An inverted U-shaped relationship between haze pollution and rural migrants' long-term residence intentions: Theoretical mechanisms and China’s evidence

doi:10.21203/rs.3.rs-3128287/v1

Download PDF

Research Article

An inverted U-shaped relationship between haze pollution and rural migrants' long-term residence intentions: Theoretical mechanisms and China’s evidence

https://doi.org/10.21203/rs.3.rs-3128287/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 22 Dec, 2023

Read the published version in Environmental Science and Pollution Research →

You are reading this latest preprint version

Despite the extensive literature on the effect of air pollution on individual’s well-being, explore the relationship between polluted haze and rural migrants' long-term residence intentions is still limited. The current study fills the gap by investigating the impact of haze pollution on rural migrants' long-term residence intentions in Chinese context. Overall, we find an inverted U-shaped relationship between the level of haze pollutants and rural migrants' long-term residence intentions. The results still hold when different haze quality indicators and using thermal inversion as the endogeneity variable. Specifically, haze pollution plays its role through two opposite mechanisms: income effect and health effect. When the size of income effect is larger than health effect, rural migrants are inclined to settle down in their host cities, vice versa. In addition, we also find that the effect of haze pollution varies across different population groups and regions.

Haze Pollution

Rural Migrants

Long-term Residence Intention

China

Haze pollution is a common side product of economics progress in most developing countries, among which China is no exception. The booming growth in Chinese economy since the opening-up reform heavily depends on the introduction of resource-intensive industries, including manufacturing, mining, fuel processing, chemical and electricity, etc (Wu et al., 2019; Chen et al., 2020). These sectors perform in an environmentally unfriendly way and consume massive natural resources. Accordingly, China’s haze pollution has become a worldwide issue as the emission of harmful atmosphere goes beyond its own borders and the negative spillovers effects are undertaken by other regions as well (D'Souza et al., 2020). The polluting industries, in combination with the use of wood and coal in stoves from millions of households makes the haze pollution situation more complex (Aunan, 2018). Depending on emission intensity and meteorological conditions, haze pollution varies across different regions. The densely populated city cluster of the Beijing-Tianjin-Hebei region has witnessed the high concentrations of PM_2.5, while the Yangtze River Delta and the Pearl River Delta experience a less extent. Generally, the air quality is better in southern areas than that in the north (Zhang et al., 2022). Sever and persistent haze pollution has become an heavy burden not only to the sustainable development of economics, but also for national health care system. In order to mitigate the adverse effects of haze pollution, China has taken a series of measures. In 2005, the central government demanded that coal-fired plants should desulfurize before releasing gas waste. In 2011, another regulation on installation of selective catalytic reduction systems (SCR) was also implemented. The issue of haze pollution reduction was put on the top agenda in Chine’s 12th Five-Year Plan for the first time, targeting at 8% cut of sulfur dioxide, 10% cut of nitrogen dioxide, 16% cut of in carbon dioxide, and 16% reduction in energy consumption by the end of 2015 (Li et al., 2016).

A negative association has been found between exposure to polluted air and individual’s physical and mental well-being (Lepeule et al., 2012; Yin et al., 2020; Guo et al., 2021). Observations on asthma and chronic obstructive pulmonary disease (COPD) sharply increases, especially among old people (Brunekreef, 2002). Other diseases, cardiovascular diseases, diabetes, neurodevelopment, cognitive impairment included, are also found to be linked to exposure to particulate pollution (Franklin et al., 2015; Thiering & Heinrich, 2015). As Global Burden of Disease (GBD) project reported, poor air quality is responsible for 1.6 million deaths in China and 4.2 million deaths worldwide in 2015 alone (Forouzanfar et al., 2016). Haze pollution related deaths comprised 17 per cent of the total deaths in China in that year, among the highest rate in the world (Song et al., 2017). In addition to the physical damage caused by poor air quality, the mental illnesses resulting from ambient haze pollution also attracts a lot of attention (Dzhambov et al., 2018; Gu et al., 2020; Sass et al., 2017). For example, Ahumada & Iturra (2021) find that the particulate matter in the air significantly decreases people’s self-reported well-being in Chile, and citizens are willing to pay 7.46–11.98 USD per month to improve the air quality. Power et al. (2015) investigates the impact of haze pollution on residents’ mental health, and their results suggest that symptoms of anxiety significantly increases with higher exposure to fine particulate matter (PM_2.5). Resulting from the polluted air, the health cost rises, either on medical treatment, or health insurance (Chang et al., 2018).

Recently, researchers has focused on whether haze pollution can influence people’s migration behavior. For instance, Germani et al. (2021) explore whether population moves away from provinces with higher levels of haze pollution emissions towards those characterized by lower levels. Considering other contextual factors associated with migration streams, the increasing environment risks related to poor air quality is still influential in shaping internal migration choices among Italian provinces. Similar evidence has also been found in other regions and migration patterns (Hu et al., 2022). It is recognized that ‘vote by foot’ rule stands in the immigration decision. Considering individual’s characteristics, younger people and males have a higher probability of migrating than their counterparts, escaping from the environmental stressors caused by haze pollution. This phenomenon can result from the personal psychological mechanism of cost-benefit judgment. The workforce as a whole, thus, can suffer from a brain drain (Khuc wt al., 2022).

Although the adverse effect of polluted air on individual’s migration decision is well documented in the literature, its effect on attracting people may not be fully explored. Haze pollution is one of the main kinds of pollution among the world which is mostly function of economic growth. The relationship between economic growth and pollution is discussed as Environmental Kuznets Curve (EKC) hypothesis (Akbostancı et al., 2009; Dinda et al., 2000; Soytas & Sari, 2009). As this theory unveils, at the first stage of the inverted U-shaped relationship, the increase in industrial activity with energy-intensive production leads to an increase in pollutant emissions. In this case, the extent of haze pollution acts as a signal indicating the economic development and employment. Although polluted environment may hamper workers move around, economic factors still counts when migrants make decisions. In Chinese context, industry structure is undergoing transformation and the relationship between air pollutants and economic development is just at the appearance of the inflection point. Thus, before reaching the highest pollution level, cities with poorer air quality still have appealing for migrations.

The current study can contribute to the eco-economics literature in two-fold ways. First, we dig into the causal inference between haze pollution and migrants’ settlement intention. Previous research mostly uncover a negative relationship between these two indicators and emphasize the adverse effect of polluted air. However, the signal effect of haze pollution suggesting high economic development and more working opportunities may be ignored. This ‘income’ effect of haze pollution functions as a pulling force, attracting rural migrants move from low-polluted areas to heavier regions. Thus, the relationship between haze pollution and rural migrants’ long-term settlement intention presents an inverted u shape rather than declined linearity. Second, this study also shed some light on the heterogeneous effect of haze pollution on different population and regions.

The remainder of this paper is organized as follows. In section 2, we present the theoretical analysis of the current study. In section 3, we describe the data source, outcome and explanatory variables, and model specifications. In section 4, the statistics description, main results, heterogeneous analysis, robustness check and mechanism test are reported. Section 5 presents discussion and conclusions of this work.

We assume that rural migrant i currently lives in the city j, and city k is his original place. rural migrant’s long-term residence decision is based on personal utility. The following equation describes his decision process.

$${\text{Intention}\text{ }}_{ij}=\left\{\begin{array}{c}1, \forall k\ne j {U}_{ij}>{U}_{ik}\\ 0, \exists k\ne j {U}_{ij}⩽{U}_{ik}\end{array}\right.$$

Where ${U}_{ij}{, U}_{ik}$represents rural migrant i’s utility when he lives in city j and k, respectively. When${\text{ }\text{Intention}\text{ }}_{ij}=1$, rural migrant tends to long-term residence in the host city j while${\text{ }\text{Intention}\text{ }}_{ij}=0$ indicating no long-term settlement plan. ${U}_{ij}$ is determined by the income y rural migrantss can earn in city j, $U=U\left(y\right)$.

Exposure to haze pollution is linked to an increase severity of physical illnesses, especially in cardiovascular related and respiratory diseases. In addition, recent evidence has also been found between haze pollution and mental illnesses (Ahumada & Iturra, 2021). This indicate that rural migrants work in more polluted areas have to pay a higher cost on their health. In this regard, we introduce health cost ${ℎ}_{j}$ into our analysis. ${ℎ}_{j}$ is correlated with the level of haze pollution ${p}_{j}$, which varies from 0 to 1. The air is more polluted when ${p}_{j}$ approaches 1. The average wage level in city j is w. Given this wage level, rural migrant’s income can be calculated by the following equation:

$$y=w-{{p}_{j}ℎ}_{j}$$

$$y\sim N\left(w-{{p}_{j}ℎ}_{j},\frac{{\sigma }^{2}}{{p}_{j}}\right)$$

In Eq. (3), ${\sigma }^{2}$ is a constant term. Assuming rural migrant is risk aversion, one’s risk coefficient is as follows:

$$\alpha =-\frac{{\text{d}}^{2}U/\text{d}{y}^{2}}{\text{d}U/\text{d}y}$$

Combine Eq. (1) to (4), rural migrant’s utility in the host city is:

$$U\left(y\right)=-\frac{{C}_{1}}{\alpha }\text{e}\text{x}\text{p}(-\alpha y)+{c}_{2}$$

Additionally, Eq. (5) can be transformed in a short form:

$$U\left(y\right)=-C\text{e}\text{x}\text{p}(-\alpha y)$$

C is a constant term. When characteristics at the city level are controlled, conditional expectation function is:

$$E\left({U}_{ij}\mid {p}_{j},w,Z\right)=-C\text{e}\text{x}\text{p}\left[-\alpha \left(w-{{p}_{j}ℎ}_{j}\right)+\frac{{\alpha }^{2}{\sigma }^{2}}{2{p}_{j}}\right]$$

Holding w fixed, Eq. (7) can be expressed as:

$$E\left({U}_{ij}\mid {p}_{j},w,Z\right)=-C\text{e}\text{x}\text{p}\left(\alpha {{p}_{j}ℎ}_{j}+\frac{{\alpha }^{2}{\sigma }^{2}}{2{p}_{j}}\right)$$

Differentiating Eq. (8) with respect to ${p}_{j}$ shows that:

$$\frac{\partial E\left({U}_{ij}\mid {p}_{j},w,Z\right)}{\partial {p}_{j}}=C\text{e}\text{x}\text{p}\left(\alpha {{p}_{j}ℎ}_{j}+\frac{{\alpha }^{2}{\sigma }^{2}}{2{p}_{j}}\right)\left(\frac{{\alpha }^{2}{\sigma }^{2}}{2{p}_{j}^{2}}-\alpha {ℎ}_{j}\right)$$

As $C\text{e}\text{x}\text{p}\left(\alpha {{p}_{j}ℎ}_{j}+\frac{{\alpha }^{2}{\sigma }^{2}}{2{p}_{j}}\right)>0$, $\frac{\partial E\left({U}_{ij}\mid {p}_{j},w,Z\right)}{\partial {p}_{j}}$=0, the equation determining the optimal settlement decision becomes:

$${p}_{j}^{\ast }=\sigma \sqrt{\frac{\alpha }{2{ℎ}_{j}}}$$

$$\frac{\partial E\left({U}_{ij}\mid {p}_{j},w,Z\right)}{\partial {p}_{j}}=\left\{\begin{array}{c}⩾0, {p}_{j}⩽{p}_{j}^{\ast }\\ <0, {p}_{j}>{p}_{j}^{\ast }\end{array}\right.$$

Consequently, when the level of haze pollution exceeds ${p}_{j}^{\ast }$, heavier haze pollution will make rural migrants move to other place, vice versa.

In sum, our analysis suggests that for risk-averse rural migrants, haze pollution has two contrary effects on their long-term residence decision in the host city. On one hand, haze pollution serves as a signal indicating rapid economic development at the expense of environment. In order to get higher expected income, rural migrants tend to settle in the more polluted areas. This is the ‘income effect’ of haze pollution. On the other hand, exposure to haze pollution significantly increase health risks and lead to a higher cost on medical care or health insurance, which may hamper their long-term residence decision. This is the ‘health effect’ of haze pollution. As show in Fig. 1, at the first stage, rural migrants’ long-term residence intention increases as air condition getting worse. When it reaches at the maximum level, people tend to escape from the polluting air and move to places with better environment.

3.1. Data

Data used in this study is from China Migrants Dynamic Survey (CMDS) for 2017, the most recent round, released by the Chinese National Health Commission. The CMDS is a high-quality household survey conducted to investigate the profile of demographic characteristics, settlement intention, health condition and public service of migrants. Respondents concluded in the survey should be more than 15 years old and stay in the host city for over 1 month. The data are widely used in migration, public health and education research.

The sample design of the CMDS is structured to cover 31 provinces of the country, the entire territory of mainland China; it is representative in terms of the major migration areas. The survey sample in 2017 totals 169989 households, providing information both at the household and individual level. Considering the registration type, reasons for migration and stay duration in the destination place, 98898 samples are finally kept for empirical analysis.

3.2. Outcome variable

The indicator for rural migrants’ Long-term residence intention derived from the CMDS is “Do you plan to live in the host city for more than 5 years”. Among the respondents, 40696 rural migrants have a plan to long-term residence in the city they work in, accounting for 41.15% of the whole sample. The distribution of rural migrants’ long-term residence intention is visualized in Fig. 2a. It is demonstrated that workers in northern and eastern areas are more inclined to stay in the host cities than those in southern and western areas.

3.3. Explanatory variable

The ground-based concentration of PM_2.5 is the indicator for haze pollution. Compared with PM_10, PM_2.5 is smaller and easier to be breathed into lower respiratory organs, lungs, bronchi, etc. Carrying pathogenic and viruses, this air pollutant is more harmful to human health. Additionally, PM_2.5 is the major source of air pollutants in China, and directly leads to the smog weather. The above two reasons make PM_2.5 a reasonable indicator for measuring haze pollution level. Data on PM_2.5 concentration used in the current study is derived from Surface PM2.5, provided by Atmospheric Composition Analysis Group, Washington University in St. Louis (https://sites.wustl.edu/acag/datasets/surface-pm2-5/). The database estimates annual and monthly ground-level fine particulate matter (PM_2.5) for 1998–2020 by combing Aerosol Optical Depth (AOD) and global ground-based observations. It is at a spatial resolution of 0.01°×0.01°. We collect PM_2.5 concentration of 371 cities in China in 2016 from this database. Figure 2b shows the spatial distribution of haze pollution, the darker colored patch indicates worse air quality. The average annual PM_2.5 concentration of the observed 371 cities is 40.977µg/m³. Among these cities, the most ten polluted cities are Kunyu, Tumushuke, Hengshui, Shijiazhuang, Xingtai, Liaocheng, Alaer, Langfang, Dezhou and Jiaozuo, whose PM_2.5 concentration are averaged at 74µg/m³, almost the twice of national level. Most of these cities are located in the northern parts of China, and the Jjingjinji Area is the center area of haze pollution.

To rule out the interference of geographical factors, we also control variables at the city level, including house price, wage level, health care, nominal GDP, and city size. In order to get the data, we match China Statistical Yearbook for Regional Economy 2017 and China City Statistical Yearbook 2017 with CMDS.

Figure 3 shows the fitting plots of non-linear relationship between haze pollution and rural migrants’ long-term residence intention. An inverted U-shaped figure is clearly presented, indicating that rural migrants’ long-term residence intention increases when haze pollution becomes heavier. After the inflection point, rural migrants’ long-term residence intention sharply decreases with air quality getting worse.

3.3. Model specification

In the current study, the Probit model for modeling binary dependent variable is employed. The benchmark model setting is as follows:

$${y}_{i}^{\ast }={a}_{0}+{a}_{1}{x}_{i}+{a}_{2}{x}_{i}^{2}+{a}_{3}{s}_{i}+{\epsilon }_{i}$$

$${y}_{i}=\left\{\begin{array}{c} 1,{ y}_{i}^{\ast }>0\\ 0,{ y}_{i}^{\ast }\le 0\end{array}\right.$$

In model (12), ${y}_{i}^{\ast }$ is a latent variable denoting rural migrants’ long-term residence intention. When ${ y}_{i}^{\ast }>0$, ${y}_{i}=1$, otherwise ${y}_{i}=0$. ${x}_{i}$ represents the level of haze pollution. ${s}_{i}$ expresses a series of control variables, for example, age, gender, education, family size, GDP, etc. ${a}_{j}(j=0, 1, 2, 3)$ represents parameters to be estimated and ${\epsilon }_{i}$ is the error item.

An important threat to identifying the causal relationship between haze pollution and rural migrants’ long-term residence intention is the potential endogeneity due to omitted variables and possible reverse causality between dependent and outcome variables. With regard to omitted variables, rural migrants’ decision is a sophisticated process; our controlled variables may miss some unobserved factors that both influence haze pollution an rural migrants’ living decision. Concerning reverse causality, large number of workers surging into cities can lead to increasing demand for infrastructure, transportation, and other services, which also cause high energy consumption and heavier haze pollution. In this regard, rural migrant’s long-term residence decision is the cause rather than result of haze pollution in the city. Hence, an instrument variable is introduced into the benchmark model to mitigate the potential bias.

An instrument should satisfy two conditions, strong correlation with the endogenous variable and exclusive restricted with the outcome variable. In our case, the instrumental variable should highly be related to haze pollution and not affect rural migrant’s long-term residence intention directly. Following Arceo et al. (2016), Sarger (2019) and Chen (2022), we choose average days of thermal inversion as the instrument variable. For one thing, the occurrence of thermal inversion generally impedes vertical air convection, less pollutants can disperse into higher atmospheric levels. When thermal inversion lasts for a long time, more pollutants will be trapped near the ground, leading to worse air quality (Hicks et al., 2016; Trinh et al., 2019). Hence, thermal inversion meets the first rule of that it is strongly positive related to haze pollution. For the other, thermal inversion is a normally natural phenomenon caused by a series of meteorological factors and can hardly influence individual’s activities in a direct way. As thermal inversion can meet the two conditions, it is the suitable instrument variable.

We propose the IV-probit model as follows:

$${y}_{i}^{\ast }={b}_{0}+{b}_{1}{x}_{i}+{b}_{2}{x}_{i}^{2}+{b}_{3}{s}_{i}+{u}_{i}$$

$${x}_{i}={c}_{0}+{{c}_{1}z}_{i}+{c}_{2}{s}_{i}+{v}_{i}$$

Where ${y}_{i}$ is observable dummy variable, ${y}_{i}^{\ast }$ represents unobservable latent variable, ${x}_{i}$ is the only endogenous variable in the model, ${z}_{i}$ stands for the instrumental variable,${ u}_{i}$and ${v}_{i}$ are error terms. $\left({u}_{i}, {v}_{i}\right)$ is subject to the normal distribution:

$$\left(\begin{array}{c}{u}_{i}\\ {v}_{i}\end{array}\right)\sim N\left[\left(\begin{array}{c}0\\ 0\end{array}\right),\left(\begin{array}{cc}1& \rho {\sigma }_{v}\\ \rho {\sigma }_{v}& {\sigma }_{v}^{2}\end{array}\right)\right]$$

The variance of ${u}_{i}$ is standardized to 1, and $\rho$ is the correlation coefficient of $\left({u}_{i}, {v}_{i}\right)$. Additionally, we also employ the Extended Probit method to test the robustness of the results. One of the advantages of this method lies in the fact that it can deal with endogeneity problem.

In addition, to test the mechanism proposed in the theoretical analysis, we also introduce a mediating effect model to do the empirical analysis. This method was first brought up by Baron & Kenny in 1986. Income level and health perception are the two mediators used in our study, in order to explore the role haze pollution plays in migrants’ settlement intention. The model specification is based on the baseline:

$${m}_{i}={d}_{0}+{d}_{1}{x}_{i}+{d}_{4}{s}_{i}+{\epsilon }_{i}$$

$${y}_{i}^{\ast }={e}_{0}+{e}_{1}{x}_{i}+{e}_{2}{x}_{i}^{2}+{e}_{3}{m}_{i}+{e}_{4}{s}_{i}+{\epsilon }_{i}$$

${y}_{i}^{\ast }$ denotes latent variable, when ${y}_{i}^{\ast }>$0, ${y}_{i}=1$, otherwise, ${y}_{i}=0$. ${x}_{i}$ is the explanatory variable, ${m}_{i}$ is mediators, including income level and health perception. Indirect effect exists when ${d}_{1}$ and ${e}_{3}$ both show significance. Additionally, when ${e}_{1}$ and ${e}_{2}$ are insignificant, there shall be complete mediating effect. Otherwise, partial mediating effect exists. One of the limitations in the method proposed by Baron and Kenny lies with its treatment for continuous dependent variable. We follow Karlson et al. (2012) by applying the KHB method to test the robustness of mediating effect.

4.1. Descriptive statistics

Table 1 presents the summary of the main variables. The results show that the rural migrants are averaged at 36 years old, more than half of them are male (57.2%), and 85.9% are married. The statistics further depict that only a small amount of workers have a high-level education (31.3%), with senior high school or college degree. Each sampled family’s have 3 people on average. The percentage of rural migrants who plan to live in the host city for a long time (41.2%) is lower than those who don’t have a settlement plan. In this regard, many rural migrants choose to work in the smaller cities rather than metropolis, and the migration type is mostly inter-provincial. Concerning the motivation of settlement, only 30.8% rural migrants consider income level as the most important factors. This indicates that the cost of living in the city, including house price, health care support, can also influence rural migrants decisions.

Table 1

Descriptive statistics of the variables.
Variables	Definition	Mean	Std.	Min	Max
long-term residence intention intention	Plan to live in the host city for more than 5 years	0.412	0.492	0	1
Haze pollution	Average PM _2.5 (µg/m³)	40.977	14.534	2.276	85.725
Age	Years	36.555	9.732	15	93
Gender	Male = 1, Female = 0	0.572	0.495	0	1
Marriage	Married = 1, Unmarried = 0	0.859	0.348	0	1
Education	Uneducated, primary and junior high = 1, Senior high, college/university = 0	0.313	0.464	0	1
Family size	No. of family members	3.230	1.159	1	10
Migration type	Inter-provincial migration = 1, Intra-provincial migration, within city = 2, Intra-city migration, across county = 3	1.669	0.753	1	3
House price	Log of house price in the host city	8.932	0.647	7.568	10.725
Wage level	Log of average wage in the host city	11.158	0.237	10.513	11.718
Health care	No. of beds in hospitals per 1000 persons in the host city	4.556	3.792	0.156	17.741
GDP	Log of GDP per capita in the host city	11.191	0.490	9.384	12.281
City size	Less than 1 million people = 1, 1–5 million people = 2, 5–10 million people = 3, More than 10 million people = 4	1.924	0.929	1	4
Settlement motivation	High income level = 1, otherwise = 0	0.308	0.462	0	1
Health education	Haze pollution related education = 1, no such education experience = 0	0.220	0.414	0	1

4.2. Baseline regression

Table 2 reports the estimation results of the Probit model. In column (1), only core explanatory variables are added into the model. In column (2), individual and family characteristics are controlled for. In column (3), city characteristics are controlled for. Column (4) include all the variables in the model setting, and results of column (5) represents the marginal effects of column (4). The results indicate a significant association between haze pollution and rural migrants’ long-term residence intention. The coefficient of quadratic term implies that as the “income effect” of haze pollution at first is larger than “health effect”, attracting rural migrants to settle down in the working place. Then the rural migrants are more likely to relocate as the haze pollution gets worse.

As far as rural migrants’ characteristics are concerned, their age also exhibit a similar inverted U-shaped relationship with long-term residence intention. When they get older, they tend to escape from the polluted areas to cities with cleaner environment. More educated rural migrants also have a tendency to live in the city, where they may find better working opportunities. This effect has also been well documented by many researches (Chen & Liu, 2016), revealing rural migrants with better human capital are more inclined to settle down in cities. Compared with women, men are more likely to settle in the host city for a longer time. Those who already get getting married and having a lager family size tend to live in the host city for long. In China, workers usually migrate from rural to urban areas. In those more developed cities, they can get better public service. For example, their offspring can enjoy high-quality education.

The nominal GDP, city size and wage level positively influence rural migrants’ long-term residence intention. The coefficients of these three variables show expected signs. However, house price in the city is a disadvantage factor when rural migrants consider a long-term stay. This results align with previous studies. Zhang et al. (2015) also finds that housing price influences people’s settlement decision through the welfare dissipation effect and the expected wealth effect in China and the effects is more prominent in the eastern area.

Table 2

Baseline results: The overall effect of haze pollution on migrants’ settlement intention.
Variable	(1)	(2)	(3)	(4)	(5)
Haze pollution	0.0224***	0.0156***	0.0242***	0.0209***	0.0077***
	(0.0012)	(0.0012)	(0.0021)	(0.0021)	(0.0008)
Haze pollution square	-0.0002***	-0.0001***	-0.0002***	-0.0002***	-0.0001***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Age		0.0356***		0.0397***	0.0147***
		(0.0031)		(0.0034)	(0.0013)
Age²		-0.0379***		-0.0440***	-0.0163***
		(0.0039)		(0.0043)	(0.0016)
Gender		0.0178**		0.0178**	0.0066**
		(0.0084)		(0.0089)	(0.0033)
Marriage		0.3372***		0.3730***	0.1383***
		(0.0157)		(0.0171)	(0.0063)
Education		0.3925***		0.4064***	0.1507***
		(0.0094)		(0.0100)	(0.0036)
Family size		0.0504***		0.0571***	0.0212***
		(0.0042)		(0.0046)	(0.0017)
Migration type		0.1909***		0.2369***	0.0878***
		(0.0054)		(0.0065)	(0.0024)
House Price			-0.1322***	-0.0814***	-0.0302***
			(0.0136)	(0.0139)	(0.0052)
Wage level			0.1325***	0.3058***	0.1134***
			(0.0378)	(0.0390)	(0.0144)
Health care			-0.0117***	-0.0149***	-0.0055***
			(0.0022)	(0.0022)	(0.0008)
GDP			0.0512***	0.0557***	0.0206***
			(0.0132)	(0.0134)	(0.0050)
City Size			0.0978***	0.1122***	0.0416***
			(0.0111)	(0.0113)	(0.0042)
Constant	-0.7760***	-2.2917***	-1.8471***	-6.0850***
	(0.0258)	(0.0624)	(0.3297)	(0.3530)
Pseudo R²	0.0044	0.0386	0.0048	0.0452
Observations	98,898	98,898	87,729	87,729	87,729
Notes: Robust standard errors in parentheses. * p < 0.01, p < 0.05, * p < 0.1. Marginal effect are reported in column (5) of Table 2.

4.3. Endogeneity test

To mitigate the possible endogeneity problem, we apply an instrument variable approach to further identify the causal effect. Table 3 reports the regression results of IVprobit model, Eprobit model and 2SLS model, respectively. Results in column (1) show haze pollution, measured with PM_2.5, is significantly and positively correlated with the thermal inversion. This means that the instrument variable meets the first criterion of correlation. Second stage results in Column (2) shows that the Wald test of exogeneity is significant at the 1% level, indicating that haze pollution is the endogenous variable. In addition, the weak identification test critical values is significant at 1% level, indicating there is no weak instrument variable problem. Besides, to obtain the economic significance of coefficients, we also calculate the marginal effects of haze pollution and its quadratic term on rural migrants’ long-term residence intention. In column (2), the coefficients of haze pollution and its quadratic term are 0.0874 and − 0.009, respectively, both at the 1% level of significance. These results also confirm the inverted U-shape between haze pollution and rural migrants’ long-term residence intention, which has been found in the baseline model.

In addition, we also introduce a Eprobit to do the endogeneity test as the IVprobit model may reduce the effectiveness of the results. Results in column (3) show that instrument variable has a significant and positive effect on haze pollution, which confirms the correlation of these two variables. Correlation test of error term is negatively significant, suggesting that haze pollution is the endogenous variable. Coefficients in column (4) depict that haze pollution has a two-phase effect on rural migrants’ long-term residence intention, with a increase effect first and then a decrease effect. Compared with the coefficient in column (2), the effect size in Eprobit model is smaller. This means IVprobit model can lead to a loss in validity.

In sum, all the models show that the polluted air is a driving force in determining rural migrants’ long-term residence intention to settle down in the host cities they work.

Table 3

Regression results with IV probit, Eprobit and 2SLS model.
Variable	IVprobit		Eprobit		2SLS
	Pollution	Intention	Pollution	Intention	Intention
	(1)	(2)	(3)	(4)	(5)
Haze pollution		0.0874***		0.0237***	0.0323***
		(0.0105)		(0.0021)	(0.0040)
Haze pollution square		-0.0009***		-0.0002***	-0.0003***
		(0.0001)		(0.0000)	(0.0000)
Instrumental variable	0.1062***		0.0954***
	(0.0006)		(0.0006)
Wald test of exogeneity		42.4000***
Weak instrumental variable test		70.2600***
Corr(e.Pollution, e.Intention)				-0.0621***
				(0.0095)
Endogeneity test					62525.4700***
Constant	-23.9198***	-5.0347***	25.1692***	-6.2287***	-1.3230***
	(2.9593)	(0.3908)	(0.1101)	(0.3533)	(0.1441)
Control variables	YES	YES	YES	YES	YES
R²	0.3797				0.0475
Observations	87,729	87,729	87,729	87,729	87,729
Notes: Robust standard errors in parentheses. * p < 0.01, p < 0.05, * p < 0.1.

4.4. Mechanism test

To test the two paths embedded in the mechanism, two mediators are used. One mediator is income, and the other one is rural migrants’ health status. Table 4 shows the results of mediating regression. Column (1) and (2) tests the effect of haze pollution on rural migrants’ income level. Column (3) and (4) examines the effect of haze pollution on rural migrants’ health conditions. Coefficients in column (1) and (3) both present positive significance at 1% significant level, indicating that haze pollution can affect rural migrant’s long-term residence intention by influencing their income and health.

Table 4

Regression results with mediators.
Variable	Income	Intention	Health	Intention
	(1)	(2)	(3)	(4)
Haze pollution	0.0013***	0.0208***	0.0194***	0.0190***
	(0.0004)	(0.0021)	(0.0003)	(0.0019)
Haze pollution square		-0.0002***		-0.0001***
		(0.0000)		(0.0000)
Income		0.0334***
		(0.0096)
Health				0.0562***
				(0.0107)
Constant	-6.7815***	-6.0244***	-2.6131***	-3.1631***
	(0.3631)	(0.3535)	(0.3868)	(0.3576)
Control variables	YES	YES	YES	YES
Pseudo R²	0.0304	0.0453	0.0550	0.0449
Total effect	0.0209***		0.0194***
	(0.0021)		(0.0019)
Direct effect	0.0208***		0.0190***
	(0.0021)		(0.0019)
Indirect effect	0.0001***		0.0004***
	(0.0000)		(0.0001)
Observations	87,729	87,729	87,085	87,085
Notes: Robust standard errors in parentheses. * p < 0.01, p < 0.05, * p < 0.1.

4.5. Robustness check

To obtain robust results, we use five different ways to do the robustness check, the results are shown in Table 5. In the baseline models, we use the question “Do you plan to live in the host city for more than 5 years” as an indicator for outcome variable; as stated in Section 4.2, the values of the dependent variable are 1 (has the plan to live in the city for more than 5 years) and 0 (otherwise). In column (1), we relax the settlement duration by asking the question “Do you plan to continue live in the host city later on?” The dummy variable is categorized by 1 (have the plan to stay) and 0 (otherwise). In column (2), we use another question “Do you plan to settle down in the city in the future?” If the rural migrants answered “yes”, this indicator takes the value of 1, otherwise it equals 0. In column (1) and (2), the results of the Probit models without considering the endogenous variable are reported. The size and direction of the core independent variable, haze pollution, correlate well with the results obtained in Table 2.

Column (3) and (4) show the results of re-estimation with replacement of core explanatory. In the baseline model, the annual average PM_2.5 concentration in 2016 is used as the proxy for haze pollution. In the robustness check, we replace it with the annual average PM_2.5 concentration in 2017. Column (3) reports the results of the estimation with new indicator for haze pollution. In column (4), we use the data from Air Quality Platform supported by China National Environmental Monitoring Centre. Although the measurement is still PM_2.5 concentration of 371 cities national wide in 2016, the data type is ground-based observations rather satellite-based estimate. The regressions are estimated with a Probit model. Results in column (3) and (4) show that polluted days have a inverted U-shape effect on rural migrants’ long-term residence intention. Although the effect size is smaller than that in the baseline model, the coefficients are significant at 1% level.

Table 5

Results of robustness check.
Variable	(1)	(2)	(3)	(4)	(5)
Haze pollution	0.0248***	0.0227***	0.0183***	0.0167***	0.0194***
	(0.0023)	(0.0023)	(0.0025)	(0.0019)	(0.0019)
Haze pollution square	-0.0003***	-0.0002***	-0.0002***	-0.0001***	-0.0001***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Constant	-3.2054***	-5.6341***	-6.3595***	-6.1016***	-3.1680***
	(0.4069)	(0.3769)	(0.3523)	(0.3540)	(0.3575)
Control variables	YES	YES	YES	YES	YES
Pseudo R²	0.0265	0.0485	0.0447	0.0447	0.0446
Observations	87,729	87,729	87,729	87,729	87,085
Notes: Robust standard errors in parentheses. * p < 0.01, p < 0.05, * p < 0.1.

4.6. Heterogeneity analysis

In the previous sections, we estimate the overall effect of haze pollution on rural migrants’ long-term residence intention. However, as a complex decision process, this effect may vary across different groups. Some people may be more sensitive to haze pollution while others care more about economical return when they make settlement decision. This section explores heterogeneous effect of haze pollution among different groups.

First, heterogeneity analysis with individual characteristics:

Figure 4a plots the difference between different generations. Migrations are classified into two groups: the old generation and the young generation. The old generation is defined as those who were born before 1980, and the others belong to the young-generation group. From Fig. 4a, it can be seen that the young generation has a higher tolerance of haze pollution. Only when the severity of haze pollution reaches around 60 µg/m³, their long-term residence intention begins to decline. In comparison, the maximum value of haze pollution old generation can bear is much lower. This may be because those elder workers have worse health conditions resulting from long time heavy work, and exposure to the ambient haze pollution even worsens their conditions. Similar evidence has also been found by other researchers that haze pollution puts more risks on the elderly (Fischer et al., 2003; Goldberg et al., 2013).

Concerning gender domain, Fig. 4b shows that compared to women, men are more sensitive to haze pollution. However, the difference between diversified gender groups is small. Notably, our result contradicts with previous researches. Although there has been no consensus on the haze pollution response between different genders (Clougherty, 2010), many scholars hold the view that women usually response more actively to haze pollution (Kim et al., 2019). As smoking is a more common phenomenon among men, the effect of haze pollution is more likely to be obscured by smoking (Sunyer et al., 2000).

For different family structure, Fig. 4c obviously shows that there is a remarkable difference disparity between married and single rural migrants, both in haze pollution tolerance and long-term residence intention. Single rural migrants can easily move to another city and find a new job as many of them are employed by the informal sector (Williams, 2015). Without the burden to support the family, they can easily make switch in their work and living space. By contrast, the couple rural migrants take more factors into consideration when they make the migration decision, working opportunities, public service provision, living cost, development potential, etc. Once they settle down in a place, it costs more for them to move to a new city. Currently, many economically developed areas experience heavier haze pollution at the same time. In searching for a better income return, married migrants can bear a higher level of haze pollution and are more inclined to living in the host city for long (Cooke, 2008).

In terms of education condition, less educated people are more sensitive to haze pollution as shown in Fig. 4d. Surprisingly, it may contradict with common thoughts that people are more vulnerable to haze pollution as their related knowledge increases (Branis and Linhartova, 2012; Ou et al., 2008). In Chinese context, more educated rural migrants are more likely to work indoors. The possibility for them to be exposed to haze pollution is much lower than those with less education experience, who often work in the open-air environment. Additionally, as more educated rural migrants usually work in the formal sector and their public service related to work is better, the opportunity cost for them to find a new job in another city is much higher. Thus, compared with rural migrants with junior school or less education, people who have a higher school or even higher degree prefer to settle in their working cities. Our findings correspond well with the research conducted by Kan et al. (2008), who also find a negative relationship between education level and sensitiveness to haze pollution.

Second, heterogeneity analysis with regional difference:

In addition to individual characteristics, the heterogeneous effects of haze pollution among different regions are also worthy of attention. For instance, the host city size and air quality differences between hometown and the current city. Figure 4e illustrates the comparison of haze pollution between rural migrants’ hometown and the city they migrate into. Notably, if the air quality in the destination city is better than that of their hometown, then rural migrants’ long-term residence intention increases with the level of haze pollution. Put it in another word, the inverted U-shape relationship is replaced by a increasing positive correlation. Figure 4f compares rural migrants’ long-term residence intention with different city size. It is not surprising that the smallest city size (less than 1 million residents) witnesses the lowest long-term residence intention and haze pollution tolerance. Among the four city size, rural migrants who work in the city with 5 to 10 million dwellers have the highest long-term residence intention and care least about haze pollution. Cities at this size have more working opportunities compared with smaller ones while having much more cleaner air compared to metropolis.

This study explores how exposure to haze pollution are associated with rural migrants’ long-term residence intention in China, one of the world’s most air-polluted countries. By combing individual-level data from a nationally representative survey and satellite-based high-resolution data on PM_2.5 concentration, we estimated the level of haze pollution in cities and its correlation with rural migrants’ long-term residence intention in the host city. Our quantitative analysis partly supports the claim that haze pollution resulting from PM_2.5 emissions exert a significant negative impact on the long-term residence intentions of working adults. However, contrary to the finding of the existing studies (Germani et al., 2021; Lai et al., 2022; Khuc et al., 2022), we find an inverted U-shaped relationship, rather than a linear one between those two indicators. According to our theoretical analysis, there are mainly two effects exist, namely income effect and health effect. Income effect refers to the signal role haze pollution play in indicating the level of economic development. In this case, higher level of haze pollution of a certain city means employment opportunities. On the contrary, health effect refers to the adverse influence polluted air may have on personal health conditions. At the first stage, rural migrants’ long-term residence intention increases as the haze pollution becomes worse. This happens because the effect of ‘income effect’ is larger than ‘health effect’. When the extent of haze pollution reaches an extreme level, rural migrants tend to leave the cities with serious air condition. At this phase, ‘health effect’ outweighs ‘income effect’.

Further, we discovered a correlation among marriage status, education background and rural migrants’ long-term residence intention. In particular, single people are much more sensitive to haze pollution and have a lower willingness to live in the host city permanently. This result is in line with many researchers who also find marriage status contribute to rural migrants’ intention of living in the working cities (Kim et al., 2020; Liu et al., 2018). Compared to the single workers, couples take more factors into consideration when making the long-term residence decision. Thus, they have higher tolerance with haze pollution when they can find more working opportunities and get higher working return. Interestingly, our results reveal that more educated people can bear higher level of haze pollution compared with their less educated counterparts. This may contradict with the prevailing literature (Cakmak et al., 2006) as it is commonly believed that educated people have more knowledge about the adverse effect of haze pollution. In our case, more educated people usually can find positions in a formal sector and work indoors, thus preventing the exposure to polluted air. This may account for the positive association between education level and tolerance for haze pollution.

Our results also add more evidence to the literature on the variation in the relationship between pollutant emissions across cities with different population densities. Increased PM_2.5 concentration has greater impact on the working adults in small and large cities, while rural migrants in medium-sized cities are less affected. In the small cities, there is lower urban population and industrial density, thus people are more sensitive to harmful pollutants. In the large cities, although haze pollution becomes worse, higher income level can compensate for rural migrants’ loss on their health expenditure.

Haze pollution is a fundamental factor influencing people’s physical and mental health. To avoid the potential negative effect of polluted air, people may take various preventive actions to mitigate its effects. For rural migrations, move to another host city is their optional choice. The current study aims to evaluate the impact of haze pollution on rural migrants long-term residence intention in Chinese context. By using the migration data sourced from CMDS 2017 and the haze pollution information provided by Atmospheric Composition Analysis Group, this paper comprehensively examined the impact of haze pollution on the long-term residence decisions of rural-to-urban migrants. This study is enriching both health economics and environmental economics research literature. Our findings confirm that there is an inverted U-shaped relationship between haze pollution and rural migrants long-term residence intention living in the host city. Put differently; cities with higher have more attractiveness for people who want to earn more money, as haze pollution is also an indicator suggesting a higher economic development level. As the air quality becomes worse and migrants’ tolerance for poor air reaches the maximum level, they attempt to move to a cleaner place. It is worth noting that haze pollution affect different population in a heterogenous way. The elder, male, less educated migrants are more sensitive to haze pollution than the other groups.

Moreover, this study is not without limitations. The data used in this study is from CMDS 2017, which is not a longitudinal one. In this regard, a fixed-effect model is not applicable to control the individual characteristics that do not change with time. The condition of haze pollution may vary over time as synoptic situation and environmental regulations on polluted air changes across year. Then, it is difficult to comprehensively explore the effects of haze pollution on rural migrants’ long-term residence intention in a dynamic way. The data and methods used in the current study is appropriate to our ends and can reach our conclusions, but improved data source and corresponding methodology may produce other different results. Thus, future researchers can extend the current work if panel data is available.

The current study proposes recommendations for the government and policy makers based on the empirical findings. First, haze pollution may not hold back labor force in the short time. However, it indeed decreases the attractiveness of cities and may hamper the process of urbanization. As central and local governments still target at housing migrations in the cities, more regulations should be implemented to improve the air quality. Second, as our empirical findings show, effects of haze pollution on different population seems to be heterogenous, different measures should be taken. For example, older people are more sensitive to poor air quality, more medical insurance for haze pollution can be inclined to the elderly. In contrast, less educated workers have a higher tendency to change the working and living place, as they probably work outdoors and have more time exposed to polluted air. Extra working subsidy can be added in their wages in order to alleviate the serious repercussions of haze pollution.

CRediT authorship contribution statement

Qingjun Zhao: Conceptualization, Methodology, Software, Visualization, Writing – original draft, Funding acquisition. Yue Wang: Conceptualization, Methodology, Writing – original draft, Writing – review & editing.

Ethical approval

This manuscript was prepared following ethical standards.

Consent to participate

The authors have voluntarily agreed to participate in this research study.

Consent to publish

The authors agree to publish the article in Environmental Science and Pollution Research.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors do not have permission to share data.

Acknowledgments

This work was supported by the Humanities and Social Sciences Research Planning Fund of the Ministry of Education of China (Grant No. 20YJA790020), Tsinghua Rural Studies PhD Scholarship (Grant No. 202108).

Ahumada G, Iturra V (2021) If the air was cleaner, would we be happier? An economic assessment of the effects of air pollution on individual subjective well-being in Chile. J Clean Prod 289:125152
Aunan K, Hansen MH, Wang S (2018) Introduction: air pollution in China. China Q 234:279–298
Akbostancı E, Türüt-Aşık S, Tunç G (2009) The relationship between income and environment in Turkey: is there an environmental. Kuznets curve? Energy policy 37(3):861–867
Baron RM, Kenny DA (1986) The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Personal Soc Psychol 51(6):1173
Branis M, Linhartova M (2012) Association between unemployment, income, education level, population size and air pollution in Czech cities: evidence for environmental inequality? A pilot national scale analysis. Health Place 18(5):1110–1114
Brunekreef B, Holgate ST (2002) Air pollution and health. The Lancet 360(9341):1233–1242
Cakmak S, Dales RE, Judek S (2006) Do gender, education, and income modify the effect of air pollution gases on cardiac disease? J Occup Environ Med, 89–94
Chang TY, Huang W, Wang Y (2018) Something in the air: Pollution and the demand for health insurance. Rev Econ Stud 85(3):1609–1634
Chen S, Liu Z (2016) What determines the settlement intention of rural migrants in China? Economic incentives versus sociocultural conditions. Habitat Int 58:42–50
Chen L, Fan Y, Guo W (2020) Relationship of economic development, family income and health status in China: the moderating role of environmental pollution perception. J Health Psychol 25(13–14):2499–2510
Clougherty JE (2010) A growing role for gender analysis in air pollution epidemiology. Environ Health Perspect 118(2):167–176
Cooke TJ (2008) Migration in a family way. Popul space place 14(4):255–265
Dinda S, Coondoo D, Pal M (2000) Air quality and economic growth: an empirical study. Ecol Econ 34(3):409–423
D'Souza C, McCormack S, Taghian M, Chu MT, Mort GS, Ahmed T (2020) An empirical examination of sustainability for multinational firms in China: Implications for cleaner production. J Clean Prod 242:118446
Dzhambov AM, Markevych I, Tilov B, Arabadzhiev Z, Stoyanov D, Gatseva P, Dimitrova DD (2018) Pathways linking residential noise and air pollution to mental ill-health in young adults. Environ Res 166:458–465
Fischer P, Hoek G, Brunekreef B, Verhoeff A, Van Wijnen J (2003) Air pollution and mortality in The Netherlands: are the elderly more at risk? Eur Respir J 21(40 suppl):34s–38s
Forouzanfar MH, Afshin A, Alexander LT, Anderson HR, Bhutta ZA, Biryukov S, Carrero JJ (2016) Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet 388(10053):1659–1724
Franklin BA, Brook R, Pope III, C. A (2015) Air pollution and cardiovascular disease. Curr Probl Cardiol 40(5):207–238
Germani AR, Scaramozzino P, Castaldo A, Talamo G (2021) Does air pollution influence internal migration? An empirical investigation on Italian provinces. Environ Sci Policy 120:11–20
Goldberg MS, Burnett RT, Stieb DM, Brophy JM, Daskalopoulou SS, Valois MF, Brook JR (2013) Associations between ambient air pollution and daily mortality among elderly persons in Montreal, Quebec. Sci Total Environ 463:931–942
Gu H, Yan W, Elahi E, Cao Y (2020) Air pollution risks human mental health: an implication of two-stages least squares estimation of interaction effects. Environ Sci Pollut Res 27(2):2036–2043
Guo W, Chen L, Fan Y, Liu M, Jiang F (2021) Effect of ambient air quality on subjective well-being among Chinese working adults. J Clean Prod 296:126509
Hicks D, Marsh P, Oliva P (2016) Air Pollution and Procyclical Mortality: Causal Evidence from Thermal Inversions. NBER working paper
Hu Z, Zhang Y, Cao J, Zhou K (2022) Longing for the Blue Sky: Urban air quality and the individual decision to immigrate. J Asian Econ 79:101437
Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N et al (2008) Season, sex, age, and education as modifiers of the effects of outdoor air pollution on daily mortality in shanghai, china: the public health and air pollution in asia (papa) study. Environ Health Perspect 116(9):1183–1188
Karlson KB, Holm A, Breen R (2012) Comparing regression coefficients between same-sample nested models using logit and probit: A new method. Sociol Methodol 42(1):286–313
Kim H, Noh J, Noh Y, Oh SS, Koh SB, Kim C (2019) Gender difference in the effects of outdoor air pollution on cognitive function among elderly in Korea. Front Public Health 7:375
Kim H, Cho J, Isehunwa O, Noh J, Noh Y, Oh SS, Kim C (2020) Marriage as a social tie in the relation of depressive symptoms attributable to air pollution exposure among the elderly. J Affect Disord 272:125–131
Khuc QV, Nguyen MH, Le TT, Nguyen TL, Nguyen T, Lich HK, Vuong QH (2022) Brain Drain out of the Blue: Pollution-Induced Migration in Vietnam. Int J Environ Res Public Health 19(6):3645
Lai W, Song H, Wang C, Wang H (2021) Air pollution and brain drain: Evidence from college graduates in China. China Econ Rev 68:101624
Li S, Williams G, Guo Y (2016) Health benefits from improved outdoor air quality and intervention in China. Environ Pollut 214:17–25
Liu W, Xu Z, Yang T (2018) Health effects of air pollution in China. Int J Environ Res Public Health 15(7):1471
Lepeule J, Laden F, Dockery D, Schwartz J (2012) Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environ Health Perspect 120(7):965–970
Ou CQ, Hedley AJ, Chung RY, Thach TQ, Chau YK, Chan KP, Lam TH (2008) Socioeconomic disparities in air pollution-associated mortality. Environ Res 107(2):237–244
Power MC, Kioumourtzoglou MA, Hart JE, Okereke OI, Laden F, Weisskopf MG (2015) The relation between past exposure to fine particulate air pollution and prevalent anxiety: observational cohort study. BMJ, 350
Trinh TT, Trinh TT, Le TT, Tu BM (2019) Temperature inversion and air pollution relationship, and its effects on human health in Hanoi City. Vietnam Environ Geochem Health 41(2):929–937
Thiering E, Heinrich J (2015) Epidemiology of air pollution and diabetes. Trends in Endocrinology & Metabolism 26(7):384–394
Sass V, Kravitz-Wirtz N, Karceski SM, Hajat A, Crowder K, Takeuchi D (2017) The effects of air pollution on individual psychological distress. Health Place 48:72–79
Song C, He J, Wu L, Jin T, Chen X, Li R, …, Mao H (2017) Health burden attributable to ambient PM2. 5 in China. Environ Pollut 223:575–586
Soytas U, Sari R (2009) Energy consumption, economic growth, and carbon emissions: challenges faced by an EU candidate member. Ecol Econ 68(6):1667–1675
Sunyer J, Schwartz J, Tobias A, Macfarlane D, Garcia J, Antó JM (2000) Patients with chronic obstructive pulmonary disease are at increased risk of death associated with urban particle air pollution: a case-crossover analysis. Am J Epidemiol 151(1):50–56
Williams CC (2015) Tackling informal employment in developing and transition economies: a critical evaluation of the neo–liberal approach. Int J Bus Glob 14(3):251–270
Wu J, Wei YD, Chen W, Yuan F (2019) Environmental regulations and redistribution of polluting industries in transitional China: Understanding regional and industrial differences. J Clean Prod 206:142–155
Yin P, Brauer M, Cohen AJ, Wang H, Li J, Burnett RT, Murray CJ (2020) The effect of air pollution on deaths, disease burden, and life expectancy across China and its provinces, 1990–2017: an analysis for the Global Burden of Disease Study 2017. Lancet Planet Health 4(9):e386–e398
Zang B, Lv P, Warren CM (2015) Housing prices, rural–urban migrants’ settlement decisions and their regional differences in China. Habitat Int 50:149–159
Zhang C, Du M, Liao L, Li W (2022) The effect of air pollution on migrants’ permanent settlement intention: Evidence from China. J Clean Prod 373:133832

Download PDF

Journal Publication

published 22 Dec, 2023

Read the published version in Environmental Science and Pollution Research →

Editorial decision: Major Revision
03 Sep, 2023
Reviewers agreed at journal
09 Aug, 2023
Reviewers invited by journal
09 Aug, 2023
Editor invited by journal
03 Aug, 2023
Editor assigned by journal
11 Jul, 2023
First submitted to journal
05 Jul, 2023

You are reading this latest preprint version

An inverted U-shaped relationship between haze pollution and rural migrants' long-term residence intentions: Theoretical mechanisms and China’s evidence

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Theoretical analysis

3. Materials and methods

3.1. Data

3.2. Outcome variable

3.3. Explanatory variable

3.3. Model specification

4. Results

4.1. Descriptive statistics

4.2. Baseline regression

4.3. Endogeneity test

4.4. Mechanism test

4.5. Robustness check

4.6. Heterogeneity analysis

5. Discussion

6. Conclusion

Declarations

References

Status:

Journal Publication

Version 1