A simple model of project selection with strategic communication and uncertain motives

I analyze a problem of project selection where two agents, privately informed of both the true value and their bias in favor of their alternatives, make non-verifiable proposals to an uninformed decision-maker. The analysis makes two contributions. First, I examine the consequences of preference uncertainty for the information aggregation problem and show that the relative attractiveness of allowing a given agent to cast the decisive vote is increasing in the stability of his preferences. Second, I analyze the costs and benefits of pre-communication disclosure of the realized biases. I find that, conditional on the decision structure, disclosing the bias of the decisive agent is never optimal while disclosing the bias of the other agent is always optimal. This result makes an asymmetric treatment of the agents in terms of disclosure policies often optimal, despite the additional benefit of disclosure, which is the ability to condition the decision-making structure itself on the disclosed biases.