Hedge trimmers

tauringhuang.

Aug 31st 2008
From Economist.com

Separating beta from alpha


HEDGE funds sound great in theory. They aim for returns that are absolute (making money whether markets rise or fall) and uncorrelated with other asset classes. But many investors are put off by the fees: often 2% a year plus a fifth of all profits.

That has prompted a search for a cut-rate version of hedge-fund strategies. Most attention has been focused at the macro level—in other words, trying to reproduce the return of overall hedge-fund indices by “factor replication”. This involves the reverse engineering of hedge-fund returns (such as the performance of the S&P 500 or credit spreads) to see which factors are most important.

As a new paper* from a leading hedge fund points out, the problem with this approach is that it will likely expose investors to many of the factors they own in the rest of their portfolios. It may also create “steak without the sizzle” portfolios, which capture none of the skill—or “alpha”, in the jargon—displayed by hedge-fund managers.

The concept of alpha is a slightly metaphysical one and resembles the “God of the gaps” familiar to Victorians. Traditionally, people were inclined to attribute natural phenomena such as earthquakes and plagues to God; eventually they discovered plate tectonics and bacteria. The role of God was steadily diminished to that which people could not explain by other means.

Similarly, fund managers used to claim all positive returns as evidence of their skill. Then investors figured out that most returns could be generated simply by owning the whole market (beta, as it is known). Beating the market then became evidence of skill. In turn, however, investors (with the help of academics) figured out that the excess return might be due to taking higher risks of exposure to a certain factor, such as smallcap stocks.

In a way, hedge funds, by virtue of their complex strategies, are one of the main perceived repositories of alpha left in the market (the average traditional fund underperforms the market, after fees). The replication concept stems from the idea that a lot of hedge-fund returns are beta, not alpha.

Adam Berger, the author of the paper, is not averse to the idea; he just believes that academics have been going about it the wrong way. Rather than replicate the hedge-fund universe, they should copy individual strategies. Merger arbitrage, for example, tends to involve managers buying shares in the target company and going short (betting on a fall in price) on the predator. Convertible arbitrage consists of a long position in the bond and a short position in the shares.

Such approaches can be replicated and are potentially less likely to be correlated with mainstream asset classes than an industry-wide approach would be. In back-testing, Mr Berger found that a diversified hedge-fund beta portfolio produced an annualised return of 12.8% over the period since 1990, two percentage points a year ahead of the S&P 500; the correlation between the two returns was just 0.2. (A figure of 1 reflects complete correlation; a broad hedge-fund index had a correlation of 0.5 with the S&P.)

A hedge-fund beta portfolio could be given a further fillip through asset allocation. Many strategies become too crowded, reducing the scope for returns (a good example being convertible arbitrage in 2003).

Mr Berger found that weighting the strategies by the funds invested (creating a bias to invest at the top of the market) turned $100 into $450 between 1994 and 2008. In contrast, an equal-weighted approach turned $100 into $650. And a risk-weighted approach (based on the volatility of the underlying strategies) turned $100 into $850.

Back-test enough data, of course, and you can always find a market-beating strategy. And some strategies (such as distressed debt) will be difficult to replicate. Nevertheless, Mr Berger produced ways of mimicking the separate strategies, including short bias, which produces a portfolio focused on the most common shorts in the market.

This approach carries some dangers. It may end up mimicking the most popular hedge-fund bets. But those bets can get crowded as many quantitative managers (including AQR) found in August 2007, when all of their models went on the blink. Since these portfolios also (according to Mr Berger) require skill in implementation, it is not clear how radically different such an approach would be from existing quant strategies. Fees would likely be lower than for a traditional hedge fund but higher than those of an exchange traded fund.

If a fully-fledged product does emerge from this paper, it will present an interesting experiment in the market. The hedge-fund industry has exploded in size, but competition has not brought down fees. Perhaps technology can eventually do so, as the extent of managerial skill becomes ever more rigidly defined.



"Is Alpha Just Beta Waiting To Be Discovered? What the Rise of Hedge Fund Beta Means for Investors”, by Adam L. Berger, Portfolio Solutions Department, AQR Capital Management. Working paper, July 2008.