Zeynep Elburz (p), Özge Ekinci

Davide Luca

Impact assessment; Social Return on Investment (SROI); Cost-benefit analysis; Social economy organizations

Francesco Scotti, Carlo Caporali, Davide Luca (p)

Ilias Kostarakos

To respond to the social and economic challenges that arose during the COVID-19 pandemic, the European Union implemented NextGenerationEU, the largest-ever stimulus package undertaken in the EU, currently worth over €650bn (at current prices). National Recovery and Resilience Plans (RRPs) are at the heart of such a package and provide member states with performance-based loans and grants to support structural reforms and investment. Plans must allocate at least 57% of their budget to green and digital transition measures, while also increasing cohesion and resilience.
Despite the policy tool's substantial size and potential, empirical evidence on how these funds are being targeted and spent is still limited. Focusing on Italy, one of NextGenerationEU's largest beneficiaries, we address two research objectives to contribute to filling this gap.
We first explore the territorial distribution of funds from Italy’s national government, in charge of overall coordination, to local administrations, by identifying the key local characteristics explaining participation in Italy’s RRP. Featuring a double-stage Heckman correction model and a dataset covering Italy’s almost 8,000 municipalities, we consider the association between funds allocated to municipal governments and variables capturing local administrative, financial, and macroeconomic characteristics. We find that funds are primarily allocated to Southern, relatively worse-off urban municipalities with better administrative capacity. Previous experience with EU Cohesion Policy funds also determines the allocation of RRP funds. The fund allocation process across different missions is also affected by the pre-existing level of infrastructures and services in the underlying sectors.
Second, we discuss whether recipient territories adopt specialisation strategies—i.e., they attract funds in policy areas where they already have strong capacity—or invest in policy domains where they have lower levels of expertise. We observe a process of specialisation in the Digital, Education, and Healthcare missions. By contrast, we find a pattern of diversification in the Social mission, in line with the convergence mechanism of the EU Cohesion Policy.
Overall, while these results suggest that monies are distributed relatively effectively, they also underscore how some of the country's weakest territories may struggle to exploit the opportunity offered by NextGenerationEU, potentially further enhancing spatial divergence.

Ilias Kostarakos (p), Angelos Alexopoulos, Petros Varthalitis, Christos Mylonakis

Zeynep Elburz

Recent empirical research has attempted to shed light on the impact that Cohesion policy, the largest investment initiative of the EU, exerts on regional economic performance. Despite the ever-increasing volume of research contributions, the literature has still not reached a consensus regarding the sign of the impact and its (statistical) significance.
The aim of this paper is to provide an ex-post evaluation of the regional effects of Cohesion policy, focusing on one of the least studied instruments: the Cohesion Fund. Using a novel dataset that covers the entire universe of Cohesion Fund expenditures since its inception, we draw on recent developments from the causal inference literature in order to estimate region- and time-specific treatment effects. Specifically, we adopt the matrix completion approach, a generalization of the synthetic control method of Abadie et al. (2010) that relaxes some key assumptions of well-established causal inference methods, like the differences-in-differences approach.
We find that, on average, the Cohesion Fund exerts a positive and persistent effect on the level of GVA per capita. The majority of the impact materializes within the first seven years that the region is under treatment. However, the region-specific analysis highlights that the results are quite heterogeneous, with the relatively poorer regions being the ones that exhibit the largest in magnitude effects. Moreover, we uncover a non-linear, inverted U-shaped relationship between the treatment intensity (Cohesion Fund expenditures as a share of output) and the size of the treatment effect. Lastly, our results indicate that –on average- recipients of the Cohesion Fund grew at a faster pace compared to the counterfactual scenario in which they do not receive the funds.