Investigating the role of health, education, energy and pollution for explaining total factor productivity in emerging economies

The paper investigates how health, education, energy use, and pollution influence total factor productivity (TFP) across 30 emerging economies. Motivated by growth literature emphasizing human capital’s central role (Barro, Mankiw, Romer, Lucas), the study argues that aggregate indicators like life expectancy and average years of schooling may miss important dimensions of human capital quality. It proposes examining disaggregated health indicators (undernourishment, access to clean water, HIV) and education by level (primary, secondary, tertiary) alongside environmental quality (air pollution), energy consumption, trade openness, ICT, and industrial share. The research aims to provide a more policy-relevant understanding of the drivers of TFP, clarify the TFP–energy–pollution nexus, and identify which components of human capital most strongly relate to productivity in emerging economies.

Prior research links human capital to TFP and growth through innovation, labor market efficiency, and complementarities (Romer, Lucas; Cole and Neumayer). Early empirical studies typically use aggregate health (life expectancy) and education (average years of schooling), which may not capture all relevant dimensions (WHO, 2002). Literature on pollution and growth is extensive (Esteve and Tamarit; Aslan and Gozbasi; Dogan and Ozturk; Mikayilov; Adedoyin; Zhang), but less work targets TFP specifically. Evidence on the TFP–energy–pollution nexus is emerging (Rath et al.; Ladu and Meleddu; Ackah and Adu; Tugcu and Tiwari; Dogan et al.), often with limited variable sets and without disaggregated human capital. Environmental quality can affect productivity via health or through production inputs, yielding ambiguous net effects (Schlenker and Walker; Graff Zivin and Neidell; Diao and Roe). ICT can enhance TFP through innovation, diffusion, and education quality improvements (Shapiro et al.; Kim and Narasimhan; Kim et al.). The paper addresses gaps by jointly modeling disaggregated health and education, environmental quality, energy, and complementary macro variables for TFP.

Design: Panel data analysis for 30 emerging economies with 30 annual observations each (900 country-year observations). Fixed effects (FE) models are estimated after a Hausman test rejects random effects (e.g., chi-square values such as 48.14 with p=0.000 across specifications). Data: TFP from Penn World Table 10.0; education (average, primary, secondary, tertiary years of schooling) from Barro and Lee (2020); health indicators from World Bank (2020): life expectancy, HIV prevalence (ages 15–49), access to at least basic drinking water, prevalence of undernourishment; environmental quality: PM2.5 exposure above WHO guideline (% population); macro controls from World Bank (2020): trade openness ((exports+imports)/GDP), energy use per capita, industrial share of GDP, ICT (fixed telephone subscriptions per 100 people). Descriptive statistics and correlation matrix indicate no severe multicollinearity (all pairwise correlations < 0.9). Model: Semi-log (log–log) FE regressions of the form: ln(TFP)_it = θ0 + θ1 ln(X_it) + θ2 ln(trade_it) + θ3 ln(energy_it) + θ4 ln(industry_it) + θ5 ln(ICT_it) + ε_it, where X denotes health or education indicators (both aggregated and disaggregated, entered in separate specifications). Three models include primary, secondary, and tertiary schooling separately; a fourth uses average years of schooling. Additional specifications incorporate health and environmental indicators (undernourishment, water access, air pollution, HIV, life expectancy). Model selection and signs are checked for theoretical consistency; FE chosen via Hausman test with p<0.05 in all reported models. The study reports coefficients and standard errors; significance denoted by conventional levels (* 10%, ** 5%, *** 1%).

- Education: In FE models, coefficients for schooling by level are positive and statistically significant, with increasing magnitude by level: primary 0.038 (p<0.05), secondary 0.045 (p<0.10), tertiary 0.091 (p<0.05), indicating tertiary education has the strongest association with TFP. Using average years of schooling yields a positive coefficient of 0.087 (p<0.05). - Health and environment: Life expectancy is positively related to TFP (approx. 0.11, significant). Morbidity-related indicators exhibit negative associations: undernourishment −0.140, lack of access to clean water −0.039, air pollution −0.073, and HIV prevalence −0.146 (all reported as adverse and significant in FE specifications). - Energy, industry, ICT: Energy use per capita, industrial share of GDP, and ICT proxy are positively associated with TFP across specifications (signs and significance consistent with theory), suggesting structural transformation and technological penetration support productivity. - Trade openness: Trade openness shows a negative coefficient in these emerging economies’ FE models, implying that openness has not translated into productivity gains, potentially due to limited technology transfer; highlights need for policies that foster spillovers. - Model diagnostics: Hausman tests consistently reject random effects in favor of fixed effects (e.g., chi-square ~33–48; p=0.000). Correlations suggest no severe multicollinearity, supporting stable estimation. - Policy-relevant synthesis: While energy use supports TFP, its benefits may be offset by pollution increases; optimizing the energy mix toward cleaner sources is necessary to realize net productivity gains.

The findings substantiate that both the quantity and composition of human capital are central for explaining cross-country TFP differences in emerging economies. Disaggregating education shows that higher-level (tertiary) skills contribute more to productivity, consistent with specialization and innovation channels. Disaggregated health indicators reveal that addressing undernourishment, water access, HIV, and air pollution is crucial because these factors depress labor effectiveness and human capital quality, thereby lowering TFP. Positive impacts of energy use, industrialization, and ICT align with mechanisms of technological adoption, structural transformation, and network externalities, but environmental degradation can counteract energy-related gains, underscoring the importance of cleaner energy and environmental governance. The negative association of trade openness with TFP suggests that, absent effective absorptive capacity and technology diffusion policies, openness alone may not enhance productivity. Overall, the results support a policy mix that simultaneously invests in health, education quality and levels, ICT-enabled learning and production, cleaner energy, and frameworks that convert openness into genuine technology transfer to boost TFP.

The study concludes that prioritizing human capital—through improved health and education at all levels—significantly enhances TFP in emerging economies. Key policy implications include targeted actions to reduce undernourishment, expand safe drinking water access, and curb HIV and air pollution, coupled with investments in ICT and alignment of education with industrial needs, especially at the tertiary level. Policymakers should also reassess the energy mix to mitigate pollution’s offsetting effects on productivity and design trade and integration policies that foster effective technology transfer. Advancing these measures supports sustainable growth and aligns with SDGs, particularly universal access to clean water and improved nutrition. A bottom-up implementation via empowered local governments and focused social safety nets is emphasized to translate these priorities into productivity and welfare gains.