The Cooper Review looks likely to back an idea I proposed some time ago. In October last year I wrote in Business Spectator:
"While still on the subject of super, I want to throw out another idea for the Cooper Review to consider. APRA has spent a lot of time recently trying to get the reporting of superannuation fund “performance” right.
But there appears to be one particularly profound problem underpinning the regulator’s approach to measuring investment performance. Take a step back and ask yourself the question, What do we actually mean by the word “performance”? Are we talking simply about raw returns, or the critical trade-off between risk and return? Surely it is the latter. Surely we are not going to encourage unsophisticated mums and dads to focus on just returns and completely overlook risk? Yet by ranking super funds by their raw performance, that is precisely the consequence of APRA’s approach.
In the research world, this is schoolboy error No. 1. You never evaluate fund performance on the basis of raw return outcomes. You always risk-adjust those returns by factor analysis or simpler measures such as Sharpe Ratios.
What is most concerning is that in APRA’s June Quarterly Statement on Superannuation Performance I can locate 32 separate references to the word “return”. There is, however, not a single reference to “risk” (or “volatility” for that matter). This seems like a very serious oversight.
Now one might respond that by making available the time-series returns APRA provides consumers with a “feel” for the probability of loss. But arriving at correct inferences regarding risk on this basis is way beyond most mums of dads. And by not even canvassing risk, the regulator is actively encouraging consumers to psychologically bury the issue in their minds.
A simple policy solution here would be for APRA to invest some time and effort developing a “risk-rating” analogous to the star ratings used by actuarial firms to rank mutual funds. The appeal of a simple risk-rating is that the regulator could embed a range of sophisticated metrics into it, which lay consumers would not need to worry about. (Of course, APRA would fully disclose their risk-rating methodology through a technical paper that the more informed could interrogate at their leisure.)
The key take-away here is that by providing a risk-rating of some kind APRA would be supplying a valuable public good to the community. They could leverage off their own sophisticated capabilities to distil a range of complex risk considerations into one easily-digested measure. And by putting risk back on the super fund agenda, they would force members to be more discriminating in making their choices as opposed to the current situation where raw returns are the only guide (and I have not bothered here to wade into the issue of the tenuous link between past performance and future outcomes that further undermines the current approach).
Interestingly, there is academic research that shows that “risk-adjusted star-ratings” such as the proxy I have proposed, which seek to incorporate more rigorous analysis of past performance, are much better guides to future outcomes than the raw benchmarks currently employed by APRA.
A final benefit of a risk-rating is that it would help compel super fund trustees to be more careful with their asset-allocation strategies. In a world where consumers can risk-adjust super fund returns, there would be far fewer rewards for those groups that load up on listed equities risk in order to chase high payoffs with much greater probabilities of loss.
One obvious complicating factor here is how one goes about measuring unlisted investment risk. But developing suitable risk proxies for unlisted assets is well within the capabilities of the regulator and should not in any way stymie the exercise."
Today, the Australian Financial Review reports that the Cooper Review is "considering the introduction of a risk-rating system for retirement funds to help members decided which scheme best suits their needs." They then quote Jeremy Cooper as stating:
"We are thinking when we are showing people investment return information, should we introduce some measure of the uncertainty or volatility, trying to use that information to project for performance. We are still wrestling with whether we should put something on the table which requires some form of standard deviation or what have you to accompany past investment return performance."
