To understand the concept, firstly ‘alpha’ has to be defined. Alpha is a return that is unsystematic and uncorrelated to a general market direction and risk, whereas beta is a systematic and market related return. The most important question is, what is ‘general market’? Do we only talk about overall stock and bond markets or are there ‘alternative betas’ like credit spreads etc. This issue will be addressed further below.
Investors usually want to have both beta and alpha in their portfolios. According to widespread opinion, beta can best be implemented using index products, with active portfolio managers providing the alpha. The ‘ideal’ portfolio therefore would consist of beta index products and alpha managers.
There are three different approaches to transport alpha.
The portable alpha concept tries to enable investors to access more and better alpha sources. There are basically three ways to achieve additional alpha.
The first solution to (portable) alpha is known as ‘alpha (directly) replaces beta (alpha for beta)’:
The approach is to replace a portion of the existing portfolio – usually by selling beta exposure – with a pure alpha investment. The big disadvantage of this approach is that the asset allocation of the portfolio has to be changed, since the alpha investment replaces the beta investment. As such, alpha is not earned on top of but instead of a systematic risk premium. Additionally, many investors or portfolios are not allowed or willing to invest in pure alpha strategies directly because of there high ‘tracking error’ or because they are defined as hedge funds.
The second solution is ‘hedging out (unwanted) beta (beta transfer)’:
In this case managers/funds will be selected based on their ability to generate alpha, without considering the nature of their systematic risk exposure. To avoid potentially unwanted beta risk, it will be swapped out of the portfolio and replaced by the ‘desired beta’ (or alternatively risk free returns). In theory, a diversified hedge fund of funds could constitute the majority of the portfolio. In practice, the portfolio would consist of mostly high tracking error long-only managers.
A key obstacle to this approach is that most investors would have to replace a substantial portion of their existing allocations, causing regulatory problems and/or significant transition costs. To accommodate the desired investment amounts, it often also means using a large number of managers with the ability to produce alpha. Additionally, the implementation of this approach requires significant use of derivatives to hedge out the undesired beta and buy the desired beta.
The third and final solution is the ‘alpha overlay’: In this case the existing portfolio with all beta exposure remains unchanged, and the performance of pure alpha managers is imported via a total return swap. As the short-term interest rate is swapped for the performance of an alpha manager, the investor receives ‘pure’ alpha. By applying this technique, the existing portfolio does not change and only one derivative could be used to provide access to the Alpha. This seems to be the most efficient and promising approach. This approach is not limited to portfolios of institutional investors but investment funds can also be structured in this way.
Most investors want to generate or buy alpha by using active long-only portfolio techniques and managers. But these may only use a limited spectrum of alpha generating instruments, dependant on the opportunity in their respective market segments and investment structures.
Active portfolio management can be most successful in an environment without artificial restrictions. Typical alpha-limiting burdens are restrictions regarding asset class, geographic allocation, short selling, use of derivatives and leverage. Hedge funds, different from traditional funds, typically operate in structures with no or very little external limitation to the implementation of a manager’s strategy, and as such they provide a well-suited structure for alpha generation.
In recent years, a vast amount of studies regarding the merits of hedge funds have been produced. For our purposes, two widely accepted results should be highlighted. First, hedge funds do not generate all of their returns through pure alpha strategies. However, systematic risk can be controlled and reduced (or even neutralised) by a portfolio diversification. Second, along with a lot of often-valid criticism, there is solid evidence that well managed hedge funds do generate alpha over time.
To show the complexity of the analysis to identify pure or true alpha, here are some facts: There are thousands of potential beta factors. At Feri we have identified some two hundred factors that are expected to potentially significantly influence investment fund returns. Our database contains almost twenty thousand so-called hedge fund time series. About six thousand of them can be considered separate single hedge funds, excluding funds of funds and different share classes of the same fund. Out of those, only 1% is currently included in Feri advised hedge fund portfolios. Most of the others have been found to have significant and stable beta exposure. Whereas a beta portfolio which is actively managed and changes market exposures over time may be considered as attractive tactical asset allocation (many global macro funds are essentially beta ‘players’), usually investors should not pay hedge fund fees for stable beta exposures.
Having accepted that portfolios of hedge funds can provide an efficient source of alpha but are usually not consisting of pure alpha, the question about residual beta exposure of such portfolios remains. There are two approaches to such residual systematic exposure. First, residual systematic exposure can be identified and eliminated by the use of index derivatives. This is the approach taken in a recent study by William Fung and David Hsieh, resulting in highly attractive performance contribution despite the loss of the residual beta return. Alternatively, the portfolio of hedge funds can be constructed in such a way that systematic exposure is neutralised in the long term by active diversification. This second approach has the advantage that any short term systematic exposure is considered as an active investment decision by the hedge fund managers, recognizing that such ‘temporary beta’ is an integral part of the alpha generation process.
Another very important feature of hedge funds is the fact that they are usually absolute return vehicles. This can be partly explained by the fact that hedge fund managers typically invest a significant part of their private wealth in their own funds. And these managers usually hate to lose their own money. Also, they often only receive significant fees if they generate absolute returns and not for beating markets whilst still losing money.
As a source of alpha such absolute return funds are more attractive than relative return or benchmark oriented funds, which also produce alpha, since the whole fund return can be used for a portable alpha concept. With relative return funds, the beta or relative return has to be eliminated from the fund first before the remaining alpha can be used for – respectively swapped into – the target portfolio.
There are certain advantages of investable hedge fund indices compared to funds of hedge funds. For example, in order to attract potential alpha from as many sources as possible, different hedge fund strategies should be incorporated. Also, usually different people should manage the different alpha generation strategies.
This leads to a hedge fund of funds as ideal source of alpha. Direct hedge fund of funds investments may be considered as active alpha sources on both levels of the fund of funds and the single fund.
Hedge fund index investments are rather “semi-passive”. But direct hedge fund of funds investments are often not eligible for – or only difficult to integrate into – investor portfolios or existing investment products. One reason for this might be the provisions against cascading fees. Whereas investable hedge fund indices also come with an additional fee level compared to single hedge funds, the ‘derivatives’ reflecting hedge fund index performance are often are easier to buy in different legislations than funds of funds directly.
In addition, single hedge funds and funds of funds need to be IFRS consolidated if institutional investors want to acquire a significant stake of them, which is often assumed to be 20% in a single entity. Derivatives do not necessarily fall under these provisions. On the other hand, most hedge fund indices are either not investable – which means that they can not be easily tracked by a derivative issuer - or are unattractive in terms of real (compared to pro forma) returns and risk.
Since 2003 and mainly in Europe, a lively debate has been going on about hedge fund indices. This discussion was motivated by consultants desire to find reliable time series to benchmark hedge funds against, and by the search for valid and widely accepted measures for general hedge fund performance. Whereas many non-investable hedge fund indices may have all kinds of well analysed and publicised biases regarding fund selection, data backfilling, etc, investable hedge fund indices can comply with the Committee of European Securities Regulators(CESR)guidelines for financial indices in general. CESR even recommends to allow derivatives on commodity and real estate indices to be included in UCITS funds which can be freely distributed Europe wide to retail investors. Investable hedgefund indices may be generally recognised as eligible investments still in 2006 for these retail investment funds.
As of early 2006, broadly diversified investable hedge fund indices which may fulfil the CESR criteria are being calculated and published by CSFB, Hedgefund Research (HFR), Standard & Poors, FTSE, Feri (ARIX index line) and Van Hedge. Dow Jones so far only publishes strategy specific indices. ARIX, the first of these indices, started in early 2002.
Today, evidence suggests that more than €10bn have been collected by so called investable hedge fund index products. After an early success mainly in 2004, net inflows in 2005 slowed down substantially.
Hedge fund indices are usually designed to be broadly diversified and to bear little risk. Also, they are usually more transparent in terms of strategy allocation and fund selection process than funds of hedge funds. But they have been criticised for not reflecting typical hedge fund returns. This is only partly true. Non-investable indices cannot be tracked since they contain too many funds including many funds that do not accept additional money. So the relevant universe for investors consists only of investable indices. They differ in the number of funds, the weighting of funds and strategies and other factors.
Most indices also differ in the number and definition of subindices, some index providers do not even provide composite indices. To generatealpha, the more representative sub-indices are available, the better the theoretical ability to generate alpha through hedge fund strategy allocation.
Werner Goricki, Dirk Soehnholz, Marcus Storr and Vincent Weber are with Feri Institutional Advisors in Bad Homburg, Germany