Eva Dettmann (p), Alexander Giebler , Antje Weyh

Guido Pelligrini

The paper presents a modification of the matching and difference-in-difference approach of Heckman et al. (1998) and its Stata implementation, the command flexpaneldid. The approach is particularly useful for causal analysis of treatments with varying dates of application and varying treatment durations (like investment grants or other subsidy schemes). Introducing more flexibility enables the user to consider individual outcome differences of the matched pairs instead of the standard comparison of the sample means. Besides, we replace the Propensity Score (PS) for matching by a combined statistical distance function that consists of a weighted average of scale specific distance measures and includes matching time information.
The introduced tool flexpaneldid for panel data implements the developed approach and commonly used alternatives like PS-Matching, CEM Matching and Fixed-effects-difference-in-difference. The novelty of this tool is an extensive data preprocessing to include panel time information into the matching approach and the effect estimation with difference-in-differences. The core of the paper gives an extended example for the use of flexpaneldid on the basis of a publicly accessible data set.

Lisa Sella (p), Valentina Lamonica , Elena Ragazzi

Eva Dettmann

We present a work on the evaluation of a pilot program addressed to foreign adolescents with low language skills and high dropout risk in junior high school. The intervention - developed by a private training centre for adolescents in Turin, Italy, jointly with state junior high schools - aims at filling their gaps both in general subjects knowledge and in character skills.
During the experiment, the treated students are transferred to the training centre for a 290-hour professional training course. It aims at restoring students’ interest in education by professionally-oriented teaching, inductive innovative pedagogic methods, and individual mentorship.
The pilot experience is to be evaluated through experimental counterfactual approach. A preliminary experience run in the recent past, had positive results, although the evaluation was based only on qualitative evidence and on the subjective assessment of the teachers. For the project evaluation we constructed a grid of profiling variables, that are monitored, for students belonging both to treated and control groups. The aim is to assess social and behavioural skills, alongside other variables concerning school performance.
Previous experimental evaluations of similar programs (see Kautz et al., 2014 for a survey) provide ambiguous results: depending on program features such as length of the treatment, role and pervasiveness of training, tutors, mentors, families’ and peers’ influence, long-term analyses showed either positive (Durlak et al., 2011; Tierney et al., 1995; Orr et al., 1994) or negative (Rodrigues-Planas, 2012) effects on some outcome variables such as behavioural attitudes, character skills, educational attainment, future wages and other labour market outcomes. For this reason the project will adopt a randomised experimental design, so as to get sound evidence on the effectiveness of the implemented activities.

Kevin Mulligan (p), Helena Lenihan , Justin Doran

Lisa Sella

Policymakers at regional, national and EU levels of governance use a variety of subsidy programmes to stimulate firms-level innovation. However, firms may apply for multiple subsidies from each of these sources simply to reduce the risk and cost of investing in innovation activities they would have engaged in even without the subsidies. To investigate whether this is the case, this paper evaluates whether receiving a mix of subsidies from regional, national and EU sources crowds-out firm-level innovation.

Augusto Cerqua , Guido Pellegrini (p)

Kevin Mulligan

Regional inequalities are extensive and have expanded during the Great Recession. As they prompt people and economic activities to migrate from lagging regions, central governments around the world transfer large amounts of resources in order to stem these phenomena. In this paper, we evaluate the effectiveness of the most extensive and long-lived experiment of income redistribution across regions and countries, i.e. the EU regional policy, at a time of economic crisis. Using a spatial regression discontinuity design, we analyze comprehensive data on all publicly funded Italian projects at the municipality level. We find a positive and significant impact on employment and number of plants in the least developed regions. However, the impact turned largely negligible when the gap between more and less intensively treated areas was relatively small. Besides, the EU regional policy appeared not to have any effect on local average income.