West African Study Finds No Detectable Link Between Carbon Intensity and Mortality

A six-year ecological panel study of eleven West African countries found no statistically detectable association between the carbon intensity of economic development and premature noncommunicable disease mortality among adult males from 2015 to 2020. The observational research evaluated whether national development pathways that decrease carbon intensity compromise or support population health gains in resource-constrained environments. Across several panel regression specifications, the researchers recorded small and imprecise estimates that could neither demonstrate a measurable health penalty nor prove equivalent health outcomes between lower-carbon and carbon-intensive economic trajectories.

The inquiry addressed longstanding debates concerning the public health consequences of pursuing low-carbon growth trajectories, frequently described as green growth, within resource-limited settings. West African nations have followed divergent carbon-intensity trends over recent years, providing an observational basis to examine health outcomes across different development paths. The authors sought to evaluate whether pursuing economic expansion with lower greenhouse gas emissions per unit of gross domestic product compromised health progress. In particular, the study analyzed whether shifts in carbon intensity correlated with changes in premature mortality from chronic diseases among working-age men.

To assess these patterns, the team compiled national panel records from the World Bank World Development Indicators database covering the 2015 through 2020 period. The eleven West African countries were retrospectively grouped into two development trajectories according to their observed changes in greenhouse gas emissions per unit of gross domestic product. Five nations achieved economic decoupling, characterized by decreasing carbon intensity, with a mean reduction of 13.8 percent. Meanwhile, six countries followed an intensifying trajectory, showing an average carbon intensity increase of 24.5 percent over the same six-year span.

The primary outcome assessed across the panel was the probability of noncommunicable disease mortality among males aged 30 to 70 years. To evaluate this metric, the researchers estimated pooled ordinary least squares, country fixed-effects, and two-way fixed-effects panel regression models. Because the dataset included only eleven country clusters, standard inference posed risks of bias; the authors therefore utilized bias-reduced CR2 cluster-robust standard errors and the wild cluster bootstrap alongside conventional cluster-robust standard errors. The investigation further incorporated a supplementary difference-in-differences trajectory comparison, Baron-Kenny mediation testing, and an effect-modification assessment evaluating government health expenditure.

Across all regression models, estimated associations between carbon intensity and male noncommunicable disease mortality remained small and statistically imprecise. In the crude pooled model, the relationship was negative but imprecise, showing a standardized beta of -0.28, a wild cluster bootstrap 95 percent confidence interval from -1.71 to 1.40, and a p-value of 0.70. Within-country estimates from fixed-effects models were similarly inconclusive. In the fully adjusted two-way fixed-effects model, the beta was 0.05, with a 95 percent confidence interval of -0.24 to 0.23 and a p-value of 0.66, compatible with mortality shifts from 0.9 percentage points lower to 0.8 percentage points higher per 0.1 kilogram carbon dioxide equivalent increase per dollar.

Supplementary analyses exploring group trajectories and potential explanatory mechanisms likewise found no statistically detectable differences. A difference-in-differences model comparing trajectory groups indicated an estimated differential change in noncommunicable disease mortality of 0.01 percentage points between decoupling and intensifying countries, with a 95 percent confidence interval between -0.80 and 0.77 percentage points and a p-value of 0.99. Neither mediation nor effect modification by government health spending was statistically detectable. While government health spending as a share of gross domestic product grew more in decoupling countries than in intensifying nations, with average relative gains of 92.3 percent versus 28.8 percent, this difference remained statistically nonsignificant.

Given the persistent statistical imprecision, the researchers highlighted that their results do not prove that lower-carbon and carbon-intensive growth pathways yield identical population health outcomes. The wide confidence intervals could not rule out small or moderate associations in either direction. The authors emphasized that determining whether economic decarbonization generates a health penalty or benefit across West Africa will require longer time series, larger samples encompassing more countries, and methodological designs capable of isolating plausibly exogenous variation in national development pathways.

Subject of Research: Carbon intensity of economic development and premature population health outcomes in West Africa

Article Title: Carbon intensity of economic development as a predictor of population health outcomes: an ecological panel analysis of West African countries, 2015–2020

Article References: Oyareme, V., Barrow, A., & Kuye, R. A. (2026). Carbon intensity of economic development as a predictor of population health outcomes: an ecological panel analysis of West African countries, 2015–2020. Journal of Health, Population and Nutrition. https://doi.org/10.1186/s41043-026-01520-6

Image Credits: AI Generated

DOI: 10.1186/s41043-026-01520-6

Keywords: carbon intensity, economic development, population health, West Africa, noncommunicable disease mortality, ecological panel analysis, greenhouse gas emissions, economic decoupling, two-way fixed effects, wild cluster bootstrap, CR2 standard errors, government health expenditure, difference-in-differences, Baron-Kenny mediation, male mortality probability, World Development Indicators

Tags:

 

Share Story:

Facebook
X
LinkedIn