Health Spending Cuts Emissions While Foreign Investment Pollutes in Sub-Saharan Africa, Study Finds

Sub-Saharan Africa stands at a crossroads in the global climate conversation. The region contributes relatively little to the world’s cumulative carbon dioxide emissions, yet it faces some of the harshest consequences of a warming planet, from prolonged droughts in the Horn of Africa to flooding across West African coastlines. Understanding what actually drives environmental degradation within the region has therefore become a central question for economists and policymakers alike. A new study published in Discover Sustainability by a team of researchers at Debre Tabor University in Ethiopia tackles that question with an unusually rigorous statistical toolkit, analyzing a balanced panel of 28 Sub-Saharan African countries over the period from 2000 to 2021. Their results challenge several assumptions that have shaped environmental policy debates in developing economies, particularly the idea that all forms of human capital investment are good for the environment.

The research team, led by Getish Birhane Bezabh, set out to disentangle the drivers of carbon dioxide emissions in a region where economic growth, population pressure, energy poverty, and foreign investment intersect in complicated ways. Rather than treating human capital as a single undifferentiated quantity, as most previous studies have done, the authors split it into two distinct dimensions: health expenditure and education expenditure. This distinction matters because the two forms of investment operate through very different channels. Health spending improves labor productivity and longevity, potentially shifting economic activity toward less polluting services, while education spending shapes technological absorption and innovation capacity over longer horizons. By separating them, the study offers a more granular picture of how social investment interacts with environmental quality than the existing literature has typically provided.

Methodologically, the paper leans on the two-step system generalized method of moments, or system GMM, an estimator designed for dynamic panel data in which the dependent variable depends on its own past values. Carbon emissions are exactly such a variable: countries with high emissions this year tend to remain high emitters next year because infrastructure, industrial structure, and energy systems change slowly. System GMM addresses two persistent problems in this setting. First, it controls for unobserved country-specific heterogeneity, such as geography or institutional history, that could otherwise bias estimates. Second, it uses lagged values of the variables as internal instruments to mitigate endogeneity, the concern that causality might run in both directions between emissions and their supposed drivers. The two-step variant applies a finite-sample correction to the standard errors, making inference more reliable when the number of instruments is large relative to the number of countries.

Because GMM estimates can be sensitive to specification choices, the authors did not rely on a single model. They complemented their baseline with Driscoll-Kraay estimation, a technique that produces standard errors robust to cross-sectional dependence, meaning that shocks common to many countries at once, such as global commodity price swings or worldwide financial crises, do not distort the results. They also re-estimated the model using an alternative system GMM specification and tested alternative proxies for human capital. Before estimation, they verified the statistical properties of their data using second-generation panel unit root tests, specifically the Pesaran CIPS and CADF statistics, which account for cross-sectional dependence that older tests ignore. These tests indicated that key variables, including the logarithm of income per capita and government health expenditure, were integrated of order one, becoming stationary after first differencing, a prerequisite for sound panel inference.

The headline findings are striking in their asymmetry. Across the baseline and the robustness checks, three variables emerged as consistently robust determinants of carbon dioxide emissions in the region: health expenditure, the logarithm of energy use, and foreign direct investment. Health expenditure was associated with lower emissions, a result the authors interpret as evidence that healthier populations and better-resourced health systems can coexist with, and even encourage, cleaner patterns of economic activity. Energy use, unsurprisingly, raised emissions, reflecting the region’s continued reliance on biomass and fossil fuels. Foreign direct investment also increased emissions, providing support for the Pollution Haven Hypothesis, the proposition that multinational firms relocate pollution-intensive production to countries with weaker environmental regulation. In Sub-Saharan Africa, the data suggest, foreign capital has too often arrived bundled with carbon-intensive industrial activity rather than clean technology transfer.

The divergence between the two faces of human capital is perhaps the study’s most consequential contribution. While health spending reliably reduced emissions across specifications, education expenditure showed no statistically significant effect on environmental quality. The authors do not conclude that education is irrelevant; rather, they suggest that the environmental payoff of education depends on its content and quality. In economies where curricula rarely address sustainability, renewable energy, or green technology, additional years of schooling may raise productivity without steering it toward cleaner production. This reading motivates one of the paper’s central policy recommendations: that governments in the region invest in green curricula so that the next generation of engineers, farmers, and entrepreneurs carries environmental literacy into the workforce. The finding also serves as a caution to researchers, since studies that aggregate human capital into a single index may mask offsetting effects and draw misleading conclusions.

The study also tested the Environmental Kuznets Curve, the influential hypothesis that environmental degradation first rises and then falls as income grows, tracing an inverted U-shape over the development path. The baseline estimation produced evidence consistent with the curve, with an estimated turning point of approximately 2,186 US dollars in constant 2015 prices, a figure that lies within the sample range of the countries studied. Taken at face value, that would suggest many Sub-Saharan African economies are approaching or have passed the income level at which growth begins to clean the air. But the authors are careful to note that this relationship weakened under the robustness analyses, meaning the Kuznets result should be treated as suggestive rather than definitive. The implication is sobering: the region cannot simply wait for growth to solve its emissions problem, because the statistical foundation for that optimism is fragile.

What, then, should policymakers take away from this work? The authors argue that environmental sustainability in Sub-Saharan Africa requires an integrated policy package rather than isolated interventions. Health systems should be developed in ways that are themselves environmentally sustainable, preserving the emission-reducing association the data reveal. Education systems should embed green content so that human capital formation translates into environmental gains. Energy policy should accelerate the transition toward cleaner sources, since energy use remains a stubborn and robust driver of emissions. And foreign investment should be governed through stronger environmental frameworks, so that capital inflows bring modern clean technology instead of displaced pollution. Together, the authors contend, these measures can advance the Sustainable Development Goals, particularly SDG 13 on climate action, while supporting the region’s broader development ambitions.

The study’s limitations and strengths deserve equal attention. It relies on publicly available secondary data from the World Bank’s World Development Indicators, which means measurement quality depends on national statistical capacity, and it examines carbon dioxide emissions rather than broader measures of environmental degradation such as deforestation or biodiversity loss. The panel covers 28 countries over 21 years, a solid but not exhaustive sample of the region. Still, the combination of dynamic panel econometrics, cross-sectionally robust standard errors, and multiple human capital proxies represents a methodological standard that few regional studies have met. As Sub-Saharan Africa’s economies grow and its energy demand expands, the question of whether development can be decoupled from emissions will only grow more urgent. This research offers a clear, data-driven message: the region’s environmental future will be shaped not by growth alone, but by the specific composition of its investments, in health, in education, in energy, and in the rules it sets for global capital.

Subject of Research: Drivers of carbon dioxide emissions and environmental degradation in Sub-Saharan Africa

Article Title: Examining the drivers of environmental degradation in Sub-Saharan Africa using system generalized method of moments

Article References: Bezabh, G. B., Tefera, M. W., Wudie, D. F., Ayalew, H. A., Baylie, M. M., & Mengistu, Y. A. (2026). Examining the drivers of environmental degradation in Sub-Saharan Africa using system generalized method of moments. Discover Sustainability. https://doi.org/10.1007/s43621-026-04997-7

Image Credits: AI Generated

DOI: 10.1007/s43621-026-04997-7

Keywords: Sub-Saharan Africa, carbon dioxide emissions, system GMM, human capital, health expenditure, education expenditure, foreign direct investment, Pollution Haven Hypothesis, Environmental Kuznets Curve, energy use, Driscoll-Kraay estimation, SDG 13