Job-training programmes with low completion rates: The case of Projoven-Peru

Projoven is an ongoing Peruvian youth job-training programme started in 1996and conducted in successive waves or ‘public calls’. The programme has twophases of instruction: i) three months of classroom training, followed by ii) threemonths of on-the-job training internship. From 1997 to 2007, less than 60% ofregistrants completed both phases. This paper estimates the effectiveness ofProjoven’s sixth Public Call, in terms of overall and formal employment andwage outcomes, while accounting for the presence of trainees with partialinstruction. Four groups are compared pairwise: the control group, the dropoutgroup with only classroom training, the dropout group with classroom trainingand internship placement, and the group completing the full programme. Ourestimations account for selection effects and the potential endogeneity of trainingcompletion. We find that in terms of overall and formal employment, theprogramme yields returns only in the short run and conditioned on internshipcompletion. The programme seems to increase trainee wages, but these effectsdiminish over time. Projoven is relatively more effective for individuals with nowork experience prior to enrolment.