Replication Data for: Education-Growth Threshold Activation: A Theory of State-Dependent Returns to Education Spending in Developing Economies

This repository contains the replication dataset for "Education-Growth Threshold Activation: A Theory of State-Dependent Returns to Education Spending in Developing Economies." The data comprise an unbalanced country-year panel of 115 developing economies across Sub-Saharan Africa (47 countries), Asia (38 countries), and Latin America (30 countries), spanning 1970–2023. Variables include: real per capita GDP growth (gdpgp); government education expenditure as a share of GDP (edug); the one-period lag of GDP growth (L.gdpgp), used as the threshold-sorting variable; and standard growth controls — gross fixed capital formation (gfcf), government consumption (gove), terms-of-trade level (tot) and volatility (stdtot), population growth (popg), political stability/institutional-quality indices (polinst, inst), log trade openness (logtop), and log initial income (loglgdpp). Data are compiled from the World Bank World Development Indicators, IMF International Financial Statistics, UNCTAD, and the Worldwide Governance Indicators, supplemented with author calculations for the terms-of-trade volatility measure and rolling growth-momentum thresholds. Full variable definitions and sources are documented in Table 1 of the accompanying paper. The dataset supports Hansen (1999) panel threshold regression, IV, and system-GMM analyses reported in the paper, including full-sample and regional (SSA, Asia, Latin America) subsample estimations.