Title

Keeping Your Story Straight: Truthtelling and Liespotting

Author(s)

Johannes Hörner Johannes Hörner (Yale University)

Xiaosheng Mu Xiaosheng Mu (Harvard University)

Nicolas Vieille Nicolas Vieille (HEC Paris)

Abstract

An agent privately observes a Markov chain online and reports it to a designer. To what patterns in the reported data should the designer pay attention? We show that, in general, keeping track of the empirical frequency of transition counts in the agent's reports is insufficient, despite the true state being Markovian. Nonetheless, we derive conditions under which any deviation that can be distinguished from truth-telling by checking the frequency of strings of an arbitrary (finite) size can be detected by "checking pairs." Further, we find that some undetectable deviations cannot be profitable, independent of the agent's preferences. Hence, we provide weaker sufficient conditions that ensure that the agent finds honesty to be the best strategy. We explore the implications of these results for the literature on (i) linking incentives, (ii) dynamic implementation, and (iii) repeated games and agency models.

Creation Date

2017-04

Section URL ID

Paper Number

2017-5

URL

https://uploads.strikinglycdn.com/files/3a60163d-859d-4a1b-8086-04354e0519c2/Truthtelling_170427.pdf

File Function

Jel

C72, C73

Keyword(s)

Detectability, testing, Markov chains, implementation, repeated games

Suppress

false

Series

13