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).

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).

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.

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).

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.

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.

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.

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.

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.