Between- and within-country correlates of tertiary enrollment: A 41-country fixed-effects panel

Countries that spend more on public education and have wider internet access tend to report higher tertiary enrollment. It does not follow that a country will see enrollment rise when it raises its own budget or expands connectivity, and it is this second, within-country question that matters for policy. Using World Bank indicators for 41 countries over 2013-2022 (N = 410 country-years), we estimate a two-way fixed-effects panel model in which all between-country variation is absorbed and the coefficients are identified from change within countries. The results differ markedly from the cross-sectional picture. Total public education expenditure (all levels) as a share of GDP is not significant (β = -0.604, p = 0.494), and internet penetration carries a small negative coefficient (β = -0.136, p = 0.019). The associations that hold up are structural: secondary-school enrollment (β = 0.231, p < 0.01), the urban population share (β = 1.529, p = 0.019) and, with a negative sign, youth labor-force participation (β = -0.790, p < 0.001). A specification with one-year lagged regressors confirms the secondary-enrollment, youth labor-force and internet results, while the urban coefficient loses precision. We read these estimates as conditional within-country associations and not as causal effects, and we discuss what they imply for higher-education policy and the SDG 4 access agenda.