MLC and Russell have both added DAA to their portfolios, joining the previously established processes at Mercer and Ibbotson. While the concept makes sense, a degree of caution does too. It seems obvious to most market participants today that equities were overvalued in 2007 and undervalued in 2009, but we need to remember that these were extreme and very rare observations. We hadn’t seen anything like the 2007 degree of over-valuation in the Australian market since 1987 (it dwarfed the tech bubble, even taking a Schiller measure of value) and then we went straight into a market that had not been as cheap since the recession of the early ’90s. Investors may have to wait a long time in their careers for such opportunities to emerge again. Furthermore, getting the timing wrong in asset allocation is much more serious than getting it wrong in stock selection because no diversification effect exists to mitigate the consequences of the error.
Increase opportunity One way around this is to increase the breadth of the opportunity set. The traditional categories of asset classes are appropriately being re-addressed. The asset allocator has sub-sectors, styles, capitalisation buckets, alphas, betas and betas prime at their disposal. They can use these to make allocations to the fundamental drivers of risk and return rather than arbitrary benchmark groupings: for example, allocations to the large-cap equity risk premium, duration and credit rather than domestic equities, domestic fixed-interest and globalfixed interest. This improves the ability to think about correlations and budgets for risk properly.
Risk management Post-crisis, many investors feel that risk management let them down. After all, the task of risk management is to dynamically quantify the amount of risk present in portfolios and to identify strategies to bring the level of risk back to the desired amount. The question is, did portfolios end up performing badly because: 1. it was not possible to actually bring risk down (that is, no strategy could have protected returns in the GFC because correlations went to one); 2. the desired level of risk was actually too high (investors didn’t really focus on risk, rather let target returns determine allocations?); or 3. investors were not actually able to properly quantify the level of risk in their portfolios? In all likelihood, it was a combination of all three. The first idea has to do with the role of diversification and underestimated correlations in a crisis. While it is true that most risk assets sold off in this particular crisis, that partly reflects the back-end of an environment where all risk assets had rallied together for a number of years as liquidity and leverage drove equities higher and credit spreads tighter.