Investable Hedge Fund Indices: Illusion or reality? August 2004 Many academic papers have tackled the failure of non-investable hedge fund indices to efficiently represent the universe of hedge funds (for a survey see for instance Duc, 2004). Indeed, hedge fund indices display such a high level of heterogeneity that it cannot be statistically concluded that they track a common phenomenon. In contrast, traditional indices such as the Dow Jones or the S&P 500, despite some differences in construction and composition, do not exhibit significant differences and statistically represent the same market, namely the US stock market. In the case of hedge fund indices, the value of each index is determined more by its composition than by an average performance of the alternative strategies that would be common to all hedge fund indices. In this article, we will focus on investable hedge fund indices. In the same way as their non-investable counterparts, they aim to provide an average picture of the hedge fund universe, with the added constraint of exclusively including open hedge funds in their composition, and are thus subject to the same lack of representativeness. Promoters of investable indices firmly disagree with this point of view and argue that they provide a higher quality standard than non-investable indices. Obviously, creating a portfolio of hedge funds in which it is possible to invest is more demanding than managing a theoretical index. While it is certain that an improvement in the quality of representativeness should cause hedge fund selection bias to decrease or, similarly, for heterogeneity between indices to diminish, this is unfortunately not the case. HIGH HETEROGENEITY BETWEEN INVESTABLE INDICES In terms of strategy allocation, the various investable indices are very different. As of November 2003, the comparative analysis of four investable indices (CSFB/Tremont, HFRX, MSCI, S&P) with respect to three broad categories of alternative strategies reveals wide differences in strategy exposures: Macro/CTA exposure varies between 13% and 22%, Long/Short exposure between 13% and 45%, and Arbitrage/Relative Value between 38% and 67%. Investable indices also differ substantially as regards the selected managers that compose the index portfolio. Of the 159 managers covered in total by the S&P, CSFB/Tremont and MSCI indices at November 2003, only 14 (i.e. 9%) were selected in more than one index, and only 2 were included in all three indices. On average, the intersection between two investable indices ranges between 5% and François Duc, Ph. D.
20%, whilst the intersection between 2 large databases ranges between 37% and 59%. Moreover, within a given strategy, investable indices differ not only in terms of weightings and of manager names, but also in the way that the selected managers implement the strategy. This heterogeneity results from a selection process that is not geared to choose representative managers, but hedge funds that are expected to generate high performance. Without going into detail and explaining the difference between each manager, three elements from the actual management of investable indices illustrate this feature. First, there are some hedge funds that apply a pure approach of a given strategy, such as, for instance, systematically investing in all announced mergers and acquisitions. This kind of approach can be seen as a naïve way of implementing the strategy and may be used as a good proxy for a passive strategy index. But, during favourable periods, managers using a naive approach tend to underperform those applying a more selective approach within the same strategy. As a result, such open hedge funds, despite their good representativeness of the strategy, are never included in any investable index. Second, the aggregated underlying positions of all the hedge fund portfolios included in the Dow Jones Convertible Bond Index, which is an investable index, has sometimes demonstrated characteristics of reverse trades, exhibiting a short bond exposure and a long equity exposure. This typically neither reflects a pure strategy nor the average approach of convertible managers. Third, the actual management of investable indices does not seem compatible with the representation of a pure strategy. At the end of February 2003, the GLC Gestalt Europe Fund replaced the Jemmco Fund in the S&P Index. Both managers follow statistical arbitrage strategies, but GLC Gestalt Europe specialises in pair trading on European stocks, whereas Jemmco applies several statistical models (including one pair-trading model) to mainly US-based equities. Hence it is difficult to comprehend how these two hedge funds can represent the same mean approach to a pure strategy or induce the same diversification effects. Note that the Jemmco Fund has recently ceased activity, which tends to prove that the fund was excluded for quality reasons and not for representativeness issues. In fact, the efforts deployed for compiling investable indices are not focused on representing the universe more efficiently, but on improving the quality of the underlying managers. Selecting hedge funds because of their expected quality will expose investors to a strong hedge fund selection bias. This bias, combined with nonhomogeneous strategy allocations and with their smaller size, will imply a larger heterogeneity between investable indices than between non-investable indices. Figure 1 compares the maximum difference in actual monthly returns of investable indices with that of noninvestable indices for 5 different strategies. Generally, the tracking error between investable indices of the same strategy is much higher than the one between non-investable indices. 2
4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 2.0% 1.3% 3.7% 2.9% Maximum absolute difference between investable CSFB/HFR indices Maximum absolute difference between non investable CSFB/HFR indices Average absolute difference between investable CSFB/HFR indices Average absolute difference between non investable CSFB/HFR indices 1.5% 1.3% 1.2% 0.5% 0.6% 0.0% Long Short Macro Convertible Market Neutral Figure 1 INVESTABLE INDICIES: NOT MORE ATTRACTIVE THAN FUNDS OF FUNDS As they are not built to represent the universe but to yield high returns, as they convey an important exposure to manager selection and as they are more heterogeneous than their non-investable version, the so-called investable indices are in fact funds of funds. Promoters of such indices argue, however, that investable indices offer four main advantages as compared to other funds of funds: 1. portfolio management is unambiguous, 2. heterogeneity between investable indices is smaller than that between other funds of funds, 3. investable indices are more profitable, 4. liquidity featured by investable indices is higher. Reality shows that the first three advantages are not accurate and that the last one is provided at a high cost either in terms of fees or risk. Our arguments are given below. Firstly, weightings within index portfolios are not so unambiguous as they are in fine provided through biased samples, however large a database or subset is considered. Moreover, the actual management of investable indices render the weightings questionable. For instance, during the Investor Fachtagung Funds of Hedge Funds conference organised in June 2004 by the Transparency Council Funds of Hedge Funds, Stephan Ewen pointed out that the equallyweighted average performance of the strategy indices was not equal to the performance of the HFRX Equally-Weighted Index from February to May 2004. Moreover, Stephan Ewen expressed surprise regarding the disappearance of the HFRX Managed Futures Index in early May 2004, which took place without any advance warning or explanation on the weightings. Secondly, the apparent higher heterogeneity of funds of funds is partly due to the fact that mono-strategy funds of funds currently represent more than 54% of the fund-of-fund universe, and partly to the fact that some multi-strategy funds have opted for a very active tactical allocation or use structural leverage. One should not mix heterogeneity and absence of a relevant classification. Moreover, as seen above, heterogeneity between investable indices is higher than that between non-investable indices. Additionally, since investable 3
indices feature short track records, there is a tendency to use either non-investable indices performance or simulated performance to measure the heterogeneity between investable indices, which results in underestimating it. Thirdly, investable indices underperform the fund-of-hedge-fund average when using non-simulated returns for comparison purposes. Indeed, having admitted that investable indices are funds of funds, it is clear that utilising extrapolated data prior to index launch would be equivalent to referring to a fund of funds simulated returns an error any sophisticated investor would be reluctant to commit. Such pro-forma data either is the result of an optimisation process (whereby it always tends to be highly favourable) or has influenced hedge fund selection. Although all selectors are aware of the fact that past performance is no guarantee of future results, there remains a natural trend to privilege those hedge funds that have generated attractive returns in the past. Therefore, it is not surprising that, as shown in Figure 2, all investable indices are poorer performers, immediately after their launch, than their non-investable counterparts, whereas this is not the case for simulated periods. Performance 20% 20.9% Non Investable Index Investable Index 15% 10% 5% 10.0% 4.9% 7.7% 8.1% 4.3% 0% CSFB/Tremont Hedge Fund (Aug-03 to Jul-04) HFR Hedge Fund (Apr-03 to Jul-04) MSCI Hedge Fund (Jul-03 to Jul-04) Figure 2 The table below compares the reported actual returns of the CSFB/Tremont, HFRX, MSCI and S&P investable indices with those of the EDHEC Funds of Funds Index1, which is found to outperform all investable indices for the various time periods corresponding to the real track of each investable index. 1 The EDHEC Funds of Funds Index is the best estimator of fund-of-hedge-fund average performance. For investable indices, we use publicly-reported performance data (such as Bloomberg) since the actual launch of each index to conduct comparisons. Thus, these non-simulated figures do not incorporate tracking fees, management fees, entrance and redemption fees charged to investors, nor tracking error-related differences. 4
Performance Period CSFB/ Tremont Investable Index 4.91% August 2003 - EDHEC Fund of Funds Index 7.77% July 2004 HFRX Equally weighted Index 7.67% April 2003 - EDHEC Fund of Funds Index 11.92% July 2004 MSCI Ivest Index 4.26% July 2003 - EDHEC Fund of Funds Index 7.49% July 2004 S&P Hedge Fund Index 12.61% September 2002 - EDHEC Fund of Funds Index 14.37% July 2004 Table 1 Finally, the liquidity offered by certain investable indices is unrealistic and infringes a cardinal risk management rule, whereby all hedge funds must maintain a degree of adequacy between the actual liquidity of the portfolio and that offered to investors. As an example, let us assume that a manager needs several months to sell a portfolio of financial assets at fair market prices. If this manager offers monthly liquidity to investors and that a large number of them simultaneously request to sell out of the fund, he will be forced to liquidate assets at prices that will disadvantage stakeholders interests. Admittedly, investable indices are generally not invested in funds but in managed accounts providing holders with daily liquidity. Nevertheless, a managed account is in reality made up of a portfolio of securities where liquidity can be completely different to that of the account. Therefore, there is no true adequacy between the liquidity of investable indices and that of underlying managed-account positions. In other words, the exposure to liquidity risk is far greater than one may tend to believe, and it is when investors will need it most urgently that liquidity will most likely be refused them. This is the reason why certain index-linked investment vehicles offer attractive liquidity but often charge dissuasive redemption fees within the promised liquidity (using sometimes bid and ask spreads) or put in place some aggregated redemption limits (redemption gates) so they can differ redemptions. NOTHING MORE THAN AN ILLUSION In conclusion, investable indices are funds of funds in disguise: Calling something an index does not make it an index said Professor Schneeweiss in the April 2003 issue of Absolute Return. Indeed, investable indices expose their investors to hedge fund manager selection bias, to index management bias and sometimes to liquidity risk. Reality shows that investable indices currently underperform the fund-of-hedge-fund average. Yet, since the majority of hedge funds that make up the fund-of-hedge-fund universe sometimes outperform the average and that investable indices behave in a similar fashion, we can expect investable indices to occasionally outperform the average. Indeed, we can assume that the quality of the hedge fund selection performed to compose investable indices is in line with that used to build up other funds of funds. The illusion is to believe that investable indices are able to provide something that other multi-strategy funds of funds cannot provide. Bibliography Duc, F., 2004, Hedge fund Indices : status review and user guide Banque Syz & Co. 5
3A S.A. Alternative Asset Advisors Rue du Rhône 30 1204 Geneva Switzerland Tel: +41 (0)22 819 9800 Fax: +41 (0)22 819 0904 E-mail: 3a@3-a.ch 6