Lorient I am reading a trading book – The Way of the Turtle – talking, among other things, about the behavioural basis of trader actions and thoughts. One of the ideas that caught my attention is the “outcome bias”. In a nutshell, we tend to judge a decision made three months ago (when people making the decision knew what was known at that time) by its public outcome today. While the author talks about how we feel miserable when we buy a stock that tanks — even though three months ago it looked set to take off — I was thinking how much is this bias influencing management appraisals. I have nothing against assessing results — in the end it is the only (mostly) measurable outcome — however I think it is a poor way of assessing management ability to make the right decisions in given circumstances.
One may argue that an insurance company will punish you by increasing your insurance rate if you get involved in three “no fault” accidents in a month (call it a “bad luck” premium increase). But it should not be the way to assess if a manager is a good manager. Otherwise we may end up with someone who got the “right outcome” once — see another excellent investment book Fooled by Randomness for details — and wonder why she does not repeat the previous performance.