THE LINK BETWEEN ENVIRONMENTAL PRESSURES AND LIFE EXPECTANCY: FROM SPATIAL DEPENDENCE TO INSTITUTIONAL CHALLENGES IN ASIA

Abstract

Introduction: Environmental pressures and institutional capacity are increasingly intertwined determinants of population health in Asia. Yet most cross-country life-expectancy studies treat countries as independent units and rarely distinguish own-country effects from spatial spillovers. This study examines the associations of carbon dioxide (CO2), nitrous oxide (N2O), and fine particulate matter (PM2.5) with life expectancy while jointly considering health expenditure, education, government effectiveness, natural-resource rents, spatial dependence, institutional moderation, and a health-expenditure pathway.
Methods: The analysis uses a panel of 39 Asian countries comprising 1,014 country-year observations in the supplied analytical results. Country fixed-effects models are estimated separately for CO2, N2O, and PM2.5. Hausman tests support fixed effects. Spatial autoregressive (SAR), spatial error (SEM), and Spatial Durbin Model (SDM) specifications are compared using a row-normalized inverse-distance weights matrix matched to 39 countries. The SDM is used to estimate spatial lags and to decompose pollutant effects into direct, indirect, and total impacts. Government effectiveness is evaluated as a moderator, health expenditure as a potential pathway, and alternative spatial-weight normalization is used for robustness.
Results: Life expectancy averages 71.84 years. In fixed-effects models, CO2 (β=0.1644, p<0.01) and N2O (β=0.1490, p<0.01) are positively associated with life expectancy, whereas PM2.5 is negatively associated (β=-0.0504, p<0.01). Health expenditure, education, and government effectiveness are consistently positive. SDM results preserve these signs: CO2 β=0.1566, N2O β=0.1513 and PM2.5 β=-0.0499 (all p<0.01). Spatial lags of life expectancy and pollutants are individually insignificant, and indirect pollutant effects are not statistically distinguishable from zero. However, total effects are significant for N2O (0.2853, p<0.05) and PM2.5 (-0.0909, p<0.05). Government effectiveness attenuates the negative PM2.5 association (PM2.5×GE=0.0196, p<0.05), while the CO2×GE interaction is positive (0.1698, p<0.01). Pollution is negatively associated with health expenditure for CO2 and PM2.5, but the supplied results do not report a bootstrapped indirect effect; therefore, mediation significance is not claimed. Core coefficient signs are preserved under alternative spatial-weight normalization.
Conclusions: The results show that PM2.5 is the most consistent environmental threat to longevity in the sample, while health expenditure, education, and effective government are robust positive correlates of life expectancy. Positive CO2 and N2O coefficients should not be interpreted as health benefits of emissions; rather, they point to the development-pollution-health bundle and possible residual confounding in macro panels. Spatial spillovers are weaker than direct effects, but the total N2O and PM2.5 impacts indicate that regional externalities cannot be ignored. Health policy in Asia should combine particulate-pollution control with stronger health-system financing, education, and governance capacity.