Publications
Partners
WHILE THE NOTION of absolute return (AR) investing is a noble and important way to control risk, we need to define the concept very specifically. AR typically refers to any type of investment approach that mandates the fund manager to take both long and short positions in an asset class (for example, equities) in order to reduce overall portfolio risk and potentially provide positive returns in both rising and falling markets.
Although there's no guarantee of a positive absolute return the ability to go long and short theoretically provides significant risk benefits to investors - that being the noble and exciting appeal of AR funds. Proponents of AR will correctly point out that long-only investing prevents active managers from utilising negative forecasts and profiting from those views through short selling.
A significant body of research now shows how constrained long-only investing is. All too often the way investment mandates are formulated also restrict the value add potential. Given those constraints the ability to more independently and flexibly invest with, in particular, dynamic short positions is enough reason to welcome new investment frameworks, such as AR.
Yet for reasons that are quite unfathomable, proponents of AR funds promote the notion that AR managers don't need benchmarks. Not having a benchmark is like having a society without laws. Major problems start occurring when you claim you don't need a benchmark.
Make up your mind: absolute versus relative returns
A benchmark is critical to (a) define the client's risk tolerance and return objectives and (b) be able to measure the value added by the manager. To further clarify that we need to formulate an appropriate framework for defining what a fair benchmark is for AR.
We view the adoption of cash returns or CPI + x% as "benchmarks" for AR funds as one of the biggest setbacks in the history of investing. That unfortunate development is also the root cause of a range of problems and misinterpretations associated with hedge funds and long/short investing.
By accepting cash or CPI + x% as benchmarks we've lost our bearings, because they don't reflect the true level of portfolio risk. Comparing an AR fund with volatility of 10%/year versus a cash benchmark with 0% volatility is like comparing Michael Schumacher with Lance Armstrong! Cash returns or CPI + x% can never act as benchmarks at all: those are merely return targets or hurdle rates, nothing more!
Thomas Schneeweis accurately summarises the problem in his seminal 1999 paper Alpha, Alpha, Who's got Alpha? "It is not appropriate to say that you have a positive alpha (net risk-adjusted return) simply because the return is greater than the risk-free rate, unless your portfolio is risk-free. Similarly, comparing the return of your fund to the S&P500 or any other benchmark is inappropriate unless your strategy responds only to the same return drivers that drive the S&P500 or the cited benchmark." - Journal of Alternative Investments , 1999
We need to match risks, not returns
So how do you benchmark an AR fund correctly? Well, the same way you benchmark any active fund. You need to identify with your client all the significant risk factors or betas upfront (eg, small caps, value, emerging markets, credit risk, etc) - also known as the risk budget. The active return generated by the fund manager will then be the residual return that remains after accounting for all those pre-defined beta returns. We assume, for lack of any other explanation, that residual return is skill (or alpha).
Some smart AR managers will correctly point out here their fund is market- or beta-neutral and therefore they aren't exposed to market risk. That is, of course, correct: but they must be exposed to other risks or betas because their fund has volatility. It's a misnomer to assume market neutral funds have no beta risks because they have a beta of zero to some arbitrary equity market index. For example, see Dopfel, F (2005).
All that market neutral really means is that the correlation to the market is low. It doesn't necessarily mean volatility is low or zero. The source of that volatility therefore needs to be explained. Managers may in some instances take the blame for lack of skill when actually it's unwanted risk in their process, despite them having excellent skills.
Precisely because there's some level of beta risk - other than market index risk - in AR funds and that this risk is most often unrelated to the skill of the active manager, explains why a non-volatile cash return or CPI + x% is such a poor "benchmark" for an AR fund. In particular, a cash benchmark is incapable of separating the manager's active risk from different market beta risks (credit risk, emerging markets, inflation, small caps, etc).
Barton Waring and Lawrence Siegel explain in their erudite paper The Myth of the Absolute Return Investor in the Financial Analysts Journal (CFA Publications, March/April 2006): "Sometimes, hedge funds are characterised as having a benchmark of cash. One certainly can imagine a hedge fund for which this is appropriate: The normal portfolio for a hedge fund with no net expected average exposures to any styles, markets or other beta factors could be correctly understood as a zero-beta portfolio and its benchmark would be cash.
"In fact, when data from actual hedge funds are evaluated most funds show persistent net positive beta exposures over time. On average, the equity beta of long/short equity hedge funds ranges between 0,3 and 0,6 and they also have some beta exposure to bonds. In effect, most hedge funds normally put fewer dollars into short positions than into long positions, and their net betas do not completely cancel and go to zero."
Every fund's true benchmark has to be a portfolio with the same generic risk drivers and risk profile of the active fund. If you outperform that "shadow portfolio" that has the same risk as your active fund then you have skill or alpha.
Below is a more colourful interpretation from Waring and Siegel, again about the fact all investors have to have a benchmark and therefore we are all relative return investors and that AR as a classification is very ill defined.
"Thus, all managers who make the effort to add special value to a portfolio, whether they want to admit it or not, must do the same thing: beat a benchmark (a normal portfolio or mix of betas). The challenge is the same for a hedge fund, a long-only manager, a market-neutral long-short manager, a traditional active manager, a quantitative active manager - whatever type of manager.
Even Warren Buffett has a benchmark, a cost of capital or blend of beta payoffs that he must beat if he wants Berkshire Hathaway to go up more than the rest of the market. So the most famous AR investor in the world is, in fact, a relative-return investor - as are all AR investors. Relative return investing may seem timid and constrained to those who do not understand the difference between beta and alpha but it is the only means through which real value can be added to portfolios. Relative return investing is the only kind of value-added investing that really exists. Get over it!"
'Absolute alpha': snake oil?
Now we've addressed how AR funds should be benchmarked and how not to, we'd like to clarify a cottage industry burgeoning within the AR universe. There's a curious sub-category of the AR universe that has the penchant and the nerve to call itself "Absolute alpha". Here are some of their registered names: "Absolute Alpha Fund", "European Alpha Absolute Fund", "Currency Absolute Alpha", "Diversified Absolute Alpha", "UK Absolute Alpha Fund," etc.
Though we haven't found any such funds by name in SA yet, AR managers also often talk about their alpha here. As mentioned earlier, the concept and philosophy of combining long and short positions are a gallant one but care must be taken in appropriately defining the fund benchmark.
Specifically, to claim you can deliver "absolute return" as well as "alpha" is simultaneously indicative of a complete lack of understanding of the basic tenets of investing, prudent benchmarking and sane investment practice. That's marketing black magic at its worst.
"Absolute alpha" is a complete absurdity. Alpha and beta have to be, by definition, relative return concepts, where alpha is the residual excess return above an appropriate risk-adjusted benchmark and beta is some representation of a benchmark. So how can alpha simultaneously be an absolute and a relative return?
Schneeweis rather snidely declares there are two types of alpha: a relative performance alpha versus a marketing alpha. The latter being conjured up to sell a product rather than being based on sound theory.
"Managers must know which path they wish to take; that is, alpha as a marketing device or as a measure of comparable risk/return performance. If managers wish to define alpha to fit their own marketing purpose and use alpha to sell a product it is understandable. However, one should never mistake a 'marketing' alpha from a relative-performance alpha.
"If the manager can choose asset positions with a higher return (but the same ex ante risk) to some comparable naive investment position then that person can be said to achieve a positive alpha. Managers may say that investors never care about relative return but only absolute return. But performance alpha is all about properly measured relative return."
The challenge therefore goes out to anybody who can prove - theoretically, empirically or otherwise - how you can generate an "absolute alpha". Almost all absolute managers claim alpha but the only alpha that exists is a relative one.
So the next time an AR manager tries to sell a pension fund "alpha", the pension fund can be confident the fund manager is selling a marketing alpha, not a performance alpha. AR managers publicly denounce all traditional benchmarks yet they claim, with a straight face, some form of alpha relative to a benchmark. Indeed, "absolute return" might exists and "alpha" also exists but they can never coexist!
The clouds are gathering
Investors and the industry will ultimately pronounce on how valuable long/short AR investing is conceptually. There's good logic and reason to support the view that every active manager will end up as a long/short manager one day. But the regulator will have the last word on how AR funds benchmark and market themselves! GIPS will have to step in and create new rules governing the "fair representation" of those funds.
Standard & Poor's press release, London 8 July 2008: "S&P Fund Services warns investors against seeking absolute return 'magic bullet'. S&P said no two absolute return funds were the same and that the risk of widespread mis-buying among investors is high." - Jeff Prestridge.
London Daily Telegraph 9 September 2008: "FSA eyes Absolute and Target return funds. FSA spokesperson: We want to better understand what consideration is given at the product development phase to risk management and treating customers fairly; to learn more about the marketing distribution of ARFs; and to assess the role of ARFs within the asset management industry as a whole."
Summary
While the philosophy of AR is a noble one and is most valuably expressed in a market-neutral way, most AR funds are far from market - let alone risk-neutral and therefore require careful benchmarking and risk management. Cash and CPI + x% benchmarks just don't cut it.
In general, the benefits to go long and short are being able to generate returns in rising and falling markets and to manage total portfolio risk more effectively than some constrained long-only mandates. That much we can salvage from the AR philosophy.
So by definition every active investor (both long-only and long-short), are relative return investors, not absolute. Absolute return - like the phrase "hedge fund" - is a complete misnomer, probably for marketing reasons, even if the principles of being able to go long and short are incredibly valuable.
Roland Rousseau
Independent researcher
ROUSSEAU spent 14 years heading quantitative and investment strategy research at Deutsche Bank and, for the past two years, has been assisting CIOs of the largest pension funds, multi-managers, hedge funds and active funds in Europe, Britain, Scandinavia and Australia on matters concerning the implementation of smart portfolio construction and the design of new investment processes and better benchmarking. He presented to 10 000 plus professional investment managers during 2008 in over 14 countries and was invited to present at more than 10 international conferences. Rousseau has also written several international reports on portfolio construction and the future of the investment industry. He left Deutsche Bank earlier this year and now works in partnership with Professor Paul van Rensburg in designing intelligent indices for smarter institutional core portfolios.
