Alternative approaches for counterfactual evaluation in the presence of a few units

The dissertation is part of the flourishing literature that estimates the causal effect of policy intervention in panel data settings with only one or a few treated and control units. In a context where many policy interventions occur at an aggregate level, the synthetic control method (SCM) literature assumes a focal role in the causal effect estimation. The thesis analyzed the SCM features, its critical issues and presented alternative approaches for the counterfactual evaluation in the presence of a few units, shedding light on how recent methodological advances contributed to this literature. The first chapter proposes a novel estimation approach that allows including units potentially affected directly or indirectly by an intervention in the donor pool: the inclusive synthetic control method (iSCM). The new methodology proposed concerns scenarios in which some of the units might be affected by spillover effects and scenarios in which there are multiple treated units and the control group is scarce and composed of units no so similar to the treated ones. The iSCM allows us to include these units safely in the donor pool and then eliminate post-intervention effects. Moreover, it allows us to estimate the potential presence of spillover effects. The second chapter illustrates how to use iSCM by estimating the economic impact of Germany reunification on West Germany's GDP per capita, allowing for spillover on Austria. The third chapter applies a recent methodological extension of the SCM, the Kernel Balancing, that improves estimation accuracy and is applicable in a wider set of contexts. Particularly, the chapter aims to estimate the regional economic impact of joining the euro area for the latecomers, i.e., the countries that adopted the euro after 2002. The use of regions as a unit of analysis and recent advancement in the SCM field allows us to tackle Optimal Currency Area (OCA) literature, considering the core-periphery dynamics.