Research on human health

At present, domestic and foreign academic research on human health has achieved rich results, and the research content closely related to this paper can be summarized in three aspects:

Firstly, research on the theoretical connotation of human health. Initially, health is mainly interpreted from the perspective of bodily functions, referring to physiological health, which belongs to a narrow concept of health [13]. Health refers to the absence of physical disease and a better state of mental and social adaptation, in addition to moral health, on top of which there are 10 other criteria that are widely recognized in society, including a strong appetite, strong resilience, and so on. Among them, the physical aspect mainly refers to the human body being healthy, including no disability and disease, nutritional standards, etc. Our research was carried out from the macro level, and mainly focuses on physical health and does not consider psychological, social, and moral factors.

Secondly, research about the measurement of human health is considered. The current measurement methods on human health can be broadly categorized into three types: (1) Using individual representative indicators to characterize human health. Zhang et al. [14] used the medical total health costs as a negative measure of human health; Xing and Tian [15] chose the number of lung cancer deaths, the number of oncology outpatient emergencies, and the number of new lung cancer cases to negatively characterize urban human health. (2) The efficiency values of multiple inputs and multiple outputs are measured. Wang and Liang [16] measured the efficiency of urban government health expenditure in China using Bootstrap-DEA and window DEA analysis with per capita budgeted fiscal health expenditure as the input variable and medical institutions per capita, beds per capita, and health technicians per capita as the output variables. Yu et al. [17] used the DEA-Malmquist model to dynamically measure the productivity of residents’ health in China’s eight comprehensive economic zones. (3) There is also some literature that constructs a comprehensive indicator system related to health measurement. Wang [18] used the weighted average of life expectancy, birth rate, and death rate to construct a health level index for each region in China as a reflection of the effect of health investment in each region. Zhao et al. [19] constructed an evaluation index system for healthy cities from four aspects, namely, health dimensions, health services, environmental dimensions, and socio-economic dimensions, based on the existing studies.

Thirdly, research on the factors that influence human health is also considered. Schultz [20] emphasized that health is an important component of human capital. Subsequently, many scholars have conducted extensive research on the importance of healthy human capital to economic growth; Fogel [21] pointed out that an improvement in food consumption could improve individual health. In recent years, a few scholars have conducted research on the causal identification between the two, including by taking a single region as the object of research and using the Grossman function model to study air pollutants’ impact on health [22]. Some scholars use panel data and econometric modeling to study the impact of air quality and other elements on human health [23] and make a comparative analysis of the differences between men and women, as well as the differences between the health effects of residents in developed and developing countries [24]. Staff Mestl [25] used Monte Carlo simulation to study the impact of indoor air pollution produced by domestic fuel combustion on the health of residents in Shanxi Province.

Trade openness impact on human health

Existing research suggests that trade openness affects human health through food. For example, Ravuvu et al. [26] found that trade significantly increased the number of healthy food products, thereby promoting health. There are also some scholars who are opposed to this; Geof et al. [27] argued that trade liberalization, although only to a certain extent, provides food for poor countries to lift them out of poverty, but it may also make other countries over-nutritional through the transmission of unhealthy diets, thus adversely affecting their health.

Second, trade openness leads to technological spillovers that promote medical innovation and improve healthcare. For example, Papageorgiou et al. [28] found that trade openness facilitates medical technology spillovers, thereby improving human health. Kshetri and Dholakia [29] found that trade improves the level of healthcare services in a country and promotes their diversification, which effectively improves the quality of health of the local population. In addition, the system of trade in healthcare services guarantees the safety and reliability of the distribution of medicines to the population, which protects the fairness of the healthcare services. Li et al. [30] found that the expansion of trade not only improves medical services but also helps to enhance the demand for them among China’s residents, which, in turn, improves the health of China’s residents.

Third, trade openness affects human health through the environment. Scholars generally agree that environmental pollution can harm the health status of residents [31–32]. From the perspective of trade, a portion of scholars believe that international trade will bring cleaner production technology to a country, which reduces the country’s pollutant emissions, thus improving the living environment and health quality of its residents; for example, through the study of US data, Levinson [33] found that trade brings technological innovation, which reduces the pollution status of the environment and improves the level of health. Zhang [34] explored the impact of trade openness on the health of the population and found that international trade does not statistically significantly effect environmental pollution, and therefore, international trade does not affect the health of the population through environmental effects. However, some scholars believe that international trade will bring more serious environmental pollution, which, in turn, will harm residents’ health. Chen et al. [35] used a panel of 31 provinces and cities from 2000 to 2013 and found that international trade reduces residents’ health through the channel of environmental pollution. Bontems and Gozlan [36] similarly found that the environmental pollution caused by international trade worsens the health of low-income classes. Lin [37] explored trade liberalization’s impact on the health of China’s residents and found that trade openness exacerbated environmental pollution in China, which, in turn, increased the prevalence of diseases and harmed the health of China’s residents.

Environmental pollution impact on human health

The impact of the environment on health mainly depends on the concentration of pollutants, the exposure–response coefficient, and the size of the exposed population in the city. The analytical method based on the exposure–response and the dose–effect principle mainly focuses on research on the loss of health of residents caused by environmental pollution. Matus et al. [38] used a policy prediction model to analyze the fact that the health and economic losses caused by China’s air pollution emissions have caused a substantial burden on the Chinese economy. Chen et al. [39] studied the impact of PM10 on the health of residents in 113 major cities in China, and the results showed that PM10 has caused greater health loss in urban residents in China. In view of the frequent occurrence of the haze phenomenon and the high concentration of PM2.5 in China, Voorhees et al. [40] analyzed the human health loss of Shanghai residents caused by PM and showed that the monetary value of all the deaths that could be avoided if Shanghai’s air quality met the national secondary standards is CNY 1.7 billion to 12 billion. Pearce and Warford [41] pointed out that the most direct consequence of environmental degradation is to jeopardize human health through various physical or mental illnesses and even premature death. In addition, the degree of harm to human health from environmental pollution depends on the level of exposure to pollution, and long-term exposure to heavily polluted areas is more likely to cause disease [42]. Chen et al. [43] found that the heating policy based on the dividing line between the Qinling Mountains and the Huaihe River increased total pollution and shortened the expected life expectancy of residents in the northern region. Cheung et al. [44] constructed instrumental variables based on wind direction and air pollution in the surrounding area and found that air pollution increases the mortality rate of the population.

Research review

It can be concluded from the above literature that although the current theoretical research and methodologies regarding environmental pollution and trade openness impacting human health have achieved some results, with these results and methods providing better policy guidance for improving human health, this area still requires further improvement and development. First, most of the research methods are based on scenario simulation, and empirical studies are limited by factors such as data availability and external conditions. Scenario simulation is based on hypothetical conditions, and the influencing factors are very limited and deviate from the real situation; this needs to be supplemented by empirical studies. Second, there are fewer theoretical approaches applied to environmental pollution and human health, and the theoretical system is not complete enough, so it is necessary to continue to deepen the research and improve the integrated research model of environmental pollution and human health. Third, existing research is limited to the relationship between the two variables; trade openness can affect public health and also cause environmental pollution problems, so it is necessary to consider the three aspects within the same framework to carry out research more comprehensively and systematically than the impact of a single variable on health. This can be derived from the different directions and degrees of impact, which can provide a solution to the complex relationship between trade, the environment, and the health of the population. We can then realize the construction of a healthy, harmonious, fair, and efficient social environment to achieve the path forward. This can provide a reference for solving the complex relationship between trade, environment, and people’s health. Based on the theoretical foundation of international trade, health economics, environmental economics, and other related disciplines, as well as the research results of related scholars from home and abroad, this research tries to analyze the impact of environmental pollution and trade openness on the human health of China’s residents in a more systematic way, which is a supplement to and improvement of existing research.

Empirical design

The health economics community has carried out more research on the factors that influence health, among which Grossman’s health production function is the most classic. Grossman [45] established a health demand model as follows:

(1)

where H denotes population health, and X denotes the factors affecting health, such as education, economy, medical conditions, etc. Filmer and Pritchett [46] developed a new health production function on the basis of Grossman’s health production function, and set the model with reference to the new health production function:

(2)

where denotes human health in period t of region i, denotes the openness in period t of region i, denotes the environmental pollution in period t of region i, and is the control variable; is the intercept term, and is the random perturbation term. In order to transform the nonlinear relationship between each explanatory variable and human health into a linear relationship, and to smooth the series to reduce the effect of heteroskedasticity on the regression results, the natural logarithmic treatment is carried out for each variable, which results in the following equation:

(3)

Variable selection

(2) Explanatory variables.

Trade openness: Existing studies generally agree that trade openness improves human health for a number of reasons: first, trade openness causes domestic residents to consume more foreign products, bringing in healthy food, medical equipment, etc., increasing nutritional intake and reducing morbidity. Second, trade liberalization leads to technological spillovers that promote medical innovation and improve medical efficiency. Some scholars have also found that the negative impact of trade openness on human health is mainly reflected in the population’s nutrition, dietary habits, food safety, and the spread of disease; the expanding scale of foreign trade creates a greater health risk that can, to a certain extent, trigger epidemics, leading to the spread of disease. This paper draws on the practice of Lin [37] to construct the effective tariff rate for imported intermediary goods at the enterprise level. The corresponding calculation formula is the following:

(4)

where is the set of products imported, is the proportion of the amount of import product, p, imported, i, in year t to the total import amount, and is the import tariff rate of product p in year t at the level of HS6 code. By weighting the tariff rates of the intermediate inputs imported by enterprises, the effective tariff rate faced by each exporting enterprise can be obtained.

Environmental Pollution (EP): For the measurement of the degree of environmental pollution, most scholars believe that PM2.5, PM10, and other solid particulate matter are the main sources of pollution that affect the quality of the environment and harm human health; however, in reality, this type of data is not easy to obtain. This paper refers to a previous study selected from different areas of industrial sulfur dioxide emissions to characterize the quality of the environment of a region; sulfur dioxide emissions caused the expansion of acid rain, reduced yields of food and cash crops, caused the corrosion and destruction of building materials, and caused the population mortality rate to increase [49–52]. The main reasons for choosing sulfur dioxide emissions are as follows: First, compared with other environmental factors, the exogeneity of the air and exogenous determinants are stronger, which can reduce the endogeneity of this paper; second, the probability of regional residents being exposed to the atmospheric environment is relatively high, which can effectively deal with the shortcomings of the use of macroscopic data; third, the main component of solid particulate matter, such as PM2.5 and PM10, that is, industrial sulfur dioxide, as an indicator of environmental quality, is also verified and recognized by many scholars.

Regression method: GMM estimation

There is a certain degree of inertia in human health, i.e., prior health conditions may have some influence on current health. The static panel data model does not take into account the health status of the previous period, and by considering the possible endogeneity between variables, we used the dynamic panel data model to verify the variables’ impact on human health. Meanwhile, considering that the ordinary standard error estimation is not accurate, the generalized moment estimation (GMM) is an effective method to overcome the endogeneity problem. Considering the limitations of the short panel data and the possible endogeneity problem that can bias the model estimation, the regression was conducted using the generalized moment estimation method. The generalized moment estimation includes two methods: system GMM and differential GMM. The differential GMM estimation method may have the problem of weak instrumental variables, and the system GMM method can more fully utilize the variable information; considering this, we selected system GMM to empirically analyze the panel data. According to the selection of the variables and methods, the theoretical model is constructed as follows:

(5)

where denotes human health, denotes trade openness, denotes environmental pollution, denotes public health expenditure, denotes environmental regulation, denotes urbanization, denotes population size, is the intercept term, and denotes the random disturbance term.

Cross-sectional dependence tests

Cross-sectional dependence is a critical issue when examining the relationship between all selected variables in a panel data model, and ignoring it may lead to serious estimation bias and size distortion. Therefore, before analyzing the smoothness of the variables, the authors first tested for the presence of cross-sectional dependence in the panel and used the Breusch–Pagan LM test and the Pesaran CD test. The results of the tests showed that the p-values were greater than 0.01 (see Table 1), proving that the cross-sectional correlation of the variables was not significant and that there was no cross-sectional correlation in the data.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Cross-sectional dependence test result.

https://doi.org/10.1371/journal.pone.0312246.t001

Unit root test

If the variable data are in a nonstationary series, the direct model estimation will show pseudo-regression results and further empirical analysis will result in false conclusions. Therefore, we adopt the ADF statistic to conduct the unit root test on the inter-provincial panel data results, as per Table 2. Under the three types of assumptions of the intercept term and the trend term, the ADF test results of the original value series fail to reject the original hypothesis of the existence of a unit root and are nonstationary. The first-order differenced series all passed the ADF test under the first two types of conditions, indicating that the differenced series is smooth, and therefore, the variables are first-order single-integrated series and can be analyzed in regression.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Unit root test results.

https://doi.org/10.1371/journal.pone.0312246.t002

Co-integration test

In this paper, Pedroni and Kao’s tests were used to conduct co-integration tests on the data taken, as per the results in Table 3; the p-value of each test method is less than 0.05. The data can pass the co-integration test, and the panel data model can be established for further analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Co-integration test results.

https://doi.org/10.1371/journal.pone.0312246.t003

Multicollinearity test

The VIF value of each variable is calculated to determine whether there is a problem of covariance among the variables. If VIF is greater than 10, this means multicollinearity exists. The results in Table 4 show that there is no multicollinearity between the variables.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Multicollinearity test results.

https://doi.org/10.1371/journal.pone.0312246.t004

Model type discrimination

Panel data regression equations, including the constant coefficient model without individual effects, the variable intercept equation, the variable coefficient equation, and the intercept term of the variable intercept equation, can be categorized into fixed effects and random effects. Before the panel data regression analysis, it is necessary to determine the type of regression equation and the type of intercept term. An F-test was used for the values in Table 5; the p-value is less than 0.05, so the developed individual fixed effect model is appropriate. Further, in order to discriminate whether the model is an individual random effect model or an individual fixed effect model, the Hausman test is introduced, and the p-value is less than 0.05, as shown in Table 5. Therefore, the original hypothesis is rejected, i.e., the established individual fixed effect model is appropriate. In addition, the object of the study is the mortality rate; there is no phenomenon of random sampling on the whole, so the establishment of an individual fixed effect model is more appropriate.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Panel data model type discrimination test results.

https://doi.org/10.1371/journal.pone.0312246.t005

Analysis of regression results

In this paper, the GMM model was constructed using panel data with the help of Stata2017 software. The GMM model is divided into differential GMM and systematic GMM, in which the systematic GMM overcomes some missing errors in the differential GMM. The results can be seen in Table 6 below.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Regression results.

https://doi.org/10.1371/journal.pone.0312246.t006

In Table 6, the results of the Hansen over-identification test show that the p-value is greater than 0.05, indicating that the instrumental variables are selected effectively. The results of the AR root test show that the P-value of AR (2) is greater than 0.05, indicating that the model does not have more than second-order serial autocorrelation. The above results show that the current model is well constructed.

Robustness testing

The method of robustness testing, which involves a data substitution method, a variable substitution method, setting control or dummy variables, and so on, does not exist singularly. In this paper, according to the model and data situation, by transforming the explanatory variables, PM2.5 was selected to replace industrial SO2 re-estimation in order to avoid the problem of obtaining accidental results due to the use of specific pollution indicators. The results of the robustness test can be seen in Table 7. After the test, the direction and significance of the regression coefficients of the variables did not particularly change, and the size of the regression coefficient value did not change much. It further confirms the results of our empirical regression earlier.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Robustness test results.

https://doi.org/10.1371/journal.pone.0312246.t007

Conclusion

We used the inter-provincial panel macro data of 30 regions to study the impact of trade openness and energy consumption on human health and drew the following conclusions:

The negative impact of trade openness on human health is reflected in the fact that trade openness may affect the nutrition of the population, their dietary habits, food safety, and the spread of diseases, but it can also bring more healthy food and medical equipment, increase nutritional intake, improve the level of hygiene, and improve human health. At the national level, the impact of trade openness on human health is not significant. From the perspective of different regions, trade openness can improve the level of human health in the eastern region but is unfavorable to the level of human health in the central and western regions.

The consumption of unclean energy sources, such as coal and oil, releases pollutants such as carbon dioxide, carbon monoxide, sulfur dioxide, and dust particulate matter, causing serious air pollution problems. Serious air pollution induces respiratory diseases and cardiovascular diseases, bringing huge losses to human health. From the results of the empirical analysis, environmental pollution reduces human health in all regions. However, the negative effect is not significant in the eastern region, which is related to the high proportion of clean energy use in the eastern region, while the central and western regions mainly have the energy industry as their dominant industry and are overly dependent on the openness and use of the energy industry, thus causing serious negative impacts on the environment, which is not conducive to the improvement of human health.

Urbanization can reduce the national population mortality rate and improve human health, and from the perspective of different regions, the coefficient of the impact of urbanization on the population mortality rate in the eastern region is significantly negative. This is consistent with the estimation results at the national level. In the central region, the coefficient of urbanization’s impact on the population mortality rate is negative, but none of them passes the significance test; in the western region, the coefficient of urbanization’s impact on the population mortality rate is negative but has a lower significance level. Increases in human health expenditures reduce population mortality and have negative coefficients in all regions. Increasing population size is not conducive to improving urban human health at the national level in the west and central regions, but it can improve the health of the population in the east. Environmental regulation policies are effective in reducing the national population mortality rate and, thus, enhancing human health, with the largest effect in the eastern region and the smallest effect in the western region, mainly due to a strong correlation with regional economic growth patterns.

Recommendations

Based on the results of the empirical analysis above, trade openness and energy consumption both have a greater impact on human health, and there is regional heterogeneity. This paper makes recommendations to improve human health using the following methods:

1. (1). First, make full use of the technological spillover effects of trade openness to improve medical care and reduce medical costs. Second, trade openness brings environmental pollution at the same time, which offsets part of the health effects and increases healthcare expenditures. In the process of deepening trade openness, we should pay attention to pollution prevention and control and emphasize the co-ordination and matching of trade and environmental policies to reduce the environmental pollution caused by trade openness and improve the health effect that trade openness is supposed to have.
2. (2). The conversion rate of coal resources should be the primary goal, and specific management measures should be formulated in conjunction with clean coal technology. Meanwhile, in northern regions with high demand for coal in winter, coal resources should be fully utilized to transform traditional inefficient heating structures and promote the use of clean coal. In addition, for heavy industrial areas, fiscal and tax incentives should be used as a means to guide enterprises to actively innovate, research, and develop new energy technologies, as well as, at the same time, regulate the emission of pollution in order to take effective measures to reward and punish practice. In the process of industrial undertakings, for the promotion of economic development in the central and western regions at the same time, it is necessary to adhere to the principle of energy saving and environmental protection, abandon the serious pollution of the industry, and actively advocate for the benign development of enterprises and healthy development, taking on the task of environmental protection.
3. (3). Improve the level of environmental regulations in less developed regions and promote the optimization and upgrading of the foreign trade structure in the central and western regions. High-energy-consuming and high-polluting import and export enterprises have chosen to locate in the central and western regions in order to economize on environmental costs, exposing the residents of the central and western regions, who have lower per capita income levels, to higher concentrations of environmental pollution, which has led to environment-related health inequalities. The level of environmental regulation in the central and western regions should be raised, and policy incentives should be provided to import and export enterprises that make innovations in green production technology.
4. (4). For precise policymaking, as there are large differences in regional endowments and conditions, it is necessary to analyze the actual situation in different regions and take into account the impact of differentiation. When formulating policies, specific issues should be analyzed, and policies should be precisely applied to the actual situation in different regions, abandoning the “one-size-fits-all” phenomenon of health trade policies that existed in the past and promoting synergistic development. In order to improve the health of the population, the eastern region should provide better public services and reduce pollution; the central region should increase government expenditure on public services, vigorously promote urbanization, stabilize employment, improve the efficiency of the use of healthcare resources, and combat pollution; the western region should invest more resources in education and develop rural revitalization to reduce the gap between urban and rural areas.
5. (5). The importance of health expenditures should be emphasized in fiscal expenditures. Among the various financial expenditures, health expenditure and education expenditure are expenditure items that prove difficult in terms of bringing economic utility and financial revenue to the government in the short term; some local governments, in pursuit of the performance view, tend to pay more attention to the items that can rapidly develop the economy and increase financial revenue and do not pay enough attention to the field of health. Therefore, it is necessary for governments at all levels to further emphasize the importance of health expenditures in fiscal expenditures and to maintain stable growth of human health expenditures in tandem with economic development and fiscal expenditures.

Shortcomings and prospects

1. (1). As both environmental pollution and trade openness may have spatial spillover effects, which, in turn, have an impact on the public health of the neighboring regions, the aim of this paper was to develop spatial research, which may bias the actual results; this could provide a direction for future research in this area;
2. (2). For the selection of variables, due to the fact that there are a lot of factors that affect the level of health, this paper does not list all of the influencing factors, referring only to previous scholars’ research to select some key factors to carry out the study. In future research, we will explore further influencing factor analysis and strive to draw different conclusions;
3. (3). In this paper, we did not test the nonlinear and threshold effects, and we will further explore the nonlinear relationship and threshold value in future research so as to provide a clearer theoretical reference for the region to formulate relevant policies.