This paper examines the impact of automation investments on employment dynamics and workforce composition using administrative data from Portugal. I exploit the lumpiness of automation imports in a difference-in-differences event study design. My results show that automation creates jobs in small firms but leads to job losses in larger ones. This pattern holds across a wide range of firm types, industries and types of automation technologies. Most importantly, automation favors low-educated, routine-blue-collar workers in routine-intensive jobs over highly skilled workers like STEM professionals. These findings challenge the view of automation as inherently skill-biased.
I presented this work recently at: FIND Seminar at Aarhus University (February 2025), Lunch Seminar at the University of Pisa (October 2024), AIEL Conference in Napoli (October 2024), Verein für Socialpolitik in Berlin (September 2024), Linked Employer-Employee Conference (LEED) in Copenhagen (June 2024), International Schumpeter Society Conference in Gothenburg (June 2024), BeNA 20-Year Jubilee Conference in Berlin (March 2024).
Co-authored with Andrea Mina and Daniele Moschella.
The debate on worker representation remains polarized. Critics argue that it harms firm performance, while proponents suggest it enhances job quality and competitiveness. This paper examines the impact of board-level employee representation in France after a 2015 mandate for firms with 1,000+ permanent workers. Using rich employer-employee data, we employ difference-in-differences and regression discontinuity designs to estimate causal effects. Our findings suggest that employee representation has negligible or small positive effects on wages and job security, with no significant impact on productivity, profit margins, or capital intensity. Crucially, we find no evidence that worker representation harms firm performance, challenging the critics' view.
I presented this work at: ASTRIL Conference in Rome (January 2025), PRIN 2022 Workshop on Inequalities and Technical Change in Pisa (November 2024). Daniele Moschella presented it at the International Schumpeter Society Conference in Gothenburg (June 2024).