Introduction

As the main focus area of economic activity and the core subject of mitigating various risks, enhancing the economic resilience of cities is a necessary means to effectively enhance the ability of their economic systems to resist risks and shocks (Fingleton and Palombi, 2013; Martin and Sunley, 2014; Cheng et al. 2022; Wang and Wang, 2021; Papaioannou, 2023; Fan et al. 2023). With the increasingly close global socioeconomic development, economic cooperation among countries is gradually moving toward a path of diversification, openness, and sharing (Zhou and Qi, 2023; Hynes et al. 2022; Jayasinghe et al. 2022; Gajewski, 2022). While countries worldwide share the fruits of economic development due to globalization, some countries have become “shock absorbers” for the cyclic regulation of the international economic system (Andrew et al. 2020; Ye and Qian, 2021; Ben and Ifergane, 2022). In particular, problems related to notable fluctuations in economic development, limited defense capabilities, and low competitiveness are particularly prominent in developing countries. These problems further exacerbate the vulnerability of cities in various countries in response to internal and external changes (Ženka et al. 2019; Wang et al. 2021; Mai et al. 2021). Therefore, how to build a strong and resilient economic system and employ resilience thinking to enhance the driving force and capacity of regional economic development has become an important research topic for countries globally, with the aim of promoting high-quality urban economic development.

The term resilience first evolved from the Latin word “resilio” (Yang et al. 2023). With the increase in the occurrence of uncertain events and external shocks, scholars have applied resilience in fields such as urban engineering resilience, ecological resilience, and economic resilience (Dario and Weterings, 2015; Paolo, 2017; Lemke et al. 2023; Du, 2023; Yu et al. 2023). The connotation of urban economic resilience has been widely investigated by the government, society, and academia, as it better conforms with the current stage of urban economic development in certain countries and the interpretation of the practical problems faced (Du et al. 2023; Gai and Yang, 2023; Hui and Tan, 2023). Urban economic resilience refers to the ability of a city to prevent and resist risks, as well as maintain efficient and sustainable economic development during a specific period (Drobniak, 2017; Tan et al. 2017; Pashapour et al. 2019; Erika and Mangirdas, 2020). Today’s world is vulnerable to severe impacts such as economic crises, epidemics, and natural disasters. Some regions may continue to maintain stable economic growth after impact, while others may suffer heavy losses and fail to recover (Deng et al. 2023; Wang et al. 2023; Lee and Wang, 2023). The reason for this difference is that countries with greater urban economic resilience often exhibit characteristics such as dynamic balance, redundant buffering, and self-healing, which can enable these countries to quickly eliminate risks and automatically adjust and recover, thus effectively resisting external shocks and mitigating internal disasters (Cheng et al. 2023; Zhou and Qi, 2023). Notably, 2008 and 2019 were critical time points in terms of global fluctuations and changes. Under the impacts of the global crisis and the COVID-19 pandemic, respectively, Compared with developed countries such as the United States and Japan, although China’s economic development has also encountered obstacles, its economic growth rate has decreased by 1.8% and 0.1% respectively compared to the previous year, But China’s long-term economic fundamentals have not changed, and the characteristics of sustained economic recovery, notable development potential, high resilience and broad space have not changed (Hao, 2023; Wang et al. 2024; Yang, 2023). For example, under the influence of the COVID-19 pandemic, the economies of the United States and Japan shrank by 3.5 and 5.3%, respectively, in 2021. On the one hand, the economy of China benefited from a series of policy documents issued by the Chinese government on “optimizing the environment, expanding the domestic demand, stabilizing growth, and promoting development”. On the other hand, it benefited from a strong labor force and high consumer market supply capacity.

Industrial structure upgrading is important for exploring the economic resilience of Chinese cities (Drobniak, 2017; Zhou et al. 2019; Betts and Buzzanell, 2022). Especially at this stage, China’s economy is transitioning from high-speed growth to high-quality development, and industrial structure upgrading plays an important role in the urban economy resilience process in different regions and stages (Zhang, 2022; Yin et al. 2023). On the one hand, industrial structure upgrading can cause an acceleration in regional industrial structure transformation from traditional high energy-consuming industries to high-tech industries. In this process, with the emergence of an advanced industrial structure and a specialized division of labor, new economic growth paths can be created, which can enhance the ability of cities to withstand market risks and can facilitate resilient growth of the urban economy (Tan et al. 2017; Cheng et al. 2023; Li et al. 2024). On the other hand, industrial structure upgrading can lead to enhancement in the service industry, but a high proportion of the service industry can easily cause the problem of hollowing out industries, which is not conducive to improving urban economic resilience (Gai and Yang, 2023; Hui and Tan, 2023). It should be noted that industrial structure upgrading entails a long and tortuous process, and its impact on urban economic resilience cannot be achieved overnight. Therefore, the causal relationship between industrial structure upgrading and urban economic resilience is not direct nor obvious. However, the impact of the former on the latter still objectively occurs through various transmission mechanisms. Moreover, due to the poor coherence and sustainability of relevant systems, this impact relationship may be repetitive and not simply linear.

From 1492 to 2023, globalization increased from the 1.0 era to the 4.0 era (Roberta et al. 2015; Aida et al. 2016; Gereffi et al. 2022). Globalization has firmly woven all countries into the network of the world system, and the commerce, economy, and system of each country have undergone tremendous transformation (Dunn, 2020; Carlos et al. 2022). In the globalization process, achieving high-quality and sustainable economic development has become a major challenge for the international community (Ngo, 2023). Since its accession to the World Trade Organization (WTO), China has received a large amount of foreign direct investment due to its own resource endowment. Foreign direct investment has, to a certain extent, accelerated the process of high-quality development of China’s urban economy, which is mainly reflected in two aspects: on the one hand, foreign direct investment can effectively promote industrial structure upgrading and optimization (Tao et al. 2023). Foreign direct investment usually leads to the introduction of technology, management, and market factors, accelerating the transformation of traditional Chinese industries to high-tech and high value-added industries and promoting sustainable development of the urban economy (Anis and Andreea, 2023). On the other hand, foreign direct investment not only provides financial support but also introduces advanced technology and management experience, providing Chinese industries with a larger market and more abundant resources (Knoke et al. 2022; Gereffi et al. 2022). Production scale increase, product quality improvement, and cost reduction further enhance the competitiveness of China’s industries (Carlos et al. 2022; Johnson and Mundell, 2023). Therefore, we must consider several questions: under the acceleration of globalization, what is the impact of industrial structure upgrading (with a focus on industrial structure upgrading and rationalization) on urban economic resilience? Are there certain stage characteristics? These questions should be explored in depth. To this end, this study links the above three aspects and focuses on determining whether there exists a threshold effect in terms of the impact of industrial structure upgrading on urban economic resilience with globalization as the threshold variable, examining its mechanism and effect, and investigating whether there is regional heterogeneity to provide an empirical basis for relevant departments to formulate targeted policies.

The remainder of this study is organized as follows: in the second section, a literature review is provided, in which the existing research on industrial structure upgrading, globalization, and urban economic resilience is summarized, providing a sound basis for this analysis. The third section constructs a theoretical framework and research hypotheses, elaborates on the theoretical basis of industrial structure upgrading, explores the impact of industrial structure upgrading and rationalization on urban economic resilience in the context of globalization, and proposes research hypotheses. In the fourth section, the research design is described in detail, including model settings, variable selection, data sources, and descriptive statistical analysis. In the fifth section, the empirical tests are introduced, including threshold effect assessment, threshold panel model-based estimation, and regional heterogeneity analysis. Finally, the sixth and seventh selections, respectively outline the conclusions, policy recommendations, and future prospects of this article.

Literature review

The study of urban economic resilience has become an important topic for countries to explore high-quality and sustainable economic development (Guo et al. 2023). Boschma (2015) first proposed the concept of economic resilience, stating that economic resilience is the ability of an economic system to absorb shocks without catastrophic changes in its basic functional organization. Thereafter, the academic community investigated urban economic resilience from the perspectives of regional and geographic economics (Wang and Ge, 2023; Wang et al. 2024). Regional economics mainly explores how to cope with the decline in the regional economy, while geographic economics largely focuses on the study of spatial differences, correlations, and influencing factors of urban economic resilience (Zhang et al. 2023). In recent years, academic research on urban economic resilience has focused on three main aspects (Jesse, 2023; Cheng et al. 2022): measurement methods, influencing factors, and the impact of industries on urban economic resilience. First, from the perspective of measurement methods for urban economic resilience, scholars have measured the urban economic resilience index based on methods such as the core variable method and comprehensive indicator method from different disciplinary backgrounds (Qiang et al. 2020; Wang and Wang, 2021). However, due to the varying focuses, there is no consensus at present. Second, from the perspective of the influencing factors of urban economic resilience, scholars have considered that fiscal gaps, geographical location conditions, and resource endowments are key factors that constrain urban economic resilience (Gan and Chen, 2021; Cheng et al. 2022; Wang and Wang, 2021). The processes of urbanization, technological innovation, economic development, and industrial agglomeration are major factors driving urban economic resilience (Du et al. 2023; Gai and Yang, 2023). Third, from the perspective of the impact of industries on urban economic resilience, existing research has focused on two main aspects: (1) from a microscopic perspective, the impact of a single industry, such as the digital industry, financial industry, or manufacturing industry, on urban economic resilience has been examined; and (2) from a macroscopic perspective, the impacts of industrial structure upgrading, adjustment, transformation, and agglomeration on urban economic resilience have been explored. For example, Feng et al. (2023) empirically determined that industrial structure rationalization and upgrading are important ways for regional integration to affect urban economic resilience. However, the policy effects of regional integration on economic resilience vary over time, by region, and by urban structure. Zhang et al. (2023) noted that regional economic resilience is closely related to the state of the industrial structure, and there exists a spatiotemporal correlation in the evolution of the two systems. In the literature review process, we could conclude that the existing research on the relationship between industry and urban economic resilience has not yet reached a consensus, both at the micro- and macroscopic levels, thus providing a theoretical basis and new ideas for this study.

Regarding the relationship between globalization, industrial structure upgrading, and urban economic resilience, studies have mostly focused on the relationship between globalization and industrial structure upgrading, as well as the relationship between industrial structure upgrading and economic resilience (Carlos et al. 2022). In terms of the relationship between globalization and industrial structure upgrading, studies have suggested that globalization can promote industrial structure upgrading through the division of labor and cooperation in the industrial structure (Dunn, 2020) and that globalization can promote industrial structure upgrading through technological innovation and industrial transformation (Ngo, 2023). Globalization can improve factor allocation efficiency and drive industrial structure upgrading by influencing the direction and quantity of factor flow. In addition, scholars have noted that the development of globalization encompasses various stages, while its impact on the industrial structure is also cyclical, which can lead to instability in the impact of globalization on industrial structure rationalization and upgrading. In terms of the relationship between globalization and urban economic resilience, there are three specific viewpoints: globalization imposes a reducing effect on urban economic resilience (Tao et al. 2023), globalization exerts an expanding effect on urban economic resilience (Anis and Andreea, 2023), and the impact of globalization on urban economic resilience is dynamic (Martin et al. 2016). In terms of the relationship between industrial structure upgrading and urban economic resilience, industrial structure rationalization and upgrading can help to reduce the impact of international markets by improving the industrial configuration and quality level, providing greater development space for adaptive structural adjustment after impact occurrence and thus continuously enhancing urban economic resilience.

In the literature, scholars have empirically evaluated industrial structure upgrading and globalization as important factors affecting urban economic resilience based on econometric models, geographic models, and spatial econometric models (Maria, 2023; Cheng et al. 2023); however, few scholars have explored the relationships among these three factors. Within the context of globalization, the transformation of the new international division of labor model profoundly affects the process of industrial structure adjustment in various countries worldwide. However, existing research has overlooked the moderating effect of globalization on industrial structure upgrading and urban economic resilience. Moreover, the literature has mostly focused on analyzing the linear effect of industrial structure upgrading on urban economic resilience, and various conclusions have been obtained. This also reflects the complexity of the relationship between the two aspects, which suggests that they may not be characterized by a simple linear relationship, namely, there may be a nonlinear relationship. Compared with the literature, the marginal contribution of this study lies in coupling industrial structure upgrading, globalization, and urban economic resilience. It was preliminarily determined that the impact of industrial structure upgrading on urban economic resilience is nonlinear, and globalization was used as a threshold variable. Panel data from 267 prefecture-level cities and above in China from 2008 to 2021 were selected, and we empirically examined the nonlinear relationship between industrial structure upgrading and urban economic resilience and studied its stage characteristics. Moreover, we conducted robustness tests.

Theoretical foundations and research hypotheses

Industrial structure upgrading is important for exploring the economic resilience of Chinese cities. Especially at the current stage, China’s economy is shifting from high-speed growth to high-quality development, and industrial structure upgrading fulfills an important role in the resilience process of the urban economy in different regions and stages (Feng et al. 2023). The theory of industrial structure was first proposed by Fisher, who divided the overall industrial structure into the primary industry, secondary industry, and tertiary industry (Zheng et al. 2023; Li, 2024). William Petty established the theory of industrial structure upgrading, which refers to the process of industrial structure transformation from lower to higher stages, and constructed a theoretical analysis framework for industrial structure upgrading, including two aspects: the rationalization and advancement in the industrial structure (Feng et al. 2023). This theoretical framework laid the foundation for subsequent related research. With the deepening and development of theoretical research, the theory of industrial structure upgrading is the result of the joint action of the two forces of industrial structure upgrading and rationalization. There is widespread consensus among scholars at home and abroad, and this approach has been widely applied in fields such as social crises, regional environments, grassroots governance, and high-quality urban development. This has provided a new research approach for exploring the impact of industrial structure upgrading on urban economic resilience.

Since China joined the WTO, the Chinese economy has been further integrated into the world economy and become an important destination for FDI worldwide. Since then, FDI has become an important manifestation of globalization (Zhang et al. 2020). Globalization not only enhances the frequent exchange of resources such as technology, capital, and labor among countries but also promotes increasingly close economic development relationships among countries. Within the context of globalization, while countries achieve effective resource allocation in their industrial structures, they also move toward advanced and rational industrial structures, thereby promoting high-quality development of urban economic resilience through the release of structural dividends. In other words, industrial structure upgrading has become an important driving force for urban economic resilience within the context of globalization (Fig. 1).

Fig. 1: Theoretical analysis framework of industrial structure upgrading and urban economic resilience with globalization as a threshold variable.
figure 1

In the figure, S denotes suppression, P denotes promotion, C denotes connection, and R denotes regular.

Full size image

Industrial structure upgrading involves a process of rationalization. The singularity of the industrial structure is not conducive to urban entities overcoming internal and external market risks within an uncertain environmental context. At the primary stage of globalization, foreign direct investment can not only solve the problem of imbalanced and insufficient development of the regional industrial structure via the utilization of local resource endowments but also reduce production costs, provide stable and long-term global competitive advantages, and promote the intensive and large-scale development of local industries by achieving a reasonable division of labor and allocation of various industries in the regional industrial chain. The entry of labor- and capital-intensive industries into developing countries creates a large number of employment opportunities, promotes the flow of surplus labor from the agricultural sector to the nonagricultural sector, and enhances regional resistance to internal and external market risks. With the improvement in the globalization level and the increase in foreign investment, the formation of a diversified industrial system in cities is accelerating. Cities rely on industrial diversification, product richness, high-added value, and asynchronous industrial cycles to avoid drastic fluctuations in output and employment, thereby enhancing the ability of the urban economic system to overcome risks and adapt to shocks. Therefore, Hypothesis 1 was proposed.

Hypothesis 1: Industrial structure rationalization exerts a significant positive impact on urban economic resilience at different stages of globalization.

Another main feature of industrial structure upgrading is the upgrade process. At the early stages of globalization, foreign direct investment in factory construction, to some extent, dealt a heavy blow to traditional industries in developing countries, reducing the ability of urban economic entities to eliminate market risks. With the increase in globalization, on the one hand, FDI can promote the return of technology, labor, and capital and accelerate the replacement of high value-added industries and low value-added industries. On the other hand, foreign direct investment can not only promote the transition from traditional resource-intensive industries to resource-intensive industries but also attract more high-quality labor and technological resources, thereby increasing the resistance of urban economic systems to risks and the ability to adapt to impact factors. Therefore, Hypothesis 2 was proposed.

Hypothesis 2: There may be significant stage differences in the impact of industrial structure upgrading on urban economic resilience between different levels of globalization.

Research design and data sources

Model settings

As mentioned earlier, industrial structure upgrading is accompanied by the reasonable flow and optimized allocation of production factors between industries and countries, which exerts a certain impact on urban economic resilience. Globalization also plays an important role in this process, but confirmation of the existence of threshold effects still requires further empirical testing. Therefore, in this study, globalization was adopted as a threshold variable, urban economic resilience was used as the dependent variable, and industrial structure upgrading and rationalization were employed as the core explanatory variables. Then, the following threshold panel model was constructed (Dou and Gao, 2023; Zheng et al. 2023):

$$\begin{array}{c}UE{R}_{it}={\beta }_{1}RI{S}_{it}\ast I(G{L}_{it}\le {r}_{1})+{\beta }_{2}\ast RI{S}_{it}\ast I({r}_{1} < G{L}_{it}\le {r}_{2})+\ldots +{\beta }_{n}RI{S}_{it}\ast I({r}_{n-1} < G{L}_{it}\le {r}_{n})\\ +{\beta }_{n+1}RI{S}_{it}\ast I(G{L}_{it} > {r}_{n})+\theta {X}_{it}+{\upsilon }_{i}+{\varepsilon }_{it},\end{array}$$
(1)
$$\begin{array}{c}UE{R}_{it}={\alpha }_{1}AI{S}_{it}\ast I(G{L}_{it}\le {r}_{1})+{\alpha }_{2}\ast AI{S}_{it}\ast I({r}_{1} < G{L}_{it}\le {r}_{2})+\ldots +{\alpha }_{n}AI{S}_{it}\ast I({r}_{n-1} < G{L}_{it}\le {r}_{n})\\ +{\alpha }_{n+1}AI{S}_{it}\ast I(G{L}_{it} > {r}_{n})+\kappa {X}_{it}+{\eta }_{i}+{\omega }_{it},\end{array}$$
(2)

where UERit denotes the urban economic resilience of city i during the t-th period; GLit denotes the globalization of city i during the t-th period as a threshold variable, with r1, r2…, rn representing n threshold values; RISit and AISit denote industrial structure rationalization and advancement, respectively; α1, α2, and αn, and β1, β2, and βn denote the regression coefficients for different threshold intervals; Xit is a series of control variables; and θ and k are the regression coefficients of the control variables. Note that the only difference between Models (1) and (2) is the core explanatory variables (without considering differences in the parameter values), which are industrial structure upgrading and industrial structure rationalization, respectively. For the sake of brevity, Models (1) and (2) are distinguished by their core explanatory variables and are thus referred to as the RIS and AIS models, respectively.

Variable selection

Explained variable

Calculation model for urban economic resilience

The regional economic foundation often determines the lower limit of a region’s ability to withstand shocks, which, to a certain extent, affects its resilience and recovery level (Wang et al. 2021). Notably, urban economic resilience is not only related to the total economic output but also closely related to the economic structure. The higher the total economic output is, the higher the urban economic resilience when facing risks. The GDP is an appropriate indicator of the total economic output of a city, while a reasonable economic structure is also a key factor in ensuring healthy and dynamic growth of the urban economy. In contrast to the former, the economic structure more strongly reflects changes in economic growth rates. The main research methods for measuring urban economic resilience are the core variable method and the comprehensive indicator method, but a consensus has not yet been reached. Considering the representativeness and continuity of the core variable method, this study refers to existing research, and the output method was used to measure urban economic resilience. This indicator can directly reflect the degree of change in the urban economy in the face of pressure and shocks (Feng et al. 2023). Compared with previous studies, this approach avoids the subjectivity of using a comprehensive indicator system, but it neglects the relationships and dependencies between factors and does not fully reflect the actual situation. The specific calculation method is as follows:

$${\rm{Urban}}\; {\rm{Economic}}\; {\rm{Resilience}}={\rm{GPD}}\times \Delta {\rm{GDP}}{\rm{V}}\times 100$$
(3)

where GDP denotes the standardized value of the GDP and ∆GDPV is the standardized value of the absolute change in the GDP growth rate in adjacent years. As the product of multiple standardized values is very small, to better visualize the differences in urban economic resilience, the regression coefficient is multiplied by 100.

Core explanatory variables

According to existing research, industrial structure rationalization and advancement were used as alternative indicators to measure urban industrial structure upgrading (Gan and Chen, 2021; Yin et al. 2023). As expressed in the equation below, rationalization of the industrial structure (RIS) mainly reflects the coupling relationship between the input and output, where i denotes the primary industry, secondary industry, or tertiary industry, Y represents industrial economic output, L represents labor input, and I represents industrial sector (i = 1,2,3). Upgrading of the industrial structure (AIS) largely reflects the proportion of the service industry, measured as the ratio of the added value of the tertiary industry to that of the secondary industry.

$$RIS=\mathop{\sum }\limits_{i=1}^{m}\left(\frac{{Y}_{i}}{Y}\right){\mathrm{ln}}\left(\frac{{Y}_{i}}{{L}_{i}}/\frac{Y}{L}\right)$$
(4)
$${\rm{AIS}}={\rm{Tertiary\; Industry}}/{\rm{Secondary\; Industry}}$$
(5)

Threshold variables

The threshold variable selected in this study is globalization (GL). After China’s accession to the WTO, China gained a large amount of foreign direct investment due to its resource endowment. With the increasingly close relationship between China’s economic development and the global economy, FDI transformed globalization into a localization force through location selection, which, to a certain extent, promoted the development of the regional economy. However, the greater the degree of closeness to the global economy is, the greater the occurrence probability of unpredictable risks, thereby exacerbating the vulnerability of regional economic resilience. Therefore, the per capita foreign direct investment amount was selected as a characterization indicator to measure globalization.

Control variables

(1) Industrial agglomeration level of industrial enterprises. We selected the total number of industrial enterprises/urban construction land area as an indicator to measure the level of industrial enterprise agglomeration (Wang et al. 2021). The agglomeration of industrial enterprises can accelerate the convergence of regional enterprises, goods, services, and highly skilled labor. The agglomeration of a large number of intermediate inputs, high-level services, and human capital imposes a significant positive effect on urban economic resilience. However, the excessive agglomeration of industrial enterprises could also lead to significant negative externalities with respect to factors such as infrastructure, residents’ health, and the ecological environment, reducing the responsiveness of urban economic systems to external shocks. (2) Local financial gap. We selected the indicator of (expenditure within the local fiscal budget—revenue within the local fiscal budget)/revenue within the local fiscal budget to measure fiscal gaps (Xiong et al. 2023). The level of local finance fulfills an important role in achieving the optimal allocation of regional resources. Research has shown that the smaller the local fiscal gap, the more capable local governments are of achieving a balance and structural optimization between the total social demand and supply when facing external shocks. In contrast, the greater the local fiscal gap is, the lower the ability of local governments to self-adjust and repair in the face of external shocks. (3) Technology investment levelFootnote 1. We selected scientific and technological investment/local general public budget expenditure as a measurement indicator (Zhang et al. 2020). Improving the technology investment level can, to a certain extent, accelerate the elimination of traditional industries and the growth of emerging industries in a given region, enhance the competitive advantage of the entire industry chain in key fields, and overcome the monopoly of core technologies in certain foreign fields, thus increasing urban economic resilience. (4) Population density. We selected the total number of permanent residents in cities within the province/the area of provincial jurisdiction as a measurement indicator (Tan et al. 2020). The population density is a key factor affecting urban economic resilience and varies across different regions. The regional agglomeration effect generated by the population density can cause various resources to gather in cities, while population aggregation can cause various high-quality resources to accumulate in cities, generating a positive agglomeration effect, which facilitates the construction of regional spatial governance systems, high-quality and intelligent public services, and infrastructure systems and has a certain significance in building urban economic resilience. When there is a turning point in the population density, the phenomenon of excessive development and utilization of regional resources may occur, requiring local governments to consume more financial and material resources to solve problems such as environmental damage, severe resource depletion, and high carrying pressure caused by population agglomeration. This exerts a significant inhibitory effect on urban economic resilience, so the total number of permanent residents in cities within the province/the area of provincial jurisdiction was chosen as a measurement indicator. (5) Infrastructure level. We selected the coverage of public transportation routes/total population as a measurement indicator (Zhang, 2022). Infrastructure, including transportation and communication facilities, provides the basic guarantee for urban economic activities. The convenience of infrastructure is directly related to the development of local economic resilience. Research has shown that improving infrastructure is not only conducive to high-quality development of the local economy but also enhances the ability of urban economic systems to overcome risks and adapt to shocks. (6) The economic development level was measured by the total GDP/total population ratio in this study (Deng et al. 2023). The economic strength and development level of a city significantly impact its resilience: the stronger the economy is, the more reasonable the structure, and the higher the innovation ability is, the greater the resilience level of a city. Research has indicated that the urban economic development level is related to the ability to withstand macroeconomic and financial risks.

Data sources and descriptive statistics

We selected panel data for 267 prefecture-level cities and above in China (Excluding Hong Kong, Macao, Taiwan, Xinjiang, and Xizang) from 2008 to 2021 for empirical analysis. The panel data were obtained from the Statistical Yearbook of Urban Construction in China and the Statistical Yearbook of Chinese Cities from 2008 to 2021. Descriptive statistics of all variables are listed in Table 1. The differences between the minimum and maximum values for measuring the urban economic resilience index were significant, indicating a significant difference in urban economic resilience among the 267 prefecture-level cities and above in China between 2008 and 2022. The maximum value of urban economic resilience was observed for Shanghai in 2019, with an urban economic resilience of 20.033, while the minimum value was obtained for Dingxi city in 2018, with an urban economic resilience of 0.000. The minimum and maximum values of the industrial structure rationalization indicator were 0.581 and 29.114, respectively, while the minimum and maximum values of the industrial structure upgrading indicator were 0.139 and 5.350, respectively. This reflects the considerable differences in industrial structure rationalization and upgrading among the 267 prefecture-level cities and above between 2008 and 2022. Similarly, there was a significant difference between the minimum value (−2.028) and the maximum value (9.970) of the globalization indicator, indicating a significant difference in globalization among the 267 prefecture-level cities and above between 2008 and 2022.

Table 1 Descriptive statistics of the variables.
Full size table

Empirical testing

Hausman test

Before conducting the Hausman test, we first applied the Durbin–Wu–Hausman (DWH) test to assess for endogeneity issues in the benchmark panel model (Zheng et al. 2023). According to the test results, the P value of the DWH test was 0.000, indicating that the null hypothesis of “all explanatory variables are exogenous” could not be rejected at a significance level of 1%. Therefore, the model did not exhibit endogeneity issues. Finally, we conducted a Hausman test of the relationship between industrial structure rationalization and upgrading and urban economic resilience. The results indicated that the P value of both models (1 and 2) in the Hausman test was 0.0000, indicating that the original hypothesis of random effects could be rejected at the 1% significance level and that the alternative hypothesis of fixed effects could be accepted. Therefore, we utilized the panel fixed effects model to estimate Models 1 and 2.

Assessing the threshold effect of industrial structure upgrading

Before establishing a specific threshold effect model, two important tests were conducted: one involved testing for the existence of threshold effects, with the aim of exploring whether the parameter spaces within different threshold intervals divided by the threshold values significantly differ; the other was using the bootstrap method for consistency testing, with the aim of determining whether the estimated threshold value is consistent with the actual value. The former is generally evaluated by the F statistic, while the latter is assessed by the likelihood ratio (LR) statistic.

The test results for the industrial structure rationalization and upgrading models with globalization as the threshold variable are provided in Table 2. First, a single threshold test was conducted, and the corresponding F values were 193.71 and 514.98, respectively, while the P values were 0.0033 and 0.0000, indicating that there exists a threshold effect in both the rationalized and advanced industrial structure models. Second, the double threshold effect existence test was performed, with corresponding F values of 55.85 and 90.97, respectively, and P values of 0.0700 and 0.0133, respectively, which were significant at the 10 and 5% levels. The original hypothesis of the existence of a single threshold could be rejected, and it could be considered that both models contain a double threshold effect. Finally, a triple threshold test was conducted, and the P values at this time indicated that the original hypothesis of the existence of double thresholds could not be rejected. Therefore, in this study, a dual threshold effect model was chosen for estimation. The model can be formulated as follows:

$$\begin{array}{c}UE{R}_{it}={\alpha }_{1}RI{S}_{it}\ast I(G{L}_{it}\le {r}_{1})+{\alpha }_{2}RI{S}_{it}\ast I({r}_{1} < G{L}_{it}\le {r}_{2})\\ +{\alpha }_{3}RIS\ast I(G{L}_{it} > {r}_{2})+\theta {X}_{it}+{\mu }_{i}+{\varepsilon }_{it}\end{array}$$
(6)
$$\begin{array}{c}UE{R}_{it}={\beta }_{1}AI{S}_{it}\ast I(G{L}_{it}\le {r}_{1})+{\beta }_{2}AI{S}_{it}\ast I({r}_{1} < G{L}_{it}\le {r}_{2})\\ +{\beta }_{3}AI{S}_{it}\ast I(G{L}_{it} > {r}_{2})+\eta {X}_{it}+{\nu }_{i}+{\sigma }_{it}\end{array}$$
(7)
Table 2 Existence test for the threshold effect in the Theil model with GL as the threshold variable.
Full size table

We used the estimation method of minimizing the sum of squares of residuals to determine the specific threshold values, and the results are provided in Table 3. The first and second threshold values for the model variables of industrial structure rationalization and upgrading (GL) were 6.8724 and 6.5514, respectively, and 7.5034 and 6.8724, respectively. Table 3 also provides the confidence intervals for each threshold value at a 95% confidence level.

Table 3 Threshold estimation results and confidence intervals for the GL variable of the RIS and AIS models.
Full size table

Next, we performed a consistency test between the estimated threshold values and the actual values. Based on the estimation results listed in Table 3, a likelihood ratio function graph was created. The horizontal axis in Fig. 2 represents the threshold value for globalization, the vertical axis represents the likelihood ratio function value, and the dashed line represents the critical value at the 95% confidence level. Here, the upper half represents the confidence interval for the first threshold of globalization, while the lower half represents the confidence interval for the second threshold of globalization. According to Hansen’s likelihood ratio test model, for LR(γ) > C(θ), the original hypothesis can be rejected. For θ = 5%, the critical value of the LR statistic is 7.35. According to Fig. 1, the threshold values of the LR statistic corresponding to industrial structure rationalization and globalization of the advanced models were significantly lower than the critical values, so the above threshold estimates could be considered true and effective.

Fig. 2
figure 2

Dual threshold values and likelihood ratio function of GL in the RIS and AIS model.

Full size image

Estimation results of the threshold panel model

Regression estimates of the impact threshold of industrial structure upgrading on urban economic resilience are provided in Table 4. Among them, the regression results for basic Models (I) and (V) indicated that the relationship between industrial structure upgrading and globalization could be divided into three intervals by the threshold variable of globalization, and there were significant differences between the different intervals. At a lower globalization level, the impact of industrial structure rationalization on globalization was significantly positive at the 1% level, while the impact of industrial structure upgrading on globalization was not significantly negative. On the one hand, this suggests that at a lower globalization level, industrial structure rationalization could significantly improve urban economic resilience. When faced with external shocks, the economic resilience of cities significantly differs. Cities with a reasonable industrial layout exhibit higher economic resilience, namely, cities with a more reasonable industrial structure can quickly adjust their industrial structure, thereby obtaining more persistent and robust economic resilience. On the other hand, this result indicates that the level of industrial upgrading in China is still low, and products are at a disadvantage in the global market competition process. At this time, globalization is still dominated by the negative siphon effect.

Table 4 Regression estimation results and robustness test results for the impact threshold of the Theil model on urban economic resilience.
Full size table

Notably, at the early stages of globalization, industrial structure rationalization is the foundation for the economy to overcome external shocks, while later, the same process is the source of sustained economic resilience. When globalization crosses the first threshold and reaches the intermediate stage, the regression coefficient between industrial structure rationalization and urban economic resilience remains significantly positive, The above research confirms the validity of Hypothesis 1. This may occur because, with the acceleration of globalization, industrial structure rationalization eliminates market barriers between regions and internationally through the rational division of labor and the allocation of various industries within the industrial chain, achieving effective resource allocation, releasing structural dividends, and providing stable and long-term global competitive advantages, thus enhancing the ability of urban economic systems to mitigate risks and adapt to shocks. The impact of an advanced industrial structure on urban economic resilience within this interval shifted from negative to positive, indicating that with the improvement in the level of the advanced industrial structure, the return of technology, talent, and capital is promoted, accelerating the replacement of high value-added industries and low value-added industries, improving the demand income elasticity of high value-added products, reducing the operational risks of urban entities, and increasing the ability of urban economic entities to withstand market risks. When globalization crosses the second threshold and reaches a high level, the impact coefficients of industrial structure rationalization and upgrading on globalization are significantly positive at the 1% significance level, with regression coefficients of 0.037 and 0.780, respectively. The main reason is that with the rapid development of globalization, the industrial structure layout becomes more reasonable, which is conducive to the formation of a diversified industrial system in cities. Cities rely on industrial diversification, product richness, high-added value, and asynchronous industrial cycles to prevent sharp fluctuations in output and employment, thereby improving the ability of the urban economic system to overcome risks and adapt to shocks. In contrast, the deepening of globalization promotes the transition from traditional resource-intensive industries to resource-intensive industries, thus vigorously enhancing the level of industrial intensification and, in return, the ability of the urban economic system to withstand risks and adapt to shocks. The above research confirms the validity of Hypothesis 2.

After benchmark regression, to ensure the reliability of the conclusions obtained, we conducted two robustness tests. First, considering that extreme values in the sample may impact the results, Beijing (which attained the highest urban economic resilience value in 2019) and Dingxi (which attained the lowest urban economic resilience value in 2018) were excluded, and the model was again regressed to obtain robustness test results (II and VI in Table 4). Second, robustness tests were performed by controlling for time and individual interaction terms, and the results of robustness tests III and VII are listed in Table 4. Finally, by including control variables for robustness testing, additional robustness test results (IV) and (VIII) were obtained, as provided in Table 4. The three additional control variables were the market size, financial development level, and cultural soft power (Guo et al. 2023; Zhang and Yao., 2023), where the market size can be measured by the proportion of the total retail sales of consumer goods to the GDP. The financial development level can be measured by the ratio of the bank loan scale to the GDP. Cultural soft power can be measured by the logarithmic value of book collection. Compared with the estimation results of the basic model, the core explanatory variables in the robustness test results—namely, industrial structure rationalization and industrial structure upgrading—basically exhibited regression coefficients of the same sign and significance within the various threshold ranges of globalization (represented by GLL, GLM, and GLH). For example, robustness test result II was relatively close to that of basic Model I, while robustness test result VI was relatively close to that of basic Model V. The sign and significance of the regression coefficients for the core explanatory variable, i.e., industrial structure rationalization, within the various threshold intervals of globalization were highly similar. However, after removing the extreme values in the sample data, the regression coefficient and significance of industrial structure upgrading exhibited significant changes within the various threshold ranges of globalization, with significance decreasing from the previous 1% level to the 5% level. However, it should be noted that although the removal of extreme values generated a certain impact in this study, the existence of individual extreme values did not affect the basic conclusions of the model. For example, robustness test result III was related to basic Model I, while robustness test result VII was related to basic Model V. The regression coefficients for the core explanatory variables of industrial structure rationalization and industrial structure upgrading did not show significant changes within the various threshold ranges of globalization, as Model VII incorporated time and individual interaction terms, and the main difference between the two models was only the fact that within the third threshold range (GLH) of globalization, the significance of the regression coefficient of Model VI slightly decreased. However, it was still significantly positive at the 5% level, and the conclusion remained significant. This indicates that, regardless of whether time and individual interaction terms were controlled, the conclusions of model analysis remained consistent. The results of robustness tests (IV) and (VIII) showed that the three additional control variables slightly but significantly affected the basic regression results. Therefore, the threshold regression model constructed based on panel data from 267 prefecture-level and above cities in China for assessing the impact of industrial structure upgrading on urban economic resilience was robust and reliable.

Analysis of regional heterogeneity

To eliminate the interference of regional heterogeneity factors and verify the moderating effect of globalization on industrial structure upgrading and urban economic resilience, we divided the 267 prefecture-level and above cities into four major regions—eastern, central, northeastern, and western regions—and conducted threshold effect testing and threshold model estimation. The specific results are listed in Table 5. The industrial structure rationalization and upgrading models in both the eastern and central regions passed the double threshold test. The threshold values for globalization of the two models in the eastern region were 6.54 and 6.88, respectively, and 6.44 and 6.88, respectively, while the threshold values for globalization of the two models in the central region were 6.63 and 6.97, respectively, and 6.97 and 7.23, respectively. The industrial structure rationalization and upgrading models in the western region and the industrial structure upgrading model in the northeastern region did not pass the double threshold test, while the industrial structure rationalization model in the northeastern region exhibited only a single threshold effect.

Table 5 Existence test and value of the threshold effect of GL as a threshold variable in the Theil model.
Full size table

The threshold panel regression results for the eastern, central, western, and northeastern regions are listed in Table 6, which indicates certain differences between the regression results for each region and the national level.

  1. (1)

    In the eastern region, when globalization fell below the first threshold, the regression coefficients of industrial structure rationalization and upgrading on globalization were 0.056 and 0.275, respectively, and both passed the 1% significance test. When globalization crossed the first threshold but remained lower than the second threshold, the regression coefficients of both were significant. After globalization crossed the second threshold, the regression coefficients of the two variables were 0.109 and 1.546, respectively, and both passed the 1% significance test. These results indicated that with the continuous improvement in globalization, the relationship between industrial structure rationalization and upgrading and urban economic resilience in eastern China exhibits an inverted U-shaped trend. The reason for this may be that the eastern region has become the preferred region for accepting foreign investment due to its inherent resource endowment and location advantages. On the one hand, with increasing foreign investment, cities in the eastern region have launched a prelude to traditional industrial transformation. The optimization and improvement in industrial and product structures have, to a certain extent, increased the overall development level of regional industries and international competitiveness while also driving the improvement in urban economic resilience. On the other hand, with the limited urban carrying capacity in the eastern region, foreign direct investment is expected to be gradually transferred to the central and western regions, resulting in the phenomenon of urban economic resilience first increasing and then decreasing.

  2. (2)

    The threshold effect of industrial structure upgrading and rationalization on urban economic resilience in the central region was similar to that in the eastern region, also exhibiting an inverted U-shaped trend. On the one hand, in recent years, the central region has become an important position for China’s development of strategic emerging industries, committed to solving problems such as industrial structure convergence and chain fragmentation, and relies on diversified industrial structures (strategic emerging industries) to accelerate the rational allocation of industrial structures and regions to enhance the synergy between the secondary industry (manufacturing) and tertiary industry (productive services), improve the correlation and cohesion between industries within the region, promote the spillover of knowledge and innovation between departments, accelerate the formation of regional industrial economies of scale, and achieve long-term and robust urban economic resilience. On the other hand, the central region has adjusted its industrial structure through specific plans such as urban renewal, accelerating the circulation of key technological elements between different industries, improving the efficiency of regional resource allocation, increasing the size of the regional economy, and taking the lead in entering the post-industrialization stage, which, to a certain extent, serves to improve urban economic resilience.

  3. (3)

    Industrial structure rationalization exerted a single threshold effect on urban economic resilience in the northeastern region. When globalization occurred below the threshold value of 6.63, the regression coefficient of industrial structure rationalization was −0.003, which is not significant. When globalization exceeded the threshold value of 6.63, the impact of industrial structure rationalization on urban economic resilience remained inhibitory, with a regression coefficient of −0.021. The reason for this may be that, as an old industrial base in China, the northeastern region exhibits a heavy industrial structure and relatively slow development, which is less favored by foreign direct investment. During the sample period, with the promotion of the Northeast Revitalization strategy and the implementation of a series of supporting measures, the evolution of the industrial structure also showed a continuous trend of improvement. However, due to the failure to achieve the ideal goal of upgrading traditional industries, they often do not exhibit market competitiveness, so the promotion effect on urban economic resilience is relatively limited.

  4. (4)

    The western region differs from other regions and did not show a threshold effect. We believe that the reason for this may be that the Western region implemented only the Western Development strategy during the early 21st century. Transportation, communication, and other infrastructure in the western region are relatively outdated, and the opportunities for foreign direct investment are lower than those in the eastern, central, and northeastern regions. As a result, the industrial spillover effect and positive externality effect in the western region are not significant. Therefore, even at the advanced stage of globalization, industrial structure rationalization and upgrading in the Western region do not significantly impact urban economic resilience.

Table 6 Regression results of the threshold effect in the Theil model on GL by subregional samples.
Full size table

Conclusions and policy recommendations

In this study, 267 prefecture-level cities and above in China were adopted as the research object, and globalization was considered a threshold variable to couple industrial structure upgrading globalization and urban economic resilience. We empirically assessed the nonlinear impact of industrial structure upgrading on urban economic resilience. The results indicated that, from a national perspective, the impact of industrial structure rationalization and upgrading on urban economic resilience exhibited a nonlinear relationship, and both showed a double threshold effect. At the initial stage of globalization, industrial structure rationalization promotes urban economic resilience, while industrial structure upgrading negatively impacts urban economic resilience. After globalization crosses the first threshold and enters the intermediate stage, industrial structure rationalization and upgrading exert a positive promoting effect on urban economic resilience, and industrial structure upgrading shifts from a negative to a positive promoting effect, which is significant. When urban economic resilience crosses the second threshold and enters the advanced stage, industrial structure rationalization and upgrading exert a significant positive effect on urban economic resilience. By removing extreme values from the sample, controlling for time and individual interaction terms, including control variables, and then modeling again, it was found that the above conclusion still holds and passes the robustness test. To empirically analyze regional heterogeneity, we divided the sample into eastern, central, northeastern, and western regions. We found that industrial structure upgrading still imposes a dual threshold effect on urban economic resilience in the eastern and central regions with globalization as the threshold variable, but the performance varied among the different regions. The threshold effect of industrial structure upgrading (including rationalization and upgrading) in the eastern and northeastern regions on urban economic resilience was characterized by an inverted U-shaped trend. In particular, the northeastern region still occurred on the left side of the inverted U-shaped curve, with only a single threshold effect. No threshold effect was observed for the western region.

Based on the research findings presented here, the following policy recommendations are proposed: firstly, we should continue to accelerate industrial transformation and upgrading, thus promoting high-quality development of urban economic resilience. We should actively promote regional industrial structure upgrading and, through the development of high value-added new economy sectors, we should embark on accelerated internal economic structure optimization, thereby enhancing urban economic resilience. Consequently, to improve the rationalization of the regional industrial structure, reasonable leading industries and strategic industries should be selected for driving economic development to support the development of leading industries, extend the industrial chain of products and enhance their added value, achieve differentiated competition and gradient transfer of industries, construct a circular economic development model, optimize the layout and management of the entire industrial chain, and establish a reasonable distribution of upstream, midstream, and downstream resources in the industry. Secondly, we should increase openness and innovation efforts to promote high-quality development through high-level openness. We should make good use of the national independent innovation demonstration zone platform, highlight open innovation, enhance innovation capabilities, promote independent innovation through open innovation, and continuously improve our competitiveness in opening up to the outside world. This would allow for increased efforts to “bring in and go out,” i.e., cultivating more market entities that are open to the outside world. We should continuously optimize the export structure, encourage and support local enterprises to “go global”, attract high-quality foreign investment, and promote the “optimal entry and exit” of opening up to the outside world. Thirdly, regional governments should implement differentiated measures based on the actual situation in their respective regions to avoid one-size-fits-all measures. The level of globalization in the eastern and central regions is already high, and it is necessary to focus on accelerating the development of high-tech industries, especially the updating and development of industries such as those in information, biology, new materials, aerospace, and ocean fields. The industrial structure and globalization levels in the northeastern and western regions are relatively low. As such, it is necessary to accelerate the replacement of new and old industries, vigorously promote the integration of informatization and industrialization, and thereby improve urban economic resilience. In the western region, it is necessary to continue industrial transfer from the eastern and central regions, optimize the regional industrial structure layout, create high-quality service platforms, focus on cultivating leading enterprises, fully leverage the siphon effect to drive industrial agglomeration, and create a bridgehead for opening up to the west, leveraging the advantages of large ports, channels, logistics, hubs, and trade. In Northeast China, efforts should be made to consolidate inventories, achieve incremental expansion, extend the industrial chain, and increase the added value. The digital, networked, and intelligent transformation of traditional manufacturing industries should be accelerated, promoting the extension of the industrial chain both upstream and downstream and finally creating a relatively complete industrial chain and cluster.

Further discussion

Industrial structure upgrading is an effective way to improve the urban economy, while globalization is an important driving force for regulating the correlation between urban industrial structure upgrading and urban economic resilience (Cheng et al. 2022). Although scholars have explored the coupling relationship between industrial structure upgrading and economic resilience, there is still insufficient research on the relationship between industrial structure upgrading, globalization, and urban economic resilience using globalization as a threshold. Compared with existing research, the similarity with previous studies lies in the fact that both have validated the nonlinear relationship between industrial structure upgrading and urban economic resilience. The difference from previous research lies in that, from a research perspective, we have taken the lead in placing industrial structure upgrading, globalization, and urban economic resilience within the same framework, and systematically elaborated on the relationship between the three, Fully considering the impact of industrial structure upgrading on urban economic resilience under the regulation of globalization is a beneficial supplement to existing theoretical research. From the perspective of research content, further subdividing the upgrading of industrial structure, using globalization as a threshold variable, exploring the threshold effect of industrial structure upgrading and industrial structure rationalization on urban economic elasticity, revealing the nonlinear relationship and stage characteristics between industrial structure upgrading and urban economic elasticity, is a strong challenge to the traditional research conclusion that there is a coupling relationship between the two. From the empirical results, although it cannot be said that the framework constructed in this article solves the “black box” problem of urban economic elasticity, it can indeed indicate that there is a linear relationship between the advancement and rationalization of industrial structure and urban economic elasticity, and it has a significant dual threshold effect and regional heterogeneity.

For developing countries, globalization is a double-edged sword. On the one hand, globalization can accelerate the development of international trade and close the global wealth gap (Xuan et al. 2023). On the other hand, globalization can create a more stimulating environment of international market competition (Shang, 2022). Research has also shown that some developing countries have deviated from their actual problems and needs, are blindly integrated into the global development system, and are constrained by factors such as industrial technology, product quality, and talent, becoming the weakest link in the global industrial chain (Carlos et al. 2022). The invasion of Western cross-border capital and industries has undoubtedly enormously impacted the economic and social development of developing countries, leading to issues such as the fragility of urban economic resilience and limited defense capabilities. Since China’s integration into the global economic system, it has adhered to a mutually beneficial and win-win strategy of opening up to the outside world. While fully leveraging its own endowments, it is adept at innovative integration of foreign and local technologies. Based on the advanced and rational industrial structure, it has increased investment in intelligent manufacturing and digital transformation, achieving the transformation from traditional manufacturing to intelligent manufacturing in China. At the same time, it enriches the content of the urban industrial system, improves industrial production efficiency and quality, expands international market competitiveness advantages, and promotes the formation of urban economic resilience. In addition, in recent years, China has also actively promoted the the Belt and Road Initiative, aiming to optimize the industrial structure system of countries along the Belt and Road through infrastructure construction, trade and investment, financial cooperation, and other means, while increasing the economic closeness of countries along the Belt and Road, so as to promote high-quality economic development of cities along the Belt and Road. This is also the main reason why China’s economy is still able to recover and maintain rapid growth under the impact of the financial crisis in 2008 and the COVID-19 epidemic. Although China still has a long way to go in enhancing urban economic resilience, it can undoubtedly provide important experiences for other developing countries to achieve the same goals.

It should be noted that this study has certain limitations. Firstly, regarding the measurement of urban economic resilience indicators, in existing research, the core variable method and the comprehensive indicator method have mostly been used to measure urban economic resilience, but no consensus has been reached thus far. The measurement indicator method selected in this study focused more notably on the representativeness and continuity of the core variable method, while the output method was used to measure urban economic resilience. Should periodicity be considered? This question may yield a direction for future research. Secondly, there may be certain shortcomings in the measurement of indicators for industrial structure advancement and rationalization, which cannot restore the comprehensiveness of industrial structure upgrading. Can industrial structure upgrading be further classified? This issue must be improved upon in the future. Finally, there were certain missing values for the research area considered in this article (especially in the western region of China). If further consideration is given to all prefecture-level cities and above, it cannot be ruled out that this may impact the conclusions of this article, which may also provide a direction for future research.