Testing Claims About Crowd Forecasts and Contrarian Investing

Summary

The document asks whether the claim that the investing crowd is usually wrong can be made precise and tested. It sketches a simple definition of the crowd as the larger of two groups betting on whether an index will rise or fall over the next period. The author then questions whether the crowd might be wrong more frequently while earning larger gains on occasions when it is right.

Testing this idea would require operational definitions of investor groups, forecast direction, outcome horizon, and what counts as being wrong or right, along with suitable positioning or survey data and subsequent index returns. The document provides no dataset, study, or findings; it is a research question rather than evidence that contrarian behavior is profitable. It also leaves open how to account for position sizes, trading costs, and the distinction between forecast accuracy and investment returns.

Key ideas

• A claim about crowd accuracy needs measurable definitions of the crowd, its forecasts, and the evaluation period.
• One possible crowd definition is the more numerous side of investors forecasting an index rise or decline.
• Forecast accuracy and the size of gains or losses are separate outcomes and should be measured separately.
• Testing contrarian profitability requires investor-position or forecast data paired with subsequent market returns.

Tags

# In investing, the crowd is wrong much more often than right


# In investing, the crowd is wrong much more often than right












I am reading Ken Fisher's Beat the Crowd, and the second sentence in the book is this:

> In investing, the crowd is wrong much more often than right.

I was wondering if there is a way to define the terms in this statement, "in investing", "crowd" and "wrong much more often than right", so that this sentence can be quantitatively tested.

My feeling is that if markets are efficient, in retail investing, this statement would imply that while crowds (assume in the simplest case that there are two sets of investors, those who bet on an index going up in the next time period, and those that bet on the index going down in the next time period, and the crowd is whichever set has larger cardinality) are wrong much more often, the volume of the gains of the crowd when they are right are far higher than when they are wrong.

$$ \mathbb{E}(Y | \text{Crowd is right}) >> \mathbb{E}(Y | \text{Crowd is wrong}) $$

Is this a testable theory? What data would be required to test this? Is there a strand of the literature which tests whether deliberately contrarian behaviour is more profitable?

Thanks.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.