Abstract

We study an agent’s incentives to discover where her talents lie before putting them to productive use. In our setting, an agent can specialize and learn about the same type of talent repeatedly, or experiment and learn about different types of talent. When talents are normally and symmetrically distributed we find that experimentation is efficient, regardless of one’s initial draw of talent. Competitive labor markets encourage experimentation whereas monopsonistic labor markets induce specialization. Relaxing our assumptions of normality and symmetry
in the distribution of talents, and allowing for human capital acquisition, provides a role for specialization in discovering talents.