Does the Running Variable Matter? A Second Look at Discontinuity Designs for Evaluating Regional Economic Development and Business Incentive Policies

Abstract

Regional economic development and business incentive programs have a prominent role in the European Union (EU). For evaluating these programs, in recent years, a growing number of studies have exploited either spatial discontinuities, set by boundaries of the targeted areas, or ranking discontinuities, based on EU-fund eligibility indexes or firm-level application scores. In light of this literature, impact evaluations are being increasingly commissioned and designed under an a-priori assumptions that discontinuity designs have superior impact identification properties. This paper argues that in a number of frequently encountered, but often unrecognized, circumstances this assumption does not hold ground. When the running variable has a weak influence on the outcome of the analysis, discontinuity designs are at risk of either unnecessarily reduce external validity or, in the presence small sample sizes, failing to achieve the complete balancing of relevant controls. In this scenario, ensuring the common support for the crucial confounders and adopting statistical matching estimators, often constitute a more viable empirical option.