If you are in the institutional investment business, you have likely heard the name “AP3” before. The SEK $228 billion (US$32 billion) fund announced in May that it would manage its alpha and beta portfolios separately from now on. According to AP3, “(The) fund’s aim by separating alpha and beta is to make the portfolio structure more flexible, improved risk diversification and higher cost efficiency in order to increase the potential return.”
Erik Valtonen, CIO of AP3, will reflect on the fund’s experience later this week at an industry gathering in Hong Kong. But today in this AllAboutAlpha.com exclusive, Valtonen tells us why portable alpha has both “merits and risks”.
Special guest contribution by: Erik Valtonen, Chief Investment Officer, AP3
Portable alpha is certainly a concept that has quickly made an entry to the every investment professional’s vocabulary. Even smaller investment managers are now talking about the merits of porting alpha. In this brief note I make a short summary of the concept, and highlight some of its merits and risks.
Alpha and beta
Return streams are traditionally divided into beta and alpha components, although it seems to be difficult to define exactly what these labels are. Obviously, the origin of the labels lays within the CAPM, where beta is a measure of the exposure to market risk and alpha is the intercept. In this spirit, beta can be defined as the return that is gained by a passive exposure to some risk premium. Traditionally, these exposures include equity and bond markets but recent years have seen the birth (or rather recognition) of a plethora of â€˜alternative betas’ ranging from small cap premium to liquidity. Alpha, on the other, could be defined as that part of the return stream that is not explained by beta . To add positive alpha the manager needs to exhibit skill (or luck) either by securities selection or by timing.
Ability to allocate between different beta factors will produce alpha, so the distinction between alpha and beta is not always clear cut, and will depend on the time horizon where the returns are viewed. As an example, a GTAA manager will typically allocate his risk budget between various beta sources. At any point of time the manager will have different beta-exposures but these are not static but are managed dynamically. The manager’s skill exhibits in his ability to find the right mix of beta exposures.
An important difference between alpha and beta is the cost of producing them. Beta production should be very cheap, while (true) alpha should have a price. These costs (fees) can and should be taken into account in portfolio design. Thus the three key variables in portfolio construction will be risk (for example VaR), information ratio and fees.
Alpha beta separation and portable alpha
The terms portable alpha and alpha beta separation are often used interchangeably. Portable alpha can either refer to the mechanism of porting alpha by transferring beta or to any stand alone vehicle that provides an alpha exposure. A few years ago when the concept of portable alpha was new, hedge funds were seen (or wanted themselves to be seen) as the main way of achieving the stand alone exposure. The more recent discussion of alternatives betas has nuanced this picture considerably. The broader term â€˜alpha beta separation’ refers to the fact that both alpha and beta have a role in portfolio construction and that a better job can be done by treating the alpha and beta as separate building blocks.
In the traditional asset management paradigm alpha and beta production are bundled together. In this long only world, alpha cannot be accessed without a beta carrier. If the beta stream can be neutralised through derivatives, that is, if the beta is transferable, a stand alone alpha can be accessed. Alternatively, the original beta can be replaced by another beta, for example, a small equity beta with a large cap beta, to get the desired beta exposure. Obviously, the beta transfer is seldom frictionless, but there are different slippage costs arising both from direct costs and bid/offer spreads as well as from imperfect replication that adds noise. Therefore, the information ratio could be slightly lower after the beta transfer.
In many cases the investment process is designed already at the beginning to be beta neutral, so that â€“ at least in theory â€“ there is no need to alter the beta exposure. An equity L/S or a relative value fund might be a typical example of a vehicle that has no systematic beta exposure .
Why is alpha beta separation interesting?
In portfolio construction the various return streams, whether called alpha or beta, are the raw material, the building blocks from which the portfolio is put together. It is intuitively clear that the more degrees of freedom you have in managing your building blocks, the better the investment process can be designed. The caveat is that the puzzle gets more complicated and that you therefore have to really know your raw material to handle it in the right way.
In an alpha-beta separated world these two sources of return can, by definition, be treated separately. This means that the alpha that would sit on top of an unattractive beta stream can still be interesting if the beta can be neutralized. Another advantage is that all return sources can be taken into account when optimising the portfolio. As alpha streams often are reasonably uncorrelated with beta sources, the risk profile can be improved by allocating to alpha.
For AP3, one the key drivers was the need to be able to manage the Fund’s beta exposure in a more dynamic fashion. As changing of the strategic beta exposure often involves physical allocation of assets, a situation where the alpha is tied to an underlying beta exposure is clearly suboptimal. For example, assume that we would like to change our exposure to Japanese equities to zero. In an alpha-beta bundled world that could not be done without discontinuing the active management of Japanese equities.
What about the risks?
Alpha beta separation is a genuinely interesting development, but is as such no guarantee that the portfolios we built will be better. The list of questions to consider is long:
- The key issue is: where do we find the alpha? Production of alpha is a skill based activity and this remains a fact even in a separated world. There has to be managers who can produce alpha, and we have to have skill in identifying those managers. As more and more money chase the alpha opportunities, the leverage might grow, increasing risk of bad outcomes or alternatively, the alpha might be diluted so that less and less remains after the costs. So there is a risk of over optimism.
- Governance budget â€“ risk management can get more complicated due to increased use of derivatives. This could also be a challenge for back/middle office functions in smaller organisations.
- Costs might increase, at least if the investor chooses prepackaged solutions
- How good is the diversification in risk scenarios?
What has AP3 done?
We started to look alpha beta separation in a more serious way in 2006. Since then, the major developments are:
- We have abandoned the old long-only active management style in our Swedish equity portfolio. Instead the portfolio has been separated into a passive core and a few L/S hedge fund like portfolios.
- We have increased the risk budget for overlay products, both for internal and external mandates. We have recently finalised a tender for GTAA and are in the process of funding the first mandates.
– Erik Valtonen, Stockholm, Sept. 7, 2007
Editor’s Note: Valtonen is speaking at Terrapinn’s Portable Alpha Asia conference in Hong Kong this Tuesday and Wednesday (at the Hong Kong Convention and Exhibition Centre). The opinions expressed in this guest posting are not necessarily those of AllAboutAlpha.com or Terrapinn.