While Paatsch was at RiskMetrics, the preferred methodology for detecting risk became a combination of ‘value at risk’ (VaR) analysis and stress-tests applied to the benchmark asset allocations disclosed by funds. “VaR is a statistical technique that is useful for measuring the chance of loss across different asset classes. It’s an imperfect number,” he said. “We were acutely aware of the limitations of mathematics and the assumptions about history replaying events accurately in distressed times. We were also aware that modelling at the asset allocation level would understate the active risk present in the actual positions that the fund holds. “But doing something was a useful way of starting the discussion about risk and avoiding the actively misleading statements on risk that dominate most fund disclosures. Surely it’s better to be approximately right than precisely wrong.” Simple measures of risk were simply tools to be used in conjunction with expected and past returns, and fees, when making a decision. “We can’t get entrenched in sophistry about the mathematics of risk measurement when the status quo is actively misleading people now and we know it!” Paatsch said. “There is a whiff of duplicity in the readiness of super funds to give fund managers strong risk parameters within individual mandates they award, and their lack of candour about the interrelationships of risk throughout the portfolio in aggregate. “If we can’t do a better job on risk transparency and management, then, in my view, we are obscuring a primary role of trusteeship.”

Join the discussion