Strategic Communication in Legislative Bargaining: An Experimental Study

Abstract

We experimentally study the impact of communication in collective bargaining. A proposer makes an initial policy offer to two voters who have private information about their ideal policies. In symmetric Perfect Bayesian equilibrium, through their ballots the voters (“senders”) reveal some preference clues valuable to the proposer (“receiver”). We examine two communication games. In Tentative, the initial proposal is cheap talk and a final proposal will always follow irrespective of the vote outcome. If the final proposal garners sufficiently many positive votes given the voting rule (i.e., majority or unanimity voting, where the proposal counts as positive vote), the final proposal supplants the status quo policy. In Definite, by contrast, an initial proposal with sufficient support is instantly implemented as the new policy, and otherwise a final proposal and ballot decide the next policy. Many of our predictions find strong support in the laboratory. First, dishonest negative voting is found approximately in the region of ideal policies where it is expected to induce more favorable final offers for the voters. Second, proposers usually make more compromising final offers after observing too few positive votes. Finally, the observed strategic communication among the proposer and voters does indeed help to avoid welfare-reducing policy gridlock and thus to improve their average payoffs, especially so for the proposer.