“Of the alpha, managers get to keep at least half of it. But there’s still substantial net alpha,” Ibbotson says.

Among the hedge funds, emerging strategies provided the most alpha for investors’ fees, generating an alpha/ fee ratio of 1.21, followed by long/ short equity with 1.16. Equity market neutral, global macro and managed futures delivered less alpha for clients’ fees, producing alph/fee ratios of 0.68, 0.58 and 0.36.

An earlier version of the paper, written in September 2009, found hedge funds produced a return of 9.9 per cent, before fees, which consisted of 1.95 per cent alpha and 4.47 per cent beta. This means that 3.48 per cent of the return was absorbed by manager fees.

In this working paper, the authors note that hedge funds are a collection of very dynamic and relatively young investment strategies, and are expected to evolve further in time. But since they now hold more than US$1 trillion, and continue to attract capital, “we cannot be assured that the high past alphas we measure are a good prediction of the future alpha in the hedge fund industry,” the researchers write.

Survivor: hedge fund land

Over the 15-year research period, more than half of the hedge funds that Ibbotson, Cheng and Zhu tracked either blew up or dropped out of the TASS database. The researchers observed that 7,413 funds were withdrawn. Worryingly, the performance of these funds does not contribute to the aggregate returns data. Because TASS monitors successful funds only, the researchers adjusted for “survivorship bias” by including the performance of the 7,000-plus dead funds in their calculations.

They also adjusted for “backfill bias”. This occurs when a hedge fund joins the database after a run of strong monthly returns, and includes this data in their performance history, bolstering their long-term return. The problem is that funds with an unfavourable return history do not include this data. This distorts the aggregate returns from the database because poor past performances are ignored, while plenty of good months are dragged into the database. Backfill data is always biased, Ibbotson says, because “managers only show it if it’s good”.

By stripping out these biases, the researchers effectively took an axe to the headline returns published by TASS: the original 13.23 per cent performance of the funds over the study period was almost halved to 7.63 per cent.

Join the discussion