Friday, July 03, 2009

The Hot Hand Fallacy

And how asset managers may mislead you.
Can you predict on which side the coin will fall looking at the following sequence: Heads (H) or Tails (T)
Most people may erroneously believe that the coin in the first sequence will land on tails. This is a very wrong assumption. The fact is that it appears as a sequence, but is actually just as random as the second sequence. Even more important is to realise that most people will perceive casual regularity of this sort as ‘n sequence of events.

According to Kahneman [1998], this was first documented by Gilovich, Vallone and Tversky [1985] as the “hot hand” fallacy. They did a study amongst professional basket ball players. The same argument can be made for place kickers in a rugby team. (Die warm skopskoen verskynsel)
Gilovich, Vallone and Tversky analysed the outcome of basketball players’ shots, both from the field and from the free throw line, in hundreds of games. In their analysis, they failed to turn up more deviations from a players’ long term shooting percentage than one would expect to occur purely from chance.
In other words, at least in professional basketball, the hot hand is an illusion. Kahneman argues that the human mind is a pattern-seeking device, and it is strongly biased to adopt the hypothesis that a casual factor is at work behind any notable sequence of events.
The hot hand fallacy is ever-present in the world of finance, where it lends unfounded credibility to fund managers who have been successful for a few years. I am always dumbfounded when fund managers publish their 3 months, 6 months and even 1-year returns.
On which is our preferred supplier of performance data for South African asset managers, they even go as far as publishing the top fund for the last 3 days. Does the best performing fund in South Africa for the last three days really deserve your millions?
No matter how exciting investing in the stock market sounds – it is actually the boring asset managers, which does not attract the lime light and does not spend a fortune on marketing their products that’s actually worth considering.
Performance is fortunately one of those things that is actually measurable – so why then, do people go and invest money based on what they heard in a radio or television add?
I am not arguing that long term past performance is the only thing to go by, but is not a better yardstick than a radio ad?
However, let us get back to the hot hand fallacy. The problem with attributing value to chance fluctuations, is that investors over reacts to any information to which their attention is drawn. You might perceive a trend where none exists.
Odean [1998b] analysed hundreds of thousands of trades made in a broking firm. He finds that when individual investors sold a share and immediately bought another – the one they sold outperformed the one they bought by 3,4% in the first year, on average. I think that the study would have been much more useful if the returns were measured over say five of ten years. One year in my opinion is very limited, but the concluded difference would very well have been even more.
Kahneman argues that this expensive overtrading may be explained in terms of these two biases: (i) people see patterns where none exists and (ii) they have too much confidence in their judgements of uncertain events.
Kahneman, Daniel. “Aspects of Investor Psycology: Beliefs, preferences and biases investment advisers should know about. “ Journal of Portfolio Management, Vol 24 No. 4, Summer 1998.