Consumers are in a weak position when it comes to assessing the risk inherent in their superannuation investments, industry experts say, given the risk measure typically used in Product Disclosure Statements is unreliable and tells only part of the story.
The Standard Risk Measure (SRM) provided in superannuation funds’ PDS and mandated by the Australian Prudential Regulation Authority (APRA) is “one of the worst risk statistics I’ve ever seen,” David Bell, outgoing chief investment officer at Mine Super, told the Fiduciary Investors Symposium.
The discussion on the topic of non-modern portfolio theory and metrics around volatility and risk highlighted the difficulty investors face in measuring and controlling risk, as well as in explaining risk to the general public.
Bell, who recently announced his resignation from Mine Super, said there was no consistency between funds in the typical PDS categorisation of assets into “growth” and “defensive”, leading to a weak starting point for consumers trying to understand these concepts in their investments.
And while not blaming APRA as its SRM measure was developed via an industry working group, he said risk is hard to explain and requires more than one metric.
“So pity the poor consumer in all of this,” Bell said. “There’s risk everywhere, they wouldn’t know what they’re exposed to do, and here we are doing a lot more [internally] but we’re all struggling to communicate this. So there’s so much more to do on the consumer side as well.”
A diagram that presents a range of outcomes could be more appropriate, he said, as Mine Super’s fixed income option currently has the same risk measure as its aggressive diversified option.
“It tells you the likelihood of a negative outcome is about the same because yields are low, but it doesn’t tell you anything about the severity of that drawdown event.”
Fellow panel member Sunsuper asset allocation manager Andrew Fisher responded to a question on how funds ensure their risk profiles remain constant through time and whether this was something that could realistically be done.
He said asset owners constantly struggle to maintain volatility levels and this was not “a solvable problem in any meaningful way”.
“You perhaps need to educate people…that the risk in markets is going to evolve through time,” Fisher said.
Internalising asset allocation has enabled Mine Super to build its own risk models to better assess the risk of investments with abnormal distributions.
Investment management firm CFM uses “very specific risk metrics” to risk-manage a portfolio of assets that do not have normal distribution curves, said CFM International president, Philippe Jordan, and the metrics vary depending on the type of investment.
Having data over a timeframe of several decades helps CFM measure realised volatility. Generally the organisation aims for a Sharpe Ratio of .6 to .8 by building a portfolio of diversified beta, and measures the statistical significance of variations over time to determine whether structural change has altered the underpinnings of its strategies.
“Unfortunately, you’ve got to pay to find out if things are significant,” Jordan said. “The higher the Sharpe ratio, the good news is that you need less time to measure significance.
“If it’s a big asset class like equities and your assumption is .3 or .4 or .2 or whatever it is, you had better have some good insight as to why it’s at that level because you’re going to need a lot of time to challenge that assumption through stats.”
CFM algorithmically controls risk in its portfolio by measuring realised volatility in what it calls ‘bins’, which are time periods that are sometimes as short as five minutes and sometimes measured in days. The organisation re-shapes the portfolio according to its measure of realised risk.
“We increase risk if the actual realised risk has dropped beyond a certain threshold and vice versa on the other side. So we’re managing a constant risk envelope.
“And the reason we do that is because we try and make our living by managing tails and trying to earn our money on the bulk of the distribution, and if you do that at variable risk, you’re going to get very strange results that are not coherent over time.”