Counterfactual Impact Evaluation of Cohesion Policy. Multi-Period Regression Discontinuity Design with New Data and Robustness Checks

Regression discontinuity designs (RDD) are widely used both in the academic literature on the evaluation of Cohesion Policy and in DG REGIO's ex-post evaluations. This choice is motivated by the 75-percent-benchmark that is applied to regional GDP per capita in pre-specified years to determine the NUTS-2 regions' Objective 1 or, later, "less developed" status, which allows for greater funding opportunities. RDD as a quasi-experimental method to estimate causal effects of cohesion policy exploits this threshold by assuming that it is "as good as random" whether a region has a GDP per capita slightly below or above that threshold. A recent RDD analysis conducted in the course of the ex-post evaluation for the 2014-2020 programming period (Huber et al., 2025), however, found that the funding intensity (funding per capita) did not differ significantly between regions with GDP per capita slightly below and above the 75 percent threshold relative to the EU average. This is a highly relevant result when it comes to analysing the EU cohesion policy's effectiveness over time. The objective of this project, therefore, is to investigate the development of the difference in treatment intensity around the 75 percent cut-off using updated data for the 2014-2020 period as well as for programming periods since 1989, and what that implies for the application of counterfactual impact evaluation methods in the context of cohesion policy.