Supplement to “ Communication and influence ”

In the baseline version of the game, (D k s), agents invest in communication before they learn the value of their local state. In the alternative time line, θ(D k s) (mentioned at the end of Section 3), agents choose communication intensities after they observe their local states. In this section, we show that both versions of the game have the same perfect Bayesian equilibrium. From Theorem 1, we know that the baseline game (D k s) has an equilibrium in linear decision functions ai = biiθi + ∑ j =i bijyij