Risk, transparency, and reputation
Again via Steve Hsu, a thought-provoking article on how to make easy money running a hedge fund.
The basic idea is that a fund manager has a good shot at generating outsized returns over a limited time period by selling insurance against a low-probability event. The reason for the high return is the low but significant risk of losing everything.
This would be no problem ordinarily. This is the same reason a treasury note pays less than a corporate bond pays less than credit card debt: the federal government is less likely than a corporation to default, and a corporation is less likely than a single individual.
The problem is in transparency: since the investors don’t know what risks the manager is taking, they can only judge on results over a finite time period. If they were explicitly asked whether they were willing to run a 10% risk of losing everything for an extra 2 points of return, they might make a different decision.
At the root of this problem is compensation. The standard “2 and 20″ structure rewards gains but does not pass through losses, and it reflects a built-in assumption that some kind of reliable algorithm, strategy, or talent lies behind the higher returns. You could argue that the same problem exists with massive CEO stock grants: risky ways of boosting the stock price may pay off for the CEO in the short term, but may hurt the company and its investors in the longer term. In both cases, the manager is highly incentivized to increase risk behind a curtain of reduced transparency.
Which brings us to reputation, which is often cited as a factor that helps to solve this problem. The idea is that the above may work short term, but the manager’s career will be over when this hidden risk-taking inevitably comes to light. But this argument loses its force when today’s pay scales are considered. Loss of reputation has got to seem less threatening to many managers if the short term reward is so high that it permanently moves the manager into a new wealth strata. Many recent articles have made the point that top pay scales are really set by cultural norms; the reputation problem is at least one good argument for lowering these norms.