Dirty Air, Weaker Growth: How Pollution Eats Into Human Capital

Environmental pollution can weaken economic performance by damaging the health of the workforce and reducing the productive value of human capital, according to a cross-country study published in Economies. The paper, "Environmental Pollution, Health, and Economic Development: A Human-Capital Perspective," by Yessengali Oskenbayev and Kanat Abdulla, examines how PM2.5 exposure, pollution-attributable mortality, education, energy use and income per capita are linked across countries.

The study finds a strong negative relationship between mortality and income in simpler empirical models. After accounting for climate and broad regional differences, the estimated coefficient on log mortality ranges from about −0.67 to −0.83. Once education and energy use are introduced, however, the relationship weakens substantially and loses statistical significance in the fullest specification.

Rather than isolating mortality as a standalone driver of economic performance, the analysis shows how closely health is bound up with education, energy use and broader economic structure. Pollution appears to affect productive capacity through a network of overlapping development conditions rather than through a single measurable pathway.

Pollution reduces more than health: it can diminish productive capacity

The authors build their argument around a human-capital model in which pollution raises the effective cost of maintaining good health. As those costs increase, health investment falls, effective human capital declines and workers become less productive.

The model treats health and schooling as complementary components of human capital. Education may increase a worker's skills, but poor health can reduce the extent to which those skills are used productively. Environmental deterioration can therefore affect both the formation of human capital and its later contribution to economic output.

The empirical analysis draws on country-level data covering PM2.5 exposure, pollution-attributable mortality, GDP per capita, average years of schooling, energy use, urbanization, precipitation and geographic characteristics. Pollution exposure and mortality measures refer to 2016, while the number of countries varies across models because missing observations are not filled in.

PM2.5 exposure is used as an instrument for mortality in the instrumental-variable estimates. The authors stress that this strategy requires caution because particulate pollution may influence income through channels other than mortality, including labor supply, productivity, cognition, migration and educational performance.

Mortality tracks sharply with income until broader development factors enter

The basic relationship between health and income is pronounced. In the ordinary least squares estimate, the coefficient on log mortality is −0.986, indicating that countries with higher pollution-attributable mortality tend to have substantially lower GDP per capita.

The instrumental-variable estimate remains strongly negative at −0.769. Adding precipitation and latitude produces a coefficient of −0.833, while replacing latitude with broader regional controls reduces the estimate to −0.669.

These results show that the mortality-income relationship is not eliminated by broad differences in climate or geography. Countries with poorer pollution-related health outcomes continue to record lower income levels even after those characteristics are taken into account.

The pattern becomes less clear once the models begin to capture additional dimensions of development. Average years of schooling reduce the mortality coefficient from −0.669 to −0.474, indicating substantial overlap between health conditions and educational attainment.

Education is not simply another control variable in this framework, but a core component of human capital and is strongly associated with both mortality and income across countries, making it difficult to attribute the observed income gradient to health alone.

Energy use reveals how difficult it is to separate health from development

The most demanding model adds energy use alongside education. Under this specification, the mortality coefficient falls further to about −0.173 and is no longer statistically significant at conventional levels.

Energy use is closely connected to the scale and structure of economic activity, which is why the authors do not treat this final estimate as a definitive measure of the health effect. Conditioning on energy removes variation that may already be closely tied to the development process the study is trying to understand.

The weakening of the coefficient is not mainly a consequence of losing countries from the sample. Common-sample comparisons show that the estimate still falls when education and energy controls are added while holding country composition more stable.

Leave-one-country-out tests produce the same general pattern. The mortality coefficient remains negative when individual countries are removed one at a time, but the fullest education-and-energy specification remains statistically insignificant throughout those tests.

The evidence points to a tightly connected development structure. Mortality, schooling and energy use capture different aspects of national economic conditions, but they also move together strongly enough that separating one from the others becomes increasingly difficult.

Cleaner air policy intersects with health, education and energy strategy

For lower-income countries with high particulate exposure and limited health capacity, the findings support combining pollution reduction with basic public-health protection and human-capital investment. Improving environmental quality can help preserve the productive value of workers rather than functioning as a separate environmental objective.

The same logic applies in countries where educational attainment has improved while air pollution remains high. Greater schooling raises the stock of human capital, but persistent health damage can reduce the productivity of that investment.

Energy policy adds another layer. Economies that depend heavily on pollution-intensive energy systems may face simultaneous pressures involving industrial development, health damage and environmental degradation. The paper does not estimate the effects of carbon taxes, emissions trading, renewable-energy subsidies or other specific instruments, so it does not rank policy options.

Several limitations narrow the claims that can be made from the results. Pollution-attributable mortality captures only the most severe health outcome and does not include morbidity, disability, cognitive impairment, school absence or reductions in workplace productivity.

The cross-sectional design also prevents the authors from controlling for persistent country-level characteristics through fixed effects. Institutions, culture, geography and other long-standing differences may still influence both health and economic performance.

The instrumental-variable strategy presents an additional difficulty because PM2.5 exposure contributes to the construction of the WHO pollution-attributable mortality measure. A strong first-stage relationship between pollution and mortality therefore cannot be treated as independent validation of the instrument.

The authors propose stronger identification strategies for future work, including panel data, within-country pollution shocks, regulatory changes, thermal inversions and wildfire exposure. They also call for broader measures of human capital covering morbidity, cognition, disability, schooling quality and labor-market productivity.