Estimating the causal impact of an intervention on efficiency in a dynamic setting

Abstract

This paper develops a novel methodology to estimate in a dynamic setting the causal impact of a policy on efficiency. Classical efficiency techniques evaluate multidimensional performance but ignore the endogeneity issues in policy evaluations. We develop an indicator which accounts for the dynamic performance of the observations and for the possible correlation between the treatment status and the efficiency score. Besides, we propose a decomposition of the indicator to disentangle the effect of the policy on the performance of the observations from the effect of the policy on the environmental harshness that the observations have to face. This innovative design allows us to introduce the notion of causality in efficiency studies and to shed light on the mechanisms underlying the inefficiency at the unit and policy level. In the application, the present study assesses the impact on the efficiency of a funding program that aims to foster educational equality.