Issue: June/August 09


Are AR benchmarks okay?

Far from it. ROLAND ROUSSEAU rips through the myths that retirement funds commonly accept.

Rousseau...put skill to the test

Rousseau...put skill to the test

Absolute-return (AR) investing is an important way to control risk, but we need to define the concept carefully. Typically, AR refers to any type of investment that allows the manager to go both long and short so that the overall portfolio risk can be reduced and potentially provide positive returns in both rising and falling markets.

Although there is no guarantee of absolute returns, the ability to go long and short could still provide significant risk benefits to investors. This is the noble and exciting aspect of AR funds. In addition, AR pundits will correctly point out that long-only investing prevents active managers from utilising negative forecasts and profiting from these views through short-selling. Much research now shows how constraining long-only investing really can be.

But things start going off the rails with the commonly-held belief that traditional long-only, benchmark-relative, investing is bad and that AR managers therefore don’t need benchmarks. The constraining influences that benchmarks place on a manager trying to beat the JSE’s Alsi or Swix indices are seen as too punishing. The benchmark is then blamed for restricting manager skill.

Not having a benchmark, however, is like having a society without morals. Major problems occur when you claim you don’t need benchmark. How would you know whether you’re getting value from your AR fund? To measure any investment skill, a benchmark is essential.

Make up your mind: absolute vs relative returns

The biggest setback in the history of investing is the adoption of cash returns, or the consumer price index plus x%, as ‘benchmarks’. By accepting them as ‘benchmarks’, we lose our bearings. Simply, they cannot reflect the risks in the active fund under scrutiny.

If your benchmark has a totally different risk profile from the AR fund with which you’re comparing it, it’s impossible to draw conclusions about how much value is being added for you as the investor. Taking for example an AR fund with annual volatility of 10% against a cash benchmark with zero volatility, is like comparing Michael Schumacher with Lance Armstrong. Cash returns, or cpi+x%, can never act as benchmarks. They are nothing more than return targets or hurdle rates.

An equity market’s passive index such as the Alsi can outperform, for years at a time, a cash or cpi benchmark. Does this mean the passive market index has ‘skill’ or has generated absolute returns? Of course not!

You are merely comparing the equity asset class, with its much higher level of risk, to a riskless cash return. This is meaningless. Yet we seem to think it acceptable and continuously repeat the mistake with AR funds. It’s a global problem that will eventually come home to roost.

Need to match risks, not returns

So how do you correctly benchmark an AR fund? You need to identify upfront with the pension fund or consultant all the significant risk factors or ‘betas’ to which you will expose the fund – such as small caps and emerging markets, also known as the risk budget – and then calculate the residual return that remains after accounting for these pre-defined beta returns. This residual return is deemed to be skill or ‘alpha’.

Smart AR managers might argue that their funds are market- or beta-neutral and therefore are not exposed to market risk. They’re right, except that the funds must nevertheless be exposed to other risks or betas because they will necessarily have volatility.

It’s misleading to contend that market-neutral funds, because they have a beta of zero to some arbitrary equity-market index, have no beta risk. The sub-prime crisis has proven that AR funds did and do indeed have risks, lots of them, in the form of common betas.

Long/short investing is undoubtedly a valuable pursuit if implemented correctly. Critically important is how we attribute the source of returns. Merely by calling a fund AR does not make it immune to risk. Excess return comes primarily from excess risk, not skill.

AR funds, even the good ones, are less risky than long-only funds. But still they’re risky, and at times might be highly risky. We can’t marvel at the term ‘absolute’ when nobody can actually see the ‘absolute’ part.

AR managers should rather accept a spade as a spade. They should deliver ‘asymmetric returns’ which is a much more tangible, easy to explain and factually accurate term. The long-short, risk-controlled, philosophy behind the inaptly-named AR concept will probably blossom. But let’s clean up the act and get the benchmarking right. Avoid creating nuclear bombs like ‘absolute alpha’ that will explode. And accept that correctly measured alpha is a zero-sum game where, for every winner, there must be a loser.

Roland Rousseau is an independent consultant focusing on portfolio construction and design of investment strategies for industry roleplayers such as pension funds, fund-of-funds and hedge funds. Having worked for 14 years at Deutsche Bank, implementing global best-practice quantitative solutions for clients, he now advises CIOs of large investment funds in the UK, Europe, Scandinavia and Australia.