Risk & Portfolio Construction: Optimising the risk factors
One of the most popular ways of implementing the insights that come from analysing portfolios according to risk factors rather than asset classes is ‘risk parity’.
Risk parity is the portfolio construction approach that seeks to identify independent risk factors and then, usually through the use of derivatives, to isolate them and apply leverage to those that exhibit lower volatility, to equalise their contributions to the overall risk of the portfolio.
And yet it is clear that the most naïve forms of risk parity waste some of the most important insights of the risk-factor analysis that are their theoretical foundation.
“When you identify risk factors you inevitably ask one of the most important questions,” says APG’s head of client risk management, Pieter van Foreest. “Is this risk rewarded positively, negatively or neutrally over time? If it’s positive you want to invest in it; if it’s neutral you don’t want to invest in it; if it’s negative you want to insure against it.
Otherwise, that question somehow falls off the table when we come to invest.”
It also falls off the table if you pursue naïve risk parity. An analysis of the main risk factors bundled in US corporate bonds – corporate credit, US Treasury duration, the US dollar and liquidity – might suggest that in most conditions most of the risk comes from corporate credit risk but that sometimes interest-rate risk becomes more important, that investors get paid handsomely for the liquidity risk but only with very short windows of opportunity, and do not get paid at all for the US dollar risk, on average, over time. That should determine which bits of this “structured product” you want to be exposed to in which periods, and to what extent.
“You should use some estimate of future returns, because otherwise you are neglecting the question of which risks are rewarded and which are not, and just assuming that all risks are rewarded in the same way,” says Van Foreest.
Rita Gamelou, portfolio manager on BlackRock’s Market Advantage Strategies fund, makes a similar point.
“The risk-factor approach helps because it means we can properly monitor the pricing of each of these fundamental risks,” she says. “At the moment, many people say that high-yield bonds look expensive, for example, because the yield is low: but look at the two components and you see that the spread looks like a decent premium for credit risk and that the yield is so low because the premium for rates is expensive. In that case it’s about not throwing out the credit-spread baby with the interest-rate bathwater.”
As such, the Market Advantage Strategies portfolio that Gamelou manages does not weight all risk factors equally, but excludes those that the team does not consider to be rewarded consistently over time and “leans into” those that are, and tactically weights to exploit shorter-term valuation opportunities.
Indeed, the portfolio maintained by Danish pension fund ATP, which often gets described as a risk-parity portfolio, is really constructed “according to risk parity principles”: it also prefers to optimise exposure to those factors that deliver the highest risk-adjusted returns over time rather than go down the non-optimised, naïve risk-parity route.
“Most investors like to take risk when and where it is rewarded the most,” as Ewout Van Schaick, head of multi-asset strategies at ING Investment Management, puts it. “When we advise clients to follow a risk-budgeted approach, we also advise them to combine this with a dynamic allocation process that incorporates some views on the market. Right now we are in a low-interest-rate environment, for example, so leveraging your portfolio of sovereign bonds might not be the best idea.”
This touches upon tail-dependency – the fact that some risk factors have fatter-tailed distributions and that an experience in the tail of one risk factor can be correlated with a similar tail event in another. Naïve risk parity applies leverage to lower-volatility risk factors not only without taking account of their pricing and expected return, but also without taking account of the distribution of their returns. This is crucially important because the risks with low volatility are almost by definition those with the fattest tails in their distributions – tails that will only get fatter with leverage applied.
“If you are taking account of risk beyond volatility it will raise the bar as to where and how much you want to apply leverage and need to apply leverage [in a portfolio built along risk-parity principles],” says Tower Watson CIO Chris Mansi.
“If we knew nothing we would start with an equal risk portfolio,” as Gamelou at BlackRock puts it. “But we increase our weight to the factors that are best-rewarded in equilibrium and decrease to those that add to the left tail.”
And Wolfgang Mader, head of asset allocation strategies at Allianz Global Investors’ Risklab agrees: “If you have any return expectations, you should do a proper asset allocation process – but then, afterwards, always control for risk clustering,” he says.
“That’s the best way to make proper use of the risk parity idea.”