Research Articles

Path dependence or path creation of mature resource-based cities: A new firm entry perspective

  • SUN Huijuan , 1, 2 ,
  • MA Li , 1, 2, * ,
  • JIN Fengjun 1, 2 ,
  • HUANG Yujin 1, 2
  • 1. Key Laboratory of Regional Sustainable Development Modeling, Institute of Geographic Sciences and Natural Resources Research, CAS; Beijing 100101, China
  • 2. College of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100049, China
*Ma Li (1975-), Associate Professor, specialized in economic geography and regional development. E-mail:

Sun Huijuan (1993-), PhD, specialized in economic geography and regional development. E-mail:

Received date: 2023-04-10

  Accepted date: 2023-11-02

  Online published: 2024-04-24

Supported by

National Natural Science Foundation of China(72050001)


Firm entry plays an important role in the industrial transformation of mature resource-based cities. This study describes the industrial evolution of resource-based cities at the firm level and uses kernel density estimation and econometric models to study the spatiotemporal characteristics and determinants of new firm entry from 2011 to 2019 in four mature resource-based cities. The results are summarized as follows: (1) New resource-based firm entry tends to be natural resource-oriented and path-dependent. The new non-resource- based firms show a high concentration in central urban areas, and the industry types are mainly wholesale and retail of resource products, cultural tourism, and equipment manufacturing. (2) Heterogeneous incumbent firms affect firm entry differently. Affected by competition and agglomeration effects, resource-based and non-resource-based incumbent firms have negative and positive impacts on new resource-based firm entry, respectively. Resource- based incumbent firm agglomeration positively influences new non-resource-based firm entry. (3) Besides incumbent firms, firm entry can also be affected by multidimensional factors, such as factor costs, economic environment, and institutional environment. Research on new firm entry can better reveal the path dependence and path creation process of the industrial development of resource-based cities from a micro-perspective.

Cite this article

SUN Huijuan , MA Li , JIN Fengjun , HUANG Yujin . Path dependence or path creation of mature resource-based cities: A new firm entry perspective[J]. Journal of Geographical Sciences, 2024 , 34(3) : 499 -526 . DOI: 10.1007/s11442-024-2215-1

1 Introduction

Resource-based cities (RBCs) are cities where industrial development and employment is highly dependent on resource mining and processing such as ore and lumber. China has 262 resource-based cities. These cities have made outstanding contributions to the economic and social perspective, but have become problematic areas, with complex contradictions in various aspects when local resources are exhausted (Jin et al., 2018; Yu et al., 2019; Li et al., 2021). In 2013, the General Office of the State Council of the People's Republic of China issued the National Sustainable Development Plan for Resource-Based Cities, which aimed to solve historical problems, accelerate the transformation and development of resource- exhausted cities, improve their comprehensive service functions, and promote their sustainable development. The entry of new firms that can bring new products and technologies to the region plays a major role in breaking the dependence on resource-based industries and promoting industrial transformation and development in RBCs. Understanding the types and paths of new firms in RBCs is important for regional industrial transformation. On the one hand, the dynamic changes in the type, composition, and quantity of new firms can not only observe the evolution of principal industries at the micro level, but also identify potential paths in the initial stage of new industries. On the other hand, studies have found that new establishments are the main agents of regional new development path renewal and creation (Neffke et al., 2018), and can significantly reduce the path dependence effect of RBCs. New products, skills, and ideas brought about by new firms can usher novelty in the local region, which is likely to recombine with or diversify the local knowledge base, and is conducive to the creation of new regional development paths (Zhou et al., 2017a).
The spatial differences in firm entry and their influencing factors are among the most important topics in economic geographic research. Based on different theoretical perspectives, scholars have discussed the influencing factors of spatial differences in the firm entry from neoclassical, behavioral, institutional, and evolutionary perspective, including market demand, labor costs, transportation costs, agglomeration effects, entrepreneurship, formal and informal institutions, firm routines and market choices (Shi and He, 2014). In empirical research, scholars have focused on firm entry in macro-industries of different spatial scales (country, city, and community) or certain industries (apparel, automobiles, biotechnology, and metal products) and discussed the effects of agglomeration economies, economic openness, institutional environment, labor cost, and firm characteristics on firm entry (Zhu and He, 2014; Frenken et al. 2015; Shi and He, 2018). However, little attention has been paid to the spatial differences and factors influencing new firm entry in RBCs. A new firm's location dynamics in RBCs is unique. Studies have found that new resource-based firms are highly dependent on natural resource endowments and are dominated by large and medium-sized state-owned firms (Mao et al., 2015); new firm entry in other industries may be squeezed out by resource-based industries and firms (Cheng et al., 2021); and new firms may be resisted by large resource-based incumbents or may branch out and recombine resources from existing local industries and firms to which they are technologically related (Grabher, 1993; Boschma, 2015). In general, existing research focuses on the location dynamics of new resource-based firms, whereas the location selection of non-resource firms and the relationship between heterogeneous firms needs to be further studied.
To better understand how RBCs break path dependence and develop new growth paths, this study focuses on new firms, the main drivers of regional industrial changes, and discusses the spatial characteristics and driving factors of the entry of new resource-based and non-resource-based firms into RBCs. This can help to better understand the industrial evolution of RBCs at the micro-enterprise level and provide guidance and suggestions for new firm entry in the study area. First, we analyze the industrial development and evolution of RBCs at the firm level and pay particular attention to new firm entries and their impact on regional industrial development. Second, we focus on the two types of new firms and their impact on the industrial evolution of RBCs. We identify two new firm types: new resource-based firms (NRFs) and new non-resource-based firms (NNRFs). The entry of the former can maintain and continue the existing resource-based development path, and the dynamic agglomeration of the latter can promote new growth paths. Further, this study describes the spatial differences and evolutionary characteristics of different types of new firms and discusses the differential effects of incumbent firm agglomeration, institutional environment, and macroeconomic context on different types of new firms, with the example of typical areas.
This study contributes to the literature in two ways. First, to fill the gap in the little attention paid to micro-firms in existing RBC studies, we constructed a theoretical framework for the industrial evolution of RBCs at the firm level and used this framework to explore the differential effects of different types of new firms on regional industrial changes. We proposed that NRF entry can strengthen and renew the resource industry growth path and that NNRF entry may create new growth paths. Second, the resource curse literature mainly focuses on how the natural resource industry or coal firms affect other industries in the economy. However, little attention has been paid to the influence of incumbent firms on new firms, especially the differential impact of resource-based and non-resource-based incumbent firms on NRFs and NNRFs. Therefore, we focused on how the agglomeration of incumbent firms in different industry types affects new firm entry to explore the interactive development of old and new industries in the industrial transformation of RBCs.
The rest of the study is organized as follows: Section 2 constructs a theoretical framework to discuss the industrial evolution of RBCs at the firm level and the differential impact of the different types of new firms on industrial evolution. Section 3 introduces basic information on the study area, data sources, and research methodology. We classified new firms (NFs) that enter resource-based cities into two types: NRFs and NNRFs. Section 4 discusses the spatiotemporal evolution characteristics of the entries of the NFs, NRFs, and NNRFs. It empirically analyzes the spatiotemporal evolution and key influencing factors of NRFs and NNRFs. Section 5 discusses the results, draws conclusions, and discusses some unresolved issues.

2 Theoretical framework

2.1 Industrial evolution and new firm entry of RBCs

Path dependence is an important feature of economic landscapes, particularly in resource-based regions. Based on the canonical path dependence model and the recently developed adaptive cycle model (David, 1985; Arthur, 1988; Martin, 2010; Martin and Sunley, 2011), we can describe regional industry evolution in RCBs as a four-phase journay and analyze firm types and population change, firm entry and exit behaviors, and their impact on local industries. Further, from the perspective of firm entry, we discussed how heterogeneous new firm entry plays an important role in the development of industrial diversification paths in RBCs as well as the important factors affecting new firm entry (Figure 1).
Figure 1 Path dependence and path creation analysis framework of resource-based cities
From the perspective of path dependence, with the industrial evolution of RBCs, the number and types of firms change continuously over time. (1) Path creation phase. In the initial stage of regional development, RBCs tend to take advantage of local natural resources (coal, oil, and forestry products) to develop resource mining industries. Some resource- based firms, cultivated locally or implanted from outside, have emerged and gathered. (2) Path development phase. Owing to the buildup of a specialized labor pool, local knowledge spillovers, and the division of labor among local firms (Martin, 2010), the spin-off dynamics and agglomeration economies (Boschma and Martin, 2007) of local resource-based firms promote the increase of the number of firms in resource-based sectors. Only a few NNRFs have emerged, belonging to the industrial support sectors (transportation and electricity) and consumer industries (building materials, agriculture, and sideline food processing). (3) Path-dependent lock-in phase. As local nonrenewable resources dwindle, the drawbacks of industrial overspecialization begin to appear, and regional development turns from positive path dependence to negative path lock-in. During this period, the composition of the firms’ population changed significantly. On the one hand, competition among resource- based incumbents intensified, and weak competitive firms either exited the industry or relocated to other regions. On the other hand, functional, cognitive, and political lock-ins of surviving resource-based incumbent firms and local governments can hinder a new firm's entry (Grabher, 1993; Hassink, 2007). (4) Path delocking phase. The depletion of natural resources and shrinking demand for energy and raw material products has led to a recession in the resource industry. Most extractive industry firms are either closed or reorganized. Surviving state-owned resource firms upgrade their technology, develop a fine chemical industry, undertake industrial transfers, and develop alternative non-resource-based industries. Therefore, the single-resource industry's development path has begun to disintegrate.
The canonical path dependence model describes a standard regional industrial evolution path, but Martin and Sunley (2011) emphasized the unpredictability and multiple potential trajectories of path development. They introduced the adaptive cycle model of complex systems theory into the evolution of industrial clusters and pointed out that business clusters are a composite system or population of entities and that the population and characteristics (technologies, products, and routines) of new and existing firms affect the evolution direction of the cluster. The concept of adaptive cycles provides interesting insights into the industrial evolution of RBCs.
First, RBCs can be more active before the decline of dominant industries, which can promote the diversification of regional industrial trajectories by supporting or introducing new firms, especially in the path-dependent lock-in and path delocking phases. However, during the path-dependent lock-in phase, new firm entry and new industry development may be impeded and restricted in several aspects: strong social networks, resulting in indifference toward fundamental renewal and opposition to newcomers; an excess of cognition between local firms, leading to a potential reduction in firm innovation, and even knowledge and technology lock-in (Broekel and Boschma, 2012); and institutional sclerosis and hysteresis resulting from relationships between mining firms and local governments, opposition from institutional conservatives, and the lack of the local government's industrial transformation motivation, impeding the development of new institutions to support the growth of new industries and new firms (Boschma, 2015). In the path delocking phase, faced with various problems in regional development, the central government supported resource-exhausted cities in fostering alternative industries by implementing special fiscal transfer payments, tax reductions, and interest subsidies. However, new industries develop slowly because of the state-owned ownership structure, weak industrial base, and distance from economic development centers. Limited financial revenue and official promotion incentives cause local governments to focus on short-term benefits, introduce many new firms, and build industrial parks. They pay less attention to soft environments, such as innovation platforms and labor pools, which are required for the sustainable development of new industries.
Second, the types of economic activities of new and existing firms determine the direction and extent of the regional industrial evolution, reminding us of the heterogeneity in the economic activities of new firms. In general, in resource-based regions, different types of new firms can induce different regional industry path changes at the firm level. The first type is NRFs, undertaking resource-processing industries closely related to natural resources and usually technologically related to pre-existing industries in RBCs. The second type consists of NNRFs, which engage in industries that are unrelated to resource-based industries. The NRFs can spur incremental and process innovation, provide the potential for inter-industry learning and new recombination, renewals, reinforcement of a regional specialized knowledge base, and promote local resource industries for further improvement. However, they frequently experience correlated demand shocks (Frenken and Boschma, 2007). The NNRFs, which are vigorously introduced and locally cultivated in resource-exhausted regions, are expected to facilitate radical innovations, recombine knowledge and technologies from cross-industries, lower regional unemployment rates, and protect regions from sudden sector-specific demand shocks. However, owing to the large cognitive gap within the local factor industry, NNRFs may face a disadvantageous position after entering the region, including (but not limited to) low levels of local knowledge, a mismatched local labor pool, difficulty accessing specialized suppliers (Zhu et al., 2021), and likely threats from resource-based incumbent firms (RIFs) and non-resource-based incumbent firms (NRIFs).

2.2 Agglomeration of incumbent firms and entry of new firms

The spatial agglomeration of incumbent firms is an important factor affecting firm entry. In the current literature, three main strands of research focus on the impact of spatial agglomeration on firm entry. The first strand of agglomeration economics literature has explored the impact of agglomeration within the same industry and across industries on firm entry. The Marshallian localization economies theory holds that labor sharing, intermediate input linkage, and knowledge spillover are brought about by the spatial agglomeration of incumbent firms in the same industry (Marshall, 1920). New firms can benefit from agglomeration effects, thereby reducing production risks and costs, improving innovation capabilities, promoting scale expansion, and increasing returns. Some industry studies have found that regional incumbent firms affect firm entry rates in industries, such as fashion design (Wenting and Frenken, 2011), hotel (Freedman and Kosová, 2012), corn ethanol (Thome and Lin Lawell, 2022), and creation (Arauzo-Carod et al., 2023). In addition, Jacobs’ externalities emphasize that a diverse industry mix can contribute to the diffusion and spillover of diverse knowledge, the sharing of infrastructure and public services, and the matching of labor and upstream and downstream firms, thereby promoting the establishment of new firms. For example, Zhang et al. (2020) found that industrial diversity has a positive impact on the entry of new firms in China's central region. Although the agglomeration economics literature emphasizes that MAR and Jacobs externalities promote the entry of new firms, relevant literature also points out that the congestion effect caused by excessive agglomeration, such as rising land and labor costs, traffic congestion, and environmental pollution, will also hinder the entry of new firms (Arauzo-Carod et al., 2005). The second strand of evolutionary economic geographic literature starts from the perspective of technological relatedness and argues that the more existing firms in a certain industry in a region, the more are the new entrants. First, new firms can be spin-offs, who inherit a large part of the capabilities of their parent firms, and tend to be located in the same region as the parent firms (Boschma and Wenting, 2007). Second, the incumbent firm's social legitimation effect can guide potential entrepreneurs to start the same type of business locally. Related studies refer to the positive effect of the number of incumbents on entry as a social legitimation effect, including cognitive and socio-political legitimacy, and argue that the more firms are active in a particular industry in a particular region, the more new firms will be created in the same industry and region (Frenken et al., 2015). The third strand of life-cycle literature focuses on the stage characteristics and industry heterogeneity of industrial development from a long-term perspective. The spatial product life cycle approach holds that in the early stage of industrial development, the entry rate of new firms is relatively high, but with product standardization and cost competition, new firms will spread from the core areas to the periphery with lower production costs (Klepper, 1996; Duranton and Puga, 2001; Neffke et al., 2011; Capasso et al., 2016). In general, more recent studies have focused on the positive effects of specialization, diversification, and related diversification on firm entry, which are mainly due to agglomeration effects, such as knowledge spillover, labor and infrastructure sharing, intermediate input linkages, market scale effects, and spinoff processes. However, they also pointed out that agglomeration diseconomies and competition effects may hinder the entry of new firms. These studies provide useful insights into the impact of spatial agglomeration on firms’ entry into RBCs.
The mechanism of spatial agglomeration in firm entry may have particularities in resource-based cities. This is due to the characteristics of resource-based firms, such as high dependence on resource endowment, low efficiency of resource and energy utilization, low technology level of primary processing, insufficient market competitiveness of some large state-owned mining firms, lack of technology and talent for the development of non- resource-based firms, and strong dependence of regional development on the resource economy (Wu et al., 2023). For example, some studies have shown that in regions dominated by traditional heavy industries, such as the Ruhr area and Saarland in Germany, the core firms of the regional coal and steel industries may block the settlement of new industries because of strong institutional lock-in (Grabher, 1993; Hudson, 1994). The relationship between the incumbents and new firms has also evolved. For example, Arbuthnott et al. (2010) qualitatively analyzed the relationship between emerging biorefinery industry actors and established actors from traditional but declining pulp and paper, chemical, and forestry industries from contestation to the complementary collaboration of Ornskoldsvik in Sweden, thereby jointly promoting regional industry renewal. Hu (2017) found that Zaozhuang entered a positive growth path in which new and old industries (resource industry and tourism) interact under the local government's innovation governance. Therefore, considering the differentiated impact of the spatial agglomeration of heterogeneous firms on firm entry in resource-based cities, it is necessary to distinguish between resource-based and non-resource- based firms, and then distinguish between resource-based and non-resource-based incumbent firms and new resource-based and non-resource-based firms.
The spatial agglomeration of heterogeneous incumbent firms in RBCs affects firm entry in several ways. First, two opposite effects of incumbent firms on firm entry in the same industry: an agglomeration effect and a competition effect. On the one hand, the high entry rate of incumbent firms has a demonstration effect, and new firms can benefit from the localization economies of incumbent firms and the transfer of knowledge and practices from selected parent firms. On the other hand, the competitive pressure exerted by new firms on incumbent companies or the strong ties formed by frequent interactions among local firms may make them reluctant to break existing interest structures and, therefore, reject new entrants. Resource-based firms in the same region may compete in the local raw material input market (raw coal and metal minerals) and downstream product markets (steel and building materials). Therefore, inter-firm competition keeps NRFs away from places where resource-based firms are highly concentrated. These competitive behaviors may be further intensified by the reduction of resources, downward energy prices, and strict environmental policies. Second, the agglomeration of incumbents from different industries has varying effects on firm entry. First, the diversification of non-resource-based incumbent firms may positively affect NRFs. This could affect in two ways: urbanization economy and market competition strategies. An urban economy with different industrial agglomerations can lead to a broader market and knowledge spillover. Furthermore, RBCs with more non-resource- based incumbent firms usually have a higher level of marketization or less competition from resource-based incumbent firms. When NRFs enter a region, they face lower entry barriers and less peer competition, or become regional industry-leading firms and obtain more local government subsidies with stronger bargaining power. Second, resource-based incumbent firms influence the NNRF entry. The development of the resource industry drives growth in the supporting industrial sector. However, because resource depletion leads to economic, social, and ecological issues, it is not conducive to the entry of NNRFs. In addition, the rapid development of the resource industry increases local fiscal revenue, which can be consciously used by the government to support non-resource industries.

3 Data and methodology

3.1 Study area

To investigate the characteristics and determinants of new firm entry in mature RBCs, we proposed a case study of the Yellow River Golden Triangle region, for several reasons:
First, the study area is located at the junction of Shanxi, Shaanxi, and Henan provinces and includes Yuncheng, Linfen, Sanmenxia, and Weinan, which are representative of RBCs in underdeveloped areas. The study area covered 57,800 km2 and had a total population of 17.4 million (Figure 2).
Figure 2 Study areas (Yellow River Golden Triangle region)
Second, the resource industries in the four cities have gradually become saturated and are in a critical period of industrial transformation. These four cities have a long history of exploitation and processing of coal and iron ore, and are important for national coal production. A group of large-scale energy and heavy chemical firms gathered here, including Shaanxi Heimao Coking Co. Ltd., and Shaanxi Longmen Iron and Steel (Group) Co. Ltd.
Third, the study area is the only national interprovincial industrial transfer demonstration area in China. In 2012, the Chinese National Development and Reform Commission approved the establishment of the Yellow River Golden Triangle Undertaking Industrial Transfer Demonstration Zone and increased policy support for finance, taxation, industry and investment, and environmental and ecological resource protection. By seizing the dividend policy and the opportunity for industrial transfer in China's eastern coastal areas, the four cities actively undertook non-ferrous metal processing, equipment manufacturing, biomedicine, new energy resource industries, and new materials industries, and introduced some large-scale firms, such as Baowu Aluminum Technology Co. Ltd., and EPOCH Technology IMECAS Co. Ltd., and cultivated a small number of Technologically Advanced Small- and Medium-sized Enterprises, such as Sanmenxia Hongxin New Material Technology Co. Ltd., and Shanxi Dayu Biological Functions Co. Ltd.
Additionally, considering the strategic timing of the demonstration zone approval in 2012 and the promulgation of the National Sustainable Development Plan for Resource-Based Cities in 2013, we selected the period 2011-2019 for our empirical study.

3.2 Data

3.2.1 Data source

The data in this study mainly included firm, spatial, and socioeconomic attributes. The firm data were obtained from the Qcc database (, which is a database that records all firms in the National Firm Credit Information Publicity System and provides firm-level information. In this study, we focused on the survival status of firms established between 2011-2019, and defined the surviving firms that were established before 2011 as incumbent firms and those established between January 1, 2011, and December 31, 2019, as new firms. The firm data acquisition and processing procedures were as follows: (1) the data were acquired on August 19, 2021, and firms were classified by industry according to the Industrial Classification for National Economic Activities (GB/T 4754-2017), and extracted from all 27024 firms with registered capital above 10 million yuan in Sanmenxia, Weinan, Yuncheng, and Linfen; (2) firm's address information was checked, and those missing were manually supplemented from the firms’ information on their official websites; (3) geographic coordinates were obtained via Baidu Map API interface based on the firm address information, and were transformed into spatial point data files in the four cities using ArcGIS 10.8 software.
For spatial data, the administrative boundary vector was primarily obtained from the National Platform for Common Geospatial Information Service, and river, highway, and railroad data were obtained from the 1:250,000 National Basic Geographic Database. Regarding the data on socioeconomic attributes, industry production value, gross domestic product (GDP) per capita, road network density, average wage of employees, fixed asset investment, and fiscal expenditure of each county these were sourced for the period 2011-2019, from the China County Statistical Yearbook and the statistical yearbooks of the four cities. Data on high-tech enterprises were acquired from the Network for the Administration of Recognition of High-tech Enterprises ( Provincial economic development zone and industrial park data were procured from the China Development Zone Audit Bulletin Catalogue (2018 edition) and the China Development Zones Network. The comprehensive utilization rates of industrial solid waste, centralized sewage treatment, and harmless domestic waste treatment were obtained from the China City Statistical Yearbook.

3.2.2 Firm classification

In this study, we divided all firms into resource-based and non-resource-based firms according to their industry type. According to the Industrial Classification for National Economic Activities (GB/T 4754-2017), industries closely related to natural resources are defined as resource based industries, such as mining and resource processing manufacturing. The mining industry includes coal mining and washing; petroleum and natural gas extraction; mining and processing of ferrous and nonferrous metal ores and nonmetal ores; mining of other ores; and ancillary mining activities. The resource processing manufacturing industry includes processing petroleum, coal, and fuels; smelting and processing ferrous and non- ferrous metals; and manufacturing metal and non-metallic mineral products. In total, the results obtained were 1477 NRFs and 25547 NNRFs (Table 1).
Table 1 Main industrial classification and number of extracted firms
Industry of NRFs Number Industry of NNRFs Number
Manufacture of non-metallic mineral products 728 Wholesale and retail trades 8152
Mining and washing of coal 157 Construction 3207
Manufacture of metal products 121 Agriculture 2976
Smelting and processing of non-ferrous metals 102 Real estate 2278
Mining and processing of ferrous metal ores 87 Leasing and commercial service 1915
Mining and processing of nonmetal ores 82 Manufacturing 1906
Mining of other ores 66 Scientific research and technology services 898
Processing of petroleum, coal, and other fuels 45 Utilities 834
Mining and processing of non-ferrous metal ores 33 Transport, storage, and postal services 715
Smelting and processing of ferrous metals 32 Administration of water, environment, and public facilities 588
Ancillary mining activities 19 Information transfer, software, and information technology services 551
Extraction of petroleum and natural gas 5 Residential services, repair, and other services 506
- - Financial revenue 419
- - Culture, sports, and entertainment 303
- - Accommodation and catering 299

3.3 Research method

3.3.1 Kernel density estimation analysis

The kernel density estimation analysis method was used to calculate the density of point or line elements in their surrounding areas and can visually characterize the degree of spatial agglomeration of a certain socio-economic activity. The calculation formula is as follows:
$f_{n}(x)=\frac{1}{n h} \sum_{i=1}^{n} K\left(\frac{x-x_{i}}{h}\right)$
where x1, …, xn are independent and identically distributed samples drawn from the overall population with the probability density function f. Additionally, K(.) is the kernel function, h is the broadband, n is the number of points in the broadband range, and x-xi is the distance from the estimated point x to the sample xi.

3.3.2 Panel regression analysis model

We used panel regression model to analyze the influencing factors, and the number of new entrants in 47 counties (cities and districts) in the study area, from 2011-2019, was taken as the dependent variable. Considering the lag in the current dependent variables for new firms, we selected them for one period. The hypothetical model was as follows:
$y_{i t}=\beta_{0}+\beta_{i} X_{i t-1}+\alpha_{i}+\lambda_{t}+\varepsilon_{i t}(i=1, \ldots, n ; t=2011, \ldots, 2019)$
where yit denotes the dependent variable, β0 is the intercept, βi denotes the coefficient of each independent variable, Xit-1 denotes the independent variable that affects the entry of new firms, αi denotes the region effect, λt denotes the year effect, and εit denotes the random disturbance term.

3.3.3 Indicator selection

We chose NRFs and NNRFs as dependent variables, and RIFs and NRIFs as core independent variables. Further, considering the complexity and diversity of the factors affecting firm entry, we selected factors of different dimensions, such as factor costs, economic environment, and institutional environment, as control variables. Finally, considering that existing studies have found that environmental regulation and industry innovation affect the agglomeration of local RIFs and NRIFs, and then produce a moderating effect between NRFs, NNRFs, and RIFs, NRIFs, we chose environmental regulation intensity and industry innovation as the moderator variables.
Independent variable. This study selected the number of incumbent firms to characterize the level of the agglomeration economy and used two indicators, resource-based and non- resource-based, to measure agglomeration economy indicators.
Moderator variables. Environmental regulations can, directly and indirectly, affect the firm entry. On the one hand, the environmental cost caused by environmental regulations is an important factor in the location selection of enterprises, especially polluting enterprises. Studies have found that improving environmental regulations reduces the number of new polluting enterprises (List et al., 2003; Zeng et al., 2023). On the other hand, environmental regulations indirectly affect new firms’ entry by influencing the dynamics of incumbent firms. Relevant theories include the Pollution Haven Hypothesis and Porter Hypothesis. The Pollution Haven Hypothesis argues that high-level environmental regulation forces high- polluting and high-energy-consuming firms to exit or relocate, while the Porter Hypothesis holds that environmental cost pressures encourage firms to innovate in the long run (Zhou et al., 2017b). Industrial innovation capability can also directly or indirectly affect new firms’ location choices. First, regions with strong industrial innovation capabilities have more specialized technical knowledge accumulation and strong knowledge spillovers. Thus, new firms can acquire specialized knowledge and talent by embedding them into local innovation networks to carry out innovation activities (Zheng and Shi, 2023). This phenomenon may be more prominent in RBCs that have relatively weak innovation capabilities. Second, a higher industry innovation capability can also improve the efficiency of resource- based and non-resource-based incumbent firms, eliminate firms with backward technology and low production efficiency, and promote the allocation of production factors between the resource and non-resource sectors, resulting in the entry of upstream and downstream firms in the mineral resources industry chain or the spin-off of new firms or industries (Zeng et al., 2021). Further, the environmental regulation intensity was measured using the average of the comprehensive utilization rate of general industrial solid waste, harmless treatment rate of domestic waste, and centralized treatment rate of sewage treatment plants in the four cities. The industrial innovation capabilities of resource and non-resource industries are measured by the number of resource and non-resource high-tech firms, respectively.
Control variables. From the perspective of factor cost, the variables mainly included labor and electricity costs to reflect the cost differences among regional firms. Employees’ wage levels generally reflect regional labor costs. Considering that the study area was dominated by heavy chemical industries and the Provincial Development and Reform Commission sets the electricity price, this price was selected to characterize the regional electricity costs of large-scale industries in each province. From an economic environment perspective, the variables include market potential, industrial structure, transportation conditions, and fixed asset investment. Market potential is measured by the per capita GDP, industrial structure is calculated as the proportion of the added value of the secondary industry in the GDP, transportation conditions are calculated as highway mileage per 100 square kilometers, and fixed asset investment is measured by the completed amount of regional fixed asset investment. From the perspective of the regional policy environment, we selected the proportions of state-owned economies and provincial development zones. Based on the availability of county-level data, we measured the proportion of state-owned economies using the proportion of employees in state-owned urban units in the urban population. We characterized the construction of development zones by identifying provincial development zones and industrial parks in each county. The definitions and descriptions of the variables are shown in Table 2, and their descriptive statistics are presented in Table 3.
Table 2 Definition and description of variables
Variable name Symbols Indicator description
Dependent variables Type of firms New Resource-based Firms NRF Number of new resource-based firm entries each year
New Non-Resource-based Firms NNRF Number of new non-resource-based firm entries each year
Independent variables Agglomeration economy Resource-based Incumbent Firms RIF Number of resource-based incumbent firms in the previous year
Non-Resource-based Incumbent Firms NRIF Number of non-resource-based incumbent firms in the previous year
Moderator variables Innovation impact Industrial innovation capability IIA Number of resource high-tech enterprises
Number of non-resource high-tech enterprises
Environmental regulation impact Environmental regulation intensity ER The comprehensive utilization rate of industrial solid waste, centralized treatment rate of sewage, harmless treatment rate of domestic waste
Control variables Factor cost Labor costs LC The average salary of on-the-job employees
Electricity costs EC Large-scale industrial electricity price
Economic environment Market potential MP GDP per capita
Industry structure IS The proportion of the added value of secondary industry and regional GDP
Transportation conditions TC Regional road network density
Fixed asset investment FAI Fixed assets investment
Provincial development zone PDZ Whether there are provincial economic development zones and industrial parks
(Yes=1, No=0)
Share of state-owned
SOE The proportion of employees in urban state-owned units in the urban population
Table 3 Descriptive statistics of independent variables
Variable Unit Minimum Maximum Mean Standard deviation
Resource-based incumbent firms pcs 0 146 31 26.37
Non-resource-based incumbent firms pcs 9 3069 289.43 415.43
Resource high-tech enterprises pcs 0 6 0.4 0.9
Non-resource high-tech enterprises pcs 0 26 1.96 3.65
Environmental regulation intensity % 62.62 98.56 80.63 9.01
Labor costs yuan 18389 72636 41301.85 11756.75
Electricity costs yuan/kWh 0.45 0.63 0.54 0.04
Market potential yuan 5857 111269 29704.49 19628.48
Industry structure % 8.5 88.90 47.17 19.16
Transportation conditions km /100 km2 45 246 116.08 40.50
Fixed asset investment 100 million yuan 3.08 751.41 98.04 104.97
Provincial development zone - 0 1 0.35 0.48
Share of state-owned economy % 2.57 26.06 11.04 4.58

4 Empirical results

4.1 Spatio-temporal evolution of new firm entry

In this study, we used the kernel density estimation analysis tool of ArcGIS 10.8 software, to analyze the point data of firm entry from 2011 to 2019 for three sub-periods: 2011-2013, 2014-2016, and 2017-2019.

4.1.1 Spatial agglomeration evolutionary characteristics of new firm entry

During 2011-2019, new firms mainly agglomerated in urban centers, demonstrating a significant difference between the central and peripheral regions (Figure 3). New firms formed three high-value zones in Yanhu district of Yuncheng city, Yaodu district of Linfen city, and Linwei district of Weinan city, each with 3213, 2868, and 2248 new firms, respectively, thereby accounting for 30.82% of all new firms in total. Among the cities, there is a prominent spatial imbalance in new firm entry in Linfen city, forming an attraction axis with Yaodu district as the core, whereas regions distant from the central city are less attractive to new firms. New entries into Yuncheng city formed a spatial pattern of high-value areas in Yanhu district, medium-value areas in Hejin city, and low-value areas in multiple counties.
Figure 3 Kernel density analysis of all new firm entry in the Yellow River Golden Triangle region
Over time, the new firm's entry shows gradual agglomeration to the urban center and southwest regions. During 2011-2013, the high-value areas for new firm entry were Yanhu district of Yuncheng city and Yaodu district of Linfen city, which formed a circular structure around the urban center, showing a significant core-periphery structure. During 2014-2016, although the number of new firms increased, the space contracted; hence, the degree of concentration increased further and some new agglomeration areas were formed. In 2017-2019, new firms were more concentrated in the southwest of the study area, presenting three agglomeration centers in Yanhu district of Yuncheng city, Linwei district of Weinan city, Yaodu district of Linfen city, and medium-value areas that expanded along the railway line in the southwest.

4.1.2 Spatial agglomeration evolutionary characteristics of new resource-based firms entry

The NRF entry shows strong features of resource and path dependence (Figure 4). On the one hand, areas dominated by the resource industry are the first choice for NRFs, such as Yanhu district of Yuncheng city, with modern chemicals and new materials; Hejin city of Yuncheng city, with coal-fired aluminum and coal-coke steel industry clusters; Hancheng city of Weinan, with coal-coke steel and electricity; and Lingbao city of Sanmenxia, with non-ferrous metal smelting and deep processing, attracting 96, 87, 83, and 72 NRFs, respectively. Furthermore, NRFs tend to develop downstream of the industrial chain and mainly manufacture nonmetallic mineral products, smelting, and processing of nonferrous metals. On the other hand, the low values of NRF entry occur mainly in resource-exhausted, resource-deficient, and economically underdeveloped areas, and the industry type is primarily the manufacturing of non-metallic mineral products, mining and processing of non-metallic ores, smelting and processing of non-ferrous metals, and mining and washing of coal. Most low-value areas are located in Linfen city, including Huozhou city (a resource-exhausted city), Houma city (an important transportation hub in southwestern Shanxi), and national-level and provincial-level poverty-stricken counties such as Yonghe, Fenxi, and Daning. For instance, Huozhou, a typical resource-oriented city, was identified as a resource-exhausted city in 2011. In recent years, although it has strongly promoted industrial adjustment and economic transformation, the process has been slow, demonstrating the difficult position of traditional coal firms and their weak inclination toward new firms. In other words, resource-exhausted areas often need external assistance because they tend to be forced to begin transforming only when the resources are exhausted and an economic recession occurs; however, an over-specialized knowledge base and heavy sunk cost causes them to have little endogenous industrial branching ability, making it difficult to achieve a path breakthrough.
Figure 4 Kernel density analysis of new resource-based firm entry in the Yellow River Golden Triangle region
Over time, the spatial agglomeration of the NRFs shows resource directedness and contagious diffusion, with different agglomeration centers moving toward different trajectories, such as strengthening, declining, and emerging. During 2011-2013, NRFs gathered in mineral resource-rich areas, including the extension of the Xiangning Coalfield in Hejin city; Longmen Town Coal Mine in Hancheng city; Huoxi Coalfield in Xiangfen, Puxian, and Hongtong counties of Yuncheng city; and the Huodong mining area in the Qinshui Coalfield. The industry types are primarily mining and washing of coal, manufacturing of non-metallic mineral products, and mining and processing of non-metal ores. During 2014-2016, the NRFs decreased in number and were spatially clustered in the southwest region. The agglomeration intensity further improved in Hancheng city in Weinan city, Hejin city, Yanhu district in Yuncheng city, and Hubin district in Sanmenxia city. However, the agglomeration level of NRFs decreased significantly in some areas dominated by the coal chemical industry, such as the Xiangfen, Puxian, and Hongtong counties of Linfen city. During 2017-2019, the agglomeration level of NRFs in Xianfen, Puxian, and Hongtong counties of Linfen city further declined, and they were no longer obvious agglomeration centers, whereas in Hupin district, Lingbao city, and Yima city of Sanmenxia city, the agglomeration level was further enhanced.

4.1.3 Spatial agglomeration evolutionary characteristics of new non-resource-based firm entry

The NNRFs are highly concentrated in city centers or regions with prospering resource industries, presenting path creation mainly in coal wholesale and retail, cultural tourism development, automobile manufacturing, and multi-path development in some areas (Figure 5). First, the high-value areas are Yanhu district in Yuncheng city, Yaodu district in Linfen city, and Linwei district in Weinan city, attracting 3118, 2816, and 2194 NNRFs, respectively. Second, the NNRF industry types include wholesale and retail (coal, coke, building materials, automobiles, real estate development, and operations), commercial services of cultural tourism development and project investment management, and a small number of manufacturing firms (general and special equipment manufacturing, the chemical industry, and automobile manufacturing). Notably, Yanhu district of Yuncheng city and Hancheng city of Weinan city both have more NNRFs and NRFs, but their development paths differ. Relying on provincial development zones and local firms, Yanhu district has attracted many new firms, including manufacturers of nonmetallic mineral products, general machinery, automobile and pharmaceutical, and cultural tourism firms. By contrast, relying on pre-existing steel, coal, and chemical industrial clusters and key firms, Hancheng city has upgraded its products and processes, extended the industrial chain, and attracted and cultivated a large number of new ceramics, energy, and materials firms.
Figure 5 Kernel density analysis of new non-resource-based firm entry in the Yellow River Golden Triangle region
Over time, the NNRFs shrank and gathered in central districts. During 2011-2013, the NNRFs were highly concentrated in Yanhu district of Yuncheng city, and there was also a certain degree of agglomeration in Yaodu district and Houma city of Linfen city, Hubin district of Sanmenxia city, and Linwei district of Weinan city. During 2014-2016, a new agglomerated center was formed in Yaodu district of Linfen city, and a contiguous agglomeration was formed in Linwei district, and counties of Pucheng and Dali. During 2017-2019, NNRFs’ agglomeration decreased in Yaodu district of Linfen city but was further enhanced in Yanhu district of Yuncheng city. Further, it increased slightly and spread to the northeast in Linwei district of Weinan city.

4.2 Influencing factors of new firm entry

First, the relevant variables were logged to avoid pseudo-regression and the explanatory variables were tested for multicollinearity. The variance inflation factor of each explanatory variable was less than 10 and there was no collinearity problem. The Hausman test was performed on the model, and the results showed that the fixed-effects model was superior to the random-effects model. The White test indicated the presence of heteroskedasticity, while the xtscc test highlighted the presence of a cross-sectional correlation. Therefore, the xtscc command was used for regression estimation to solve the heteroscedasticity and cross- sectional correlation problems in the panel data.

4.2.1 The influence of agglomeration factors

In Table 4, we ran our model for the NRFs and NNRFs samples, and performed a robustness test by changing the time periods and indexes. In Model 1 (Table 4), the agglomeration of resource-based incumbent firms has a negative and significant impact on NRF entry, whereas the diversification agglomeration of non-resource-based incumbent firms has a positive and significant impact at the 1% level, with a coefficient of 0.805, indicating that NRFs are more inclined to choose areas where diversified firm agglomeration occurs, rather than specialized firm agglomeration in RBCs. On the one hand, RIFs have a significant negative impact on NRFs, which indicates that the competition effect between RIFs and NRFs exceeds the positive externality agglomeration effect in mature resource-based cities. Considering the fierce competition effect of resource-based incumbent firms and the agglomeration diseconomy effect of resource industry development, NRFs tend to stay away from over-specialized areas. This phenomenon requires vigilance. When the high concentration and strong network connections between local RIFs hinder the entry of NRFs, if RIFs lack channels to knowledge from the outside region, they will further move toward cognitive, functional, and institutional locks, which will lead to a decline in the resource-based industry and cities. On the other hand, regions with more NRIFs attract more NRFs, which can be explained from the perspective of agglomeration economies and market competition. From the perspective of agglomeration economies, the diversified agglomeration of different industries can provide broader markets and opportunities for NRFs (Van and Suddle, 2008).
Table 4 Regression of factors influencing the NRFs and NNRFs during 2011-2019
Variables lnNRF lnNNRF
(1) (2)
lnRIF -0.634** 0.439***
(-2.81) (4.04)
lnNRIF 0.805*** -0.352
(3.59) (-1.58)
IIA 0.0533** -0.0137
(2.40) (-1.67)
ER -0.0292*** 0.00587***
(-11.35) (3.68)
lnLC -0.721 -0.0504
(-1.67) (-0.21)
lnEC -3.923*** -0.315
(-6.16) (-0.67)
lnMP 0.558*** 0.241
(3.96) (1.37)
lnIS -0.320** 0.189
(-2.66) (1.23)
lnTC -0.218 0.969*
(-0.19) (2.22)
lnFAI 0.187*** 0.0706
(3.99) (1.65)
PDZ 0.127 0.0116
(0.74) (0.19)
SOE 0.0208** -0.00309
(2.63) (-0.65)
Year dummies Yes Yes
Region dummies Yes Yes
Number of observations 423 423

Note: Standard errors are mentioned in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

From a market competition perspective, resource-based cities with NRIFs can be divided into two types: those with more RIFs and NRIFs and those with fewer RIFs and more NRIFs. The resource-based cities with more RIFs and NRIFs, which usually has a more active private and individual economy and a higher level of marketization, can alleviate the lock-in effect of resource industries, and promote the entry of non-resource industries (Zhu and Lin, 2022). Regarding those with fewer RIFs and more NRIFs, because there are fewer RIFs in the region, NRFs face less competitive pressure and are more likely to become leading local firms and obtain government subsidies. Therefore, although NRFs still need to enter resource industries because of resource dependence and technology lock-in, they are more inclined to choose areas with low agglomeration of resource-based incumbents (specialization agglomeration), that is, areas with diversification agglomeration.
To better understand how diversification agglomeration affects NRFs, we divided NRIFs into agricultural, manufacturing, utilities, transportation, and cultural tourism firms, based on the main transformation path of the resource-based city and the development situation of the study area. This study examines the different influences of NRIFs in different industries on the location choices of NRFs. The econometric results are presented in Table 5. The estimates show industry heterogeneity in the impact of NRIF agglomeration (Models 3-7). On the one hand, the agglomeration of incumbent agricultural firms has a positive and significant effect on NRFs. Generally, the areas where agricultural firms gather have a low level of industrialization and a large amount of rural surplus labor owing to low agricultural labor productivity and limited land resources. Meanwhile, subject to the early immature resource view and the unsustainable development concept, based on the advantages of local mineral resources used to develop resource industries, this has become one of the simplest and most direct means for resource-based cities to promote regional industrialization and urbanization. On the other hand, utilities incumbent firms have a negative effect. Considering that the supply of electricity and heat in the study area mainly comes from coal, and electricity consumption is closely related to high-energy-consuming industries, the location choice of NRFs should not only consider competition with resource-based incumbent firms regarding industry capabilities but also competition with utility incumbent firms in regional resources, such as fuel, and physical and human capital. Owing to this dual competitive pressure, NRF entry is inhibited.
Table 5 Regression of factors influencing the NRFs during 2011-2019
Variables lnNRF lnNRF lnNRF lnNRF lnNRF
(3) (4) (5) (6) (7)
lnRIF -0.634** -0.614** -0.683** -0.607** -0.625**
(-2.81) (-2.93) (-3.26) (-2.50) (-2.91)
lnNRIF_agriculture 0.384***
lnNRIF_manufacturing -0.219
lnNRIF_utilities -0.399**
lnNRIF_transportation 0.202
lnNRIF_cultural tourism 0.159
Year dummies Yes Yes Yes Yes Yes
Region dummies Yes Yes Yes Yes Yes
Number of observations 423 423 423 423 423

Note: Standard errors are mentioned in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

In Model 2 (Table 4), RIF agglomeration has a positive impact on NNRF entry, whereas NRIF has no significant impact. In particular, NRIF agglomeration has little impact on NNRF entry, which may be because NRIFs have not yet formed an agglomeration economy, or could be related to the diversified industry types of NRIFs. Thus, RIF agglomeration promotes NNRF entry. There are several possible explanations for these results. First, large state-owned resource-based firms can generate considerable local fiscal revenue, which can be used to cultivate and develop new regional industries, such as state-owned capital investment companies, and increase financial support for new industries. Second, RIF agglomeration can attract NNRFs through industrial chain extensions and ecological industrial chain construction, such as coal chemical firms and firms that utilize comprehensive waste resources. Third, under pressure from enterprise survival and the guidance of government policies, RIFs may cooperate with non-resource-based firms or create NNRFs independently to diversify into non-resource-based industries. However, due to a lack of prior experience in non-resource industries and the pressure of transformation, RIFs usually choose to enter industries with lower technical and capital thresholds.
To better analyze how the agglomeration economy affects the NNRF entry, we examined the effect of RIFs on the location choice of different types of NNRFs. Table 6 reports the econometric results (Models 8-12). The estimates show that resource-based incumbent firm agglomeration has a positive and significant impact on new manufacturing and cultural tourism firms, indicating that old industries (resource industries) and new industries (manufacturing and cultural tourism) can develop mutually during the process of industrial transformation in RBCs. In Model 10, resource-based incumbent firm agglomeration impedes new utility firms, which is consistent with the negative impact of incumbent utility firms on NRFs in Model 5, confirming a strong competition between them. Additionally, in Models 8-12, manufacturing, utilities, and transportation incumbent firms all hinder new firms in this sector, indicating that the demonstration effect of a high firm entry rate in new industry sectors is not significant, while the competitive pressure exerted by incumbents is more significant.
Table 6 Regression of factors influencing the NNRFs during 2011-2019
Variables lnNNRF_
cultural tourism
(8) (9) (10) (11) (12)
lnRIF 0.219 0.549*** -0.155** 0.0406 0.255**
(1.12) (8.53) (-2.33) (0.42) (3.05)
lnNRIF_agriculture -0.104
lnNRIF_manufacturing -0.516***
lnNRIF_utilities -0.302*
lnNRIF_transportation -0.369*
lnNRIF_cultural tourism -0.149
Year dummies Yes Yes Yes Yes Yes
Region dummies Yes Yes Yes Yes Yes
Number of observations 423 423 423 423 423

Note: Standard errors are mentioned in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

4.2.2 The influence of moderator variables

There are differences in the influences of industry innovation ability and environmental regulation intensity on the NRF and NNRF entries. In Models 1 and 2 (Table 4), industry innovation capability has a positive and significant impact on the NRF entry. A region with more resource-based high-tech enterprises may have more opportunities to update its local knowledge base, improve reorganization potential, and facilitate regional branching, thereby attracting more NRFs. Second, environmental regulation intensity has an important effect on the NRFs and NNRFs; however, industry heterogeneity exists. The environmental regulation intensity is significantly negative for NRF entry at the 1% level, indicating that stronger environmental regulation policies will restrict NRF entry. This may be because resource-based firms are pollution-intensive and prefer areas with less restrictive environmental policies, whereas strict policies have a greater impact on their development in the short term (Zou et al., 2022). On the contrary, environmental regulation intensity has a positive and significant impact on NNRF entry, indicating that enhancing the environmental regulation intensity can promote the entry of NNRFs, largely because places with high environmental regulation intensity will eliminate or crowd out the existing heavily polluting firms and restrict their entry through multiple means, such as setting entry thresholds and strengthening environmental supervision, thereby indirectly promoting the establishment and entry of NNRFs.
In addition, industry innovation capability and environmental regulation intensity can affect the entry of new firms by acting on incumbent firms. Therefore, this study adds the interaction terms of RIF, NRIF, industry innovation ability, and environmental regulation intensity to the regression model. The results are shown in Table 7. For NRF entry, the results showed that the interaction terms did not pass the significance tests. For NNRF entry, the coefficient of the lnRIF_IIA is -0.044 and has passed the significance test at the 1% level, indicating that the innovation capability of the resource industry can weaken the promotional effect of RIFs on NNRF entry. This is because RIFs can diversify into non-resource industries under competitive pressure through government promotions. However, if RIFs’ industry innovation ability is relatively strong, they will be more inclined to focus on the research and development of new products and upgrades of existing products rather than developing unrelated industries that may face greater risks. The coefficient of the lnNRIF_ER is negative and significant at the 5% level, which indicates that the improvement in environmental regulation strength can strengthen the relationship between NRIF and NNRF entry. This may be because areas with stronger environmental regulations have more NRIFs, thus reducing the willingness of the NNRFs to enter.
Table 7 Test results of the moderating effect
Variables lnNRF lnNNRF
(13) (14)
lnRIF -0.659*** 0.462***
(-3.67) (4.31)
lnNRIF 0.874*** -0.354
(3.56) (-1.63)
IIA 0.0170 -0.0111
(0.52) (-1.00)
lnRIF_IIA 0.0295 -0.0440***
(0.67) (-5.25)
lnNRIF_IIA 0.00305 0.000758
(0.51) (0.16)
ER -0.0274*** 0.00394**
(-9.93) (2.45)
lnRIF_ER -0.00248 0.000858
(-0.55) (0.80)
lnNRIF_ER 0.00837 -0.00384**
(1.25) (-2.96)
lnLC -0.532 -0.0658
(-0.98) (-0.26)
lnEC -4.388*** -0.208
(-4.89) (-0.67)
lnMP 0.378*** 0.268
(3.54) (1.45)
lnIS -0.180 0.156
(-1.34) (0.94)
lnTC -0.102 1.015**
(-0.09) (2.57)
lnFAI 0.161** 0.0837*
(2.47) (1.96)
PDZ 0.0761 0.0497
(0.51) (0.82)
SOE 0.0205** -0.00436
(2.77) (-0.94)
Year dummies Yes Yes
Region dummies Yes Yes
Number of observations 423 423

Note: Standard errors are mentioned in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

4.2.3 The influence of control variables

The control variables that affect NRF entry include electricity costs, market potential, industrial structure, fixed assets, state-owned economy, and environmental regulation intensity as shown in Table 4. First, the electricity costs have a negative and significant impact at the 1% level, and the coefficient is -3.923, indicating that if the cost of electricity increased by 1%, the number of new firms in the resource category decreased by 3.923%. This could be because NRFs mainly belong to high-energy-consuming industries, such as mining and washing of coal, extraction of petroleum and natural gas, and smelting and processing of ferrous metals, which consume large amounts of electricity. Higher industrial electricity prices increase the production costs of high energy-intensive firms, resulting in the decline in firm productivity (Moerenhout et al., 2019; Qu et al., 2022). Lower industrial electricity prices induce the development of high energy-consuming firms, which may become a political tool for local authorities to promote the development of energy-intensive industries (Elliott et al., 2019). Second, the fixed asset investment is significantly positive at the 1% level, indicating that it is one of the important factors affecting NRF entry. The development of capital-intensive industries, such as coal and metallurgy, is characterized by large investment volumes and long payback periods and is highly dependent on capital input. Therefore, large fixed asset investments can provide a supportive environment for regional capital-intensive industries, and attract more NRFs. Third, the share of the state-owned economy is significantly positive at the 5% level, with an effect coefficient of 0.0208, which indicates that it is favorable for the entry of NRFs.
The control variables affecting NNRF entry include transportation conditions. Most factors show no significant impact on NNRF entry, and NNRFs paid more attention to transportation conditions when selecting a location. A high-density transportation network implies convenient traffic conditions that decrease transportation costs, strengthen the flow and agglomeration of production factors, and facilitate connections with external markets. This is especially true for fresh agricultural products, agricultural product processing, manufacturing industries with high transportation requirements, and the tourism industry, which are sensitive to the cost of traveling distances and the time required.

5 Discussion and conclusion

5.1 Discussion

Path dependence is an important feature of the economic landscape in RBCs, and new firms play a significant role in breaking path-dependence and creating new paths. Moreover, for RBCs, based on whether the industry is related to a pre-existing resource industry, we identified two types of new firms: NRFs and NNRFs. The NRFs may persist or renew a regional path, and the NNRFs may import or create a new regional path, which depends on interactions with incumbent firms, the institutional environment, and the socioeconomic context. However, little information is available on how the entry of new firms affects the industrial path development of RCBs. To fill this gap, we proposed an industrial evolution process of RBCs at the firm level based on existing research and put forth an empirical analysis to explore the spatial evolutionary characteristics and driving factors of different types of new firm entries, hoping to promote RBCs’ path renewal and new path creation by attracting new firms to enter the market.
Based on our empirical analysis, we find that the heterogeneity of incumbent firms affects new firm entries in different ways. First, RIFs negatively affect the entry of NRFs, suggesting that the competition effect between RIFs and NRFs exceeds the positive externality agglomeration effect during the mature period of industrial development in RBCs. In contrast to previous studies that emphasize the positive impact of the positive externalities of firm spatial agglomeration on firm entry (Frenken et al., 2015), we find that competition effects among resource-based firms in the mature development stage are significant, which will keep NRFs away from the areas where RIFs gather. Therefore, there is an urgent need to develop alternative industries (Zhang et al., 2011). Second, NRIFs have a positive effect on NRFs’ entry, which may be due to the location selection of NRFs based on market factors. However, few studies have focused on this issue, and more practical research is required. Finally, we find that RIFs have a positive impact on NNRF entry, and these NNRFs mainly belong to the manufacturing industry, extending along the coal and metal mineral industry chain (coal chemical industry and comprehensive utilization industry of waste resources) and cultural tourism industry with a low technical threshold and strong employment absorption capacity. Our findings provide two insights. In the industrial transformation of RBCs, it is possible to explore a transformation path for benign interactive development between old and new industries while considering the optimization and upgrading of traditional industries and the cultivation and support of emerging industries. For mature RBCs, there are usually two new paths: related and unrelated diversification. Related diversification is an extension of the resource industry chain and the construction of the resource ecological industry chain, relying on local knowledge bases and technological innovation, such as the development of the green coking industry cluster in Yuncheng city. Unrelated diversification is primarily the cultural tourism industry. Considering the historical and cultural value of industrial heritage, the low-skill level and capital-driven characteristics of the current tourism industry development, and the government's financial and policy support for the tourism industry, some large-scale resource-based firms have begun to diversify into unrelated tourism industries, such as the Shanxi Jianlong Iron and Steel Cultural and Creative Park and Long Steel Company Longmen Steel Scenic Area. A similar result was obtained by Hu and Zhang (2018), who found that owing to the strategic policies of local governments and the cognitive changes of local coal firms in Zaozhuang city, coal firms started to be involved in the tourism industry or invest in the hotel industry, catering industry, and scenic spot construction. Second, our findings confirm, at the firm level, that the transformation of most resource-based cities in China tends to take a diversified path of extending the industrial chain of resource-based industries and developing alternative tourism industries (Zhang and Yu, 2023). However, we find that resource-based firms play a more active role in the diversified transformation of RBCs.
We find that mature resource-based cities are inclined to strengthen and renew the development paths of resource industries and lack endogenous motivation to create new growth paths for non-resource-based industries. First, for areas where regional natural resources have not been exhausted because of the dual characteristics of natural resource dependence and knowledge specialization, NRFs will still enter the market. However, the number of NRFs is decreasing owing to the intensification of the competitive effect. Second, there are many NNRFs in the commercial service, agriculture, and tourism industries and a relatively small number of manufacturing firms, mainly in the chemical, food, and equipment manufacturing industries. Additionally, the agglomeration effect of non-resource-based industries is not significant. Third, once local mineral resources are gradually depleted, the endogenous potential for new regional industrial path development tends to be low in resource-based regions with backward economies shown as low returns of investment and technical bases (Trippl et al., 2018). For example, Huozhou city once had as many as 60% of its urban population engaged in the mining industry, and the coal industry accounted for more than 70% of its GDP. However, after it was identified as a resource-exhausted city in 2011, only a small number of NRFs and NNRFs entered Huozhou city during 2011-2019, such as manufacture of non-metallic mineral products, wholesale and retail coal, and equipment manufacturing.
Furthermore, focusing on non-resource industries, we found that interactive relationships between new and old industries, efficient transportation networks, and strong policy guidance are conducive to promoting the development of new regional industries. First, the sustainable transformation of RBCs is a process involving multi-subject participation and industrial interaction. Resource industry path renewal can promote economic development and increase fiscal revenue, and part of the capital accumulation can be reinvested in new growth path creation under the combined action of an effective government and effective market. In contrast, when RBCs begin to transform after the recession of the resource industry, relatively stagnant economic development and social inertia hinder the reform of old industries and the entry of new industries, and the region usually falls into the dilemma of resource industry recession and slow growth, especially regarding new industries. Therefore, local governments should seize this critical period of industrial transformation and promote the interactive development of new and old industries. Second, environmental regulations can promote non-resource-based industry creation and force the industrial transformation of RBCs. By increasing a firm's environmental costs, environmental regulations may cause resource-based firms to move to areas with relatively less restrictive environmental policies, force energy-intensive resource-based firms, or pollution-intensive non-resource-based firms to develop technology-intensive and knowledge-intensive firms, thereby hindering resource industry development and creating new growth paths (Lu et al.,2019). In general, sustainable transformation of RBCs requires industrial diversification, industrial upgrading, and responsive infrastructure improvement.

5.2 Conclusion and policy recommendations

In summary, this study discusses the development paths of resource-based and non-resource- based industries in RBCs by demonstrating the spatial characteristics and driving factors of new firm entries. First, there are significant spatial differences in the entry of the different types of new firms. We find that the NRF entry shows natural resource dependence and developmental path dependence, whereas NNRFs present path creation mainly in coal wholesale and retail, cultural tourism development, and automobile manufacturing. Second, there were differences in the factors affecting the entry of NRFs and NNRFs. The agglomeration economy, factor costs, economic environment, and policy environment affect NRF entry, whereas the agglomeration economy and economic and policy environments affect NNRF entry. Moreover, we find that diversification agglomeration has a more significant effect than specialization agglomeration in the middle and late periods of specialized areas, and that NRFs tend to choose areas with a lower agglomeration of resource-based incumbents (specialization agglomeration), that is, areas with diversification agglomeration. The resource and path dependence and self-reinforcement of NRFs’ location choice, are more likely to accelerate RBCs into a locked-in path, which requires attention.
However, there are still several issues worthy of in-depth discussion: (1) The division of the influencing factors of resource-based and non-resource-based firms of different industrial types still requires detailed analysis. The sample size of 1477 NRFs was small, which may have interfered with the comparison of influencing factors. To comprehensively explore the development paths of different industries in the region, the new non-resource-based firm types selected for this study include agriculture, manufacturing, and the service industry, but it may also lead to insufficient exploration of the differentiation of the impact factors of NNRFs in manufacturing firms. Future research should explore manufacturing firms for a comparative analysis. (2) Firm spatial dynamics include entry, exit, and migration behaviors. Considering path dependence and path creation in regional development, this study focuses only on entry behavior. Future research should explore exit behavior to portray dynamic changes in regional firms more accurately. (3) There are significant differences in firm size, capital intensity, and ownership. The analysis of firm heterogeneity and its influencing factors needs to be further strengthened.
Based on the above research conclusions, this study proposes some policy suggestions to better promote the transformation of mature resource-based cities. First, the industrial transformation of mature resource-based cities must consider upgrading traditional resource industries and cultivating alternative industries to promote the diversification of industries. By promoting the entry of resource and non-resource new firms, RBCs can continuously update the resource industry's development path and cultivate emerging industries. Second, RBC transformation requires the collective agency of firms and governments. Firms can build knowledge channels outside and within a region; acquire, absorb, and utilize new knowledge; and promote innovation and green development. The government can stimulate the vitality of existing firms and attract new ones by improving infrastructure construction, increasing support for new industries, strengthening the supervision of resource-based industries, and optimizing the regional innovation environment. Finally, interactive relationships between the new and old industries, efficient transportation networks, and strong policy guidance are conducive to promote the development of new regional industries. Therefore, RBCs should seize the critical period of industrial transformation, promote the early interactive development of new and old industries and the spatial agglomeration of new firms and industries by optimizing the construction of transportation infrastructure, and use environmental regulation policies to force industrial transformation.
Arauzo-Carod J M, Coll-Martínez E, Turcu C, 2023. Location attributes explaining the entry of firms in creative industries: Evidence from France. The Annals of Regional Science, 1-32.

Arauzo-Carod J M, Teruel-Carrizosa M, 2005. An urban approach to firm entry: The effect of urban size. Growth and Change, 36(4): 508-528.


Arbuthnott A, Eriksson J, Wincent J, 2010. When a new industry meets traditional and declining ones: An integrative approach towards dialectics and social movement theory in a model of regional industry emergence processes. Scandinavian Journal of Management, 26(3): 290-308.


Arthur W B, 1988. Self-reinforcing mechanisms in economics. In: Anderson P W, Pines D (eds.). The Economy as an Evolving Complex System. Reading, Mass: Addison-Wesley, 9-31.

Boschma R, 2015. Towards an evolutionary perspective on regional resilience. Regional Studies, 49(5): 733-751.


Boschma R, Martin R, 2007. Constructing an evolutionary economic geography. Journal of Economic Geography, 7(5): 537-548.


Boschma R, Wenting R, 2007. The spatial evolution of the British automobile industry: Does location matter? Industrial and Corporate Change, 16(2): 213-238.


Broekel T, Boschma R, 2012. Knowledge networks in the Dutch aviation industry: The proximity paradox. Journal of Economic Geography, 12(2): 409-433.


Capasso M, Cefis E, Frenken K, 2016. Spatial differentiation in industrial dynamics. The case of the Netherlands (1994-2005). Tijdschrift voor Economische en Sociale Geografie, 107(3): 316-330.


Cheng Z, Li X, Wang M, 2021. Resource curse and green economic growth. Resources Policy, 74: 102325.


David P A, 1985. Clio and the economics of QWERTY. The American Economic Review, 75(2): 332-337.

Duranton G, Puga D, 2001. Nursery cities: Urban diversity, process innovation, and the life cycle of products. American Economic Review, 91(5): 1454-1477.


Elliott R, Sun P, Zhu T, 2019. Electricity prices and industry switching: Evidence from Chinese manufacturing firms. Energy Economics, 78: 567-588.


Freedman M L, Kosová R, 2012. Agglomeration, product heterogeneity and firm entry. Journal of Economic Geography, 12(3): 601-626.


Frenken K, Boschma R, 2007. A theoretical framework for evolutionary economic geography: Industrial dynamics and urban growth as a branching process. Journal of Economic Geography, 7(5): 635-649.


Frenken K, Cefis E, Stam E, 2015. Industrial dynamics and clusters: A survey. Regional Studies, 49(1): 10-27.


Grabher G, 1993. The weakness of strong ties:The lock-in of regional development in Ruhr area. In: Grabher G (ed.). The Embedded Firm: On the Socioeconomics of Industrial Networks. London: Routledge, 255-277.

Hassink R, 2007. The strength of weak lock-ins: The renewal of the Westmünsterland textile industry. Environment and Planning A, 39(5): 1147-1165.


Hu X, 2017. From coal mining to coal chemicals? Unpacking new path creation in an old industrial region of transitional China. Growth and Change, 48(2): 233-245.


Hu X, Zhang W, 2018. Institutional evolution and regional economic resilience: A comparison of two resource-exhausted cities in China. Geographical Research, 37(7): 1308-1319. (in Chinese)


Hudson R, 1994. Institutional change; cultural transformation, and economic regeneration:Myths and realities from Europe's old industrial area. In: Amin A, Thrift N (eds.). Globalization, Institutions, and Regional Development in Europe. Oxford: Oxford University Press, 196-216.

Jin F, Chen L, Yang Y et al., 2018. The classification and growth pattern of industrial bases in China. Journal of Geographical Sciences, 28(12): 1757-1780.

Klepper S, 1996. Entry, exit, growth, and innovation over the product life cycle. The American Economic Review, 86(3): 562-583.

Li Q, Liu S, Yang M et al., 2021. The effects of China's sustainable development policy for resource-based cities on local industrial transformation. Resources Policy, 71: 101940.


List J A, Millimet D L, Fredriksson P G et al., 2003. Effects of environmental regulations on manufacturing plant births: Evidence from a propensity score matching estimator. Review of Economics and Statistics, 85(4): 944-952.


Lu Y, Wang J, Zhu L, 2019. Place-based policies, creation, and agglomeration economies: Evidence from China's economic zone program. American Economic Journal: Economic Policy, 11(3): 325-360.

Mao X, Liu Y, He C, 2015. Spatial pattern dynamics of resource-based industry in China. Journal of Natural Resources, 30(8): 1332-1342. (in Chinese)


Marshall A, 1920. Principles of Economics. London: Macmillan.

Martin R, 2010. Roepke lecture in economic geography: Rethinking regional path dependence: Beyond lock-in to evolution. Economic Geography, 86(1): 1-27.


Martin R, Sunley P, 2011. Conceptualizing cluster evolution: Beyond the life cycle model? Regional Studies, 45(10): 1299-1318.


Moerenhout T S, Sharma S, Urpelainen J, 2019. Commercial and industrial consumers’ perspectives on electricity pricing reform: Evidence from India. Energy Policy, 130: 162-171.


Neffke F, Hartog M, Boschma R et al., 2018. Agents of structural change: The role of firms and entrepreneurs in regional diversification. Economic Geography, 94(1): 23-48.


Neffke F, Henning M, Boschma R et al., 2011. The dynamics of agglomeration externalities along the life cycle of industries. Regional Studies, 45(1): 49-65.


Qu J, Li K, Ye Z et al., 2022. The impact of differential power pricing policy on firm productivity in China: Evidence from iron and steel firms. Journal of Asian Economics, 101478.

Shi J, He C, 2014. Research progress in spatial dynamics of firms. Progress in Geography, 33(10): 1342-1353. (in Chinese)


Shi J, He C, 2018. Determinants of the spatial variation of new firm formation in China's metal product industry. Geographical Research, 37 (7): 1282-1296. (in Chinese)


Thome K E, Lin Lawell C Y C, 2022. Corn ethanol in the Midwestern USA: Local competition, entry and agglomeration. Journal of Economic Geography, 22(6): 1247-1273.


Trippl M, Grillitsch M, Isaksen A, 2018. Exogenous sources of regional industrial change: Attraction and absorption of non-local knowledge for new path development. Progress in Human Geography, 42(5): 687-705.


Van Stel A, Suddle K, 2008. The impact of new firm formation on regional development in the Netherlands. Small Business Economics, 30(1): 31-47.


Wenting R, Frenken K, 2011. Firm entry and institutional lock-in: An organizational ecology analysis of the global fashion design industry. Industrial and Corporate Change, 20(4): 1031-1048.


Wu K, Zhang W, Zhang P et al., 2023. High-quality development of resource-based cities in China: Dilemmas breakthroughs. Journal of Natural Resources, 38(1): 1-21. (in Chinese)


Yu J, Li J, Zhang W, 2019. Identification and classification of resource-based cities in China. Journal of Geographical Sciences, 29(8): 1300-1314. (in Chinese)


Zeng G, Lu L, He J, 2021. Impact of ecological innovation on the economic transformation of resource-based cities. Resources Science, 43(1): 94-103. (in Chinese)


Zeng J, Zhou Q, Yang D, 2023. Estimating the economic impact of intensifying environmental regulation in China. Environmental and Resource Economics, 1-26.

Zhang W, Wang D, Yu J, 2011. Study on the development path and pattern of substituted industry in resource-based cities. Bulletin of Chinese Academy of Sciences, 26(2): 134-141. (in Chinese)

Zhang W, Yu J, 2023. Policy evolution and transformation effect analysis of sustainable development of resource-based cities in China. Journal of Natural Resources, 38(1): 22-38. (in Chinese)

Zhang Y, Hu Z, Miao C, 2020. Relationship between different agglomeration types and the dynamics of new enterprises in the central region. Economic Geography, 40(5): 155-164. (in Chinese)

Zheng J, Shi L, 2023. Local innovation ability, regional innovation highland and gradient evolution of industrial geography. China Industrial Economics, (5): 43-60. (in Chinese)

Zhou Y, He C, Zhu S, 2017a. Does creative destruction work for Chinese regions? Growth and Change, 48(3): 274-296.


Zhou Y, Zhu S, He C, 2017b. How do environmental regulations affect industrial dynamics? Evidence from China's pollution-intensive industries. Habitat International, 60: 10-18.


Zhu J, Lin B, 2022. Resource dependence, market-oriented reform, and industrial transformation: Empirical evidence from Chinese cities. Resources Policy, 78: 102914.


Zhu S, Guo Q, He C, 2021. Strong links and weak links: How do unrelated industries survive in an unfriendly environment? Economic Geography, 97(1): 66-88.


Zhu S, He C, 2014. Global, regional and local: New firm formation and spatial restructuring in China's apparel industry. GeoJournal, 79(2): 237-253.


Zou H, Duan X, Wang L et al., 2022. The effects of environmental regulation on chemical industry location: Evidence from the region along the Yangtze River, China. Growth and Change, 53(2): 800-822.