Persuasion via Weak Institutions: A Comparative Analysis of Ex-post IR Bayesian Persuasion and Credible Persuasion

In this article, we explore the concept of Bayesian persuasion in economic models, specifically focusing on the design of optimal information disclosure policies. Bayesian persuasion allows a sender to commit to sending any distribution of messages as a function of the state of the world, which can be useful in various scenarios. However, full commitment power is not always reasonable or possible in real-world situations due to the receiver’s inability to monitor the sender’s signaling scheme. To address this limitation, we propose a model that considers partial commitment and compare it to credible persuasion, another concept in economic models.
We begin by defining Bayesian persuasion and its key assumptions. In this context, the sender chooses a signaling scheme ϕ and sends it to the receiver, who can observe the distribution of signals induced by ϕ. We then examine the utility of both the sender and receiver under different scenarios. Next, we assume that Assumption 2.1 holds and show that VexpostBP⪰VCheaptalk, indicating that our model falls within the scope of Bayesian persuasion with partial commitment.
To further understand this concept, let’s consider an analogy. Imagine you’re trying to convince your friend to try a new restaurant for dinner. You could send them a signal (e.g., a positive review) that represents your belief about the restaurant’s quality. If they can observe this signal, they may be more likely to trust your recommendation and choose to dine there as well. However, if you cannot observe their reaction or response to your signal, it becomes more challenging to commit to sending a specific signal without worrying about being caught in a lie.
In economics, this concept of partial commitment is essential in understanding how senders and receivers interact in various scenarios. By considering the receiver’s perspective and the limitations of full commitment power, our model provides a more realistic and practical approach to understanding Bayesian persuasion.