Automating Food Production: Evidence From the Dutch Food Industry

Abstract

Automation technology has the potential to replace a significant part of the labour force in the food industry. With data from 2002 to 2023, I investigate the causal treatment effects of the adoption of automation technology in the Dutch food industry. Automation technology in the food industry is used in a variety of tasks, such as logistics, safety, quality and production. I present a theoretical model to outline relationships that are tested in the empirical analysis. The outcome variables of interest are gross value added, average wage, employment and the labour share. The empirical analysis finds that the effect of automation on these outcomes is ambiguous and the majority of the estimates are not statistically significant. I study the relationship under various samples of food industry firms and automation definitions. For the full sample, long-run changes in these outcomes are largely uncorrelated with automation, while food processors show a long-run decline in gross value added and a rising labour share. A chained difference-in-differences design shows that adoption raises gross value added and lowers the labour share and wages among larger firms and food processors. However, most of the effects of automation are not statistically significant. The absence of strong effects of automation could be explained by investments that monitor and control production steps, rather than displace worker roles in production.