I want to first thank everyone that has commented on the posts so far – reading reactions is one of the greatest rewards from writing the blog. The comments yesterday from PD and Pat Burns highlighted some central issues and intelligent questions surrounding fund and manager analysis, which was great. Today, I want to address what I believe to be the single biggest issue in this field and why, because I never saw away around it and its implications, I successfully removed myself from this business after a long period. The issue is this:
The quantitative results of any large portfolio analysis are predetermined by the market activity of the time slice being studied.
Stated more simply, the most risk-averse strategies will generate the best returns in a bad market by protecting downside and the reverse will also be proven true, that managers with the highest risk tolerance will be highest rated during periods of strong market returns. The conventional method of mitigating this dilemma is to “risk adjust” the results, comparing nominal returns with volatility, historically mean variance but “Post-Modern” concepts like target variance and target semi-variance are also used. I would argue that taken in context, risk adjusted returns can be useful but there are large companies with hundreds of well paid experts and armies of academics willing to argue both sides of that argument.
If we go back to my central point, the decision on which fund manager deserves assets in the current environment is determined almost entirely by the decision on whether to emphasize three or five year numbers in the historical analysis. The five year numbers, while including a couple of months of more welcome 2006 results will also contain the later 2007 onset of disaster and the terrors of ‘08. The three year numbers, despite the recent sell-off, will be much more favorable to risk-tolerant managers as they will be dominated by the post-March 2009 recovery period and the, in hindsight, relatively sanguine 2010 and 1H 2011. Risk-averse managers will look (relatively) great in the analysis of five year data but will likely look mediocre in the three year. (Admittedly, the manager’s decision on precious metals exposure throws a big, big wrench into the spreadsheets, but that is a separate, asset class issue that deserves it own discussion. )
Which is more useful, the three or five year numbers? We are only going to know after the next 24 months or so. THAT is the central dilemma. The analysis can help understand the performance characteristics of a fund but this is only useful if combined with a correct, forward-looking market call. If markets are going to be crap for the next two years, the analysts that used the five-year numbers will look like geniuses. A strong equity market recovery in 2012 will make them look like morons.
Everybody wants the same thing in the end, high relative returns with a minimum number of sleepless nights. Human psychology being what it is, however, investors are often their own worst enemies in realizing their goals. Risk-averse investors, for instance, should want to underperform the benchmark in a bull market – it implies a strategy of risk management that will protect them when, inevitably, the benchmark heads lower. One of our favorite phrases was “no one’s managing risk in the index”. Of course the benchmark’s going to outperform in a sustained bull market. What happens in practice, investors getting frustrated by trailing the index for a couple of years, and then switching their money into more aggressive pools right before a downturn, goes a long way in explaining why most people lose money.
The “time slice” issue and behavioral economics makes the selection of investment management or advice a tricky thing. For myself, I have reached a point where literally don’t care about the benchmark. I have specific future financial requirements and am tailoring my personal investments to achieving those – what happens to the MSCI Global benchmark over some random period of time is of merely passing interest.
I will end off with a point that doesn’t really fit in, but has fascinated me forever: the S&P 500 is actively managed. (I always picture the kid from The Matrix saying “there is no spoon” when I write that). There are criteria by which stocks are removed and replaced – that’s the active part – and somehow this process most often generates higher returns than Ivy League mangers who have been studying markets for their entire adult lives. Huh.