Digital Assets Report


Like this article?

Sign up to our free newsletter

Barclays Capital Equity Gilt Study 2006, Chapter 4: Implications and limitations of hedge fund returns, by Sree Kochugovindan

Related Topics

This extract from BarCap’s Study examines the historical performance of hedge fund returns, and considers the relative merits/limitations o

This extract from BarCap’s Study examines the historical performance of hedge fund returns, and considers the relative merits/limitations of the available indices.

The potential for hedge funds to generate positive returns in any market environment has raised the popularity of the asset class among institutional investors.

Despite a lacklustre performance in 2005, attractive returns at the start of 2006 have rekindled publicity over the merits of the asset class. However, there are also increasing doubts over the ability of hedge funds to generate alpha. Limited transparency and the lack of accurate performance indicators have led to a growing debate over whether hedge funds are in fact providing alpha, or just providing alternative betas.

Have the returns been overestimated? What are the limitations of the existing indices?

In the previous chapter, we followed an academic approach in order to distinguish between the alpha and beta components of hedge fund returns. In this chapter, we take a closer look at the relative merits of the return measures themselves.

At first glance, hedge funds appear to provide higher returns and lower volatility in comparison to more traditional assets. Between 1994 and 2005, the average annualized return of the S&P 500 was 10.6% with a return/risk ratio of 0.5. Over the same period, the CSFB/Tremont Hedge fund index possessed a slightly higher annualised average return of 11% and a return/risk ratio of 1.1.

The CSFB equity long/short index also outperformed the S&P with an average annual return of 13.6% and a return to risk ratio of 0.9. However, hedge fund indices are prone to some well-known biases, which may provide a distorted picture of the true returns. The three main biases most frequently mentioned in the literature are, selection bias, survivorship bias and instant history bias.

Selection bias can arise as a result of voluntary reporting. Index providers rely upon hedge fund managers to voluntarily submit return data. Not all managers are willing to provide this information, so the index will not be truly representative of the hedge fund industry.

Managers may prefer not to report during periods of particularly poor performance. Meanwhile, managers that have reached their funding limits may not report results as they are closed to new investments. As the performance of funds that choose to report may be systematically different to other funds, voluntary reporting is likely to skew the data.

Instant history bias occurs when a fund first enters the index. New funds may wish to accumulate a track record, waiting for strong performance before volunteering to report to an index provider. Once these funds enter, the funds performance history is instantly backfilled, producing an upward bias.

Most index providers only include information on funds still in operation during the current sample period. Survivorship bias is introduced as funds that have closed simply fall out of the index. This leads to an upward bias in index returns and a downward bias in historical volatility. There are many studies that have attempted to quantify this bias. Estimates range from 2% to 4% on aggregate index returns.

A study by Edhec business school found that a large share of survivorship bias originates from the smallest funds in the database. By comparing data between 1994 and 2003, they estimated that the impact of survivorship bias on an equally weighted index was 2.44%, while for the same period, asset weighted returns were only biased by 0.85%. Thus, one way of analysing the hedge fund industry and reducing survivorship bias could be to examine asset-weighted indices.

Another method of reducing the above data biases is to monitor the returns of fund of hedge funds. Fund of funds data are less vulnerable to all three of the aforementioned biases. If a fund of funds invests in a hedge fund that chooses not to report to any database, or closes at a later date, the returns of that particular hedge fund will still be reflected in the historical performance of the fund of funds manager. Thus, survivorship bias and self-selection bias can be significantly reduced. Instant history bias is also reduced.

If a fund of funds manager invests in a new hedge fund, the history of that hedge fund will not be included. All we see are the historical returns of the fund of funds. As a result, it is possible to reduce all three of the main biases that distort returns.

Analytical issues can still arise as a result of the short data history. Hedge fund returns data only begins in the 1990s; in the case of equity/long short funds this coincides with a long bull run and only a few years of decline. When considering any of the investment styles, there is no way of accurately analysing fund performance across different market environments.

Style classifications may also introduce distortions. It is not always possible to determine the specific strategy or style adopted by the fund manager. There may be an overlap in style groups making it difficult to allocate a specific fund into one particular group. Classification bias can be aggravated if the index provider does not interview the fund managers, but allows the fund to classify themselves.
In short, although there are risks of being misled by the biases inherent in the index construction, it is possible to gain a better indication of hedge fund performance by using asset weighted or fund of hedge funds data. The reduction of return biases in fund of funds indices allows us to provide more reliable comparisons with traditional asset classes. We include real fund of fund returns in a portfolio with the S&P, 10 yr Treasuries, cash, property and commodities.

For the sake of comparison, we run one unconstrained optimisation with fund of funds and one without. The results suggest that inclusion of fund of funds in an unconstrained portfolio lowers the risk at each return level by almost 1%.

Comparing the optimal weights across the two portfolios, we find that the inclusion of fund of funds comes at the expense of equities and commodities, as the weights of these more volatile assets are reduced in favour of hedge funds.

Portfolio optimisation analysis is inherently backward looking as optimal weights are selected on the basis of historical returns. Going forward, the main concerns over the future growth of the hedge fund industry relate to capacity constraints.

Recent years have witnessed huge financial inflows into hedge funds. The more money that flows into hedge funds, the more likely it is that trades and strategies will be crowded.

Assuming that hedge funds are exploiting market inefficiencies supplied by other less skilled investors, there will inevitably be a natural cap on the potential size of the hedge fund industry. Weaker returns over the last two years are a possible indication that this trend may already have begun. However, a low volatility environment combined with market dislocations may also have been responsible.

Evidence of capacity constraints have been found by Agarwal, Daniel and Naik (2003). Using data from 1994 to the end of 2000, they found that large funds with large inflows were more likely to display weak future growth. Another way of spotting diminishing opportunities is to see if absolute returns are in decline, or if the spread between returns of the best and worst performing managers narrows. Analysis in 2004 by Watson Wyatt indicated that fixed income, statistical and convertible arbitrage funds were experiencing either lower returns, narrowing manager spreads or both.

Convertible arbitrage is often highlighted as an example of the impact of capacity constraints. Hedge funds have a very strong presence in the convertibles market, so there is greater pressure for managers to be highly skilled in order to extract excess returns.

Macro funds on the other hand operate in many different markets. There is still a tendency for managers to focus on the same trade, but not to the same extent as in some of the arbitrage funds.

In short, although some strategies may have begun to feel the impact of capacity constraints, this is not the case across the broad spectrum of styles.

In summary, there are many biases inherent in the construction of hedge fund indices. However, these indices still perform a crucial role in providing a benchmark and an indication as to how the industry is developing. Understanding the biases allows us to at least try and control for them, via more reliable indices, such as fund of hedge funds.

Concerns over capacity constraints and the risk of restraining future growth, have been extensively discussed. Certain strategies may be more vulnerable to capacity constraints than others. However, there still exist funds with access to a broad range of markets and instruments, and which are unlikely to be reaching their capacity limits in the near future.

Even if we expect hedge fund returns to decline as the industry grows, the broad spectrum of styles available to fund of fund managers, still provide the opportunity to extract alpha.

To download a pdf copy of the Equity Gilt Study 2006 by Barclays Capital, visit

Like this article? Sign up to our free newsletter

Most Popular

Further Reading