Benchmark. The very word sounds solid and reliable. Obsessive attention to tracking error, though, perpetuated by the more emotive term ‘benchmark risk’, leaves the underlying structure unscrutinised. So while there is a belief that a small tracking error provides safety, little thought is given to the risk associated with the benchmark itself and the risks inherent in sticking to it.
Common wisdom has it that low tracking error (and hence low benchmark risk) equals low-risk investing. Wrong, says Richard Timberlake, chairman of Investment Manager Selection in London. Instead it is simply a measure of how close the benchmark is being adhered to and equating benchmark risk with tracking error is a non sequitur in that it assumes, incorrectly, that the benchmark is a risk-free investment.
According to Timberlake, benchmark risk (henceforth the risk of deviating) is a gimmick invented to serve the investment manager not the pension fund. Sticking close to the benchmark is never going to produce stunning outperformance but then managers are rarely fired for average performance. “Benchmark risk has very little to do with any form of risk to pension trustees. It’s not a real risk, it’s a risk to the management group deviating too far from the peer group and losing the account,” he says.
Consider, as an example, three investment types – strong growth bias, strong value bias and market related. Market related has a low deviation from the index and is therefore deemed low risk. Timberlake says the other two strategies will deviate more from the benchmark but that if bought in equal measure will produce the same deviation as the market related strategy.
“Anyone who has studied in detail the academic work on returns and the alpha added by good managers will see that it is only added by deviating from the benchmark,” he says. Or in other words, investors are potentially losing out by mimicking the benchmark, a scenario that Timberlake says is borne out clearly at major turning points.
Consider the fortunes of Japan and the US in the past 12 years. By definition, a benchmark includes the most of what is overinflated at the time and the least in what is undervalued at the time. In December 1998, the Japanese market made up 44% and the US 29% of the MSCI World index. Ten years on Japan was down at 10% and the US up to 50%. “In 10 years you would have really suffered if you’d frozen on that asset allocation based on the benchmark, just as you would have more recently if you’d frozen TMT at 40% of your portfolio.”
And striving for a tiny tracking error can carry a hefty price. Alan Brown, CIO of State Street Global Advisors, says that adhering too rigorously to an index can be counter-productive. When Dimension Data entered the FTSE 100 it promptly rose from £6 to £10 and fell back to its pre-entry price a few days later. Managers following the 100 religiously and buying on the up could boast both a tiny tracking error and hefty losses.
While the ‘myth’ of benchmark risk is talked about by academics and a few contrarian managers, Timberlake says few managers running institutional business talk about it as it’s not in their interests. “The asset management business has become extremely self serving and not client serving. Asset managers allow their institutional management groups and marketing departments to perpetuate this idea that the benchmark itself is a risk-free zone whereas, by definition, it is a high risk zone.”
The larger fund managers are apparently being hoisted up by their own size. “It’s interesting that some of the brightest people now prefer to run hedge funds in small boutiques limiting themselves to a few hundred million under management. Why? Because it’s well known that once you’re running £20 billion, it’s hard to add value. You’re forced into blue chips and benchmarks,” he says.
Some pension funds have recently started taking a more relaxed attitude towards tracking error. Eighteen months ago E18bn MN Services dropped tracking error and introduced bands within which managers can fluctuate. “We do not want to restrict the manager. It’s really important that they can follow their own style and not feel too restricted by a certain range. You might get a manager who is anxious about doing something they are doing for all their other clients but not for you as they are worried about tracking error. We do not want to impose that kind of restriction on them,” says fund manager Bas Tiaan de Kreij.
Daniel Gloor, head of asset management at the Kanton of Zurich pension fund says they have to be measured against some kind of benchmark but that tracking error is too theoretical and is given far too much emphasis. Rather it is more important to look at the risks taken and whether they are paying off and achieving alpha. The fund runs small and mid cap funds about which he says: “I couldn’t care less if the tracking error is 5% or 10%.”