Cultural diversity - JP Bouchaud
https://www.nature.com/articles/s41598-021-95914-7
Everybody has heard about the “wisdom of crowds”, which posits that the average belief of a large crowd is more accurate that most individual beliefs.
But consider the following paradox: if listening to the crowd is indeed more rewarding than individually scouting for information, the emergent belief should soon become entirely self-referential and totally disconnected from ground truth. The wisdom of crowd effect thrives on some form of diversification, which is no longer present precisely when too many people adopt the same opinion based on vox populi.
Fortunately, we do not all agree. Some of us continue to look for information, even when it is costly to do so, while others can happily free ride. How can this division of labour be evolutionary stable? This is the conundrum that Benoit de Courson proposed to solve in
The two key ideas are (a) individual information seekers are not only striving for accuracy but also for audience, which brings them prestige, consulting fees, etc. that compensate for their lower success rate (b) some followers understand the value of deviant predictions: even if less accurate on average, the opinion of information seekers is genuinely relevant because it is independent from the opinion of others.
When included in an evolutionary model, these two features allow, in a certain region of parameter space, for the stable coexistence of three “species”: plain vanilla followers, followers valuing dissenting opinions, and genuine information seekers. Cultural diversity and wisdom of crowds are then mutually compatible. Interestingly, this socially beneficial equilibrium disappears when the signal that information seekers attempt to capture is too weak. In this case, information seekers go extinct; wisdom of crowds produces strongly polarized opinions based on thin air.
This whole story is reminiscent of the Grossman-Stiglitz efficient market paradox. The distinction between information seekers (analysts) and followers is indeed a relatively faithful description of the ecology of financial markets. But financial markets are notoriously difficult to predict and quacks do proliferate as it is hard to distinguish them from genuine experts.
Grossman & Stiglitz proposed that financial markets are close to efficiency, with a small residual error that allow information seekers to cover their costs while allowing the aggregation process to perform its task. I am much less optimistic and rather believe with Fischer Black that market prices are within a factor of 2 of fundamental prices. Only when mispricing becomes large enough will information seekers become relevant and drive back prices to more reasonable values – albeit sometimes abruptly!