Jamie Dimon's testimony this week about how JP Morgan Chase ended up losing over $2B due to bad decisions compounded by bad decisions got me thinking about risk, risk management, and incentives. And how, at the end of the day, it's the incentives that matter.

Now, Dimon has asserted that they will "claw back" some of the pay from the traders that caused the $2B loss. Really? How are they going to do that? Dock their bonus by $1M/year for the next 2,000 years? Make them send their kids to public school and confiscate the St. Mark's tuition?

They can't, in any meaningful way, get the money back from the traders. And they know it. And the traders know it. Which is a big part of the problem. The incentives for making risky trades are just too great -- if you win, you get literally millions in bonus. If you lose, worst, worst case, you lose your job. So why not take a spin? Or keep doubling down?

A long, long time ago, I worked on the risk management desk. Our job was to analyze the overall risk position of the bank every night, and report to management. Since I worked at a relatively conservative Japanese commercial bank, it wasn't uncommon for overly aggressive trades to be unwound or the risk somehow shifted off to another, more risk tolerant party. (Of course, the Japanese banks had their own issues with underwriting commercial properties, so no one is perfect in this department.)

However, as more and more of the earnings from the banking system come out of the trading floor, the power of the risk manager declines. It's simple -- risk management is just the annoying group of pointy-haired people that get in the way of the next great innovation. So, smarter and smarter people get into trading, where the money is made, and people who would someday want to be traders get into risk management.

So, how do we change this system? In my opinion, it's not about banks being too big. Or derivatives too complex. Trying to litigate these types of things is a pointless game -- smart bankers will figure out the loopholes a lot faster than politicians can close them. No, the answer instead needs to come from the bank's owners (shareholders) and managers.

It's critical that the long-term view come into play when managing a bank. And unfortunately, given the short-term nature of the stock market, just paying people in stock doesn't immediately fix this problem. Instead, we need to figure out a system where people are paid over the risk horizon of the investments they make.

So, for example, if a trader enters into a 7-year USD/JPY currency swap, and that trade is held on the books for seven years, the payout should be over seven years. If they are making highly leveraged overnight or day trades, the payment can be made daily, but their notional value should be limited to the amount they are allowed to risk.

Risk managers should become the partners of traders, not the gatekeepers -- risk managers should be empowered to help keep score and accurately portray to everyone, from the trading floor all the way up to senior management, what the value at risk is at any given time.

The way we're working right now is clearly not working. Now, I don't know if the suggestions above will make a difference, but I think we need to start thinking about how the system can be reformed from within. Without some kind of reform, the public may lose even more trust in the banking system, leading Washington to place restrictions and policies in place that benefit no one.

• Good Measurement Needs Good Leadership (August 20, 2012)