Skip to content

CES Dissertation Workshop

Machines and Workers: Evidence from Winners and Losers of Technology Subsidies


April 9, 2021
12:00pm - 1:15pm
Virtual/RSVP Required
April 9, 2021
12:00pm - 1:15pm
Virtual/RSVP Required

The Dissertation Workshop is a graduate educational seminar open only to graduate students and their advisors. CES invites graduates students who are interested in attending this workshop or in presenting their research, to contact CES Dissertation Workshop coordinators Clare Bradford Anderson and Armando Miano. All workshops will be held virtually during the 2020-2021 academic year. Zoom links for the workshop will be distributed a day in advance to registered participants.

About

Joonas Tuhkuri will present a paper (co-authored with Johannes Hirvonen and Aapo Kivinen) that revisits a classic question: Do technologies replace workers and increase skill demand at the firm level?

The research design estimates the causal effects of technologies by contrasting winning and losing firms of a technology-subsidy program for manufacturing firms in Finland. The co-authors use natural language processing (NLP) on the subsidy application text data to match the applicants on similarity and estimate the effects of specific technologies.

To formalize the research design and identification assumptions, they develop a dynamic technology-adoption model with adjustment costs. Their data track workers and firms over time and measures multiple technologies and skills. The evidence shows that after winning a technology subsidy, firms invested sharply more in technologies, hired more workers, but did not change their skill composition. They found no worker replacement or skill-bias across multiple skills and technologies, such as automation vs. non-automation.

The study contrasts with the conventional view that new technologies replace workers and increase skill demand. Based on the evidence, they argue that the critical point is technology vs. how technology is used. The sample firms use technologies to change output — expand, export, and launch new products — rather than change work. The results are robust to several alternative explanations evaluated in light of the evidence.

Sponsors

Close