2.2.1. ICPC and industrial resilience.

As a key initiative for promoting the digital transformation of cities, the ICPC policy enhances industrial resilience through three main channels. First, it broadens the application and deepens the penetration of information technology in pilot cities, thereby accelerating the digital transformation and intelligent upgrading of industrial chains. This transformation strengthens the industrial chain’s resilience by improving information accessibility and flow, making it less vulnerable to external shocks [32,33]. Second, the ICPC policy, as a critical driver of new consumption and a lever for economic circulation, promotes the upgrading of consumption structures and the expansion of domestic demand by stimulating intelligent, personalized, and high-end consumption. This, in turn, drives the extension and development of industries [24,34].

Moreover, from the perspective of resource agglomeration and informatization, the ICPC policy accelerates the development of information infrastructure in pilot cities and fosters the integration of information technology across various industrial chain links. This facilitates the agglomeration and sharing of innovative resources, while providing enterprises with a more efficient external environment, thus lowering the costs of entrepreneurial activities. The resulting denser industrial chain network enhances the innovation capacity of urban industries, promoting effective collaboration and increasing adaptability to market fluctuations, ultimately improving resilience [35]. Therefore, we propose the following hypothesis:

H1: The ICPC policy can effectively enhance the urban industrial resilience.

2.2.2. Mechanism hypothesis.

Based on the specific objectives of the ICPC policy and the analysis outlined in the previous sections, this study seeks to explore the mechanisms through which the ICPC policy enhances industrial resilience. The analysis will focus on three key aspects: information development, entrepreneurial agglomeration, and digital finance, as illustrated in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Mechanisms of the ICPC policy on urban industrial resilience.

https://doi.org/10.1371/journal.pone.0323101.g001

(1) Information development effect. The ICPC policy explicitly outlines several targeted measures aimed at accelerating the improvement of information infrastructure and enhancing the informatization of public services in pilot cities. For instance, the policy emphasizes the construction of high-speed broadband networks, the deployment of 5G infrastructure, and the promotion of smart city platforms [30,31]. These measures not only facilitate the overall informatization development of these cities but also create a robust foundation for digital transformation across various industries. Firstly, an increase in the level of informatization facilitates the digital transformation and intelligent upgrading of industries, significantly improving the speed of information acquisition, transmission, processing, and feedback across various links of the industrial chain [23]. This enables enterprises to respond more rapidly to market changes and adjust their production strategies accordingly, thus enhancing the adaptability and flexibility of the industrial chain.

Secondly, informatization promotes efficient collaboration and optimized resource allocation among the different nodes of the industrial chain. Through data sharing and information integration, supply chain management and collaborative innovation capabilities are strengthened, reducing uncertainties and risks in the production process [36]. Particularly in the face of emergencies or external shocks, a high level of informatization enables the industrial chain to quickly mobilize and integrate resources, forming a rapid response mechanism that mitigates potential economic losses [37]. The deep application of information technology also allows enterprises to innovate production processes and optimize operational models, further enhancing the competitiveness and sustainable development capacity of the industrial chain. Therefore, we propose the following hypothesis:

H2: The ICPC policy enhances the urban industrial resilience by promoting the information development.

(2) Entrepreneurial agglomeration effect. Firstly, ICPC cities foster new digital consumption formats and, coupled with promising economic development prospects, generate a substantial number of job opportunities for the region. This attracts the agglomeration of production factors such as labor and capital, providing essential resource support for entrepreneurial activities. Secondly, The ICPC policy creates a digitalized, information-driven business environment that effectively reduces information asymmetry and lowers entrepreneurial costs, thereby stimulating the agglomeration of entrepreneurial activities in cities [24]. Specifically, these policies equip cities with efficient digital infrastructure, significantly enhancing the information mobility and resource-sharing capabilities among enterprises. This enables startups to rapidly access market information, technical support, and financial resources [5,38], accelerating the industrialization of innovative outcomes. Additionally, the expansion of information consumption continuously generates new market demands, attracting more enterprises to enter related fields.

Existing literature indicates that the agglomeration of entrepreneurial activities effectively enhances the resilience of urban industrial chains [39,40]. When startups with digital technology advantages enter the market, increased information flows and technical exchanges encourage firms to overcome inertia and strengthen collaborative innovation, promoting closer cooperation and resource-sharing among upstream and downstream participants in the industrial chain. This coordinated and efficient collaboration network mitigates industrial risks and improves the overall ability of the industrial chain to withstand market fluctuations and external shocks [41]. Furthermore, entrepreneurial agglomeration promotes diversification and specialization within the industrial chain. The refinement of the intra-industry division of labor enhances the overall efficiency of urban industries and strengthens complementarity and coordination within the industrial chain, enabling it to maintain greater adaptability and resilience in complex and volatile market conditions [40,42]. Therefore, the following hypothesis is proposed:

H3: The ICPC policy enhances the urban industrial resilience by promoting entrepreneurial agglomeration.

(3) Digital finance effect. The implementation of the ICPC policy significantly enhances the digitalization and inclusiveness of financial services. First, by promoting the integration of municipal e-government platforms with digital payment systems, the policy fosters the widespread adoption of mobile payments, contactless transactions, and intelligent financial services. Specific initiatives, such as subsidies for digital payment infrastructure and tax incentives for fintech firms, have accelerated the deployment of financial technologies in urban economies. Second, the establishment of government-led credit information sharing platforms, particularly through municipal big data centers, improves financial institutions’ ability to conduct data-driven credit assessments and risk management. This reduces information asymmetry, lowers transaction costs, and expands the accessibility of financial services. Third, regulatory policies under the ICPC framework encourage collaboration between traditional financial institutions and fintech enterprises, facilitating innovation in digital lending, supply chain finance, and inclusive credit products.

With the expansion of digital financial infrastructure, enterprises, particularly small and medium-sized enterprises (SMEs), can access financing and other financial services more seamlessly. On one hand, this reduction in financing costs alleviates financial strain on capital chain, thereby strengthening the resilience of industries to risks [7]. Concurrently, financial technologies, such as digital inclusive finance, utilize big data, artificial intelligence, and other digital tools to mitigate resource misallocation and effectively facilitate inter-firm collaboration and resource sharing [43]. This establishes a solid foundation for improving the coordination and adaptability of urban industries [44]. Consequently, industrial chains maintain stability and resilience amid market fluctuations and external shocks through more agile capital operations and resource allocations. On the other hand, digital finance, with its high accessibility, reduces transaction costs and enhances payment convenience, promoting the upgrading of residents’ consumption capacity and structure [45]. This consumption upgrade not only expands market demand but also spurs new demands and business formats, prompting enterprises to engage in technological innovation and product optimization. This accelerates the upgrading and renewal of products and services, driving the industrial structure towards high-end and diversified development [44]. Such optimization strengthens industries’ adaptability and recovery capacity in response to external risks, further enhancing overall resilience. Based on the above analysis, the following hypothesis is proposed:

H4: The ICPC policy can effectively enhance the urban industrial resilience by promoting the development of digital finance.

3.2.1. Explained variable.

Urban industrial resilience refers to the capacity of an industrial system to maintain stability and continuity in the face of external shocks, such as economic crises, natural disasters, or supply chain disruptions [47]. This resilience not only reflects the system’s ability to withstand risks but also encompasses its speed of recovery and adaptability to new environments post-crisis. Given the multifaceted nature of resilience, a single indicator may inadequately capture its complexity, necessitating a multidimensional comprehensive assessment. Scholars generally acknowledge that both shock resistance and transformational recovery capacity are critical dimensions of urban industry resilience [46,48]. Therefore, this study, drawing on the frameworks established by Di et al.[48] and Li [46], measures resilience comprehensively through the dimensions of shock resistance and transformational recovery capacity, utilizing available prefecture-level city-data.

Shock resistance is assessed using the city’s industrial diversity index, which is represented by the reciprocal of industrial concentration, specifically the Herfindahl-Hirschman Index. This metric evaluates the relative significance of various industries within the urban economy, reflecting the capacity for resource allocation and risk diversification in response to external shocks. Additionally, transformational recovery capacity is measured by the number of authorized invention patents—a metric widely recognized in the academic community [17,20,49]. Its theoretical foundation is primarily reflected in three aspects. First, the number of authorized invention patents reflects the development level of the industrial chain in terms of innovation resource investment, the construction of an innovative environment, and skill upgrading. Enhanced innovation accelerates the development and refinement of new products, thereby improving the efficiency of industrial renewal and iteration [46,50]. Second, patent counts indicate the diversification of technological innovation and knowledge accumulation within a region. As regional technological capabilities and reserves become increasingly diversified, the depth and breadth of technological knowledge expand, enriching the regional technological structure. This not only signifies breakthroughs in individual technologies but also demonstrates an overall improvement in regional technological levels, thereby providing a solid knowledge base for industrial entities to adjust and restore their interrelationships to a normal or even more optimal state in the face of potential market risks and uncertainties [40]. Third, high levels of patent activity are often accompanied by the emergence of new technologies and industries, which can accelerate the diversification and adjustment of the regional industrial structure. Under the pressure of external shocks, new technologies often become an important tool for industries to seek breakthroughs and rapid development [37,51]. In summary, using the number of authorized invention patents as an indicator for transformational recovery capability not only directly reflects the level of technological innovation but also reveals the underlying mechanisms by which regions construct complex technological networks and promote industrial diversification, fully embodying their capacity to transform and upgrade in response to external environmental changes.

Finally, following the approach of Yang and Liu [17], this study employs the panel entropy weight method to comprehensively calculate the comprehensive index of Res (where, the weights of shock resistance and transformational recovery capacity were weighted 0.159 and 0.841, respectively). The calculation steps are as follows:

First, data standardization:

(2)

(3)

Where, X’ represents the standardized value of indicator for city in year ,  = 1, …, ; = 1, …, ; = 1, …, .

Second, calculation of indicator proportion:

(4)

Third, calculation of information entropy:

(5)

Fourth, calculation of information utility value:

(6)

Fifth, determination of indicator weights:

(7)

Sixth, calculation of the comprehensive industrial resilience score:

(8)

3.2.2. Explanatory variable.

The explanatory variable in this study is , a dummy variable used to measure whether a city has been designated as an ICPC pilot. If a city is included in the list of ICPC pilots, takes a value of 1 in the year it is included and for all subsequent years; otherwise, it takes a value of 0.

3.2.3. Control variables.

Building on the practices of previous studies [5,30,31], to ensure the reliability of the benchmark regression results, this paper controls for a series of variables that may potentially affect the urban industrial resilience. Specifically, these include: economic development level (), measured by the natural logarithm of the city’s per capita GDP; financial development (), proxied by the ratio of total loans from banks and other financial institutions to GDP, following common practices in existing research; human capital (), measured by the natural logarithm of educational expenditure; fiscal pressure (), quantified as the ratio of general government revenue to GDP; average wage (), measured by the natural logarithm of the average wage.

4.5.1 Excluding the effects of other policies.

To address the potential impact of concurrent policies during the sample period on the baseline results, this study adopts the approach of Zhang et al. [30] by carefully reviewing the policy objectives in place during the study timeframe. We specifically exclude interference from three policies that are closely related to information infrastructure development: “Broadband China,” “National E-commerce Demonstration Cities,” and “Big Data Comprehensive Pilot Zones.” Table 4 presents the estimated effects of the ICPC policy on industrial resilience after incorporating each of these three interfering policies in Columns (1) to (3). The results show that after controlling for the influence of other information policies, the coefficients of the remain positive and significant at the 1% level. This indicates that the ICPC policy continues to effectively enhance industrial resilience, even when accounting for the effects of concurrent policies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Excluding other policy effects.

https://doi.org/10.1371/journal.pone.0323101.t004

4.5.2. Heterogeneity treatment effect test.

Since the pilot cities were not established simultaneously but constructed in phases, the estimation results of the baseline model may be subject to TEFW fixed effects bias. To address this issue, this study employs two key approaches: Bacon decomposition and heterogeneity-robust estimators. First, the Bacon decomposition method proposed by Goodman-Bacon [59] is used to examine potential heterogeneity in treatment effects. The results, presented in Panel A of Table 5, show that the weight of the inappropriate treatment effect (Later T vs. Earlier C) is only 6.3%. This indicates that the heterogeneity of treatment effects in the overlapping DID model is minimal, suggesting that the benchmark regression results are robust.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Heterogeneity treatment effect results.

https://doi.org/10.1371/journal.pone.0323101.t005

Second, to further mitigate the potential impact of heterogeneous treatment effects on the findings, three state-of-the-art robust estimators from recent research are applied. These include: (1) The method proposed by De Chaisemartin and D’Haultfoeuille [60], which addresses the bias caused by staggered treatment adoption and provides unbiased estimates under complex time distributions. (2) The interaction-weighted estimator developed by Sun and Abraham [61], which accounts for heterogeneous treatment effects across groups and time periods. (3) The interpolation-based counterfactual method introduced by Borusyak et al. [62], which corrects for TWFE estimation bias and ensures robustness in multi-period settings. As shown in Panel B of Table 5, the regression coefficients of Did_inf are all significant at the 1% level across all three methods. These results confirm that information consumption has a significant positive effect on industrial resilience, even after accounting for potential biases caused by heterogeneous treatment effects. In conclusion, by addressing TEFW bias and heterogeneous treatment effects through Bacon decomposition and multiple robust estimators, strong evidence is provided to support the positive impact of information consumption on industrial resilience.

4.5.3. Other robustness tests.

To further validate the robustness of the baseline conclusions, we conduct a series of robustness checks from several perspectives:

1. (1). Addressing selection bias. Policymakers might have prioritized cities with advanced economic development and information infrastructure when selecting information consumption pilot cities, leading to inherent differences between pilot and non-pilot cities. To address this, we use the control variables in Model (1) as covariates and employ caliper matching and nearest neighbor matching methods to match the treatment group annually. The results of the PSM-DID analysis are presented in Columns (1) and (2) of Table 6.
2. (2). Winsorization. To avoid potential biases caused by extreme values of the explained variable, this study applies a 1% winsorization treatment to both the upper and lower tails of the distribution and re-estimates the model using the trimmed sample. The results are shown in Column (3) of Table 6.
3. (3). Exclusion of special cities. Recognizing the significant differences in economic development and resource allocation between municipalities such as Beijing, Shanghai, and other cities, this study excludes these four municipalities from the sample and re-runs the regression. The results are presented in Column (4) of Table 6.
4. (4). Stringent clustering. To further mitigate the potential impact of heterogeneity among cities on the results, this study adjusts the clustering method to a stricter provincial level and re-estimates the model. The results are displayed in Column (5) of Table 6.
5. (5). Controlling for provincial-year interaction effects. City-level data are often influenced by various provincial policies that vary over time. These provincial-level factors, if uncontrolled, may act as omitted variables and bias the estimation results. Therefore, we control for the interaction terms between provinces and years in the baseline model to reduce errors caused by omitted variables. The results are reported in Column (6) of Table 6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Other robustness tests.

https://doi.org/10.1371/journal.pone.0323101.t006

The results consistently show that, regardless of the method used to address sample selection bias, mitigate the influence of extreme values, or control for omitted variables, the coefficient for remains positive and significant at least at the 5% level. This indicates that the positive effect of the ICPC policy on enhancing industrial resilience holds and is robust across a variety of robustness checks.

5.2.1. Locational heterogeneity.

There are notable regional disparities in economic development, industrial structure, and the degree of information technology adoption, which may result in uneven effects of the ICPC policy across different areas [31]. Columns (1) to (4) of Table 8 display the regression results for the eastern, central, western regions, and central-western respectively. The findings reveal that the policy has no significant impact on industrial resilience in the western regions, while it significantly enhances the resilience of urban industrial chains in both the eastern and central regions. Importantly, the regression coefficient for the eastern cities is the largest and statistically significant at the 1% level. To further examine whether the policy effects differ significantly across regions, we conducted coefficient difference tests between groups. The results indicate that the policy effect in eastern regions is significantly higher than that in central and western regions at the 1% level, while the difference between central-western regions is also statistically significant. These findings highlight the existence of locational differences in policy effectiveness across regions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Regional location heterogeneity test results.

https://doi.org/10.1371/journal.pone.0323101.t008

Possible explanations for this include the stronger economic foundations, optimized industrial structures, and advanced information technology development in eastern regions, which allow them to better leverage the opportunities presented by the ICPC policy. In these cities, information consumption drives the widespread adoption of information technology, facilitating the digital transformation of industrial chains and enhancing innovation capacity [7,21]. In contrast, the central and western regions face challenges such as lower levels of economic development, more monolithic industrial structures, and weaker information technology infrastructure, limiting the effectiveness of the pilot policies due to insufficient external support. The relatively underdeveloped information infrastructure in these regions hinders the penetration and application of information technology in industries [54]. Enterprises in these areas often lack the necessary technical support and financial investment for their digital transformation, leading to limited improvements in industrial resilience through information consumption. As a result, the positive effects of the ICPC policy on industrial resilience are not significantly observed in these cities.

5.2.2. Urban scale heterogeneity.

Large-scale cities and small-to-medium-sized cities differ significantly in terms of economic foundations, market demands, and levels of informatization, which may result in varying outcomes of the ICPC policy in enhancing industrial resilience. To account for these differences, we conduct heterogeneity tests to assess the specific policy impacts across different city sizes, ensuring the precision and effectiveness of policy implementation. Following the classification method of Liu et al. (2024) [5], cities are categorized into large-scale and small-sized cities based on their permanent resident populations. Columns (1) and (2) of Table 9 present the regression results for the impact of the ICPC policy on industrial resilience in large and small cities, respectively. The results indicate that the regression coefficients of the ICPC policy (Did_inf) are significantly positive for both large and small cities, though the coefficient for large cities is notably higher. Specifically, the estimated coefficient for large cities is 0.025, significant at the 1% level, while for small cities, the coefficient is 0.008, significant at the 10% level. This suggests that the ICPC policy has a more pronounced effect on enhancing industrial resilience in large cities. Moreover, the coefficient difference test confirms that the impact of the ICPC policy is significantly greater in large cities than in small cities at the 1% significance level.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 9. City size heterogeneity test results.

https://doi.org/10.1371/journal.pone.0323101.t009

Possible reasons for this include the more advanced infrastructure, richer industrial resources, and stronger innovation capacities typically found in large cities, which may amplify the positive impact of the ICPC policy [5,37]. Large cities generally have better access to financial resources, a higher concentration of technology-intensive industries, and a more dynamic business environment, all of which contribute to a stronger policy response. In contrast, small cities may face constraints such as weaker industrial foundations, lower levels of digitalization, and limited access to high-quality human capital, which could hinder the full realization of policy benefits. The significant disparity in policy effects across city sizes underscores the importance of tailored policy strategies that consider the unique characteristics and development needs of different urban areas.