The direct impact effects of data factor on regional economic resilience are manifested across multiple dimensions, realized through various institutional arrangements and practical pathways. Firstly, abundant data factor can contribute to enhancing information transparency and decision-making scientificity. Enterprises can utilize various digital platforms and sensors to collect real-time data on market dynamics, consumer demands, supply chain conditions, and other aspects, directly aiding in the scientific formulation of production plans and market strategies to cope with complex and diverse economic shocks. Government departments can also leverage big data technology to integrate data resources from different fields and departments, comprehensively and timely understanding economic operating conditions, including macroeconomic indicators, industrial development trends, and regional economic disparities, to accurately predict economic trends and formulate more targeted and scientific policy measures, improving the effectiveness and precision of macroeconomic regulation. Secondly, the market-oriented allocation of data factor, characterized by the establishment of data trading platforms, can significantly enhance regional economic resilience by strengthening economic resistance and recovery capabilities, adaptability and adjustment capabilities, and innovation and transformation capabilities. This allocation method reduces information asymmetry, optimizes the flow efficiency of production factors, and enables the regional economic system to exhibit greater stability and recovery capacity when facing external shocks. Thirdly, the agglomeration of data factor formed through the establishment of comprehensive big data pilot zones can have a direct and significant growth effect on regional economic resilience. This agglomeration effect not only promotes the large-scale utilization of data resources but also fosters a virtuous cycle of technological innovation and industrial upgrading, injecting new vitality into regional economic development. Finally, the construction and improvement of the data factor market itself, as a fundamental institutional arrangement, provides a fundamental guarantee for the full release of data value by establishing data property rights, pricing, and trading rules, and its construction process is itself an important force driving economic resilience enhancement ([18]). In summary, as a new type of critical production factor, data factor can directly enhance the ability of regional economic systems to resist external shocks and strengthen the robustness and resilience of economic development. Based on this, this paper proposes the following hypothesis:

H1:The development of data factor has a positive driving effect on the enhancement of regional economic resilience.

The role of data factor in enhancing regional economic resilience through the “innovation-driven effect” has been supported by numerous empirical studies. The core mechanism lies in the fact that the agglomeration of data factor injects strong impetus into regional innovation systems. Specifically, firstly, the agglomeration of data factor can effectively guide governments and enterprises to increase R&D investment. Taking the establishment of national-level big data comprehensive experimental zones as an example, such policies have significantly attracted the clustering of high-tech enterprises by providing financial support and creating a favorable innovation environment, thereby expanding the scale of regional innovation activities ([27]). Secondly, data factor helps uncover innovation points and enhance the scientificity and efficiency of the innovation process. Vast amounts of market data, user data, and technological data contain rich information. Through big data analysis, artificial intelligence, and other technological means, enterprises and research institutions can discover potential market demands, technological pain points, and innovation opportunities from these data, conduct precise market research and technological route planning, and optimize R&D decisions, thereby reducing the uncertainty and trial-and-error costs of innovation activities. Thirdly, the circulation and sharing of data factor can promote the formation of knowledge spillovers and a collaborative innovation ecosystem. The construction of infrastructure, such as data trading platforms breaks down information barriers, provides a platform for in-depth interactions among enterprises, universities, and research institutions, accelerates the creation and diffusion of new knowledge and technologies, and speeds up the commercialization process of innovation achievements. In summary, data factor can systematically enhance regional scientific and technological innovation capabilities through multiple pathways, and innovation, as the core driving force for economic systems to cope with shocks and achieve transformation and upgrading, will ultimately be transformed into stronger resistance, adaptability, and recovery capabilities of regions in the face of external risks, namely, higher economic resilience. Based on this, this paper proposes the following hypothesis:

H2:Data factor can enhance economic resilience by promoting innovation-driven development.

As the core production factor of the digital economy, data factor significantly enhances the risk resistance and recovery adjustment capabilities of economic systems by empowering the development of digital finance, becoming an important pathway for enhancing economic resilience ([15]). Firstly, data factor can drive the development of digital finance through the “property rights determination-trading-financialization” pathway, thereby enhancing economic stability. Data capitalization can transform data into tradable assets, empower digital financial innovation in credit assessment and risk pricing models, optimize capital allocation efficiency, improve the precision and inclusiveness of financial credit services, and effectively enhance the resistance and recovery capabilities of regions in the face of external shocks ([10]; [4]). Secondly, digital finance can alleviate information asymmetry between enterprises and financial institutions, reduce factor transaction costs, assist financial institutions in scientifically developing financial products and services that better meet market demands, and enhance the adaptability and transformation capabilities of regional economies through mechanisms such as improving entrepreneurial activity, strengthening enterprises’ ability to cope with risks, and narrowing the urban-rural income gap. It should be noted that the promoting effect of digital finance on economic resilience exhibits structural differences and time-series increasing characteristics, with its empowering effect being more significant in regions with higher levels of technological innovation and deeper digital integration ([11]). In summary, data factor can systematically enhance the comprehensive capabilities of regional economies in resisting shocks, rapid recovery, and structural transformation by driving the deepening development of digital finance, providing key support for building a highly resilient economic system. Based on this, this paper proposes the following hypothesis:

H3:Data factor can enhance economic resilience by improving digital finance.

As the core driving force for economic development, data factor significantly enhances regional economic resilience by promoting the optimization of industrial structure, a key mediating pathway. Data factor can freely flow and be shared among different industries, breaking down industry boundaries, promoting cross-industry integration, and giving rise to new business forms such as smart manufacturing and industrial internet; they can combine with other technologies and resources to breed a series of emerging industries, such as big data, cloud computing, artificial intelligence, and blockchain industries; they can also drive the digital transformation of traditional agriculture and service industries, achieving improvements in production efficiency, cost reduction, and product quality. Industrial structure optimization is not a single-dimensional upgrade but a systematic evolution of the industrial system towards higher efficiency, stronger adaptability, and better coordination, playing an irreplaceable role at different stages of economic resilience. Specifically, during the resistance period, data-empowered industrial structure optimization enhances the resistance of economic systems to external shocks by improving resource allocation efficiency and risk dispersion capabilities; entering the recovery period, the optimized industrial structure accelerates resource reorganization and path innovation, promoting rapid economic recovery through its good industrial linkages and techno-economic connections; and during the adjustment period, industrial structure optimization provides stronger structural transformation capabilities for economic systems, driving the formation of new growth paths ([28]). Industrial structure optimization plays a key mediating role in enhancing economic resilience in the region, and data factor systematically enhances the resistance, recovery, and reconstruction capabilities of regional economies through this pathway, laying a solid industrial structure foundation for building a highly resilient economic system ([26]). Based on this, this paper proposes the following hypothesis:

H4:Data factor can enhance economic resilience by optimizing industrial structure.

This paper empirically examines the impact of data factor on economic resilience by constructing a benchmark regression model as shown in Equation (1), aiming to provide empirical evidence for understanding its role in responding to economic shocks.

where, i and t represent province and year, respectively; RES i t denotes the level of economic resilience; DAT i t is the core independent variable, representing the development of data factor; C i t is a series of control variables; α 0 , α 1 , and α 2 are the parameters to be estimated in the model; μ i and λ t represent province and time fixed effect, respectively; and ε i t is the random disturbance term.

To verify the mediating roles of three mechanism variables, namely, innovation-driven effect, digital finance level, and industrial structure optimization, models as shown in Equation (2) and Equation (3) are constructed to test the mediating effects of these variables ([8]).

where, M i t represents the mediating variable, which includes innovation-driven effect (IDE), digital finance level (DFL), and industrial structure optimization (STR), respectively. To depict the regional innovation level, the proportion of internal R&D expenditure in gross domestic product (GDP) is used as a proxy variable for the innovation-driven effect. The digital finance directly utilizes the Digital Inclusive Finance Index released by Peking University. For measuring industrial structure optimization, this paper adopts the method used by [5] and constructs a composite indicator based on the labor productivity and value-added share of each industry. The specific formula is shown in Equation (4):

where, j represents the primary, secondary, and tertiary industries; G i j t denotes the proportion of the value-added of the j-th industry in region iin year t to the regional GDP; L i j t represents the labor productivity of employees in the j-th industry in regioniin year t, that is, the ratio of the value-added of each industry to the total employment of that industry. The meanings of the remaining variables in Equation (2) and Equation (3) are the same as those in Equation (1).

3.2.1. Dependent Variable: Economic Resilience (RES)

This study utilizes the economic resilience index and its data constructed by [6]. The index establishes a comprehensive evaluation system comprising 26 indicators across four core dimensions: risk resistance, risk recovery, adaptive adjustment, and innovation transformation.

3.2.2. Independent Variable: Development of Data Factor (DAT)

To overcome the direct quantification challenges of data factor due to their implicit characteristics and external environmental dependencies, this study constructs a multidimensional comprehensive evaluation system based on the research frameworks of [16] and [22]. The system encompasses three dimensions: foundational support, application level, and transformation efficiency (see Table 1). This indicator system is designed to systematically unveil the entire production and application process of data factor, spanning from foundational conditions to practical applications and culminating in ultimate value creation. In doing so, it comprehensively reflects the development of regional data factor, with measurements conducted through the entropy-weighted TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) method.

Table 1. Indicator system for the development of data factor.

Continued

3.2.3. Control Variables

To control for the influence of other potential factors on economic resilience and mitigate omitted variable bias, this study selects the following variables as control variables: 1) Industrialization level (IND): measured by the proportion of industrial added value in regional GDP; 2) Urban economic density (ECO): characterized by GDP per unit of land area; 3) Financial development level (FIN): represented by the proportion of financial industry added value at the end of the year in regional GDP; 4) Urbanization (URB): measured by the proportion of urban population in the total population; 5) Degree of fiscal decentralization (REV): measured by the ratio of local budgetary revenue to budgetary expenditure.

The data used in this paper consists of panel data from 30 provinces in China spanning from 2011 to 2023 (due to data availability, the sample excludes Tibet, Hong Kong, Macau, and Taiwan regions of China). Data on economic resilience, data factor, and control variables are sourced from the China Statistical Yearbook, China Statistical Yearbook on Science and Technology, China Population and Employment Statistical Yearbook, as well as statistical yearbooks of various provinces, statistical bulletins on national economic and social development of various provinces, and Mark Database (https://www.macrodatas.cn/). Missing values are filled in using the linear interpolation method. Descriptive statistical results for each variable are presented in Table 2. Based on the regional classification standards of National Bureau of Statistics of China, this paper divides China’s 30 provinces into four major regions (eastern, central, western, and northeastern) for analysis.

Table 2. Descriptive statistics of variables.

Continued

Overall, significant disparities exist in the development of economic resilience across regions, with the eastern region demonstrating relatively strong performance, while the western and northeastern regions lag behind. This reflects the uneven regional development characteristics and underscores the importance of in-depth exploration of mechanisms to enhance economic resilience. From the perspective of the core explanatory variable, the development level of data factor in the eastern region is significantly higher than that in other regions, reflecting the leading position of this region in digital transformation. Regarding mediating variables, the overall regional differences in the innovation-driven effect are relatively low, whereas digital finance levels and industrial structure optimization exhibit more pronounced gradient differences. Additionally, various control variables, such as the level of industrialization, urbanization, and fiscal decentralization, also exhibit substantial structural differences across regions, providing an important foundation for subsequent regional comparisons and empirical analyses.

To investigate the impact of development of data factor on economic resilience, this paper establishes a two-way fixed-effects model for analysis. As shown in Table 3, the model presented in Column (1) does not control for time and province fixed effect and does not include control variables. The coefficient for the development of data factor is 0.836, which is significant at the 1% level. After introducing two-way fixed effects in the model shown in Column (2), the coefficient remains at 0.855, with no change in significance, demonstrating strong robustness. Further, after incorporating control variables into the model shown in Column (3), the regression coefficient for the development of data factor remains significantly positive, indicating its stable promoting effect in enhancing economic resilience. This thereby validates the research hypothesis H1 proposed in this paper.

Table 3.Benchmark regression results.

Continued

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. Values in parentheses are standard errors. The same applies below.

To ensure the reliability of the benchmark regression conclusions, this paper conducts a series of tests targeting potential endogeneity issues and estimation robustness.

4.2.1. Endogeneity Test

To mitigate endogeneity issues in the benchmark model, this paper employs the Two-Stage Least Squares (2SLS) method, selecting the number of fixed-line telephones in 1984 as an instrumental variable for the development of data factor ([28]). The rationale for this choice is as follows. On the one hand, due to the significant path-dependence mechanism in technological development, the early fixed-line telephone networks, as physical carriers for information transmission, shaped the layout of subsequent Internet infrastructure and the trajectory of technological adoption. Studies by researchers such as [7] suggest that the popularization of the Internet in China technically originated from the widespread adoption of fixed-line telephones, and regions with historically high fixed-line telephone penetration rates are also likely to have high Internet penetration rates in recent years. Therefore, selecting this variable as an instrumental variable satisfies the relevance hypothesis. On the other hand, with technological iterations, the direct impact of traditional communication tools like fixed-line telephones on the current economy has diminished, meeting the exogeneity condition. Since this variable is cross-sectional data and cannot be directly used in a panel model, it is multiplied by the number of Internet users in the previous year to construct an instrumental variable with a temporal dimension. To enhance the robustness of the test results, the first-order lag of the development of data factor is also used as another instrumental variable ([24]). The test results are shown in Table 4. The regression results in Column (1) and Column (3) both indicate a significant correlation between the instrumental variable and the independent variable. Meanwhile, both the under-identification test (Kleibergen-Paaprk LM) and the weak identification test (Cragg-Donald Weak Wald F) significantly reject the null hypothesis, indicating that the selection of instrumental variables is reasonable and effective. Based on this, the estimation results in Column (2) and Column (4) show that the impact of data factor on economic resilience remains significantly positive at the 1% level, suggesting that the core conclusion of the benchmark regression still holds after controlling for endogeneity issues.

Table 4.Test results (2 SLS).

4.2.2. Time Lag Test

To further examine the credibility of the benchmark regression results, this paper applies a second-order lag to the core independent variable (the development of data factor) to capture the persistence of its impact and reduce concurrent interference. The test results are shown in Table 5. The regression results in Column (1) indicate that after introducing the lagged term, the development of data factor still significantly promotes economic resilience, which is consistent with the benchmark regression conclusion.

Table 5. Robustness test results.

4.2.3. Adjusting the Sample Size

Considering that the sample size may affect estimation accuracy, this paper employs the bootstrap method for repeated sampling to enhance the robustness of the inferences. The results in Column (2) of Table 5 indicate that the estimated coefficient of the development of data factor remains significant in the simulated samples, suggesting that the core conclusion of the benchmark regression is less affected by sample randomness.

4.2.4. Replacing the Independent Variable

To mitigate potential interference from variable measurement methods on the results, this paper re-measures the development of data factor using principal component analysis and conducts regression tests again. The regression results in Column (3) of Table 5 show that the core conclusion of the benchmark regression still holds, indicating that the conclusion exhibits good robustness to different measurement methods for independent variables.

4.2.5. Replacing Control Variables

To control for the impact of potential omitted factors that evolve over time on economic resilience, this paper interacts all control variables in the benchmark model with time trend terms (using 2010 as the base year) to form a new set of control variables for estimation. The regression results in Column (4) of Table 5 demonstrate that, after considering the temporal heterogeneity of variable impacts, the coefficient for the development of data factor remains significantly positive.

4.2.6. Adjusting the Sample Scope

Finally, to enhance the generalizability of the benchmark regression conclusion, this paper re-estimates the model after excluding samples from municipalities directly under the central government, which may possess unique characteristics. The regression results in Column (5) of Table 5 reveal that, after adjusting the sample, the positive impact of development of data factor on economic resilience remains robust, indicating that the core conclusion of the benchmark regression exhibits good generality across different sample scopes.

To thoroughly investigate the regional heterogeneity in the impact of development of data factor on economic resilience, this paper conducts examinations from three dimensions: the geographical location, the development of data factor, and the foundational level of economic resilience itself.

Firstly, from the results of the regional heterogeneity test (see Table 6), there are significant regional differences in the impact of development of data factor on economic resilience. In the eastern, central, and western regions, the coefficients of development of data factor are all positive at the 1% significance level, indicating a robust enhancing effect on economic resilience. In the northeastern region, the coefficient of development of data factor fails to pass the significance test, preventing us from making statistically meaningful inferences. Among them, the central region has the highest coefficient, reflecting that data factor may have a stronger marginal effect in this region. This outcome may be related to the characteristics of the industrial structure and development stage in the central region. In recent years, the central region has been at a critical juncture of transitioning from traditional industries to digital and intelligent ones. The input of data factor can significantly enhance its industrial efficiency and optimize resource allocation, thereby unleashing greater incremental effects. In comparison, the eastern region has already reached a relatively mature level in terms of digital infrastructure and utilization of data resources, resulting in relatively diminishing marginal effects. Meanwhile, constrained by its relatively poor foundational conditions and level of industrial agglomeration, the western region experiences a relatively weaker driving effect from the data factor. In summary, the impact of data factor development on economic resilience exhibits notable spatial gradient characteristics, suggesting that relevant policies should be tailored to regional realities and implement differentiated strategies when promoting data factor development.

Secondly, to examine the heterogeneity in the impact of differences in the level of data factor development on economic resilience, this study divides the sample into a “high-level data factor development group” and a “low-level data factor development group” based on the data factor development level of each province, using the national average as the dividing standard. Regression tests are then conducted separately to reveal differences in the impact of development of data factor on economic resilience at different development stages. The results are shown in Columns (1) and (2) of Table 7. In the high-level data factor development group, the coefficient of DAT is 0.785, while in the low-level group, it is 0.957, indicating that the promoting effect of data factor development on economic resilience is more pronounced in regions with lower levels of data factor development. This may imply that in regions with weaker foundations in data factor, improving the level of data factor development can bring greater marginal benefits, thereby more effectively enhancing economic resilience.

Table 6. Results of geographical location heterogeneity test.

Table 7.Results of heterogeneity test on development of data factor and economic resilience.

Finally, to further analyze the impact mechanism of differences in economic resilience levels on how the development of data factor empowers the enhancement of economic resilience, this study divides the sample into a “high-level economic resilience group” and a “low-level economic resilience group” based on the varying levels of economic resilience across provinces. Subsequent analyses are then conducted separately to explore the differences in the enhancing effect of data factor development on economic resilience at different stages of economic resilience development. The results are shown in Columns (3) and (4) of Table 7. The results indicate that the coefficient of development of data factor in the high-level economic resilience group is significantly larger than that in the low-level group. This not only suggests that in regions with weaker economic resilience, the enabling role of data factor may be constrained by local overall development conditions but also implies that in regions with inherently stronger economic resilience, data factor can be more efficiently absorbed and utilized, thereby further consolidating and enhancing their economic resilience, presenting a development trend of “the strong getting stronger.”

In summary, we can identify a dual pathway through which the data factor exerts its influence. First, in regions with a relatively low initial level of data factor development, the development of data factor exhibits a pronounced “catching up” effect by addressing foundational shortcomings. Given that the digital infrastructure and data resource systems were originally relatively weak, the investment in the data factor represents a process of building from scratch or strengthening from a weak base, which can significantly enhance information circulation efficiency and resource allocation capabilities in the short term, thereby yielding higher marginal benefits. Second, in regions with inherently high levels of economic resilience, the development of data factor primarily manifests as a synergistic effect that “adds icing on the cake.” These regions typically possess a more robust innovation ecosystem and complementary conditions, such as an abundant human capital base, active innovation entities, and high levels of capital accumulation, enabling the data factor to integrate more rapidly and fully into industrial chains and governance systems, and efficiently translate into improvements in economic resilience.

To delve deeper into the pathways through which the development of data factor influences economic resilience, this paper conducts a mechanism test from three dimensions: innovation-driven development, digital finance, and industrial structure optimization, based on the mediating effect models presented in Equation (2) and Equation (3). The results are shown in Table 8. Specifically, Columns (1) and (2) examine the mediating role of the innovation-driven effect (IDE). The development of data factor significantly promotes innovation-driven development, which in turn significantly enhances economic resilience. This indicates that data factor can strengthen the adaptability and recovery capacity of the economic system by stimulating scientific and technological innovation activities. Research Hypothesis H2 is thus validated. Columns (3) and (4) investigate the transmission mechanism of the digital finance level (DFL). The development of data factor significantly elevates the digital finance level, and digital finance, as a modern financial infrastructure, further exerts a significant positive impact on economic resilience. This demonstrates that digital finance is an important channel through which data factor empower economic resilience. Research Hypothesis H3 is validated. Columns (5) and (6) analyze the mediating role of industrial structure optimization (STR). The development of data factor significantly drives the evolution of the industrial structure towards a more advanced level, and industrial structure optimization also significantly enhances economic resilience. This reflects that the development of data factor indirectly elevates the level of economic resilience by optimizing resource allocation among industries and promoting industrial upgrading. Research Hypothesis H4 is validated.

Table 8.Results of the mechanism test.

In summary, the results of the mechanism test indicate that the development of data factor not only directly promote the enhancement of economic resilience but also indirectly contribute through three pathways: stimulating innovation vitality, improving digital financial services, and promoting industrial structure optimization, forming a multi-dimensional and interconnected transmission mechanism.