Empirical Evidence for the Impact of Environmental Quality on Life Expectancy in African Countries

Background. Protecting the health of citizens is a central aim of sustainable development plans, due to the effect of health on social and economic development. However, studies show that environment-related diseases adversely affect the health status of a people, and this situation is worse for African countries. The Sustainable Development Goals (SDG) targets have included reducing environment-related deaths since 2015. However, there is a lack of empirical findings focused on the effects of environmental quality on life expectancy in Africa. Objectives. The present study examined the impact of environmental quality on life expectancy in 24 African countries. Methods. Time-series data ranging from 2000 to 2016 was used and the panel autoregressive distributed lag (ARDL)–dynamic fixed effect (DFE) model was employed to analyze the data. Results. The results confirmed that, in the long run, improvements in environmental quality significantly increased life expectancy in the studied African countries during the study period. A unit increment in environmental performance index (EPI) and ecosystem vitality (EV) increased the life expectancy of Africans by 0.137 and 0.1417 years, respectively. Conclusions. To the best of the authors' knowledge, this is the first empirical (econometric) study using a broad measurement (indicator) of environmental quality to investigate its impact on life expectancy in African countries. The study recommends that the introduction of environmentally friendly economies (like renewable energy, land, water, and waste management), legal, socio-economic, demographic, and technological measures are essential to reduce environmental pollution and improve life expectancy in Africa. Competing Interests. The authors declare no competing financial interests.


Introduction
According to the Alma-Ata Declaration, protecting human health is vital for human welfare and social and economic development. 1,2,3 Healthy people can positively contribute to the economic growth and development of a country, directly and indirectly, by providing labor to different sectors. This improves per capita income and reduces poverty and income inequality. Governments should actively seek to preserve their people's lives and reduce the incidence of unnecessary mortality and avoidable illnesses. 4 Multiple factors (economic, political, social, and environmental) affect human health. Globally, around one-quarter of all diseases and deaths (about 13 million deaths each year) are caused by environmental problems. 4-7 Studies show that air pollution results in seven million deaths per year globally, more than 90% of the world's population breathes polluted air, and almost 3,000 million people still depend on polluting fuels. 7, 8 Studies also show that more than half of the world's population is living with unclean water, poor sanitation and hygiene, which causes more than 800,000 deaths each year. 7, 8 Furthermore, the lack of effective environmental management results in 400,000 deaths annually from malaria and 700,000 deaths from other vector-borne diseases. 7,9 Statistics show that 23% of all deaths globally are environment-related. Specifically, there were 3.8 million environment-Research to Africa; according to the WHO, 7 the continent is fighting the world's biggest public health crisis. Around 2.2 million deaths each year are reported to be caused by environment-related diseases in Africa. Air pollution alone is responsible for 600,000 deaths every year across the continent. In Africa, approximately 66% of children live in homes where solid fuels are used for cooking and heating, 11 which was responsible for 400,000 deaths in the region in 2017. 12 Countries in sub-Saharan Africa (SSA) are also affected by serious environmental problems, such as soil erosion, deforestation, desertification, insect infestation, and wetland degradation. 13 In addition, carbon dioxide (CO 2 ) emissions are an important environmental problem in SSA countries, at over 0.8 metric tons per capita since 1990 and reaching a maximum of 0.93 metric tons in 2003. 14 Rapid environmental degradation is another crucial problem in the Horn of Africa. 15 According to Mutai, 16 overgrazing, deforestation, shortage of water, loss of biodiversity, and industrial pollution in urban areas are the most serious environmental problems in East Africa and the Horn of Africa.
Policymakers regard the link between environmental quality and health status as a cause of increasing global concern and their 2030 agendas include plans to reduce risks to health from the environment. The global Sustainable Development Goals (SDGs) comprise seventeen goals in all. The third goal, titled 'Good Health and Well-being' , is concerned with issues that directly and indirectly concern the health status of humans. 10 In particular, the third and sixth goals focus on the environment and health. 17 Another SDG target is the reduction of the environmental impact of cities on human health by minimizing air, water, and soil pollution and putting into practice appropriate waste management. 18 However, few empirical studies exist that assess the impact of environmental quality on life expectancy in African countries, even though these countries support people that are often disproportionately exposed to environment-related diseases. The few studies that do exist are often out of date. They therefore do not use the latest methodologies, and their measurements of environmental quality are too specific, i.e., carbon dioxide (CO 2 ), sulfur dioxide (SO 2 ) , or other indicators. All of the above considerations lead to gaps in the literature, methodology, and appropriate measurements that this study aims to overcome. Specifically, this study addresses the methodology deficit by employing panel data autoregressive distributed lag (ARDL)-dynamic fixed effect (DFE) estimation, and the issue of measurement by using broad indicators (environmental performance index (EPI), and ecosystem vitality (EV)) to examine the impact of environmental quality on life expectancy in the case of 24 African countries from 2000 to 2016. Due to unavailability of data especially on target variables (environmental quality) this study is limited to only 24 African countries.

Empirical literature review
This section presents empirical studies on the impact of the environment on human health in general even though most studies used proxy variables (life expectancy, mortality rate, and health expenditure) to measure health status. Table 1

Methods
Depending on the nature of the data, panel time series models are classified as static models (pooled OLS, fixed effects (FE), and random affects (RE)) or dynamic models (generalized methods of momentum (GMM) and ARDL). Unlike static models, dynamic models capture the dynamic nature of the data. The ARDL model differs from other dynamic models in that it can be used if the variables in the model have the same, different, or mixed order of integration. Furthermore, it provides both long-run and short-run estimation results at the same time 32,33 and, even if endogeneity problems exist, ARDL models provide reliable coefficients. 34 The DFE estimator is one of the methods used for estimation of panel ARDL. The DFE estimator imposes homogeneity restrictions on both the long-run and short-run estimation coefficients, although it allows the intercept to be heterogeneous. 35

Equation 2
Equation 3 Equation 4 Research Results Table 3 provides descriptive and econometric results which are inputs and preconditions for more advanced analysis. Specifically, it has information about descriptive statistics, crosssectional dependence, unit root, and co-integration tests of the models.
Following the basic diagnostic and panel data econometric tests which are presented above, Table 4 presents the long-run and short-run estimation results.
In addition to the results in Table 4, a causality test was conducted to check the robustness of the results (Table 5).
Even though the present study estimated linear models (Equation 2), it also estimated the adopted log-linear model of Grossman,44 Fayisa and Gutama, 45 and Mutizwa and Makochekanwa 2 to check the robustness (sensitivity analysis) of the results in Table 6 and 7.  Table 3). The results confirm the existence of a long-run relationship among the variables in both models, which leads the study for long-run and short-run estimations.

Long-run and short-run results
The present study estimated two different models to obtain robust and reliable results concerning the impact of environmental quality on life expectancy. Both models showed that improvements in environmental performance and ecosystem viability positively and significantly increased life expectancy. A unit increment in EPI and EV increases life expectancy in Africa by 0.137 and 0.1417 years, respectively (Table 4). When we compare these results with other variables, their influence seems small, but this might be due to differences in measurement. That means that all independent variables are measured in either percentage or monetary terms; however, EPI, EV, and POLITY 2 are indices and the dependent variable (life expectancy) is measured in number of years. Therefore, these measurement differences can make the effect of EPI and EV on life expectancy smaller than those variables measured as a percentage.
The estimated results can be explained by the fact that improvements in environmental quality can reduce pollution (air, noise, chemicals, water) and the loss of natural areas, which also contribute to a substantial reduction in rates of environmentrelated diseases. This study suggested one resource which is available in Africa to improve environmental quality and life expectancy.   Table 4).
The positive impact of per capita GDP on life expectancy agrees with the relationship described by the "Preston curve", which implies that, on average, individuals born in wealthier countries can expect to live longer than those born in developing countries. 55 In other words, when the economy improves, the per capita income of individuals increases and hence they have better health.
The present study also shows unemployment significantly increases life expectancy, which is an unexpected but not surprising result. This finding may be due to the inappropriateness of the global definition and measurement of unemployment in developing countries like those in Africa. Even though "unemployed" is generally defined as applying to a person who is actively searching for employment but unable to find work, due to the lack of well-structured and more formalized labor markets, most "unemployed" people in Africa are participating in the informal economy, which hides the adverse effects of unemployment. According to the International Labor Organization (ILO), 56 in Africa, 85.8% of employment is informal, which suggests that unemployed people in Africa can gain income by participating in the informal economy and hence they can improve their health status. Contrary to the long-run results, in the short run, EPI and EV significantly reduced life expectancy in the sampled countries. This suggests that, in the short run, during the economic transformation from agricultural to industrial economies, manufacturing companies have limited waste management facilities or strategies and generally dump waste in rivers. Furthermore, in the short run, environmental quality in African nations is poor due to weak environmental protection strategies and awareness.
Nonetheless, urbanization positively significantly increases life expectancy in African nations, both in the short and long term. Government expenditures also makes a positive contribution to life expectancy in the short term. However, unemployment negatively reduces life expectancy of African people in the short run.

Beyene, Kotosz
Environmental Quality and Life Expectancy in African Countries  Research which might be due to severe income inequality in African countries.
In both estimated models, the coefficients of one lagged period of the error correction term (ECM) were highly significant and fell between 0 and -1, which implies the existence of a long-run relationship among the variables. The coefficients of ECM are -0. 0423 and -0. 0410 for Models 1 and 2, respectively, which implies every year about 4% (too slow to be equilibrium) of the deviation from the long run is corrected.

Robustness check analysis using alternative models
The present study conducted the Granger causality test in order to confirm whether environmental quality has a potential causality for life expectancy. The results demonstrated that the null hypothesis that change in both EPI and EV does not homogeneously cause LIFEXP is rejected at a 1% level of significance, suggesting that improved environmental quality can increase life expectancy (Table 5).
Since per capita GDP, which was used in the previous section, is comparable with either GDP growth (GDPGR) or GDP per capita growth (GDPPCG),  3 and 4) are estimated by substituting GDPPCG by GDPPC (Table 6). The study also estimated the equations using the natural logarithm of life expectancy (LNLIFEXP) and GDPGR ( Table 7).

Beyene, Kotosz Environmental Quality and Life Expectancy in African Countries
As an alternative result, the study used a natural log of life expectancy and GDPGR because all variables in the models were measured as either an index or percentage. Therefore, the study used the natural logarithm of life expectancy and GDPGR to consistently measure all variables and interpret the regression coefficients as elasticity (percentage).
Similar to the previous results, the robustness analysis also conducted a cross-sectional dependence test, panel unit root tests, and cointegration test before estimation. The result confirms that there is no cross-sectional dependence in any of the models and all alternative variables (LNLIFEXP, GDPPCG, and GDPGR) are stationary at level (I(0)). In addition, the Kao cointegration test confirms the longrun relationship among the variables in the model. To save space, this study does not report those results here, but they are available from the authors.
The long-run and short-run estimated results show that both EPI and EV have a positive and significant impact on life expectancy in African nations, which is consistent with the previous findings. The results in Table 6 confirm that a unit increase in EPI and EV increases life expectancy by 0.1134 and 0.137 years, respectively. Similarly, a unit increment in EPI and EV leads to a rise in life expectancy by 0.18 and 0.37%, respectively (Table 7).

Conclusions
Protecting human health from risk is vital for human welfare, and social and economic development. Among the factors affecting human health, low environmental quality is a leading cause of death in LMIC countries. Currently, the SDGs have many targets for reducing environment-related deaths. Therefore, this study aimed to examine the impact of environmental quality on life expectancy in selected African countries between 2000 to 2016, employing the panel ARDL -DFE estimation technique. In long-run models, the study found that improvements in EPI and EV significantly increased life expectancy in African countries. The study recommends that African countries should give high priority to improving environmental quality by developing policies to stimulate environmentally friendly economies (green economies) using resources such as renewable energy (electricity generation, air and water heating/cooling system), effective land management (reforestation), and encouraging optimum use of water resources. In addition, the waste management systems of African nations should be improved. Furthermore, developing and applying strict rules and regulations to protect the environment and other technological measures can help in improving the quality of the environment as well as human health. Achieving rapid and sustainable economic growth (especially in productive sectors) is essential to provide basic health facilities, a better quality of life, and to increase longevity for the people of Africa.
In addition, African states should work to improve the general level of education of their people, which would make it easier to raise awareness of the earth's natural resources. Africans also need demographic measures like a reduction in birth rates, which can substantially reduce population density and environmental pollution and improve life expectancy.
Finally, although the present study has tried to satisfactorily address the literature, measurement, and methodological aspects of the effect of environmental quality on life expectancy, it has limitations. Lack of data on environmental quality limits the time frame of this study to conditions after 2000, and therefore it did not take into account major policies and environmental agreements in the 1980s and 1990s such as health policy on primary health care, the United Nations framework on climate change, and the Kyoto protocol. In addition, due to data constraints on the main target variables (environmental quality) this study is limited to 24 African countries. Furthermore, it would have been interesting to estimate either one or two-step system GMM. However, the life expectancy (LIFEXP) variable is not integrated (I(0)) and all other variables are integrated at the first difference (I(1)) ( Table 3); hence, this study could not employ the GMM estimation for an additional robustness check. Therefore, future studies should consider these factors in their analyses.