Targeting and Effectiveness of Location‐Based Policies

This paper provides new insights into the design of effective location‐based policies. In the context of European regional policy, we apply machine learning algorithms to predict regions likely to underutilize EU funding and identify the main determinants of their low absorptive capacity. Using a regression discontinuity design (RDD), we show that EU funds are less effective in regions with low predicted absorptive capacity, while boosting output and employment in high‐capacity regions. Our analysis further reveals that these low‐capacity regions do not necessarily coincide with the least developed ones and are not constant over time, varying across programming waves and depending on the specific funds considered. This highlights the role of institutional and administrative factors beyond income levels. The proposed approach enables early identification and targeted interventions to enhance regional spending capacity using publicly available data, offering a practical tool for improving policy design and implementation.