Research

Publications

"Equalising the effects of Automation? The Role of Task Overlap for Job Finding", joint with Sabrina Genz and Diego Dabed, published at Labour Economics, 96, 102766, October 2025.

Abstract:

This paper investigates whether task overlap can equalise the distributional effects of automation for unemployed job seekers displaced from routine jobs. Using a language model, we establish a novel job-to-job task similarity measure. Exploiting the resulting job network to define job markets flexibly, we find that only the most similar jobs affect job finding. Since automation-exposed jobs overlap with other highly exposed jobs, task-based reallocation provides little relief for affected job seekers. We show that this is not true for more recent software exposure, for which task overlap lowers the inequality in job finding.

"Job Polarization: Its History, an Intuitive Framework and some Empirical Evidence", joint work with Maarten Goos, Anna Salomons and Marieke Vandeweyer, The Oxford Handbook of Job Quality. Warhurst, Chris, Chris Mathieu, and Rachel E. Dwyer, eds. Oxford Handbooks. Oxford, New York: Oxford University Press, June 2022.

Abstract:

This chapter first gives an overview of the literature on job polarization: The idea that job markets are polarizing into lovely and lousy jobs at the expense of middling jobs originates from the early 1980s. During the following decades, the phenomenon of job polarization was documented for many advanced economies and its drivers became better understood. To summarize the intuition that economists currently have about these drivers, the chapter then provides a simple intuitive framework and some empirical evidence in support of it. The contribution of this framework is that it explicitly derives critical assumptions about key structural parameters for understanding job polarization: the direct substitutability between differently-skilled workers and new digital technologies in production, and scale-effects driven by technology-induced changes in the prices of goods and services for consumers. Finally, the importance of these channels is documented by empirical evidence from 13 advanced economies. Our analyses imply that job polarization is not a natural law that must hold at all times and in all places, but that it critically depends on the nature of technological progress, the substitutability between differently-skilled workers and new technologies in production, and the substitutability between goods and services in consumption.

"Routine-biased technical change: Individual-level evidence from a Plant Closure", joint work with Maarten Goos and Ronja Roettger, published at Research Policy, Volume 50, Issue 7, September 2021, 104002.

Abstract:

Routine-biased technical change (RBTC) argues that digitisation decreases job opportunities for workers withroutine task competencies, but increases job opportunities for workers with nonroutine task competencies. While there is considerable evidence for RBTC at the aggregate level, its effects on individual workers are yet to be fullyunderstood. Therefore, this paper uses unique survey data of workers at a large car plant who became un-employed when the plant closed. In line with the RBTC hypothesis, we find that re-employment probabilities 1,5years after the plant’s closure are substantially higher for workers with nonroutine task competencies and withdigital skills. Moreover, for the subset of individuals who were re-employed 1,5 years after the plant’s closure,we find that the nonroutine content of job tasks is higher, wages are lower, and contracts are less permanent.Finally, our paper shows that a crude age-based early retirement policy that was negotiated as part of the plant’sclosure and that ignores workers’ skills, results in significant foregone employment of older workers withnonroutine task competencies.

"Markets for Jobs and their Task Overlap", joint work with Maarten Goos, Anna Salomons and Bert Willekens. published at Labour Economics, Volume 61, December 2019, 101750.

Abstract:

We show that tightness in markets for jobs for which an unemployed job seeker fully qualifies in terms of her task competencies is predictive of her unemployment duration. This suggests that the labour market is organized along jobs and their task content. We also find that unemployed job seekers do not compete in markets where they possess only part of the required task competencies, suggesting that task overlap across jobs is unimportant for worker mobility between job markets. This implies that adverse task-biased shocks are likely to have pronounced distributional consequences across workers with different task competencies. To illustrate this, we quantify the impact of technological progress that automates routine tasks, showing that this imposes substantial adjustment costs that are highly unevenly distributed across unemployed job seekers with routine versus non-routine task competencies.

Working Papers

"Automation and Collective agreements" , joint work with Sabrina Genz (submitted at Fiscal Studies) link to latest version

Abstract:

This paper empirically examines how collective bargaining agreements influence firms’ automation decisions and employment dynamics. Using novel administrative data on Dutch firms and workers, we link detailed job-level collective bargaining coverage to firm-level automation expenditures. Our analysis yields two main findings. First, firms covered by firm-level collective bargaining invest more in automation than uncovered firms, suggesting that collective agreements create cost-incentives for automation. Second, firms that were initially covered by firm-level agreements tend to experience smaller employment growth, which can contribute to the aggregate decline in collective agreement coverage.

"Types in Job Search Effort" , joint work with Maarten Goos (submitted) latest version available at SSRN

Abstract:

Public employment services are increasingly using online job platforms to help unemployed job seekers find jobs. This paper shows that these policies implicitly target those who benefit the most from them. The reason is that search intensity on online job platforms is characterized by "types": more-educated job seekers with better language and digital skills use the online job platform more intensively throughout their unemployment spell. Moreover, these high-types in online search effort are also more likely to take-up in-person job search assistance and, if they do, are more responsive to it. Using data from a public employment service, the paper provides causal evidence for the existence of types and presents a job search model to rationalize its findings.

Work in Progress

"Tightness and wage growth in the Netherlands" , joint work with Melanie Kühnast, Bas Scheer, Marente Vlekke en Alice Zulkarnain

Bas Scheer. Policy summary in Dutch

Abstract:

This study examines the relationship between labor market tightness and wage growth in the Netherlands between 2010 and 2024, using data on hourly wages by industry sector and province. Our analysis reveals that workers in tight labor markets, such as ICT, experience approximately 1.2 percentage points higher wage growth compared to moderately tight sectors like manufacturing, with job switchers gaining an additional 1 percentage point premium. Importantly, we find substantial heterogeneity across the wage distribution: the association between labor market tightness and wage growth is approximately 2.5 times stronger for workers in the bottom quartile compared to all workers, while workers in the top half of the wage distribution show no significant relationship except when changing jobs. These findings suggest that labor market tightness operates as a mechanism for wage compression, disproportionately benefiting low-wage workers by improving their outside options and enabling better job matching.

Google Sites

Report abuse