Active Labor Market Policy Effects in a Dynamic Setting

This paper develops and implements a method to identify and estimate treatmenteffects in a dynamic setting where treatments may occur at any point in time. Bycombining the standard matching approach to the timing-of-events approach, itdemonstrates that the effect of the treatment on the treated at a given date can beidentified although non-treated may be treated later in time. The approach builds on a"no anticipation" assumption and the assumption of conditional independence betweenthe duration before treatment and the duration before exit. To illustrate the approach,the paper studies the effect of training for unemployed workers in France, using a richregister data set. Training has little impact on unemployment duration. Thecontamination of the standard matching estimator due to later entries into treatment islarge if the treatment probability is high.