Information-signaling in pre-trial bargaining: An experiment

This paper provides experimental evidence on the effectiveness of self-penalizing commitments as signaling mechanisms in pre-trial bargaining under asymmetric information, aimed at reducing trial rates. Using an online experiment (N=2,041), we design a novel two-type signaling game with asymmetric information, where one party possesses private information regarding the strength of its case, and proposes a take-it-or-leave-it settlement offer to an uninformed party. In the baseline scenario, the uninformed party must decide whether to accept the offer, or reject it and proceed to a costly trial. In our novel signaling treatments, the informed party can credibly commit ex ante to a costly monetary penalty contingent upon losing at trial. Consistent with theoretical predictions, our results show that self-penalizing commitments reduce trial rates. Informed parties are more likely to commit when they have strong cases; and uninformed parties interpret the commitment as a signal of merits, and are more likely to settle less-generous offers when these are coupled with a commitment. However, we also document departures from theory: effect magnitudes are generally smaller than expected; the commitment device is both under- and over-used across offer ranges; and, when committing is possible, buyers sometimes penalize its omission by rejecting otherwise generous offers.