Persistent Doubt: An Examination of Hedge Fund Performance

• Date: 01 September 2016
• Published date: 01 September 2016
• DOI: http://doi.org/10.1111/eufm.12070

Persistent Doubt: An Examination of

Hedge Fund Performance

Mar

ıa de la O. Gonz

alez

School of Economic& Business Sciences, University of Castilla–LaMancha, Plaza de la Universidad 1,

02071, Albacete, Spain

E-mail: mariao.gonzalez@uclm.es

Nicolas A. Papageorgiou

HEC Montreal, 3000 Cote Ste Catherine, Montreal, H3T2A7, Canada

E-mail: nicolas.papageorgiou@hec.ca

Frank S. Skinner

Department of Economics and Finance, Brunel University, Uxbridge, London, UB8 3PH, United

Kingdom

E-mail: frank.skinner@brunel.ac.uk

Abstract

We examine whether performance persistence is suspicious. Top quintile portfolios

formed on the Sharpe ratio, alpha, and information ratio persistently outperform

similarly constructed mediocre third quintile portfolios throughout our sample

period, but performance is more modest and less persistent when portfolios are

formed on the excess manipulation-proof performance measure (EMPPM). By

selecting funds formed on ranking by Sharpe and information ratios, investors also

select funds that have persistently doubtful performance according to the doubt

ratio. In contrast, portfolios formed on alphas and especially the EMPPM have

much less excess and persistent doubt.

Keywords: hedge funds, performance measures, manipulation-proof performance

measure, doubt ratio

JEL classification: G1, G11, G23, G24

We thank the Editor (John Doukas) and an anonymous referee for excellent comments and

suggestions. We are also grateful to Ephraim Clark, Richard McGee, Frank McGroarty,

Thomas and Julia Henker, the seminar participants of the Adam Smith Seminar Series, and

the conference participants of Infinity and FEBS for their helpful comments. We use hedge

fund risk factors from David A. Hsieh’s data library and gratefully acknowledge David A.

Hsieh for making these data publically available. This work was supported by Junta de

Comunidades de Castilla–La Mancha and Ministerio de Econom

ıa y Competitividad (grant

numbers PEII-2014-019-P and ECO2014-59664-P, respectively). Any errors are the

responsibility of the authors.

European Financial Management, Vol. 22, No. 4, 2016, 613–639

doi: 10.1111/eufm.12070

© 2015 John Wiley & Sons, Ltd.

1. Introduction

Hedge fund performance and performance persistence are a controversial issue. Many

authors, such as Ammann et al. (2013) and Boyson (2008), find that hedge funds

persistently deliver positive alphas out of sample. Other research is more critical,

claiming that performance persistence depends upon the measure used (Capocci et al.,

2005), that performance persistence is attributable to passive investments in illiquid

assets (Brandon and Wang, 2013), and even that performance persistence for top funds

does not exist at all (Slavutskaya, 2013). Recently, Dichev and Yu (2011) document a

sharp reduction in buy and hold returns for a very large sample of hedge funds and

commodity trading advisors, from 18.7% from 1980 to 1994 to 9.5% from 1995 to 2008.

Clearly, there are doubts concerning the performance, performance persistence, and

measurement of the performance of hedge funds.

Meanwhile, Goetzmann et al. (2007) demonstrate that common performance

measures such as the Sharpe ratio, alpha, and information ratio can be subject to

manipulation, deliberate or otherwise. To address this issue, the authors develop a

manipulation-proof performance measure (MPPM), so called because it is robust to the

underlying return distribution and to gaming by hedge fund managers. Additionally,

Brown et al. (2010) develop the doubt ratio. This measure is a diagnostic statistic derived

from the MPPM and indicates whether the reported returns from hedge funds are

suspicious. Together, the MPPM and the doubt ratio allow us to examine recent hedge

fund performance and shed light on whether top hedge funds achieve superior

performance, whether this performance persists, whether performance persistence is

doubtful, whether doubts concerning hedge fund performance persist, and whether

performance persistence is down to skill, manipulation or luck.

We test for persistence in hedge fund performance by forming portfolios according to

five different performance measures that are further stratified by quintile and strategy.

The five measures are the Sharpe ratio; the alpha from Fung and Hsieh’s (2004) hedge

fund asset pricing model augmented by an eighth factor for emerging markets; the

information ratio, obtained as the t-statistic from Fung and Hsieh’s alphas; the MPPM of

Goetzmann et al. (2007); and the doubt ratio of Brown et al. (2010). The Sharpe ratio,

information ratio, and alpha enable us to compare our results with the recent literature

and allow us to highlight how the picture changes when we evaluate the performance and

performance persistence of top quintile hedge funds with the MPPM and doubt ratio.

Each portfolio is formed monthly, between January 31, 2001, and December 31, 2010,

using the prior month’s performance and performance is tracked fortwo years. We thus form

120 overlapping out-of-sample portfolios for each strategy and quintile, each of which is

held for 2 years. This ensures that we only use information available on the date that the

portfolios are formed so that our subsequent examination of monthly out-of-sample

performance mimics the investment activities of portfolio managers. We then test for

superior performance using bootstrap t-tests.

This testing strategy allows us to address several questions. Do hedge funds deliver

consistent superior performance, on average, throughout the sample period and in the

bull and bear markets around the December 31, 2006, pivot date? Can top-performing

hedge funds persist in delivering superior performance? If so, for how long will this

superior performance persist? Are the answers to these questions sensitive to the choice

of performance measure? Specifically, is there any evidence that the hedge fund industry

manipulates common performance measures? Is the newly developed MPPM resilient to

© 2015 John Wiley & Sons, Ltd.

614 Mar

ıa de la O. Gonz

alez, Nicolas A. Papageorgiou and Frank S. Skinner