Einde inhoudsopgave
The Decoupling of Voting and Economic Ownership (IVOR nr. 88) 2012/3.3.3.0
3.3.3.0 Introduction
mr. M.C. Schouten, datum 01-06-2012
- Datum
01-06-2012
- Auteur
mr. M.C. Schouten
- JCDI
JCDI:ADS597131:1
- Vakgebied(en)
Ondernemingsrecht / Rechtspersonenrecht
Voetnoten
Voetnoten
Vermeule, supra note 19, at 6.
Nicholas R. Miller, Information, Individual Errors, and Collective Performance: Empirical Evidence on the Condorcet Jury Theorem, 5 Group Decision & Negotiation 211, 214 (1996) (noting that '[s]tandard Condorcet Jury Theorem variants assume that individual choices are statistically independent'); Thompson & Edelman, supra note 17, at 149-50 (referring to the condition of voter independence but not explaining why this condition is fulfilled in the case of voting shareholders. In footnote 5, Thompson & Edelman refer to a study by Krishna Ladha (cited infra note 69) suggesting that one can 'relax the independence' of the voters to allow for some level of correlation, but they do not offer an argument for why we can assume that voting shareholders are sufficiently independent).
Krishna K. Ladha, The Condorcet Jury Theorem, Free Speech, and Correlated rotes, 36 Am. J. Pol. Sci. 617, 621 (1992).
See, e.g., Org. for Econ. Co-Operabon and Dev, OECD Principles of Corporate Govemance 38 (2004) (suggesting that shareholders should be encouraged to cooperate and coordinate to overcome collective action problems).
Sunstein, supra note 16, at 54; see also Jeremy Waldron, The Wisdom of the Multitude: Some Reflections on Book 3 Chapter 11 of Aristotle's Politics, 23 Pol. Theory 563, 564 (1995) (discussing the epistemic benefits of deliberation as described by Aristotle). The idea of the wisdom of the multitude resonates in one court's characterization of shareholder resolutions as the 'fruit of deliberation.' Netherlands Supreme Court, 15 July 1968, NJ 1969, 101 (Wijsmuller).
Vermeule, supra note 19, at 6-7, 18-23; see also Sunstein, supra note 16, at 54, 78-80 (discussing positive and negative effects of deliberation on judgments); See Miller, supra note 68, at 214 (noting that '[d]eliberation and mutual influence can be thought of as having two effects: first, they increase average individual competence . . . . [S]econd, they reduce the 'effective number' of group members. The first effect increases collective competence, while the second reduces it, so the net effect is difficult to predict'); Page, supra note 41, at 213 (demonstrating that deliberation acts as a double-edged sword by increasing individual accuracy but reducing diversity within the group).
What we do know is that the Jury Theorem requires statistical independence, not causal independence. This means that voter A and voter B's judgments are allowed to both depend on the same exogenous factor (such as the judgment of an opinion leader), as long as their judgments do not depend on each other. If their judgments do depend on each other, for example, if B always bases his vote on how A votes, statistical independence is lost. See David M. Estlund, Opinion Leaders, Independence, and Condorcet's Jury Theorem, 36 Theory & Decision 131, 132-35 (1994). Unfortunately, we have limited understanding of whether and when causal dependence undermines statistical independence. See Vermeule, supra note 19, at 6.
Lu Hong & Scott E. Page, Problem Solving by Heterogeneous Agents, 97 J. Econ. Theory 123 (2001); see also Page, supra note 41, at 197 (offering a detailed explanation).
This is a simplification of Hong and Page's sophisticated account of cognitive diversity, which deserves a more detailed discussion than is possible within the physical confines of this Chapter.
Lu Hong & Scott E. Page, Interpreted and Generated Signals, 144 J. Econ. Theory 2174 (2009) (establishing the negative correlation result for independent interpreted signals); see also Page, supra note 41, at 197-235.
Independent voting is a "crucial engine" behind the Jury Theorem.1 Standard Jury Theorem variants simply assume independence, and advocates of the Theorem as a theoretical foundation for corporate voting have implicitly done the same.2 But independence is by no means a given. According to one commentator, the main weakness of the Jury Theorem "is that its assumption of independence is unreasonable. Independent voting requires that there be no opinion leaders, that voters do not communicate, and that they do not possess common information, culture, religion, beliefs, or other elements that could lead to correlated votes."3
This quote suggests a number of issues that could compromise shareholder independence. One is communication between shareholders, or "acting in concert," which is widely seen as a means to overcome collective action problems and strengthen shareholder voice.4 To be sure, deliberation might elicit perspectives and information and thus improve the judgment of the deliberating parties.5 But it is unclear whether and to what extent deliberation compromises independence, and "[a]bsent any general account of this, the basic reach of the Jury Theorem is not well understood."6 More generally, it is unclear to what extent the condition of independence can be relaxed.7
Fortunately, the path-breaking work of Lu Hong and Scott Page, which moves beyond the Jury Theorem, enables us to see why independent voting is so important. In the previous two sections we have looked at informed voting and rational voting, mechanisms that can be seen as building blocks of individual voter competency. Now, we need to look at how putting together individually competent voters can lead to an even more competent group. In other words, we have to look at the circumstances that cause one plus one to equal three. That is where independence comes in.
The core insight delivered by Hong and Page is that putting people together in a group can be a means to leverage their individual competence if and when people have diverse cognitive skills.8 And the good news is that people generally do have diverse cognitive skills We just need to ensure that they apply those skills when making a prediction, for example, about which option maximizes shareholder value. When they do, they can for present purposes be said to independently make a prediction—even if they did not initially receive "independent signals" about which option maximizes shareholder value.
What do we mean by cognitive diversity? When there is cognitive diversity, it means that people facing an issue differ in the analytical steps they take to arrive at a prediction. They may look at different dimensions of the same issue; they may come to different interpretations of what they see even when they look at the same dimension; and by using different prediction models, they may come to different predictions even if they share an interpretation.9
To see why different predictions are beneficial, let's return to the ABN Amro case. The suggestion that Fortis shareholders may have attached too much weight to management's track record of prior acquisitions implicitly assumed that in making a prediction about whether the acquisition of ABN Amro would maximize shareholder value, each shareholder focused on management's track record. But when there is cognitive diversity that is not necessarily the case. While some shareholders may look at management's track record, others may look at different dimensions of the issue, such as the resulting financing burden. When an investor focuses on management's track record and sees a strong track record, he may predict that the acquisition will be a success. But when an investor focuses on financing and sees a heavy burden, he may predict that the acquisition will be a failure. The predictions will be negatively correlated; when one investor is wrong, the other will be more likely to be right. Hong and Page demonstrate, mathematically, that negatively correlated individual judgments result in more accurate collective judgments.10
The bottom line is that in assessing independent voting as a mechanism of voting efficiency, we should focus on phenomena that may cause diversity breakdowns. The remainder of this section discusses three such phenomena: correlated biases, information cascades, and opinion leaders.