Research
Research
Working papers
Automation and the Changing Composition of Skill Demand with Mark Hellsten and Sarah Schroeder (R&R at Labour Economics)
This paper provides new evidence on how automation reshapes firms’ demand for skills, not only by changing the occupational composition, but also by reshaping what existing jobs require. Using matched data on firm-level automation investments and detailed job vacancy postings from Denmark, we extract multidimensional skill profiles through natural language processing and decompose changes in skill demand into within- and between-occupation components. Within-occupation adjustment is a quantitatively important margin, accounting for 14–39% of total skill demand change depending on skill type and occupational group. Drawing on a task-based framework that links automation to shifts in multiple skill types within occupations, we estimate the causal effect of automation using a staggered difference-in-differences design. The effects are heterogeneous across the occupational hierarchy: among managers and professionals, automation increases the demand for soft skills, shifting the within-occupation skill mix toward interpersonal and cognitive competencies; among production workers, adjustment operates primarily through reduced hiring rather than changes in skill requirements, while retraining intensity rises by 5 percentage points. Our findings highlight that automation operates through multiple adjustment margins, with implications for training policy and labour market resilience.
Who Adopts AI? Evidence on Firms, Technologies and Workers with Mariola Pytlikova, Sarah Schroeder, Magnus Lodefalk
Using two waves of nationally representative Danish firm surveys linked to employer--employee administrative registers, we study how adoption varies across artificial intelligence (AI) and related advanced technologies. We show that AI adoption is highly technology-specific. While firm size and digital infrastructure predict adoption broadly, workforce composition operates through distinct channels: STEM-educated workforces predict core AI adoption, whereas non-STEM university-educated workforces are associated with generative AI adoption, indicating different human capital complementarities. The factors associated with adoption differ from those predicting deployment breadth: firm size and digital maturity matter for both, whereas workforce composition primarily predicts adoption alone. Machine learning and natural language processing are deployed across multiple business functions, whereas other advanced technologies remain concentrated in specific operational domains. Individual-level evidence provides a foundation for these patterns, with awareness of workplace AI usage concentrated among managers and high-skilled workers. Self-reported AI knowledge is higher among younger and more educated individuals. Finally, commonly used occupational AI exposure measures vary substantially in their ability to predict observed adoption, with benchmark-based measures outperforming patent-based and LLM-focused alternatives. These findings show that treating AI as a monolithic category obscures economically meaningful variation in who adopts, what they deploy, and how well existing measures capture it.
Offshoring and Tasks: Skill Substitution or Retraining?
This paper examines how offshoring reshapes firms’ task structure and skill composition. I distinguish between two margins of skill adjustment: reallocating workers through external labor markets and retraining incumbent workers into new occupations within the same firm. The analysis shows that offshoring firms use these two margins on different segments of the task structure: they retrain manual and production workers who move to other jobs within the organization, while hiring younger and more educated workers for cognitive tasks. The results suggest that internal labor markets and retraining can facilitate structural transformation, though they may not fully address skill gaps arising from changes in task demand.
AI Unboxed and Jobs: A Novel Measure and Firm-Level Evidence from Three Countries with Erik Engberg, Holger Görg, Farrukh Javed, Hildegunn Kyvik Nordås, Martin Längkvist, Magnus Lodefalk, Natália Monteiro, Sarah Schroeder, Aili Tang
IZA DP No. 16717; RFBerlin DP 14/24
Advances in artificial intelligence may displace some tasks, augment others, and reshape firms’ skill demand. We develop a Dynamic AI Occupational Exposure (DAIOE) measure for 2010–2023 that translates performance improvements across 140 AI benchmarks into occupation-specific exposure scores. Using matched employer–employee data from Denmark, Portugal, and Sweden in a shift-share design, we find limited associations between AI exposure and total employment but substantial skill upgrading: a one-standard-deviation increase in exposure is associated with 0.18 (Denmark) and 0.32 (Portugal) more high-skilled workers per low-skilled worker. Adjustment patterns vary across AI capabilities, worker groups, and firm characteristics, consistent with skill-biased organisational change rather than broad displacement.
Selected Work in Progress
The Employment Effects of AI Adoption with Mariola Pytlikova, Sarah Schroeder