Persuasion Meets Delegation

Abstract

Abstract. There are two common ways for a principal to influence the decision making of an agent. One is to manipulate the agent's information (persuasion problem). Another is to limit the agent's decisions (delegation problem). We show that, under general assumptions, these two problems are equivalent; so solving one problem solves the other. We illustrate how the methods developed in the persuasion literature can be applied to address unsolved delegation problems by considering monopoly regulation with a participation constraint. JEL Classification: D82, D83, L43Keywords: persuasion, delegation, regulation Date: 7th April 2018.Kolotilin: School of Economics, UNSW Business School, Sydney, NSW 2052, Australia. E-mail: akolotilin@gmail.com. Zapechelnyuk : School of Economics and Finance, University of St Andrews, Castlecliffe, the Scores, St Andrews KY16 9AR, UK. E-mail: az48@st-andrews.ac.uk.We are grateful to Tymofiy Mylovanov, with whom we are working on related projects. We thank

Extreme Points and Majorization: Economic Applications

“…In recent, independent work, Arieli et al (2020) also studied a Bayesian persuasion problem via an extreme points approach and considered maximization on a majorizing set of functions. 6 Our majorization approach thereby clearly reveals the close connection between delegation and Bayesian persuasion and their respective optimal mechanisms, and shows that the equivalence between delegation and persuasion mechanisms obtained for a subset of mechanisms by Kolotilin and Zapechelnyuk (2019) extends to all randomized mechanisms. We also illustrate how results obtained in one strand can be immediately applied to the other.…”

“…REMARK 2: Certain Bayesian persuasion problems give rise to the same class of optimization problems (see Section 4.4 below), and this allows us to extend the equivalence observed in Kolotilin and Zapechelnyuk (2019) to stochastic delegation and to general persuasion mechanisms. As an illustration of this equivalence, we now provide a sufficient condition for a deterministic delegation mechanism to be optimal by applying a result in Dworczak and Martini (2019) 45 Our result also extends a result by Kovac and Mylovanov (2009).…”

“…Hence, the basic structure of their respective optimal mechanisms is identical. Kolotilin and Zapechelnyuk (2019) have recently established a formal equivalence between optimal delegation and Bayesian persuasion for the case where the set of policies for the principal was exogenously restricted to deterministic delegation mechanisms and to monotone partitional signals, respectively. Our majorization characterization immediately implies that this equivalence holds without any restrictions on the policy space: optimal signal structures for Bayesian persuasion that are not monotone partitional correspond to randomized optimal delegation mechanisms.…”

“…In particular, we obtain analogous results if and for a strongly concave function b . Closely related utility functions have been used, for example, by Amador and Bagwell (2013) and Kolotilin and Zapechelnyuk (2019).…”

Bayesian persuasion and information design: perspectives and open issues

“…For example, Dworczak and Martini (2019) explain how price theory can be used to interpret and analyze Bayesian persuasion. Kolotilin and Zapechelnyuk (2018) establish an equivalence result between Bayesian persuasion and delegation and show how the former can be used to improve upon results on the latter. Kleiner et al (2020) explain how the same mathematical technique can be applied to several different economic problems, including auction and contest design, Bayesian persuasion, delegation, and decision making under uncertainty.…”