The Fairness and Influencing Factors of the Spatial Distribution of Public Services in Beijing–Tianjin–Hebei

Abstract

Achieving the equalization of public services is one of the main goals of regional highquality development at this stage. It plays an important guiding role in optimizing the spatial pattern of a metropolis. Taking Beijing–Tianjin–Hebei as the research area, this paper used hotspot analysis, spatial measurement, and other methods to conduct quantitative analysis and empirical research based on POI data, census data, etc., to identify regional public service centers and population distribution centers, and compared and matched them. This study built public service evaluation indicators, comprehensively evaluated the public service level of each district and county in the region, established a spatial error model, and empirically analyzed the differences in the comprehensive public service level of 200 districts and counties in the Beijing–Tianjin–Hebei region. The results show that there is a certain degree of mismatch between the comprehensive hotspots of density of public service facilities and the hotspots of permanent populations in the Beijing–Tianjin–Hebei region. The spatial differentiation of the comprehensive level of public services is obvious, showing the characteristics of “high, middle, and low in the north” and “high in the core urban area and low in other districts and counties”. Government policies, economic development level, and the level of urbanization have a significant positive effect on the public service level. The article also discussed the policy implications of the conclusions, as well as corresponding countermeasures and suggestions.

1. Introduction

Public service facilities are the basis for providing public services, including social infrastructure such as education, medical care, sports, and commerce. As public service facilities are traditionally public goods provided by the government, their supply scale, service level, spatial distribution, and accessibility are related to the social equity and fairness of urban public resource distribution. These facilities are an important factor reflecting the quality of life of urban residents, and an important way to achieve common human development, social prosperity, and sharing the fruits of social development [1,2,3].

The Beijing–Tianjin–Hebei region covers an area of 218,000 square kilometers and has a permanent population of more than 100 million. Based on the capital Beijing, the political, economic, cultural, scientific, and technological education status of the Beijing–Tianjin–Hebei region is prominent in the country. However, at the same time, the characteristics of “focusing on the city, ignoring the countryside” and “focusing on the capital, ignoring the surrounding areas” are also obvious, and “urban disease” and a “poverty belt” coexist [4]. At the same time, the allocation of public service resources is also, to a large extent, uncoordinated, and exhibits a cliff-like gap. This is reflected in the difficulties in sharing medical and health resources, the excessive gap in education resources, and the uneven fringe benefits of residents with different registration status [5]. Excessive disparities in the development of public services have greatly affected regional social equity and harmony. Since the coordinated development of Beijing–Tianjin–Hebei was put forward as a major national strategy in 2014, the planning outline and specific policies have been introduced and implemented in succession. One of the important starting points is to ease the non-capital core functions and solve the “big city disease” in Beijing. The key approach is to optimize the spatial layout, upgrade, and transfer industries, which are inseparable from the core role of public service facilities.

Since 1968, when Teitz first proposed the location theory of public service facilities [6], research on urban public service facilities has piqued the interest of foreign scholars. With the maturation of spatial-econometric analysis methods, academic research has become more detailed and in-depth. Researchers have conducted both theoretical and empirical studies that focus on the accessibility of public service facilities [7,8], socio-economic impacts [9,10], spatial equity [11,12], and location selection [13,14]. These studies have yielded corresponding research findings. Due to limitations in data acquisition, research on the allocation of domestic urban public service facilities lags behind in terms of both content and depth. Research on the fairness of public service space allocation can be roughly divided into two categories: The first category includes the range method, concentration curve method, Lorenz curve, Gini coefficient [15], circle analysis method [16,17], accessibility model [18,19,20,21,22], and comprehensive evaluation index system [23], and other methods used to describe the spatial fairness of public service facilities. Among these approaches, the method of constructing public service evaluation indicators to study the equity of public services has been widely adopted in recent years. The second category includes methods that focus on the factors affecting fairness, and explain the mechanism of the interaction between supply and demand, with a view to contributing to building a more equitable public service system. Based on different analyses of public service levels, the literature has explored the factors affecting fairness from various aspects, including economic level, industrialization level, openness, government policies, government administrative capacity, residents’ values and preferences, and historical factors [11,14,24,25,26,27,28,29,30]. A review of the relevant literature indicates that there are more studies providing qualitative and statistical descriptions, and fewer studies on the spatial differences and influencing factors of the public service level using methods of measurement and empirical testing. In terms of spatial scale, most studies are conducted at the scale of individual cities. In China, relevant studies have mainly focused on megacities such as Beijing, Shanghai, and Guangzhou. There are few studies on the coordinated development of regional public services based on contact with surrounding regions and horizontal comparison. Most existing studies tend to use traditional, statistical, distribution assumptions and regard the study area as an independent and homogeneous unit, whereas the possible relevance and heterogeneity among research units are seldom considered.

This paper selected the Beijing–Tianjin–Hebei region to study different types of public service facilities, from the district and county level and street level, in order to achieve small-scale and multitype analysis. This is conducive to horizontal comparison of public services in three provinces and cities using the same research method and indicator system. In this context, this paper analyzed the spatial distribution and supply level of public service facilities in the Beijing–Tianjin–Hebei region and conducted empirical research on the influencing factors to provide a basis for the formulation of public service integration policies.

2. Data Sources and Research Methods

2.1. Data Sources

This paper selected the administrative divisions of Beijing, Tianjin and Hebei Province as the research area and took six types of public service institutions as the research objects: primary schools, middle schools, medical and health institutions, libraries, museums and gymnasiums. The data on public services and social and economic development at the district and county level were mainly from the statistical yearbook. The POI data at the street level came from August 2018 (The data for 2016 at the district and county level in this paper are from the statistical yearbook, which provides the latest data at that level. Due to the real-time updating of POI data, it is impossible to achieve full matching in all years. However, unless there is a large-scale and rapid urban-pattern adjustment, the change in public service facilities will remain relatively small. Therefore, we use the two complements to conduct research at different scales) and were crawled from Gao de map, screened for the required public service elements after de-weighting and calibration, and imported into ArcGIS. The public service facility data at the street and township level was obtained by linking with the map to count the number of various public service facilities in each street or township. The street-scale population data came from the sixth population census in 2010.

2.2. Research Methods

2.2.1. Hotspot Analysis

This paper used the local Getis-Ord G* index to analyze the spatial distribution density characteristics of six types of public service facilities (the number of certain types of facilities per street area) in the street unit. The specific methods are as follows: Firstly, the density of public service facilities is calculated by taking streets as a unit. Then, the ArcGIS spatial statistics tool module is used to conduct hotspot analysis on the density of various public services calculated above. Finally, the hotspot analysis results of the six types of facilities are assigned and summarized. If the Z-score of the local Getis-Ord G* index of the density of various public service facilities is greater than 2.58, it indicates a very significant high-value cluster (the original assumption of no spatial autocorrelation can be rejected at the 1% level). If the Z-score is greater than 1.96 and less than 2.58, it indicates a significant high-value cluster (significant at the 5% level significance). The above two types of areas are hotspots, and a certain type of facility is assigned a value of 1 if it meets the conditions of a hotspot. The results of the six types of facilities at the street level are summarized to obtain a comprehensive map summarizing the density hotspots of public service facilities.

2.2.2. Spatial Metrological Analysis

Moran’s I index is a commonly used method to measure spatial autocorrelation and can be divided into the global Moran I index and local Moran I index. This paper used the global Moran I index to test the spatial dependence of the comprehensive evaluation level of public services, providing a basis for adopting the spatial measurement model.

3. Spatial Distribution Characteristics of Public Services and Population Matching Analysis in the Beijing–Tianjin–Hebei Region

3.1. Identification of the Public Service Comprehensive Centre

Against the background of the integration of Beijing, Tianjin and Hebei and efforts to allocate some of the non-capital functions to areas outside of Beijing, it is particularly important to consider the overall regional public service allocation level and spatial relations in Beijing, Tianjin and Hebei. This paper provided statistics on the spatial distribution of six types of public facilities at the street level in Beijing, Tianjin and Hebei, including primary schools, middle schools, medical and health institutions, libraries, museums and gymnasiums, and performed a hotspot analysis. The number of hotspot clusters, in descending order, are middle schools, primary schools, healthcare institutions, libraries, museums, and gymnasiums. The five hotspots of primary school facility density are distributed in Beijing, Tianjin, Shijiazhuang, Tangshan and Handan. The eight main hotspots of middle school density are distributed in Beijing, Tianjin, Baoding, Shijiazhuang, Handan, Zhangjiakou, Tangshan and Qinhuangdao. Among these, Zhangjiakou, Tangshan and Qinhuangdao have smaller hotspots and lower significance. The four major hotspots of medical and health institutions are located in Beijing, Tianjin, Shijiazhuang and Tangshan. The hotspots of library and museum density are very similar, with three main hotspots distributed in Beijing, Tianjin and Shijiazhuang. The density of gymnasium facilities has only two hotspots located in Beijing and Tianjin. The six types of public service facility density hotspots are assigned a value of 1 if a certain type of facility meets the hotspot conditions. The results of the assignment of the six types of facilities at the street level were summarized in the figure below to provide a comprehensive summary of public service facility density hotspots. A street summary with a value of 6 indicates that the street is identified as a hotspot under all six types of facilities.

In the summary map (Figure 1b) showing the areas in which facility density meets the conditions to be considered hotspots, the high values are mainly distributed in Beijing, Tianjin and Shijiazhuang. Among 2952 township streets in the Beijing–Tianjin–Hebei region, 249 township streets are identified as hotspots according to six facility densities, indicating that the concentration of various public services in and around these areas is high.

3.2. Identification of Permanent Population Centers

By revealing the local Getis-Ord G* index values that identify the hotspots of the permanent population density, Figure 2 shows that there are five hotspots in total in Beijing, Tianjin, Tangshan and Shijiazhuang. Tianjin has two hotspots, but they are close to each other.

3.3. Comparative Analysis

The distribution of public service facilities reflects the supply of public service resources, while the distribution of the permanent population reflects the demand for public service resources. To visually compare the distribution of hotspots for public service facility density and resident population density, we compared streets that meet the criteria for being identified as public service facility density hotspots in four or more of the six types of facilities with resident population density hotspots. This was shown in Figure 3.

As seen from Figure 3, the red area marked the overlap of comprehensive hotspots of public service facility density and hotspots of permanent population. These areas are mainly found in three areas on the map. The first area of overlap includes all streets in Dongcheng District and Xicheng District of Beijing, most streets in Chaoyang District and Haidian District, and some streets in Daxing District and Fengtai District. The overlapping area is surrounded by a small number of streets that are only comprehensive hotspots of public service facility density. The second area of overlap includes all the streets in the six downtown areas of Tianjin and most of the streets in the four districts around the city. To the east of the cluster, there are also a small number of streets that are only hotspots for public services. At the same time, there are many streets on the east and west sides that are only hotspots for permanent population density. The third area of overlap includes the Jingxing mining area in Shijiazhuang city, Hebei Province, and most of the towns and streets in Jingxing County, as well as some towns and streets in Jingxing County, Pingshan County and Luquan District. In Luquan District, the eastern part of the area is only a comprehensive hotspot of public service facilities. In addition, Fengnan District, Yutian County, and their nearby areas are large regions that are only hotspots for permanent residents. The red-marked area represents the matching and overlapping of facilities and population hotspots. The public service facilities in the green area are more adequate than the population, while the yellow area is densely populated but lacks the distribution of public service facilities. Therefore, this last area should be the key priority area for the development of public service facilities.

4. Comprehensive Evaluation of the Public Service Level in the Beijing–Tianjin–Hebei Region

4.1. Index Selection and Weight Determination

For the evaluation of public service levels, some scholars only considered the perspective of supply and calculated the score of various public service facilities by building evaluation indicators, such as the number, quality and satisfaction of public service facilities [17]. Others built the index from the perspective of supply (pressure) and demand (pressure) [28,29]. According to the supply and demand characteristics of public service facilities, and following the principles of scientific logic and accessibility, the evaluation index system of the Beijing–Tianjin–Hebei public service level was obtained based on the existing research on public service facilities in domestic and foreign literature, as well as theoretical analysis. The comprehensive level of public services was measured from education, healthcare and sports, as shown in

Table 1. Evaluation Indicator System of Public Service Level.

Type of Facility	Specific Metrics	Unit	The Nature of the Indicator
Educational facilities	Number of primary schools per 10,000 people	place	positive
	Number of secondary schools per 10,000 people	place	positive
	Number of full-time primary school teachers per 10,000 people	person	positive
	Number of full-time secondary school teachers per 10,000 people	person	positive
Medical facilities	Number of medical institutions per 10,000 people	piece	positive
	Number of medical technicians per 10,000 people	person	positive
	Number of beds per 10,000 people	sheet	positive
Cultural and sports facilities	Number of libraries per 10,000 people	piece	positive
	Number of museums per 10,000 people	piece	positive
	Number of gymnasiums per 10,000 people	piece	positive

Note: According to the Beijing Regional Statistical Yearbook 2017, Tianjin Statistical Yearbook 2017, and Hebei Economic Yearbook 2017, the data of each district and county are collated and converted into unified indicators.

.

To overcome the defect that subjective weighting is arbitrary, this paper uses the entropy weighting method to determine the corresponding weight for each measurement index (

Table 2. Weights of Public Service Level Evaluation Indicators.

Specific Metrics	Entropy	Variance Coefficient	Weight
Number of primary schools per 10,000 people	0.9985	0.0015	0.1168
Number of full-time primary school teachers per 10,000 people	0.9986	0.0014	0.1094
Number of secondary schools per 10,000 people	0.9992	0.0008	0.0614
Number of full-time secondary school teachers per 10,000 people	0.9987	0.0013	0.1003
Number of medical institutions per 10,000 people	0.9985	0.0015	0.1129
Number of beds per 10,000 people	0.9989	0.0011	0.0854
Number of medical technicians per 10,000 people	0.9984	0.0016	0.1226
Number of libraries per 10,000 people	0.9987	0.0013	0.1007
Number of museums per 10,000 people	0.9988	0.0012	0.0913
Number of gymnasiums per 10,000 people	0.9987	0.0013	0.0994

):

4.2. Spatial Characteristics of the Comprehensive Level of Public Services in the Beijing–Tianjin–Hebei Region

According to the index weight, the comprehensive score was obtained by multiplying the standardized data and weight of 200 districts and counties in Beijing, Tianjin and Hebei. Using ArcGIS software, the evaluation score of the public service level in Beijing, Tianjin and Hebei was visualized, and the spatial distribution map of the public service comprehensive level in Beijing, Tianjin and Hebei was drawn according to the natural fracture method (Figure 4).

The figure shows that the spatial heterogeneity of the comprehensive level of public services in the Beijing–Tianjin–Hebei region is obvious. In general, the comprehensive service level of the northern districts and counties is relatively high, while the low-level districts and counties are mostly distributed in the central region. High-level districts and counties are widely distributed in Zhangjiakou city. In addition, they are mostly distributed in the core urban areas of major cities, such as Dongcheng District and Xicheng District of Beijing, Heping District of Tianjin, Jingxing Mining Area of Shijiazhuang City, and Beidaihe District of Qinhuangdao City.

Quyang County, Dongguang County, Guantao County, Wei County and Jize County are poor counties and are among the bottom 20 counties in terms of comprehensive level. Improving the availability of basic public services will improve the efficiency of poverty alleviation. Thus, it is necessary to increase policy support and financial support for the development of public services in these areas. Zhuozhou and Gu’an (already included in Xiong’an New Area) are located in the “green economic circle around the capital” and are responsible for relieving the pressure on the capital city. In the process of undertaking the transfer of capital industry, these areas will benefit from the spillover effect of talent, technology, information and other elements from the capital. However, at present, the comprehensive ranking of public services in these two areas is relatively low (ranked 19th and 17th, respectively). Therefore, supporting infrastructure and public services must be provided while developing industries. The standards of Beijing and Tianjin in education, healthcare, sports, transportation and other public services can be matched to ensure that talents and industries are attracted to these areas, remain in these areas and develop well. This will achieve “industry–city integration,” utilize geographical advantages, enhance the vitality of economic development, and maintain sustainable development.

5. An Empirical Study on the Influencing Factors of Public Service Levels in the Beijing–Tianjin–Hebei Region

Based on the above analysis of the spatial layout and agglomeration characteristics of public services in the Beijing–Tianjin–Hebei region, a sample of 200 districts and counties in the region was taken into account, and the spatial effects were considered. The spatial econometric model was used to quantitatively evaluate and analyze the influencing factors of the comprehensive level of public services in the districts and counties in the Beijing–Tianjin–Hebei region.

5.1. Variable Selection and Model Setting

Pigou, a representative of welfare economics, proposed that an increase in national income would lead to an increase in economic welfare. Additionally, a reduction in income inequality would also increase the overall welfare level [31]. In 1954, Samuelson demonstrated the necessity of government provision of public services based on the market failure theory. He argued that government supply could achieve “Pareto optimality” in the allocation of social resources [32]. An important aspect of the government’s policy to address income inequality is the establishment of social security and social service facilities. It can be observed that public services, as a manifestation of social and economic welfare, are directly impacted by national income and government policies. Based on this premise, this paper took into account the level of economic development and government policies as the foundation. By combining the relevant literature and selecting appropriate indicators, the study aimed to identify the factors that influence the level of public service. Based on an analysis of differences in public service levels, the existing literature explores the factors that affect equity from various perspectives, including economic level, degree of industrialization, degree of openness, government policies, administrative capacity, residents’ values and preferences, and historical factors [33,34,35,36]. In terms of empirical research, most researchers selected indicators that reflect the level of economic development, such as per capita GDP, GDP per capita coefficient, household income, total industrial output, and economic growth rate. In terms of government policies, factors such as per capita fiscal expenditure, proportion of financial expenditure allocated to basic public services, local financial capacity, investment level in basic public services, government’s preference for public services, and degree of fiscal decentralization, were all important considerations. Indicators, such as urbanization level and population density, were used to measure population. Some scholars also studied the impact of foreign investment and the degree of marketization on public services [37]. Summarizing the existing literature, this paper selected five factors to study: government policy, economic development level, degree of marketization, degree of openness, and level of urbanization. This paper is based on the evaluation results of public services in 200 districts and counties in Beijing, Tianjin and Hebei, as described in Part 4. Since the evaluation score is between 0 and 1, to simplify the analysis, the original score was multiplied by 100 and converted into the percentile system. The comprehensive evaluation score of public services (Score 100) is taken as the dependent variable. The independent variable selection indicators include government policy, economic development level, marketization level, openness level and urbanization level. Specifically, this paper uses the per capita value (Policy_edu_medical_pc) of the sum of education and health expenditure in the general public budget expenditure to represent the government policy, which is a direct factor affecting the level of public services. The GDP per capita is used to measure the level of regional economic development, which is an internal factor affecting the development of regional public services. The total retail sales of consumer goods (Market) represent the degree of marketization. The per capita actually utilized foreign capital (FCaptial_pc), which is used to measure the level of foreign capital utilization and the degree of openness of a region. In addition, Urban Rate is introduced as a control variable (

Table 3. Variable definition and descriptive statistics.

Variable		Average Value	Standard Deviation	Minimum	Maximum	Unit
Comprehensive level of public services	Score 100	21.4	7.95	8.51	57.11	point
The level of economic development	GDP_pc	4.92	3.88	1.32	28.61	Ten thousand yuan per person
Government policy	Policy_edu_medical_pc	0.23	0.16	0.07	1.5	Ten thousand yuan per person
The degree of marketization	Market	150.35	288.91	10.06	2654.51	100 million yuan
Degree of openness to the outside world	Fcaptial_pc	126.64	259.99	0	2375.55	USD/person
Level of urbanization	UrbanRate	57.1	21.93	27.1	100	%

).

When the data involves regional factors, the established regional data model exhibits spatial heterogeneity and spatial dependence in the observations, which violates the Gauss–Markov hypothesis in the classical regression model. Considering that the research object in this paper may have a strong spatial dependence, a pretest of spatial autocorrelation is required before establishing the model. If the spatial effect is significant, it needs to be included in the model analysis framework, and an appropriate spatial measurement model should be selected for estimation. When quantifying the location factors of sample data, we can use adjacency to reflect a group of regional observation sets. The distance between any two points can also be calculated in this way. This distance generally uses longitude and latitude to represent the location in the analytic space to calculate the geographic distance. Economic distance can be similarly calculated. In this paper, spatial weights were generated based on geographical adjacency. The specific adjacency relationship is called “queue contiguity”; that is, two regions have a common edge or vertex defined as adjacent.

First, according to the comprehensive evaluation level of public services (Score 100), the global spatial autocorrelation test was carried out. The global Moran’s I index was 0.328, and the p-value was 0. The original hypothesis of “no spatial autocorrelation” was strongly rejected. At the same time, the index was positive, indicating significant positive spatial correlation. Next, a scatter plot of the Moran Index was drawn (Figure 5). It can be seen from the figure that most of the scattered points of the Moran index in 200 districts and counties are located in the first and third quadrants, indicating that high values are clustered together, as well as low values, resulting in a “positive spatial autocorrelation.” The LISA cluster diagram in Figure 6 indicates various clusters that have passed the significance test. Among them, red refers to significant high-value clusters, mainly distributed in the north, while low-value clusters are distributed in the middle and south.

Finally, to compare the suitability of the spatial error model (SEM) and the spatial lag model (SLM), a spatial dependence test is conducted. This includes the Lagrange multiplier test and the robustness form of the spatial error and spatial lag models. The results are shown in

Table 4. Statistical Values of the Spatial Dependency Test.

TEST	VALUE	p
Lagrange Multiplier (lag)	27.010	0.000
Robust LM (lag)	0.103	0.749
Lagrange Multiplier (error)	30.342	0.000
Robust LM (error)	3.335	0.068

. The LM error based on the Queen spatial weight matrix is numerically larger than the LM lag. LM error passed the significance test at the 1% level, and R-LM error was significant, while R-LM lag was not. Therefore, according to the criteria, the spatial error model is preferred.

The spatial error model is as follows:

$$\mathrm{S}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{e}100=\alpha +{\beta }_{1}GD{P}_{pc}+{\beta }_{2}Polic{y}_{edu}\_medical\_pc+{\beta }_{3}Market+{\beta }_{4}FCaptial\_pc+{\beta }_{5}UrbanRate+\mathsf{\lambda }\mathrm{W}\mathsf{\mu }+\mathsf{\epsilon }$$

(1)

where W is the spatial weight matrix and λ is the parameter of spatial correlation error.

5.2. Empirical Test Results and Analysis

Through the comparison of OLS regression and SEM regression results based on Queen’s adjacent spatial-weight matrix (

Table 5. Regression Results of OLS and SEM Models.

VARIABLES	OLS	SEM
GDP_pc	0.3520 *	0.4011 **
	(0.2020)	(0.1996)
Policy_edu_medical_pc	15.1457 ***	12.9852 ***
	(3.5843)	(3.4387)
Market	−0.0018	−0.0028
	(0.0025)	(0.0025)
Fcaptial_pc	−0.0035	−0.0028
	(0.0028)	(0.0025)
UrbanRate	0.0981 ***	0.0968 ***
	(0.0267)	(0.0279)
Constant	11.3098 ***	11.7788 ***
	(1.5263)	(1.8005)
lambda		0.4235 ***
		(0.0752)
Observations	200	200
R-squared	0.2491	0.2471

Note: ***, **, * represent significance at the levels of 1%, 5%, and 10%, respectively.

), it can be seen that, compared with OLS, the test value of the SEM has been improved to some extent. The estimated value of the spatial autoregressive coefficient (lambda) of the introduced error term is 0.4235, which is significant at the 1% level, indicating that the comprehensive level of public services in the Beijing–Tianjin–Hebei region shows a relatively obvious spatial effect. The empirical model considering spatial error correlation is more appropriate.

From the perspective of the spatial autoregressive coefficient, government policies, the economic development level and average urbanization water have significant positive effects on the public service level. Compared with OLS regression, the coefficient of government policies and urbanization level in the spatial-error model decreased after introducing spatial factors, while the coefficient of economic development level increased to some extent, which means that if spatial factors are not taken into account, the model will overestimate the impact of government policies and urbanization on the comprehensive level of public services and underestimate the impact of economic level. The per capita actual utilization of the foreign capital variable measuring the degree of openness to the outside world, did not pass the significance test. This is because, at the district and county level (especially in the counties), the actual utilization of foreign capital is low or even zero. The difference between counties is small, and the fluctuation between years is large, unlike the ranking of GDP and other indicators.

The government policy is significant at the 1% level, which indicates that the improvement of the public service level in the Beijing–Tianjin–Hebei region largely depends on the government’s planning, investment in public services and other policies. Even against the background of innovative mechanisms and methods for the government to purchase public services, it is still necessary to highlight the leading role and forward-thinking outlook of the government in the field of public goods provision, and increase support for the public service industry. The economic development level is significant at the 5% level, which indicates that the improvement of the public service level basically requires the support of a sufficient economic foundation. Through benign interactions and improvements in economic development and public services, the region’s capacity can be effectively enhanced, generating a scale effect in public service supply and allowing more people to enjoy better public services.

6. Conclusions and Policy Recommendations

Firstly, the assignment summary of six types of facility density hotspots based on the street scale shows that the comprehensive hotspots of facility density are mainly distributed in Beijing, Tianjin and Shijiazhuang. The analysis of the matching degree of public service supply and demand shows that there are three clusters in some streets of Beijing, Tianjin and Shijiazhuang, Hebei Province, which are simultaneously identified as hotspots of public service facility density and hotspots of permanent population. Some areas are only identified as hotspots in either facility density or permanent population. There is a mismatch between supply and demand in these areas. Either the distribution of public service facilities is insufficient, or the supply is greater than the demand, and public facilities are not fully utilized.

Secondly, in terms of the comprehensive evaluation of public services, the spatial differentiation of the comprehensive level of public services is clear. On the whole, the comprehensive service level of the northern districts and counties is relatively high, while the districts and counties with low-level services are mostly distributed in the central region. The high-level districts and counties are widely distributed in Zhangjiakou city. In addition, these areas are mostly located in the core urban areas of major cities.

Thirdly, in terms of empirical research on the influencing factors of public service, the Moran index shows that there is a “positive spatial autocorrelation” in the comprehensive level of public service in the Beijing–Tianjin–Hebei region. The SEM regression results, based on Queen’s adjacent spatial-weight matrix, showed that the comprehensive level of public service in the Beijing–Tianjin–Hebei region exhibits a relatively significant spatial effect. The economic development level and urbanization’s average water have a significant positive effect on the public service level.

The implementation of the city’s strategic positioning and the coordinated development of Beijing, Tianjin and Hebei have given rise to new and higher requirements for Beijing’s population evacuation and the spatial layout of public service resources in the three provinces and cities. Based on the research conclusions of this paper, the following suggestions are proposed:

(1)

Scientific planning is needed to complement the shortfall in public service allocation. The present research revealed that the comprehensive hotspots of facility distribution in the Beijing–Tianjin–Hebei region from the perspective of multiple centers are highly concentrated. However, there is a certain degree of mismatch with the hotspots of permanent population distribution. Differences between the permanent population and various public service facilities leads to a deficiency in urban functions. Some regions are densely populated but have an insufficient supply of public service facilities, which make public service resources scarce and result in a lack of corresponding service function support. Therefore, it is necessary to consider the distribution characteristics of the comprehensive level of public services in the Beijing–Tianjin–Hebei region from a macro-perspective, which can be summarized as “high in the north, low in the middle” and “high in the core urban area, low in other districts and counties.” From a micro-perspective, it is necessary to define the public services that need to be supplemented in various streets and towns. For the “non-hotspots” of public service configuration, it is necessary to increase investment, engage in scientific planning, address the weaknesses, promote the balanced development of various public services, and promote the integration of public services in the Beijing–Tianjin–Hebei region.

Public service facilities are affected by the characteristics of fixed costs and location, and lack adjustability after completion. The reason for the shortage of public service resources is largely due to the failure to effectively predict future public service demand and reasonably allocate public service facilities. Therefore, more attention should be provided to improving the level of effective supply and governance, rather than unilaterally emphasizing the rigid control of urban population size and development speed. When evaluating the allocation level of public service resources at the subdistrict, township or district/county level, in addition to the current population, the impact of further population expansion, periodic adjustment of population structure and the liberalization of the second-child policy on the allocation increment should also be considered. Some scholars, based on the population forecast after the implementation of the “comprehensive two-child” policy, had calculated the number of public service resource gaps in public education, healthcare and other aspects in China’s provinces and regions from 2016 to 2030. The results showed that the average annual growth rate of resource demand for preschool education, primary and secondary schools in Beijing and Tianjin is at the forefront of the country. The phenomenon of a staged surge is prominent, and the allocation pressure is high. Hebei Province has a large gap between the supply and demand for educational facilities, but the annual allocation intensity is low [23]. At the same time, when expanding capacity and constructing new buildings, it is important to fully consider the characteristics of public service demand. For example, medical service is a general demand that runs throughout people’s lives. Education service is a phased demand, and the demand reaches its peak in the short term, but there is also a period of decline. Avoiding the waste of public service resources has become a problem that needs further consideration in follow-up research.

(2)

Cultivate multi-center public service functions from point to area. To redistribute the population and industries from the overpopulated areas of the capital in an orderly fashion, it is necessary to consider the “breakthrough” of public services and guide the aggregation of market resources by allocating public service resources to achieve the role of guiding the spatial distribution of the population. The comprehensive evaluation level of public services in Anxin County and Xiong County lags behind and does not conform to the high standards for strategic positioning of Xiong’an New Area. Therefore, these areas should be the key areas for the development of a future comprehensive service center. From point to area, it is important to highlight the radiating role of a public service center, in terms of its functions, and to apply successful experiences to the surrounding districts and counties to achieve the overall improvement of the public service level, and promote the maturity of the multi-center spatial structure of public service functions.

(3)

Develop the economy and consider the policy and market simultaneously. The empirical analysis shows that the main factors affecting the level of public services are economic development, government policies and the level of urbanization. The improvement of the public service level basically requires the support of an economic foundation. Through benign interaction and improve economic development and public service levels, regional carrying capacity can be effectively enhanced. This generates a scale effect in public service supply, allowing more people to enjoy better public services. In terms of public service supply, we can learn from the experiences of developed countries and success stories in China. We should highlight the leading role and foresight of the government in the field of public goods provision, increase support for public services, improve the efficiency of resource allocation by improving a diversified supply system under the market competition mechanism, and further optimize the efficiency and level of public service supply.

(4)

Establish a cooperative mechanism for the supply of public services in the Beijing–Tianjin–Hebei region. Firstly, establish a pattern of mutually beneficial interests. We will explore ways to establish a mutual compensation mechanism for financial support of public services among the three regions. For example, we can explore the new approach of “money follows people” in the financial subsidies for medical insurance and endowment insurance in the three regions of Beijing, Tianjin, and Hebei. Secondly, establish an integrated and standardized integration system. The goal is to standardize public services across the Beijing–Tianjin–Hebei region, enabling mutual recognition and alignment of standards within the area, and expand the field and scope of standardization in public service. Thirdly, we will utilize big data to facilitate the sharing of resources among public services and enhance their complementary functions. The coordinated development of the Beijing–Tianjin–Hebei region should begin with improving the level of information supply for integrating public services. The government can utilize the Internet and other information technology resources to enhance the efficiency of public service delivery. By leveraging the benefits of the “Internet plus” model, convenient services can be provided, promoting the ease and intelligence of public service supply.

China is seeking to enter the stage of high-quality development, which demands greater efficiency in resource allocation. Therefore, future research on the spatial arrangement of public service facilities should focus on this aspect. Existing research can be further developed and refined. Based on the latest seventh census data, it is possible to analyze the alignment of educational, medical, cultural, and sports facilities with the age structure of the population in the Beijing–Tianjin–Hebei region. By considering the economic development situation and trend in the area, future research can investigate the degree of match between the spatial distribution of public service facilities and the spatial pattern of future economic development. Furthermore, considering the perspective of users, using a combination of questionnaire and spatial analysis, conducting a thorough analysis of the rationality of the spatial distribution of urban public service facilities, taking into account both the quantity and quality of supply, should also be a focus for further research.