What Happens to Workers at Firms that Automate?
James Bessen, Maarten Goos, Anna Salomons, Wiljan van den Berge
Review of Economics and Statistics 107(1), 125–141
Abstract
We study how automation affects the workers initially employed at automating firms, using a novel measure of firm-level automation cost and Dutch administrative microdata covering the universe of firms and workers. Workers at automating firms see a persistent fall in days worked and labour income in the five years after automation, driven by an increased probability of separating from the firm. These costs are borne disproportionately by incumbent workers, are not offset by greater benefit uptake, and resemble the costs of mass-layoff events — even though automating firms are growing.
Research question
What are the labour-market consequences for incumbent workers when their firm automates?
Why it matters
Debates about automation usually focus on whether jobs disappear in aggregate. The more immediate question for workers and policymakers is what happens to the specific people at firms that adopt automating technologies — and whether existing social insurance cushions the blow. It largely does not.
Main findings
- Automation raises the probability that incumbent workers leave the firm, lowering days worked and annual earnings for up to five years.
- The income losses look like those from a mass layoff, despite automating firms expanding rather than contracting.
- Existing benefits replace only a small share of the lost earnings, leaving workers exposed to most of the adjustment cost.