Research opportunity · Aggregate null or distributional shift? Employment

Why do young-worker estimates in AI-exposed jobs differ between the US and the Nordics?

In short: has this been done?

Related work: Partly addressed. Related studies cover parts of this question. The in-depth check found none that does the whole design. Not finding one is not proof that none exists; what the existing work leaves out is set out below.

How far checked: Checked in depth. A first-pass screen on 12 October 2026, then a second, deeper literature check on 13 October 2026. Result of that check: Question stood after this check; reworded later (see Changes). Every search and check so far was done by AI; no person or domain expert has checked them yet.

The searches, the nearest work found and what the search could not reach

This page shows RO-LD02-01 version v3, whose record never changes. It is the current version; if the opportunity is revised, the newer version appears at its main page.

Question

Estimates of how young workers in AI-exposed occupations have fared since 2022 range from large relative declines in US payroll and Census data to near-zero or only emerging gaps in Nordic registers. When one exposure measure, the same age bands, occupation coding and event-study specification are applied to US data and to at least one Nordic population register, how much of the difference comes from the data used (payroll records, surveys or registers) and how much from differences between the countries (occupation mix, education composition, sensitivity to interest-rate rises, remote-work exposure and cohort size)?

Gap type: Untested beyond the original settingOne PhD chapter or paperLow costRelated work: Partly addressed

Why it matters

The strongest displacement-like signal in this debate comes from US young workers. Norwegian and Finnish registers mostly find small or no effects, while Swedish data and a later Norwegian replication show emerging gaps, so the Nordic picture is itself mixed. Each paper uses its own data source, exposure index, age cut-offs and model, so readers cannot tell whether the countries really differ or the studies simply measure different things. A harmonised comparison that separates measurement from country would tell us whether the US result is a general early warning or a US-specific pattern, which is the question supervisors, policymakers and graduates are actually asking.

The debate lists this among what we still do not know: Why do estimates of young-worker effects in AI-exposed jobs differ between US datasets and Nordic data, and how much of that comes from the data rather than the countries?

Evidence that motivates it

Each claim is shown as it read when this version was published.

A possible design

Write one pre-registered specification (one exposure index plus one alternative, the same age bands, an occupation crosswalk to ISCO-08, a monthly or quarterly cell-level Poisson event study with occupation and period fixed effects). Run it on US public data (CPS basic monthly files, ACS) and on at least one Nordic register through an existing collaborator or remote-access system (for example microdata.no in Norway). The US results that motivate the question come from private payroll and Census administrative records, which this study cannot rerun, and a survey-based US estimate is a different measurement. So first test whether the specification on CPS/ACS reproduces the published US payroll and Census patterns; where it does not, report the gap as a measurement effect, not a country effect. Where possible, also run the specification on the same Nordic country’s labour force survey, to measure the register-versus-survey difference within one country. Then decompose what remains of the US–Nordic difference by re-weighting the US occupation and education mix to the Nordic distribution, adding each occupation’s sensitivity to the 2022 interest-rate rises and its remote-work exposure, and adjusting for the size of entering cohorts. Report what share of the gap each step removes, what remains unexplained, and the precision of each US cell.

Data needed

US: CPS and ACS public microdata (free), and the published US payroll and Census estimates as benchmarks. Nordic: one population register with occupation and age, via remote access (Norway’s microdata.no, used by the Norwegian study below, is open to researchers at approved institutions; eligibility needs checking) or a co-author with Statistics Sweden, Finland or Denmark access; where possible, the same country’s labour force survey microdata, whose access rules also vary. Published exposure crosswalks (Eloundou et al. and one alternative). An occupation-level interest-rate sensitivity measure, built from pre-2022 industry composition, and a remote-work exposure measure.

Feasibility

One PhD chapter or paper. Low only once register access is confirmed: the US public data are free, but the Nordic side needs an application through a collaborator or remote-access system, and fees and eligibility rules vary by country. Confirm access before committing to the project. Two conditions decide feasibility and should be checked before pre-registration. First, confirmed register access. Second, US cells precise enough to compare: CPS cells for 22–25-year-olds in narrow occupations are small, so test whether pooling months or broader occupation groups gives usable precision, and compare at a coarser level if it does not. Making US SOC and ISCO codes comparable is the other main hurdle.

Has this been done?

Partly done already: related work covers some of this question (first-pass screen of 12 October 2026, checked again on 13 October 2026).

Separate single-country studies exist for Norway (two papers), Finland and Sweden, and the US paper runs many robustness checks, including education and interest-rate controls. The Norwegian and Finnish authors offer institutional explanations (unions, employment protection, demographics) for why their results differ from the US but do not test them. Multi-country papers with a common specification now exist, but they cover four English-speaking countries (where remote-work exposure absorbs most of the GenAI effect on junior hiring) or job postings across 84 countries without any age dimension. No study was found that runs one identical specification across the US and several Nordic registers for young workers and decomposes why the estimates differ; such a decomposition should now also test remote-work exposure alongside occupation mix, education, interest-rate sensitivity and cohort size. The Swedish paper also finds a widening 22-25 gap from 2024 with failed pre-trend tests, the Norwegian 22-25 replication shows an emerging gap from spring 2025, and a Banque de France analysis reports a large relative fall in permanent hiring of 21-29-year-olds in the most exposed French occupations, so the European picture is itself not uniformly null.

How strong this evidence is

Nearest work found

Literature check

Question stood after this check; reworded later (see Changes). No study was found that runs one specification on US data and Nordic registers and decomposes the difference. The closest new multi-country work uses common specifications but on other country sets or outcomes: Lambert & Schindler (US, UK, Canada, Australia; junior hiring share, remote-work exposure absorbs most of the GenAI effect) and Huang et al. at the World Bank (84 countries including the Nordics, job postings, no age dimension). These narrow nothing in the core design, but the remote-work result identifies a further candidate explanation the decomposition should include.

Searched: web (extended); arxiv.org; ssrn.com; osf.io; nber.org; iza.org; oecd.org; econstor.eu; cepr.org; Swedish- and Danish-language web; rfberlin.com (opened); thedocs.worldbank.org (opened). Checked by Claude (SocioLens Project Lead) with AI research agents, by web search; not a person and not a domain expert.

The screen before this check

Partly done already: related work covers some of this question (first-pass screen of 12 October 2026).

Separate single-country studies exist for Norway (two papers), Finland and Sweden, and the US paper runs many robustness checks, including education and interest-rate controls. The Norwegian and Finnish authors offer institutional explanations (unions, employment protection, demographics) for why their results differ from the US but do not test them. No study was found that runs one identical specification across the US and several Nordic registers and decomposes why the estimates differ. The Swedish paper also finds a widening 22–25 gap from 2024 with failed pre-trend tests, the Norwegian 22–25 replication shows an emerging gap from spring 2025, and a Banque de France analysis reports a large relative fall in permanent hiring of 21–29-year-olds in the most exposed French occupations, so the European picture is itself not uniformly null.

Search queries used
  • generative AI young workers employment register data Sweden OR Finland OR Netherlands exposed occupations 2025 2026
  • Norway register AI exposure young workers employment near zero working paper
  • cross-country comparison AI exposure entry-level employment young workers administrative data harmonised
  • harmonized multi-country replication "AI exposure" young workers employment Denmark Norway Sweden Finland United States same specification
  • why AI young worker employment effects differ US versus Europe institutions explanation decomposition Brynjolfsson canaries replication
  • generative AI young workers employment multi-country comparison United States Europe same exposure measure 2026 Denmark Norway Sweden Finland Netherlands

Literature check, 13 October 2026:

  • cross-country comparison generative AI early-career employment Denmark Norway Sweden United States harmonized specification 2026
  • AI exposure young workers employment Europe versus United States why different institutions decomposition (searched only on arxiv.org, ssrn.com, papers.ssrn.com, osf.io, nber.org, iza.org, docs.iza.org)
  • "Canaries in the Coal Mine" replication Denmark register data young workers AI exposed occupations
  • Lambert Schindler 2026 AI exposure remote work junior share hiring four countries job postings
  • unga arbetstagare AI-exponerade yrken sysselsättning registerdata 2026
  • "Has AI Widened Employment Gaps" Hernæs Kostøl cites OR extends OR comparison United States
  • unge arbejdstagere kunstig intelligens eksponerede erhverv beskæftigelse registerdata Danmark 2026 analyse
  • multi-country AI exposure entry-level employment labour force survey ages 20-29 OECD countries event study ChatGPT (searched only on arxiv.org, papers.ssrn.com, ssrn.com, osf.io, oecd.org, econstor.eu, cepr.org)
Cite this page

SocioLens project. (2026, October 22). Why do young-worker estimates in AI-exposed jobs differ between the US and the Nordics? [Research opportunity RO-LD02-01, version v3]. SocioLens. https://sociolens.eu/opportunities/ro-ld02-01/v3/

@misc{sociolens_ro_ld02_01_v3,
  author = {{SocioLens project}},
  title = {Why do young-worker estimates in AI-exposed jobs differ between the US and the Nordics?},
  year = {2026},
  date = {2026-10-22},
  howpublished = {\url{https://sociolens.eu/opportunities/ro-ld02-01/v3/}},
  note = {Research opportunity RO-LD02-01, version v3}
}

The link points to this version, whose record never changes. For a finding, cite the original study as well: SocioLens summarizes research and does not replace it.

Provenance

RO-LD02-01 v3. Drafted by Claude (SocioLens Project Lead), AI-drafted from the LD-02 debate on 2026-10-12. Checked by an independent AI reviewer with fresh context (not a person), 12 October 2026: every nearest-work citation checked against its source page; corrections applied. Not yet reviewed by a domain expert. This is a research hypothesis, not a finding.

Changes

All research opportunities