Risk & Portfolio Construction: Theory and practice
A survey conducted in September 2011 by Allianz Global Investors suggests that at least one-third of European institutional investors see the merit in analysing portfolios by risk categories as opposed to asset classes.
However, the extent to which this happens on the ground looks limited. In Denmark, ATP has, for a long time, allocated its assets to five ‘risk classes’ – interest rates, credit, corporate earnings, inflation and the oil price – as well as taking the concept further by taking account of ‘secondary risk factors’, such as curve steepness in the interest-rate exposure and both liquid and illiquid investments to capture inflation risk. Moreover, three Dutch pension funds – the industry-wide grocery trade fund, the corporate scheme of Kas Bank and the scheme for Cordares employees – are set to maintain a ‘shadow allocation’ organised along the lines as set out in a recent report into this question published by Dutch trade union FNV Bondgenoten, ‘Investment risk – an approach aimed at controlling risks in pension fund investment policies’.
But as Chris Mansi, CIO at Towers Watson, suggests, there remain all sorts of decisions left to be made once allocations to risk factors have been decided – such as how much credit risk we want to take from investment grade or high yield, for example, or, like ATP, how much inflation risk we want from liquid or illiquid investments.
“A policy is required to make these decisions, and those policies tend to be founded in a more traditional way of thinking about asset allocation,” he says. “Many investors will sense that they understand an allocation to an asset class better than an allocation to a risk factor. But it’s also about measurement: how can you tell if someone has done a good job if they have just gone away and put together what they say is a portfolio of diversified risk factors?”
That last point is important because regulators work within the traditional framework of assessing risk exposure by asset class. Ewout van Schaick, head of multi-asset strategies at ING Investment Management, says that this is the main objection he gets from institutional investors who feel unable to break free of the asset-class chains.
Wolfgang Mader, head of asset allocation strategies at Allianz Global Investors’ Risklab, agrees that the complexity of risk factor-based investing could conflict with regulatory constraints, but also be a barrier to entry in itself.
“It’s not even the difference between buying a meal and buying the ingredients for a meal,” as he puts it, “but rather the difference between buying the ingredients for a meal and buying the chemical elements for the ingredients.”
One major European investor – Norway’s Government Pension Fund Global – has shied away from going down the risk-factor route, even after a 2009 report into its active management performance recommended that exposure to “systematic risk factors” should be included in the fund’s benchmark.
Complexity appears to have been the ostensible reason. In a letter to the finance ministry investment manager Norges Bank said that it would be straightforward enough to account for simpler risks such as small companies, emerging markets and well different types of credit using investable and testable indices, but warned against “introducing systematic risk factors to the reference portfolios that weaken transparency and testability”.
Still, Norges Bank also argued, in its letter, for a distinction between real and nominal assets, which would have represented a step towards the kind of portfolio managed by ATP.
The fund’s Strategy Council, however, remained concerned to preserve that practicality of the traditional asset class-based benchmark and noted that “labelling equities as real assets and fixed-income instruments as nominal assets does not make them purely so” – equities have suffered during bouts of deflation and high inflation, for example.
Of course, a critic might observe that this is precisely why investors ought to be thinking in terms of risk factors like inflation rather than asset classes like equities – because these risks undoubtedly exist and current thinking misses the ways in which they cut across asset-class boundaries. The Strategy Council, by contrast, wrote that “these issues can be addressed adequately within an asset based framework”, albeit with enhanced portfolio performance analytics that “embrace multiple risk factors”.
Analysis based on the new lines is worthwhile, the fund seems to have concluded, even if full implementation might be problematic. It is a position Van Schaick at INGIM sympathises with.
“There is nothing wrong with investing along the lines of asset classes, as long as you recognise that there are some common factor risks driving a number of different asset classes,” he says. “That is why looking at common factors does provide useful insight – it gives you an indication of this commonality.”