3.3.3.3 Opinion Leaders

As the reference to Warren Buffet suggests, information cascades often originate with opinion leaders. In stock markets, the main opinion leaders are analysts, whose status as experts is so undisputed it may lead investors to put aside their own views.1 In terras of corporate voting, it is clear that if increasing numbers of shareholders base their voting decision on the judgment of a perceived expert, the probability that they collectively choose the correct option will increasingly depend on the probability that the expert chooses the correct option.2

The obvious expert that shareholders rely on is the board, which is generally perceived to have superior information as far as the firm is concerned. And of course, shareholders will be fully aware of the board's opinion. The mere fact that the board proposes a merger signals to shareholders that the board believes such merger would be beneficial. The board will next distribute proxy materials making the case for the merger, for example, by projecting synergies. These synergies will often be difficult to verify. In the absence of counterfactual information, many shareholders will be inclined to rely on the board's recommendation and vote for the merger. Skeptics, meanwhile, may prefer to vote with their feet by selling shares instead of voting against the merger—that is, they may prefer "exit" over "voice," in Albert Hirschman's terminologycausing self-selection among voters.3 Together, these insights may help to explain why shareholders almost never vote acquisitions down.4 If this explanation were correct, it would be troubling from an epistemic perspective. As we have seen, boards too can sometimes be wrong. Their expert opinions may be skewed by overconfidence, their real motive to merge may be empire building, and so forth. To prevent a board's flawed judgment from translating to erroneous policies, it is key that shareholders vote on the basis of information obtained independently, or at least processed independently, rather than blindly following the board's recommendation.

By now, the reader will have noticed there is something of a trade-off between voter competence and voter independence. It may be possible to increase the accuracy of the average voter 's prediction by, for example, publishing an expert's opinion. But this increase in average accuracy will come at the cost of a reduction in prediction diversity. Scott Page has formalized this trade-off in his Diversity Theorem, which shows that average individual accuracy and prediction diversity are both equally important.5 When it comes to corporate voting, it is difficult to measure average individual accuracy and also prediction diversity. Nevertheless, the Theorem is important for the analysis of voting efficiency because it cautions against overestimating the importance of accuracy and underestimating the importance of diversity.

The trade-off becomes a real issue when policymakers have to decide, for example, to what extent they should facilitate participation by voters with relatively low competence. To see the issue, consider the role of retail investors in the stock market. The fact that they have relatively linie information and suffer from bounded rationality suggests that, from a theoretical perspective, we might be better off excluding them from the market. After all, if retail investors invested solely through mutual funds, the competence of the average market participant, by definition an institutional investor, would be higher (and retail investors wouldn't lose as much from trading on noise). At the same time, however, there would be less prediction diversity. Gregory La Blanc and Jeffrey Rachlinski, in their paper In Praise of Investor Irrationality, make essentially this point when they argue that excluding noise traders could result in less accurate prices because in a market consisting only of institutional investors there would be a greater risk of correlated biases.6

The competence of the average voter could, again from a theoretical perspective, similarly be increased by excluding retail investors from corporate governance. In a way, discrimination by competence already takes place when the right to put items on the agenda is restricted to large shareholders.7 More subtly, policymakers could discourage retail shareholder participation by allowing practical barriers to the exercise of voting rights to persist.8 The preceding analysis suggests that a downside to such a policy would be that prediction diversity is reduced. As a consequence, it is far from clear that shareholders as a group would be right more often, and there is even a risk that on balance, epistemic quality would decrease.