Saturday, August 1, 2009

Capitalization, not compensation



The way to control excessive risk taking is through proper capitalization and proper cost of capital, not with compensation restrictions. Compensation practices should be left to the shareholders.

From CNBC:
A new bill giving the SEC power to directly limit compensation for Wall Street employees will help put an end to a culture of excessive risk-taking, Congressman Barney Frank told CNBC Tuesday.

What is "excessive risk" anyway? The answer is always "it depends". If you buy a $250,000 home, are you taking too much risk? How about a $500,000 home? A million dollar home? At some point the purchase looks like excessive risk. But that point is different for everyone, isn't it?

The amount of risk a financial institution is allowed to take should be proportional to the firm's capitalization. The whole reason AIG, Citi, Lehman, RBS, etc. became insolvent has to do with the fact that these firms were completely undercapitalized for the amount of risk they were taking.

Here is a dirty secret that has existed on Wall Street for years. Compensation excesses were generally directly linked to outsized returns on capital. That is businesses at many financial firms engaged in "regulatory capital arbitrage". They used the firm's low capital requirements or found loopholes in regulatory capital to be able to take significant risks with minimal capital usage. They took advantage of cheap capital the banks made available to them. And they got paid for it.

One of the key issues is that risk takers tend to get the benefit of cheap capital the financial institution is able to access. Back in 1996 a junior emerging markets trader at a large investment bank was purchasing Ukrainian bonds yielding 40%. The bank was financing these at LIBOR. He made over $2 million in bonus that year. When asked what value he was adding, the answer was "you are barking up the wrong tree. I made millions for the firm last year." But in reality a kid can do this: (40% - 6%) x purchase amount = large P&L (using firm's capital). What's wrong with this picture? The answer is improper financing charges to use the firm's capital. If you lent that guy money, would you charge LIBOR? No, you would want to be compensated for the risk you are taking. So should the financial institution.

There is a reason these bonds were yielding 40% - they defaulted a couple of years later. If the bank said that in order to buy a bond like this, the firm will charge the desk 25% financing charge (instead of LIBOR), the situation changes. All of a sudden, neither the return looks that spectacular, nor is the compensation so ridiculous. It may still be a worthwhile risk to take, but this type of approach creates a more leveled playing field and creates appropriate incentives.

With proper capitalization and capital charges, return on capital becomes more modest and so does compensation. To the extent someone can generate outsized revenues without taking much risk (and reserving much capital), they should indeed receive outsized compensation. And the government should have no right to restrict that person's pay, as only the shareholders can decide on the compensation structure. Obviously if the government owns the firm, they are free to do with it what they want - including running it to the ground. Product innovation, client service, market making, etc. are some examples of areas where incentives are important. Otherwise we are quickly marching toward socialism.

The key of course is to watch those areas of the firm that generate higher than normal returns, and make sure the capital and the financing charges that the business uses are sufficient for the risks they take. That's the type of role a bank regulator should focus on instead of trying to decide which bonuses will be permitted this year.