Firm size and the automation wage premium

Abstract We investigate the impact of investment in automation-related goods on wages and wage inequality across firms of different sizes in Italy from 2011 to 2019. We integrate datasets on trade activities, firm, and worker characteristics for the population of Italian importing firms and quantify the automation wage premium for adopting firms, which stands at approximately 10%. However, Mincer-type wage regressions reveal that automation adopters pay approximately 3% higher wages after controlling for worker sorting. We then disentangle the impact of automation adoption on wages from selection into adoption by estimating the effects on adopters’ wages and wage dispersion within a difference-in-differences framework, exploiting import lumpiness in product categories linked to automation technologies (including robots). We find a positive average adoption effect on adopters’ average wages, which stabilizes at around 4% 5 years after an automation spike. Wage increases at small firms primarily explain this effect, while wages at medium and large firms remain stagnant after adoption. Importantly, wage gains coincide with an increase in within-firm wage dispersion in small firms, with wage variance rising by around 7.5%. This result comprises heterogeneous effects not only across firms but also across worker groups: decomposition analysis reveals that wage benefits are concentrated among specific worker categories—employees aged 45 and above, managers, and white-collar workers. Other worker categories experience stagnant wages, although no group shows a negative wage effect.