The Event-History Approach to Program Evaluation

This paper studies the event-history approach to microeconometric programevaluation. We present a mixed semi-Markov event-history model, discussits application to program evaluation, and analyze its empirical content.The results of this paper provide fundamental insights in what can be learnedfrom longitudinal micro data about, for example, the effects of training programsfor the unemployed on their unemployment durations and subsequentjob stability. They can guide the choice of particular models and methods forthe empirical analysis of such effects.<p>This discussion paper resulted in a publication in: D.L. Millimet, J. Smith, and E. Vytlacil (Eds.), chapter 2 in '<I>Advances in Econometrics</I>', 2008, volume 21: Modeling and Evaluating Treatment Effects in Econometrics, 33-55, Elsevier Science, Amsterdam.