Research on the Impact of Sustainable Urbanization on Urban Rural Income Disparity in China

Abstract

There is a large gap between China’s urban and rural income. The urban rural income disparity (URID) is particularly prominent. This paper deeply studies the interaction between sustainable urbanization (SU) and the URID in China by taking the construction of a new urbanization plan with Chinese characteristics suggested in China’s 14th Five-Year Plan as the research background. Firstly, the paper puts forward the hypothesis of an inverted U-shape between SU and the URID through impact path analysis. Secondly, a nonlinear panel regression model (NPRM) and a threshold regression model (TRM) are constructed to test the hypothesis. Research findings: (1) The NPRM shows that the quadratic coefficient is significant and negative, so the assumption is true, and an inverted U-shaped relationship exists. (2) The result of the TRM shows that SU passes the single threshold test, and the hypothesis of a U-shaped relationship is tenable. (3) Economic development has narrowed the URID; economic openness and road network construction have increased the URID; education has no significant impact on the URID.

1. Introduction

Since opening up to the outside world, China has achieved rapid growth in social stability, economic growth, people’s living standards, ecological protection and other aspects. However, China now faces many challenges if it wants to make breakthroughs. In the pursuit of common prosperity, China’s regional development shows a trend of polarization. Particularly, urban residents receive substantially more labor compensation than rural residents do. The income gap between urban and rural dwellers reached its highest level in recent years in China in 2007, when it was 3.14-fold greater. Even though the situation has improved recently, it is still serious [1].

Urbanization construction drives the development and upgrading of rural industries and ensure the employment of the rural labor force. Additionally, it helps residents increase their income. Moreover, the continuous influx of rural labor into cities has accelerated urbanization construction and improved the income level of urban residents [2]. Therefore, urbanization construction has a great impact on regional residents’ income. We discovered that the Chinese government has blindly pursued the rate of urbanization after sorting through the history of the growth of urbanization in China. Urbanization is only evaluated by the urbanization of land. As a result, regional development faces many economic, social and environmental problems [3]. In order to avoid slowing down or even hindering regional urbanization, the Chinese government must improve the quality of regional urbanization construction. China’s 14th Five-Year Plan report proposes that the new urbanization be planned with Chinese characteristics [4]. Therefore, this paper focuses on the relationship between SU and the URID.

Taking the construction of a new urbanization plan with Chinese characteristics proposed in China’s 14th Five-Year Plan as the research background, this paper deeply studies the relationship between SU and the URID. Firstly, this paper hypothesizes through an impact path analysis. Secondly, an indicator system for evaluating sustainable urbanization is constructed. Then, an NPRM and a TRM are constructed to test the hypothesis. Finally, according to the research conclusions and the current regional development situation, effective measures to narrow the URID are proposed.

2. Literature Review

The relationship between urbanization and the URID is a key topic to study. While sorting through the existing research results, it was found that the relationship between urbanization and the URID is mainly shown in the following three categories.

First, urbanization can narrow the URID [5]. Lang et al. found that industrialization and urbanization can increase rural residents’ income, thus reducing the URID [6]. Sulemana et al. used African countries to study the relationship between urbanization and the URID and discovered that urbanization can narrow the URID [7]. Sato et al. found that urbanization can promote the exchange of labor capital by studying the influence of urbanization on consumption and production. The full play of labor capital externalities can narrow the URID [8]. Lucas found that urbanization can improve public services, optimize industrial structure and improve agricultural labor productivity; it can narrow the URID [9]. In addition, some research results show the relationship between urbanization and narrowing the URID from human capital, population policy, and production factor perspectives [10,11].

Second, urbanization widens the URID [12]. Robinson et al. found that when more and more farmers flow into cities, the supply of urban infrastructure is short. As a result, more government funds are invested in cities and towns rather than rural areas. It increases the income of urban residents, reduces the income of rural residents, and further widens the URID [13]. Wan’s research results show that although urbanization construction attracts rural residents into urban areas, the knowledge level of this part of the labor force is relatively low and can only be engaged in basic industries. The income is only half of the income of urban residents and they cannot enjoy the welfare security of urban residents, which has widened the URID [14]. Binkai took China as the research object and found that China’s urbanization construction seriously wastes production resources, and the urban–rural and industrial layouts are unreasonable, which is not conducive to narrowing the URID [15]. Albrecht’s research found that it is difficult for rural residents to integrate into the city by promoting in-depth exchanges. The urbanization of land is not equal to the urbanization of citizens, which expands the URID [16]. His and Larson’s research found that urbanization has accelerated the rapid development of the urban real estate industry. Moreover, the price of urban commercial housing has increased too quickly. High-income groups have gained higher income by purchasing real estate, further widening the URID [17,18].

Third, the relationship between urbanization and the URID is uncertain [19]. At the beginning of urbanization, the inflow of farmers into cities is restricted by conditions such as roads and public transportation. This makes it possible for farmers with certain skills or high quality to complete urbanization, which widens the URID. With the maturity of urbanization construction, the labor value of farmers has increased and the remuneration of urban and rural labor forces has gradually become consistent, narrowing the URID. A few research results point out that the relationship between urbanization and the URID is U-shaped or inverted U-shaped. Few mathematicians use time series or panel data to study the relationship between urbanization and the URID and obtain a U-shaped relationship [20,21] or an inverted U-shaped relationship [22,23].

To sum up, the existing research results on urbanization and the URID are rich, providing a good reference for this article. However, the discussion on SU is relatively scarce. There are fewer research results on the relationship between SU and the URID. In order to narrow the URID, this paper innovatively constructs the evaluation index system of SU. Based on the analysis of the impact path, it is proposed that the relationship between SU and the URID is an inverted U-shaped relationship. The nonlinear panel regression model and threshold regression model are constructed to test such a hypothesis.

3. Impact Mechanism Analysis

3.1. Concept of Sustainable Urbanization

Simon Kuznets defined “urbanization” in 1989 as a process of urban change accompanied by economic development [24]. In other words, it refers to the changes in the distribution structure of urban and rural populations. In the past, the Chinese government unilaterally pursued the speed of urbanization. Urbanization, limited to land, has caused a series of problems. In order to avoid problems such as difficulties in economic upgrading and the intensification of social contradictions caused by the development of urbanization, regional governments need to pay attention to the quality of urbanization construction. SU has attracted more and more attention. Compared with the concept of urbanization, there are great differences in SU. The main differences are reflected in the five aspects of development goals, development concepts, and promotion subjects, as shown in

Table 1. Comparison between urbanization and Sustainable Urbanization.

	Urbanization	Sustainable Urbanization
Development objectives	Take population urbanization and land urbanization as the development goal	Take the balanced development of the economy, society, public services, and ecological protection as the goal
Development concept	Pay attention to the scale and speed of urbanization	Pay attention to the speed, quality, and urban–rural balance of urbanization
Promotion subject	Government	Government, enterprises, and individuals
Promotion mode	Emphasize the urbanization scale of a single city	Pay attention to the cooperative development of urbanization of cities of different scales
Promotion strategy	Nationwide unified urbanization construction strategy	Formulate urbanization strategies according to local conditions in different regions

[25].

3.2. Analysis of the Impact Path of Sustainable Urbanization

When the level of SU is low, the government unilaterally pursues population and land urbanization. In this process, production resources are continuously concentrated in cities and towns. In addition, the rapid growth of regional manufacturing and service industries leads to a rapid increase in the income of nearby residents. However, the urban labor force has adequate educational opportunities and professional production skills and can also invest in the service industry. Therefore, under the policy focusing on the manufacturing and service industries, urban labor remuneration increases rapidly. The rural labor force has weak professional skills and is mainly engaged in manual labor. Only a small number of rural laborers can engage in higher-income jobs. Farmers’ income is relatively small [26]. Therefore, when the level of SU is low, the URID expands. With the continuous progress of SU, production resources are gradually invested in rural infrastructure and agriculture. While accelerating urban construction, the government also needs to constantly improve rural construction to avoid widening the URID. Cities help rural areas and industry helps agriculture, effectively promoting the upgrading of rural industries and improving the quality of living standards of residents. Therefore, the URID gradually narrows [27].

To sum up, we hypothesize that there may be an inverted U-shaped relationship between SU and the URID in China.

4. Research Design

4.1. Sustainable Urbanization

4.1.1. Evaluation Index System for Sustainable Urbanization

The connotation of sustainable development was first put forward in Our Common Future, published in 1987, and the research on sustainable development has increased with each passing day. In order to measure the level of SU in China’s provinces, this paper summarizes the existing evaluation system of sustainable development [28,29]. To build the SU index system, we selected a total of 30 indicators, as shown in

Table 2. Evaluation index system of Sustainable Urbanization.

Target Layer A	Guideline Layer B	Element Layer C	Indicator Layer D	Unit	Indicator Attributes
Sustainable development level of urbanization	Resource sustainability B1	Resource Endowment C1	Energy reserves D1	million tons	Positive
			Urban water supply D2	million cubic meters	Positive
			Arable land per capita D3	Acre/person	Positive
		Utilization C2	Energy consumption D4	Tons of standard coal/RMB 10,000	Inverse
			Crude oil production D5	million tons	Positive
	Economic sustainability B2	Economic Scale C3	GDP growth rate D6	%	Positive
			GDPD per capita D7	RMB/person	Positive
		Industry Structure C4	Share of primary sector D8	%	Inverse
			Share of secondary sector D9	%	Inverse
			Share of tertiary sector D10	%	Positive
		Economic benefits C5	Fiscal revenue D11	billion	Positive
			Social Labor Productivity D12	million Yuan/person	Positive
			Industrial output tax rate D13	%	Positive
		Innovation driven C6	Proportion of scientific and technical staff D14	%	Positive
			R&D expenditure D15	%	Positive
	Social sustainability B3	People’s quality of living standards C7	Wages per employee D16	Yuan	Positive
			Road area per capita D17	m2/person	Positive
		Social service functions C8	Number of doctors D18	person	Positive
			Investment in education D19	million	Positive
			Bus occupancy per 10,000 people D20	units per 10,000 people	Positive
			Daily water consumption per capita D21	liter	Positive
	Ecosystem sustainability B4	Environmental pollution C9	Industrial wastewater discharge D22	million tons	Inverse
			Industrial waste gas emissions D23	tons	Inverse
		Environmental Governance C10	Industrial wastewater treatment rate D24	%	Positive
			Industrial waste gas treatment rate D25	%	Positive
			Domestic wastewater treatment rate D26	%	Positive
			Integrated industrial solid waste rate D27	%	Positive
		Ecological construction C11	Urban green coverage D28	%	Positive
			Green space per capita D29	sqm/person	Positive
			Proportion of environmental protection investment in GDP D30	%	Positive

.

4.1.2. Evaluation Method for Sustainable Urbanization

Different measurement methods have different emphases. Generally speaking, they can be divided into subjective empowerment and objective empowerment. Objective empowerment is more easily recognized by the academic community [30].

First is the standardization of indicator data. As the evaluation index system involves 30 indicators, and different indicators involving different orders of magnitude and units of measurement cannot be directly compared, the original data of each indicator must be dimensionless.

The dimensionless formula for the positive indicator is

$${\mathrm{x}}_{\mathrm{ij}}^{}=\frac{{\mathrm{x}}_{\mathrm{ij}}-\min \left({\mathrm{x}}_{\mathrm{j}}\right)}{\max \left({\mathrm{x}}_{\mathrm{j}}\right)-\min \left({\mathrm{x}}_{\mathrm{j}}\right)}$$

(1)

The dimensionless formula for the negative indicator is

$${\mathrm{x}}_{\mathrm{ij}}^{}=\frac{\max \left({\mathrm{x}}_{\mathrm{j}}\right)-{\mathrm{x}}_{\mathrm{ij}}}{\max \left({\mathrm{x}}_{\mathrm{j}}\right)-\min \left({\mathrm{x}}_{\mathrm{j}}\right)}$$

(2)

Second, we measure the weight of year i indicator values for indicator j

$${\mathrm{g}}_{\mathrm{ij}}=\frac{{\mathrm{x}}_{\mathrm{ij}}^{}}{{{\displaystyle \sum }}_{\mathrm{i}=1}^n{\mathrm{x}}_{\mathrm{ij}}^{}}$$

(3)

Third, we measure the information entropy value and information utility value

$${\mathrm{e}}_{\mathrm{j}}=-\mathrm{k}{{\displaystyle \sum }}_{\mathrm{i}=1}^n{\mathrm{g}}_{\mathrm{ij}}{\mathrm{lng}}_{\mathrm{ij}} \mathrm{k}=\frac{1}{\ln n} {\mathrm{d}}_{\mathrm{j}}=1-{\mathrm{e}}_{\mathrm{j}}$$

(4)

Fourth, we measure the weight of the j-th indicator

$${\mathrm{v}}_{\mathrm{j}}=\frac{{\mathrm{d}}_{\mathrm{j}}}{{{\displaystyle \sum }}_{\mathrm{j}=1}^n{\mathrm{d}}_{\mathrm{j}}}$$

(5)

Fifth is the measurement of the combined level

$${\mathrm{W}}_{\mathrm{j}}={\displaystyle \sum }_{\mathrm{j}=1}^n{\mathrm{v}}_{\mathrm{j}}{\mathrm{x}}_{\mathrm{ij}}^{}$$

(6)

4.2. Measurement Model Construction

In order to test the hypothesis, this paper first constructs a nonlinear panel regression model based on the panel data of China. Secondly, Hansen’s method is used to establish the threshold regression model of SU and the URID [31].

4.2.1. Index System of Measurement Model

Based on the analysis of the existing research results of urbanization and the URID, as well as the impact mechanism of SU on the URID, this paper selects six variables [32,33], as shown in

Table 3. Index system of measurement model.

Variable		Specific Description	Symbol
Interpreted variable	Urban rural income disparity	The ratio of urban residents’ income to rural residents’ income	URID
Core explanatory variable	Sustainable urbanization level	Evaluation results of sustainable urbanization	SU
Control variables	Economic development level	GDP per capita	PGDP
	Educational level	Per capita years of education	EDU
	Economic openness level	Proportion of total imports and exports in GDP	PTIE
	Road network construction level	Road area per 10,000 people	PRA

.

(1)

Interpreted variable: The ratio of urban residents’ income to rural residents’ income is used to express the URID.

(2)

Core explanatory variable: The evaluation results of sustainable urbanization are used to reflect the level of SU in various provinces in China.

(3)

Control variables: (1) We select GDP per capita to reflect the economic level. Based on the Kuznets hypothesis, the URID will show a nonlinear relationship with the economy [34]. (2) The per capita education years are selected to reflect the regional education level. The opportunities and quality of the urban population to receive education are far higher than a husbandman, and education is the main way to improve workers’ production skills, which has a great impact on residents’ income [35]. Therefore, this indicator expands the URID. (3) The ratio of total imports and exports to GDP is selected to represent the degree of regional economic openness. The higher openness of the economy and the increase in export trade and foreign direct investment brings more employment and investment opportunities. However, export trade and foreign direct investment play greater roles in urban areas and have less of an impact on rural areas [36]. Therefore, this indicator expands the URID.

(4)

We select the per capita road area to reflect the construction level of the regional road network. A convenient road network is an important way to connect cities, towns and rural areas. It promotes not only urban economic growth but also drives rural development. Developed road network construction promotes the development of rural industries and provides convenience to farmers to work in cities [37]. Therefore, this indicator narrows the URID.

4.2.2. Nonlinear Panel Regression Model

To test the nonlinear relationship between $SU$ and the $URID$, we introduce the quadratic term of SU [38] and consider other variables that affect the $URID$, adding control variables such as $PGDP$, $EDU$, $PTIE$ and $PRA$ to establish a nonlinear regression model, as shown in Formula (7). i and t represent the i-th province and the t-th year. ${\alpha }_{\mathrm{i}}$ represents the intercept. ${\mu }_{it}$ represents the error value. ${\epsilon }_{\mathrm{it}}$ represents a regional effect. ${\delta }_{it}$ represents a time effect. $URI{D}_{it}$ represents the $URID$ in province i in year t. $S{U}_{it}$ represents the level of $SU$ in province i in year t. $PGD{P}_{it}$, $ED{U}_{it}$, $PTI{E}_{it}$ and $PR{A}_{it}$, respectively, represent per capita $GDP$, per capita education years, the proportion of total imports and exports to $GDP$, and per capita road area. All variables are logarithmically processed to eliminate the influence of absolute value difference.

$$lnURI{D}_{it}={\alpha }_{\mathrm{i}}+{\beta }_{1}lnS{U}_{it}+{\beta }_{2}lnS{U}_{it}^2+{\beta }_{3}lnPGD{P}_{it}+{\beta }_{4}lnED{U}_{it}+{\beta }_{5}lnPTI{E}_{it}+{\beta }_{6}lnPR{A}_{it}+{\mu }_{it}+{\epsilon }_{it}+{\delta }_{it}$$

(7)

4.2.3. Threshold Regression Model

To confirm that there is a nonlinear relationship between SU and the URID, we analyze the influence of SU in different development stages on the URID. We take SU as the threshold variable and construct a threshold regression model, as shown in Formulas (8) and (9) [39]. The SU in the formula refers to the threshold variable and η refers to the threshold.

$$URI{D}_{it}={\alpha }_{\mathrm{i}}+{\beta }_{1}S{U}_{it}+{\beta }_{2}PGD{P}_{it}+{\beta }_{3}ED{U}_{it}+{\beta }_{4}PTI{E}_{it}+{\beta }_{5}PR{A}_{it}+{\mu }_{it}+{\epsilon }_{\mathrm{it}}+{\delta }_{\mathrm{it}} \left(SU\le \eta \right)$$

(8)

$$URI{D}_{it}={\alpha }_{\mathrm{i}}+{\beta }_{1}S{U}_{it}+{\beta }_{2}PGD{P}_{it}+{\beta }_{3}ED{U}_{it}+{\beta }_{4}PTI{E}_{it}+{\beta }_{5}PR{A}_{it}+{\mu }_{it}+{\epsilon }_{\mathrm{it}}+{\delta }_{\mathrm{it}} (SU>\eta )$$

(9)

4.3. Data Interpretation

According to the model built above to study the relationship between SU and the URID, we obtain the values of various indicators in the model from the China Statistical Yearbook, China Rural Statistical Yearbook and Wande Database. For missing data, interpolation and moving average methods are used to fill the slots. The variables involved in price factors are calculated at constant prices in 2010.

5. Results and Discussion

5.1. Descriptive Statistics of Variables

This paper makes a descriptive statistical analysis of the relevant data, and the results are shown in

Table 4. Descriptive statistics results.

Variable	Mean	Std. Dev.	Maximum	Minimum	Kurtosis	Skewness	JB [p.]
URID	2.87	0.62	5.61	1.85	4.87	1.21	254.6 [5.3 × 10−56]
SU	0.5	0.16	0.9	0.19	3.10	0.58	36.23 [1.40 × 10−8]
PGDP	3.31	2.72	16.4	0.25	5.50	1.45	396.8 [6.70 × 10−87]
EDU	8.41	1.31	12.95	2.95	5.17	−0.39	144 [5.40 × 10−32]
PTIE	0.27	0.32	1.49	0.01	5.92	1.93	633.9 [2.00 × 10−138]
PRA	7.3	1.83	13.1	3.27	3.10	0.54	31.46 [1.50 × 10−7]

. We also test the normality of the selected variables through the JB test method. The six variables selected reject the original hypothesis and do not obey the normal distribution.

5.2. Unit Root Test

We use LLC, ADF, and IPS to test the unit root of the variables selected. In

Table 5. Unit root test results.

Variable	LLC Test	ADF Test	IPS Test	Conclusion
URID	−4.747 ***	13.148 ***	−3.427 ***	Stability
SU	−7.127 ***	7.279 ***	−2.949 ***	Stability
PGDP	−10.771 ***	2.769 ***	−4.142 ***	Stability
EDU	−9.752 ***	11.182 ***	−8.638 ***	Stability
PTIE	−4.630 ***	9.189 ***	−3.123 ***	Stability
PRA	−8.248 ***	5.342 ***	−4.594 ***	Stability

*** p < 0.01.

, we can see that the six variables selected pass the unit root test, showing that the original series data are stable and can be used for subsequent regression analysis.

5.3. Results and Discussion of the Nonlinear Panel Regression Model

To test this study’s hypothesis, we first carry out the Hausman test on the model, and the results are shown in

Table 6. Hausman test results.

Test Summary	Chi-Sq. Statistic	Prob > chi2
Cross-section random	72.81	0.000

[40]. The result shows that the chi-squared value is 72.81. The corresponding probability value is 0.000. Therefore, we choose the fixed-effect model.

As shown in

Table 7. Fixed-effect regression results.

Variable	Coefficient
$lnSU$	4.351 **
$lnS{U}^2$	−3.182 ***
$lnPGDP$	−6.43 × 10−6 ***
$lnEDU$	−0.221
$lnPTIE$	0.514 ***
$lnPRA$	−3.221 ***
R2	0.902
Prob > F	0.000

*** p < 0.01, ** p < 0.05.

, the regression results show that the model fits the sample data well. R² is 0.902. The corresponding p values of $lnSU$, $lnS{U}^2$, $lnPGDP$, $lnPTIE$, and $lnPRA$ are all less than 0.05. These variables have a significant impact on the URID. The p value of $lnEDU$ is greater than 0.1, so this variable has no significant impact on the URID at the 10% significance level.

The coefficient of $lnSU$ on $lnURID$ is positive, with a value of 4.351. It shows that by improving SU, the URID also increases. The coefficient of $lnS{U}^2$ on $lnURID$ is negative, with a value of −3.182. The results show that the impact of SU on the URID is not linear but inverted U-shaped. When the level of SU is low, the government unilaterally pursues population and land urbanization. Production resources are continuously concentrated in cities and towns, and industries and services are vigorously developed. The income of urban labor increases rapidly. The rural labor force has weak professional skills, mainly engaged in manual labor, with relatively small income growth. As a result, the URID gradually widens. With the continuous improvement of the level of SU, production resources are gradually invested in rural infrastructure and agriculture. Because cities help rural areas and industry helps agriculture, it effectively promotes rural industries. Therefore, the income of farmers gradually increases. The URID gradually narrows. Further calculation shows that when the value of SU is greater than 0.6836895, the relationship between SU and the URID begins to turn. That is, when the value of SU is greater than 0.6836895, the URID decreases with the increase in SU. This paper analyzes the spatial and temporal differences of 31 provinces in China and finds that by 2020, 6 of the 31 provinces in China had SU values greater than 0.6836895. They are Shanghai, Beijing, Tianjin, Guangdong, Jiangsu and Zhejiang. According to the social and economic development status of the six provinces and cities, it is found that the six provinces and cities are all regions with rapid economic development in China. The construction of urbanization in these regions has been accelerated, and the primary stage of urbanization has been previously completed. However, with the blind construction of early urbanization, a series of economic and social problems has gradually emerged. The six provinces and cities are focusing on SU. Compared with before, SU was greater than 0.6836895, and the URIDs of the six provinces and cities were reduced. Take Jiangsu Province as an example. In 2000, the SU of Jiangsu Province was 0.36, and the URID was 1.87. With the growth of SU, the URID also expands. By 2015, the SU of Jiangsu Province increased to 0.685, and the URID reached the highest value of 2.57. Then, the URID of Jiangsu Province continued to shrink with the growth of SU. By 2020, the SU of Jiangsu Province increased to 0.706, and the URID shrunk to 2.25.

Among the control variables, the $lnPGDP$ influence coefficient is significantly negative. $lnPGDP$ has a significant inhibitory effect on the URID, but the impact is small. Each unit of increase in $lnPGDP$ reduces the URID by 0.00000643 units. With the continuous improvement of the regional economic level, the public security measures provided by the government are also improving, and the support for low-income people also increases to narrow the URID. $lnPTIE$ influence coefficient is significantly positive. For each unit increase in $lnPTIE$, the URID increases by 0.514 units. It shows that $lnPTIE$ significantly promotes the URID in China’s provinces. A higher the degree of economic openness, i.e., an increase in export trade and foreign direct investment, brings more employment and investment opportunities to cities and towns. However, the impact on rural areas is small, so the URID widens. The influence coefficient of $lnPRA$ on the URID passes the significance test and is negative. For each unit increase in $lnPRA$, the URID decreases by 3.221 units, showing that $lnPRA$ has a significant inhibitory effect on the URID. Convenient road construction is a bridge linking cities and rural areas. It promotes not only urban economic growth but also drives rural development to narrow the URID.

5.4. Results and Discussion of Threshold Regression Model

To test the hypothesis, this paper first tests the model when SU has a single threshold, double threshold, and triple threshold. The F statistic in the threshold test and the p value obtained by the Bootstrap method are shown in

Table 8. Threshold effect test (bootstrap = 10,000, 10,000, 10,000).

Variable	Threshold	Fstat	Prob.
SU	Single	210.74	0.000
	Double	−48.38	1.000
	Triple	14.16	0.683

. The threshold variable SU passes the single threshold test. The corresponding threshold results are shown in

Table 9. Threshold estimator (level = 95).

Variable	Single Threshold Value	Lower	Upper
SU	0.2090	0.2050	0.2110

. The estimated threshold is 0.209. Therefore, SU has a significant threshold effect on the URID.

The threshold regression results are shown in

Table 10. Threshold regression results.

Variable	Coefficient
$PGDP$	−0.013 **
$EDU$	0.065
$PTIE$	0.341 ***
$PRA$	−0.030 ***
$SU (SU\le 0.209$)	6.685 ***
$SU (SU>0.209)$	−2.577 ***
Cons	3.319 ***
R-squared	0.897
Prob > F	F = 0.000

*** p < 0.01, ** p < 0.05.

. When SU is less than 0.209, the impact coefficient is 6.685, which passes the test at a 1% significance level. It shows that SU has a significant promoting effect on the URID. When SU is greater than 0.209, the influence coefficient is −2.577, showing that SU has a significant inhibitory effect on the URID. This result is consistent with the fixed-effect regression model. It further verifies that the relationship between SU and the URID is an inverted U-shape. When the level of SU is lower than the threshold, the government unilaterally pursues population and land urbanization. All kinds of production resources flow into cities and towns. The government vigorously develops manufacturing and service industries, and the income of urban residents increases rapidly; thus, the URID increases. When the level of SU is higher than the threshold, the regional government gradually promotes the common promotion of towns and villages and reduces the gap between the two. The government pays attention to developing modern agriculture, improving rural public facilities and raising farmers’ income, leading to the URID gradually decreasing.

6. Conclusions and Suggestions

The main conclusions of the paper are as follows: (1) The fixed-effect regression model shows that the quadratic coefficient is significantly negative, and the relationship between SU and the URID has an inverted U-shape. (2) The threshold regression model shows that SU passes the single threshold test, and the influence coefficient is positive and then negative. It also proves that the assumptions about SU and the URID are valid. (3) Economic development narrows the URID. Economic openness and road network construction expand the URID. Education has no significant impact on the URID. (4) Promoting SU is an important way to narrow the URID in China.

China should promote SU and take advantage of the relationship between SU and the URID. China must strengthen the synchronous growth of urban and rural areas, comprehensively increase farmers’ income, and narrow the URID. Based on the previous research, we propose the following four suggestions to reduce the URID: (1) Promote SU and build an urban–rural cooperative relationship in line with regional development. Some provinces in China are still at the beginning of urbanization, and the level of SU is low. In the western provinces of China, extensive urbanization has aggravated the URID. Therefore, the provinces must adjust their urbanization strategy in time and replace the priority development strategy of cities with the mutual promotion and common development strategy of cities and rural areas. The region should implement the strategy of industry-supporting agriculture. (2) Promote the transformation of industrial structure and improve rural human capital and further study the improvement of farmers’ income through regional and service industries. Vigorously promote the construction mode of agricultural intensification and cooperation. The region should focus on building a complete industrial chain, improving the product value chain, and expanding the income chain to accelerate the realization of regional agricultural modernization. The government needs to continuously increase farmers’ wage income and operating income. Improve the quality of farmers’ production, increase skill training, and create more employment opportunities for rural residents. (3) Regional governments should improve the registered residence system and rural infrastructure. The government needs to speed up the improvement of the registered residence mechanism and break the restrictions of the registered residence system on rural labor mobility. The government should guide the rational distribution of production resources and implement the differentiated registered residence policy. A more convenient environment for the market-oriented flow of the labor force needs to be conducted to improve farmers’ income. The government and enterprises must speed up the improvement of rural infrastructure. The government should also promote the comprehensive coverage of rural medical, education and other basic services so that more farmers can enjoy high-quality medical care, education, old-age care and other services. This will reduce the impact of unstable factors such as rural residents’ living standards, education, old-age care, medical care, and unemployment. (4) We should strengthen vocational education and skills training for farmers and promote their stable employment. A resource base for farmers’ vocational skills training needs to be established. Integrate all kinds of vocational skills training resources in the region and share training resources nationwide to meet the training needs of migrant workers at different levels. Secondly, the government should increase its investment in the training of farmers. Increase the investment in training funds to ensure the training costs for skilled personnel required by key industries are in place. Establish a training base for flexible and diversified order-based and orientation training for migrant workers according to market demand.

This paper focuses on whether there is an inverted U-shaped relationship between SU and the URID. However, the selection of control variables in this empirical analysis model is based on the research results of other scholars. We do not analyze the influencing factors of the URID. In the next stage, we will continue to study the relationship between SU on the URID in combination with the research on the factors affecting the URID. Moreover, we will further study the nonlinear relationship between SU and DI by using Markov switching [41], smooth transition regression [42], neural networks [43], and the fully modified ordinary least squares method [44].