Dynamic Analysis of Regional Integration Development: Comprehensive Evaluation, Evolutionary Trend, and Driving Factors

Abstract

Regional integration, as an essential measurement for solving unbalanced and uncoordinated regional development, plays an important role in achieving regional sustainable development. In this study, we aimed to construct a systematic research framework to facilitate the development of regional integration. Using 31 prefecture-level cities in the urban agglomeration in the Middle Reaches of the Yangtze River (MYR-UA) as case sites, this study applied box plots, kernel density estimation, GIS visualization tools, Markov chains, and geographic detectors to reveal the dynamic spatiotemporal evolution and factors influencing regional integration from 2009 to 2018. The results indicate that the level of regional integration and its subsystem development is suboptimal in MYR-UA; the temporal characteristic shows an upward fluctuating trend, and the spatial distribution shows remarkable spatial correlation and clustering characteristics. Additionally, we found that the level of regional integration development in MYR-UA has both “path dependence” and “self-locking” effects, and the spatial lag type has a crucial impact on the degree of regional transfer stability. The dominant factors affecting regional integration development include the GDP per capita, economic openness, industrial structure, proportion of education expenditure in fiscal expenditure, urbanization rate, proportion of environmental investment in fiscal expenditure, population density, capital flow, information flow, and technology flow. Finally, based on the findings of this study, policy recommendations for promoting regional integration are proposed.

1. Introduction

Regional integration development is a long, complex, and hierarchical process, investigated extensively since the work of A. Smith [1]. Early regional integration refers to regional economic integration, mainly concentrated in the field of transnational regional economic integration [2,3]. With the current trends, regional integration has gradually developed a wider scope and deeper level of application, limited to economic integration and extending to ecological integration, spatial integration, education integration, infrastructure integration, and industrial integration [4,5,6,7,8,9]. The purpose of regional integration is to remove barriers between regions and solve the problem of unbalanced regional development [10].

Regional integration meets the endogenous needs of China’s development, and the government regards it as an essential development strategy [11]. The Chinese government has undertaken long-term explorations and efforts to promote regional integration. Since China’s reform and opening up, the eastern region has taken the lead in development, relying on its superior geographical location and policy advantages. The Pearl River Delta, Yangtze River Delta, and Beijing–Tianjin–Hebei have become the most developed regions in China in all aspects. Simultaneously, regional development disparities have gradually emerged [12]. Therefore, the Chinese government successively implemented regional development strategies, such as the Development of Western China, the Rise of Central China, and the Revitalization of Northeast China, to promote economic growth in lagging regions and achieve coordinated regional development [13]. With the continuous promotion of coordinated regional development policies, the degree of regional development imbalance in China has significantly reduced. Especially after the completion of poverty eradication, China’s regional development gap has further narrowed [14].

However, the problem of unbalanced and uncoordinated regional development in China has been long-standing. Despite numerous efforts by the Chinese government, such as implementing poverty alleviation and regional coordinated development strategies, as well as promoting new urbanization centered on people, the unbalanced development between provinces and cities has not been eradicated in a short period [15]. Maintaining existing achievements, further promoting the process of regional integration, and truly achieving coordinated regional development is the challenge we need to solve at this stage and in the longer term in the future.

Since the regional integration strategy was proposed, it has received attention from various disciplines, such as urban-rural planning, human geography, and economics. At the empirical level, previous studies have focused on measuring integration levels from a particular perspective. Liao, Abdullahi and Duan et al. measured the integration level from the perspectives of green innovation efficiency, economic efficiency, technology transfer, and factor flow [16,17,18,19]. At the theoretical level, scholars have researched the development characteristics, influencing factors, and realization paths of regional integration [20,21,22]. Most of these studies revolve around national, provincial, and city cluster scales, but there are few studies on integration development at the municipal scale [16,18,22]. In addition, the existing literature mostly evaluates the integration development process from a single dimension [17], which cannot reflect the comprehensiveness of the integration system. Moreover, these studies often treat research objects as independent individuals, ignoring the correlation between different fields.

This study aimed to construct a systematic research framework for regional integration development to complement the existing literature. Based on a comprehensive consideration of theory and practice, we constructed an evaluation system for the integration process in five dimensions: the economy, public services, urban–rural development, ecology, and space. We took 31 prefecture-level cities in MYR-UA as case sites and conducted empirical tests using geographic and econometric methods (Figure 1). This study has three main contributions: (1) Constructing an innovative indicator evaluation system for measuring the process of regional integration, and conducting empirical tests using MYR-UA as a case study, proving that the evaluation model has strong scientific and practical applicability. (2) Not only did we summarize the dynamic evolution characteristics of the overall integration process of MYR-UA, but we also explored the spatiotemporal evolution laws of the subsystems of the integration process, which previous research has not achieved. (3) The geographic detector used in this study is more effective than the regression models used in previous studies on influencing factors. It is a new method in the field of urban geography to explore influencing factors, which can effectively identify key factors that affect the process of regional integration, and thus provide more targeted suggestions for promoting the development of regional integration.

The remainder of this paper is organized as follows. Section 2 presents the literature review. Section 3 introduces and describes this study’s data and methodology. Section 4 presents the results of the study. Section 5 discusses the results of this empirical study. Finally, Section 6 concludes this study and provides policy implications.

2. Literature Review

The concept of regional integration was first introduced by the Dutch economist Tinbergen in his book “International Economic Integration”, where he pointed out that regional integration is the process of creating an optimal economic structure by weakening and eliminating factors that hinder the effective functioning of economic activities through mutual collaboration and unification between countries or regions [23]. In 1962, American economist Balassa expanded Tinbergen’s definition. He defined regional integration as the process of eliminating differential treatment between countries or regions within a region through various measures [24]. Subsequently, geography, planning, and a political economy created different definitions of regional integration. The connotation of integration has also broken through the narrow sense of economic integration and has gradually expanded to include infrastructure, ecology, institutions, politics, and culture [4,5,6,7,8,9,25]. For example, Li et al. argued that regional integration should focus on economic aspects and the coordinated development of regional ecology [6]. According to Fang and Zhang, regional integration is the coordinated and unified development of the economy, society, culture, ecology, governance, and space among cities in geographically adjacent regions, relying on developed transportation infrastructure and communication networks [8]. Tang et al. emphasized the complexity and multidimensionality of regional integration and pointed out that it contains dimensions such as economic integration, market integration, ecological integration, industrial integration, and social integration [25].

With the deepening of regional integration research, the evaluation method of regional integration development has become mainstream academic research. To scientifically measure the level of regional integration, the construction of an evaluation system and selection of a measurement model are crucial. On the one hand, scholars have conducted quantitative data analysis by constructing evaluation models to explore the development status of regional integration. Zhong et al. constructed a regional integration indicator system from the perspective of the economy, society, and environment and evaluated the regional integration level in the Yangtze River Delta from 2000 to 2015 [26]. Alemayehu and Haile applied the gravity model to measure the level of economic integration in the African region [27]. Based on accessibility analysis, Guo et al. discussed the potential for urban integration development in the northeast region using city and road network data [28]. Qian et al. constructed an evaluation index system of the urban–rural integration development level from the perspective of factor flow and evaluated the urban–rural integration development of 31 Chinese provinces using a joint weighted model of GI and CRITIC [29]. On the other hand, qualitative research methods have also been used to explore the evaluation system of regional integration. For example, Abbas adopted a textual analysis method to construct an evaluation system for regional integration [30]. Song used the Delphi method to build a performance evaluation system for coordinated regional development and took the Hefei Metropolitan Area as a case to conduct an empirical analysis [31].

Based on the evaluation of regional integration, the academic community focuses on exploring the factors influencing regional integration to promote its development of regional integration better [31,32,33,34,35,36,37]. Song stated that the degree of information and capital flows within a region have an essential impact on the level of integration development [31]. Zhang and Sun indicated that regional policies have a strong influence on the integration process. Governments should formulate effective policies to facilitate the regional integration process [32]. Xu et al. confirmed that the opening of high-speed rail has a positive effect on the development of regional integration [33]. Yang et al. used structural equation modeling and regression modeling to verify the role of science and technology innovation in influencing the regional integration process [34]. Ruan and Zhang demonstrated a better effect of information flow on strengthening regional economic linkages using fuzzy-set qualitative comparative analysis (fsQCA) [35].

The literature review shows that the relevant research results on regional integration are relatively fruitful and can provide references for subsequent research; however, problems worthy of further exploration persist. First, previous studies mostly evaluated the level of regional integration development from the perspective of a single dimension, which cannot reflect its multidimensional attributes. As a comprehensive variable system, the evaluation and measurement of regional integration and development must be considered comprehensively and systematically. Second, current studies mainly focus on the spatial differentiation characteristics of regional integration development and lack a method for the long-term monitoring and identification of its dynamic transfer law. Third, most studies have focused on overall regional integration development, and few studies have explored its subsystems. Finally, research on the factors influencing regional integration development is mainly based on qualitative rather than quantitative methods. In particular, there is a lack of comparison and dynamic analysis of the influencing factors in different years.

This study constructs a systematic evaluation system based on existing studies on regional integration and the actual situation of case sites. The integration development levels of 31 prefecture-level cities in MYR-UA from 2009 to 2018 are then measured. The spatiotemporal characteristics of the integration development level are demonstrated with the help of box plots, kernel density curves, GIS visualization tools, and Markov chains. Finally, the factors influencing the integration development level are comprehensively detected with the help of geographic detectors.

3. Materials and Methods

3.1. Study Area

MYR-UA is an important part of the Yangtze River Economic Belt, and it mainly comprises the Wuhan Metropolitan Area (WMA), Chang-Zhu-Tan Urban Agglomeration (CZT-UA), and Poyang Lake City Group (PLCG) (Figure 2). MYR-UA has a land area of approximately 326,100 km² and a total population of 125 million. In 2018, its regional GDP exceeded CNY 8 trillion, and its GDP per capita reached CNY 104,800. It is a large area with a dense population and robust transportation and economic activities. Under the powerful support of national policies, MYR-UA has become a key region for implementing the strategy named the Rise of Central China, which involves deepening reform and opening up in an all-around way and promoting new urbanization. In recent years, with the rapid development of MYR-UA, its immature integration development mechanism, the weak radiation-driving ability of its central city, and unreasonable structure and spatial layout of its industries have aroused public concern. Moreover, the problems of environmental pollution and the unbalanced development of urban-rural areas need to be addressed. Therefore, this realistic context is of great theoretical significance for providing practical guidance to promote the high-quality development of each city in MYR-UA to explore the spatiotemporal evolution characteristics and influencing factors of regional integration.

3.2. Data Sources

Based on the research objectives of this study, data from 31 prefecture-level cities in MYR-UA from 2009 to 2018 were collected. The data were obtained from the China City Statistical Yearbook, China Statistical Yearbook for Regional Economy, China Environmental Statistical Yearbook, China Urban-Rural Construction Statistical Yearbook, Hunan Statistical Yearbook, Hubei Statistical Yearbook, Jiangxi Statistical Yearbook, and the statistical yearbooks of various cities, spanning the period from 2010 to 2019. Some data were obtained from the Statistical Bulletin of National Economic and Social Development of various cities for the years 2009–2018. Missing data were filled in using a linear interpolation method. In addition, to eliminate the influence of price factors, GDP per capita and other indicators related to price factors in the above index system were uniformly deflated according to the constant price in 2009. For indicators in foreign currency, we converted the currency to RMB units according to the RMB exchange rate of the corresponding year. The administrative boundary vector map used in this study was obtained from the National Basic Geographic Information Center.

3.3. Research Methodology

3.3.1. Index System Construction

The core concept of regional integration development is that the structural relationship, combination state, and spatial distribution of various production factors, such as population, economy, infrastructure, ecological environment, and public services, within a specific geographical area reach their optimal configurations at different spatial scales [16,17,18,32]. The change in the regional integration pattern is a vertical and multidimensional development process that moves from “shallow to deep” [33]. According to the principles of science, objectivity, comprehensiveness, and feasibility, and based on the existing research on regional integration development [16,17,18,23,24,25,26,27], this study constructed an indicator system that can reflect the level of integration development from five dimensions (Table 1).

Economic integration forms the basis of regional integration development [3]. Referring to Zhao and Zhang’s study, we selected GDP per capita, economic openness, industrial structure, and economic development deviation to measure the level of economic integration [38,39]. GDP per capita and economic openness represent the level of regional economic development, while industrial structure and economic development deviation represent the regional industrial layout and industrial labor productivity level, respectively.

Public service integration is essential for regional integration development [40]. The purpose of public service integration is to continuously optimize the supply of public services, improve the co-construction and sharing mechanism of public services, and promote the sharing of educational resources, cross-regional medical and health services, the joint development of cultural industries, and the interconnection of social security [41,42,43]. Therefore, the proportion of education expenditure, transportation construction expenditure, medical and health expenditure, and social security expenditure in fiscal expenditure were selected to assess the public service integration subsystem.

Urban–rural integration is a fundamental goal of regional integration development [44]. Urban-rural integration is a process of narrowing the gap between urban and rural areas, and involves the urban–rural population, residents’ consumption, residents’ income, and other dynamic spatial balances [5,45]. In the urban–rural integration subsystem, the urbanization rate, the proportion of per capita income of urban and rural residents, and the Engel coefficient proportion of urban and rural residents were selected for measurement. The urbanization rate represents the level of regional urbanization. The proportion of the per capita income of urban and rural residents represents the income gap between them. The Engel coefficient proportion of urban to rural residents represents the difference in living standards between urban and rural residents.

Ecological integration is an important mechanism in regional integration and development [6]. Adhering to green development and building an ecological civilization will contribute to the sustainable development of socially productive forces and improve people’s quality of life [46]. To realize ecological integration, it is necessary to control regional environmental pollution, improve resource utilization, and increase investment in environmental construction [47,48,49]. In this study, industrial wastewater discharge, energy consumption per unit of GDP, industrial SO₂ emissions, the comprehensive utilization rate of industrial solid waste, the investment proportion of environmental pollution treatment, and green coverage in constructed areas were the factors selected to study the ecological integration subsystem.

Spatial integration is an essential component of regional integration development [50]. Spatial integration is a deep-seated manifestation of regional factor agglomeration, and population, capital, information, and technology are the most influential factor flows [51,52,53]. Therefore, the population density, capital flow, information flow, and technology flow were selected to study the spatial integration subsystem.

Table 1. Description of relevant variables.

Objectives	Criterion Layer	Index Layer	Unit	Index Type	Literature Source	Weight
Level of integration development	Economic integration	GDP per capita	yuan	+	Zhao and Zhou [38,39]	0.085
		Economic openness	%	+		0.161
		Industrial structure	%	+		0.023
		Economic development deviation	——	−		0.003
	Public service integration	Proportion of education expenditure in fiscal expenditure	%	+	Feng, Alessia, and Yu [41,42,43]	0.015
		Proportion of transportation construction expenditure in fiscal expenditure	%	+		0.094
		Proportion of medical and health expenditure in fiscal expenditure	%	+		0.044
		Proportion of social security expenditure in fiscal expenditure	%	+		0.029
	Urban-rural integration	Urbanization rate	%	+	Yang and Ma [5,45]	0.016
		Proportion of per capita income of urban and rural resident	%	−		0.007
		Engel coefficient Proportion of urban and rural residents	%	−		0.004
	Ecological integration	Industrial wastewater discharge	tons	−	Pan, Xu, and Hu [47,48,49]	0.014
		Energy consumption per unit of GDP	tons/yuan	−		0.010
		Industrial SO2 emissions	tons	−		0.025
		Comprehensive utilization rate of industrial solid waste	%	+		0.020
		Proportion of environmental investment in fiscal expenditure	%	+		0.080
		Green coverage in constructed areas	%	+		0.023
	Spatial integration	Population density	km2/person	+	Fajle, Liu, and Dai [51,52,53]	0.099
		Capital flow	%	+		0.071
		Information flow	%	+		0.076
		Technology flow	%	+		0.101

Table 1. Description of relevant variables.

Objectives	Criterion Layer	Index Layer	Unit	Index Type	Literature Source	Weight
Level of integration development	Economic integration	GDP per capita	yuan	+	Zhao and Zhou [38,39]	0.085
		Economic openness	%	+		0.161
		Industrial structure	%	+		0.023
		Economic development deviation	——	−		0.003
	Public service integration	Proportion of education expenditure in fiscal expenditure	%	+	Feng, Alessia, and Yu [41,42,43]	0.015
		Proportion of transportation construction expenditure in fiscal expenditure	%	+		0.094
		Proportion of medical and health expenditure in fiscal expenditure	%	+		0.044
		Proportion of social security expenditure in fiscal expenditure	%	+		0.029
	Urban-rural integration	Urbanization rate	%	+	Yang and Ma [5,45]	0.016
		Proportion of per capita income of urban and rural resident	%	−		0.007
		Engel coefficient Proportion of urban and rural residents	%	−		0.004
	Ecological integration	Industrial wastewater discharge	tons	−	Pan, Xu, and Hu [47,48,49]	0.014
		Energy consumption per unit of GDP	tons/yuan	−		0.010
		Industrial SO2 emissions	tons	−		0.025
		Comprehensive utilization rate of industrial solid waste	%	+		0.020
		Proportion of environmental investment in fiscal expenditure	%	+		0.080
		Green coverage in constructed areas	%	+		0.023
	Spatial integration	Population density	km2/person	+	Fajle, Liu, and Dai [51,52,53]	0.099
		Capital flow	%	+		0.071
		Information flow	%	+		0.076
		Technology flow	%	+		0.101

3.3.2. Entropy Value Method

As an objective assignment method, the entropy method is often used in comprehensive evaluation studies of multiple indicators [54]. The principle of the entropy method is to determine the indicator weight according to the size of the information entropy of the indicators, and the greater the weight, the greater the influence on the evaluation system. The advantage of the entropy weighting method is that it can eliminate the interference of subjective factors and render the measurement results more scientific and reasonable [55]. In this study, the entropy value method was used to measure the weight of each index. The specific calculation steps were as follows.

First, owing to the different measurement units and attributes of each indicator, the range method was used to standardize the original data. The formula is as follows:

$$\mathrm{Positive} \mathrm{indicator}: {x^{\prime }}_{ij}=\frac{x_{ij}-min\left(x_{ij}\right)}{\mathit{max}\left(x_{ij}\right)-min\left(x_{ij}\right)}$$

(1)

$$\mathrm{Negative} \mathrm{indicator}: {x^{\prime }}_{ij}=\frac{\mathrm{m}\mathrm{a}\mathrm{x}\left(x_{ij}\right)-x_{ij}}{\max \left(x_{ij}\right)-\mathrm{m}\mathrm{i}\mathrm{n}\left(x_{ij}\right)}$$

(2)

where ${x^{\prime }}_{ij}$ represents the standardized value of the indicator i in the year j, $x_{ij}$ is the original value of the ith evaluation object corresponding to the jth index, and $max\left(x_{ij}\right)$ and $min\left(x_{ij}\right)$ are the maximum and minimum values of each index, respectively.

Second, the proportion of the ith city in the jth index is calculated as:

$$P_{ij}=\frac{{x^{\prime }}_{ij}}{\sum _{i=1}^n{x^{\prime }}_{ij}}$$

(3)

Third, the entropy value of the jth index is calculated as:

$$e_j=-k{\sum }_{i=1}^n{p}_{ij}ln(p_{ij})$$

(4)

where $k=\frac{1}{lnn}$ and n represents the sample size.

Fourth, the difference coefficient of the jth index is calculated as:

$$g_i=1-e_j$$

(5)

Fifth, the weight of each indicator is calculated as:

$$w_j=\frac{g_j}{\sum _{j=1}^m{g}_i}$$

(6)

Sixth, the comprehensive score for each year is calculated as:

$$s_i={\sum }_{j=1}^m{w}_j{x^{\prime }}_{ij}$$

(7)

3.3.3. Kernel Density Estimation

Kernel density estimation is a non-parametric method used to estimate the density function of random variables. Compared with parametric methods, it has a weaker spatial dependence on the model and is now commonly used to describe the phenomenon of uneven distribution of research objects [56]. This study applied kernel density estimation to compare and analyze the distribution characteristics of the integration development level and its subsystem development level over different years. The calculation formula is as follows:

$$f\left(x\right)=\frac{1}{nh}{\sum }_{i=1}^{n}K(\frac{x_i-x}{h})$$

(8)

where $k(·)$ is the kernel function, $(x-x_i)$ is the distance from the estimated point $x$ to the $x_i$ of the sample point, $h$ is the broadband, and $f\left(x\right)$ represents the value of $f$ at $x$ as estimated by the sample point. In this study, the Gaussian kernel function was used to analyze the spatial–temporal evolution of the regional integration development level. The Gaussian kernel function is expressed as follows:

$$K\left(x\right)=\frac{1}{\sqrt{2\pi }}exp\left|-\frac{x^2}{2}\right|$$

(9)

3.3.4. Markov Chain

A Markov chain is a discrete Markov process for both time and state that can measure the state of an event’s occurrence and its development trend by constructing a state transfer probability matrix [57]. This method is capable of discretizing continuous data into k types and then calculating the probability distribution and transfer of the corresponding type. According to the state of regional integration development, the traditional Markov chain model can construct an N × N-order Markov probability transfer matrix to analyze the temporal evolution characteristics of regional integration development. Assuming that $P_{ij}$ is the transition probability of the regional integration development of a given unit in the study area from state $E_i$ to state $E_j$ from year t to year t + 1, the value can be estimated using Equation (10) as follows:

$$P_{ij}\left(E_i\to E_j\right)=\frac{n_{ij}}{n_i}$$

(10)

where $n_{ij}$ denotes the total number of regional units as the state type of the regional integration development transitions from $E_i$ to $E_j$, and $n_i$ denotes the number of regional units with $E_i$ occurring at the i level.

The traditional Markov process fails to reflect the effects of spatial factors. Therefore, a “spatial lag” is introduced into the traditional Markov analysis process to effectively examine the dynamic spatial evolution characteristics of the regional integration development level. Referring to Chen et al. [58], this study used a spatial adjacency weight matrix for measurement. Theoretically, an N × N transition probability matrix can be decomposed into an N × N × N transition probability matrix.${ P}_{ij}$ represents the probability of a research unit moving from state i in year t to state j in year t + 1 with spatial lag type${ N}_i$. The spatial lag type of a spatial unit is determined by its spatial lag value, which is the spatially weighted average of the attribute values of the neighboring regions of the spatial unit. The formula is:

$$Lag=Y_i{W}_{ij}$$

(11)

where $Y_i$ is the attribute value of the spatial unit and $W_{ij}$ is the element of the i-th row and j-th column of the spatial weight matrix W, which is the relationship matrix between the spatial unit and its neighboring regions.

3.3.5. Geographic Detector

A geographic detector is a set of statistical methods for detecting spatial dissimilarities and revealing the driving forces behind them, with capabilities such as factor detection and interaction detection [59]. In addition, a geographic detector has the advantage of immunity to covariance among the independent variables, which is applicable to the detection of numerous driving factors [60]. The spatial differentiation of regional integration is dynamically formed by the complex synergy of economic, social, cultural, spatial and ecological factors in the process of regional development. Therefore, a geographic detector is suitable for the identification of the many influencing factors in this research. This study examined the factors influencing the level of regional integration development with the help of a geographic detector and identified the dominant influencing factors. The calculation formula is as follows [60]:

$$q=1-\frac{\sum _{i=1}^m{n}_i{\sigma }_i^2}{n{\sigma }^2}$$

(12)

where q is the factor determining value with a range of [0, 1]; the larger the q value, the stronger the influence of the factor. The term n is the sample size, $n_i$ is the number of cities in the layer, and ${\sigma }^2$ and ${\sigma }_i^2$ are the variances of the entire area and area i, respectively.

3.3.6. Dynamic Change Model of Influence Factors

This study used the dynamic change model of the impact factor to measure the degree of dynamic change in the impact factor of integration development and analyzed the change in each impact factor. The formula is as follows [61].

$$IDM=\frac{q_{n+1}-q_n}{q_n}\times 100\%$$

(13)

where IDM is the dynamic value of the factor, and $q_{n+1}$ and $q_n$ are the determinants of the impact factor in years n + 1 and n, respectively.

4. Results

4.1. Temporal Evolution Characteristics of Regional Integration Development

4.1.1. Comprehensive Temporal Evolution Analysis

Based on the calculation results of the integration development index, a box map and a kernel density curve were drawn to present the time-series evolution characteristics of the integration development level between 2009 and 2018. As shown in Figure 3a, the mean value of the integration development level of each prefecture-level city in MYR-UA shows a fluctuating upward trend, with a small increase in the early stage and a large increase in the later stage. The scatter plot shows that most cities are distributed in the low-value area, but over time, the number of cities in the high-value area continues to increase, and the gap between regions widens. However, the randomness and complexity of the distribution of integration development levels among regions can change over time. Therefore, Stata 15 software (https://www.stata.com/, accessed on 15 June 2022) was used to plot kernel density curves to reveal the time-series dynamic evolutionary characteristics of the integration development level; the results are shown in Figure 3b. From the perspective of the distribution location, the kernel density curve shifts to the right as a whole, and the magnitude of the shift is small at first and then becomes large, indicating that the integration development level shows an increasing trend from slow to fast. From the shape, the kernel density curve evolves from bimodal to unimodal, indicating that the polarization levels are alleviated. In addition, the height of the main peak of the kernel density curve decreases, the width increases, and the tail on the right side continues to extend. This shows that high-value regions with integration development levels are increasing, and the gap between regions in the integration development levels is widening.

4.1.2. Temporal Evolution Analysis of Subsystems

According to the calculation results of each dimension of the integration development index, the corresponding box plots (Figure 4a,c,e,g,i) and kernel density curves (Figure 4b,d,f,h,j) were drawn for analysis.

(1)

Economic integration subsystem. The change trend in the mean value of the level of economic integration development is consistent with the overall trend, showing fluctuating growth (Figure 4a). Figure 4b shows that the kernel density curve shifted significantly to the right between 2009 and 2018, and the magnitude of the shift in the two stages was even, indicating that the level of economic integration was steadily improving. Moreover, the kernel density curve was always unimodal, indicating that there was only a single polarization phenomenon. The height of the main peak of the curve shifted to the right and then decreased, and the width of the crest increased, indicating that the difference in the level of economic integration development in various regions gradually expanded.

(2)

Public service integration subsystem. The mean value of the level of public service integration development was low, fluctuating between 0 and 0.03 with insignificant changes. The differences in the level of public service integration development among the cities in the study area narrowed and showed a convergent trend (Figure 4c). Furthermore, as shown in Figure 4d, the curve shifted slightly to the right between 2009 and 2018, indicating that the level of public service integration development was growing slowly. In terms of shape, the curve had a noticeable right-trailing characteristic, indicating that the level of public service integration development in most cities in MYR-UA was clustered in low-value areas, and only a few cities were close to high-value areas. In terms of the peak, the height of the main peak increased, and the crest narrowed, illustrating that the regional differences in the level of public service integration development in various cities was gradually narrowing.

(3)

Urban–rural integration subsystem. From 2009 to 2018, the mean value of the level of urban–rural integration development was low, but it showed a trend of steady growth (Figure 4e). Figure 4f further shows that the kernel density curve for 2009 had a unimodal distribution with a long left trailing tail, indicating that the level of urban-rural integration development was low in some areas. From 2009 to 2013, the left tail of the curve became shorter, and the curve showed a standard inverted “U” distribution, meaning that the regions with low levels were improving. From 2013 to 2018, the right tail of the curve became longer, showing that some cities with a high level of urban-rural integration development appeared. In terms of the peak, there was no significant change in the crest; however, the peak of the curve increased after decreasing. This demonstrates that the gap in the level of urban–rural integration development in each city experienced a decreasing–increasing process.

(4)

Ecological integration subsystem. The mean value of the level of ecological integration development showed a trend of first decreasing and then increasing, finally converging in the range of 0.05–0.1 (Figure 4g). Figure 4h further shows that the kernel density curve shifted to the left between 2009 and 2013, and the curve shifted to the right between 2013 and 2018. This means that the level of ecological integration development experienced a process of “decrease–increase”. From this shape, we see that the curve evolved from a left trailing shape to a right trailing shape, indicating that some cities with a low level of ecological integration development were gradually improving. From the peak, we can see that the peak rose and the crest narrowed significantly from 2009 to 2018, meaning that the gap in the level of ecological integration development of each city was decreasing.

(5)

Spatial integration subsystem. The mean value of the level of spatial integration development fluctuated and increased from 2009 to 2018 (Figure 4i). Furthermore, Figure 4j shows that the kernel density curve shifted slightly to the left between 2009 and 2013 and to the right between 2013 and 2018. This shows that the level of spatial integration development showed a trend of first decreasing and then increasing, which is consistent with the conclusion drawn from the box plot. In terms of shape, the curve undergoes the change process of “unimodal-multimodal-unimodal”, the peak of the curve undergoes the process of “increase-decrease”, and the width of the crest undergoes the process of “narrowing–widening”. This shows that the gap in the level of spatial integration development of each city showed a trend that first narrowed and then expanded.

4.2. Spatial Pattern of Regional Integration Development

4.2.1. Comprehensive Spatial Pattern Analysis

To further analyze the spatial differentiation of the integration development level of each city in MYR-UA, the natural breakpoint method in ArcGIS10.8 was used to classify the integration development into four levels: low, mid-low, mid-high, and high. We then selected 2009, 2013, and 2018 as the time points for analysis (Figure 5).

As shown in Figure 5, the integration development level of each city in MYR-UA between 2009 and 2018 showed a trend toward better development. Over time, the number of low-level areas decreased, and the number of mid- and high-level areas increased, showing a spatial distribution pattern of high-level areas in the southeast and low-level areas in the northwest. Specifically, in 2009, after the global financial crisis, the integration development level was relatively low, and cities northwest of the study area were low-level areas. At this stage, only Wuhan and Nanchang reached mid-level. In 2013, with the implementation of “Several Opinions of the State Council on Vigorously Implementing the Strategy of Promoting the Rise of the Central Region”, the mid-level (including mid-low- and mid-high-level) areas of integration development further expanded and spatially distributed in clusters. Changsha and Yingtan developed into new mid-level areas. However, it is worth noting that a high-level area had not yet appeared. In 2018, the “Opinions of the Central Committee of the Communist Party of China and the State Council on Establishing a More Effective New Mechanism for Coordinated Regional Development” clearly stated that it is necessary to accelerate regional integration development. At this time, the spatial pattern of the integration development level changed significantly, with a large number of mid-low-level and above areas increasing, accounting for 67.7% of the total number of research units. Wuhan, Nanchang, Yingtan, Changsha, and Xinyu developed into high-level areas, and Jingdezhen, Ji’an, Pingxiang, Zhuzhou, Xiangtan, Hengyang, and Huangshi developed into mid-high-level areas. At this stage, the integration development level in MYR-UA presents significant characteristics of spatial agglomeration and spatial correlation.

4.2.2. Spatial Pattern Analysis of Subsystems

Similarly, we used the natural discontinuity method to divide economic integration, public service integration, urban-rural integration, ecological integration, and spatial integration into four levels: low, mid-low, mid-high, and high. We then selected 2009, 2013, and 2018 as the time points for analysis (Figure 6).

As shown in Figure 6, the level of each subsystem of development in MYR-UA increased from 2009 to 2018, but there were significant differences in the scope of spatial expansion and spatial evolution patterns. Specifically, economic integration development in the early stages is dominated by low-level development. Only Xinyu and Yingtan were classified as high-level development areas, while Wuhan and Changsha were considered as mid-high-level development areas (Figure 6a). Affected by the global economic crisis, most cities had not yet recovered their level of economic development and the process of economic integration was low. From 2009 to 2013, the high- and mid-low-level areas expanded further, but the low-level areas still accounted for a large proportion. High-level areas were spatially scattered, whereas mid-level areas were spatially clustered. At this stage, the three provincial capitals (Wuhan, Nanchang, and Changsha) actively responded to the economic integration development policy, and economic integration development entered high-level areas. From 2013 to 2018, the spatial pattern of economic integration development changed significantly, and the mid-level area expanded rapidly, becoming the dominant type of spatial distribution. The high-level areas did not change during this period; the mid-high-level areas were clustered around the high-level areas, and the mid-low-level areas were clustered around the mid-high-level areas, with obvious hierarchical characteristics.

The change in the spatial pattern of public service integration is obvious, forming a spatial pattern of staggered high and low distributions (Figure 6b). In 2009, public service integration development mainly occurred at a low level, and mid- and high-level areas were scattered. Tianmen paid attention to the coordinated development of the regional public service level as early as the Eleventh Five-Year Plan period, so it took the lead in reaching the status of a high-level area. In 2013, all types of areas, except for low-level areas, increased. Overall, the level of public service integration development in the WMA was highest during this period. In 2018, the mid-high-level regions expanded further to the WMA, covering the southern cities of the entire WMA. The mid-low-level areas were concentrated in the northern part of the PLCG and southern part of the CZT-UA. During this stage, only three cities were high-level areas: Ezhou, Changde, and Xinyu.

The spatial polarization of urban-rural integration development is apparent, showing a spatial distribution pattern of high in the west and low in the east (Figure 6c). In 2009, the level of urban–rural integration development was extremely low. Only six cities were considered mid-level areas, and high-level areas were not formed. Spatial polarization characteristics initially appeared between 2009 and 2013. Xiangyang, Yichang, Jingmen, Tianmen, Xiaogan, Qianjiang, Xiantao, Yueyang, Xiangtan, and Yingtan developed from low-level to mid-low-level areas; Wuhan, Ezhou, Jingdezhen, Pingxiang, and Zhuzhou developed from mid-low-level to mid-high-level areas, and Nanchang and Changsha developed from mid-low-level to high-level areas. From 2013 to 2018, the phenomenon of spatial polarization intensified. The mid-high-level areas moved northwest of the WMA and southwest of CZT-UA. The high-level areas also included three cities: Wuhan, Pingxiang, and Xinyu. The spatial distribution pattern of “high in the west and low in the east” was fully formed.

The spatial agglomeration of ecological integration development is remarkable; there are different degrees of improvement or deterioration in each prefecture-level city, and spatial heterogeneity is obvious (Figure 6d). In 2009, after the global economic crisis, each city was eager to restore industrial production and develop its economy, ignoring the importance of ecological integration. Therefore, the development of ecological integration across the entire study area was mainly at the low and mid-low levels. From 2009 to 2013, the ecological integration development level of WMA declined, and the ecological integration development levels of PLCG and CZT-UA increased, showing a spatial distribution pattern of high in the southeast and low in the northwest. From 2013 to 2020, with the vigorous promotion of ecological civilization construction and green development concepts, the ecological integration development level of most cities markedly improved, the mid-high- and high-level areas further spread to the southwest of the study area, and regional differences expanded.

The spatial integration development presents a “core-edge” structure with “Wuhan–Changsha–Nanchang” as the core (Figure 6e). In 2009, the polarization of spatial integration development was noticeable, and the gap between core and peripheral cities was large. Only Wuhan, Nanchang, and Changsha were considered high-level areas. In 2013, the differences among the three major urban agglomerations and cities within urban agglomerations expanded further. Among them, WMA had the highest level of spatial integration development, forming a “core–edge” structure with Wuhan as the core. CZT-UA was next, showing a spatial pattern of high in the south and low in the north, and the lowest level of spatial integration was in PLCG, which showed a spatial pattern of low in the middle and high in the north and south. In 2018, the level of spatial integration was significantly enhanced. Only Xiangyang, Yichang, Jingmen, Xiantao, Changde, and Yichang were considered low-level areas. CZT-UA had the highest level of spatial integration development, followed by PLCG and WMA.

4.3. Spatial Pattern of Regional Integration Development

The above analysis explored the evolution trend of the integration development level of each city in MYR-UA from a spatial and temporal perspective but could not accurately reflect the internal flow direction and the probability of state transition of the integration development of each research unit. Therefore, this study used the Markov transition matrix to explore the dynamic evolutionary characteristics of the level of regional integration development. According to the relevant literature [44], the level of regional integration development is divided into four levels: the low level (I) is below 25%, the mid-low level (II) is between 26 and 50%, the mid-high level (III) is between 51 and 75%, and the high level (IV) is above 75%. The specific calculation results of the traditional and spatial Markov chains are listed in

Table 2. Traditional Markov and Markov probability transition matrix of the integration development level.

Spatial Lag Type	t/t + 1	I	II	III	IV
No lag	I	0.654	0.321	0.026	0.000
	II	0.158	0.500	0.303	0.039
	III	0.000	0.066	0.754	0.180
	IV	0.000	0.047	0.047	0.906
I	I	0.724	0.241	0.034	0.000
	II	0.125	0.625	0.188	0.063
	III	0.000	0.286	0.714	0.000
	IV	0.000	0.042	0.042	0.917
II	I	0.520	0.440	0.040	0.000
	II	0.300	0.450	0.200	0.050
	III	0.000	0.071	0.786	0.143
	IV	0.000	0.118	0.000	0.882
III	I	0.762	0.238	0.000	0.000
	II	0.174	0.391	0.391	0.043
	III	0.000	0.059	0.824	0.118
	IV	0.000	0.000	0.100	0.900
IV	I	0.333	0.667	0.000	0.000
	II	0.000	0.588	0.412	0.000
	III	0.000	0.000	0.696	0.304
	IV	0.000	0.000	0.077	0.923

Note: No lag represents the traditional Markov chain probability transition matrix.

.

There are three typical characteristics of the Markov transition probability matrix, except for the spatial lag factor. First, there is a club convergence phenomenon at the level of regional integration development in MYA-UA. The transition probabilities on the diagonal lines were greater than those on the non-diagonal lines. The probability of maintaining the original level after one year for low-level, mid-low-level, mid-high-level, and high-level cities is 65.4%, 50%, 75.4%, and 90.6%, respectively. This indicates that the integration development level in each grade was relatively stable, showing an evident convergence trend. Second, the probability value on the right side of the nondiagonal line is greater than the probability value on the left side, indicating that the probability of upward transfer of the integration development level is larger than the probability of downward transfer, which is consistent with the abovementioned results regarding the significant improvement in the integration development level in MYA-UA. Third, there is a possibility of cross-level transfer in the integration development level of each prefecture-level city; however, the probability of level transfer between adjacent types is higher. This shows that integration development is a gradual process and that leapfrog development is difficult to achieve.

As there is a certain geographical correlation in the integration development between regions, it is necessary to consider spatial factors to establish a spatial Markov transition probability matrix. The spatial Markov chain probability transfer matrix shows that the transfer of integration development levels in MYR-UA exhibits a specific spatial dependence under the action of a geographical spatial effect. Specifically, the four transition probability matrices differ for the different spatial lag types. This shows that when there are differences in the integration development level of neighboring cities, the probability of the transfer of the integration development level of cities differs. In addition, the diagonal elements of the transition probability matrix are smaller than the non-diagonal elements under different spatial lag types. This indicates that under the spatial spillover effect, the probability of the integration development level undergoing “grade locking” decreases, which is more obvious under the condition of a type IV lag. Furthermore, different lag types had different effects on the same grade. For example, the probability of transition from low-level to mid-low under the high-level lag type is 66.7%, which is significantly greater than that under the mid-high lag type. Finally, the same lag type has different effects on different grades. For example, under the condition of high-level lag, the probabilities of low-level, mid-low-level, and mid-high-level achieving an upward transition are 66.7%, 41.2%, and 30.45%, respectively, showing a decreasing trend. This implies that the transition probability is affected by the lag type and initial value of the integration development level.

4.4. Driving Factors of Regional Integration Development

By analyzing the spatiotemporal evolution pattern of the integration development level in MYR-UA, it was found that the integration development level shows a positive development trend in MYR-UA in general, albeit with a possibility of degradation. To further enhance the integration development level of MYR-UA, the major driving factors of the integration development level must be explored. This study uses a geographic detector to examine the impact of 21 evaluation indices on integration development levels. First, we imported the values of each evaluation index in 2009, 2013, and 2018 into ArcGIS 10.8 and classified them into five categories using natural breaks. The purpose was to transform the independent variables into type variables. The transformed data were then imported into the geographic detector to obtain the results of the influence factor of the level of regional integration development in

Table 3. Results of impact factor.

Detection Dimension	Detection Index	q in 2009	q in 2013	q in 2018	Average
Economic integration	GDP per capita	0.415	0.461	0.406	0.427
	Economic openness	0.510	0.454	0.409	0.458
	Industrial structure	0.559	0.597	0.654	0.603
	Economic development deviation	0.227	0.088	0.030	0.115
Public service integration	Proportion of education expenditure in fiscal expenditure	0.322	0.134	0.241	0.232
	Proportion of transportation construction expenditure in fiscal expenditure	0.207	0.088	0.066	0.120
	Proportion of medical and health expenditure in fiscal expenditure	0.072	0.217	0.274	0.188
	Proportion of social security expenditure in fiscal expenditure	0.130	0.274	0.174	0.193
Urban-rural integration	Urbanization rate	0.750	0.589	0.672	0.670
	Proportion of per capita income of urban and rural resident	0.160	0.058	0.232	0.150
	Engel coefficient proportion of urban and rural residents	0.112	0.224	0.172	0.169
Ecological integration	Industrial wastewater discharge	0.290	0.075	0.177	0.181
	Energy consumption per unit of GDP	0.099	0.144	0.234	0.159
	Industrial SO2 emissions	0.095	0.064	0.050	0.070
	Comprehensive utilization rate of industrial solid waste	0.143	0.159	0.231	0.178
	Proportion of environmental investment in fiscal expenditure	0.245	0.313	0.176	0.245
	Green coverage in constructed areas	0.085	0.031	0.070	0.062
Spatial integration	Population density	0.299	0.393	0.309	0.333
	Capital flow	0.407	0.401	0.443	0.417
	Information flow	0.361	0.265	0.636	0.420
	Technology flow	0.287	0.329	0.402	0.339

.

(1)

In the economic integration subsystem, the q-values of GDP per capita, economic openness, and industrial structure were relatively high, and only the q-value of economic development deviation was low. The q-value of the industrial structure increases year by year, and the q-values of GDP per capita and economic openness first increase and then decrease, but the change is not significant, and the q-values always remain above 0.4. This indicates that GDP per capita, economic openness, and industrial structure have a positive driving effect on the economic development of each city, which, in turn, has a fundamental influence on the integration development level.

(2)

In the public service integration subsystem, the q-value of each indicator changed significantly, indicating that their impact effects were not sufficiently stable. Among them, the q-value of education expenditure is relatively high and plays an important role in the public service integration subsystem. Education development is an important dynamic in promoting regional economic win–win dynamics, cultural integration, and complementary resources. Therefore, increasing investment in education funding and improving education development have a crucial impact on enhancing regional integration development.

(3)

In the urban-rural integration subsystem, the q-value of the urbanization rate is relatively high, with q-values of 0.750, 0.589, and 0.672 at the three time points, which are all above 0.5, indicating that the urbanization rate has a significant influence on the integration development level. The urbanization rate represents the proportion of the urban population to the total population and reflects the level of urbanization of a region. Existing studies have pointed out that with the advancement of urbanization, urban production, lifestyle, and urban civilization will continue to spread to rural areas, thereby narrowing the gap between urban and rural development and promoting urban-rural integration [45]. Therefore, in the process of regional integration and development, it is necessary to pay close attention to the urbanization rate.

(4)

In the ecological integration subsystem, the q-value of the environmental pollution treatment investment is relatively high, showing a trend of first increasing and then decreasing, with an average value of 0.245, indicating that environmental pollution treatment investment has a vital impact on the integration development level. The Development Plan for the Midstream City Cluster of the Yangtze River clearly states that ecological prioritization and green development represent the principles and requirements for coordinated development in MYR-UA. Investment in environmental treatment plays a key role in the construction of an urban ecological civilization and green development.

(5)

In the spatial integration subsystem, the q-values of capital and information flows are high, and the growth rate is high. In 2018, the q-values of both indicators were 0.443 and 0.636, respectively. This illustrates that capital flow and information flow have an increasing impact on integration development levels. Capital flow is conducive to strengthening interregional economic ties and narrowing the development gap between regions, but it may also exacerbate the development differences between regions, resulting in a situation in which “the strongest get stronger, and the weakest get weaker”. Information flow can promote interregional information dissemination and exchange, thus making the connections between regions closer.

In summary, the integration development level is influenced by numerous factors, and over time, the effect of each influencing factor also changes. To ensure that the integration development level can develop well for a long time, it is necessary to monitor the main influencing factors. Referring to the relevant literature [60], the impact factors were divided into primary and secondary impact factors based on whether the mean value of their q-value was greater than 0.2. We further used the dynamic change model of influence factors to measure the degree of spatiotemporal change in the main influencing factors to categorize them into enhancing, stabilizing, and weakening factors. The results are presented in

Table 4. Results of main influencing factors.

Detection Factor	Degree of Change	Factor Type
GDP per capita	−2.169	Weakening Factor
Economic openness	−19.804	Weakening Factor
Industrial structure	16.995	Enhancing Factor
Proportion of education expenditure in fiscal expenditure	−25.155	Weakening Factor
Urbanization rate	−10.400	Weakening Factor
Proportion of environmental investment in fiscal expenditure	−28.163	Weakening Factor
Population density	3.344	Stabilizing Factor
Capital flow	8.845	Stabilizing Factor
Information flow	76.177	Enhancing Factor
Technology flow	40.070	Enhancing Factor

.

Table 4 shows that industrial structure, information flow, and technology flow are the leading factors in the spatiotemporal evolution of the integration development level. The impacts of population density and capital flow on the spatiotemporal evolution of the integration development level are in a continuous steady state. GDP per capita, economic openness, the proportion of education expenditure in fiscal expenditure, urbanization rate, and the proportion of environmental investment in fiscal expenditure have negative effects on the spatiotemporal evolution of the integration development level. Therefore, in the process of integration development, each city in MYR-UA needs to continuously account for enhancing factors such as industrial structure, information flow, and technology flow. At the same time, each city also needs to be alert to the degradation effects of weakening factors such as GDP per capita, economic openness, the proportion of education expenditure and environmental investment in fiscal expenditure.

5. Discussion

5.1. Analysis of Evaluation System

Previous studies were more likely to measure the level of regional integration from a single perspective, such as economy, infrastructure, urban–rural development, and ecology [16,17,18,19,27,28,29]. However, these studies cannot accurately reflect the multidimensional attributes of regional integration, and were also unable to comprehensively measure the integration development process of a certain region. Drawing on the connotation of regional integration and existing evaluation indicators, and taking into account the actual situation in China, this study constructs a new comprehensive evaluation system for regional integration development. The index system contains the content of regional integration in the traditional sense, covers the content of regional integration extension, and is suitable for measuring the comprehensive level of regional integration development. The evaluation system constructed in this article not only draws on the research of Liao et al. [16], Duan et al. [18], Chen et al. [19], but also supplements and extends their research. In addition, Guo et al. pointed out that regional integration will continue to enrich with the development of the times [28]. Therefore, the evaluation system constructed by this research institute is also based on the current situation in China. In the future, it is necessary to further explore more effective evaluation mechanisms based on contemporary characteristics of regional integration.

5.2. Analysis of the Spatiotemporal Dynamic Evolution

Previous studies have focused more on analyzing the time-series characteristics of the level of integration development in a region [8,18,25,62,63,64], lacking visualization in space. Furthermore, few studies have monitored integration development trends and transfer patterns of a region over time. This study complements and improves on this aspect. First, we explore the spatiotemporal pattern of the integration development level of 31 prefecture-level cities in MYR-UA with the help of box plots, kernel density estimation curves, and GIS visualization tools, which enriches quantitative research on regional integration development. From the time series, the integration development level in MYR-UA showed a fluctuating upward trend. The results of Li’s study support this trend [65]. From the spatial distribution, the integration development level of each city in the southern part of the study area is higher than that in the northern part, and the north–south differentiation is more prominent after evolution, which is consistent with the results of Liu [66]. This difference can be attributed to the natural background, government policy, and history of development. Specifically, the cities with low levels of integrated development in the northern part of the study area are Xiangyang, Jingmen, Tianmen, Qianjiang, Huanggang, and Jingzhou. The topography of these cities is complex and includes hills and plains. Different topographic conditions cause the internal development of these cities to vary significantly [67], adversely affecting their integration development. Additionally, the economic development level, traffic conditions, information networks, and scientific and technological level of these cities are relatively backward [68], which weakens the basic conditions for integration development. In contrast, cities in the southern part of the study area have a flat terrain, convenient transportation, and good economic foundations [69], thus providing a firm foundation for regional integration development.

Second, we applied a combination of the traditional Markov chain and spatial Markov chain to explore the dynamic transfer processes and the law of integration development levels of 31 prefecture-level cities in MYR-UA. This approach can compensate for the shortcomings of past research that uses only traditional Markov chains [70] and clarify the improvement direction of the integration development of each city in MYR-UA. This is the difference between this study and previous studies [57]. The above research results show that there is a club convergence phenomenon in the integration development of each city in MYR-UA. The trend of the upward shift of the integration development level is greater than that of the downward shift, and the possibility of a cross-level shift is small. This suggests that the enhancement of the integration development level is a continuous, pressure-filled, and gradual process, which is in line with the evolutionary law of development in waves and spirals. In addition, the transfer of the integration development level of each city shows specific spatial dependence under the condition of geospatial effects. This suggests that regional integration development does not proceed independently and is often influenced by neighboring cities. The results of Li’s study support this trend [71]. Therefore, to formulate relevant policies effectively, the government needs to comprehensively consider the integration development level of neighboring cities.

5.3. Analysis of the Influencing Factors

Most previous studies on the factors influencing regional integration have used a qualitative approach [31,32]. This study used quantitative analysis to validate and complement the findings of previous studies. By using a geographic detector, we verify Song and Ruan’s claim that information and financial flows have a significant impact on regional integration [31,35]. The view that increased investment in infrastructure and science and technology innovation has a significant promoting effect on the regional integration process is also confirmed in this study [33,34]. In addition, we found that GDP per capita, economic openness, industrial structure, urbanization rate, and population density play important roles in the regional integration process. Then, we referred to Peng’s research and finally divided the influencing factors into dominant and secondary influencing factors based on whether the mean of the q value was greater than 0.2 [59]. Finally, with the help of the dynamic change model of influencing factors, the dominant influencing factors were classified into weakening influencing factors, stabilizing influencing factors, and enhancing influencing factors. Stabilizing and enhancing influencing factors have a long-term sustainable impact on regional integration development, which is a key concern for the future. Weakening influence factors also have a crucial impact on regional integration development, but over time, their influence will gradually diminish or even disappear. Therefore, policymakers must dynamically adjust relevant policies according to changes in the actual situation.

6. Conclusions and Suggestions

This study explored the evaluation methods and factors influencing regional integration. We analyzed the integration development level and influencing factors of 31 prefecture-level cities in MYR-UA. Policy suggestions were also proposed to promote the regional integration process. This study complements the relevant theoretical research on regional integration development, providing practical guidance for promoting the process of regional integration development and achieving sustainable urban development.

From the time series, the level of regional integration development in MYR-UA between 2009 and 2018 showed an overall upward fluctuating trend. This corresponds to the actual situation. Between 2009 and 2018, the government enacted several policies to coordinate regional development. Notably, the level of integration in different cities was highly differentiated. This may have been influenced by factors such as the economic level, regional policies, and geographic location. The level of each subsystem’s development also showed a fluctuating growth trend, but the growth rate was different. Specifically, the degree of economic, ecological, and spatial integration was relatively high. The level of urban–rural integration was growing at a faster rate. Changes in public service integration were not obvious.

From the perspective of spatial distribution, the level of regional integration development in MYR-UA showed remarkable spatial correlation and spatial clustering characteristics. The integration development level of each city in the southern part of the study area was higher than that in the northern part, and the north–south differentiation was more prominent after evolution. Therefore, in the future, the government should focus on the integration development of cities in the northern part of the study area.

The results of the Markov transfer matrix show that the level of regional integration development in MYR-UA displays “path dependence” and “self-locking” effects. The transfer of the integration development level usually occurs between adjacent levels, and the probability of a cross-level transfer is small. However, there was also the possibility of a downward shift. The government needs to focus on monitoring and responding to this situation in advance. Additionally, the dynamic change process of the integration development level is strongly influenced by regional factors. This indicates that regional integration is not a process of independent development. The government should pay attention to integration development within cities and between neighboring cities.

The dominant factors affecting regional integration development include GDP per capita, economic openness, industrial structure, proportion of education expenditure in fiscal expenditure, urbanization rate, proportion of environmental investment in fiscal expenditure, population density, capital flow, information flow, and technology flow. This conclusion can provide a reference for governments to formulate effective regional integration development measures and construct a feasible integration development model.

This study presents the following suggestions to promote the integration development process. First, it improves infrastructure construction and enhances public services. The research results and the previous literature confirm the importance of infrastructure construction in the regional integration process. Therefore, governments should focus on the development of infrastructure. Specifically, the government can promote the coordinated development of education in a region by increasing investment in education in rural areas [7]. Moreover, strengthening the construction of roads, railroads, and information networks in remote areas promotes intra-regional population flow, information flow, and the circulation of goods, eventually narrowing the regional development gap [9].

Second, it strengthens science and technology innovation and implements innovative development strategies. The results confirm that science and technology innovation is an important driving force for regional integration. Therefore, government budgets should focus on innovation investment in science and technology. Specifically, it should vigorously implement the talent introduction policy to introduce scientific and technological talent and retain them [60]. Additionally, a digital information platform was established. Digital information technology has become a new means of regional integration development through digital information resources and technical means to understand the demands and dilemmas of urban economic, ecological, and social development in a timely, accurate, and efficient manner [65].

Third, innovation in institutional mechanisms breaks traditional geographical boundaries and promotes the free flow of factors in the region. Improving regional cooperation mechanisms helps to promote the common development of developed and less-developed regions. By effectively managing the relationship between the government and the market, market forces determine price trends, respect the laws of market development, and promote the formation of a unified, open, and orderly competitive market system [72].

Finally, we optimize and upgrade the industrial structure and establish a scientific industrial division of the labor collaboration system. These results indicate that industrial structure is an important positive factor that affects regional integration development. For the sustainable development of a city, the government should promote the transformation and upgrading of industries. The development of sustainable, green, and low-carbon industries should be encouraged to promote harmonious development of the economy, society, and ecology [6]. In addition, this 0encourages and promotes employment in tertiary industries and supports the development of non-polluting or low-polluting enterprises, such as service, light, and innovative industries [61].

Although the results of this study have made substantial contributions to the evaluation of the driving factors of regional integration, some limitations point to the need for future research. First, owing to the availability of data, this study only selected 21 evaluation indicators based on five dimensions, namely economy, public services, urban and rural development, ecology, and space, and there may be other indicators not yet covered. In the future, we will apply big data (such as satellite remote sensing image data, pollution monitoring station data, tourism flow big data, enterprise big data, and POI data) to further enrich the evaluation index system for regional integration. Second, geographic detection is an effective method for exploring the critical drivers of regional integration. However, it is not possible to analyze the effects of government policies, leader capacity, and other factors on regional integration. In the future, we will use a combination of qualitative and quantitative methods to refine the factors influencing regional integration. Third, this study analyzed the future trends of the integration development level of 31 prefecture-level cities in MYR-UA, but a reasonable coping mechanism has not yet been constructed because of the limitation of the article length. In future research, we plan to construct corresponding prevention and control mechanisms to facilitate the integration process. Finally, Western countries such as the United States, the European Union, and Russia have accumulated a large amount of mature experience in the development of regional integration. In future research, we will further explore the reference significance of these mature experiences for China.