March 27, 2007 - Why Gurus Go to Extremes
Are stock market forecasters prone to hyperbole? Is there logic to predicting plunges and melt-ups at probabilities unjustified by rigorous empirical analysis? In their February 2007 paper entitled "Probability Elicitation, Scoring Rules, and Competition among Forecasters", Kenneth Lichtendahl, Jr. and Robert Winkler apply game theory to model the behavior of forecasters who pit themselves not only against the data, but also against each other. In other words, they examine the logical behavior of a forecaster whose reward depends not only on own accuracy but also on the accuracies of competing forecasters. When forecasters compete, they conclude that:
- Forecasters who want to do better than other forecasters rationally assert extreme probabilities. They are willing to sacrifice raw forecasting performance to optimize relative performance. Forecasters who are more competitive tend to greater hyperbole.
- This strategy of exaggerating probabilities makes forecasters in general behave as if overconfident.
- Decision makers can adjust for this motivational bias by revising the probabilities asserted by experts toward 50%.
In summary, forecasters trying to beat other forecasters tend to take extreme public positions that reflect the motivational bias of competition. An investor considering the public forecasts of gurus should probably shift asserted probabilities away from 0% and 100% toward 50%.
Note that the gurus may be unaware of this bias.
More broadly, being correct at the highest possible frequency is not the only game of forecasters. For example, their forecasts may be:
- Marketing ploys designed to capture fearful or greedy naive prospects, safe in the knowledge that quack detection is difficult when it hard to distinguish skill from luck. (One "big call" may anecdotally outweigh an otherwise extended record of mediocrity.)
- Attention-getters for the media, designed to boost readership/viewership and thereby advertising revenue.
- Attempts to manipulate others in support of existing or planned trades.
The tendency to hyperbole seems obvious in the discourse of other fields, such as politics and sports. Perhaps there is a common underlying imperative driving the inhabitants of all environments, both physical and abstract, to expand into every possible niche.
For related research, see Blog Synthesis: The Wisdom of Analysts, Experts and Gurus. See especially our blog entries of 2/16/07 on the diversity and persistence of quacks and 1/17/07 on the apparent lack of relationship between forecasting accuracy and attention.
For examples of extreme stock market forecasts, browse some individual guru forecasting records.
No comments:
Post a Comment