Over the period of 1997–2006, hedge funds generated higher gross alphas on their long-equity holdings relative to other institutional investors, which suggests that hedge funds have a superior ability to select underpriced securities.
What’s Inside?
The authors compare the gross alphas generated over the period of 1997–2006 with the long-only holdings of seven categories of institutional investors: hedge funds, advisers, pensions, banks, insurance firms, indexers, and specialty/other. A five-factor model that incorporates market risk, size risk, value risk, momentum risk, and illiquidity risk is used to assess the alphas over the course of three separate market conditions: bubble, deflation, and recovery.
How Is This Research Useful to Practitioners?
The research is limited to long-only holdings as reported in SEC Form 13F filings because of the overall lack of transparency around hedge fund holdings and returns. Long positions are merely one component of hedge funds’ overall strategy and total returns, but that component provides a measure of alpha that is comparable with other institutional investors to assess stock selection ability.
The gross alpha generated by each class of institutional investor was positive, but the alpha generated by hedge funds was statistically significant and economically higher than the others. This finding suggests that hedge funds are better at identifying underpriced securities. This finding is also consistent across all three market conditions (bubble, deflation, and recovery eras). The authors determine a high portfolio churn rate and active share deviation from benchmark holdings to be correlated with hedge funds’ superior alpha because funds with the highest churn and active share generated the highest alphas.
In addition to the highest mean alpha, hedge funds generated the highest mean return and highest risk measurements for both volatility and semi-deviation. Standard deviation of market betas for hedge funds is nearly double that of other institutional investors, suggesting that market timing plays a large role in hedge fund strategy. Hedge fund equity holdings are generally characterized as small value firms with high betas and high liquidity. Hedge funds apparently do not provide market liquidity because they load negatively on the illiquidity risk factor, which is intuitive given the high churn of their portfolios.
This research may encourage fund-of-funds managers to decompose the performance of funds into long versus short performance and further into active share versus timing of risk-factor loadings. The fund-of-funds managers may be able to build a superior bundle if they buy the components of various funds.
How Did the Authors Conduct This Research?
Sample data are from the period of 1997–2006 and are broken down into market regimes, or eras, as follows: bubble era from 1997 through Q1 2000, deflation era from Q2 2000 through Q1 2003, and recovery era from Q2 2003 through 2006. The timing of these regime changes is determined by stock market returns, statistical evidence based on tests of parameter stability determined by cumulative sums of the squares of recursive residuals, the National Bureau of Economic Research’s official measures of business cycles, and popular belief supported by the timing of national media coverage.
Ken French’s data library website is the source of returns for the risk-free rate (T-bills) along with the returns of four risk factors: market, firm size, book to market, and momentum. Returns attributable to the fifth risk factor, illiquidity risk, are from the Orissa Group.
Returns for the long-only holdings of institutional investors are obtained from Form 13F filings via the Global Equity Ownership Feed of Thomson Reuters, and institutions with less than $100 million long equity are filtered out. Alphas are deemed robust to the non-Gaussian (fat-tailed, leptokurtic, heteroskedastic) nature of real-world returns.
The authors’ hypothesis that alphas are equal across institutional classes is rejected by the analysis of the variance test and the Kruskal–Wallis equality-of-populations rank test (Journal of the American Statistical Association 1952). The Fisher–Hayter test (Journal of the American Statistical Association 1986) confirms that hedge fund alpha is significantly different. Testing robustness includes re-estimating alpha by using the Fama–French three-factor and five-factor alpha models over 52-week rolling windows.
Abstractor’s Viewpoint
The relative lack of governing regulations imposed on hedge funds, along with a fee structure that incentivizes more aggressive risk taking, lends itself to hedge funds pursuing alpha-maximizing strategies. Such strategies may be inappropriate for institutional investors whose objectives may be benchmark tracking, minimizing specific risk, or risk-adjusted portfolio management. Further research could expand the time frame examined beyond 2006 to include the crash of 2008 and the rebound period of monetary easing. It would be particularly interesting to have such thorough quantitative analysis of recent returns (2009–2014), during which time the S&P 500 Index has outperformed benchmark hedge fund indicators.