Séminaire COURNOT – Rafael Treibich (Univ. of Southern Denmark)

Détails de l'événement :

“Achieving Consensus Under Majority Voting”

Résumé : Consensus decision-making, whereby collective decisions are adopted unanimously without an explicit vote, is ubiquitous in international organizations, boardrooms, expert advisory councils, and many other types of committees. Yet, in most cases, consensus is not formally required ”by law”, but follows instead from a norm of collective behavior. The prevalence of consensus may seem puzzling, especially in large committees, where preferences over decisions are rarely unanimous. In practice, however, consensus appears to be an effective way of making decisions, and does not suffer from the gross inefficiency associated to respecting unanimous preferences. This apparent contradiction can be explained by the repeated nature of collective decisions, which greatly alleviates the cost of achieving consensus. In order to investigate the possibility of consensus decision-making, and how it relates to the underlying formal voting rule, we consider a model of repeated collective decision-making under complete information. At the optimal equilibrium, agents engage in a form of implicit logroll, sometimes voting against their preference in order to achieve the efficient outcome. The repeated interaction thus generates a higher degree of consensus than implied by the underlying preferences. Increased levels of consensus can be achieved at equilibrium either by rejecting proposals that would not generate the support of all countries (“forcing” consensus), or by allowing countries to reduce the scope of the proposed reform (consensus through “compromise”). In both cases, we investigate the optimal equilibrium as a function of the formal voting rule and discount factor. We illustrate our results in a simpler model where utilities can be additively separated into a fixed private component and a variable public component.