Recent experimental evidence suggests that some people dislike telling lies, and tell the truth even at a cost. We use experiments as well to study the socio-demographic covariates of such lie aversion, and find gender and religiosity to be without predictive value. However, subjects’ major is...

We present a game-theoretical model that accounts for abundant experimental evidence from games with non-binding communication (‘cheap talk’). It is based on two key ideas: People are conditionally averse to break norms of honesty and fairness (i.e., the emotional cost of breaking a norm is...