International Business & Economics Research Journal – November 2009 Volume 8, Number 11 95 New Evidence On Hedge Fund Performance Measures Komlan Sedzro, University of Quebec in Montreal, Canada ABSTRACT Hedge funds are still relatively unfamiliar to most investors despite the intense popularity they have enjoyed in recent years. Measuring the performance of these financial instruments using traditional methods is, however, problematic, since their returns do not follow a normal distribution. In this study, we consider rankings obtained with the Stochastic Dominance (SD) method and compare them with ranks produced using Sharpe Ratios, Modified Sharpe Ratios, and Data Envelopment Analysis. We also explore the advantages highlighted by the literature of the Data Envelopment Analysis (DEA) method in relation to traditional measures like Sharpe ratio and Modified Sharpe ratio. Our results show that classic performance measures are better correlated with SD than DEA results. Keywords: Data Envelopment Analysis, Performance Evaluation, Stochastic Dominance, Hedge Funds. 1. INTRODUCTION ollowing the stock market turbulence of the recent years, alternative investments have become the latest trend among traders. The number of hedge funds (HF) and the value of assets under HF management rose from 600 funds managing $38 billion in 1990 to more than 10,000 funds valuing more than $1,500 billion in 2006 (Crockett, 2007), representing growth of over 1000% in 16 years. The size of the HF industry justifies this study, as well as the differences in the results obtained using traditional measures to evaluate the performance of this asset class. Indeed, many studies have shown that HFs outperform both mutual funds and traditional investments, based on returns or risk-adjusted returns (Cottier 2000; Liang 1999; Caglayan and Edwards 2001; and Agarwal and Naik 2000a). However, some authors (Ackermann, McEnally, and Ravenscraft, 1999, Brown, Goetzmann, and Ibbotson 1999) report a less conclusive performance for hedge funds. These diverging observations could be because the results are based on different performance measures. Indeed, financial literature proposes several tools to measure the performance of mutual or hedge funds. To start with, there are the traditional measures, like the Sharpe, Jensen or Treynor ratios (Elton and Gruber, 1995; Sharpe, 1966). Measures like the Modified Sharpe ratio suggested by Favre and Galeano (2002) that take into account skewness in return distributions also exist but are more complex. Moreover, Gregoriou, Sedzro, and Zhu (2005) have recently proposed Data Envelopment Analysis (DEA), a non-parametric method often used to assess the efficiency of public sector decision-making units, as a useful measure for evaluating hedge fund performance. Another more general measure is the stochastic dominance (SD) used by Wong, Phoon, and Lean (2008) to compare the performance of Asian hedge funds. Although involving time consuming pairwise comparison between any pair of hedge funds, SD is a more powerful performance measure because it incorporates information on the entire return distribution (Post, 2003; Davidson and Duclos, 2000). Therefore, the objective of this study is to test the effectiveness of traditional measures and the DEA scores to assess hedge fund performance using stochastic dominance as benchmark. More specifically, we will compare hedge fund performance using the Sharpe ratio, Modified Sharpe and DEA, with the rankings obtained using stochastic dominance. To our knowledge, this question has not been raised in previous studies. F