Defensive inversion A crisis that calls into question the fabric of our financial system is likely to affect all risk assets in any case. It is clearly not true, however, that diversification did not work. One only needs to look at the role genuinely defensive assets played in protecting returns. The rub, however, is that investors had taken too much risk with their ‘defensive assets’ making them not defensive at all. Sub-investment grade credit portfolios with long-weighted average lives are not defensive and proved this during the GFC. But at the time, investors’ focus was on chasing yield and hitting return targets, rather than on how much risk they were taking. In many cases, this meant that the desired level of risk ended up being too high. One corporate pension fund in the research noted that “if we had set our asset allocation strategy in terms of the risk we really faced, expected returns would be lower because we would have had to follow a more conservative policy.”
The ability to quantify risk needs elaboration and goes to the heart of our understanding of risk and the quality of the risk measurement tools that were (overly?) relied upon. The research says that the crisis brought attention to risks previously not properly accounted for – liquidity risk, counterparty risk, leverage risks, systemic risks – and that in future the solution is to better recognise and measure these risks. Perhaps this is just rearview mirror risk management. No doubt the next crisis will highlight different risks that will become the talk for the ensuing couple of years. The very real risk that DAA, discussed above, is trying to manage is price risk. How long is a cycle? Can equities underperform over a five-year period? Certainly. Over 10 years? Surely not. It’s time in the market, not timing the market, isn’t it? Not great consolation for British, American, European and Japanese investors who have all lost money over the first decade this century, mainly because of price risk. So the crisis has prompted us to re-think the quantification of risk. Outside of using DAA for price risk, one other way investment professionals are doing this is by recognising that returns are not normally distributed. Neither are they completely random. They are, in fact, fat-tailed.
Returns are fat-tailed Unfortunately, that makes most of the optimisation and value-atrisk analysis relied upon for the past 20 years not very useful. It’s a shame because they looked so robust, so thorough and so very quantitative. From now on the plan is to improve risk analysis through stress testing, conditional value-at-risk, copula functions for covariance measurement, extreme value theory and the like. There is no doubt that these tools will improve our understanding of risk, but it is likely that the full solution lies in using these together with good judgement and, most importantly, realistic expectations. Financial crises will usually do two things. First, they give rise to new thinking that evolves tools and strategies to allow the better management of money over the longer term. Second, they remind investors of some simple principles that were always known but sadly forgotten.