Relational Contracts and the Theory of Autocracy

Abstract

Standard organizational economics has studied the problem of a supervisor (principal) incentivizing effort from a subordinate (agent) without using perfectly enforceable contracts---instead, ``relational contracts". However, this literature has always taken for granted which player is the ``principal" and which players are ``agents", something that cannot be disregarded in dictatorships. Without a sufficiently strong and interested third-party to backup the principal's role at the rightful principal, agents have a nontrivial opportunity to usurp that position of power away from the principal. In many cases, this is their only viable outside option. If they were to quit on the dictator without challenging their power, that may be such a betrayal that leads to certain death/imprisonment. We study the problem of a dictator searching for heterogeneously competent agents and engaging in relational contracting with them, with the backstop to punish ruler deviations being an attempted coup, which with some probability leads that agent to becoming the new principal and restarting the game.