Research data - Extractive Wealth, Industrial Strain? Testing the Dutch Disease Hypothesis in Sub-Saharan Africa

This dataset accompanies a study testing whether natural-resource extraction produces Dutch-disease effects in Sub-Saharan Africa through changes in manufacturing and wider sectoral reallocation. It provides a research frame covering 48 Sub-Saharan African countries from 1980 to 2024, comprising 2,160 country-year rows and 52 observed and derived variables. Actual coverage varies by indicator and country. The data were obtained from the World Bank’s World Development Indicators, July 2026 release, using machine-readable distributions provided through Our World in Data. The variables include manufacturing, agriculture and services value added as shares of GDP; sectoral growth and constant-price value added; total natural-resource rents; oil, natural-gas, mineral and coal rents; trade openness; foreign direct investment; gross capital formation; inflation; GDP per capita; population; the real effective exchange rate; and the domestic price level relative to the United States. Derived measures include aggregate extractive rents, hydrocarbon rents, mineral-and-coal rents, lags, annual changes, transformed inflation, logarithmic price and exchange-rate variables, and real sectoral value added per capita. Extractive rents are calculated only where all four rent components are observed, while lags and annual changes require consecutive country-year observations. No values were interpolated, extrapolated or imputed; blank cells represent missing observations. The workbook contains 13 worksheets: README, results dashboard, processed country-year panel, codebook, descriptive statistics, main regressions, sectoral and mechanism models, robustness tests, extractive-rent heterogeneity estimates, leave-one-country-out results, missingness assessment, country summary and source documentation. The preferred specification applies country and year fixed effects, control variables and country-clustered standard errors. It contains 983 observations from 41 countries over 1990–2022. A one-percentage-point increase in lagged extractive rents is associated with an approximately 0.111-percentage-point reduction in manufacturing’s share of GDP. The relationship is robust across alternative specifications and appears mainly hydrocarbon-led. However, manufacturing growth and trend-adjusted real manufacturing output do not decline consistently. The evidence therefore supports a partial, composition-based form of Dutch disease rather than conclusive absolute manufacturing contraction. The dataset is suitable for research on resource dependence, structural transformation, manufacturing, deindustrialisation, sectoral development and comparative economic performance in Sub-Saharan Africa. Results are observational associations and should not be interpreted as definitive causal effects.