This paper considers the human side of mechanism design, the behavior of economic agents in gathering and processing information
and responding to incentives. I first give an overview of the subject of mechanism design, and then examine a pervasive premise
in this field that economic agents are rational in their information processing and decisions. Examples from applied mechanism
design identify the roles of perceptions and inference in agent behavior, and the influence of systematic irrationalities
and sociality on agent responses. These examples suggest that tolerance of behavioral faults be added to the criteria for
good mechanism design. In principle-agent problems for example, designers should consider using experimental treatments in
contracts, and statistical post-processing of agent responses, to identify and mitigate the effects of agent non-compliance
with contract incentives.
Keywords Mechanism_design - Principal-agent_problem - Incentives - Agent_compliance
JEL Classification D000 - D600 - D610 - D710 - D800 - C420 - C700
This paper was first presented in the Jean-Jacques Laffont Lecture, Toulouse, October 2006, and a condensed version was presented
on the occasion of Leo Hurwicz’s 90th birthday, April 2007. I am indebted to Eric Maskin for useful comments, to Florian Heiss,
Charles Manski, Rosa Matzkin, and Joachim Winter for collaborative research on control of response errors in economic surveys,
and to the National Institute on Aging for research support.
An erratum to this article can be found at
http://dx.doi.org/10.1007/s10058-009-0085-8