Exploring the scope of neurometrically informed mechanism design

Abstract

A basic goal in mechanism design is to construct mechanisms that simultaneously satisfy efficiency, voluntary participation, and dominant strategy incentive compatibility. Previous work has shown that this is impossible, unless the agents and planner have sufficient information about each other and common knowledge. These results have remained largely theoretical because the required information is generally not available in practical applications. However, recent work has shown that these limitations can be overcome in simple settings, using neurometric technologies that provide noisy signals of subjects' preferences that can be used in the mechanism design problem. Here we build on this work by carrying out two new experiments designed to test the extent to which these Neurometrically Informed Mechanisms (NIMs) can be applied to more complicated and realistic environments. We find robustness to large type and action space and to the degrees of loss and risk-aversion observed in most of our sample.

Files

Additional details