Overeducation, defined as having more schooling than what is necessary for the job, is a labor market phenomenon that has special incidence in Spain. In principle, overeducation needs not to be a problem if it is transitory. But when it becomes persistent it can have important negative effects in the economy. In this paper I use duration models to investigate the persistence of overeducation using data from the Spanish Continuous Sample of Employment Histories (Muestra Continua de Vidas Laborales or MCVL, in Spanish). Controlling for observable characteristics, I find evidence for the hypothesis that overeducation is indeed a persistent phenomenon in Spain. Further research will be carry out to control for unobserved heterogeneity.