In the mutual fund industry, which is based on a passive investment approach, and where respecting the tracking error is an inevitable notion, the use of indices is necessary in order to play on exposure to the market. In the hedge fund universe, where it is frequently said that performance is extracted from managers, reflecting active asset management, the implementation of hedge fund indices may be surprising, because the notion of index is commonly associated with the notion of passive management.

However, picking the best performers in the hedge fund universe appears to be a very challenging task. Hedge funds tend to be extremely secretive about their performance and their investment strategy, making it very difficult for investors to differentiate between returns explained by the style of the fund (i.e. beta drivers) and returns generated by the skill of the manager (i.e. alpha drivers). On the other hand, according to different academic studies (for example Naik and Agarwal (2000), or Ko, Ko and Teo (2004)) the existence of performance persistence is not well-established. In an attempt to rationalise the whole investment process, institutional investors increased pressure on the industry to provide them with hedge fund benchmarks at the beginning of the 90s.

Traditional performance measures such as the Sharpe ratio do not account for hedge fund risks (e.g. exposure to multiple risk factors, extreme risks, etc.). In other words, traditional performance indicators cannot inform investors about hedge funds’ risk-adjusted performance. For this reason, investors turned to multi-factor models to measure hedge fund alphas. Unfortunately, traditional multi-factor models also fail to properly account for the specific characteristics of hedge funds (e.g. dynamic and non-linear exposure to risk factors). Building on Glosten and Jagannathan’s (1994) contingent-claim-approach, some attempts have been made to capture hedge funds’ non-linear exposure to risk factors through the use of options. However, though theoretically robust, these models are characterised by high model risk (i.e. problem of misspecification). A practical alternative to this approach consists of using factors embedding hedge funds’ original risk characteristics. Hedge fund indices therefore appeared to be ideal candidates to serve as pseudo-risk factors. Such models must however be handled with care. Investors have to bear in mind that the relevance of the results strongly depends on the quality of inputs.

In this respect, the development of peer indices (i.e. non-investable indices) has been a strong response to the need of investors for a better understanding of hedge fund performance. A multitude of “boutiques” specialised in hedge funds (HFR, CSFB/Tremont, MAR/Hedge, etc.) have launched their own indices relying on different databases (HFR, TASS, MAR/Hedge, etc.), following varying construction methodologies and diverse management principles. Furthermore, it is worth noting that index providers have themselves made considerable efforts towards increasing their own level of transparency. Some of them now dispose of an independent index committee, and most of them publish the composition and the details of the construction methodology on dedicated websites.

More recently, firms specialized in hedge funds accompanied by respected traditional financial institutions (e.g. S&P, MSCI, Dow Jones, FTSE) have launched series of investable hedge fund indices offering investors a low-cost solution compared to funds of hedge funds (henceforth FoHF) to gain exposure to hedge fund strategies. Generally based on platforms of separate accounts rather than on hedge fund databases, this new generation of indices has been able to provide investors with improved liquidity (up to weekly) and increased transparency (i.e. managed accounts allow for full transparency and daily pricing). In addition, for all these indices, the composition, construction methodology and management principles are overseen by an independent committee and disclosed to the public. The success of these indices was almost immediate and assets managed by investment vehicles linked to these indices rapidly reached 10 billion dollars.

In Anson (2003), the author highlighted the demand of institutional investors for relative returns. After an overview of hedge fund indices, the author concluded that the alternative industry is still maturing, and that hedge fund indices are not good indices since they provide investors with a somewhat confusing picture of the performance of hedge fund strategies. He thus suggested that investors pay particular attention to the adequacy of the index characteristics, and the investor’s objective. In the meantime, however, index providers have improved their construction methodology and management principles. In parallel, new series of investable hedge fund indices have been launched. Our contribution will thus be to revisit the issue of hedge fund strategy benchmarks in light of these new elements, and answer the question that what implicitly asked in Anson (2003) but could not be answered by the author due to lack of elements. Can investors in the alternative arena measure the relative return of hedge funds? In other words, can investable and/or non-investable hedge fund indices provide investors with useful tools for performance measurement?

This study proposes an overview of hedge fund strategy benchmarks, on the basis of academic studies, the points of view of practitioners and documentation from index providers. The remainder of the article is organized as follows. The first section discusses the relation between purposes and requirements. The second section gives an overview of “non investable” hedge fund indices while the third section examines the latest industry development, namely “investable” hedge fund indices.