Spatial pattern and drivers of urbanization in China’s mid-level developing urban agglomeration:A case study of Chang-Zhu-Tan

Ynhu He,Yi Lin,Guohu Zhou,＊,Yixun Zhu,Ki Tng

aKey Laboratory of Geospatial Big Data Mining and Application,Department of Resources and Environmental Science,Hunan Normal University,Changsha,410081,China

bInstitute of Real Estate Administration,Hunan Normal University,Changsha,410081,China

cHunan Planning Institute of Land and Resources,Changsha,410007,China

ABSTRACT

Urban agglomeration has become the main form of regional spatial organization in China.While most of the existing studies of urban agglomeration in China have focused on the eastern coastal areas,urban agglomeration with mid-level development in the rest of the country has been overlooked.To better understand the urbanization process of the mid-level developing urban agglomeration,this study investigated the clustering pattern and the drivers of both urban population and firm dynamics during 2005–2015 in the Changsha-Zhuzhou-Xiangtan(Chang-Zhu-Tan)urban agglomeration of China using the methods of kernel density estimation and geographic detection.Our results show that centralization was obvious,although decentralization also occurred in Chang-Zhu-Tan,and that the spatial agglomeration was promoted by several factors,such as administrative resources,location advantage,labor cost,and consumption capacity.Some problems hindering the development of this region were also discovered:administrative resources played a critical role in urbanization because small towns and villages did not receive enough attention,and the effect of local policy was not as beneficial as expected.These findings partly explain the relatively slow development of mid-level developing urban agglomerations and have important implications for promoting healthier urbanization.

ARTICLEINFO

Keywords:

Urbanization

Spatial pattern

Urban agglomeration

Firm clustering

Driving mechanism

China

1.Introduction

Urbanization is one of the most important human activities,driving economic development,changing the urban form and rural landscape,affecting human well-being,and creating enormous impacts on local,regional,and global environments(Turner et al.,1990;Antrop,2004;Long et al.,2018;Gu,2019;Li and Liu,2019).Since the second half of the 20thcentury,the trend of world urbanization has become irreversible.In the future,it will continue to increase,and Asia and Africa will become the main areas of urban population growth.China is expected to account for approximately 10% of the growth in the world’s urban population between 2018 and 2050(United Nations,2018).Therefore,China’s urbanization has received worldwide attention.

Many studies have been conducted on the process and mechanism of urbanization in China.These studies have shown that China has experienced rapid urbanization and urban expansion in the last four decades(Fang,2018;Zhao et al.,2019),with urban population gathering in large and medium-sized cities and the scale of urban construction land expanding rapidly(Dai et al.,2014).The process of urbanization is closely related to all aspects of economic and social development as well as the flows and changes of population,resources,and economic factors(Chen et al.,2015).Both geographical and socio-economic factors play important roles in determining urbanization patterns,e.g.,gross domestic production,population distribution,industrial agglomeration,consumption market,investment level,technological progress,and land use policies(Long et al.,2007;Zhou and He,2007;Ma and Xu,2010;Ma et al.,2016;You and Yang,2017;Li et al.,2018).There are great differences in the development foundations and conditions of urbanization in different regions in China,which will inevitably affect the hierarchical scale of urban systems,the relationship between urban and rural areas,and the regional role of central cities(Lu,2013).Therefore,China’s urbanization processes have significant regional differences(Chen et al.,2010;Wang et al.,2014;Li et al.,2018),with significant spatial agglomeration(Jiang et al.,2016).Urbanization in eastern China has entered a new stage characterized by innovation,compound drivers,people orientation,an emphasis on quality,and coordinated development between urban and rural areas(Chen et al.,2015;Zhou et al.,2019).The level of urbanization in the central region is generally low,and the cities and towns are relatively small with different geographical characteristics(Fang et al.,2015;Chen and Xie,2018).However,urbanization in the western region is facing a special dilemma,as it is experiencing failure in the context of backwardness(Wang et al.,2014).In summary,China’s urbanization is a highly complex process with scale dependence and regional differences(Gu,2012).

In recent years,the combination of institutional changes,marketization,and globalization has brought a new process of urbanization and urban-rural interaction,resulting in a new form of human settlements in China(Ma,2002;Pannell,2002;Tian et al.,2005).Urban agglomeration,a highly spatial form of integrated cities,has become the main form of regional spatial organization in China’s urbanization.An urban agglomeration has a mega-city at the center and three or more metropolitan areas or large cities forming the core region.The core region is connected to daily commutable peripheral areas via highly developed transportation and other infrastructure networks,forming a spatially compact,economically related,and regionally integrated urban entity(Fang and Yu,2017).As scholars and officials have acknowledged,urban agglomerations have contributed greatly to China’s urbanization and economic development in the last four decades(Fang,2018).Many studies have been conducted on the spatial structure and dynamics of urban agglomerations(Hu,1998;Xu et al.,2000;Ning,2006;Yu and Wu,2006;Wang et al.,2008;Zhang et al.,2008;Fang and Zhang,2014;Zhao et al.,2014;Yao et al.,2017;Feng and Wang,2018;Gu,2019).Urban agglomeration,as a main form of urban spatial organization,is characterized by intensive towns and populations,a prominent role of central cities,close economic and social ties between cities,and a high degree of openness(Ning,2011;Huang,2014;Chen et al.,2015).The development of the high-tech industry,the application of information technology,the construction of rapid transportation networks,the rapid growth and structural optimization of the economy,and regional integration have become the main driving forces promoting urbanization(Wang and Cheng,2010;Wang and Fang,2011;Peng and Wang,2015;Wang and Feng,2016;Zheng et al.,2016).Geographical,cultural,and administrative proximity also exert a significant influence on the formation and evolution of urbanization patterns(Gao et al.,2019).

Previous studies were helpful in capturing the spatial features of urban agglomerations and exploring the urbanization process.However,most of these studies focused on high-level urban agglomerations and paid less attention to mid-level ones.Many urban agglomerations with different degrees of development constitute an integrated system in China and play different roles(Fang et al.,2005).According to the existing research,the Yangtze River Delta,Pearl River Delta,and Beijing-Tianjin-Hebei region in the eastern coastal area of China have entered a high-level stage of development(Chen et al.,2015;Lu,2015;Zhou et al.,2019)with high density,strong element flows,and powerful influence and competitiveness.However,the Changsha-Zhuzhou-Xiangtan(Chang-Zhu-Tan),Shandong Peninsula,Chengdu,Hohhot-Baotou-Ordos,Wuhan,Guanzhong,and other urban agglomerations are still at the mid-level stage(Fang et al.,2005;Fang,2018).While the high-level agglomerations constitute China’s main participants in global competition and cooperation,the mid-level agglomerations are the key to promoting the coordinated development of all regions.Therefore,we need to explore them more.The Chang-Zhu-Tan urban agglomeration,the first to consciously implement regional economic integration in China,is located in the central region of China.It was approved as the national comprehensive reform pilot area for a resource-friendly and environmentally friendly society in 2007.As an important part of the urban agglomeration in the central reach of the Yangtze River,it was listed as one of the key urban agglomerations for cultivating the new urbanization strategic pattern.Now,it plays an important role in linking West and East and connecting South and North.At present,some research has been carried out on the Chang-Zhu-Tan urban agglomeration.Self-organization and hetero-organization have worked together in its formation and development(Peng and Lin,2015;Zhou et al.,2018).The urbanization process and spatial structure are influenced by the market economy,spatial competition,industrial upgrading and spatial transfer,and planning and regulation(Tang and Su,2010;Tang et al.,2018).It has made great progress in the last few decades,but the urban network system is not yet perfect(Peng and Lin,2015;Chen et al.,2016),and the degree of regional integration is not very high(Chen and Song,2011;Xu et al.,2015;Tang et al.,2018).The comprehensive development degree is at the mid-level(Fang et al.,2005),which to a certain extent limits its role in the national development strategic pattern.To better understand its development process as a mid-level urban agglomeration, we explored the dynamic changes and driversof urbanization patterns.

Therefore,this paper aims to reveal the urbanization pattern characteristics and driving mechanism of the Chang-Zhu-Tan urban agglomeration from a new perspective.Unlike previous studies,we took both the populations and firms into account and detected the interaction relationship between them,with the findings for one confirming the findings for the other.Because urbanization is the agglomeration and diffusion process of capital,talent,information,and technology in urban and rural areas(Hu,1998),analysing the agglomeration and diffusion process of factors is an effective way to understand urbanization(Ning,2011).As micro-carriers of urban and rural mobility,populations and enterprises have always played an important role in urbanization,affecting the flow of capital,technology,talent,and information.In the process of urbanization,the increase in population size promotes the spatial structural evolution of urban agglomeration(Sun et al.,2017)and the concentration and development of enterprises(especially those in the manufacturing and service industries)(Gu,2019;Zhang et al.,2019;Li et al.,2020).Thus,the spatial distribution of populations and enterprises is a direct manifestation of the urbanization model(Mao et al.,2015;Wu et al.,2015).The mechanism of urbanization is explored by analysing the factors that affect the aggregation and diffusion of populations and enterprises and their interaction process.Their spatial distribution is the direct manifestation of the urbanization model and the micro-carrier of urban-rural flows.The results of research on each entity confirms the other(He et al.,2019).Therefore,this paper is based mainly on population and firm data of the Chang-Zhu-Tan urban agglomeration,analyses the spatial agglomeration status and the influencing factors,and discusses the inter-relationships and interaction degrees among the factors to reveal the characteristics of the urbanization pattern and mechanism.On this basis,we identify problems with urbanization that may partly explain why the area is not developing more rapidly and propose some suggestions for healthier urbanization.

2.Materials and methods

2.1.Study area

Chang-Zhu-Tan,located in the central reaches of the Yangtze River(Fig.1),has a population of 1.05×106and covers an area of 2.81×104km2(Hunan Provincial Bureau of Statistics,2006,2019).It includes the 3 prefecture-level cities of Changsha City,Zhuzhou City,and Xiangtan City,which comprise 12 districts,5 county-level cities,and 6 counties.This region is at a mid-level of development and is experiencing rapid economic growth and urbanization with a per capita GDP of 1.06×105CNY,an urbanization level of 72.8%,per capita disposable income of urban residents of 4.66×104CNY,and per capita disposable income of rural residents of 2.47×104CNY in 2018(Hunan Provincial Bureau of Statistics,2006,2019).It has been listed as a key urban agglomeration to be fostered in the national plan of new urbanization in China.

2.2.Date sources

Fig.1.Location of Changsha-Zhuzhou-Xiangtan(Chang-Zhu-Tan)in Hunan Province.

This paper relied mainly on population and firm data.The main sources of population data are the economic and social statistics statements of townships(towns and sub-districts)and the statistical yearbooks of districts or counties for 2015.Here,we focused only on the urban population,as it is the core subject of urbanization in urban agglomerations.The firm data were from the legal entity basic information database of Hunan and the Corporate Yellow Pages.The Industrial Classification for National Economy Activities(GB/T4754-2017)divides industries into 20 types in the national economy.This study focused on manufacturing(MF)and high-end services(HS,including research and technology services,finance,and information transmission and computer services)firms.

Other data were used to measure the indicators of influencing factors.Specifically,socio-economic data were obtained from the economic and social statistics statements of townships(towns and sub-districts)and the relevant statistical yearbooks,and topographical data were obtained from Google maps. To accurately reflect the spatial heterogeneity of populations and firms, we selected the data scale at the township level(towns and sub-districts).

2.3.Methodology

First,we analyzed the agglomeration status of populations and firms by the kernel density estimation method to reflect the urbanization pattern characteristics in the Chang-Zhu-Tan urban agglomeration.Then,following the qualitative analysis and selection of the main influencing factors of population and firm agglomeration,we measured the correlation between these factors and the population or firm agglomeration scale by Geographical Detector(GeoDetector)to reveal the driving mechanism of urbanization in the Chang-Zhu-Tan urban agglomeration.

2.3.1.Kernel density estimation

Kernel density estimation is essentially a sophisticated form of locally weighted averaging of the distribution(Tukey,1977).It can be used to estimate the distribution intensity of sample points and represent them by a smoothed continuous surface to help analyze the presence of clusters or regularities in the parameter distribution(Gatrell et al.,1996).It has been applied in several empirical studies using ArcGIS.We used the kernel density estimation method to analyze the agglomeration pattern of firms with a search radius of 5000 m.

The calculation equation of the kernel density method can be expressed as:

where f(s)is the kernel density calculation function at space position s;h(h=1,2,3,…)is the distance attenuation threshold(i.e.,bandwidth);n is the number of elements with a distance from position s of less than or equal to h;and k represents the spatial weight function;s–ciis the distance from the estimated point s to the sample ci.

2.3.2.Geographic detection

GeoDetector is a new set of statistical methods used to detect spatially stratified heterogeneity and reveal the driving factors behind it without a linear hypothesis(Wang and Xu,2017);it has been applied in many fields in the natural and social sciences.It includes four detectors:a risk detector,a factor detector,an ecological detector,and an interaction detector(Wang et al.,2010).This paper used a factor detector to detect explanatory variables for the population and firm pattern and an interaction detector to analyze the interactive relationships between variables,both of which were measured by a q value defined as follows:

where l is the stratum of independent variable x,that is,the classification of x;nhis the unit number of stratum h;n is the total unit number in the whole region;andand σ2are the variances for dependent variable y of stratum h and for the whole region,respectively.The q value is between 0 and 1.In this research,q represents the degree of influence exerted by the influencing factors on the spatial pattern of the population or firm.The greater the q value,the stronger the influence of the factor;q=1 indicates that the factor completely explains the spatial pattern,whereas q=0 implies that the factor is completely irrelevant to the spatial pattern.

The interaction detector identifies the interaction between two factors by analyzing the relationship of q(x1),q(x2),and q(x1∩x2),that is,the q value when we overlay factor x1and factor x2.The interaction can be defined as follows(Wang et al.,2010):

2.3.3.Selection and calculation of indices

The key to factor and interaction detection is choosing the right indicators to explain the influencing factors of population and firm patterns.In the research,we hypothesized that the spatial patterns of populations and firms are always influenced by three dimensions:geographical environment,market condition,and government policy.Accordingly,taking into account indicator relevance and data availability,we selected several indicators from these three dimensions(Table 1).For the geographical environmental dimension,two factors,location and resource condition,were considered and measured by the location advantage index(LAI)and the resource potential index(RPI),respectively;for the market condition and investment dimension,three factors,labor cost,consumption capacity,and investment capacity,were selected and measured by annual average wage of employees(AAWE),per capita retail sales of consumer goods(PCRSCG),and ratio of public expenditures to income(RPEI),respectively;for the government policy dimension,administrative resources and distinctive policy were the main factors and were measured by the administrative-level index(ALI)and the policy advantage index(PAI),respectively.Among the seven indicators,AAWE,PCRSCG,and RPEI can be derived directly from the statistical yearbooks,while the others must be obtained through integrative calculation.It is undeniable that regional and sectorial collaboration,cultural environment,and other social factors also had an important impact on the urbanization pattern.Due to the difficulty in quantification,they were not included in the index system,but they were reflected in the discussion part.

2.3.3.1.Resource potential index(RPI).Natural resources are an important foundation for regional urbanization and socio-economic development.The regional differences in natural resource endowment will lead certain resource-dependent industries or functions to be concentrated in specific regions,thereby promoting the formation of regional differentiation patterns.Land use capability(LUC)and landscape resource advantage(LRA)are two important indicators that reflect the potential of resources.LUC was used to account for the potential of land development for agriculture,industry,housing,and so on.It is impacted by several factors,such as topography,soil,and hydrology.In this research,considering that topography is the most important factor in urban-rural development and construction,we evaluated LUC using topographical data based on the best available data sources.According to the studies on the suitability evaluation of urban construction land(Tan,2016)and considering the mid-low mountainous and hilly landscape,we defined areas with a slope below 20°and an elevation below 600 m as suitable development areas and calculated the proportion of suitable development areas in the total land area as the LUC.The advantages of landscape resources depended mainly on natural resource characteristics,cultural heritage,and local products.Whether a town has won a reputation as a historical and cultural site,a famous tourist destination with special landscapes,or an area of geographical interest to a certain extent could reflect the advantages of landscape resources.The rule of assignment was as follows:a town nationally famed for its history and culture,as a tourism destination with special landscape resources,or as being particularly habitable received 2 points;a town provincially famous for its history and culture,as a tourism destination with special landscape resources,or as being particularly habitable received 1 point;and a nationally protected product with specific geographical indications received 2 points.Then,we summed the LUC and the LRA to obtain the RPI:

where RPI is the resource potential index;LUC is the land use capability;LRA is the landscape resource advantage;and α and β are the weights.Because LUC and LRA are equally important, they are assigned the same weight by using the subjective assignment method,i.e.,α=β=0.5.

2.3.3.2.Location advantage index(LAI).The location advantage of a unit is reflected in its potential and the degree to which it experiences spillovers from neighboring central urban areas.Therefore,we integrated the radiation force of the neighboring central urban area and the accessibility of the central urban area from the unit to calculate the LAI.

The influence of node cities reflects mainly their agglomeration and diffusion capacity,which not only are related to the comprehensive economic strength and market consumption potential but also depend on the flow of people,cargo,information,and so on.Therefore,we used GDP,the average annual salary of on-duty employees,the retail sales of consumer goods,and the added value of transportation,warehousing,and postal services to quantify the spillover based on the availability of data.The calculation formula is as follows:

where Ejrepresents the spillover effect of central urban area j;Virepresents the influencing factors of Ej;and αiis the weight of the factor.In this formula,the weight of each factor αiis calculated by the entropy weight method.The weight of GDP is 0.2641;the weight of the average annual salary of on-duty employees is 0.0666;the weight of the retail sales of consumer goods is 0.3376;and the weight of the added value of transportation,warehousing,and postal services is 0.3317.The calculation results are shown in Table 2.

Table 1Influence factors and indicators of the spatial patterns of populations and firms.

Table 2Calculation results of the spillover force(Ej).

Accessibility accounted for the convenience and efficiency of spillover from the central urban areas to the units and included two components:traffic accessibility and administrative accessibility.Traffic accessibility was measured by the inverse of the shortest time and distance between the central urban area and the unit and was derived from Amap.Administrative accessibility reflected the influences of the administrative division and regional policy on the spillover effect.It was defined by subjective assignment based mainly on the administrative span.

In urban agglomerations,one unit is always affected by overlapping radiation from multiple central urban areas,so we calculated the LAI with the following equation:

where Tijand Aijrepresent traffic accessibility and administrative accessibility,respectively,between central urban area j and unit i.

Note that because the urban district of Changsha City is the first-level center of Chang-Zhu-Tan,we assumed that the spillover effect occurs internally and set Tijand Aijboth to 1.0;the spillover effect of the urban districts of Xiangtan City and Zhuzhou City were mainly came from the urban districts of Changsha City,so Aijwas set to 0.8;the county towns and the urban districts of county-level cities mainly received spillover from the urban districts of three prefecture-level cities,so Aijwas set to 1.0 in the same prefecture-level city and 0.7 otherwise;the townships(towns)obtained multi-level spillover from the urban districts of the three prefecture-level cities,so Aijwas set to 0.8 in the same prefecture-level city and 0.6otherwise;and Aijwas set to 1.0 for the county town(urban district)of the located county(county-level city).

2.3.3.3.Administrative-level index(ALI).Cities and towns with distinct administrative hierarchies in China are divided into five levels:provincial capital,general prefecture-level city,county-level city,county,and township.The administrative hierarchy affects the allocation of urban resources(Li et al.,2016),and the size and growth of cities and towns are closely related to the administrative level(Wei,2014;Qin and Liu,2016).The administrative hierarchy index was used mainly to reflect the priority of resource allocation obtained by regional units under the influence of administrative hierarchy.We obtained the ALI by subjective assignment based on the government level to which it was subordinate.According to the five-level administrative management system,the ALI was set to 5 for units subordinate to the urban districts of the provincial city(Changsha City),4 for units subordinate to the urban districts of the non-provincial prefecture-level cities(Zhuzhou City and Xiangtan City),3 for units located in the urban districts of county-level cities,2 for units located in county towns,1 for units located in other townships,and 0 for units located in other towns.

2.3.3.4.Policy advantage index(PAI).China is in the transition period of institutional reform and policy innovation,and the adjustment of various institutions and policies will have an impact on the urbanization pattern and process.The PAI was used mainly to reflect national or provincial policy support for regional units.To assess the influence of policy,we primarily considered the demonstration pilot for building a resource-saving and environmentally friendly society(two-type society),the development of Hunan Xiangjiang New Area,and the national-or provincial-level development parks.The units covered by national policies were assigned 2 points,and those covered by provincial policies were assigned 1 point.The values of the PAI were determined by the cumulative score.

3.Results

3.1.Agglomeration characteristics of the urban population

Statistical data showed that the urban population increased yearly in 2005–2015.The urban population of Chang-Zhu-Tan was 6.19×106in 2005,representing approximately 48% of the total population in this region and reached 9.66×106in 2015,approximately 68% of the total population.According to the maps shown in Fig.2,the urban population distribution was characterized by spatial concentration.Most of the population was clustered in the north-central part of Chang-Zhu-Tan,especially in the urban districts of the three prefecture-level cities(i.e.,Changsha City,Zhuzhou City,and Xiangtan City)and some county-level cities(i.e.,Liuyang City and Liling City)(He et al.,2019).Moreover,from 2005 to 2015(Fig.3),the urban population of 312 towns(sub-districts)increased,mainly in towns located in the urban areas of prefecture-level cities and county-level cities as well as some county towns.In contrast,the urban population declined in 43 towns located mainly in suburban areas(sub-districts).

Fig.2.Spatial distribution of urban populations in Chang-Zhu-Tan in 2005(a)and 2015(b).

Fig.3.Spatial change in urban populations in Chang-Zhu-Tan from 2005 to 2015.

Fig.4.Kernel estimation results of manufacturing(MF)firms in 2005(a)and 2015(b).

3.2.Clustering characteristics of firms

According to the incomplete collected data,both types of firms more than doubled in number from 2005 to 2015 in Chang-Zhu-Tan.The number of MF firms rose from 2556 in 2005–6022 in 2015,and the number of HS firms increased from 1502 in 2005–3962 in 2015.The resultant kernel maps(Figs.4 and 5)provided evidence of firm clustering.The largest clusters of both types of firms occurred in the central urban areas,especially in the urban districts of Changsha City,Zhuzhou City,and Xiangtan City,and these grew over time.By comparing the kernel maps,we found that there was a spatial heterogeneity:the MF firms clustered mainly in urban areas,a decentralizing trend towards the city suburbs appeared,and HS firms showed the highest affiliation with urban centers.

Fig.5.Kernel estimation results of high-end services(HS)firms in 2005(a)and 2015(b).

3.3.GeoDetector of urban population and firms

To explore the driving forces of urban population agglomeration and firm clustering,we used the seven indicators listed in Table 1.Because GeoDetector is always valid for categorical data,we classified five indicators(i.e.,RPI,LAI,AAWE,PCRSCG,and PAI)into five grade zones by natural break classification using ArcGIS,while classified the other two indicators(i.e.,RPEI and ALI)into two grade zones based on professional experience(Fig.6).Moreover,to quantitatively test the inter-relation between population and firms,we added the number of firms(NF)as a factor in the geographical process of detecting the urban population,which was also classified by natural break classification.

The results of the factor detector regarding urban population agglomeration(Table 3)showed that in terms of the q value,the indicators can be sorted as follows:ALI(0.630)>NF(0.334)>PCRSCG(0.333)>LAI(0.280)>RPEI(0.180)>AAWE(0.160)>RPI(0.080)>PAI(0.060).According to the results of the interaction detector(Table 4),the q value of every pair of indicators of urban population distribution was greater than that of any single indicator,particularly in the interactions of PCRSCG and PAI,RPEI and NF,ALI and RPI,ALI and PAI(Xu et al.,2018).

Two types of firm were detected separately.According to the results of the factor detector for MF firms(Table 5),the q values of the indicators were ranked as follows:ALI(0.26)>LAI(0.17)>PCRSCG(0.15)>AAWE(0.12)>RPEI(0.10)>RPI(0.06)>PAI(0.05).The q values of the factors for HS firms were ranked as follows:ALI(0.44)>PCRSCG(0.42)>LAI(0.39)>AAWE(0.34)>RPEI(0.28)>PAI(0.10)>RPI(0.07).In addition,according to the results of the interaction detector(Tables 6 and 7),we calculated a total of 21 pairs to capture the interactions between any two of the seven factors for every kind of enterprise.The q values of every indicator were significantly heightened after the interaction.

Table 3The q values of the factor detector for urban population agglomeration.

Table 4The q values of the interaction detector for urban population agglomeration.

Table 5The q values of the factor detector for manufacturing(MF)and high-end services(HS)firms.

Table 6The q values of the interaction detector for MF firms.

Table 7The q values of the interaction detector for HS firms.

4.Discussion

4.1.Continuous centralization trend of urbanization

Urbanization is a process of continuous population migration to cities and concentrated development of non-agricultural industries(manufacturing and service industries)(Gu,2019).Based on the analysis results of the spatial pattern of the urban population and firms in 2005 and 2015,we can deduce the spatial organization pattern of urbanization and its dynamic changes in Chang-Zhu-Tan.The centralization trend in the spatial organization of urbanization was evident and prefecture-level cities,especially Changsha City,played an important role as the main centers(He et al.,2019).However,the role of small towns in urbanization was not fully realized;thus,the diffusion effect was still limited to rural areas(He et al.,2019).That is,decentralization coexisted with centralization,but centralization was still the leading trend in this mid-level developing urban agglomeration.

4.2.Administrative resources as the most critical factor

The q value calculated by the factor detector indicated the relative importance of the driving forces of urban population agglomeration and firm clustering.Administrative resource is the strongest explanatory factor for both of them,as its indicators obtained the maximum q values in the factor detector.The administrative hierarchy explained 68% of the distribution of the urban population,44% of HS firms,and 24% of MF firms.This finding is in line with “urbanization with Chinese characteristics”under the current hierarchical administrative system(Chen et al.,2010).Corresponding to the local administrative hierarchy,most of the urban agglomerations cover four levels of urban administrative units,i.e.,prefecture-level cities,county-level cities,county towns,and general towns;a few may consist of five or six levels additionally covering provincial-level cities and deputy provincial-level cities. The higher-ranking urban units are more likely to attract investments and obtain more opportunities for economic development(Chen et al.,2010;Chen and Partridge,2013)making them eligible for the allocation of various public resources.Therefore,there is always sufficient capital,adequate infrastructure,convenient facilities,high-quality social services,and rich recreation resources in prefecture-level cities,especially provincial cities such as Changsha City,compared with county-level cities as well as in county towns compared with other small towns and villages.All of these are the main factors attracting population and firms and distributing them in a spatially heterogeneous manner under the administrative hierarchy.

4.3.Surprisingly limited influences of local policy

The role of distinctive policies,reflecting the intervention of government in socio-economic development,cannot be ignored in the process of urbanization.However,the results for the detected q values showed the least influence on the urban population and firms in Chang-Zhu-Tan.The results of distinctive policy impacts were not as beneficial as expected(Lai,2008;Liu,2008).The policies,regardless of whether they encompassed a demonstration pilot for building a “two-type”society,the development of a national Hunan Xiangjiang New Area or the construction of development parks,played a limited role in driving the agglomeration of the urban population and firms.One reason may be the lag effect(Zhang et al.,2013);another reason,which is possibly the main reason,is the low effectiveness of policy-making and implementation.However,whether policies truly work on resource and environmental protection in urbanization must be further examined and confirmed.

Fig.6.Classification results of the indicators. (a), RPI (resource potential index); (b), LAI (location advantage index); (c), PAI (policy advantage index); (d), NF (number of firms); (e), ALI (administrative level index); (f), RPEI (ratio of public expenditures to income); (g), PCRSCG (per capita retail sales of consumer goods); (h), AAWE (annual average wage of employees).

4.4.Interaction of multiple factors to the urbanization pattern

The quantitative expression of the interaction detector illustrated that urban population agglomeration and firm clustering were affected by several interacting factors.First,not only administrative resources but also locational advantages,labor cost,and consumption capacity were important drivers.As consumption capacity always reflected the quality of life and market potential,a high quality of life indicated that a location has become more attractive to people,while market potential was the most important factor for the location choice of firms.According to the q value of PCRSCG,consumption capacity was the second most important factor in HS firm clustering(q=0.42)and the third most important factor in urban population agglomeration(q=0.28)and MF firm clustering(q=0.15).In general,the better the location advantage is,the more development opportunities will be obtained so that people and firms will tend to concentrate in areas with more location advantages,such as urban centers or suburbs.Thus,the location advantage was the second most important factor in MF firm clustering,the third most important factor in HS firm clustering,and the fourth most important factor in urban population agglomeration,with q values of 0.17,0.39,and 0.28,respectively.In addition,labor cost was the fourth most important factor for all types of firms because it was the main component of production cost.Second,the results of the interaction detector also showed that factors did not operate independently.The pairwise interaction between factors significantly enhanced the influence of each factor,especially when the administrative hierarchy interacted with other factors,such as consumption capacity,investment capacity,labor cost,and distinctive policy,even when the q values of population agglomeration reached 0.75,0.73,0.73,and 0.72,respectively.

Above all,we not only examined the impact to analyze which factors had a greater impact on urban population agglomeration and firm clustering but also detected the interaction between urban population and firms.We found that population agglomeration and firm clustering were interactive and inter-related processes.The spatial distribution of firms indicated the spatial supply of employment and the spatial demand for labor,which directly affected the migration of the urban-rural population;while the spatial distribution of the population determined the market potential and the supply of labor,which directly affected the location of firms.This finding is consistent with the detection results in which the NF factor had the second greatest impact,at 0.334,on urban population agglomeration,and it was greatly improved when interacting with other factors.Indirectly,consumption capacity and labor salary,which are bound up with urban population,also influenced firms with a high q value.On the whole,none of these factors played a decisive role in population agglomeration and firm clustering;it was their resultant force that promoted spatial agglomeration.

4.5.Implications for promoting healthy urbanization

From the above discussion,we not only understood the urbanization pattern and its drivers but more importantly discovered the critical problems in the urbanization process of the mid-level urban agglomeration.The results implied that the small towns and villages had not received enough attention,as the administrative resources played a critical role in urbanization,leading directly to investment inequality,and the local policy effect was not as positive as expected.These findings partly explained the reasons that the area is not developing more rapidly.These findings have important implications for the healthy development and management of urbanization in the mid-level urban agglomerations.First,local policies should be designed to improve the driving effects on rural areas in urbanization.The leading centralization trend also implies that there is insufficient agglomeration in small towns and a relatively lower spillover effect on rural areas,partly because of policy biases(Chan,2010).However,the sustainability of urbanization is definitely based on the synergetic development of urban and rural areas.Therefore,the local government should take specific measures to improve urbanization in rural areas,including facilitating industrial development in small towns and rural areas and guiding infrastructure and public service resources to favor small towns and rural areas to improve their comprehensive service ability.Second,the decision-maker should alleviate the inequalities of urban areas because of the administrative hierarchy.According to the factor analysis,the flow and allocation of capital,information,and technology are still subject to the administrative hierarchy;thus,many urban areas at the lower level cannot develop to their full capacity.This may even produce social problems because of inequity.Therefore,local governments should pay more attention to the equality of public resource allocation across urban areas at all levels,including transportation,education,health care,and investment policies.Bottom-up urbanization strategies such as public scrutiny and participation should work together with top-down policies(Li et al.,2018).Additionally,it is necessary to break down administrative boundaries and limitations to promote the in-depth cross-administrative collaboration in advance.Changing the constraining structures of local governments in the hierarchical system would be a fundamental solution for sustainable urbanization in the future(Li et al.,2015).Third,local governments should objectively assess the effectiveness of policies.To avoid the unnecessary waste of public resources and possible market disruption,it is necessary to sustain more effective policies and adjust less effective policies over time.The ultimate goal of all policies should be to increase employment,improve quality of life,and achieve sustainable development.Most importantly,all decisions should be based on the consideration of multiple influences of various factors,such as population agglomeration,firm clustering,location condition,labor cost,and consumption capacity.

5.Conclusions

Urban agglomeration has become the main form of regional spatial organization in the urbanization of China.The distribution of populations and firms is an effective representation of urbanization patterns in urban agglomerations.To understand the process and mechanisms of urbanization,this paper explores the features and factors influencing population and firm distribution.Based on the example of the Chang-Zhu-Tan urban agglomeration,the research results showed that centralization trends of the urban population and firms were obvious during 2005–2015,mainly leading to clusters in the central urban areas of Changsha City,Zhuzhou City,and Xiangtan City,and that a trend of decentralizing to the city suburbs had occurred.We also found that spatial agglomeration was promoted by forces resulting from several factors,such as administrative resources,location advantage,labor cost,and consumption capacity,which enhance each other.In particular,our study revealed that population agglomeration and firm clustering were interactive and inter-related processes.Moreover,we found that the factor with the most explanatory power was administrative resources,so the order of agglomeration sizes was absolutely in line with the hierarchy of prefecture-level cities,county-level cities,and county towns.This finding indicated that the integrated development of urban agglomerations was still constrained by the administrative system.Surprisingly,policies,regardless of whether they addressed the demonstration pilot of building a “two-type”society,the development of Hunan Xiangjiang New Area or the construction of development parks,played a limited role in driving the agglomeration of urban populations and firms.This result implied that some problems existed in policy-making and implementation.Therefore,to improve healthy urbanization,our study suggested that governments should pay more attention to the effectiveness of policies,the equality of public resource allocation,in-depth collaboration to break down administrative boundaries,and similar issues.

In general,this study took the typical Chang-Zhu-Tan urban agglomeration as an example to discuss the characteristics and driving forces of urbanization patterns.Its main research contributions were twofold:first,it paid attention to the relative lag of urbanization development in mid-level developing urban agglomerations,determines the core administrative factors,and puts forward targeted suggestions;second,it used the data of urban populations and non-agricultural firms to interpret urbanization from the perspectives of both industrial urbanization and population urbanization and their interaction to reveal the urbanization pattern and dynamics more objectively.We sincerely hope that this study will make a valuable contribution to understanding the process of urbanization in urban agglomerations at a mid-level of development to improve the health and sustainability of urbanization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Date availability statement

Some data used during the study were provided by a third party(The economic and social statistics statements of townships and the basic information database of legal entities).Direct requests for these materials may be made to the provider,as indicated in the acknowledgements.

Acknowledgements

This research was supported by the National Natural Science Foundation of China(41301192),the Natural Science Foundation of Hunan Province(2020JJ4056),and the Key Project of Education Department of Hunan Province(19A333).The authors would also like to thank the Hunan Provincial Bureau of Statistics and Development Research Center of the People’s Government of Hunan Province for providing the economic and social statistics and firm data.