Do funds of hedge funds really add value for investors?

Funds of hedge funds remain the most oft-used funnel to pour money into hedge funds. So does using this conduit increase chances of profit?

In the aftermath of the bursting of the internet bubble, institutional investors suffered considerably from the long bear markets, confirming that geographic, sector and (traditional) style diversification generally turn out to fail precisely when they are needed the most (see, for example, Longin and Solnik (1995)).

In an attempt to address this issue, an increasing proportion of investors have showed interest in the diversification benefits of alternative investment strategies. Most of them are now planning to include hedge fund strategies in their global asset allocation to reduce downside risk, especially in times of market turmoil. However, hedge fund investing is not straightforward.

Hedge fund strategies are significantly more sophisticated than the typical buy-and-hold strategies implemented in the traditional world, and information in the alternative arena is scarce. As a result, traditional investors rarely have the resources (for example, time, knowledge, network, assets under management) to invest directly into single hedge funds. Instead, they generally gain exposure to hedge fund strategies through third parties like funds of hedge funds.

They can thus benefit from the experience of professional fund of hedge funds managers, and from the diversification effect offered by large investment pools.

However, by externalising not only strategic and tactical style allocation decisions but also the fund selection process, investors enter into a principal/agent relationship with the fund of hedge funds manager. Investors are therefore exposed to both visible (that is, management and incentive fees) and invisible costs (that is, agency costs). Our objective in this article is to find out whether funds of hedge funds manage, against this backdrop of costs, to add value through strategic style allocation and/or active management (that is, tactical style allocation and fund picking).

The remainder of this article is organised as follows. In the first section, we will see how the performance of funds of hedge funds can be decomposed and we will elaborate on the construction of funds of hedge funds' strategy benchmarks.

In the second section, we will see how one can use these strategy benchmarks to measure the value added by funds of hedge funds at the strategic style allocation level and through active management - tactical style allocation and fund picking.

In the third and final section, we will briefly present the empirical results obtained in Amenc and Vaissié (2005).

Decomposing Performance

The performance of a fund of hedge funds, like that of any other investment vehicle, can basically be broken down into two components: normal and abnormal returns. While the first component is the fair reward from the market for a long-term exposure to risk factors - risk premia - the second component is due to the talent of the fund manager, that is, factor timing and fund picking.

Nevertheless, although they are more transparent than single hedge funds, funds of hedge funds do not readily communicate detailed positions. It is therefore technically difficult to identify the key drivers of the performance of a specific fund of hedge funds. A pragmatic solution to circumvent the model risk issue involves using style factors. By doing so, the performance of a fund of hedge funds can either be attributed to its investment style or to active management - tactical style allocation and fund picking. We obtain the following decomposition:

Where is the return of the fund of hedge funds at time t, is the strategic exposure of the fund of hedge funds to style factor k, is the return of style factor k at time t, a is the intercept term and white noise.

The investment style of a fund of hedge funds is defined by its strategy benchmark, that is, its long-term exposure to the different investment styles (see Kuenzi (2003) for a formal definition). Unfortunately, this information is rarely available on a real-time basis, if at all. We are thus generally compelled to run a return-based style analysis (RBSA) to estimate the strategy benchmark (see Sharpe (1992) for more details on the style analysis). As shown in Fung and Hsieh (1998), traditional asset class indices are little help in capturing the variability of hedge fund returns.

We will therefore replace them with pseudo-risk factors embedding the specific risk features of hedge funds, namely hedge fund indices.

In an RBSA, the quality of the output is directly linked to the quality of the inputs. Therefore we need to use a set of indices that are both collectively exhaustive and mutually exclusive in order to obtain reliable results. For this reason, we suggest selecting the Edhec Alternative Indexes to proxy the return of hedge fund strategies (see Amenc et al. (2004) for a detailed discussion on the appealing properties of these indices). Furthermore, when constraints are imposed on weights, for example portfolio and/or short selling constraints, the confidence interval of the different model parameters are not readily available anymore.1

In an attempt to address this issue and improve the robustness of style models, one can implement the procedure introduced in DiBartolomeo and Lobosco (1997) to identify the style factors that present significant explanatory power. The factor loadings resulting from this RBSA are subsequently used to construct the strategy benchmark.

Measuring Added Value

At first sight, once the strategy benchmark has been constructed, it is straightforward to measure the value added by a fund of hedge funds. We simply calculate the difference between the average performance of the fund of hedge funds and that of its strategy benchmark.

However, active management is not the only source of added value for funds of hedge funds. Indeed, the first source of added value is the choice of the strategy benchmark itself. As highlighted in the literature (see Amenc and Vaissié (2005)), the strategic allocation of funds of hedge funds is the main determinant of their performance. Considering the strategic style allocation as an exogenous variable would therefore make no sense and lead to a biased estimate of the value-added.

Measuring the value added by a fund manager at the strategic allocation level raises the question of the choice of the strategy benchmark's benchmark or neutral portfolio. Unfortunately, there is no clear answer in the literature to the neutral portfolio selection problem.

Following equilibrium logic, market capitalisation might be used to construct the neutral portfolio. However, information on market capitalisation is not available in the alternative arena. Following capital preservation logic, the minimum risk portfolio appears to be a natural candidate, but this approach already requires a certain expertise in terms of asset allocation, and as a result, gives a poor proxy for the behaviour of uninformed investors. We therefore suggest using as a neutral portfolio a portfolio that is evenly invested across the different hedge fund strategies.2

Once we have the neutral portfolio, and the strategy benchmark at our disposal, we can estimate the value added by the fund of hedge funds. To do so, we simply proceed as follows:

Where is the return of the fund of hedge funds at time t, (respectively ) is the strategic exposure of the fund of hedge funds (respectively of the neutral portfolio) to style factor k, is the return of style factor k at time t, and a is the intercept term.

Empirical Results

In Amenc and Vaissié (2005), the authors analyse the performance of 97 funds of hedge funds belonging to the Alternative Asset Centre (AAC) database and showing a continuous track record from January 1997 through to December 2004. Their first conclusion is that a majority (57%) of funds of hedge funds succeeded in adding value.

On average, these funds of hedge funds even added as much as 3.50% per annum over the neutral portfolio. However, what is interesting to note is that while almost 90% of the funds of hedge funds in the sample added value at the strategic asset allocation level (2.1% per annum on average), only 31% succeeded in doing so through active management.

The good news is that these 31% added 3.25% per annum on average; the bad news is that the remaining 69% destroyed 2.4% per annum on average (see illustration 1 below for detailed results).

In other words, while strategic allocation seems to be a relatively safe, though rewarding, exercise, active management, on the other hand, appears to be a double-edged sword. This is an interesting finding as in practice most fund of hedge funds managers continue to neglect the former to concentrate their efforts on the latter (see Edhec (2003)).

As we have seen in this article, funds of hedge funds basically dispose of two sources of added value, namely strategic style allocation and active management (that is, tactical style allocation and fund picking). Empirical results indicate that the majority of funds of hedge funds, namely 57%, do add value. As a result, we can say that the fund of hedge funds market offers appealing investment opportunities for investors.

However, contrary to what they generally claim, it is at the strategic allocation level and not through active management that fund of hedge funds managers add most of the value. Indeed, only a small proportion of fund of hedge funds managers, namely 31%, appear to have succeeded in adding value through active management. In other words, in the alternative arena, just like in the traditional world, strategic allocation appears to be the key determinant of the performance of a fund.

Based on this evidence, it would not be surprising to see the development, in the near future, of funds of hedge funds made up of investable hedge fund indices and focusing on the portfolio-construction process.

author: Mathieu Vaissie, edhec

Mathieu Vaissié, research engineer with the Edhec Risk and Asset Management Research Centre.

Footnotes

1 Please note that performance measurement biases (for example survivorship bias, instant history bias) may have a significant impact on the performance of the strategy benchmark. Particular attention must therefore be paid to this issue and appropriate adjustments should be made to underlying indices when necessary.

2 See footnote 1, above.

references

Amenc, N., Martellini, L., and M. Vaissié, 2004, "Indexing Hedge Fund Indices", in Intelligent Hedge Fund Investing, ed. Barry Schachter, Publisher: RiskBooks.

Amenc, N., and M. Vaissié, 2005, "Determinants of Funds of Hedge Funds' Performance", working paper.

DiBartolomeo, D., and A. Lobosco, 1997, "Approximating the Confidence Intervals for Sharpe Style Weights", Financial Analysts Journal, 53, 4, pp 80-85.

Edhec, 2003, "Edhec European Multi Management Practices Survey".

Fung, W. and D. A. Hsieh, 1998, "Performance Attribution and Style Analysis: From Mutual Funds to Hedge Funds", working paper.

Longin F., and B. Solnik, 1995, "Is the Correlation in International Equity Returns Constant?", Journal of International Money and Finance, Vol.14, pp 3-26.

Sharpe, W.F., 1992, "Asset Allocation: Management Style and Performance Measurement", Journal of Portfolio Management, 18, 2, pp 7-19.