Should The Value Perspective be employing a ninja or some other type of assassin? We only wonder because of a recent piece of analysis by empirical research partners that included the following graph, which shows the average returns recorded by companies in the year after their Chief Executive left office because of retirement, resignation or death..
Date as at October 2013
The chart uses data going back to 1977 and, as you can see, the companies whose chief executives died have easily done best. still, before you think about hiring a sniper as a way of boosting your portfolio performance by a few percentage points, we should quickly point out this chart is actually a much clearer illustration than one normally gets to see of the difference between correlation and causation.
We believe this chart shows more about the former rather than the latter. There is no serious suggestion here that, if you went out and assassinated the chief executives of all the companies you own, your portfolio would suddenly outperform. Well, that may depend on the company and the chief executive – but then that is precisely the point.
It is not the chief executives’ deaths that create the outperformance here but some other relationship – one that is undefined and unknowable from this chart. However, what should be clear from the chart – or at least we would hope there are few people reading this who would conclude otherwise – is that what looks to be the explanatory variable may really be a spoof or red herring.
With the above chart, it seems probable the data has been skewed – perhaps because the number of chief executives who die in office is likely to be a very small sample. Often, however, the data contained in a chart will look utterly plausible and you will feel certain you are being shown a causation when, in reality, it is actually measuring something entirely different that you cannot see or has been created through a small data set, odd time horizon or some other influencing factor.
The lesson here is that charts are trotted out all the time in finance and investment that fail to include, say, a range of the highest and lowest variables that could provide a missing piece to the puzzle. Clearly we would like to think this does not happen on The Value Perspective but, as a general rule, when confronted with a chart you should ask yourself whether it is really showing what it purports to.