← Papers

Highly digitalizable sectors did not generate net job gains during the COVID-era but paid more and went remote: wages rose by about €0.52/hr (≈4.6%) and remote work surged by ~41 percentage points compared with less-digitalized activities.

Digital transformation and labor market indicators in the EU: Evidence from the COVID-19 shock using difference-in-differences

Nataliia Bieliaieva, Oleksandr Rozhko, Iuliia Padafet, Svitlana Cherkasova, Semen Blahun, Tetyana Kharchenko, Dmytro Poroshyn · May 12, 2026 · Problems and Perspectives in Management

openalex quasi_experimental medium evidence 8/10 relevance DOI Source PDF

Using a DiD design on 2018–2024 EU quarterly data, sectors with higher digitalization potential saw no aggregate employment change, experienced an average hourly wage increase of €0.52 (≈4.6%), and a roughly 40.7 percentage-point rise in remote work.

Type of the article: Research ArticleAbstractDigital transformation has emerged as a key driver of structural change in labor markets worldwide, especially in the aftermath of the COVID-19 shock. In the European Union, the pandemic particularly accelerated the adoption of digital technologies and remote work across economic activities. This study estimates the causal effect of the digitalization potential of economic activity (proxied by a binary classification into highly and less digitalized groups based on telework feasibility and digital intensity) on three labor market indicators: employment, hourly wages, and remote work. Using the COVID-19 shock as a quasi-natural experiment within a difference-in-differences (DiD) framework, the empirical analysis draws on quarterly panel data for a consistent sample of 27 EU Member States (excluding the United Kingdom) over 2018–2024 (N = 36,685). The results indicate that higher sectoral digitalization potential (telework feasibility and digital intensity) does not significantly affect aggregate employment levels, as evidenced by a near-zero DiD coefficient (0.06, p ≈ 0.98). In contrast, it has a statistically significant positive effect on wages, with a DiD coefficient of 0.52 €/hour (p &amp;lt; 0.001), corresponding to an increase of approximately 4.6% in the wage gap between highly and less digitalized activities. The strongest effect is found for remote work: the DiD estimate is 40.74 percentage points (p &amp;lt; 0.001). Remote work rose from 17.6% to 82.1% in highly digitalized sectors, compared with only 1.3% to 6.6% in less digitalized economic activities.AcknowledgmentThis article was prepared within the framework of the research project “Modelling the impact of economic digitalisation on public health in Ukraine in the context of preserving human capital” (State Registration No. 0126U001085).

Summary

Main Finding

Using the COVID-19 shock as a quasi-natural experiment in a difference-in-differences (DiD) design, the paper finds that sectoral digitalization potential (binary classification into highly vs less digitalized activities based on telework feasibility and digital intensity) had (1) no detectable effect on aggregate employment, (2) a statistically significant positive effect on hourly wages, and (3) a very large positive effect on remote work adoption.

Key point estimates: - Employment: DiD = 0.06 (p ≈ 0.98) — essentially no effect. - Hourly wages: DiD = €0.52/hour (p < 0.001) — about a 4.6% increase in the wage gap favoring highly digitalized sectors. - Remote work: DiD = 40.74 percentage points (p < 0.001). Remote work in highly digitalized sectors rose from 17.6% to 82.1%; in less digitalized sectors it rose from 1.3% to 6.6%.

Key Points

Treatment definition: binary sectoral classification into highly vs less digitalized based on telework feasibility and measures of digital intensity.
Outcomes analyzed: employment, hourly wages, and incidence of remote work.
Large and sustained shift toward remote work in highly digitalized activities following COVID-19.
Wage premium increased for highly digitalized sectors, while aggregate employment levels remained stable.
Results are based on EU-wide panel data and exploit the COVID shock for identification.

Data & Methods

Empirical strategy: difference-in-differences (DiD) using the COVID-19 shock as a quasi-natural experiment.
Data: quarterly panel covering 27 EU Member States (United Kingdom excluded) over 2018–2024.
Sample size: N = 36,685 (quarter–country–sector observations as reported in the abstract).
Treatment measure: binary indicator combining telework feasibility and digital intensity at the sectoral/activity level.
Outcomes: sectoral employment, hourly wages, and share of remote work.
(Details on specific control variables, fixed effects, and standard-error clustering are not reported in the abstract.)

Implications for AI Economics

Complementarity with digital skills and AI adoption: the wage premium and strong remote-work uptake in digitalized sectors suggest rising returns to digital/remote-capable skills that will matter for AI-driven task reallocation.
Limited short-run aggregate employment losses: no measurable aggregate employment decline implies COVID-accelerated digitalization may have shifted tasks and jobs rather than caused net job destruction at the macro level — but distributional reallocation across occupations and regions is likely.
Inequality and labor-market sorting: increased wage gaps and remote-work differentials can amplify earnings inequality and affect geographic labor-market dynamics (e.g., firm location, commuting, regional wages).
Policy relevance: targeted reskilling/upskilling, remote-work regulation, tax and social-protection adjustments, and regional policies to manage uneven benefits of digitalization are important policy levers.
Research agenda for AI economics: differentiate effects of general digitalization from AI-specific automation; analyze heterogeneous impacts by skill, occupation, firm size, and region; study medium- to long-run employment dynamics and productivity responses; and use worker- and firm-level data to trace task reallocation and distributional consequences.

Acknowledgment: the article was prepared under the project “Modelling the impact of economic digitalisation on public health in Ukraine in the context of preserving human capital” (State Registration No. 0126U001085).

Assessment

Paper Typequasi_experimental Evidence Strengthmedium — The DiD design over a large multi-country quarterly panel and large N provides plausible causal leverage, especially for sharp outcomes like remote work; however, strength is limited by reliance on a binary treatment measure, potential violations of parallel trends or differential pandemic shocks across sectors/countries, and the abstract’s lack of detail on robustness checks, controls, fixed effects, clustering, and event-study validation. Methods Rigormedium — The study applies a standard quasi-experimental method to rich panel data (2018–2024, 27 EU states) which is appropriate and potentially rigorous, but the abstract does not report key methodological details (pre-trend tests, covariates, fixed effects structure, standard error clustering, heterogeneity/robustness analyses), and the coarse two-group classification may mask within-sector variation. SampleQuarterly panel data for a consistent sample of 27 EU Member States (United Kingdom excluded) from 2018 to 2024, yielding 36,685 observations; sectors are classified into two groups (high vs less digitalization potential) based on telework feasibility and digital intensity; outcomes analyzed are sector-level employment, hourly wages, and share of remote work. Themeslabor_markets adoption IdentificationDifference-in-differences (DiD) that uses the COVID-19 shock as a quasi-natural experiment to compare outcomes in sectors classified as 'high digitalization potential' versus 'less digitalized' (binary classification based on telework feasibility and digital intensity) across a panel of 27 EU Member States (2018–2024); identification rests on a parallel trends assumption in the pre-COVID period and on the sectoral binary treatment capturing differential exposure to the pandemic-driven digitalization shock. GeneralizabilityGeographic: limited to EU Member States (UK excluded); results may not generalize to non-EU contexts with different labor market institutions., Shock-specific: identification leverages the COVID-19 pandemic, so effects may reflect pandemic-specific demand and policy shocks rather than steady-state digitalization impacts., Measurement: binary sectoral classification (high vs low digitalization potential) may be coarse and conceal within-sector and occupation-level heterogeneity., Temporal: covers 2018–2024 (short-to-medium term); long-run adjustment dynamics beyond this window are unobserved., Policy heterogeneity: cross-country differences in pandemic policies and safety nets could confound estimates if not fully controlled.

Claims (4)

Claim	Direction	Confidence	Outcome	Details
Higher sectoral digitalization potential (telework feasibility and digital intensity) does not significantly affect aggregate employment levels. Employment	null_result	high	aggregate employment levels	n=36685 0.06 0.48
Higher sectoral digitalization potential has a statistically significant positive effect on wages (hourly wages). Wages	positive	high	hourly wages	n=36685 0.52 €/hour 0.48
Higher sectoral digitalization potential strongly increased remote work: DiD estimate 40.74 percentage points (p < 0.001); remote work rose from 17.6% to 82.1% in highly digitalized sectors versus 1.3% to 6.6% in less digitalized sectors. Adoption Rate	positive	high	share of remote work (percent of work done remotely)	n=36685 40.74 percentage points 0.8
The study classifies economic activities into a binary grouping (highly digitalized vs less digitalized) based on telework feasibility and digital intensity and uses COVID-19 as a quasi-natural experiment within a DiD framework on quarterly panel data for 27 EU Member States (2018–2024, N = 36,685). Other	other	high		n=36685 0.8