EUVIMEDEUVIMED entdecken
Crossref

Algorithmic Profiling: Effective Tool for Targeting Active Labour Market Policies?

Rainer Eppel, Ulrike Huemer, Helmut Mahringer, Lukas Schmoigl

LABOUR · 2026-12 · 40 (4) · 432-451

DOI: 10.1111/labr.70013 ↗

Kurz gefasst

Fachartikel in der Zeitschrift LABOUR, online erschienen am 27.04.2026. Verfasst von Rainer Eppel und 3 weiteren Autor:innen. Die Arbeit nennt 77 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Digitisation has sparked interest in automated decision making in Public Employment Services (PES). We evaluate an algorithmic profiling model in Austria that predicts the reemployment prospects of unemployed individuals to classify and assign them to active labour market policies. Our analysis shows that reallocating resources from jobseekers with low to medium prospects, as proposed by the PES, does not yield the expected efficiency gains. We find no systematic evidence that programmes are less effective for individuals with low predicted employment prospects than for those with medium prospects. These findings caution against crude algorithmic profiling and highlight the need for nuanced targeting strategies that prioritise the most disadvantaged jobseekers. JEL Classification: J24, J64, J68

Zitieren

APA

Eppel, R., Huemer, U., Mahringer, H., & Schmoigl, L. (2026). Algorithmic Profiling: Effective Tool for Targeting Active Labour Market Policies?. LABOUR, 40(4), 432-451. https://doi.org/10.1111/labr.70013

Vancouver

Eppel R, Huemer U, Mahringer H, Schmoigl L. Algorithmic Profiling: Effective Tool for Targeting Active Labour Market Policies?. LABOUR. 2026;40(4):432-451. doi:10.1111/labr.70013

BibTeX
@article{Eppel2026,
  author = {Eppel, Rainer and Huemer, Ulrike and Mahringer, Helmut and Schmoigl, Lukas},
  title = {Algorithmic Profiling: Effective Tool for Targeting Active Labour Market Policies?},
  journal = {LABOUR},
  year = {2026},
  volume = {40},
  number = {4},
  pages = {432--451},
  doi = {10.1111/labr.70013}
}
Kostenlos starten →