Hedge Fund Indexes: Benchmarking the Hedge Fund Marketplace

Transcription

1 Hedge Fund Indexes: Benchmarking the Hedge Fund Marketplace Introduction by Mark Anson, Ph.D., CFA, CPA, Esq. 1 CalPERS Investment Office 400 P Street Sacramento, CA mark@calpers.ca.gov Initially, hedge funds were the domain of high net worth individuals. However, in the latter half of the 1990s, large institutional investors discovered the charms of these investments. Endowments were first, followed by corporate and public pension plans. As more and more institutional investors entered the hedge fund arena, they demanded many of the investment parameters from their traditional long-only programs. Generally, with respect to external investment managers, institutional investors demand three things 2 : 1. A well-defined investment process. 2. Transparency. 3. Relative Returns. It is the last requirement, relative returns, that we explore in this paper. Relative returns are one of the primary reasons for index construction. We begin by reviewing issues regarding the construction of hedge fund indexes. Next we compare and contrast the hedge fund indexes in existence. Last, we consider some issues regarding the selection of hedge fund indexes. Hedge Funds as an Investment Before discussing hedge fund indexes, a threshold question must be addressed: should investors consider hedge funds as part of a diversified portfolio? Considerable research has pursued this issue, and the answer is consistently, yes. As Brown, Goetzmann and Ibbotson (1999) note, the most interesting feature of the hedge fund industry is that an investment in a hedge fund is almost a pure bet on the skill of a specific manager. Hedge fund manages tend to seek out arbitrage or mispricing opportunities in the financial markets using a variety of 1 This paper represents the insights and opinions of the author and not the author s employer. 2 See Anson (2001). 1

2 cash and derivative instruments. They tend to take small amounts of market exposure to exploit mispricing opportunities, but employ large amounts of leverage to extract the greatest value. The key point is that hedge fund managers pursue investment strategies unfettered by conventional financial market benchmarks. Their investment styles are alpha-driven rather than betadriven. In Exhibit 1 we show that hedge funds have favorable risk/return benefits compared to traditional stocks and bonds. We present the Sharpe Ratios for several hedge fund indexes as well as stock market indexes and US Treasury bonds and observe that in all cases, the hedge fund indexes have higher Sharpe Ratios than either stocks or bonds. Furthermore, the diversification benefits of hedge funds is documented by many studies. For example, Goldman Sachs & Co. and Financial Risk Management Ltd. (1999, 2000) in two reports covering two different time periods of and , study the interaction of hedge fund returns with traditional asset classes. In their first study they find that a portfolio of 60% S&P 500, 30% Lehman Aggregate Bonds and 10% Hedge Funds outperformed the Pension Plan Index of 60/40 stocks/bonds by 78 basis points with a reduction in portfolio standard deviation by 31 basis points. In their second study (which included the turbulent year of 1998) they find that the portfolio with hedge funds outperformed the 60/40 Pension Plan Index by 48 basis points, but volatility increased by 14 basis points. Similar results are found in other studies. Lamm (1999) suggests that hedge funds may act as a cash or fixed income substitute for a diversified portfolio because of their low volatility and high absolute return. Purcell and Crowley (1999) find that the inclusion of hedge funds in a diversified portfolio can increase the expected return by as much as 200 basis points. Edwards and Liew (1999) find that an unconstrained optimization including stocks, bonds and a hedge fund of funds selects an allocation of 84% to the hedge fund of funds, 7% to the S&P 500 and 10% to long term bonds. In summary, hedge funds have consistently demonstrated that ability to add value to a traditional portfolio of stocks and bonds. With respect to hedge fund indexes, they serve two key purposes. First, they serve as a proxy for the hedge fund asset class. This is important for asset allocation studies. Second, hedge fund indexes can serve as performance benchmarks to judge the success or failure of hedge fund managers. However, as we demonstrate below, there are many differences among the several hedge fund index products offered. We begin with a discussion on index construction. Issues with Index Construction 2

3 In this paper we provide information on 10 hedge fund indexes. Each index is based on a different number of hedge funds, ranging from 60 to over 2,000. Most of these indexes use simple averages, while others use capital-weighted indexes. Also, some index providers collect the underlying data themselves, while others allow the hedge fund managers to enter the data. Still other hedge fund indexes include managed futures while some do not. In sum, there are many different construction techniques of hedge fund indexes. We discuss the challenges of implementing these methodologies below. The Size of the Hedge Fund Universe One of the problems with constructing a hedge fund index is that the size of the total universe of hedge funds is not known with certainty. Depending on which report you choose, there are 5000, 6000, or 7000 hedge funds in existence with assets ranging from $500 billion to $1 trillion. 3 The uncertainty regarding the true size of the hedge fund industry stems from the fact that it is an unregulated industry. 4 Hedge funds are not required to register with the Securities and Exchange Commission and are not required to report or publish their performance data. 5 Consequently, hedge funds have no need to report their performance to an index provider. They can enjoy relative secrecy. This is in contrast to their mutual fund counterparts. Mutual funds are regulated investment companies that are required to register with the SEC. In addition, investment advisers to mutual funds are also required to register with the SEC. In fact, mutual funds are considered public investment companies that issue public securities (mutual fund shares) on a continual basis. Therefore, they are required by law to report and publish their performance numbers to the SEC and the public. A good example of the unknown size of the hedge fund universe was demonstrated by Liang (2001). He studied the composition of indexes constructed by two well-known providers: TASS and Hedge Fund Research Inc. 6 At the time of his study there were 1,162 hedge funds in the HFR index and 1,627 hedge funds in the TASS index. He found that only 465 hedge funds were 3 See Daniel Collins, Alternative Vehicles Open to Retail, Futures, December 1, 2002; Robert Clow, Investors Pile into Alternative Fund Strategies, Financial Times, April 30, 2002; Robert Clow, Hedge Fund Data are no More than a Rough Guide, Financial Times, March 14, 2002; and Chipping Away at Hedge Funds, The Boston Business Journal, December 13, This unregulated nature, at least in the United States, may soon end. In a recent statement, SEC Commissioner Roel Campos stated that several regulations are being considered including registration of hedge funds and hedge fund managers, as well as restrictions on short -selling and hedge funds ability to leverage their portfolios. 5 The one exception is managed futures funds. This is a sub-class of the hedge fund industry managed by commodity trading advisors and commodity pool operations. CTAs and CPOs must register their hedge funds with the Commodity Futures Trading Commission and publish the performance results. 6 The Tass database is now used by the joint venture of CSFB/Tremont Advisors. 3

4 common to both hedge fund indexes. Further, of these 465 common hedge funds, only 154 had data covering the same time period. Another problem with measuring the size of the hedge fund universe is that the attrition rate for hedge funds is quite high. Park, Brown, and Goetzmann (1999) and Brown, Goetzmann and Ibbotson (1999) find that the average life of a hedge fund manager is 2.5 to 3 years. The short half-life of the average hedge fund means that there will be considerable turnover on an annual basis with respect to hedge fund index construction. In conclusion, the hedge fund universe is not known with certainty and very little overlap exists between hedge fund index providers. Data Biases There are several data biases associated with hedge fund indexes. The first, is the well-known survivorship bias. Survivorship bias arises when constructing a hedge fund index today based on hedge fund managers that have survived the time period of study and are available for index construction. Those hedge fund managers that have not survived are excluded from the index construction. This can bias the performance of an index of hedge funds upwards, because presumably, the remaining hedge funds survived as a result of their superior performance. This bias is also common with mutual fund studies. However, the lack of a regulatory environment for hedge funds creates the opportunity for other data biases that are unique to the hedge fund industry. In addition to survivorship bias, there are three other biases that may affect hedge fund index construction. First, there is selection bias. Essentially, because they are unregulated, hedge fund managers have a free option to report their data. They can pick and choose when and with whom to report their data. Selection bias also pushes hedge fund index returns upward because it is the better performing hedge fund manager who will choose to exercise their option and report their performance to an index provider. 7 Closely related to selection bias is instant history or backfill bias. Instant history bias occurs because once a hedge fund manager begins to report his performance to an index provider, the index provider backfills the hedge fund manager s historical performance into the database. Again, because it is more likely that a hedge fund manager will begin reporting his performance after a period of good performance, this bias pushes index returns upward. Last, there is liquidation bias. Frequently, hedge fund managers go out of business or shut down an unsuccessful hedge fund. When this happens, these managers stop reporting their performance in advance of the cessation of operations. In other words, several months of poor performance is lost because 7 A contrary argument can be made for selection bias: that is, good hedge fund managers choose not to report their data to hedge fund index providers because they have no need to attract additional assets. 4

5 hedge fund managers are more concerned with winding down their operations than they are in reporting their performance to an index provider. 8 In total, these biases can add up to 3% to 4% of annual performance enhancement to a hedge fund index. It is important to take note of these biases because they cannot be diversified away by constructing a portfolio of indexes because all indexes suffer from these biases (see Fung and Hsieh, 2000). Exhibit 2 details the size of these biases from several recent studies on hedge funds. Strategy Definition and Style Drift Strategy definitions can be very difficult for index providers. An index must have enough strategies to capture the broad market for hedge fund returns. Index providers determine their own hedge fund strategy classification system, and this varies from index to index. Consider a hedge fund manager that goes long the stock of a target company subject to a merger bid and short the stock of the acquiring company. The strategy of this hedge fund manager may be classified alternatively as merger arbitrage by one index provider (e.g., HFR), relative value by another index provider (e.g. MSCI), or event driven by still another index provider (e.g. CSFB/Tremont). Further complicating strategy definition is that some hedge fund managers may be hard to classify. Most hedge fund managers are classified according to the disclosure language in their offering documents. However, consider the following language from an actual hedge fund private placement memorandum: Consistent with the General Partner s opportunistic approach, there are no fixed limitations as to specific asset classes invested in by the Partnership. The Partnership is not limited with respect to the types of investment strategies it may employ or the markets or instruments in which it may invest. Where to classify this manager? Relative Value? Global Macro? Possibly, Diversified. Unfortunately, with hedge funds this type of strategy description is commonplace. The lack of specificity may lead to guesswork on the part of index providers with respect to the manager s strategy. Alternatively, some index providers may leave out this manager because of lack of clarity (e.g. Zurich Hedge Fund Indices), but this adds another bias to the index by purposely excluding certain types of hedge fund managers. In sum, there is no established 8 The flip-side to liquidation bias is participation bias. This bias may occur for a successful hedge fund manager who closes his fund and stops reporting his results because he no longer needs to attract new capital. 5

6 format for classifying hedge funds. Each index provider develops its own scheme without concern for consistency with other hedge fund index providers. Even if an index provider can successfully classify a hedge fund manager s investment strategy, there is the additional problem of strategy drift. Again, because of the unregulated nature of hedge fund managers, there is no requirement for a hedge fund manager to notify an index provider when his investment style has changed. Let s continue with our example of merger arbitrage managers. With the sluggish US economy, the market for mergers has declined significantly over the past two years. There are simply too few deals to feed all of the merger arbitrage manager mouths. Consequently, many of these managers changed their investment style to invest in the rising tide of distressed debt deals which are counter-cyclical from mergers and acquisitions, or expanded their investment portfolio to consider other corporate transactions such as spin-offs and recapitalizations. Unfortunately, all too often, once a hedge fund manager has been classified as merger arbitrage it will remain in that category despite significant changes in its investment focus. Investability A key issue is whether a hedge fund index can be or should be investable. This is an issue for hedge funds that is distinct and different from their mutual fund counterparts. Mutual funds are public companies. They can and do continually offer their shares to the public. Capacity issues are virtually non-existent. However, hedge funds generally do have capacity issues as certain strategies only work well within certain limits of investment capital. This means that hedge fund managers often refuse further capital when they have achieved a maximum level of assets under management. Consequently, it is very difficult for hedge fund indexes to remain investable when the underlying hedge funds close their doors to new investors. A related issue is whether hedge fund indexes should be investable. The argument is that an investable index will exclude hedge fund managers that are closed to new investors and therefore, exclude a large section of the hedge fund universe. Most index providers argue that to be a truly representative index that acts as a barometer for hedge fund performance, both open and closed funds should be included. The trade-off, therefore, is between having as broad a representation as possible of hedge fund performance versus having a smaller pool of hedge fund managers that represent the performance that may be accessed through investment. Hedge Fund Indexes 6

7 In this section we provide summary information on ten hedge fund index providers. These indexes vary as to number of hedge fund managers, types of strategies employed, and investability. Exhibit 3 summarizes the key attributes of the hedge fund indexes. We discuss some of their differences below. Fees All of the hedge fund indexes listed in Exhibit 3 calculate hedge fund performance net of fees. However, there are two issues related to fees that can result in difference performance than portrayed by a hedge fund index. First, incentive fees, that portion of hedge fund remuneration that is tied to the performance of hedge funds, are normally calculated on a quarterly or annual basis. However, all of these indexes provide month by month performance. Therefore, on a monthly basis, incentive fees must be estimated and subtracted from performance. The actual fees collected at quarter or year end may be very different than the monthly estimates. Second, hedge funds are a form of private investing. Indeed, virtually all hedge funds are structured as private limited partnerships. As a consequence, the terms of specific investments in hedge funds often are not negotiated in a consistent manner among different investors or across different time periods. The lack of consistency means that the net-of-fees returns earned by one investor may not be what another investor can negotiate. In fact, the more successful the hedge fund manager, the greater the likelihood that he will increase his fee structure to take advantage of his success. The end result is that index returns may overstate what a new investor can obtain in the hedge fund market place. At least one index provider has offered an investment product tied to the performance of the index. Credit Suisse First Boston (CSFB) in conjunction with the Tremont hedge fund index has offered an investable CSFB/Tremont product tied to the total return of the Tremont hedge fund composite. When first introduced, this product was initially offered for a fee of 1%, but competition may soon lower this fee. Turnover Most of the turnover with respect to hedge fund indexes tends to be one-sided. That is, the index composite grows as more hedge funds report their performance to the index provider. However, some hedge funds go out of business or close their fund to new investors and cease reporting their returns. This can lead to several of the data biases presented in Exhibit 2. In sum, turnover tends to be low, with more hedge fund returns added to the composite over time. Performance 7

8 Exhibit 4 demonstrates the historical performance of the 10 indexes. 9 The most striking observation is that the risk/return performance of the 10 indexes varies significantly. The highest return is associated with the Tuna funds (average annual return of 16.35%) and the lowest associated with MSCI (7.62%). Also the standard deviation of annual returns ranges from 14% (Van Hedge) to 3% for the S&P Hedge Fund index. We also include as an additional reference, the risk and return of the S&P 500, the Russell 1000 and 2000 stock indexes, 10 year US Treasury Bonds, and the Barclays CTA Index (managed futures index). 10 We can see that US Treasury Bonds and managed futures offer about the same risk and return relationship, modest returns but with low volatility. We include Treasury bonds to provide a low risk/low return alternative to stocks, and we include the Barclays CTA Index because some of the hedge fund indexes include managed futures, while others exclude this hedge fund subclass. All three of the stock indexes appear as outliers on this chart with average returns of 9% to 11%, but with volatility significantly higher than that for the hedge fund indexes, at 19% to 20%. Exhibit 4 underscores our earlier comments regarding the diversity of index construction and the fact that the size of the hedge fund universe is not known with certainty. Further, the wide range of historical risk/return performance carries over to the hedge fund sub-indexes. In Exhibit 5, we present the historical risk/return profile for equity long/short indexes. 11 If anything, there is even more variability. All of this means that when choosing a hedge fund composite index or sub-index, an investor must use care to ensure that the chosen index is representative of her hedge fund investment program. For example using the Zurich long/short equity hedge index to measure the performance of a program that resembles more the economic parameters of the MSCI long/short equity hedge index could lead to inaccurate conclusions regarding the performance of the program. Correlation across Hedge Fund Indexes and Stock Indexes Exhibit 6 presents a table of correlation measures between the hedge fund indexes and the stock indexes. We also include the Barclays CTA index in the table. We omit the S&P Hedge Fund Index and the MSCI Hedge Fund Index 9 In the case of Zurich Capital, we used the average performance across its five sub-indexes, and for MAR/CISDM we used the performance of its Fund of Funds index. Data for S&P and MSCI are based on pro-forma performance prior to The Barclays CTA Index is a composite of all Commodity Trading Advisors in the Barclays CTA database that have at least 4 years of performance data. About 300 CTA managers are included in the index and the index is equally weighted. CTA trading styles are a subset of the hedge fund market place and are often referred to as Managed Futures. 11 Equity long/short hedge fund sub-indexes are also referred to as equity hedge (HFR and EACM), long/short hedged (Tuna funds), and market neutral long/short (MAR and Van Hedge). 8

9 from this analysis because of their shorter, pro-forma track records. We also include the Barclays CTA index to represent the managed futures asset class that is excluded from half of the hedge fund indexes. The variability of historical risk/return profiles is demonstrated in the correlation coefficients between the hedge fund indexes. The coefficients range from a high of 0.98 (Tuna funds/hfr) to a low of 0.67 (Tremont/Zurich). Most of the correlation coefficients are in the range of 0.8 to 0.9. Compared to equity stock indexes, these correlations are low; the correlation between the S&P 500 and the Russell 1000 stock indexes is 0.99, although the correlation between the S&P 500 and the Russell 2000 is only about This simply underscores the fact that the hedge fund known universe is not known with certainty and while certain indexes may capture similar parts of the universe, there is still a wider variation among hedge fund index returns than among equity index returns. It is worthwhile to note that the Barclays CTA composite index is negatively correlated with both hedge fund and stock indexes. With respect to hedge fund indexes, a large component of these strategies include convergent, or arbitrage, type managers. That is, many hedge fund managers engage in arbitrage trades where they expect the prices of two securities to converge over time. These type of strategies are known as short volatility or convergent trading (see Anson, (2002), Lo, (2001), and Weisman (2002)). However, managed futures strategies tend to be long volatility trades, or divergent trading (see Fung and Hsieh, 1997). As a result, managed futures makes a good diversifying agent for other hedge fund styles. Last, it is interesting to note that the hedge fund indexes are much more highly correlated with small cap stocks (represented by the Russell 2000) than with large cap (S&P 500) or mid cap stocks (Russell 1000). 12 We offer two suggestions for this observation. First, the small cap stock market is generally considered to be less efficient than large or mid cap stocks. To the extent that this market is driven by the similar pricing anomalies that drive hedge fund investment returns, a higher correlation of returns would be expected. Second, hedge fund managers may concentrate is small cap stocks as a way to avoid the phenomenon of crowded shorts. That is, hedge fund managers prefer to short stocks that do not already have a large short interest typically associated with larger capitalized companies. Asset versus Equal Weighted An asset weighted index is susceptible to disproportionate representation from large funds that have a very large gain or loss in any given time period. Additionally, an asset-weighted index can be distorted by errors in reporting by larger funds. Further, some of the largest funds choose not to report their data to 12 I am indebted to the referee, Matt Moran of the Chicago Board Options Exchange for pointing this out to me. 9

10 public databases, and it may be difficult to interpret an asset weighted index return that does not include some of the larger hedge funds. Equal weighting has the advantage of not favoring large funds or hedge fund strategies that attract a lot of capital (like global macro or relative value). Investors may be prone to chasing either returns or the latest hedge fund flavor of the year. This can distort a market capitalization index because the returns of a market-cap index will be influenced by the flows of capital. Most hedge fund index providers argue that a hedge fund index should be equally-weighted to reflect fully all strategies. Yet there are two worthwhile arguments for an asset-weighted hedge fund index. First, smaller hedge funds can transact with a smaller market impact. An asset weighted index would more accurately reflect the full market impact from the hedge fund universe as it conducts its transactions. This is all the more important for hedge fund managers because of the nature of the high portfolio turnover associated with their frequent and opportunistic trading patterns. Second, many other asset classes are benchmarked against capital weighted indexes. The S&P 500 and the Russell 1000, for example, are cap-weighted equity indexes. This is important because large institutional investors use these cap-weighted indexes in their asset allocation decision models. Therefore, to compare on an apples -to-apples basis, hedge fund indexes should also be capweighted when used for asset allocation decisions. Index Diversification The size of the ten hedge fund indexes varies from 60 funds to over Most index providers have a single composite index with the exception of MAR and Zurich Capital. However, MAR provides a Fund of Funds median index that acts as a proxy for its total universe, while Zurich Capital does not. Each index provider constructs several sub-indexes so that the performance of specific hedge fund strategies can be tracked more closely. But what is the right size of an index? For instance, does 60 funds offer sufficient diversification such that the idiosyncratic risk of individual managers are diversified away? Two studies have examined the issue of the proper diversification for hedge funds. Henker (1999) finds that the majority of idiosyncratic risk associated with equity long/short hedge funds can be diversified away with as little as 10 funds, while most of the risk is diversified away with about 20 funds. Similarly, Park and Staum (1999) find that fund of funds about 95% of hedge fund idiosyncratic risk can be diversified away with 20 hedge funds. Therefore, each hedge fund index listed in Exhibit 2 should provide a well-diversified benchmark of hedge fund performance. 10

11 However, another question that should be asked is how many hedge funds are necessary in an investment program to produce a correlation with a chosen hedge fund index that is sufficiently high? This is important since hedge fund indexes may be used for asset allocation purposes and the resulting hedge fund investment program should meet the expectations of the asset allocation study. Lhabitant and Learned (2002) examine several hedge fund strategies, and find that an investment program of 20 hedge funds captures 80% to 90% of the correlation with the chosen hedge fund index. Summary Benchmarks serve two useful purposes. First, they provide a yardstick for measuring performance of an asset class or an individual external manager. Second, they can be used in asset allocation studies to determine how much to allocate among broad asset classes. Benchmarks are tools for building and monitoring portfolios. With respect to hedge funds, investors have a wide variety to choose from. Unfortunately, there is a lack of consistency in the construction of hedge fund indexes. This lack of consistency was demonstrated in the wide range of risk and return measures in Exhibits 4 and 5 as well as the wide range of correlation coefficients in Exhibit 6. This creates two distinct problems. First, given the large range of performance among the hedge fund indexes, an investment manager who invests in hedge funds can significantly out-perform or under-perform her bogey by the choice of hedge fund index. Second, asset allocation studies that are driven by the risk/return trade-off of different asset classes may over or under allocate to hedge fund investments based on the simple choice of hedge fund index. Some variability among hedge fund indexes is good, but too much can result in misleading asset allocation decisions. Last, all of these indexes suffer from several data biases that can boost returns by 3% to 4%. In summary, the world of hedge fund performance measurement is still maturing. Currently, there are many indexes to choose from, each with its own pros and cons. Also, the consistency among hedge fund indexes is considerably less than that than for equity indexes. Perhaps the best way to choose a hedge fund index is to first state clearly the risk and return objectives of the hedge fund investment program. With this as their guide, investors can then make an informed benchmark selection. Bibliography 11

12 Carl Ackermann, Richard McEnally, and David Ravenscraft. The Performance of Hedge Funds: Risk, Return, and Incentives. The Journal of Finance, June Mark Anson. Should Hedge Funds be Institutionalized? The Journal of Investing, Fall Mark Anson. Symmetrical Performance Measures and Asymmetrical Trading Strategies: A Cautionary Example. Journal of Alternative Investments, Ross Barry. Hedge Funds: A Walk Through the Graveyard. Working paper, Applied Finance Center, MacQuarie University, March Stephen Brown, William Goetzmann and Roger Ibbotson. Offshore Hedge Funds: Survival and Performance, The Journal of Business, vol. 72, Franklin Edwards and Jimmy Liew. Hedge Funds versus Managed Futures as Asset Classes. The Journal of Derivatives, Summer William Fung and David Hsieh. Performance characteristics of Hedge Funds and Commodity Funds: Natural versus Spurious Biases. The Journal of Financial and Quantitative Analysis, vol. 35, Empirical characteristics of Dynamic Trading Strategies: The Case of Hedge Funds. Review of Financial Studies, October Goldman Sachs & Co. and Financial Risk Management Ltd. The Hedge Fund Industry and Absolute Return Funds, The Journal of Alternative Investments, Spring Hedge Funds Revisited, Pension and Endowment Forum, January Thomas Henker. Naïve Diversification for Hedge Funds. The Journal of Alternative Assets, Winter R. McFall Lamm, Jr. Portfolios of Alternative Assets: Why Not 100% Hedge Funds? The Journal of Investing, Winter 1999, pp Francois -Serge Lhabitant and Michelle Learned. Hedge Fund Diversification: How Much is Enough? The Journal of Alternative Investments, Winter Bing Liang. Hedge Funds: The Living and the Dead. The Journal of Financial and Quantitative Analysis,

13 Andrew Lo. Risk Management for Hedge Funds: Introduction and Overview. Financial Analysts Journal, Nov./Dec James Park and Jeremy Staum. Fund of Funds Diversification: How Much is Enough? The Journal of Alternative Assets, Winter James Park, Stephen Brown and William Goetzmann. Performance Benchmarks and Survivorship Bias for Hedge Fund and Commodity Trading Advisors. Hedge Fund News, August David Purcell and Paul Crowley. The Reality of Hedge Funds. The Journal of Investing, Fall 1999, pp Andrew Weisman, Dangerous Attractions: Informationless Investing and Hedge Fund Performance Measurement Bias, The Journal of Portfolio Management,