An expert can covertly acquire information about the state of the world before communicating with a decision maker in order to influence her action. The expert's information acquisition is unrestricted and costless but her ability to prove to the decision maker what she privately learnt is limited. I study how the verifiability of the expert's acquired information affects equilibrium information acquisition and transmission. Even when acquired information is only partially verifiable, I prove an unravelling result: all equilibria in which the expert influences the decision maker involve full revelation of the expert's private information. I compare different verifiability environments and prove a comparative statics result, characterizing the sense in which more verifiability is never worse for the expert. I then study optimal verifiability environments, giving necessary and sufficient conditions for optimality for each of the two agents. Expert-optimal environments are credibly rich in the sense that, even when facing a sceptical decision maker, the expert has access to a rich language to communicate her information. I show that this is akin to her having a large amount of commitment power. The optimum for the decision maker restricts the expert’s ability to credibly communicate intermediate results, inducing the expert to acquire and disclose full information in equilibrium.
I investigate the effect of cheap talk in sender-receiver games where the sender wants the receiver to always take the same action, irrespective of the state of the world, while the receiver’s preferences are state-dependent. These games model many economic situations such as investment, voting or purchase decisions. I consider environments where the interests of the players are either more likely aligned, or more likely conflicting. Using a simple theoretical analysis I show that communication can harm in settings where conflict of interests is less likely. This is due to increased skepticism resulting from receivers’ aversion to being deceived (“sucker aversion”). However, when interests are more likely conflicting, communication can help due to senders’ lying aversion. I run experiments to test these predictions and find that communication has a significantly positive effect on payoffs when interests are more likely conflicting. This cannot be explained by lying aversion only; social preferences seem to also matter. On the other hand, when interests are more likely aligned, opposite to previous findings, communication does not significantly increase receivers’ skepticism and it does not affect payoffs.
We investigate the welfare effect of disclosure of conflicts of interest when an
expert advises a client. In a model with verifiable information and uncertainty
about the direction of the expert's conflict of interest and the informedness of
the expert, we show that disclosure of the expert's bias is counterproductive
when the magnitude of the expert's bias is not too large and the likelihood of
the expert being informed is low. Moreover, the harm of disclosing the expert's
conflict of interest is more significant when there is a larger uncertainty about
the nature of the expert's conflict of interest.
We study markets where the seller can acquire conclusive evidence about the quality of his asset by conducting costly tests. We consider both the cases of uninformed seller and (perfectly) informed seller and identify the policy that maximizes market information for each case. A regulatory policy determines whether (i) the seller's test activities are observable by the market and (ii) the disclosure of evidence obtained is mandatory. If the quality of the asset is unknown to the seller, covert testing and voluntary disclosure incentivize the seller to acquire evidence and constitute an optimal policy. In contrast, if the seller is informed, overt testing and mandatory disclosure together enable signaling and maximize market information. Our results suggest that different regulatory practices observed in different markets could be explained by difference in sellers' private information.