Course descriptions and skill supply: an exploration of green content and uniqueness

This paper develops a novel empirical framework to measure the skill content of higher education programmes. Using natural language processing techniques, we link the official descriptions of Italian degree programmes to the ESCO taxonomy of labour-market skills, constructing a high-dimensional skill provision matrix covering more than 45,000 programme-year observations over 2013–2022. We exploit this skill-based representation in two applications. First, we map the distribution and evolution of green skills across disciplines and territories. Second, we construct measures of programme-level uniqueness and examine their association with first-year enrolment. While abstract skill-based uniqueness is not significantly related to enrolment, a geographically weighted measure–capturing differentiation relative to proximate alternatives–is positively and robustly associated with student demand. The proposed methodology provides a scalable and flexible tool to open the ‘black box’ of curricular content and can be readily extended to a wide range of applications, including the analysis of skill alignment, institutional adaptation, and the evolving geography of higher education provision.