Like the first-mover in an ultimatum game, the principal is a first-mover with foreknowledge of the agent’s “rational” best response function. The solution to the “principal’s problem” is to choose a contract that maximizes the principal’s expected profit given the agent’s marginal efficacy and marginal effort cost. However, this paper reports experiments that show that principals make large concessions toward an equal division outcome. As in ultimatum games, agents are at times willing to punish principals who are perceived as being overly acquisitive. Variations in agent effort cost and effectiveness that should (theoretically) produce qualitatively different game-theoretic equilibria have little impact on outcomes.