The Impact of Climate Change on the Urban–Rural Income Gap in China

Abstract

Based on the annual average climate data and economic and social data from 262 prefecture-level cities in China from 2001 to 2019, this paper explores the impact of climate change on urban–rural income inequality and its mechanisms using fixed-effects (FEs) and mediated-effects (MEs) models. This study finds that (1) climate change has an inverted U-shaped relationship with the urban–rural income disparity; (2) climate change can affect the urban–rural income disparity by influencing urban and rural income levels, the regional degree of urbanization, and the labor force employment structure; (3) the impact of climate change on the urban–rural income gap is heterogeneous in East, Center, and West China; and (4) extreme heat can widen the urban–rural income gap, and extreme drought can narrow the urban–rural income gap. Climate change has a significant impact on the urban–rural income gap, and there is a need to continue to promote urbanization and the optimization of the employment structure of the workforce, reduce the vulnerability of rural residents to climate change, and narrow the urban–rural income gap.

1. Introduction

Using 1850–1900 as a baseline, the change in the global average surface temperature to date (2011–2020) is 1.09 °C. The average annual land area precipitation has increased in the temperate regions of the Northern Hemisphere, while precipitation has decreased in arid subtropical regions [1]. Global climate change brought about by industrial civilization is affecting and will continue to affect natural ecosystems and socioeconomic systems in the long term. The increasing probability of events such as sea level rise, frequent climate extremes, damage to human health, humanitarian crises, and increased poverty has once again called on all sectors to pay more attention to climate change [2,3,4]. Although a series of measures have been taken by international organizations and governments since 1992 to address and mitigate the potential impacts of climate change, it is clear that it will take time for these agreements to be implemented and even to take effect while climate change continues to advance [5]. Over the past 60 years, the rate of warming in China (0.26 °C/decade) has been higher than the global average for the same period (0.15 °C/decade), the average annual precipitation has been on an increasing trend (5.5 mm/decade), and extreme weather and climate events have tended to be more frequent and stronger [6]. Between 2001 and 2019, China’s average annual temperature and rainfall showed increasing trends, and we should expect that they will continue to increase in the future (Figure 1). China is developing rapidly via industrialization and modernization; does the climate change induced by these activities have an impact on economic and social development? How can we effectively respond to the positive or negative impacts of climate change? These issues have become the focus of attention of the academic community and have been studied accordingly.

The urban–rural income gap is an economic phenomenon that exists in every country in the world and is an inevitable part of the industrialization process [7]. At an early stage of industrialization, the urban–rural income gap tends to widen due to the uncoordinated or inequitable income distribution caused by the urban–rural dual-structure system, differences in industrial development, factor allocation differences, and policy bias [8,9,10]. As the industrialization process continues to accelerate, the social problems caused by the urban–rural income gap become increasingly prominent, and people pay increasing attention to the importance of narrowing the urban–rural gap to promote social equity; thus, they try to promote social equity through policy bias, the optimization of industrial structure, and the innovation of industrial patterns. They have tried to improve farmers’ incomes through policies, the optimization of industrial structure, the innovation of industrial pattern, financial support, increases in investments in education, etc., to narrow the income gap [11,12,13]. In China, however, the unique urban–rural dual structure has led to a large difference in the income levels of urban and rural residents [14] with the average annual income of urban residents still far higher than the average annual income of rural residents (Figure 2). Narrowing the income gap between urban and rural areas has become a major issue that China needs to address urgently.

There is no doubt that climate change has a more significant and direct impact on income levels. When a Ricardian approach was used to measure the effect of climate on land prices, it was found that higher temperatures in all seasons except fall would reduce average farm values, while more precipitation outside of fall would increase farm values [15]. An analysis of cross-country data and subnational data for 12 countries in the Americas revealed a significant negative correlation between income and temperature, both within and across countries, with a 1 °C increase in temperature decreasing the per capita income by 1.2–1.9% [16]. Due to the differences in the farmers’ perceptions of climate change, the farmers’ net rice incomes will be very sensitive to marginal changes in temperature and rainfall [17]. However, most of the above studies took the perspective of overall income and the income of farmers; there are relatively few studies on the income gap between urban and rural areas, and only some scholars believe that climate change and climate policies will affect the income gap between urban and rural areas. Climate change will reduce food availability and employment in the agricultural sector in Egypt’s major food-producing regions, thereby affecting the health and well-being of the poorest communities and increasing the inequality between rural and urban areas [18]. While there is no uniform conclusion on the impact of climate policies on the income gap between urban and rural residents, some scholars argue that carbon pricing and carbon tax policies can reduce income inequality through income-recovery mechanisms; however, others argue that the negative impact of these policies on rural households is much greater than the impact on urban households and may thus increase the urban–rural income gap [19,20,21,22].

Related studies on the impacts of climate change also have some relevance to the discussion in this paper. First, climate change will affect economic growth. Scholars have incorporated climate-change-related indicators into macroeconomic models, such as the environmental Kutznets curve hypothesis, the optimal economic growth model, and the endogenous growth model, to explore the impact of climate change on economic activities, arguing that climate change will have an impact on the rate of economic growth, especially in the short term, and that high temperatures have huge negative effects on economic growth in poor countries [23,24,25,26,27]. Second, climate change will affect agricultural production. Scholars argue that agricultural systems are more sensitive to sudden changes in precipitation and temperature and that, in countries with high proportions of households whose incomes are heavily dependent on agriculture, climate change affects food security and agriculture by affecting food quantity and quality, water availability and quality, pests, disease, and pollination, making it more difficult to eradicate famine and hunger [28,29,30]. Third, climate change affects industrial output. Scholars studying the impact of temperature change on industrial production based on country-level industry output data argue that higher temperatures affect output in the industrial sector and that there is an inverted U-shaped relationship between temperature and industrial output. In terms of its impact mechanism, climate change may indirectly affect industrial output by affecting labor productivity, working hours, total factor productivity, and fixed-asset investment [31,32,33]. Fourth, climate change affects population mobility. Climate change can change people’s behavioral preferences to move to other areas with relatively better climatic conditions, leading to climate migration, and this is a key response to environmental and nonenvironmental changes and stresses. For example, crop failures caused by rainfall can result in large-scale rural migration [34,35].

It is precisely because existing studies have paid less attention to the impact of climate change on the urban–rural income gap that we can do more in this regard. At the same time, on the basis of existing research, we can also study the mechanisms through which climate change affects the urban–rural income gap. Therefore, this study (a) analyzes whether climate change affects the urban–rural income gap in China by using data on the economic and social development of China’s prefecture-level cities and meteorological data; (b) validates the mediating roles of urban–rural income level, the degree of urbanization, and the employment structure of the labor force in the relationship between climate change and the urban–rural income gap; (c) estimates the heterogeneity of the impact of climate change on the urban–rural income gap; and (d) analyzes the impact of extreme weather events on the urban–rural income gap.

This paper differs from the previous literature in several ways. First, it provides research on the impact of climate change on income disparity. As mentioned above, the existing literature focuses mainly on the impact of climate change on income levels with little attention paid to the impact of climate change on regional income inequality. This paper adds to the limited literature on the impact of climate change on income disparities between urban and rural areas in China. Second, this paper provides the pathways through which climate change affects the income gap between urban and rural areas in China. Climate change affects the income gap by directly affecting the income level and through other channels, and studies on the indirect path of impact are often neglected. Finally, previous studies mostly focused on the cross-country level or sub-administrative regional level. However, China has a variety of climate types, and there are more obvious climate differences between different prefecture-level cities at the same time, which should be emphasized. For this reason, this paper examines China’s tertiary administrative regions, which can more accurately reduce the errors caused by climate differences and can provide the government with more differentiated solutions for different prefecture-level cities.

The remainder of this paper is organized as follows: Section 2 describes the possible impact mechanisms of climate change and presents the premise hypotheses; Section 3 describes the data used as well as the econometric model settings and variable descriptions; Section 4 reports the estimation results and conducts the analysis; Section 5 provides an analysis of extreme weather events; Section 6 presents our discussion; and Section 7 provides a conclusion.

2. Theoretical Framework and Hypotheses

2.1. Direct Impact

Climate change will affect the urban–rural income gap by affecting the incomes of urban and rural residents. On the one hand, climate change affects the income of rural residents by affecting crop production, thus affecting the urban–rural income gap. The natural environmental conditions required for crop growth are relatively harsh, and crops cannot be produced properly at extremely low or high temperatures and with minimal or great rainfall. In drought-prone areas, declines in cereal, horticultural crop, and livestock production and employment losses will result in lower farmer incomes [36]. On the other hand, climate change affects the income of urban residents by affecting labor productivity and labor hours, thus affecting the urban–rural income gap. The impact of climate change on labor productivity is not the same among different labor groups; the high levels of exposure of low-skilled laborers to climate change and rising temperatures will reduce the supply of low-skilled labor, hence narrowing the wage gap between high-skilled labor and low-skilled labor [37]. Since climate change will have an impact on the incomes of both urban and rural residents, and there may be differences in the effects of this impact, it will lead to different evolutionary trends in the urban–rural income gap.

Hypothesis 1 (H1):

Climate change may directly affect the urban–rural income gap by affecting urban–rural income levels.

2.2. Indirect Impact

Climate change will affect the urban–rural income gap by influencing the degree of urbanization. It has been found that, when cities are able to absorb more surplus rural labor, an increase in climate harshness will induce rural–urban migration, thus increasing the share of the urban population in the total population, i.e., the degree of urbanization [38,39]. Meanwhile, there is an inverted U-shaped relationship between urbanization and the urban–rural income gap, i.e., urbanization pulls up the urban–rural income gap in the early stage of economic development but can reduce the urban–rural income gap in the later stage, thus improving the income distribution [40,41]. Therefore, when the climate is poor, rural residents will migrate to cities to promote urbanization, and the urban–rural income gap will widen before urbanization reaches a certain stage; when urbanization reaches a later stage, the urban–rural income gap will narrow. When the climate is more suitable, urban and rural populations may choose to return to rural areas, the urbanization process will develop slowly or even decrease, and the urban–rural income gap will show a widening trend.

Climate change will affect the urban–rural income gap by influencing the structure of the employment of the labor force. Global climate change will reduce labor productivity by increasing outdoor and indoor heat loads and potentially harming the health and productivity of millions of working people [42]. In turn, higher temperatures will drive a shift in the labor hours of non-agricultural sectors for both men and women [43]. A significant decrease in labor productivity occurs at either relatively low or high temperatures, and the labor force chooses to shift from industries that are highly exposed to climate change (agriculture) to industries with relatively favorable climatic conditions (non-agricultural industries), i.e., the severity of the climate affects labor mobility between agricultural labor and non-agricultural labor. In a general sense, the transfer of labor from primary to secondary and tertiary industries is conducive to promoting the improvement of agricultural labor productivity, and the transfer of rural labor to secondary and tertiary industries for employment is conducive to increasing the income levels of the workers, i.e., the transfer of rural labor to industries and services will lead to a reduction in the urban–rural income gap.

Hypothesis 2a (H2a):

Climate change may indirectly affect the urban–rural income gap by affecting the degree of urbanization.

Hypothesis 2b (H2b):

Climate change may indirectly affect the urban–rural income gap by influencing the structure of the employment of the labor force.

3. Methods

3.1. Data

This paper aims to explore the impact of climate change on the urban–rural income gap; therefore, it is necessary to obtain relevant annual meteorological data for the study area. First, we obtained daily site data from the daily value dataset V3.0 of China’s surface climate data. Second, after retaining the corresponding latitude and longitude information, the daily meteorological data were processed using an annual average to obtain the annual average meteorological data. Once again, the inverse distance weighting method was used to interpolate the annual data. Finally, the data were partitioned and statistically analyzed according to administrative divisions and concatenated to obtain the annual meteorological data for each city. The data were then interpolated using the inverse distance weight method for each year, and the data were then divided by administrative division and stitched together to obtain year-by-year meteorological data by municipality. The urban and rural income data and economic and social development data of the prefecture-level cities were obtained from the China Urban Statistical Yearbook (2001–2020), the EPS Global Statistics/Analysis Platform, the provincial Statistical Yearbooks (2001–2020), and the Statistical Bulletin of National Economic and Social Development (2001–2020) of each prefecture-level city. According to the purpose of this study, prefecture-level cities were selected as the study area in this paper, and after removing invalid data and missing samples, the sample data of 262 prefecture-level cities from 2001 to 2019 were finally compiled, including the annual average meteorological data and economic and social development data.

3.2. Models

3.2.1. Benchmark Model

This paper selected Chinese prefecture-level cities from 2001 to 2019 as the research sample to explore the impact of climate change on the urban–rural income gap. Considering the differences in the magnitudes of the data, the raw data were logarithmically processed to eliminate the magnitudes. Meanwhile, factors of climate change, such as differences in temperature and rainfall, have significant effects on farmers’ income levels, and there is a nonlinear relationship between temperature and rainfall and income levels due to the nonlinear relationship between climate and crop yield [44,45]. We can assume that there will also be a nonlinear effect of climate change on the urban–rural income gap, so the squared term of the meteorological change data was also included in the model. Thus, the baseline model was constructed as follows:

$${\mathrm{lnGgp}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }\mathrm{lncliamte}}_{\mathrm{it}}+{\mathsf{\alpha }\mathrm{lnclimate}}_{\mathrm{it}}^2+{{\displaystyle \sum }}^{ }{\mathrm{lnZ}}_{\mathrm{it}}\mathsf{\delta }+{\mathsf{\gamma }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(1)

In Equation (1), i denotes the prefecture-level city and t denotes the year. ${\mathrm{lnGap}}_{\mathrm{it}}$ denotes the logarithm of the urban–rural income gap in year t of city i. In this article, the main proxy variables for this indicator are the urban–rural income ratio and the Theil index. ${\mathrm{lnclimate}}_{\mathrm{it}}$ denotes the logarithm of the climate data in year t of city i and the estimation coefficient, including the average annual temperature and average annual rainfall. $\mathsf{\beta }$ denotes the impact of climate change on the urban–rural income gap. The regression coefficient $\mathsf{\alpha }$ denotes the square term of the logarithm of the temperature climate data. When $\mathsf{\alpha }$ > 0, it indicates a positive U-shaped relationship between temperature (or rainfall) and the urban–rural income gap; when $\mathsf{\alpha }$ < 0, it indicates an inverted U-shaped relationship between temperature (or rainfall) and the urban–rural income gap. When $\mathsf{\alpha }$ is 0, there is a linear relationship between temperature (or rainfall) and the urban–rural income gap. ${\mathrm{Z}}_{\mathrm{it}}$ represents the logarithmic value of the control variables. ${\mathsf{\gamma }}_{\mathrm{t}}$ denotes a fixed-time effect, ${\mathsf{\mu }}_{\mathrm{i}}$ denotes an individual-fixed effect, and ${\mathsf{\epsilon }}_{\mathrm{it}}$ denotes the error term.

3.2.2. Mediating-Effects Model

In Equation (2), M represents the intermediary variable. To verify Hypotheses H1, H2a, and H2b, this article selects the urban–rural income level, urbanization rate, and labor employment structure as potential intermediary mechanisms. Among them, the urban–rural income level includes the urban and rural income levels, respectively, and the labor employment structure is calculated using the proportion of non-agricultural employment to employment. The other variables are consistent with the previous text.

$${\mathrm{lnM}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }\mathrm{lnclimate}}_{\mathrm{it}}+{\mathsf{\alpha }\mathrm{lnclimate}}_{\mathrm{it}}^2+{{\displaystyle \sum }}^{ }{\mathrm{lnZ}}_{\mathrm{it}}\mathsf{\delta }+{\mathsf{\gamma }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(2)

In this paper, Equations (1) and (2) are estimated using fixed-effects and mediated-effects models with the operations implemented in Stata 16.0 software.

3.3. Variables

In the benchmark model, this study referred to the indicators used in previous studies and selected the ratio of the per capita disposable income of urban residents to the per capita disposable income of rural residents in prefecture-level cities as the dependent variable [46,47]. On one hand, as the explanatory variable, the ratio can reflect the urban–rural income gap well and demonstrate the corresponding characteristics; on the other hand, since the data obtained are the absolute values of the year, taking the ratio can effectively eliminate the influence of price changes. Meanwhile, in order to verify the robustness of the benchmark model, we referred to previous studies and selected the Terrell index as a proxy variable for the urban–rural income gap to verify the model again. The Theil index is an indicator for calculating the degree of inequality between individuals or regions obtained according to the concept of entropy in information theory, and it can be used to decompose the income gap into the income gap within population groups and the income gap between population groups. In addition, the Theil index is more sensitive to changes in the high-end and low-end income groups [48,49]. The formula for calculating the Theil index is as follows:

$${\mathrm{Theil}}_{\mathrm{it}}={{\displaystyle \sum }}_{\mathrm{j}=1}^2\left[\frac{{\mathrm{p}}_{\mathrm{ij},\mathrm{t}}}{{\mathrm{p}}_{\mathrm{i},\mathrm{t}}}\right]\ln \left[\frac{{\mathrm{p}}_{\mathrm{ij},\mathrm{t}}}{{\mathrm{p}}_{\mathrm{i},\mathrm{t}}}\backslash \frac{{\mathrm{z}}_{\mathrm{ij},\mathrm{t}}}{{\mathrm{z}}_{\mathrm{i},\mathrm{t}}}\right]$$

(3)

In Equation (3), $\mathrm{j}=1,2$ represents urban and rural areas, respectively; ${\mathrm{z}}_{\mathrm{ij},\mathrm{t}}$ represents the total population of a region i; and ${\mathrm{p}}_{\mathrm{ij},\mathrm{t}}$ represents the total income of a region i.

The core explanatory variable of interest in this paper is the climate change in the region. Climate change refers to any change in climate over time, whether the cause is due to natural variability or the result of human activities, and differences in the statistics of climate elements such as temperature and rainfall are usually used to reflect these changes. Therefore, the regional average annual temperature and average annual rainfall data were selected as the core explanatory variables in this paper.

In this paper, the following variables were selected as control variables to join the research model based on data availability. On one hand, this paper controls the important regional economic characteristic variables that affect the urban–rural income gap between regions, including the per capita gross regional product, local public finance balance ratio, and the per capita total retail sales of consumer goods. On the other hand, this paper controls for important regional social characteristic variables that affect regional differences, including population density, the number of students in higher education per 10,000 people, and the number of beds in medical institutions per 10,000 people.

The variables involved in the further discussion of this paper include the disposable income per urban resident, the disposable income per rural resident, the rate of urbanization, and the percentage of people employed in non-agricultural industries. Descriptive statistics for the variables are displayed in

Table 1. Descriptive statistics of variables.

Variable	Description	Mean	Std.Dev.
Dependent variable
Ratio	Urban–rural income ratio	2.572	0.651
Theil	Theil index	0.0000814	0.4489857
Key variable
Tem	Average annual temperature	14.50	5.013
Rain	Average annual rainfall	1016	534.7
Control variable
GDP_people	GDP per capita	35,950	95,569
DP	Population density	428.9	349.9
Finance	Ratio of local financial revenues to expenditures	0.491	0.227
Student	Number of students enrolled in higher education colleges per 10,000 people	145.1	191.9
Hospital	Number of beds in medical institutions per 10,000 people	36.06	14.79
Consuming	Total retail sales of social consumer goods per capita	12,328	11,619
Mechanism variable
IncUrb	Logarithm of per capita disposable income of urban residents	19,136	11,059
IncRur	Logarithm of per capita disposable income of rural residents	8172	5684
Ur	Urbanization rate	0.422	0.183
In	Percentage of employees in non-agricultural industries	96.85	6.752

and Figure 3 and Figure 4.

4. Results

4.1. Results of the Benchmark Analysis

Table 2. The results of the benchmark model.

	ln_Ratio
	(1)	(2)	(3)
ln_Tem	0.484 ***	0.597 ***	0.563 ***
	(0.04)	(0.07)	(0.06)
ln_Rain	0.676 ***	1.112 ***	1.039 ***
	(0.07)	(0.07)	(0.07)
ln_Tem2	−0.090 ***	−0.092 ***	−0.084 ***
	(0.01)	(0.02)	(0.01)
ln_Rain2	−0.056 ***	−0.087 ***	−0.082 ***
	(0.01)	(0.01)	(0.01)
ln_GDP_people		−0.019 *	−0.076 ***
		(0.01)	(0.01)
ln_DP		−0.093 ***	−0.092 ***
		(0.00)	(0.00)
ln_Finance		−0.051 ***	0.014
		(0.01)	(0.01)
ln_Student		0.059 ***	0.053 ***
		(0.00)	(0.00)
ln_Hospital		−0.067 ***	−0.037 ***
		(0.01)	(0.01)
ln_Consuming		−0.090 ***	−0.098 ***
		(0.01)	(0.01)
Constant	−1.727 ***	−2.013 ***	−1.258 ***
	(0.24)	(0.24)	(0.26)
Year-fixed effect			Yes
Individual-fixed effect		Yes	Yes
N	5016	4911	4911

Note: Robust standard errors are shown in parentheses. * p < 0.1, and *** p < 0.01.

reports the results of the baseline regression of climate change on the urban–rural income gap at the prefecture level. The results in columns (1) through (3) all indicate that there is an inverted U-shaped relationship between temperature and rainfall and the urban–rural income ratio, i.e., the urban–rural income ratios of the prefectural cities will first increase and then decrease as the temperature and rainfall increase. According to the calculation results in column (3), the inflection point values for temperature and rainfall are 28.5 °C and 564.2 mm, respectively (see Appendix A for a detailed calculation). This shows that, when the average temperature of the year is higher than 28.5 °C, the temperature increases and the urban–rural income gap decreases; when the average temperature of the year is lower than 28.5 °C, the temperature decreases and the urban–rural income gap decreases. When the average rainfall of the year is higher than 564.2 mm, the rainfall increases and the urban–rural income gap decreases; when the average rainfall of the year is lower than 564.2 mm, the rainfall decreases and the urban–rural income gap decreases. Compared with 2022, we can see that China’s average annual temperature (10.5 °C) is still far from the inflection point value, but the average annual rainfall (606.1 mm) has already exceeded the inflection point value [6]. With the trend of climate evolution shown in Figure 2, the average annual temperature and average annual rainfall will continue to increase, and the increase in temperature will lead to a widening of the urban–rural income gap, while the increase in rainfall will lead to a narrowing of the urban–rural income gap.

4.2. Robustness Test

This study used the Theil index to replace the urban–rural income ratio as a proxy variable for the urban–rural income gap.

Table 3. The results of the robustness tests.

	Theil
	(1)	(2)	(3)
ln_Tem	0.104 ***	0.134 ***	0.129 ***
	(0.01)	(0.02)	(0.01)
ln_Rain	0.198 ***	0.342 ***	0.318 ***
	(0.02)	(0.02)	(0.02)
ln_Tem2	−0.023 ***	−0.021 ***	−0.020 ***
	(0.00)	(0.00)	(0.00)
ln_Rain2	−0.015 ***	−0.026 ***	−0.024 ***
	(0.00)	(0.00)	(0.00)
ln_GDP_people		−0.003	−0.017 ***
		(0.00)	(0.00)
ln_DP		−0.026 ***	−0.026 ***
		(0.00)	(0.00)
ln_Finance		−0.016 ***	0.001
		(0.00)	(0.00)
ln_Student		0.011 ***	0.010 ***
		(0.00)	(0.00)
ln_Hospital		−0.013 ***	−0.007 **
		(0.00)	(0.00)
ln_Consuming		−0.022 ***	−0.025 ***
		(0.00)	(0.00)
Constant	−0.657 ***	−0.847 ***	−0.634 ***
	(0.06)	(0.06)	(0.06)
Year-fixed effect			Yes
Individual-fixed effect		Yes	Yes
Observations	4521	4451	4451

Note: Robust standard errors are shown in parentheses. ** p < 0.05, and *** p < 0.01.

reports the results of the benchmark regression based on the Theil index. The results in columns (1) through (3) show that the effects of temperature and rainfall on the Theil index are significant at the 1% level of significance, and the signs of the primary and secondary terms are the same as in Table 2 with only the degree of effect varying. This regression result indicates the reliability of the previous estimation results, i.e., the baseline regression results are robust and again prove the existence of an inverted U-shaped relationship between temperature and rainfall and the urban–rural income gap.

4.3. Mechanism Test

Table 4. The results of the mechanism tests.

	(1)	(2)	(3)	(4)
	ln_IncRur	ln_IncUrb	ln_Ur	ln_In
ln_Tem	−0.474 ***	0.089 ***	−1.129 ***	0.496 ***
	(0.05)	(0.03)	(0.19)	(0.04)
ln_Rain	−1.049 ***	−0.010	−1.337 ***	−0.034
	(0.08)	(0.06)	(0.14)	(0.02)
ln_Tem2	0.077 ***	−0.007	0.221 ***	−0.095 ***
	(0.01)	(0.01)	(0.04)	(0.01)
ln_Rain2	0.086 ***	0.004	0.099 ***	0.002
	(0.01)	(0.00)	(0.01)	(0.00)
ln_GDP_people	0.278 ***	0.202 ***	0.150 ***	−0.000
	(0.03)	(0.02)	(0.03)	(0.00)
ln_DP	0.081 ***	−0.011 **	0.075 ***	0.011 ***
	(0.01)	(0.00)	(0.01)	(0.00)
ln_Finance	0.063 ***	0.077 ***	0.060 ***	−0.002
	(0.01)	(0.01)	(0.02)	(0.00)
ln_Student	−0.046 ***	0.007 **	−0.005	0.010 ***
	(0.00)	(0.00)	(0.01)	(0.00)
ln_Hospital	−0.054 ***	−0.091 ***	0.473 ***	−0.026 ***
	(0.02)	(0.01)	(0.02)	(0.01)
ln_Consuming	0.173 ***	0.075 **	0.099 ***	0.010 **
	(0.04)	(0.03)	(0.03)	(0.00)
Constant	7.702 ***	6.444 ***	0.569	3.962 ***
	(0.34)	(0.24)	(0.49)	(0.09)
Year-fixed effect	Yes	Yes	Yes	Yes
Individual-fixed effect	Yes	Yes	Yes	Yes
Observations	4911	4911	4452	4908

Note: Robust standard errors are shown in parentheses. ** p < 0.05, and *** p < 0.01.

reports the results of the regression of climate change on the mediating mechanism variables. The results in column (1) show that temperature and rainfall have a positive U-shaped relationship with the per capita disposable income of rural residents, decreasing and then increasing with increasing temperature and rainfall. Based on the results in column (1), the inflection point values for temperature and rainfall were calculated to be 21.7 °C and 445.3 mm, respectively, which may be due to the fact that, when the temperature and rainfall are far from the inflection point values, crops are outside of the optimal growth temperature and rainfall ranges and yields decrease. The results in column (2) show that temperature has a positive linear effect on urban income, and the disposable income per capita of urban residents increases with an increase in temperature. This may be due to the fact that, when the temperature gradually increases, the urban labor population, labor efficiency, and labor time decrease, leading to an increase in the unit labor income that, in turn, leads to an increase in the income of urban residents.

Column (3) shows that there is a positive U-shaped relationship between temperature and rainfall and urbanization, i.e., urbanization decreases and then increases with increasing temperature or rainfall. When the urbanization rate is added to Equation (1), there is still an inverted U-shaped relationship between temperature and rainfall and the urban–rural income ratio with inflection point values of 27.7 °C and 579.5 mm, respectively, and these are similar to the results of the benchmark regression. Meanwhile, the coefficient of the urbanization rate is negative, indicating that the urban–rural income gap decreases with an increase in the urbanization rate. Therefore, when temperature and rainfall increase, the urban–rural income gap first widens as the urbanization rate decreases and then narrows as the urbanization rate increases.

Column (4) shows that temperature has an inverted U-shaped relationship with the proportion of employees in non-agricultural industries with the proportion of employees in non-agricultural industries increasing and then decreasing with an increasing temperature. The effect of rainfall on the proportion of employees in non-agricultural industries is not significant. When adding the proportion of employees in non-agricultural industries into Equation (1), there is still an inverted U-shaped relationship between temperature and the urban–rural income ratio. Meanwhile, the positive coefficient of the proportion of employees in non-agricultural industries indicates that the urban–rural income gap widens with an increase in the proportion of employees employed in non-agricultural industries. Therefore, when the temperature rises, the urban–rural income gap widens with the increase in the proportion of persons employed in non-agricultural industries and then narrows with the decrease in the proportion of persons employed in non-agricultural industries.

4.4. Heterogeneity Test

There are significant differences in economic and social development among the eastern, central, and western regions of China. The eastern region is located on the coast and has taken the lead in development with its external advantages and has achieved significant results in urbanization and modernization development; the central and western regions are lagging in economic and social development due to geographical location constraints, are relatively backward in infrastructure construction and other aspects, and have different abilities to adapt to climate change. Therefore, the impact of climate change on the urban–rural income gap in the eastern, central, and western regions may differ to some extent.

Table 5. The results of the heterogeneity tests.

	ln_Ratio
	Eastern Region	Central Region	Western Region
ln_Tem	0.880 ***	0.514 ***	−0.076
	(0.19)	(0.09)	(0.20)
ln_Rain	−1.983 ***	−0.604	1.311 ***
	(0.24)	(0.37)	(0.11)
ln_Tem2	−0.135 ***	−0.088 ***	0.008
	(0.04)	(0.02)	(0.04)
ln_Rain2	0.134 ***	0.038	−0.096 ***
	(0.02)	(0.03)	(0.01)
ln_GDP_people	−0.088 ***	−0.089 **	−0.141 ***
	(0.01)	(0.04)	(0.03)
ln_DP	−0.089 ***	−0.047 ***	−0.078 ***
	(0.01)	(0.01)	(0.01)
ln_Finance	0.068 ***	0.052 ***	0.113 ***
	(0.02)	(0.02)	(0.02)
ln_Student	0.042 ***	0.029 ***	0.025 ***
	(0.00)	(0.01)	(0.01)
ln_Hospital	−0.068 ***	−0.112 ***	−0.007
	(0.02)	(0.02)	(0.03)
ln_Consuming	0.027	−0.033	−0.076 ***
	(0.02)	(0.03)	(0.02)
Constant	7.943 ***	4.130 ***	−0.985 **
	(0.85)	(1.23)	(0.42)
Year-fixed effect	Yes	Yes	Yes
Individual-fixed effect	Yes	Yes	Yes
Observations	1808	1828	1275

Note: Robust standard errors are shown in parentheses. ** p < 0.05, and *** p < 0.01.

reports the regression results for subgroups based on regional divisions. The results in column (1) show that, in the eastern region, temperature has an inverted U-shaped relationship with the rural–urban income gap; the rural–urban income ratio increases and then decreases with temperature with an inflection point value of 26.0 °C. Rainfall has a positive U-shaped relationship with the rural–urban income gap; the rural–urban income ratio decreases and then increases with an increase in rainfall with an inflection point value of 1634.4 mm. Column (2) shows that, in the central region, temperature has an inverted U-shaped relationship with the rural–urban income gap; the temperature increases and then decreases with an inflection point value of 18.5 °C. Rainfall does not have a significant effect on the rural–urban income gap. In the central region, temperature has an inverted U-shaped relationship with the urban–rural income gap, which first increases and then decreases with the temperature with an inflection point value of 18.5 °C. The effect of rainfall on the urban–rural income gap is not significant. Column (3) shows that the effect of temperature on the urban–rural income gap is not significant in the western region. It shows that rainfall has an inverted U-shaped relationship with the urban–rural income gap, and the urban–rural income gap first decreases and then increases with an increase in rainfall, and the inflection point value is 923.4 mm.

5. Extreme Climate and the Urban–Rural Income Gap

Climate change is a relatively slow process, but extreme weather, as a statistically rare weather event, may also have an impact on the urban–rural income gap. Extreme weather is both risky and an impairment, and by affecting the labor supply of households, it has, in turn, a significant impact on household income and productive investment [50,51]. Therefore, in further research, we will consider the impacts of extreme heat, extreme cold, extreme rainfall, and extreme drought events on the urban–rural income gap.

5.1. Method

We referenced previous studies and used percentage-based threshold metrics to measure extreme weather events. For example, extreme weather was identified as instances in which the daily climate was greater than the 95th percentile (or less than the 5th percentile) of the climate reference period. Extreme high temperatures (or extreme low temperatures) may be expressed as the sum of the number of days in a year in which the daily temperature is greater than (or less than) the 95th (or 5th) percentile of the sequence of daily temperatures over a standard time period, and extreme precipitation (or extreme drought) may be expressed as the sum of the number of days in a year in which the daily precipitation is greater than (or less than) the 95th (or 5th) percentile of the sequence of daily precipitation over a standard time period [52,53]. Our specification is as follows:

$${\mathrm{lnGap}}_{\mathrm{it}}={\mathsf{\omega }}_0+{\mathsf{\omega }\mathrm{EC}}_{\mathrm{it}}+{{\displaystyle \sum }}^{ }{\mathrm{lnZ}}_{\mathrm{it}}\mathsf{\phi }+{\mathsf{\mu }}_{\mathrm{it}}$$

(4)

In Equation (4), Gap represents the urban–rural income gap, which is replaced with the urban–rural income ratio; and EC represents the sum of the number of days during the year during which different climate extremes occur, including extreme heat, extreme cold, extreme rainfall, and extreme drought. z represents the control scalar, which is the same as in (1).

5.2. Results

Table 6. Results: the impacts of extreme weather events on the urban–rural income gap.

	ln_Ratio
	(1)	(2)	(3)	(4)	(5)	(6)
high_Tem_num	−0.006 **		−0.006 **
	(0.00)		(0.00)
low_Tem_num		0.000	0.000
		(0.00)	(0.00)
high_Rain_num				0.001		0.001
				(0.00)		(0.00)
low_Rain_num					0.001 ***	0.001 ***
					(0.00)	(0.00)
ln_GDP_people	−0.057 ***	−0.057 ***	−0.057 ***	−0.057 ***	−0.055 ***	−0.055 ***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
ln_DP	−0.009	−0.010	−0.009	−0.010	−0.010	−0.010
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
ln_Finance	0.018 **	0.018 **	0.018 **	0.018 **	0.019 **	0.019 **
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
ln_Student	0.004	0.004	0.004	0.004	0.003	0.003
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
ln_Hospital	−0.099 ***	−0.099 ***	−0.100 ***	−0.099 ***	−0.100 ***	−0.100 ***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
ln_Consuming	−0.012	−0.012	−0.012	−0.012	−0.013 *	−0.013 *
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Constant	1.826 ***	1.711 ***	1.821 ***	1.705 ***	1.659 ***	1.643 ***
	−0.12	−0.12	−0.13	−0.12	−0.11	−0.12
Year-fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Individual-fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Observations	4911	4911	4911	4911	4911	4911

Note: Robust standard errors are shown in parentheses. * p < 0.1, ** p < 0.05, and *** p < 0.01.

reports the impacts of extreme weather events on the urban–rural income gap. In columns (1) through (3), we find that the number of extreme heat days has a significant negative effect on the urban–rural income gap, and when more extreme heat events occur, the urban–rural income gap decreases. This may be due to the fact that extreme heat pushes rural to urban migration and agricultural employment to non-agricultural migration, which are the same as the previous findings. Extreme low temperatures have no effect on the urban–rural income gap. In columns (4) through (6), extreme rainfall has no effect on the rural–urban income gap. Extreme drought has a positive effect on the urban–rural income gap, and when more extreme drought events occur, the urban–rural income gap widens. This may be due to the fact that extreme drought reduces the income of the agricultural population but does not affect the income of the urban population.

6. Discussion

This study investigates the impact of climate change on the urban–rural income gap in prefecture-level cities in China. The data were obtained from the China Daily Values of Surface Climate Data Dataset V3.0 and the China Urban Statistical Yearbook (2001–2020). Fixed-effects and mediated-effects models were used to test the relationships and mechanisms between climate change and the urban–rural income gap.

At the prefecture-level city level in China, climate change has an impact on the urban–rural income gap, and this impact shows an inverted U-shaped relationship with the calculated inflection point values of the climate data being 28.5 °C and 564.2 mm when the urban–rural income ratio is used as a proxy variable for the urban–rural income gap. Therefore, a better response to climate change may be an important means of mitigating income inequality, but it is highly correlated with individual vulnerability to climate change, especially for rural individuals [54,55].

Second, it is worth noting that climate change can affect the urban–rural income gap through both direct and indirect channels. In the direct channel, climate change can affect the urban–rural income gap through the income levels of residents in different regions, especially the income levels of rural residents, which is consistent with previous studies [56,57]. In the indirect channel, climate change will affect the urban–rural income gap by influencing the urbanization process and the employment structure of the labor force. A severe climate environment will accelerate the urbanization process and promote population mobility [39,58]. This is an important guideline for adjusting the urbanization process and optimizing the employment structure of the workforce according to the extent of climate change.

The third important finding suggests that there is significant heterogeneity in the impact of climate change on the urban–rural income gap in East, Central, and West China, and understanding regional climate heterogeneity is important for setting national or regional climate policy goals and designing scenarios [59]. This is the reason for separating our discussion by region.

Finally, we measured the impact of extreme weather on the urban–rural income gap using the number of days in a year that extreme weather occurs. We found that extreme heat reduces the urban–rural income gap. The urban poor who are exposed to high temperatures do not actually increase their purchases of consumer goods, such as air conditioners, and are more vulnerable to high temperatures [60]. Extreme droughts widen the urban–rural income gap, but increased levels of natural biodiversity may mitigate this effect. The conservation of natural diversity would, therefore, be an important way to reduce the urban–rural income gap [61].

Our study has important policy implications. First, by clarifying the impacts and mechanisms of climate change on the urban–rural income gap, it is conducive to better integrating climate change into economic and social development in all regions and better realizing sustainable economic development. Second, the mechanism of the impact of climate change provides us with the possibility to analyze it further. By accelerating the urbanization process and optimizing the employment structure of the workforce, the vulnerability of rural groups to climate change will be effectively reduced. Finally, China’s measures to narrow the income gap between urban and rural areas will effectively help its people better cope with the century-old problem of global warming.

7. Conclusions

Climate change is affecting and will continue to affect human production and life. There is sufficient evidence from our research to suggest that climate change will have an impact on the urban–rural income gap in China. This paper provides some proof that there is an inverted U-shaped relationship between climate change and the urban–rural income gap. Climate change can affect the urban–rural income gap both directly (by affecting the income level of residents) and indirectly (by affecting the degree of urbanization and the employment structure of the labor force). The impact of climate change on the urban–rural income gap in China differs in the eastern, central, and western regions. Further research found that extreme weather also affects the urban–rural income gap with extreme heat narrowing the urban–rural income gap and extreme drought widening the urban–rural income gap. These findings can help the government formulate climate and economic policies that can help to narrow the urban–rural income gap.

The current study has limitations that need to be addressed further. First, using the third-level administrative unit as the study area enabled us to better ascertain the actual situation of China’s development. However, whether the impact of climate change on the urban–rural income gap can be better reflected if the fourth-level administrative unit (county level) is used as the study area is a topic that deserves in-depth discussion. Second, this paper mainly explored the impact of climate change on the urban–rural income gap within China, and in the future, we can continue to explore the differences between China and other countries to discover more general patterns. Finally, climate change is not just about temperature and rainfall, and it will be interesting to study indicators such as insolation, wind speed, and humidity.