Wednesday, March 19, 2008

Are Hedge Funds a Scam?

(Naked Capitalism) Martin Wolf comes perilously close to declaring hedge funds a scam in his article today, "Why today’s hedge fund industry may not survive." However, in a departure for the normally clearthinking Wolf, the article conflates issues that would have benefitted from being discussed separately.

First, he seems shocked, shocked that investment managers can make a lot of dough while delivering mediocre returns. Where has he been? It's almost a truism that active managers don't earn their fees; numerous analyses have found that ones that outperform don't sustain it over time. It's just that hedge funds have elevated the fee extraction to unimagined levels.

What Wolf points out is that hedge funds that pursue low risk strategies (ones that exploit predictable patterns that reverse only rarely, say something as mundane as the fact that the yield curve is generally positive) and then leverage it heavily, a hedge fund manager can deliver years of good returns (good these days is in the mid teens; the eagerness of institutional investors to put their money with hedge funds has lowered return requirements). But those strategies will eventually hit that one-in-ten year wall when their strategy blows up and wipes the investors out.

Wolf argues that it's hard for investors to discern luck from skill, which is accurate and vexing. He then invokes George Akerlof's theory of lemon markets, which posits when there is information asymmetry (ie, the seller inevitably knows more about his product than the buyer) that bad supply eventually drives out good. Customers come to recognize that there is only overpriced junk on offer, and the market dies. However, Akerlof's thought example of a lemon market was used cars, and it is still thriving. So Wolf's prediction of hedge funds dying off, or the industry shrinking considerably, seems a bit premature (although costly borrowings alone will make strategies that depend on high gearing much less viable).

Wolf misses quite a few issues: first, there are a number hedge fund strategies that don't use much or any leverage. The term hedge fund initially referred to the ability to go long or short; now as one wag observed, it's a compensation scheme, not an investment strategy. What is pernicious is not hedge funds generally, but the particular pattern that Wolf describes, of using leverage to make a mediocre-but-highly-predictable-most-of-the-time strategy look sexy.

Second, the industry has gotten very clever in selling institutional investors on attributes other than return (that's why they now accept those mid-teen returns as good, when before, the reason to pay those exorbitant fees was the expectation of 25+% returns) was that they deliver synthetic beta (aka alternative beta), or uncorrelated returns, which makes them a desirable addition to a portfolio. But this is a real scam; you can create synthetic beta cheaply, so there is no reason to pay 2 and 20 for it.

What might lead the hedge fund industry to shrink considerably is if a large number of funds in a fairly wide range of strategies were to blow up in a fairly compressed period, say a year to eighteen months, That would give the product such a taint as to deter risk averse investors (particularly institutional investors) for quite a while.

From the Financial Times:
Hardly a week goes by without the implosion of a hedge fund. Last week it was Carlyle Capital, with an astonishing $31 of debt for each dollar of equity. But we should not be surprised. These collapses are inherent in the hedge-fund model. It is even conceivable that this model will join securitised subprime mortgages on the scrap heap.

Every week, 50 of the world’s most influential economists discuss Martin Wolf’s articles on
Getting away with producing adulterated milk is hard; getting away with an investment strategy that adds no value is not. That was the point made by John Kay, in a superb column last week (this page, March 11). With the “right” fee structure mediocre investment managers may become rich as they ensure that their investors cease to remain so.

Two distinguished academics, Dean Foster at the Wharton School of the University of Pennsylvania and Peyton Young of Oxford university and the Brookings Institution, explain the point beautifully*. They start by asking us to consider a rare event – that the stock market will fall by 20 per cent over the next 12 months, for example. They assume, too, that the options market prices this risk correctly, say at one in 10. An option costs $0.1 and pays out $1.

Now imagine that we set up a hedge fund with $100m from investors on the normal terms of 2 per cent management fees and 20 per cent of the return above a benchmark. We put our $100m in Treasury bills yielding 4 per cent. We also sell 100m covered options on the event, which nets us $10m. We put this $10m, too, in Treasury bills, which allows us to sell another 10m options. This nets another $1m. Then we go on holiday.

There is a 90 per cent chance that this bet will pay off in the first year. The fund then grosses $11m on the sale of the options, plus 4 per cent interest on the $110m in Treasury bills, for a handsome 15.4 per cent return. Our investors are delighted. Assume our benchmark was 4 per cent. We then earn $2m in management fees, plus 20 per cent of $11.4m, which amounts to over $4m gross. Whatever subsequently happens, we need never give this money back.

The chances are nearly 60 per cent that the bad event will not occur over five years. Since the fund is compounding at a rate of 11.4 per cent a year after fees, we will make well over $20m even if no new money is attracted into this apparently stellar enterprise. In the long run, however, the bad event is highly likely to occur. Since we have made huge profits, our investors have paid us handsomely for the near certainty of losing them money.

The immediate response may be that so naked a scam is inconceivable. Well, imagine a fund that leverages investors’ money by borrowing massively in short-term money markets in order to purchase higher-yielding paper. Assume, again, that the premium gives a correct estimate of the risk. With sufficient leverage, this fund, too, is likely to make profits for years. But it is also very likely to be wiped out, at some point. Does this strategy sound familiar? It certainly should by now.

We can identify two huge problems to be solved. First, many investment strategies have the characteristics of a “Taleb distribution”, after Nicholas Taleb, author of Fooled by Randomness. At its simplest, a Taleb distribution has a high probability of a modest gain and a low probability of huge losses in any period.

Second, the systems of reward fail to align the interests of managers with those of investors. As a result, the former have an incentive to exploit such distributions for their own benefit.

Professors Foster and Young argue that it is extremely hard to resolve these difficulties. It is particularly difficult to know whether a manager is skilful rather than lucky. In their telling example, the chances are more than 10 per cent that the fund will run for 20 years without being exposed. In other words, even after 20 years the outside investor cannot be confident that the results were not being generated by luck or a scam.

It is also tricky to align the interests of managers with those of investors. Obvious possibilities include rewarding managers on the basis of final returns, forcing them to hold a sizeable equity stake or levying penalties for underperformance.

None of these solutions solves the problem of distinguishing luck from skill. The first also encourages managers to take sizeable risks when they are close to the return at which payouts begin. Managers can evade the effects of the second alternative by taking positions in derivatives, which may be hard to police. Finally, even under the apparently attractive final alternative it appears that any clawback contract harsh enough to keep unskilled managers away will also discourage skilled ones.

It is obviously best not to pay the manager, as a manager, at all, but rather to invest alongside him, as at Berkshire Hathaway, Warren Buffett’s investment company. But we still have the challenge of knowing whether the manager is any good. We know this today of Mr Buffett. Fifty years ago, that would have been very hard to know.

What we have then is a huge “lemons” problem: in this business it is really hard to distinguish talented managers from untalented ones. For this reason, the business is bound to attract the unscrupulous and unskilled, just as such people are attracted to dealing in used cars (which was the original example of a market in lemons). The lemons theorem states that such markets are likely to disappear. The same may happen to today’s hedge-fund industry.

Now consider the financial sector as a whole: it is, again, hard either to distinguish skill from luck or to align the interests of management, staff, shareholders and the public. It is in the interests of insiders to game the system by exploiting the returns from higher probability events. This means that businesses will suddenly blow up when the low probability disaster occurs, as happened spectacularly at Northern Rock and Bear Stearns.

Moreover, if these unfavourable events – stock market crashes, mortgage failures, liquidity freezes – come in stampeding herds (because so many managers copy one another), they will say: “Nobody could have expected this, but, now that it has happened to all of us the government must come to the rescue.”

The more one believes this is how an unregulated financial system operates, the more worried one has to become. Rescue from this crisis may be on the way, but what about next time and the time after next?

No comments: