"The ABCs of Hedge Funds: Alphas, Betas, and Costs”, Ibbotson, Chen and Zhu (2011). Flashcards
- What question(s) are the researchers trying to answer?
In the paper, the authors attempt to measure and classify the sources of hedge fund returns and look to examine whether or not hedge fund managers, on average, produce returns that justify their fees.
- What prior evidence is there from other studies in relation to this question?
This study updates Brown, Goetzmann, and Ibbotson (1999), who found statistically significant positive alphas after fees in the hedge fund industry over 1989 - 1995.
Kosowski, Naik and Teo (2007) were the first to use Bayesian methods to analyze hedge fund performance and also found evidence of positive alphas after fees across hedge funds from 1994 – 2002, although this wasn’t statistically significant.
Hsieh, Naik and Ramadorai (2007) looked at funds-of-funds over three separate periods, finding significantly positive alpha in one period and insignificant positive alpha in the other two.
- What methodology, data sample and time period do they use?
The authors use monthly hedge fund return data from January 1995 to December 2009, from the Lipper TASS database. From this dataset of over 13,000 funds, they exclude funds that don’t report in $US, funds that report returns gross of fees and funds-of-funds. This gave them a sample of 6,169 funds.
They adjust hedge fund return data to remove commonly-cited bias problems such as back-fill bias (where a hedge fund manager only decides to report returns when they go public – invariably with their winning funds) and survivorship bias (where the data only includes funds that did well enough to survive).
They then break down the returns into alpha (portion of return generated by the manager), beta (portion of return generated by the markets) and cost (portion of return that goes to fees).
- What are their key findings?
The authors decomposed their estimated pre-fee 1995-2009 hedge fund return of 11.13% into fees (3.43%), alpha (3%) and beta (4.7%).
The authors found that the average hedge fund manager produced significantly positive alpha during the time period studied and added value in both bear and bull markets. The year-by-year results show that alphas were positive for every year except 1998.
This consistently high alpha is quite remarkable given the variety of market conditions over the period: the 1990s bubbles, the 2000–02 bear market, the 2003–07 bull market, and the recent global financial crisis. The annual results confirm that hedge funds have added a significant amount of alpha to stock, bond and cash portfolios over the period. The results also show that hedge funds exhibit tactical asset allocation skills, especially in reducing beta exposures during bear markets. For example, the estimated stock beta exposure was quite low during the 2000-02 bear market. Although hedge funds did not avoid the beta exposure in 2008, they nevertheless kept their positive alpha throughout the financial crisis of 2008–2009.
The results show a significant level of alpha in the hedge fund industry and a low correlation with other asset classes, underlining hedge funds’ abilities in adding alpha and reducing beta during bear markets.
Though estimated fees exceeded alpha, fees net of alpha averaged about 43 bps, less than the fees charged by most mutual funds. Moreover the hedge funds provided excellent diversification benefits.
- What is different about the findings of this study relative to prior evidence?
This study differs from previous studies in many ways and can be seen to add real value to the growing literature on hedge fund performance.
This is an admirably careful and meticulous study of hedge fund performance. With access to data from 1995-2009, they analyzed a relatively complete 15-year dataset and the authors were diligent in correcting for both survivorship bias (by including dead funds) and backfill bias (by excluding backfill data). They also included both lagged betas and contemporaneous betas to control for the impact of stale pricing on hedge fund returns (as outlined in Asness, Krail and Liew, 2001).
This is unlike previous articles (namely Brown, Goetzmann and Ibbotson, 1999), who attempted to estimate the impact of survivorship bias without a complete sample of dead funds and recognized the potential selectivity biases in their database. Thus, their survivorship estimate is substantially higher than that of most other researchers (Brown, Goetzmann and Ibbotson 1999; Fung and Hsieh, 2000), but similarly high to Aggarwal and Jorion (2010).
The results emphasize the extent of survivorship and backfill biases in hedge fund data - the sample of funds that existed at the end of the sample period had a compound return of 14.88% net of fees. Including dead funds reduced this return to 11.72%; excluding the backfill data further reduced the return to 7.7% net of fees.
This study also suggests that size of fund does matter with bigger funds outperforming small funds. The largest 1% of funds had a return of 11.59%, the largest 20% of funds had a return of 9.24%, and the smallest 50% of funds had a return of 6.85%. The large funds, however, also had commensurately higher risk. It’s widely speculated among other researchers that hedge funds with large AUM are likely to underperform because their size makes it difficult for managers to find enough investment opportunities to generate superior returns and/or because transaction costs increase with size. However, this is not borne out in the results.
