Posted by: Aaron Pressman on August 30, 2006
CXO Advisory Group has posted a write-up of a fascinating paper by MIT’s resident financial genius, Professor Andrew Lo, and grad student Jasmina Hasanhodzic. The paper examines the performance of 1,610 hedge funds from 1986 to 2005 in 11 investment categories to determine how much of the returns could be attributed to manager smarts (known in the trade as alpha) versus exposure to basic market movements in bonds, stocks, currencies and so forth (known as beta).
Overall, managers were responsible for 61% of the performance of the hedge funds studied on average, a vindication perhaps for the high fees charged, but the results were more interesting broken down by category.
For convertible arbitrage and managed futures, the average contribution of managers was a net negative to performance. At the other end of the spectrum, for dedicated short selling funds, managers contributed 226% of net performance (since the price of stocks and commodities generally rise, they have a lot to make up for). Investors seeking exposure to the uncorrelated benefits of convertibles or managed futures must pick a top-tier manager or else seek a less costly, index-like vehicle.
Following up on that notion, Lo then sought to replicate the hedge fund strategies using portfolios of futures and forward contracts on the S&P 500, Lehman AA corporate bond index, U.S. dollar index, Goldman Sachs Commodity Index and a Lehman corporate to Treasury yield spread index. The clones weight each of the five market exposures to match each of the real hedge funds on a monthly basis. Over the same almost-20 year period, the mindless clones exceed the actual funds in four categories, nearly match their performance in four categories and fall short in three categories.
Still, although the academic research is promising, Lo concedes that it’s just a start on the road to constructing actual clones that well-to-do or institutional investors could use. The model didn’t account for transaction costs and the simplistic five-factor design could be augmented by including more market exposures. “We are cautiously optimistic that the promise of our initial findings will provide sufficient motivation to take on these practical challenges,” Lo and Hasanhodzic conclude. Motivation indeed.