Beyond the SUTVA: how industrial policy evaluations change when we allow for interaction among firms

Developed countries have used several place-based policies to address the socioeconomic underdevelopment of their lagging regions ranging from tax exemptions to soft loans. Among these place-based policies, investment subsidies to private firms are one of the most popular in the EU. However, the empirical evidence to date is mixed and there is no general consensus on the effectiveness of such policy. Most evaluation studies have focused on the policy impact on subsidized firms, whereas the possible spillovers on other firms have been mostly overlooked. This is due to the dependence of these analyses on the Stable Unit Treatment Value Assumption (SUTVA), i.e. they assume away any possible interaction among firms. There are several situations in which this assumption is not plausible; however, there are severe empirical difficulties in disentangling the spillover effects from more relevant confounding factors. In this paper we propose evaluation strategies capable of detecting potential positive and negative spillovers. In presence of spillovers, the use of eligible but unsubsidized firms as control group will yield biased ATT estimates. We propose to build a counterfactual scenario drawing firms from the pool of firms located in non-assisted areas as similar as possible to the eligible areas. In addition to the ATT, our approach allows to estimate spillover parameters that contrast the positive agglomeration effect with the negative cross-sectional substitution and the crowding-out effect. Econometrically we adopt a Matching difference-in-differences (DID) using the recent coarsened exact matching (CEM) that dominates commonly used existing matching methods in its ability to reduce imbalance, model dependence, and bias. Our application concerns the Italian Law 488 (L488) and contrasting positive and negative spillover effects we find positive spillovers in terms of investments and negative spillovers in terms of employment, but these estimates are not statistically significant. This demonstrates that the positive effects of the subsidies on the subsidized firms' growth in terms of investments and turnover are not engendered to the detriment of the unsubsidized firms. The main methodological contribution to the literature is our novel evaluation strategy that can be adopted in other contexts to evaluate the micro spillovers of other policy instruments.