There has been the strong view that there is wisdom within crowds. Research has shown that a group will do better than the individual at predicting. The collected wisdom and ensemble from the group will be closer to reality than any one person.
A closer look suggests that group dynamics have some major problems with making good decisions. There is a big difference between a statistical group, wisdom from the average forecast of a group, and the deliberative group as a predictor. Generally, we don't poll members of group and form their average. The normal approach is that the group will deliberate and come together to solve a problem, make a decision, or prediction.
In the case of asset management, the core group is the investment committee and when you look closely at the mechanics of an investment committee, you will likely conclude that many are dysfunctional and not very good at their core task.
The problem with most groups is that they can amplify errors through behavioral biases no different than individuals. There can be cascade effects that opinions are swayed by a few that push the rest to a decision. There can be polarizations that is driven to greater extremes, and there is often focus on shared information and not on new information. Not much hope for the group, yet there are some simple changes that can support better group thinking.
The solutions follow a simple framework that emphasizes using data, rewarding successes, and ensuring equal weight toward alternative opinions.
The power of models or even an ensemble of models is that they will not have the problems of group dynamics. The issues associated with group dynamics and potential solutions is discussed in the book Wiser: Getting Beyond Groupthink to Make Groups Smarter by Cass Sunstein and Reid Hastie. Everyone should work on improving group dynamics but there is clear value with allowing an ensemble of models serve as the group.
No comments:
Post a Comment