Hedge funds have come into public view in recent years as a result of their
growth in numbers and the publicity about their successes and failures. Largely
unregulated and for the most part restricted to individual investors, hedge
funds have features absent in mutual funds that influence their performance.
Using historical data, the authors find that hedge funds net of fees
consistently outperform mutual funds but are more volatile than mutual funds.
Furthermore, they are unable to consistently beat the market on a risk-adjusted
basis, indicating average gross outperformance equal to the fees paid. Incentive
fees explain some of the higher performance but not the increased volatility.
In 1997, assets of hedge funds totaled more than $200 billion. Although fewer in number
and smaller in size than mutual funds, hedge funds have grown in recent years as an
alternative investment vehicle for wealthy individual investors and institutional
investors. A number of features characterize hedge funds and distinguish them from
mutual funds, including a largely unregulated organizational structure, flexible
investment strategies, relatively sophisticated investors, and strong managerial
incentives and investment participation.
Hedge funds are also characterized by strong performance incentives. Their successes and,
in cases such as Long-Term Capital Management, their failures have led to increased
scrutiny by the general investing public. The authors examine the performance of hedge
funds with the objective of linking performance to a number of characteristics, with a
particular focus on the hedge fund incentive fee structure. Besides the focus on
incentive fees, a number of other factors distinguish this research from prior research
on hedge funds. These factors include the use of a large sample consisting of offshore
as well as U.S. funds and the use of monthly instead of annual return data. A broad set
of performance metrics are used and several data-conditioning bias analyses are
completed.
The authors construct the sample from two publicly available hedge fund databases that
contain voluntarily reported return data and include both existing and defunct hedge
funds. Gross monthly return data are adjusted for incentive and management fees. Funds
in the sample had at least 24 months of returns in order to have a sufficient number of
observations to measure risk and risk-adjusted returns. More than 500 funds had monthly
returns for calendar years 1994 and 1995. Besides the two-year sample, the authors also
find results for funds with four, six, and eight years of monthly returns ending
December 31, 1995.
The organizational features of hedge funds should help align the interests of fund
managers and investors. These fund manager features include large-percentage ownership
in the funds, incentive pay as a substantial portion of total compensation, and
liability exposure incurred for being general partners of the funds. These managers also
have substantial latitude and flexibility with regards to investment strategy and
investment alternatives, such as leverage, options, and short selling. The authors
suggest that these features, which are not generally available to mutual fund managers,
produce a clear performance advantage over mutual funds.
For the overall sample and time period examined, the average hedge fund Sharpe ratio was
21 percent higher than comparable mutual fund Sharpe ratios. This advantage over mutual
funds was achieved despite higher total risk. Hedge funds, however, are unable to
consistently beat the market. When compared with eight standard market indexes, the
results are mixed because the time period, index choice, and hedge fund category are all
strong influences. On average, the ability to earn superior gross returns is about equal
to the incentive and administrative fees charged. Although hedge funds appear to offer
little advantage over indexing, the authors suggest that the generally low correlations
of hedge funds with most other asset classes make them a potentially valuable addition
to many investors' portfolios.
The authors also investigate six related data-conditioning biases. Termination and
self-selection bias are the most powerful and, working in opposite directions, remove
any significant effect of survivorship bias on the data. Specifically, funds that
terminated had significantly lower median performance measures than extant funds. The
group of funds that voluntarily ceased to report their returns on average outperformed.
The authors state that several caveats about the results are warranted. The time period
is short, and the diverse investment options make it difficult to classify funds and
identify correct benchmarks. Systematic risk was not estimated with a high degree of
confidence. Also, the approaches used to measure incentive fees did not control for such
complications as varying policies on fee-allocation mechanisms and the treatment of new
and existing investors.