Hedge Fund Replication by TrueBeta
The first fully independent, fully transparent approach
Hedge Fund Replication - an Overview                               info@true-beta.com

Hedge Fund Replication: A Brief Background

There is much confusion about the nature of hedge funds, the meaning of the word “hedge” in the context, the expectation for so-called “absolute returns”, and the notion of alpha.

Aside from their fee structure, and a promise to deliver attractive absolute returns, hedge funds generally have little in common. A few make long-term investments; most buy and sell incessantly. Some trade individual stocks; others place bets on entire industries and markets. Some rely on human intuition to identify plum investments; many use computer software programs to ferret out profitable trades.

It is therefore not immediately obvious how their returns could be replicated.

Academics have been studying the potentially replicable properties of hedge fund returns for well over ten years. William Fung, of the London Business School, and David Hsieh, of Duke, in particular have done pioneering research in the field, as has Henry Kat of the Sir John Cass Business School.

It is not the purpose here to review this research in detail, as it by now is relatively well known. The main conclusion is that the research has clearly shown that hedge fund returns can indeed be replicated in a variety of ways discussed in more detail below.

But let's first note a fundamental point about the nature of hedge fund returns observed by Akihiko Takahashi of the University of Tokyo [Takahashi 2008]:

Hedge fund returns closely match stock returns in up markets but decline less than stock returns in down markets. This in turn leads to higher returns than stocks in the long run.

This then appears to be the fundamental value proposition of hedge funds as an asset class.

The TrueBeta Value Proposition

By implication, hedge fund returns contain a beta component, which allows us to model the risk exposures in a framework of both long and short exposures. This in turn allows us to gain access to those returns through a passive, index-like investment. The challenge is to do it in a way that is rigorous, fully transparent and allows for practical market execution. This sums up the value proposition of TrueBeta.

Alpha vs. Beta

"We define alpha as the part of hedge fund return that cannot be explained by the exposure to systematic risk factors (beta) in the capital markets, and it is thus the return part that stems from the unique ability and skill set of the hedge fund manager" [Jaeger and Wagner, 2005].

Alternative Beta vs. Market Beta

This definition can be further refined, as part of the risk-return profile of hedge funds can be explained by the investment tools and strategies available to hedge fund managers. Already in 1997, Fung and Hsieh extended Sharpe’s factor models to hedge fund styles by the introduction of short-selling, leverage and derivatives.

The generic application of these provide what can be called alternative beta, or strategy beta. In other words, to the extent that a manager's return is due to the systematic application of replicable quantitative techniques, it is not alpha. This can be conceptually summarized as follows:

In practice therefore, we can define alpha as the portion of hedge fund returns that cannot be replicated.

We believe this has fundamental implications for the investment process, and that a replicator can serve as a true benchmark for hedge fund managers.

There is good reason to believe that the average alpha extracted by hedge fund managers is destined to decline,

  • As more money chases a limited number of market inefficiencies, those inefficiencies should decrease or even disappear.

  • As the barrier of entry to the hedge fund industry continues to decline, less skilled managers will enter the market, which will tend to dilute average performance.

Replication Approaches

The replication approaches can be divided into three:

  • Factor-based Approach: Aims to replicate month-to-month hedge fund returns, by first identifying the risk factors that drive hedge fund returns, and then by capturing dynamically changing exposures by rolling regression. This approach is most useful in providing exposure to general hedge fund market returns

  • Pay-Off Distribution. Aims to replicate the risk profiles of hedge funds over the long term, the volatility of the returns, their correlation with the stock market, and the likelihood of suffering extreme losses. The utility of this approach is mainly in a long-term portfolio context, as it does not specifically aim to replicate the returns on a month-by-month basis

  • Mechanical Duplication: Aims to create a “naive" strategy that emulates the underlying investments of a particular hedge fund style. Merger arbitrage for example, can be closely replicated by buying the ten biggest acquisition targets and selling the ten biggest acquirers at any given time

Most currently available replication offerings are based on a factor approach. They differ significantly  however in terms of methodology, and in particular in the level of transparency and use of manager discretion in their investment process.

The TrueBeta proposition offers full transparency to clients, in a 100% rules driven framework (please visit the Methodology section of this web site). We believe this sets TrueBeta apart from other replication offerings available in the market place today.

Hedge Fund Indices and Liquid Hedge Fund Indices

There are several Hedge Fund Index providers, HFR, CS Tremont, and MSCI among them, who calculate strategy and overall hedge fund returns based on data provided to them monthly by a universe of both open and closed hedge funds. As each provider has its own universe, the hedge fund indices are in effect partially overlapping peer-groups. They suffer from a variety of well-known biases, e.g. survivorship bias, backfilling bias, and selection bias, variously estimated to over-state hedge fund returns by 3-4%.

In attempting to replicate hedge fund returns by replicating hedge fund indices, we therefore embark on an inherently imprecise exercise. It is however the only practical approach, which meets a need as demonstrated by the success of the so-called liquid hedge fund indices, which have attracted almost $2bn in assets (before the recent turmoil).

Several Hedge Fund Index providers offer these liquid versions of their broad market indices. They are typically managed account platforms, combining strategies in such a way as to approximate the returns of the broad index. The underlying funds have agreed to provide monthly or in some cases weekly liquidity.

These platforms however necessarily consist of only open funds, which moreover are constrained in their investment opportunities by the requirement to provide monthly or weekly liquidity. Consequently they have often produced unsatisfactory correlations and/or performance compared to their target indices (please also see the Performance section of this web site).