Work

Towards Sustainable Labour Market Intelligence: A Dual Analytical Framework for Analysing Regional Vacancy-Pressure Dynamics

Year: 2026
Type: article
Abstract: Persistent regional labour shortages have become a major challenge for sustainable regional development, labour market resilience and evidence-based policymaking. Despite recent advances in labour-shortage measurement, comparatively limited attention has been devoted to analytical frameworks that compare proportional and signed percentage-point deviations of regional vacancy pressure from corresponding national benchmarks. This study proposes a Dual Analytical Framework (DAF) that integrates the Labour Deficit Index (LDI) and the Regional Vacancy Gap (RVG) within a sequential evidence-based analytical workflow to identify, prioritise and investigate persistent regional vacancy-pressure dynamics. Using quarterly Romanian Job Vacancy Survey data at the NUTS 2 level, the sectoral analysis covers 19 NACE Rev. 2 economic sectors over 2008Q1–2025Q4, whereas the occupational analysis covers nine ISCO-08 major occupational groups over 2011Q1–2026Q1. The framework combines two benchmark-based vacancy-pressure measures expressed through proportional and signed percentage-point scaling with integrated sectoral and occupational screening, followed by Space–Time Cube analysis, Forest-based Forecasting, Emerging Hot Spot Analysis (EHSA), and Time Series Clustering. The statistical screening identified Sections E, G and K as the economic activities exhibiting the most persistent above-benchmark regional vacancy-pressure dynamics, with Section E selected as the representative case for advanced spatial–temporal modelling. Forecasting performance was region-specific: the Forest-based Forecast achieved lower validation RMSE than the naïve benchmark in three of the eight NUTS 2 regions, whereas the naïve benchmark produced the lower mean RMSE across regions. While Time Series Clustering identified three distinct regional temporal profiles, Emerging Hot Spot Analysis did not detect statistically significant emerging hot- or cold-spot patterns for Section E under four alternative neighbourhood specifications. Because the spatial application contains only eight NUTS 2 regions, this result is interpreted as non-detection under the tested configurations rather than as evidence that spatial dependence is absent. These findings demonstrate the complementary interpretation of proportionally and percentage-point scaled vacancy-pressure indicators and highlight the added value of integrating statistical screening, GIS-based spatial–temporal modelling and machine-learning forecasting within a unified and reproducible analytical framework. Beyond the Romanian case study, the proposed methodology provides a transferable approach for labour market intelligence, sustainable regional development and evidence-based regional policy under the green, digital and demographic transitions.
Language: English
Open Access status: gold
Estimated APC paid: $2,896
OpenAlex ID: W7225893219
FWCI:
Cites: 42
Cited by:
Related to: 10