Can Urban Green Transformation Reduce the Urban–Rural Income Gap? Empirical Evidence Based on Spatial Durbin Model and Mediation Effect Model

Abstract

The urban green transformation is the basis for the green development of China’s economy, and the reduction of income inequality between urban and rural areas is necessary to ensure stable economic growth. Therefore, ensuring green and sustainable economic development, while taking into account social equity, is of practical importance for China to achieve comprehensive high-quality development. This paper constructs a spatial Durbin model and a mediating-effects model to examine the spatial effect of urban green transformation on the urban-rural income gap (URG) and its mechanism of action based on panel data of 265 cities in China from 2006 to 2018. It also divides cities by geographical location and urban population size to further investigate the heterogeneity of the impact of the urban green transition on URG. The study found that (1) there is a significant positive spatial correlation for the URG in China, and the urban green transition can reduce the URG, and the results of the study remain reliable after a series of robustness tests. (2) Urban green transformation can reduce the URG through technological innovation effects and digital effects. (3) Urban green transformation significantly reduced the URG in eastern regions and cities of considerable size and above and had no significant impact on the URG in other cities. The study results demonstrate the possibility of reconciling urban and rural economic development and environmental friendliness at the same time.

1. Introduction

China’s economic development has entered a new normal, but the problem of unbalanced and insufficient regional development has become increasingly prominent, and the income gap between urban and rural areas has widened significantly [1]. Data on the Gini coefficient for the last 40 years published by the Statistical Information Management Centre of the National Statistics Office shows that China’s Gini coefficient gradually increased from 0.2494 in 1982 to 0.4650 in 2019, which exceeded the warning line of 0.4 published by the United Nations and belongs to the range of significant income disparity, and the income disparity shows a trend of gradually increasing [2]. China is a vast country, and the problem of uneven development in various regions is still very prominent. The problem of widening income disparity between urban and rural areas may be even more severe [3]. To address this issue, the grand report of the 19th National People’s Congress on 18 October 2017 proposed a rural revitalization strategy. It calls for strengthening cities to lead rural areas, promoting joint urban–rural development, establishing urban–rural complementarity, and achieving common prosperity in the region [4]. The full implementation of the strategy fully reflects the importance of narrowing the URG in promoting the coordinated development of urban and rural areas, achieving shared prosperity for all people, and promoting the steady progress of the “double-cycle” development pattern [5]. Therefore, how to improve the current situation of income disparity and achieving coordinated regional development is an urgent task that governments at all levels cannot avoid [6].

In the new normal of quality development in China, ecological pollution has been one of the reasons for the widening income gap between urban and rural areas [7]. Environmental pollution causes equal harm to people’s health in different income brackets, but the ability to mitigate this harm varies between income brackets [8]. Therefore, when environmental pollution is more severe, people with lower incomes suffer a more significant loss of physical and mental health due to pollution than those with higher incomes. This will lead to a decline in their labor force participation rate and a reduction in employment opportunities, in addition to other hazards, which will cause the income inequality of the population to become increasingly severe in the long run [9,10]. This leads us to consider whether the widening URG in regional development can be alleviated by promoting a green transformation of China’s cities [11]. The Central Committee of the Communist Party of China has called for the full implementation of the socialist rural revitalization path with Chinese characteristics by adhering to the path of green development and promoting sustainable development strategies [12,13]. Governments at all levels have responded positively, using the green transformation of cities to try to break the dilemma of the widening income gap between urban and rural areas [14]. Urban green transformation is a critical factor influencing the URG, so how effective is the promotion of green urban transformation on the URG? What are the mechanisms through which the urban green transition affects the URG? There are few direct answers to the above questions in recent research. Therefore, based on the perspective of green and sustainable development, this paper explores the impact of urban green transformation on the URG in China and its mechanism of action. In China’s modernization, this paper aims to promote the pace of green transformation in Chinese cities, coordinate regional development, improve the current situation of income disparity, and explore the way forward for the shared prosperity of all people.

The marginal contributions of this paper are as follows: (1) Revealing the mechanism of the impact of urban green transformation on the URG and the mechanism of action and enriching the theoretical basis of the URG based on the background of the sustainable development strategy and from the perspective of cities promoting green development; (2) exploring the heterogeneity analysis of the impact of urban green transformation on the URG through the spatial Durbin model, which deepens the study of the URG by classifying cities with different geographical locations and urban population numbers; (3) this paper extends the sample data to urban data to discover the relationship between urban green transformation and the URG in an attempt to develop more targeted policy solutions for collaborative urban–rural regional development and shared prosperity at the urban level.

The rest of the article is as follows: Section 2 reviews the relevant literature. Section 3 presents a theoretical analysis and outlines the hypotheses. Section 4 shows the model construction and variable descriptions. Section 5 illustrates the empirical results and analyses them. Section 6 provides research conclusions and Policy Implications. The research framework is shown in Figure 1.

2. Literature Review

Current research on the URG has been conducted in two main areas. On the one hand, different factors contribute to the differential impact of the URG. The adverse effects of inclusive digital finance [15,16,17], international trade [18,19,20], discriminatory government behavior [21], and administrative systems [22,23] on reducing the URG are highlighted. In addition, urbanization [24,25], structural optimization [26,27], and technological progress [28,29] have a positive impact on reducing the URG. On the other hand, there is literature on the effect of environmental pollution on the URG. In recent years, most studies have shown that environmental pollution contributes to the widening of the URG [30]. Wang et al. (2019) pointed out that environmental pollution causes homogeneous damage to the health of urban and rural residents [31]. Yet, the ability to mitigate this damage differs between urban and rural residents, which in turn affects the size of the URG. Katsouyanni (1997) and Brook (2004) also believe that pollution harms human health [32,33]. Zhang and Hu (2020) further found that the strength of the effect of environmental pollution in widening the income gap is somewhat reinforced when the early urban–rural level gap is significant [34]. In this context, Hu (2016) proposed that promoting the green transformation of the industry can alleviate the negative impact of environmental pollution on the income gap between urban and rural areas [35].

As governments at all levels pay increasing attention to urban green transformation, scholars at home and abroad have gradually enriched their research on green transformation, mainly from two perspectives: Qualitative research and quantitative research. On the one hand, qualitative research focuses on several parts such as the conceptual definition [36,37,38], measurement methods [39,40,41], the evolution process [42,43], and path exploration [44] of urban green transformation. On the other hand, quantitative studies have focused on factors such as environmental regulation [45,46], financial development [47], and foreign trade [48,49] regarding the urban green transformation. In addition, some scholars have further examined the environmental effects of the green transition. Hou (2018) observed the green transformation within the industrial sector as a significant driving force in reducing pollution emissions [50]. Zeng (2021) undertook a comprehensive exploration of the environmental effects of green transformation through China’s eco-parks [51]. The results of the study show that the green transition is conducive to reducing the environmental pollution. Promoting a green transition is an effective way not only to encourage environmental optimization but also to achieve economic development. Qin (2019) used a CGE model to estimate the costs and benefits of promoting a green transition, and the study showed that the economic benefits of a green transformation far outweigh the costs [52]. However, few scholars have studied the economic effects of urban green transformation, especially the impact of urban green transformation on the URG. Indeed, the urban green transition has a significant impact on the URG. The green transformation process has adjusted the allocation of resources, improved the quantity and quality of agricultural products, reduced agricultural costs, improved agricultural production efficiency, and provided more employment for farmers [53], all of which have contributed to improving farmers’ income. Cities are at the forefront of implementing the green transition. Has the urban green transition raised farmers’ incomes? This is the critical question that this paper seeks to answer.

Exploring the available research, it is easy to see that, firstly, separate studies are richer considering the green transition of cities and the URG. However, studies on the impact of urban green transformation on urban–rural income disparity and its mechanism of action still need to be supplemented and improved. Secondly, the impact of urban green transformation on the URG varies considerably depending on the characteristics of the city’s geographical location and urban population size, but few studies have explored the heterogeneous impact. Finally, most of the existing studies use provincial-level data as a sample, and no study has yet examined the spatial effects of urban green transformation on the URG at the city level. Therefore, using panel data from 265 Chinese cities from 2006 to 2018, this paper systematically explores the spatial impact, mechanism of action, and heterogeneity analysis of urban green transformation on the URG by constructing the spatial Durbin model and mediation effect model to propose more scientific and targeted policy recommendations.

3. Theoretical Analysis and Research Hypotheses

3.1. The Direct Impact of the Urban Green Transition on the URG

Since China put forward its sustainable development strategy of adhering to green and low-carbon development, governments at all levels have actively promoted the urban green transformation, which has narrowed the income gap between urban and rural areas [54]. Firstly, under the call of the central government to vigorously advocate green and sustainable development, governments at all levels have followed suit and paid more and more attention to the urban green transformation, and environmental pollution regulation is undoubtedly an important breakthrough and key element in achieving the green transformation [55]. As an important spatial carrier for the flow of factors, cities facilitate the flow of resources between regions and are important fronts for developing regional green economies. This is often inseparable from the economic benefits created by the development of inter-urban industrial spatial agglomerations, but it brings with it heavy environmental pollution problems resulting from the movement of factors [56]. Therefore, under the pressure of green transformation, the environmental pollution regime tends to increase the production costs of urban enterprises, leading to damage to their economic efficiency, which has a significant impact on the economic income of urban residents [57]. For rural areas, the environmental pollution problems associated with daily farming production patterns are inherently lower than those in cities. The costs of treatment are less than those in cities when faced with environmental pollution. In addition, the green and organic agricultural products produced by rural residents through the green transformation have also expanded the income channels of rural residents, thus effectively reducing the income gap with urban residents to a certain extent [58]. Secondly, the urban green transition attaches greater importance to environmental protection in the countryside. A good ecological environment ensures that the quality of production and life in the countryside does not deteriorate and is a prerequisite for safeguarding healthy human capital [59]. On the one hand, a good ecological environment not only ensures the survival rate of crops but also improves the quality of rural products, thus being able to increase the business income of rural residents [60]. On the other hand, a good ecological environment can reduce the possibility of rural residents suffering from major diseases and reduce the burden of expenditure on medical care, thus narrowing the URG [61]. In addition, the process of achieving urban green transformation often involves upgrading from traditional industries to green industries, which requires a large amount of labor. Thus, the promotion of urban green transformation can provide opportunities for the migration of agricultural labor to the manufacturing or tourism sectors [62]. This will not only re-optimize the allocation of resources but will also increase farmers’ incomes to a certain extent, thus facilitating the reduction of the URG [63]. Finally, when the urban green transformation develops to a certain extent, the externalities generated by urban development will gradually diminish or manifest themselves as diseconomies [64]. The focus on equity and efficiency at all levels of government and the policy bias and tilt will see a gradual bias towards the injection of foreign capital into rural areas [65]. This will lead to the optimization of the industrial structure in rural areas and the gradual diffusion and penetration of factor resources such as responsive supporting policies, highly qualified personnel, and advanced technology into rural areas [66]. Not only will it solve the problem of hollowing out the countryside, but it will also boost the economic development of the rural areas, ultimately leading to an increase in the income of rural residents and reducing the gap between their incomes and those of urban residents. As Chinese cities have different factor endowments, factors such as capital, technology, infrastructure levels, and talent can move freely across regions, leading to neighboring cities having similar characteristics. Moreover, the URG is often the result of a combination of geographical and social factors [67]. This joint relationship does not disappear because of the division of provincial boundaries, so the URG may be characterized by spatial correlation, the following hypothesis was proposed:

Hypothesis 1 (H1).

Urban green transformation can significantly reduce the URG and has a spatial spillover effect.

3.2. Indirect Effects of Urban Green Transformation on the URG

Numerous studies have shown that the green transformation is an important path for a country’s high-quality economic development, that high-quality economic development can effectively contribute to infrastructure development, especially new digital infrastructures such as Internet development and artificial intelligence, and that improving digitization can effectively reduce the URG [68]. Firstly, the urban green transition reduces the cost of farming for farmers by facilitating the development of the Internet, thereby reducing the URG [69]. The development of the Internet has broken through information barriers, reduced the cost of access to information, optimized the allocation of production factors, improved the efficiency of the use of agricultural capital, and better released the potential dividends of narrowing the URG [70]. Internet development allows farmers to effectively gain agricultural information, such as agricultural production techniques, market demand for agricultural products, pest and disease control, agricultural production materials, etc., through the Internet and integrated agricultural production information service platforms. Farmers have access to more information related to agricultural output inputs, which reduces information search costs [71]. The popularization of the Internet has brought farmers richer information resources on the economy, policies, and society, increasing the accessibility of information and helping residents to quickly grasp information beneficial to their value development, which is conducive to farmers making optimal market decisions and obtaining timely and accurate market information, thereby reducing production costs, increasing agricultural incomes, and narrowing the income gap between urban and rural areas [72]. Secondly, the development of the Internet has widened the marketing channels for agricultural products, increasing farmers’ income and thus narrowing the income gap between urban and rural areas [73]. With the support of the Internet platform, agricultural development can break the shackles of geography, take advantage of local natural resources and factor costs, form extensive and convenient market access, develop new businesses such as e-commerce for agricultural products and agricultural sharing economy, change the sales model of agricultural and sideline products, and broaden sales channels for rural residents to improve their income [74]. The new economic development of “Internet+” has lowered the threshold for entrepreneurship, reduced the burden of starting a business, and provided new financing channels, increasing income generation for residents and allowing for the effective allocation of social resources [75]. Finally, the Internet promotes employment for farmers and reduces the income gap between urban and rural areas [76]. The increasing integration of the Internet with market economy activities has given rise to a new online economy similar to webcasting, which has created many employment opportunities that are not bound by geographical constraints. Rural residents can access a wider range of employment opportunities through the use of the Internet [77]. Farmers earn farm income and wage income through part-time or full-time online employment and other income, which promotes diversification of non-farm employment in rural areas and boosts the wage income of rural residents [78]. The Internet has also provided a new environment for rural entrepreneurship that is open and dynamic, breaking the previous curing model of “living off the land” [79]. The establishment of e-commerce platforms has, on the one hand, widened the sales channels of agricultural products, effectively reduced the transaction costs of agricultural products, improved the efficiency of agricultural product sales, and generally increased the income of rural residents [80]. On the other hand, a group of professional e-merchants has been generated to complete the classification of local agricultural products on their own and to form scale sales on e-commerce platforms to reduce sales costs [81]. In addition, the development of the Internet has reduced the information asymmetry between farmers and buyers, helped farmers gain access to fairer information resources, improved the efficiency of agricultural trade, and increased farmers’ income, thus reducing the URG [82]. The following hypothesis was proposed:

Hypothesis 2 (H2).

Urban green transformation contributes to reducing the URG by promoting digital levels.

When the urban green transformation reaches a certain stage of development, the demand for new production equipment and advanced technologies for their internal structural optimization and production greening behavior will increase accordingly. This forces the level of technological innovation to rise and further research and development of advanced production technologies to reduce the URG [83]. Firstly, the urban green transformation has brought about an increase in technological innovation that can transform the structure and production methods of traditional agriculture into a new agricultural model that can maximize agricultural labor productivity to reduce the URG [84]. Technological innovation has transformed hand tools and animal-powered agricultural implements into advanced and applicable agricultural machinery and equipment, and the increased productivity of agricultural labor has led to higher levels of land output and agricultural commodity rates, resulting in higher incomes for farmers [85]. Technological innovation has contributed to the emergence of a large number of cooperatives in rural areas that integrate production, processing, packaging, and marketing [86]. This not only deepens the specialization of farmers’ production, but also increases the per capita ownership of resources by farmers who remain in the agricultural economy, increasing their business income and therefore reducing the URG [87]. Secondly, technological innovation has improved the quantity and quality of agricultural products to reduce the URG [88]. Advanced agricultural technology can enrich the variety of agricultural products, improve the quality of agricultural products, optimize the structure of agricultural products, and increase the level of effective supply of agricultural products [89]. New varieties and better-quality agricultural products become more elastic in demand, the price of agricultural products rises, and the income of farmers rises [90]. Technological innovation also promotes deeper processing of agricultural products, which increases the commercialization of agricultural products and makes them more valuable per unit, thus increasing farmers’ household business income to reduce the URG [91]. Finally, technological innovation will improve farmers’ labor skills and thus reduce the URG [92]. The diffusion of new agricultural technologies allows farmers to acquire the skills to optimize their labor to secure plentiful crops. Farmers effectively contribute to the knowledge spillover effects of technology competition spillover, technology imitation spillover, and technology diffusion spillover through technological innovation to reduce the URG [93]. However, technological innovation is often characterized by high investment, high risk, and long research times. Rural residents tend to adopt technology introduction as a means of production and farming due to their low level of human capital [94]. As a result, technological innovation tends to increase the economic income of rural residents at a faster rate than urban residents, thus reducing the income gap between urban and rural areas. The following hypothesis was proposed:

Hypothesis 3 (H3).

Urban green transformation promotes a reduction in the URG by driving up the level of technological innovation.

4. Model Construction and Description of Variables

4.1. Model Construction

4.1.1. Constructing the Spatial Weight Matrix

The accuracy of the model estimates depends, to a large extent, on the setting of the spatial weight matrix. The construction of the spatial weight matrix needs to comply with the basic requirement that spatial dependence decreases with increasing “distance” [95]. The distance can be measured in terms of real physical lengths or the interconnectedness of things, such as the proximity of economic or social relationships [96]. Anselin and Florax (2004) argue that the spatial weight matrix should match the true spatial structure of the sample as closely as possible to obtain a better interpretation [97]. Regional income levels are very closely related to geographical location and economic development levels; therefore, a spatial inverse distance weighting matrix is constructed in this paper. The inverse of the geographical distance between the two cities is used as an indicator. The spatial inverse weighting matrices are set as follows:

$${\mathrm{W}}_{\mathrm{ij}}=\{\begin{array}{c}1/{\mathrm{d}}_{\mathrm{ij}} , \mathrm{i}\ne \mathrm{j}\\ 0, \mathrm{i}=\mathrm{j}\end{array}$$

(1)

where $\mathrm{d}$ is the distance between cities.

4.1.2. Constructing Spatial Correlation Model

The use of spatial econometric models presupposes that variables are spatially correlated [98]. Therefore, this paper uses Moran’s I index to conduct a bilateral test for global spatial correlation of the variable URG to identify the significance of Moran’s I index by whether the Z statistic is significant. The global Moran index is used to measure the correlation and aggregation state of the space as a whole and takes values in the range of (−1,1). The value of the Moran index is greater than 0, indicating that there is a positive correlation between spatial units, showing spatial agglomeration; the value of the Moran index is less than 0, indicating a negative correlation between spatial units, showing spatial diffusion; the value of the Moran index is equal to 0, indicating that there is no correlation between spatial units, and the observed values of each indicator show independent random distribution. The calculation formula is as follows:

$${\mathrm{I}}_{\mathrm{morans}}=\frac{\mathrm{n}{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{ij}}\left({\mathrm{X}}_{\mathrm{i}}-\overline{\mathrm{X}}\right)\left({\mathrm{X}}_{\mathrm{j}}-\overline{\mathrm{X}}\right)}{{\mathrm{S}}^2{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{W}}_{\mathrm{ij}}}$$

(2)

where ${\mathrm{S}}^2={\displaystyle \sum }_{\mathrm{i}=1}^{\mathrm{n}}{\left({\mathrm{X}}_{\mathrm{i}}-\overline{\mathrm{X}}\right)}^2$ is the sample variance; $\overline{\mathrm{X}}=\frac{1}{\mathrm{n}}{\displaystyle \sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{X}}_{\mathrm{i}}$ is mean value of URG; n is the total number of cities; ${\mathrm{X}}_{\mathrm{i}}$ and ${\mathrm{X}}_{\mathrm{j}}$ represent the level of URG for i and j, respectively; and ${\mathrm{W}}_{\mathrm{ij}}$ is the spatial weight matrix.

4.1.3. Constructing Spatial Durbin Model

Spatial econometric models generally cover the spatial lag model, the spatial error model, and the spatial Durbin model. The spatial lag model is used primarily to analyze the presence or absence of spatial effects of the explained variables. The spatial error model is used to analyze the spatial effects of omitted variables. The spatial Durbin model is a general form of the spatial lag and spatial error models that can be used to analyze the spatial effects of the explained variables based on the explanatory variables. It can also be used to address the problem of adding or missing variables and reduce the bias in the parameter estimates of the independent variables and the error terms [99]. Urban economic development cannot be separated from the cross-regional movement of production factors, especially the large-scale movement of labor. To comprehensively examine the impact of green transformation between local cities and spatially connected cities on the URG, this paper adopts the Spatial Durbin model as follows:

$${\mathrm{URG}}_{\mathrm{it}}={\mathsf{\beta }}_0+\mathsf{\rho }{\displaystyle \sum }_{\mathrm{j}=1}^{\mathrm{N}}{\mathrm{W}}_{\mathrm{ijt}}{\mathrm{URG}}_{\mathrm{it}}+{\mathsf{\beta }}_1{\mathrm{Eco}}_{\mathrm{it}}+{\mathsf{\beta }}_2{\displaystyle \sum }_{\mathrm{i}\ne \mathrm{j}}^{\mathrm{N}}{\mathrm{W}}_{\mathrm{ijt}}{\mathrm{Eco}}_{\mathrm{it}}+{\displaystyle \sum }_{\mathrm{k}=1}^6{\mathsf{\delta }}_{\mathrm{k}}{\mathrm{X}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(3)

In Equation (3), $\mathrm{i}$ denotes a prefecture-level city, $\mathrm{t}$ denotes a year, ${\mathrm{URG}}_{\mathrm{it}}$ is the urban–rural income gap; ${\mathrm{Eco}}_{\mathrm{it}}$ is the urban green transformation; ${\mathrm{W}}_{\mathrm{ijt}}$ is the spatial weight matrix; ${\mathrm{X}}_{\mathrm{it}}$ denotes a series of control variables, including the level of economic development (GDP), population density (POP), advanced industrial structure (Ind), level of financial development (Fin), fiscal decentralization (Fd), and level of infrastructure(Infra); $\mathsf{\rho }$ denotes the spatial spillover coefficient of the urban–rural income gap; ${\mathsf{\beta }}_1$ denotes the coefficient to be estimated; and ${\mathsf{\epsilon }}_{\mathrm{it}}$ denotes the random disturbance term.

4.1.4. Constructing Mediation Effect Model

The mediation effect model is used to test the effect of an independent variable on a dependent variable through a particular variable. According to the theoretical analysis above, urban green transformation can have an indirect effect on the URG by affecting technological innovation and digitization levels, respectively, as mediation effects. Therefore, based on the test for mediation effects proposed by Li (2020) [100], the mediation effect model was constructed as shown in Figure 2.

$${\mathrm{URG}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{Eco}}_{\mathrm{it}}+{\mathsf{\beta }}_{\mathrm{n}}{\mathrm{X}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(4)

$${\mathrm{M}}_{\mathrm{it}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{Eco}}_{\mathrm{it}}+{\mathsf{\alpha }}_{\mathrm{n}}{\mathrm{X}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(5)

$${\mathrm{URG}}_{\mathrm{it}}={\mathsf{\phi }}_0+{\mathsf{\phi }}_1{\mathrm{Eco}}_{\mathrm{it}}+{\mathsf{\theta }\mathrm{M}}_{\mathrm{it}}+{\mathsf{\phi }}_{\mathrm{n}}{\mathrm{X}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(6)

where i and t denote city and year, respectively; ${\mathrm{URG}}_{\mathrm{it}}$ is the urban–rural income gap; ${\mathrm{Eco}}_{\mathrm{it}}$ is the urban green transformation; ${\mathrm{M}}_{\mathrm{it}}$ denotes the mediating variables, which are the level of digitization and technological innovation, respectively; ${\mathrm{X}}_{\mathrm{it}}$ denotes control variables; and ${\mathsf{\epsilon }}_{\mathrm{it}}$ denotes the random error term.

According to the procedure for the mediation effect test, Equation (4) is a test of whether the urban green transition has reduced the URG. If the coefficient ${\mathsf{\beta }}_1$ is significant, then the urban green transition significantly reduces the URG, as the theoretical hypothesis above suggests. Equation (5) tests whether the green transformation of cities significantly affects the level of technological innovation and digitization. If the coefficient ${\mathsf{\alpha }}_1$ is significant, it indicates that the urban green transition brings about the impact of both effects. Equation (6) is a test of whether the effects of urban green transformation and mediating variables on the URG at the same time are significant. If the coefficient $\mathsf{\theta }$ is significant, it indicates a mediation effect between the three variables. If ${\mathsf{\phi }}_1$ does not pass the significance test, it indicates a full mediation effect. If it passes, it indicates a partial mediation effect. This paper again verifies the existence of a mediation effect by using the Sobel test and the Bootstrap test.

4.2. Variable Selection and Data Sources

4.2.1. Explained Variable

Regarding the urban–rural income gap (URG), much of China’s income inequity is expressed in the URG [101]. Indicators of URG disparity are usually expressed in three ways: The ratio of disposable income per capita of urban residents to income per capita of rural residents, the Gini coefficient, and the Theil index. When using the Theil index as an indicator of the URG, although the population income gap is taken into account, the Thiel coefficient is affected by the population weights [102]. The Gini coefficient is the most desirable measure of the three, but the lack of statistical data for each city makes it impossible to calculate the Gini coefficient [103]. Therefore, this paper refers to Yu and Wang’s (2021) study and uses the ratio of urban residents’ per capita disposable income to rural residents’ per capita disposable income to measure, which can reflect the urban and rural income situation more intuitively [104]. The higher the URG disparity ratio, the more serious the urban–rural income disparity situation.

4.2.2. Explanatory Variable

Regarding the urban green transformation (Eco), referring to the study by Feng (2020), this paper uses green total factor productivity to measure the urban green transformation [105]. The indicator not only considers the important impact of substitution effects between energy consumption, labor, and capital in the urban green transformation but also takes into account the environmental problems caused by the pollution emissions accompanying the development and use of energy, reflecting the essence and content of the urban green transformation. This paper measures green total factor productivity using an SBM model that considers undesired outputs, where the input factors include capital inputs, labor inputs, and energy inputs. Capital input (K) is based on the perpetual inventory method and is measured against the city’s capital stock for a base period of 2005; Labor input (L) is measured by the number of people employed in the city, and energy input (E) is measured by the per capita consumption of electricity in the city. The output factor is measured as a pollution index for the city, which is derived by combining sulfur dioxide emissions (S), soot emissions (Q), and wastewater emissions (W) for the city using the entropy method.

4.2.3. Mediating Variables

(1) Level of digitization (Internet). Referring to Wang and Hao (2018), this paper uses urban internet penetration to measure digitization [106]. The spread of the Internet has greatly facilitated the accessibility of information between regions in China, helping to remove barriers between different regions, promoting the flow and allocation of factors between different regions, and laying a solid foundation for the development of the digital economy. When cities have higher internet penetration rates, the level of digitization is stronger.

(2) Technological innovation (Patent). Referring to Popp (2005), this paper uses the total number of invention patents granted in cities to measure the level of technological innovation [107]. Technological innovation measures a city’s ability to innovate, and the number of patents is a common indicator of technology levels used in research. In China, patents include invention patents, utility model patents, and design patents, which can objectively reflect a city’s original innovation capability and comprehensive competitiveness in science and technology [108]. When the total number of patents granted for inventions is higher, it represents a higher level of technological innovation in the city.

4.2.4. Control Variables

(1) Level of economic development (GDP). Economic development and income disparity show an inverted U-shape. In the later stages of economic development, economic growth and further development of markets will reduce the URG [109]. Then, adjusting the long-term goal of social development from the traditional pursuit of the economic growth rate to the overall harmonious development of society can reduce the URG. Referring to Solomonov (2021), this paper uses the ratio of the city’s gross regional product to the total population at the end of the year to measure the level of economic development [110].

(2) Population density (POP). Referring to Tang (2022), this paper uses the ratio of the number of people in the city to the area of the administrative area to measure the population density [111]. When the ratio is higher, it represents a higher population density.

(3) Advanced industrial structure (Ind). The optimization of the income distribution cannot be separated from the process of transformation of the industrial structure from lower to higher levels [112]. A sound industrial structure can attract more rural labor to move, thereby reducing income inequality between urban and rural areas. In this paper, we refer to Zhou’s (2013) study and use the ratio of the total tertiary sector output to the total secondary sector output in cities to measure the level of advanced industrial structure [113].

(4) Level of financial development (Fin). Finance, at the heart of the modern economy, contributes to economic growth and has an impact on the URG. On the one hand, urban residents have an advantage over rural residents in terms of the affordability of financing costs and the ability to repay loans, so financial development may have contributed to the URG. On the other hand, financial development generates more agricultural support services, increases farmers’ incomes, and reduces the URG [114]. Based on this and drawing from Huang and Zhang’s research (2020), this paper measures financial development in terms of year-end balances of loans to financial institutions in cities [115].

(5) Financial decentralization (Fd). In the new round of reform of China’s fiscal and taxation system, more autonomy has been given to local governments. The increasing level of fiscal decentralization has led to the strengthening of local fiscal autonomy, which has a significant positive impact on reducing the URG. Referring to Chen (2020), this paper uses the ratio of per capita urban fiscal expenditure to the sum of per capita central fiscal expenditure [116], per capita province-level fiscal expenditure, and per capita urban fiscal expenditure to measure fiscal decentralization.

(6) Level of infrastructure (Infra). Referring to Sun (2017), this paper measures the infrastructure level using the ratio of urban paved road area to population [117]. The level of urban infrastructure is higher when the paved road area per capita is greater.

4.2.5. Data Sources and Descriptive Statistics of Variables

The disposable income of rural residents in the statistical indicators of urban–rural income disparity in this paper, which the statistics department used to refer to the net income per rural resident before 2013, was changed to disposable income per rural resident from 2013 onwards. Since no specific conversion method has been published, and since the differences between the two are small and the trends are consistent, the authors uniformly refer to this statistical indicator as disposable income per rural resident. This paper uses panel data for 265 prefecture-level cities in China from 2006 to 2018, with data from the China Statistical Yearbook, the China City Statistical Yearbook, the statistical yearbooks of each city in previous years, the Economy Prediction System database (EPS database), and the China Stock Market & Accounting Research Database (CSMAR database). All indicators are treated logarithmically, except for the indicators Internet, Ind, and FD. For some missing data, this paper complements them by interpolation. The descriptive statistics of variables are shown in

Table 1. Variable description statistics.

Variable	Obs	Mean	Std.Dev	Min	Max
URG	3445	0.9013	0.2293	0	3.3556
Eco	3445	0.0529	0.2175	−1.6129	1.6475
GDP	3445	10.4125	0.7152	4.5951	15.6752
POP	3445	5.7556	0.9012	1.5476	7.8816
Ind	3445	0.8519	0.4183	0.0943	4.2441
Fin	3445	16.2713	1.1909	13.0707	20.2846
Fd	3445	0.2589	0.1528	0	0.8807
Infra	3445	2.2967	0.5949	−1.1712	4.6856
Internet	3445	66.0248	81.4136	0.0237	766
Patent	3445	6.6043	1.6929	1.7918	11.8475

.

5. Empirical Results and Analysis

5.1. Spatial Correlation Test

5.1.1. Global Spatial Correlation Test

This paper constructs a spatial inverse distance weighting matrix and uses the Moran index model to test the spatial correlation. The global Moran values of the URG for each city in China during the study sample period are all significantly positive at the 1% confidence interval, indicating a significant positive spatial correlation and some spatial clustering of the URG in China. The results of the test are shown in

Table 2. Spatial correlation test for URG.

Time	URG
	Moran’s I	z	p-Value
2006	0.088	16.572	0.000
2007	0.083	15.97	0.000
2008	0.102	19.073	0.000
2009	0.107	19.914	0.000
2010	0.102	19.900	0.000
2011	0.129	23.832	0.000
2012	0.122	22.735	0.000
2013	0.115	21.331	0.000
2014	0.122	22.746	0.000
2015	0.095	17.872	0.000
2016	0.105	19.586	0.000
2017	0.096	18.061	0.000
2018	0.101	18.871	0.000

.

5.1.2. Local Spatial Correlation Test

To further explore the spatial clustering distribution state of the URG, this paper examines the spatial clustering characteristics of the URG through a local Moran scatter plot under a spatial inverse distance weighting matrix. Figure 3 shows the Moran scatter plot of the URG in 2006 and 2018. The figure shows that most cities fall in the first quadrant of high–high agglomeration and the third quadrant of low–low agglomeration, indicating that cities with higher URG are surrounded by high-value cities and cities with lower URG are surrounded by low-value city shares. There is a positive spatial correlation between urban and rural income gaps.

5.2. The Direct Impact of the Urban Green Transition on the URG

5.2.1. Validation of the Applicability of the Model

To verify the validity of the SDM model chosen in this paper, the LM, LR, and WALD tests need to be performed on the model set. The estimation results are shown in

Table 3. Spatial model selectivity test.

Test Methods	Value	p-Value	Test Methods	Value	p-Value
Moran’s I	47.363	0.000
LM-error	1758.320	0.000	LR-SAR	581.30	0.000
LM-error(robust)	1143.344	0.000	LR-SEM	−82.35	1.000
LM-lag	642.339	0.000	WALD-SAR	33.93	0.000
LM-lag(robust)	27.363	0.000	WALD-SEM	50.85	0.000

. The tests of LM-error, LM-error(robust), LM-lag, and LM-lag(robust) all reject the original hypothesis at the 1% significance level, indicating that the spatial error model and the spatial lag model alone may be biased in examining the spatial spillover effect of the URG, so the choice of the spatial Durbin model is reasonable. The LR and WALD tests show that the SDM model does not degenerate the SAR and SEM models, further demonstrating the validity of the model set up in this paper. Therefore, the spatial Durbin model is chosen to explore the impact of urban green transformation on the URG.

5.2.2. Analysis of Spatial Regression Results

This paper uses the SDM model for empirical analysis and regresses the urban green transition on the urban–rural income gap by adding control variables one by one, and the results are shown in

Table 4. The direct impact of Eco on the URG.

Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Eco	−0.0633 ***	−0.0339 **	−0.0430 ***	−0.0322 **	−0.0329 **	−0.0332 **	−0.0325 **
	(0.0144)	(0.0138)	(0.0149)	(0.0140)	(0.0140)	(0.0140)	(0.0140)
GDP		−0.0851 ***	−0.0756 ***	−0.0801 ***	−0.0975 ***	−0.0864 ***	−0.0884 ***
		(0.0058)	(0.0062)	(0.0059)	(0.0068)	(0.0087)	(0.0088)
POP			−0.0364 ***	−0.0255 ***	−0.0330 ***	−0.0335 ***	−0.0326 ***
			(0.0050)	(0.0047)	(0.0051)	(0.0051)	(0.0051)
Ind				0.0021	−0.0096	−0.0075	−0.0070
				(0.0077)	(0.0083)	(0.0083)	(0.0083)
Fin					0.0195 ***	0.0218 ***	0.0209 ***
					(0.0043)	(0.0044)	(0.0045)
Fd						−0.0716 **	−0.0765 **
						(0.0353)	(0.0355)
Infra							0.0084
							(0.0064)
W×Eco	−0.0306	0.4409 ***	−0.6640 ***	0.1109	0.1898	0.1805	0.2122
	(0.1732)	(0.1696)	(0.1822)	(0.1716)	(0.1713)	(0.1713)	(0.1741)
$\mathsf{\rho }$	2.6903 ***	3.5465 ***	0.9261 ***	2.6302 ***	2.6208 ***	2.6188 ***	2.6175 ***
	(0.0285)	(0.0500)	(0.0202)	(0.0378)	(0.0388)	(0.0390)	(0.0391)
δ2	0.0302 ***	0.0276 ***	0.0320 ***	0.0281 ***	0.0278 ***	0.0278 ***	0.0277 ***
	(0.0007)	(0.0007)	(0.0008)	(0.0007)	(0.0007)	(0.0007)	(0.0007)
R2	0.0281	0.2434	0.1089	0.1604	0.1843	0.2097	0.1951
N	3445	3445	3445	3445	3445	3445	3445

Note: ***, ** indicate significance at 1%, 5% confidence intervals, respectively; () indicates standard error.

. The correlation coefficient $\mathsf{\rho }$ for the URG is significant at the 1% level, indicating a significant inter-urban-regional interaction between the urban and rural income gaps. The local URG impacts the URG between neighboring cities, and the URG has a spatial effect. The regression results for the core explanatory variables show that the urban green transition has a significant effect on the suppression of URG at least at the 5% level. This suggests that, spatially, the URG is significantly inhibited by the urban green transition, which supports research hypothesis 1 of this paper.

In terms of control variables, the economic development level coefficient is significantly negative, indicating that a higher level of overall economic development in the city can reduce the URG in the city. The coefficient on population density is significantly negative, indicating that higher regional population density indicates a smaller URG. The coefficient of advanced industrial structure is negative, indicating that a higher level of advanced industrial structure can reduce the URG, but the results are not significant. The coefficient on the level of financial development is significantly positive, indicating that a higher level of financial development promotes the URG. The coefficient of fiscal decentralization is significantly negative, implying that a high level of fiscal decentralization reduces the URG. The coefficient on infrastructure development is negative but not significant, suggesting that there may also be some heterogeneity in the extent to which higher levels of urban infrastructure reduce the URG.

5.2.3. Heterogeneity Analysis

(1) Geographical Heterogeneity

The above findings validate that urban green transformation can significantly reduce the URG. However, China is a vast country, and cities in different geographical locations differ greatly in terms of their level of economic development, level of infrastructure development, level of financial support, population density, and foreign direct investment. Therefore, this paper divides the urban panel data into eastern, central, and western regions according to different geographical locations and examines the differential impact of urban green transformation on the URG. The empirical results are shown in

Table 5. Estimated results of geographical heterogeneity.

Variable	Eastern	Central	Western
Eco	−0.0653 ***	−0.0110	−0.0366
	(0.0187)	(0.0224)	(0.0341)
GDP	−0.0029	−0.1326 ***	−0.1679 ***
	(0.0131)	(0.0169)	(0.0194)
POP	−0.1125 ***	−0.0296 ***	−0.0226 *
	(0.0086)	(0.0079)	(0.0126)
Ind	0.0216 **	−0.0323 **	−0.0636 **
	(0.0106)	(0.0132)	(0.0270)
Fin	0.0381 ***	0.0463 ***	0.0440 ***
	(0.0064)	(0.0073)	(0.0131)
Fd	−0.1106 **	0.1291 *	−0.2835 ***
	(0.0477)	(0.0759)	(0.0805)
Infra	0.0245 ***	0.0172	−0.0059
	(0.0090)	(0.0107)	(0.0164)
W×Eco	−0.1549	−0.7396 ***	0.0036
	(0.1295)	(0.1962)	(0.2683)
$\mathsf{\rho }$	0.7996 ***	0.7743 ***	0.7888 ***
	(0.0491)	(0.0569)	(0.0561)
δ2	0.0188 ***	0.0263 ***	0.0355 ***
	(0.0007)	(0.0010)	(0.0020)
R2	0.1396	0.1100	0.0789
N	1391	1378	676

Note: ***, **, * indicate significance at 1%, 5%, and 10% confidence intervals, respectively; () indicates standard error.

, where the coefficient of urban green transition is significantly negative at the 1% level in the eastern region, reducing the URG. In the central region and the eastern region, the coefficient of urban green transformation is negative but insignificant, indicating that the effect of urban green transformation in narrowing the URG is insignificant.

(2) Heterogeneity in Urban Population Size

Based on the Notice of the State Council on the Adjustment of the Standard for the Classification of City Size, this paper classifies cities with a population of less than 1 million as small and medium-sized cities, cities with a population greater than or equal to 1 million and less than 5 million as large cities, and cities with a population greater than 5 million as mega cities, using the resident population in 2014 as the standard. Based on the above classification criteria, we examine the heterogeneous impact of the green transition on the URG in cities with different population sizes. As shown in

Table 6. Results of estimating heterogeneity in urban population size.

Variable	Mega-Cities	Large Cities	Small and Medium Cities
Eco	−0.0697 ***	−0.0378 *	0.0278
	(0.0206)	(0.0204)	(0.0689)
GDP	−0.1045 ***	−0.1044 ***	−0.0271
	(0.0147)	(0.0128)	(0.0553)
POP	−0.0927 ***	−0.0371 ***	0.1043 ***
	(0.0121)	(0.0069)	(0.0218)
Ind	0.0186	−0.0456 ***	0.0356
	(0.0187)	(0.0112)	(0.0243)
Fin	0.0243 ***	0.0133	−0.1957 ***
	(0.0093)	(0.0085)	(0.0321)
Fd	0.0260	−0.0451	−0.3761 **
	(0.0585)	(0.0525)	(0.1753)
Infra	0.0446 ***	0.0209 **	−0.0124
	(0.0092)	(0.0096)	(0.0265)
W×Eco	−0.3559 *	−0.4009 *	0.0803
	(0.1886)	(0.2070)	(0.1718)
$\mathsf{\rho }$	0.4651 ***	0.8568 ***	−0.0772
	(0.1181)	(0.0379)	(0.1543)
δ2	0.0196 ***	0.0365 ***	0.0132 ***
	(0.0008)	(0.0011)	(0.0016)
R2	0.1998	0.1013	0.6696
N	1235	2067	130

Note: ***, **, * indicate significance at 1%, 5%, and 10% confidence intervals, respectively; () indicates standard error.

. In mega cities, the coefficient of urban green transformation is significantly negative at the 1% level, indicating that urban green transformation can significantly reduce the urban income gap. Among large cities, the coefficient of urban green transition is significantly negative at the 10% level and is smaller than that of mega-cities, suggesting that the suppressive effect of urban green transition on the URG is significantly greater in mega cities than in large cities. For small and medium-sized cities, the coefficient on urban green transition is positive but insignificant, suggesting that the green transition in small and medium-sized cities may contribute to the URG.

5.2.4. Robustness Tests

In order to analyze the accuracy of the conclusions obtained under the above full-sample conditions, robustness tests were carried out in the following aspects: (1) Change in the measurement of core explanatory variables: To avoid bias in the results due to measurement methods, this paper uses a single factor energy intensity measure of urban green transition, obtained using the ratio of energy consumption to GDP, and uses the same spatial Durbin model and other variables to test the impact of urban green transition on the URG. The regression results, as can be seen in

Table 7. Robustness test.

Variable	(1)	(2)	(3)	(4)	(5)
Eco	−0.1040 **	−0.0391 ***	−0.0551 ***	−0.0435 ***	−0.0580 ***
	(0.0426)	(0.0135)	(0.0152)	(0.0153)	(0.0163)
GDP	−0.0884 ***	−0.0883 ***	−0.1174 ***	−0.0898 ***	−0.1143 ***
	(0.0087)	(0.0086)	(0.0089)	(0.0093)	(0.0121)
POP	−0.0338 ***	−0.0160 ***	−0.0291 ***	−0.0351 ***	−0.0313 ***
	(0.0052)	(0.0045)	(0.0041)	(0.0049)	(0.0046)
Ind	−0.0049	−0.0129	−0.0386 ***	−0.0153 *	−0.0455 ***
	(0.0082)	(0.0081)	(0.0086)	(0.0090)	(0.0093)
Fin	0.0187 ***	0.0151 ***	0.0292 ***	0.0260 ***	0.0338 ***
	(0.0046)	(0.0044)	(0.0048)	(0.0049)	(0.0052)
Fd	−0.0411	−0.0692 **	−0.1196 ***	−0.1112 ***	−0.1081 ***
	(0.0364)	(0.0350)	(0.0333)	(0.0352)	(0.0381)
Infra	0.0134 **	0.0085	0.0025	0.0019	−0.0027
	(0.0064)	(0.0062)	(0.0068)	(0.0071)	(0.0073)
W×Eco	−2.8286 ***			2.0659 ***	0.0011
	(0.4381)			(0.4107)	(0.0448)
$\mathsf{\rho }$	2.6341 ***	3.5011 ***	0.9496 ***	−0.9301 ***	−0.1731 ***
	(0.0373)	(0.0506)	(0.0138)	(0.2950)	(0.0313)
δ2	0.0273 ***	0.0274 ***	0.0331 ***	0.0326 ***	0.0379 ***
	(0.0007)	(0.0007)	(0.0008)	(0.0008)	(0.0009)
R2	0.1628	0.2330	0.1302	0.0042	0.2375
N	3445	3445	3445	3445	3445

Note: ***, **, * indicate significance at 1%, 5%, and 10% confidence intervals, respectively; () indicates standard error.

-(1), show that the coefficient on urban green transformation is negative at least at the 5% significance level, indicating the robustness of the regression results above. (2) Replacement of the estimation model: In addition to the spatial Durbin model, spatial error models (SEM) and spatial lag models (SAR) are also used to test spatial effects. Therefore, to test the robustness of the model, this study replaces the estimated model and uses the SAR model and SEM model to examine the impact of urban green transformation on the URG. The estimated results are shown in (2) and (3), where the coefficient on urban green transformation is negative at least at the 1% significance level under the SAR and SEM model regressions, indicating the robustness of the regression results above. (3) Replacing the spatial weight matrix: To check the robustness of the results, we have used the spatial geographical weight matrix and spatial economic weight matrix. The estimated results are shown in (4) and (5), where the coefficient of urban green transformation is significantly negative, indicating that urban green transformation significantly reduces the URG, which further validates the robustness of the previous empirical analysis.

5.3. Indirect Effects of Urban Green Transformation on the URG

The results of the spatial Durbin model estimation above show that the urban green transition significantly reduces the URG. So how is the urban green transition reducing the URG? Based on the theoretical analysis and research hypotheses, it is clear that urban green transformation may affect the URG through the role of digital and innovation. As can be seen from regression (1) in

Table 8. Mediation effect test.

Variable	(1)	(2)	(3)	(4)	(5)
	URG	Internet	URG	Patent	URG
Eco	−0.0570 ***	8.6972 **	−0.0547 ***	0.1516 ***	−0.0439 ***
	(0.0158)	(4.1594)	(0.0157)	(0.0552)	(0.0151)
M			−0.0003 ***		−0.0863 ***
			(0.0001)		(0.0046)
GDP	−0.1613 ***	−1.4213	−0.1617 ***	0.2729 ***	−0.1378 ***
	(0.0084)	(2.2242)	(0.0084)	(0.0295)	(0.0081)
POP	−0.0318 ***	8.0671 ***	−0.0297 ***	0.4514 ***	0.0071
	(0.0044)	(1.1699)	(0.0045)	(0.0155)	(0.0047)
Ind	−0.0651 ***	21.9027 ***	−0.0592 ***	0.0593 *	−0.0599 ***
	(0.0089)	(2.3392)	(0.0090)	(0.0310)	(0.0085)
Fin	0.0247 ***	45.2935 ***	0.0368 ***	0.9183 ***	0.1039 ***
	(0.0047)	(1.2493)	(0.0056)	(0.0166)	(0.0062)
Fd	−0.0497	58.3551 ***	−0.0341	0.4471 ***	−0.0111
	(0.0319)	(8.4110)	(0.0320)	(0.1112)	(0.0305)
Infra	−0.0034	−8.2358 ***	−0.0056	0.2098 **	0.0147 **
	(0.0074)	(1.9481)	(0.0074)	(0.0259)	(0.0071)
Cons	2.4411 ***	−717.9031 ***	2.2496 ***	−14.4330 ***	1.1962 ***
	(0.0695)	(18.3235)	(0.0834)	(0.2432)	(0.0943)
Sobel test		−0.0023 *		−0.0131 ***
		(0.0012)		(0.0048)
Indirect effect		−0.0023 **		−0.0131 **
		[−0.0046, −0.0000]		[−0.0236, −0.0026]
Direct effect		−0.0547 ***		−0.0439 **
		[−0.0900, −0.0194]		[−0.0776, −0.0102]
R2	0.2411	0.5813	0.2448	0.8294	0.3102
N	3445	3445	3445	3445	3445

Note: ***, **, * indicate significance at 1%, 5% and 10% confidence intervals, respectively; () indicates standard error.

, the total effect of urban green transition on the suppression of URG is 0.0570. It passes the 1% significance level test, which is consistent with the previous findings. Regressions (2) and (3) represent the estimated results of urban green transformation on urban Internet penetration and the estimated results of urban green transformation and urban Internet penetration on URG, respectively. From regression (2), it can be seen that the regression coefficient of the urban green transformation on the digital effect of cities is significantly positive at the 5% level, indicating that the urban green transformation promotes urban Internet penetration, which is conducive to enhancing the digital effect of cities. In regression (3), the regression coefficient of the digitization effect on the URG is significantly negative at the 1% level, indicating that the digitization effect can effectively reduce the URG, which confirms hypothesis 2. In summary, the Sobel test is significant at the 10% level, indicating that urban green transformation can reduce the URG by increasing the urban digital effect. The Bootstrap test showed a mediation effect of −0.0023, and there is a mediating effect. Regressions (4) and (5) indicate the test results of the innovation effect, where (4) shows that the regression coefficient of the urban green transformation on the urban innovation effect is significantly positive at the 1% level, indicating that the urban green transformation promotes the level of urban technological innovation. The regression coefficient of the urban innovation effect on the URG in regression (5) is significantly negative at the 1% level, indicating that the level of urban innovation can effectively reduce the URG. In summary, the Sobel test is significant at the 1% level, indicating that urban green transformation can reduce the URG by increasing the urban innovation effect, which confirms hypothesis 3. The Bootstrap test showed a mediation effect of −0.0131, providing evidence that there is a mediating effect.

6. Discussion

6.1. Discussion on the Spatial Regression

The spatial regression in Table 4 shows that the urban green transformation significantly reduced the income gap between urban and rural areas. An analysis of the reasons for this may be that, firstly, under the pressure of urban green transformation, for cities and towns, environmental pollution policy increases the production costs of urban enterprises and reduces the economic effects of enterprises, resulting in relatively less income and narrowing the income gap between urban and rural areas [118]. For rural areas, the environmental pollution problems brought about by daily farming production patterns are inherently lower than those in cities, the costs of treatment, when faced with environmental pollution, are less than those in the cities, and the green and organic agricultural products derived from the green transformation of rural residents have also expanded the income channels of rural residents, thus effectively reducing the income gap with urban residents to a certain extent [119]. Secondly, the process of urban green transformation often involves upgrading from traditional industries to green industries, which requires a large amount of labor [120]. The promotion of urban green transformation can provide opportunities for the transfer of agricultural labor to the non-agricultural sector, which can not only re-optimize the allocation of resources but also increase farmers’ income to a certain extent, thus facilitating the reduction of the income gap between urban and rural areas [121]. In addition, the green transformation of urban development has enabled rural areas to enjoy preferential policies to attract quality talents and advanced technologies, which further promote rural economic development and narrow the income gap between urban and rural areas [122].

6.2. Discussion on the Heterogeneous Influence

Table 5 shows the impacts of Urban Green Transformation on the URG in the eastern, central, and western regions. This may be due to the fact that cities in the eastern region have developed infrastructure, technology levels, and highly skilled personnel that can facilitate digitization and innovation levels. Therefore, cities have the highest level of green transformation development and can best alleviate the URG [123]. The “latecomer advantage” in infrastructure and technology in the central region has a slightly weaker dampening effect on the URG than in the eastern region. In contrast, in the central region and the western region, its backward technological innovation, labor force skill level, and infrastructure are not conducive to the development and application of information and communication technology [124]. Enterprises are predominantly labor-intensive and resource-intensive, the financial sector is reluctant to lend to enterprises, and there is a massive exodus of labor, resulting in the slow development of digitization and technological innovation levels [125]. This is not conducive to the transformation of the rural labor force from traditional agriculture to modern agriculture, manufacturing, and services, thus reducing URG.

Table 6 shows the impact of the urban green transition on the URG in mega cities, large cities, and small and medium-sized cities. This may be due to the more significant economic agglomeration effect of mega cities, where cities with larger populations are better equipped in terms of infrastructure level, education level, and technology level, driving up farmers’ income, which in turn reduces the URG [126]. For small and medium-sized cities, this may be because the smaller the city, the less developed the infrastructure, the lower the level of technological innovation, and the relatively more backward development of the Internet, which is not conducive to increasing farmers’ income [127].

6.3. Discussion of the Results of the Transmission Mechanism

Tests of transmission mechanisms such as Table 8 show that the urban green transformation can indirectly reduce the URG through technological innovation and digital levels. This may be due to, on the one hand, urban green transformation for Internet development not only accelerating the speed of dissemination of central and local government policy documents but also disseminating advanced production technologies through knowledge spillover effects, reducing the cost of access to information in rural areas [128]. With the support of Internet platforms, agricultural development can break the shackles of geography and take advantage of local natural resources and factor costs to form wide and convenient access to markets [129]. Developing new business models such as e-commerce for agricultural products and the agricultural sharing economy can transform the sales model of agricultural and sideline products and broaden sales channels for rural residents. The continuous integration of the Internet with market economy activities has given rise to a new online economy, such as webcasting, which has created a large number of employment opportunities that are not bound by geographical constraints [130]. Residents can access more employment opportunities through the use of the Internet, promoting employment among farmers and enabling the reduction of the URG [131]. On the other hand, the urban green transformation promotes technological innovation and transforms the structure and production methods of traditional agriculture into a new agricultural model that can maximize agricultural labor productivity to reduce the URG [132]. Advanced agricultural technology can enrich the variety and improve the quality of agricultural products, thus increasing farmers’ total income. Technological innovation will also promote the deep processing of agricultural products and increase the unit value of agricultural products, thus increasing farmers’ income. Innovation leads enterprises to establish agro-industrial alliances with farmers, promptly guiding farmers to adjust crop varieties, providing agricultural cultivation techniques according to market conditions, improving farmers’ labor skills, guiding most farmers to achieve common prosperity, and narrowing the scope of agricultural employment [133].

7. Conclusions and Policy Implications

Based on panel data from 265 Chinese cities from 2006 to 2018, the article constructs a spatial Durbin model and a mediation effect model to empirically examine the spatial effects, transmission mechanism, and heterogeneity analysis of urban green transformation on the URG. The study results show that, (1) in general, there is a significant spatial correlation between urban and rural income gaps, and the urban green transition has a significant spatial inhibitory effect on the URG. (2) The transmission mechanism test shows that the urban green transition reduced the URG through two mechanisms: The technological innovation effect and the digital effect. (3) Heterogeneity analysis shows that, for different urban locations, the suppressive effect of urban green transformation on the URG is higher in the eastern regions than in the central and western regions. For different urban population sizes, the suppression effect of urban green transformation on the URG is significantly greater in mega cities than in large cities, while the suppression effect of urban green transformation on the URG is not significant in small and medium-sized cities.

Based on the above findings, this paper proposes the following policy implications to better reduce the URG through urban green transformation:

(1) Deepen the digital transformation and actively broaden the channels for farmers to increase their income; specifically, broadening the channels for selling agricultural products by digitally establishing cross-domain agricultural internet platforms, agricultural apps, and green and organic agricultural products derived from the green transformation. Strengthen the backbone network of agricultural logistics and the construction of cold chain logistics systems, promote collaboration in agricultural production, supply and the storage, transport, processing, and marketing of agricultural products, and promote the establishment of “Internet+” modern agriculture.

(2) Strengthen the capacity building of independent innovation and adjust the redistribution of resources. Integrate resources across industries, and unite enterprises, universities, scientific research institutions, and other multi-party organizations to participate in the construction of innovation networks and breakthroughs in key areas and important industry technical problems. Government policies are biased and tilted, foreign investment is gradually introduced in favor of rural areas for injection, responsive supporting policies, highly qualified personnel, and advanced technology, and other factor resources gradually spread and penetrate rural areas to guide farmers to adjust the structure of agricultural production, improve the efficiency of the use of funds, and lead the majority of farmers to common prosperity.

(3) “Tailor-made” policies to implement coordinated regional development. Increasing the urban green transformation in the east and central regions can more effectively reduce the URG in the east and central regions. In contrast, the governments of the western regions should continue to implement the strategy of revitalizing the development of the west and the rise of central China, and gradually improve the environment for green transformation in the western regions. For cities on larger scales, green transformation should incorporate indicators such as green GDP into the mechanism for long-term effective assessment to avoid vicious competition caused by the government seeking an increase in gross output value and provide support for cities to narrow the URG. For cities with smaller scales, they should break the division between regions, promote win–win cooperation between regions, and facilitate the flow of resources such as crops, agricultural products, and farmers’ labor to achieve the purpose of narrowing the URG in a gradual and progressive manner.

Although this study systematically examines the spatial impact and transmission mechanisms of urban green transformation on the URG, there are still certain limitations. This paper only considers two mediating effects, namely, technological innovation and digitization level, suggesting that there are other mediating effects of urban green transformation that deserve further exploration.