Growth in Southern Africa Still Runs on Carbon, 25-Year Study Finds

Economic growth across the Southern African Development Community between 1997 and 2021 has come largely at the expense of environmental quality, according to a new open-access study published in Discover Energy. Using heterogeneous dynamic panel estimation techniques, researchers Nazatul Faizah Haron, Isaac Akpemah Bathuure and Insah Baba analysed World Bank World Development Indicators for SADC member countries and found a stable long-run relationship between economic performance and a set of environmental metrics, including per capita carbon dioxide emissions, emissions from electricity and heat generation, industrial and construction emissions, total carbon emissions and unemployment.

The study’s central empirical finding concerns emissions from manufacturing and construction. In the pooled mean group (PMG) long-run estimates, a one-unit increase in these emissions is associated with approximately a 1.024 increase in gross domestic product, holding other factors constant. The authors interpret this direct, statistically significant link as evidence of the carbon-intensive character of growth in the region, where industrial output expands through fossil fuel-dependent processes. This pattern, they argue, is consistent with the early, upward-sloping phase of the Environmental Kuznets Curve, in which industrialisation drives output and emissions upward together.

The Environmental Kuznets Curve hypothesis proposes an inverted U-shaped relationship between income and environmental degradation: pollution worsens in the initial stages of economic growth but improves once income crosses a certain threshold, as cleaner technologies and stricter regulation take hold. The model tested this by including a squared income term, with the hypothesis requiring a positive coefficient on income and a negative coefficient on squared income. The authors report partial confirmation of the hypothesis, describing a “pollute-first, clean-later” development pattern typical of developing economies, while cautioning that many SADC countries still appear to be in the early phase where growth and rising emissions remain tightly linked.

Not every environmental variable behaved as expected. Per capita carbon dioxide emissions and emissions from electricity and heat production were not statistically significant in the preferred long-run specification, and renewable energy consumption also showed no significant long-run effect on economic growth. The authors attribute the marginal role of renewables to carbon lock-in, in which existing fossil fuel-based infrastructure impedes the transition to a cleaner economy, combined with strong industrial reliance on energy-intensive processes and current energy inefficiencies. Electricity generation in the region still relies heavily on coal-based systems, particularly in South Africa, while the hydropower potential of several member states remains largely underdeveloped.

Unemployment emerged as a significant macroeconomic factor. Its long-run coefficient was negative and statistically significant at −0.541, indicating that higher unemployment constrains economic growth through underutilisation of human resources and weaker aggregate demand. The study highlights persistently high joblessness in the region, exceeding 28 percent in South Africa and around 19 percent in Angola, and argues that growth which fails to create jobs cannot be considered inclusive or sustainable. The correlation analysis reinforced this picture: per capita carbon emissions were positively correlated with unemployment at 0.5722, while renewable energy consumption showed a notable negative correlation with unemployment at −0.3693.

Methodologically, the study applied the Mean Group and Pooled Mean Group estimators developed by Pesaran and Smith and by Pesaran and co-authors, with a Dynamic Fixed Effects model as a robustness check. The PMG estimator constrains long-run coefficients to be homogeneous across countries while allowing short-run dynamics and adjustment speeds to vary, an approach the authors justify by the shared structural features of SADC economies, including regional integration and comparable development challenges, alongside substantial differences in economic structures and energy mixes. A Hausman test comparing the estimators yielded a chi-squared statistic of 6.17 with a p-value of 0.4049, supporting the PMG restriction. The error-correction term was negative and significant in both the MG and PMG models, at −0.85 and −0.64 respectively, confirming that deviations from the long-run equilibrium are corrected over time.

The authors are explicit that the design does not support strict causal inference. Because economic growth and environmental outcomes are endogenously linked, and because the PMG framework cannot fully eliminate reverse causality or omitted variable bias, the results should be read as long-run associations rather than causal effects. System Generalized Method of Moments and instrumental variable techniques could address endogeneity more directly, but the authors report that instrument proliferation undermined these alternatives: in the unrestricted specification the instrument count of 308 exceeded the 16 cross-sectional units, and even after collapsing the instrument matrix to 53, the Hansen test remained at a p-value of 1.000, indicating probable overfitting. System GMM was therefore used only as a robustness exercise.

Robustness checks extended to sub-period analyses before and after the 2008 global financial crisis. Variables including manufacturing and construction emissions, electricity and heat emissions, and agricultural methane were omitted from the restricted sub-period models because structural zeros in later years caused convergence problems. Even with these omissions, the error-correction term remained negative and significant across specifications, and the short-run coefficient of per capita carbon emissions stayed positive and significant, supporting the association between carbon emissions and growth dynamics across both periods. The authors note that the DFE robustness model produced no statistically significant long-run coefficients, so its results serve only as a comparison rather than a primary basis for inference.

The study situates these findings in a region acutely exposed to climate change. Mozambique and Zimbabwe rank among the world’s most climate-exposed countries, and in 2019 Tropical Cyclone Idai killed more than 1,000 people and caused damages estimated at 2.2 billion US dollars across Zimbabwe, Malawi and Mozambique. At the same time, the subregion contributes less than 3 percent of global cumulative emissions, and its mean per capita carbon emissions were the lowest among the variables examined. Descriptive statistics also showed agricultural methane emissions with the highest mean value, and the third-highest mean carbon emissions from heat and power generation on the African continent.

On the policy side, the authors recommend scaling up industrial decarbonisation, promoting cleaner production technologies, improving energy efficiency and strengthening environmental regulation balanced with fiscal incentives and technical assistance. They point to South Africa’s Renewable Energy Independent Power Producer Procurement Programme as a benchmark for attracting private investment in wind and solar, and to the Southern African Power Pool as a platform for regional electricity market integration. Citing joint research from IRENA, RES4Africa and the UN Economic Commission for Africa, they note that Africa’s energy sector could create around 23 million jobs by mid-century under a 1.5-degree climate scenario, with almost 73 percent of those positions classified as green jobs. The study’s limitations include data availability constraints that shortened coverage for some countries and periods, and the restriction of findings to SADC economies, which limits generalisability to other subregions.

Subject of Research: Long-run relationship between economic growth and environmental sustainability in SADC countries

Article Title: Economic advancement and environmental sustainability in the SADC region

Article References: Haron, N. F., Bathuure, I. A., & Baba, I. (2026). Economic advancement and environmental sustainability in the SADC region. Discover Energy, 6(1), Article 93. https://doi.org/10.1007/s43937-026-00187-8

Image Credits: AI Generated

DOI: 10.1007/s43937-026-00187-8

Keywords: SADC, economic growth, carbon emissions, Environmental Kuznets Curve, renewable energy, unemployment, panel data, PMG estimator, ecological modernization, Southern Africa, climate vulnerability, energy transition