We examine communication between an informed sender and an uninformed receiver with a presence of a strategic fact-checker. The sender makes a claim about an issue to persuade the receiver to approve the sender's proposal. The fact-checker has its own goal and chooses a stochastic fact-checking policy that checks sender's claims. Checking a claim is costly and, with some probability, can fail to verify whether the claim is true or false. Full fact-checking is optimal when the cost is below a threshold. Otherwise, no fact-checking is optimal. We characterize the cost threshold as a function of fact-checker's preferences.
The receiver need not prefer a fact-checker with preferences aligned with the receiver to one with opposed preferences. Adding multiple fact-checkers does not necessarily improve communication even when all fact-checkers are willing to fully check by themselves. For intermediate cost of checking, having multiple fact-checkers can lead to underprovision of fact-checking due to free riding.
Many online platforms intermediate trade between sellers and buyers using data records of the buyers' personal characteristics. How much value do such intermediaries derive from each record? Is this value higher for specific buyers? What are its properties? We find that an important component of the value of a data record is a novel externality that arises when a platform pools records to withhold information from the sellers. Ignoring this externality can significantly bias our understanding of the value of data records. We then characterize a platform's willingness to pay for more data, thereby establishing a series of basic properties of the demand side of data markets. Our analysis combines modern information design with classic duality methods and applies to a large class of principal-agent problems.
This paper presents the optimal editorial policy for state-owned media who can manipulate information
flow from a strategic informed elite to an uninformed receiver. The receiver attempts to match the state
of the ruler’s competence with a binary action. If the elite’s and audience’s preferences are too distant
from each other, then the editorial policy is uninformative. Otherwise, the media signal whether the
state is higher or lower than a threshold which depends on the elite’s preferences. The media benefit
from a more lenient elite, as long as the elite is not too lenient. The media are worse off when the receiver
is more critical of the ruler, whereas the elite generally is better off when the receiver is more critical.
When the receiver has private information about how critical he is, I characterize the lower bound on
the media’s payoff obtained within the class of restricted editorial policies. I identify a condition on the
distribution of receiver’s private information that implies the media’s payoff attains this lower bound.
Work in Progress
Subjective Uncertainty and Contract Dissolution (with Renee Bowen and Malte Lammert)