Strategic Learning with Asymmetric Rationality

TL;DR

This paper models a strategic interaction where a fully rational sender controls information flow and a boundedly rational receiver with memory constraints learns and incentivizes decision-making, revealing systematic behavioral patterns.

Contribution

It introduces a framework for analyzing strategic communication with asymmetric rationality and bounded memory, characterizing optimal protocols and behavioral phenomena.

Findings

Behavioral patterns like disengagement and polarization emerge systematically.

Optimal protocols balance information control and incentive costs.

The model captures procedural rationality in strategic interactions.

Abstract

This paper analyzes a dynamic interaction between a fully rational, privately informed sender and a boundedly rational, uninformed receiver with memory constraints. The sender controls the flow of information, while the receiver designs a decision-making protocol that uses a finite state space to learn and to provide incentives. We characterize optimal protocols and quantify the scope for manipulation and the incentive cost of guarding against it. We show that distinctive behavioral patterns that might otherwise appear erratic or psychologically driven -- such as information disengagement, opinion polarization conditional on the same information, and indecision near the decision point -- emerge as systematic equilibrium responses to asymmetric rationality and information. The model provides an expressive framework for procedural rationality in strategic settings.