Treatment Evaluation in the Case of Interaction Within Markets

We extend the standard evaluation framework to allow for interactions betweenindividuals within segmented markets. An individual's outcome depends not only onthe assigned treatment status but also on (features of) the distribution of the assignedtreatments in his market. To evaluate how the distribution of treatments within amarket causally affects the average effect within the market, averaged over the fullpopulation, we develop an identification and estimation method in two steps. The firstone focuses on the distribution of the treatment within markets and betweenindividuals and the second step addresses the distribution of the treatment betweenmarkets. We apply our method to data on training programs for unemployed workersin France. We use a rich administrative register of unemployment and training spellsas well as the information on local labor demand that is used by unemploymentagencies to allocate training programs. The results show that the average treatmenteffect on the employment rate causally decreases with respect to the proportion oftreated in the market. Our analysis accounts for unobserved heterogeneity betweenmarkets (using the longitudinal dimension of the data) and, in a robustness check,between individuals.