Truth by consensus : a theoretical and empirical investigation

Truthful reporting about the realization of a publicly observed event cannot be guaranteed by a consensus process. This fact, which we establish theoretically and verify empirically, holds true even if some individuals are compelled to tell the truth, regardless of economic incentives. We document results from an experiment where subjects routinely misreported a commonly known event when they could monetarily gain from it. Relying on majority consensus did not help uncover the truth, especially if complying with the majority granted small personal monetary gains. This highlights the difficulties in relying on shared consensus protocols to agree on specific events, and the importance of institutions with trusted, impartial observers.