Séminaire ERMEES – Sofiene OMRI

Détails de l'événement :

Titre: Firms’ hiring bottlenecks in EU member states : Ordered probit model Vs. Ordered forest estimator

Abstract: This paper looks at the critical aspects of hiring difficulties markedly associated with skills shortage in EU member states’ establishments. Similarly, differences across NACE Rev.2 sectors in terms of recruitment bottlenecks were also identified. Our empirical investigation draws on the estimation of marginal effects using econometric ordered choice model (i.e., ordered probit) and a Machine learning technique (i.e., ordered forest estimator). The comparison between the two methods proves that the ordered forest is a robust flexible alternative to classical ordered choice models. The former overcomes the multidimensionality curse. It identifies non-linearities and multicollinearity between covariates. Further, it relaxes distributional and functional assumptions through its data-driven approximation approach. The results suggest that firms’ innovative and digital character, employee motivation, skills mismatch, competitiveness, hiring criteria and HR management schemes are key factors of hiring difficulties. Overall, the paper contributes empirically to the existing ordered choice modelling techniques by suggesting a data-centered approach as a flexible alternative.

Agenda des séminaires ERMEES : https://www.beta-economics.fr//seminaires-ermees-2022-2023/