This study aims to map the complex pathways through which higher education contributes to socio-economic development and to identify which disciplinary structures are most critical for different development outcomes.
We employ a two-stage analytical approach. First, structural equation modeling (SEM) is used on panel data from OECD countries (1998–2022) to test the direct and indirect (via industry and technology) effects of higher education on a multi-dimensional “High-Quality Development” index. Second, machine learning algorithms (Random Forest and XGBoost) are used to analyze the relative importance of different educational disciplines (STEM, Social Sciences etc.) for economic, environmental and health-related development goals.
The SEM results reveal that higher education’s direct effect on development is negative in the short term, but it exerts strong positive indirect effects by fostering industrial output and technological advancement. The machine learning analysis indicates that STEM disciplines are paramount for economic growth, agricultural disciplines for green development and health disciplines for population health outcomes.
This study uniquely integrates macro-level path modeling (SEM) with micro-level feature importance analysis (ML) to provide a holistic view of the education–development nexus. While based on OECD data, the derived framework offers a valuable evidence-based reference for policymakers in developing economies to strategically prioritize higher education investments aligned with specific national development stages.
