Morningstar Advisor - Spring 2007 - 17

of this flexibility to manage their reported returns.3 Whether or not returns are “managed,” illiquid securities in a portfolio will often lead to an appearance of lower return volatility than is really the case. Smoothing reduces the reliability of the statistics that we most often use when evaluating funds. It lowers volatility, raises Sharpe ratios, raises alphas, and lowers betas.4 The latter two effects are particularly important for measuring how much a fund hedges. Applying conventional regression analysis to smoothed returns overstates the return and diversification benefits of many hedge funds. Because of this effect, many studies apply a modified form of regression analysis to hedge fund data. However, these are not the sorts of analyses an advisor is likely to encounter. The statistical analysis that advisors typically come across regarding hedge funds need to be taken with a large grain of salt. Fortunately, smoothing does leave a rather large statistical footprint on return data, so an advisor can use serial correlation as one gauge of how large a grain to use. Serial correlation is the correlation between a return series and itself lagged one or more periods. Before looking any further at a report on a hedge fund’s historical performance, first ask how much serial correlation is in the data. The higher the serial correlation, the more advisors need to downsize the reported alpha and upsize the reported beta. The greater the serial correlation in a fund's return history, the more its hedging benefits tend to be overstated in a conventional analysis. To see just how much serial correlation hedge fund data contains, we estimated the onemonth serial correlation coefficient for all of the funds in the Morningstar hedge fund database for which there were at least 60 consecutive months of returns (a little more than 1,700 funds). We found that close to one out of five funds had a serial correlation coefficient of at least 30%, a significant minority of a large sample of funds. Before recommending a fund, advisors must achieve a clear understanding of how, and how much, individual hedge funds actually hedge, how much alpha they provide, and how their returns are estimated, reported, and analyzed. After all, looks can be deceiving. K Paul D. Kaplan, Ph.D., CFA, is vice president of quantitative research at Morningstar, where Jodie Gunzberg, CFA, is vice president of hedge funds. 1 We thank Jim Knowles of York Investment Strategies (Toronto), a founder of what is now Morningstar Canada, for his comments. 2 A fund that enables investors to reduce markets exposure in a portfolio context, such as a dedicated short fund, is providing a hedge, so that the hedge fund label makes sense, even though it may not use hedging strategies in and of itself. 3 See “Do Hedge Funds Hedge?” by Clifford Asness, Robert Krail, and John Liew, Journal of Portfolio Management, Fall 2001. 4 Understanding beta causes alpha to be overstated if the average excess return on the market was positive over the estimation period. If it were negative, alpha would be understated. Got a tough quantitative problem? Can’t get your head around a complex investing concept? Every quarter, Morningstar’s research team will tackle a question from readers. E-mail your queries to magazine_editor@morningstar.com Smoothed Returns Nearly 20% of all hedge funds have a relatively high serial correlation, indicating that their benefits may be exaggerated. 23.71% 23.65% 18.84% 9.68% 10.84% 5.86% 4.00% .06% –0.6 .06% –0.5 .46% –0.4 .87% –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4 1.51% 0.5 .17% 0.6 .17% 0.7 0% 0.8 .12% 0.9 1.0 Range of Serial Correlation MorningstarAdvisor.com 17
http://MorningstarAdvisor.com
Morningstar Advisor - Spring 2007

Table of Contents for the Digital Edition of Morningstar Advisor - Spring 2007
