Look at return patterns and how returns are generated.
It is estimated that there are now close to 7,0001 hedge funds managing close to $1 trillion. We estimate that the number of firms is growing almost 30% annually in the United States and the assets they manage are growing more than 40% annually. Similar to concerns about the myriad of mutual fund choices, investors face a similar dilemma about how to choose a successful hedge fund. Unlike mutual funds, which are regulated and must meet stringent reporting and valuation criteria, most hedge funds are not registered, causing confusion about performance, fees, holdings and risk-taking. Thus, selecting a hedge fund can be a daunting task, but not an impossible one for the educated investor.2
As in the selection of any investment, the investor should understand the nature of the investment including expectations of returns and risk. In this article we will focus on various important aspects of hedge fund performance. We will point out the danger in relying on performance as a selection criteria, noting the wide gap in performance among different hedge fund strategies as well as the surprisingly large performance gap within the same strategy during the same time periods. As part of the performance discussion, we will discuss the biases that exist in databases, as well as the impact that fees and expenses can have upon reported returns.
A subsequent article will review the other crucial component of any investment choice-risk. The focus of the article will highlight why traditional and reported measures of risk may not be appropriate when considering a hedge fund investment.
Returns-What Are People Buying?
The term hedge fund is loosely defined and covers a number of different strategies. The performance objective of a hedge fund manager is to produce a targeted return or absolute performance, regardless of the underlying trends in the financial markets. Hedge funds use a variety of tools to implement their strategies, including the use of short selling, derivatives and leverage. Differences in implementation and the varied use of financial instruments lead to a wide difference in performance among hedge funds. Importantly, there are biases in performance databases that can be misinterpreted, leading to incorrect and potentially damaging decisions about which strategy to use or even if investing in hedge funds is appropriate.
Performance Data Biases
Databases may be biased because the data upon which they rely is supplied by the funds themselves. While database providers check the data for reasonableness, the performance is not audited. We make reference to close to 7,000 hedge funds in the marketplace, yet the largest databases consist of only about half that number. While we can speculate why funds choose not to submit their data, we do believe that the databases are a representative universe.
An important result of any fund's discretion to participate in the databases is survivorship bias. If a firm closes its strategy because it is successful or disappears because it has failed, the performance numbers disappear. Also, many new hedge funds do not report their performance data until their success is proven, at which point these firms may "backfill" the data for the past years of performance-causing history to change. There are also those hedge funds that never report results and eventually fail. Those performance numbers were never reported and hence were not captured in any database. Academic studies suggest that survivorship and backfill biases could reduce the reported mean returns 1.4% to 4% annually3. In recent years it also has been estimated that there may be up to 600 hedge funds annually going out of business. Out of a universe of 7,000 firms, that is a significant portion that may not be reporting results. Hence, survivorship bias and lack of reporting standards are significant issues.
In the last and final caveat emptor, the average statistics, which are calculated from databases, may be additionally biased if these databases calculate averages based on the size of the firm's assets under management. This is another example of survivorship bias, which skews the data by weighting the most successful firms (larger asset size) more heavily. Most practitioners compensate for this bias by using databases that show results equally weighted across all funds so that large funds do not dominate results. While not as serious as capital-weighting the databases, another bias may creep in as managers submit performance data on identical funds. Specifically, there may be two or more funds managed by the same firm in the same strategy with the same or nearly the same returns, thereby skewing results towards that series of funds. Identical, or clone, funds are more likely to be established by more successful fund groups. When looking at overall information in peer comparisons, ask whether or not the peer universe is equal- or asset-weighted, and whether or not cloned funds are prevalent in the data.
Aside from the performance data itself, one should be aware of the impact of fees. During a period of strong performance, fees may not appear to be a significant issue. However, during a period of mediocre performance, fees may tip a strategy into posting a loss. In a traditional investment approach, negative returns do occur from time to time without too much damage to the underlying business of the investment firm. However in a hedge fund, where performance fees are typical, a negative after-fee return may cause severe hardship to the business. Therefore, understanding the fee structure of any fund strategy is important. High-maintenance funds, in particular, may be more vulnerable during periods of subpar performance.
The overall conclusion is that measuring hedge fund performance from databases is skewed towards success and may overstate the average results by a fair amount. The degree of skew is debatable and a subject of intense academic study. In addition, investors should insist on looking at performance that is reported net of fees and expenses in order to get a true picture of the returns that the manager has delivered to clients.
What Does Published Performances Mean?
Given all of the quality issues highlighted above, why use databases at all? The reason is that they are a good starting point from which to eliminate strategies and/or firms that do not fit other criteria. Databases provide the basis for extensive risk analysis and to establish benchmarks, which can be useful in ongoing stewardship. The performance results used in this article have been "scrubbed" so that all results are net of fees and the averages are equally weighted, eliminating the large-firm bias. The clone strategies and fund of funds data have been removed. Also, each fund's return history has been inspected and obvious errors corrected or the fund has been removed from the data set. In addition, some funds report only on a quarterly basis, but data has been shown as monthly data. The quarterly returns have been eliminated or the funds have been removed. The resulting database we used for analysis contains 1,262 funds available to U.S. tax-paying resident investors. Eliminating the other biases discussed previously takes significant due diligence and manager review-the qualitative assessment that is key to selecting any investment advisor.