Place-Based Investment Incentives and Technological Upgrading in Lagging EU Regions: Province-Level Evidence on Inventive Activity from Italy’s Mezzogiorno

Description

This dataset contains the replication materials for the article “Place-Based Investment Incentives and Technological Upgrading in Lagging EU Regions: Province-Level Evidence on Inventive Activity from Italy’s Mezzogiorno”, submitted to Economic Modelling. The study evaluates the effects of Italy’s investment tax credit for the Mezzogiorno on regional inventive activity over 2016–2019 using a long province-level panel and a range of Difference-in-Differences, Synthetic DiD, matching, event-study, heterogeneity, robustness, and falsification analyses. The repository includes Stata code and analysis-ready data needed to reproduce the full empirical workflow for all numbered tables and figures in the article and appendix. The main province-level dataset covers 91 harmonised Italian NUTS-3 provinces over 1980–2019. Additional datasets support technological-field and Pavitt-class heterogeneity, interregional inventor-collaboration analysis, applicant-based patent regionalisation, citation-normalisation checks, cohesion-policy robustness, and placebo exercises. The principal patent and inventor measures are derived from the European Patent Office PATSTAT® database. Because the original PATSTAT data are subject to licensing restrictions and cannot be redistributed, the repository provides synthetic replacements for PATSTAT-derived variables. These synthetic files preserve the identifiers, panel structure, variable names, and relevant distributional characteristics required to execute the replication code, but they are not intended to reproduce the published patent-based numerical estimates exactly. Exact replication of those results requires authorised access to the original PATSTAT data. The cohesion-policy robustness analysis likewise uses a synthetic province-level replacement for the OpenCoesione-based exposure measure because the underlying data used in the study are not freely available for redistribution. The repository is organised into Raw Data, Analysis Data, and Output folders, together with a master Stata script and individual do-files for data construction, descriptive statistics, main estimates, event studies, heterogeneity analysis, production-side and collaboration results, and appendix robustness and placebo tests. Running the master do-file executes the full public replication workflow in sequence.

Institutions

• University of Genoa

  Liguria, Genoa

Categories

Licence