Aproximación con algoritmos evolutivos de la frontera de Pareto de un modelo dinámico de agente-principal con acciones discretas

We use Evolutionary Algorithms to approximate the Pareto frontier of a dynamic Principal-Agent model with discrete actions. The Pareto frontier that we obtain is concave, a result that is due to the asymmetry of information between the Principal and the Agent. As we move in the Pareto frontier towards contracts that favor the discounted expected utility of the agent, we obtain compensation schemes in which the Principal primarily assumes the risk inherent to the productive activity. Given changes in the discount factor of the Principal and the Agent and in the relative cost of exerting an additional effort unit, the Pareto frontier preserves its topological characteristics. When the Principal and the Agent are more patient, both of them obtain higher values of discounted expected utility and the relationship generates higher economic surplus. When it is costlier to exert an additional effort unit, the Principal tends to bear the majority of the risk inherent to the productive activity along the Pareto frontier.