Age-Coded Job Advertisements Inflate Labour Shortages in African Economies: Machine-Learning Evidence from a Seven-Country Study

Purpose: This study interrogates the paradox of employer-reported “labour shortages” in labour-abundant African economies. It advances the claim that shortage signals are partly institutional outputs: they arise when screening rules narrow the effective labour pool, rather than reflecting exogenous skill scarcity. Design/methodology/approach: Drawing on labour market segmentation, information economics, and critical institutionalism, we analyse 10,432 job advertisements scraped monthly (January 2024–June 2025) from leading portals in seven Anglophone African countries. A rigorously validated support-vector-machine classifier distinguishes explicit numeric age ceilings from implicit youth-coded cues to construct an Age-Coded Hiring Index (ACHI). We triangulate ACHI with employer-reported workforce-constraint indicators from the World Bank Enterprise Surveys and labour-underutilisation (LU4) from ILOSTAT, estimating fixed-effects and interaction models to test whether age-coded screening predicts shortage complaints most strongly where latent labour supply is greatest. Findings: Age-coded screening is pervasive in vacancy texts: approximately 15–20% of postings impose numeric age caps and a much larger share deploys implicit youth signals. Higher ACHI is robustly associated with stronger shortage complaints net of underutilisation and macro controls, and the relationship steepens under high labour slack, consistent with an institutional mechanism in which screening rules convert latent labour supply into perceived scarcity. Originality/value: Conceptually, the paper reframes “shortage” indicators as partially endogenous to screening rules and to employers’ definition of “suitability,” rather than treating them as market facts. Empirically, it introduces a replicable NLP-based measure of exclusionary screening from vacancy text, enabling cross-country tests of institutional scarcity dynamics in low- and middle-income contexts. Practical implications: The results imply that diagnostic and policy responses to “shortages” should not presume supply failure alone; they should also examine how recruitment criteria restrict the recognised labour pool and thereby shape shortage measurement itself.