UCLA professor offers wild solution to bad bank behavior: Scare them straight
Nov 16th 2009 12:30PM
Updated Dec 4th 2009 4:45PM
A UCLA professor has come up with a wild idea to help prevent the next financial meltdown. He doesn't propose expensive regulations and mechanisms of enforcement on financial actors. Instead, he suggests that swift and decisive punishment of the biggest instigators of financial mayhem could scare the industry straight.
Before getting into the details of his proposal, let me introduce the man and his new book, from which this idea springs. Professor Mark A.R. Kleiman heads UCLA's Drug Policy Analysis Program. Years ago, we shared an apartment while I was working at a summer job in Cambridge, Mass., and he was teaching at Harvard's Kennedy School of Government. Kleiman also runs a great blog -- The Reality Based Community. I contacted Kleiman a few weeks ago after seeing his new book, When Brute Force Fails, from Princeton University Press, in the window of the Harvard Book Store. Here are his email responses to my questions.
DailyFinance: Why did you write your book?
Because I'm both horrified and excited. Horrified that we have both an intolerable level of crime and an appalling rate of incarceration, and excited that we now know how to have less of both. My hope is to help the public and political conversation about crime and justice catch up with the progress being made by creative people actually doing work in the field.
What is the "news" in your book?
That we could have half as much crime, and half as many people behind bars, five years from now as we have today, without spending any more money or doing anything we don't already know how to do.
What is the central principle about how to deter bad behavior (and possibly to motivate good behavior)?
The central principles of deterrence are familiar to anyone who has ever successfully raised a child or trained a puppy: Make the rules clear and enforce them consistently and quickly. The current criminal justice system fails to do that, and tries to substitute severity for swiftness and certainty. That's the "brute force" that fails, and that got us into our current trap.
It's not just that swiftness and certainty work better than severity; severity is actually the enemy of swiftness and certainty, because a severe punishment is too expensive to hand out often (one twenty-year sentence costs as much as twenty one-year sentences) and because the amount of "due process" required to impose a severe penalty slows it down. The ideal is to impose the minimum effective dose of punishment; if you can make it quick and consistent, that dose turns out to be pretty small, measured in days, not years.
Even using minimum effective doses, in general you can't punish everybody for every offense: "zero tolerance" as a slogan reflects zero intelligence on the part of the person using it and has zero chance of working. So you need to concentrate enforcement: Targeted zero tolerance can work, as long as you make it very clear just what it is you're not going to tolerate, and deliver that message directly to the people whose behavior you're trying to change.
As the Lord said to Moses, don't just strike the rock; talk to the rock. Swiftness and certainty are what you're trying to achieve; concentrated enforcement and direct communication are how you can achieve it at finite cost.
That's when the second key idea kicks in: positive feedback, which leads to "tipping" phenomena. If everyone's breaking the rules, breaking the rules is pretty safe, because the enforcement machinery faces too many targets. If most people are behaving, it's a lot more risky for those who misbehave.
So by concentrating enforcement until you get one set of offenses or offenders, or one geographic area, under control, you can then free up resources that can be used to attack the next problem.
How could this principle be applied to financial actors?
Say you're a bank regulator trying to control the behavior of mortgage originators and mortgage brokers. If they're all running around writing liar loans, it's hard for you to do enough cases to make that behavior expensive for all of them at once.
But if [you] take a sample from each broker or originator and figure out what fraction of the loans from each one are bogus, you can then go to the worst tier and say to each firm, "If your error rate isn't down to the 75th percentile in two months, we're coming after you." Try to force a race to the bottom. Then make an example of the worst-behaving firm, and tell everyone else that they really, really don't want to be next.
The same applies if you're a bank working with a bunch of brokers, or a packager of mortgage-backed paper working with a bunch of banks. You can't audit every loan, but you can take a sample and tell all of your counter-parties that the outfit with the highest rate of garbage is going to get black-balled.