in jail has an obvious appeal. But if you're thinking about how best to regulate the financial system going forward, "tons more jail" is not an obviously right answer. Judge Rakoff cites Enron as a prosecutorial success story, and financial-scam rules and penalties were significantly strengthened in the aftermath of Enron. And then, y'know, 2007-2008 happened. Increasing the costs of losing the game is one obvious way to respond to financial gamesmanship, but it's not the only way. Getting rid of the game altogether might be another possibility.

1 I mean it is not particularly original to me. The rules-vs.-principles distinction is long-standing in thinking about financial regulation, and accounting, and law generally.

2 Here is a baffling article about "hedging" that might convey a flavor of that. Basically, regulators don't want banks to do "proprietary trading," which is risky, but do want them to be able to "hedge," which reduces risk, or whatever. And so now there is a fight to the death over what the definition of "hedging" will be, and exactly what activities will and won't be covered. And amusingly some people on both sides want to avoid defining it at all.

3 Here is a terrible citation for that proposition, enjoy!

4 I mean, don't believe that too literally, there are plenty of rules in Europe. Basel is in Europe, for one thing.

5 I am partial to the views of James Q. Whitman, my former law professor, who traces the difference to the age of revolutions in the late 18th and early 19th centuries. The American and French revolutions both were about equality, but the American one leveled down -- "no more aristocrats!" -- while the French one leveled up -- "everyone's an aristocrat!" So in pre-revolutionary Europe, aristocrats who did bad stuff were beheaded, while commoners were hanged. After the revolutions, America got rid of beheading, but France got rid of hanging.

6 Or, for an example closer to home, we frequently use securities disclosure rules to punish theft, because it's easier to prove "you took this money and didn't mention it in your 10Q" than "you took this money for illicit personal purposes and weren't planning to give it back."

7 Those quotes are from Rakoff characterizing prosecutors' positions -- not actually from prosecutors.

8 Or take his weird history here: "Thus, in the 1970's, in the aftermath of the 'junk bond' bubble that, in many ways, was a precursor of the more recent bubble in mortgage-backed securities, the progenitors of the fraud were all successfully prosecuted, right up to Michael Milken." But Milken got done for stupid technical violations and tax evasion, not inflating a junk bond bubble or whatever. Nobody said "junk bonds are bad for America, let's punish Milken for them." Or, I mean: They said precisely that. It's just that they punished him for hypertechnical stuff because the bad-for-America thing is not actually a crime. (Also 1970's whatever.)