Optimal incentive schemes when only the agents' ''best'' output matters to the principal

Standard principal-agent models assume that the principal's payoff is a function of the total output of all agents. In many real-world situations, however, the principal's payoff is based solely on the ''best'' of the agents' outputs (e.g., the first agent to make an innovation, the most creative advertising campaign, or the cheapest product design). The results obtained from such a model differ from the standard results in a number of respects. For instance, even when identical agents perform identical tasks, the optimal incentive scheme will often differ across agents. Also, the principal may want to increase the variance of the agents' output or reduce the correlation of output across agents, even when the agents are risk averse.