Assessing the Impact of Financial Aids to Firms: Causal Inference in the Presence of Interference

We consider the problem of estimating causal effects of a policy intervention implemented in Tuscany (a region in Italy) on handicraft firms accounting for the presence of possible spill-over effects. In the presence of spill-over effects, the no-interference assumption, usually invoked in the potential outcomes literature, is violated. We address this problem, by explicitly modelling interactions among units. This approach involves specifying which firms interact with each other, and model the magnitudes of these interactions. We allow intensity of interactions to depend on a distance metric, based on geographical distance and firms’ characteristics. Our results show that the impact of interference can be considerable and should not be neglected.