Passive Hedge Funds

H/T Matt Levine

Mikhail Tupitsyn

Monash Business School

Paul Lajbcygier

Monash University – Department of Banking & Finance

August 10, 2015

Abstract:

We show that most hedge fund managers are passive, not active. Active management should be manifest through nonlinear exposure to the systematic risk factors that drive hedge fund returns. In order to demonstrate managerial skill enhanced performance should accrue as a consequence of active management. Using generalized additive models we find that approximately two-thirds of hedge funds exhibit only linear factor exposures and hence are “passive”. What’s more such “passive” managers tend to outperform “active” managers. Finally, we also show that many “active” managers, despite initial nonlinear risk exposures, eventually become “passive”.

Passive Hedge Funds – Introduction

The question at the heart of this study is simple: do most hedge fund managers generate returns through managerial skill? The answer, according to our work is no. Most hedge fund managers rely on “passive” linear risk exposures to generate their returns and, paradoxically, they outperform most “active” managers which try to deploy skill.

To demonstrate skill, a hedge fund manager must generate enhanced performance through active management. Such skill should be manifest through nonlinear exposure to the systematic risk factors that drive hedge fund returns and as a consequence of active management outperformance should ensue. We ascertain whether nonlinear systematic risk exposures drive hedge fund returns, and if so, what economic outperformance is generated as a consequence.

Prior work has challenged the notion of hedge fund managers possessing skill (Fung and Hsieh 1997; Hasanhodzic and Lo 2007), suggesting instead that they extract a significant part of their returns from “passive” linear systematic risk exposures. Since nonlinear hedge fund returns maybe difficult to measure using standard linear multi-factor models, we use Generalized Additive Models (GAMs) (see Hastie and Tibshirani (1990)) to detect nonlinear systematic risk factors that drive hedge fund returns. We estimate GAMs using the same systematic risk factors (i.e., explanatory variables) as used in a popular linear factor model — the six-factor Hasanhodzic and Lo (2007) model (HL6) — as well as factors derived from Agarwal and Naik (2004) and from Vrontos, Vrontos, and Giamouridis (2008).

By utilizing GAMs to model hedge fund returns we obtain some interesting results: First, we show that when hedge funds are grouped in the same style portfolios, nonlinear risk exposures are more pronounced in styles that focus on exploiting arbitrage opportunities and relative security mispricing, consistent with findings of Mitchell and Pulvino (2001), Fung and Hsieh (2002b), and Agarwal and Naik (2004). Second, we investigate the prevalence of nonlinear patterns in risk exposures of individual funds using GAMs and find that the majority of funds, roughly two-thirds, exhibit only linear exposures, while nonlinear features are present in the exposures of around one-fifth of funds; the rest of the funds have insignificant exposures to any systematic risk factors and are deemed to be market-neutral funds. Third, in order that we may evaluate the impact of nonlinear risk exposures on hedge fund performance, we construct three portfolios of funds for each style: a portfolio of exclusively linear exposed funds; another of nonlinear funds; and, a final portfolio of market-neutral funds. Here, our results suggest that nonlinear funds are (on average) inferior to linear funds in terms of raw and risk-adjusted returns and also have higher negative tail risk. This provides evidence against hedge fund managers’ claims of skill leading to superior returns. Finally, we also analyse the persistence of hedge funds to risk exposures in an attempt to verify whether performance patterns observed among nonlinear, linear, and market-neutral funds can be exploited by investors to generate profits; we find that the majority of nonlinear funds that survive over the long term tend to alter their risk exposures and eventually become linear funds.

In conclusion, consistent with the notion of efficient markets removing abnormal profits, our results suggest that most hedge funds are “passive” and generate returns consistent with linear risk factor exposures. Paradoxically, such “passive” hedge funds outperform “active” hedge funds which have nonlinear factor exposures.