Public Data Elements and Enterprise Digital Transformation: A Quasi-Natural Experiment Based on Open Government Data Platforms for Sustainable Urban Planning

Abstract

In the era of digital economy, the open sharing of public data elements has become a key engine driving the digital transformation of enterprises. This article is based on a quasi-natural experiment using government data platforms launched from 2011 to 2022, and it uses the asymptotic difference-in-differences method to empirically test the mechanism and impact of public data elements on enterprise digital transformation. It has been discovered that making data items publicly available dramatically increases the degree of enterprise digital transformation. According to mechanism analysis, corporate digital transformation is primarily impacted by releasing public data elements through two channels: information resource sharing and removing financial restraints. According to heterogeneity analysis, public data elements significantly impact enterprise digital transformation when they are located in regions with high data openness quality, when they are located in regions with substantial levels of digital economic vitality, and when they are available during businesses’ growth and maturity stages. The enabling influence of public data elements on organizational digital transformation was positively mitigated via a fully functional digital infrastructure and cutting-edge digital products. This study has confirmed the driving value of the market-oriented allocation of data elements in the digital transformation of enterprises, providing a theoretical basis for improving the government’s data openness system and unleashing the productivity of data elements. It also provides practical inspiration for enterprises to seize digital opportunities.

1. Introduction

Driven by the global digital wave and the modernization of state governance, China’s new urbanization process and the construction of “digital China” form a historic intersection. According to the data from the National Bureau of Statistics in 2023, the urbanization rate in China had reached 66.16% (https://data.stats.gov.cn/files/html/vchart/vchart_001/vchart_001.html accessed on 6 March 2025), but the “urban disease” caused by the mismatch of land resources still plagues new first-tier cities. At the same time, the People’s Republic of China’s Outline of the 14th Five-Year Plan and the 2035 Vision of the National Economic and Social Development suggests “expanding the safe and orderly opening of basic public information data” to facilitate the orderly circulation of public information data. This will speed up the sharing and exchange of public data and encourage the thorough docking of public and enterprise data. In December 2022, the Central Committee of the Communist Party of China and the State Council noted in their Opinions on Building a Data Base System to Maximize a Better Role of Data Elements that “for public data generated by Party and government organs at all levels, enterprises and institutions in the course of performing their duties by the law or in the course of providing public services, we shall strengthen the convergence, sharing and open development, reinforce the integrated authorized use and management, promote interconnection and interoperability, and break down the data silos.” In December 2023, the National Data Bureau released the Data Elements X Three-Year Action Plan (2024–2026), which also noted the importance of “expanding the supply of public data resources” and optimizing the distribution of data resources. In this context, public data elements, as the key media connecting physical space and digital twins, are reconstructing a new urban governance paradigm of “people–land–production” coordinated development, which provides a new way to solve the structural contradiction between the tight constraints of land resources and high-quality development requirements in the process of urbanization. Therefore, the realization of a high-quality supply of public data elements is of great significance for better releasing data value, reshaping urban spatial patterns, and promoting sustainable land use.

The rational allocation of urban land resources depends on the construction of smart cities [1,2]. The Guiding Opinions on Deepening the Development of Smart Cities and Promoting the Digital Transformation of Cities, issued by the National Development and Reform Commission, aims to promote the digital transformation and intelligent development of cities, emphasizing data integration, development, and utilization as the core and progressing through the digital transformation and construction of cities. As the micro-subject of urban economic activities, enterprises should digitize the operational data accumulated by enterprises, such as those regarding industrial agglomeration, land use efficiency, carbon emissions, etc., to provide a scientific basis for decision making for the government, to help optimize the development of existing land, and to reduce inefficient land use and resource waste [3]. In October 2024, the Central Network Information Office and other departments compiled the Implementation Guide for the Collaborative Transformation and Development of Digital Greening, aiming to better guiding enterprises to carry out the collaborative transformation and development of digital greening. It can be said that strengthening the digital transformation of enterprises is not only related to the remodeling of their own operating models, but it also has important strategic significance for China’s urban planning and land resources management and the realization of green and sustainable development [4].

A public data element is a data element that is provided by the government and utilized by enterprises in their production. It has two characteristics. First, from the perspective of public resources, public data elements display public attributes, which can significantly reduce the threshold of enterprise access to resources. Second, from the perspective of the value of data elements, public data exhibits the high value of data elements, which can effectively empower enterprise production. It is worth noting that at this stage, the data management ability of Chinese enterprises is relatively low compared to the management efficiency of the production and operation organization process, and the emergence of the government’s open data platform makes it possible for enterprises to obtain high-quality data resources. Data resources are one of the important assets driving the digital transformation of enterprises, and the two characteristics of public data elements aid in the integration of data elements into the production and operation of enterprises. The question remains regarding whether public data elements can promote enterprise digital transformation. What is the mechanism behind this process? Until now, there has been little corresponding research published in the literature. The study of the elements influencing the digital transformation of businesses is one of two major areas in the literature strongly tied to the findings presented in this paper. In the era of digital economy, the digital transformation of enterprises is not only related to their own survival and development, but also to the rational use of urban land. “Digital transformation” describes how businesses integrate computing, communication, connection, and information technology to alter organizational characteristics and eventually achieve organizational transformation. Three broad categories can be used to classify the determinants of enterprise digital transformation. The first is the institutional environment factors, which can speed up the digital transformation of businesses. These include mixed ownership reform [5], intellectual property administrative protection [6], business environment [7], regional digital infrastructure policy [8], state-owned participation [9], low carbon development policy [10], and more. The second aspect is the industrial environment; according to Verhoef et al. (2021) [11], the shift in industrial form will encourage businesses to go digital. According to Liu et al. (2024) [12], consolidating the digital services sector successfully lowers operating expenses for businesses and boosts their investment in innovation, eventually encouraging digital transformation. As a final consideration of the social environment, non-family shareholders facilitate the digital transformation of businesses, primarily through their human and capital effects in family businesses [13]. Furthermore, according to some academics, labor protection can help businesses acquire top-notch human capital, which will support business digital transformation [14].

The second area includes studies on how public data elements affect the economy. Scholars have primarily examined the macro-level impact of public data on economic development, pointing out that it can significantly boost economic growth [15,16]; at the meso-level, it can improve government governance efficiency and increase the vitality of innovation and entrepreneurship, which is conducive to improving the resilience of the urban industrial chain and enhancing the ability to withstand risk [17]; at the micro-level, some researchers empirically tested the impact of public data elements on corporate stock dividends [18], total factor productivity [19], and the contribution of new quality productivity [20]. In fact, public data elements have initially empowered the actual economy. Retail entrepreneurs use the New York City Business Atlas database to advise their funding and site selection [16], but currently, there is no standardized measurement method for choosing indicators for different types of public data. Some researchers employ the progressive twofold difference approach to empirically examine the economic effects of public data elements, using the “government data platform online” as a quasi-natural experiment [17,19]. In order to measure public data elements, some researchers have also employed comprehensive indicators. For example, they have developed an assessment framework for public data openness from the viewpoint of regular citizens based on the organizational requirements and characteristics of government datasets [21] and have evaluated the World Justice Project (WJP) Index [22]. In conclusion, few papers have thoroughly examined the mechanism and influence of public data elements on the digital transformation of organizations, even though previous research began to focus on the micro impact of these elements.

To empirically examine the impact of public data elements on enterprise digital transformation, this study exploits a quasi-natural experiment based on the staggered launch of municipal-level government data platforms across China from 2011 to 2022. A multi-period difference-in-differences (DID) approach is employed to estimate the causal effect of public data availability on firm-level digital transformation outcomes. This empirical strategy enables us to control for time-invariant unobserved heterogeneity and common shocks across regions and years.

Our findings reveal that public data elements can inspire companies to undergo digital transformation. Its transmission mechanism lies in the fact that public data elements enhance the level of digital transformation of enterprises by sharing information resources and alleviating financing constraints. Heterogeneity analysis shows that in regards to high-quality data openness, areas with high-energy digital economy experiencing growth and maturity of enterprises, the role of public data elements in promoting enterprises’ digital transformation is more significant. Analysis of the regulatory role of public data found that complete digital infrastructure and advanced digital products positively regulate the role of public data elements in promoting the digital transformation of enterprises. In addition, corporate investment plays a negative regulatory role, while innovative investment plays a positive regulatory role.

This article systematically analyzes the role and mechanism of public data elements in promoting the digital transformation of enterprises using the incremental difference method. Firstly, in terms of research perspective, previous studies have not yet given enough attention to the impact of public data elements on enterprise digital transformation [18,19,20]. Secondly, in terms of the impact mechanism, this article proposes and verifies that public data elements mainly influence digital transformation through information resource sharing effects and financing constraint mitigation effects, which helps clarify the mechanism of the role of public data elements in digital transformation. Thirdly, the role of the carrier was examined. Previous studies have typically focused on the public data elements themselves, neglecting the prerequisite conditions for the public data elements to play a role [17,19]. This paper further explores the requirements for the public data elements to play a role through the moderating effect and notes that public data elements can contribute to the economic growth of enterprises. Through the moderating impact, this study further examines the prerequisites for the function of public data components. It highlights that the role of public data elements necessitates a comprehensive digital infrastructure and cutting-edge digital products to accompany it.

2. Background Information from Institution and Theoretical Theories

2.1. The Setting of the Institution

The goal of creating a single, open platform for national government data by the end of 2018 was stated in the State Council’s Outline of Action for Promoting the Development of Big Data, which was released in September 2015. In the context of big data governance, the building of digital government has become the foundation for the open data platform’s efficient operation. In June of 2022, the State Council released the Guiding Opinions on Strengthening the Construction of Digital Government, pledging to develop the systematic architecture of data openness, realize the download and service of open data, and transform the form of digital government. The Central Committee of the Communist Party of China and the State Council released their Opinions on Building a Data Base System to Maximize a Better Role of Data Elements in December of that year. This document adopts explicitly “adhering to the sharing of common use and releasing the value dividend” as the working principle. It investigates the creation of a mechanism for the equitable distribution of the advantages of opening up public data resources in order to assist businesses in managing the risks and difficulties associated with the digitalization transformation process. To activate the potential of data elements and support the sustainable development of the economy, the National Data Bureau and other departments released the Three-Year Action Plan for Data Elements X (2024–2026) in December 2023. This plan encourages the organization of public data authorization and operation mechanisms in key areas and relevant regions.

The creation of public data elements is a top priority for national policy. Local governments have responded favorably by implementing region-specific plans and actions for developing public data, establishing high-quality, high-level, and high-standard public data open platforms and enabling the market and social actors to develop and use data, thus making available the benefits of government data. The “digital divide” is lessened when data publishers, such as government departments, can publish data more methodically and effectively through open government data platforms, and data acquirers, such as businesses, research institutions, and the general public, can access and use the data. This is because open government data platforms offer market participants services that are both widely available and free from prejudice. Second, the open government data platform’s high-quality data has reduced the threshold for use and enhanced data use convenience. As of August 2023, excluding Hong Kong, Macao, and Taiwan, 226 provincial and local governments had set up open platforms for local government data (the data comes from the Research Report on Open Data Utilization of China Government (2023), published by School of Information Management, Huazhong Normal University). As an illustration, consider Beijing’s public data open platform, which is currently accessed by 115 data-providing units and comprises 18,573 datasets, totaling over 7186 million pieces of data. Regarding data usage, the platform has opened 15,014 interfaces, with over 320 million platform accesses and over 370,000 downloads and usages, with the most extensive dataset possessing a capacity of over 840,000 pieces of data (https://data.beijing.gov.cn/index.html, accessed on 6 March 2025). The public data open platform’s extensive coverage and facilitation have reduced the “access gap” between businesses and public data, enabling the digital transformation of businesses.

2.2. Analysis of the Impact of Public Data Elements on Enterprise Digital Transformation

Information economics focuses on the economic phenomena of incomplete information and asymmetric information. The analysis of incomplete information can be divided into optimal information search theory and information dissemination theory [23]. The analysis of asymmetric information is usually based on the time dimension. Among the typical early manifestations of asymmetric information are adverse selection and signal transmission. The later manifestation of asymmetric information mainly involves moral hazard and principal-agent issues. As mentioned above, the two important sectors of information economics are incomplete information and asymmetric information, but the existence of these two sectors usually leads to increased expenditure in the process of completing the prefetch goal, which is information friction.

Public data elements have two characteristics: one is their public attributes, which can significantly reduce the threshold for enterprises to obtain resources. The other is that public data displays the high-value attributes of data elements, which can effectively empower enterprise production.

The impact of public data elements on enterprise digital transformation can be explained from two aspects, i.e., the incentive effect and reverse pressure effect. First, at the incentive level, the acquisition and utilization of public data by enterprises helps to enhance the endogenous driving force of their digital transformation. Initially, from the perspective of public attributes, digital transformation relies on a large number of high-quality data resources as its basis. The opening of public data provides enterprises with basic data resources with strong authority, wide coverage, and high versatility. It not only significantly reduces the cost of enterprises to obtain information, but also promotes the efficient flow of data elements, effectively supplementing the shortcomings inherent in the internal data of enterprises, thereby providing key support for enterprise digital transformation [24]. Based on the theory of information economics, the openness of public data elements effectively reduces the cost of enterprises to obtain external information and improves the openness and transparency of this information, thereby alleviating the problem of information asymmetry in the market, to a certain extent. The reduction in information friction not only optimizes the company’s perception of the external environment, but also enhances the company’s confidence and willingness to make digital transformation decisions under uncertain conditions [25]. Secondly, from the perspective of high-value attributes, public data elements can be deeply integrated with traditional production factors such as land, labor, and capital, reshaping the approach of enterprises to allocating production factors and the logic of value creation. By embedding the entire production process, public data not only improves resource allocation efficiency, but also promotes the intelligence, refinement, and coordination of production methods, thereby injecting new impetus into the digital transformation of enterprises [26,27]. Public data elements promote enterprises to change their production management models, make up for the management’s shortcomings regarding cognition and information processing, reduce information frictions faced by management, improve information transparency and decision-making efficiency between organizational levels, and thus reduce opportunity cost losses caused by planning errors [24]. Finally, by reducing information friction, public data elements significantly promote digital technology innovation in enterprises, thus promoting the digital transformation of enterprises. Digital technology innovation provides a technical foundation for and reduces the difficulty of enterprise digital transformation. On the one hand, the opening of public data has improved the accessibility of enterprises to acquire and process digital resources such as high-quality data, metadata, and technical documents, fundamentally reducing the difficulty of data access and integration and effectively alleviating information asymmetry and processing obstacles. This reduction in information friction is of great significance to the rapid development and efficient deployment of digital technology innovation [28]. On the other hand, public data elements also exhibit certain governance functions in the governance structure, which helps alleviate the delegation–agent problems within the enterprise [29]. By providing improved information transparency and enhanced supervisory capacity, public data helps mitigate managerial short-termism and self-interest, streamlines resource allocation processes, and reduces inefficient behaviors such as defensive management, thereby facilitating greater investment in digital innovation initiatives by firms.

Second, the widespread application of public data elements by competing companies has significantly aggravated the intensity of market competition, which in turn forces companies to accelerate the pace of digital transformation [30]. The sharing and inclusiveness of public data effectively breaks through the high entry threshold of traditional production factors, significantly lowers market entry barriers, and allows more emerging companies to quickly enter the market, thereby intensifying competition among market entities. On the other hand, public data elements have improved the transparency and service efficiency of government operations, optimized the business environment, and promoted the formation of a more fair and efficient market order. This optimization of the institutional environment has further attracted external enterprises to enter the market and increased the overall competitiveness [31]. Against this background, if traditional enterprises respond slowly in digital transformation, they will face the risk of being marginalized or even eliminated from the market, and their survival pressure will increase significantly [32]. Therefore, the intensification of market competition has become an important external driving force for motivating enterprises to accelerate digital transformation, making enterprises pay more attention to improving profitability and maintaining competitive advantages through digital means [33]. Accordingly, this paper proposes the following hypothesis:

H1. 

Public data elements can facilitate enterprise digital transformation.

2.3. Mechanism Analysis

In the age of the digital economy, digital transformation is an inevitable trend that must be embraced for enterprises to adapt to the sustainable development of the economy, and its secret is the complete use of data. However, information friction increases the difficulty of digital transformation. Relying on the basic theory of information economics, this paper divides the types of information friction that hinder the digital transformation of enterprises into two categories: one is friction with the public system, such as the increase in costs caused by redundant administrative approvals, and the other is friction with market entities, such as the deterioration of financing capacity due to uncertainty in the market environment. Public data elements comprise a type of universal and shared public resource that can lower market costs and lessen friction with the public system by increasing the degree of data sharing. On the other hand, public data elements are still fundamentally data elements, and the economic value they release can maximize business cash flow, ease financing constraints, and lessen friction with market entities. Thus, from the standpoint of information friction and in conjunction with the economic properties of public data, this research examines the theoretical mechanism of public data elements influencing the digital transformation of organizations.

2.3.1. Effects of Information Resource Sharing

The digital transformation of businesses can be facilitated by public data elements that optimize the external information environment of businesses, improve transparency, and lessen the congestion of production resources brought on by governmental information lag. According to previous research, enhancing the external information environment reduces information barriers to digital transformation by improving the availability, validity, and integrity of data resources [34]. As an essential public resource, public data elements are non-scarce data elements comprising public goods in the market [35], which can simultaneously maximize resource sharing within the government and information resource sharing between the government and businesses. This is why the information resource-sharing effect exists. The government will first be compelled to restructure the government informatization system due to the openness of public data elements. This will make internal information communication within government departments more efficient and information transmission more truthful, which can significantly improve the transparency of government operations [31,36]. Second, the limited and subpar internal datasets of businesses and their inability to handle data have emerged as significant barriers to digital transformation. The public data factors it offers can help to improve the level of data resources, continuously promote the flow and sharing of data factors, facilitate the overflow of high-quality data resources to enterprises, lower the threshold of data factor use, and empower the digital transformation of enterprises. The government data platform, established by government departments, is crucial in promoting and perfecting the data factor market service system.

2.3.2. Effects of Mitigating Financial Constraints

In order to improve the degree of enterprise digital transformation, public data elements can assist businesses in overcoming the crowding-out problem of financial resources through the relieving effect of financing constraints. According to existing research, enterprise digital transformation depends on resolving the enterprise funding constraint challenges. The primary reason for the existence of the financing constraint relief effect is that, as a type of data element, the acquisition of public data elements first and foremost lowers the cost of businesses looking for high-value data [24]. It also directly improves the business’s operational and management capabilities, which in turn increases the business’s negotiating power in financing activities. Additionally, public data elements offer an opportunity screening mechanism [37] that enables firms to prioritize “generally better candidates” [38] when making investment decisions. That is, enterprises rely on public data to identify the most profitable projects from a large amount of information. Promising initiatives are typically more profitable, and when businesses can give investors precise and reliable information, they are more likely to back these promising projects, which increases the possibility of receiving corporate financing. Last but not least, financial institutions can better understand an enterprise’s solvency and risk levels by analyzing public data, such as industrial and commercial administrative penalties, enterprise registration information, and enterprise credit status. Financial institutions can create a thorough and accurate enterprise portrait, lower the assessment cost and financing risk rating of an enterprise, and then tend to provide enterprises with more significant amounts of financing or more favorable financing conditions, thereby relieving the financing constraints of enterprises. Businesses may now access more liquid capital and use it more loosely, thanks to removing financing restrictions, which is improves digital transformation. As a result, one crucial mechanism of action for public data elements to support businesses’ digital transformation is the relief of financial constraints.

Based on the above analysis, this paper proposes the following hypotheses:

H2. 

Regarding friction with the public system, public data elements can facilitate enterprises’ digital transformation through the information resource sharing effect.

H3. 

Regarding friction with market players, public data elements can facilitate firms’ digital transformation through financing constraint mitigation effects.

2.4. Public Data Elements and Enterprise Digital Transformation: Heterogeneity Analysis

The impact of public data elements on enterprise digital transformation has varying effects based on the level of high and low open data quality, high and low digital economy market dynamics, and the enterprise growth cycle because there are differences in the efficiency of businesses regarding their use of public data elements, i.e., different open data quality, different growth cycles, and different digital economy dynamics.

Firstly, the extant research generally agrees that the quality of data openness has a robust correlation with its economic value regarding the impact of public data elements on the digital transformation of enterprises with varying data openness quality [39]. Higher data openness means that the government provides businesses with better public data elements. High-quality public data elements have clear advantages over low-quality data in terms of data accuracy, timeliness, and completeness, i.e., accurate data can give businesses a solid information base and help them make the right decisions and take appropriate actions; complete data can give them comprehensive information and help them better understand business operations and customer needs; timely data can increase the operational efficiency and response speed of businesses; and timely data can reduce incorrect decisions and losses caused by information lag. Therefore, businesses can better construct their digital transformation when they employ high-quality public data elements.

Second, the degree of advanced digital technology development and corresponding technology use capabilities are significant factors for the emergence of heterogeneity regarding the influence of public data elements on the digital transformation of enterprises in regions with varying levels of digital economy vitality. On the one hand, areas with a high digital economic vitality display more developed digital technology innovation s. Digital technology innovation will first emerge in regions with high digital economic vitality characterized by significant economic strength and risk-taking capacity. This is because implementing digital technology innovation is extremely risky and requires entities to make significant expenditures and bear unknown risks. In addition, regions with high digital economic vitality exhibit greater integration between the digital and physical economies, which can consistently lead to the emergence of new business models and forms. This makes these regions ideal for the long-term growth of digital innovation technology. Local business digital transformation will be aided by the rise and advancement of digital technological breakthroughs. However, areas with a high level of digital economic vitality are better able to use technology. Innovative technology’s contribution to digital transformation is contingent upon the infrastructure that supports it. Highly active regions have more substantial technical usage capabilities and a more comprehensive infrastructure related to digital aspects, allowing them to discover more significant value in public data, supporting the digital transformation of businesses.

Furthermore, the layered integration of data elements is more economically useful due to the incremental returns to scale of the data elements. Enterprise digital transformation can be facilitated by the rich accumulation of data elements in areas with strong digital economic vitality, providing more technological use possibilities [40]. Therefore, public data elements display a more evident function in boosting the digital transformation of businesses in high digital economic vitality regions than in low digital economic vitality regions.

Last but not least, the absorption and transformation of digital resources are key to the digital transformation of organizations in terms of the impact of public data elements on the digital transformation of enterprises in various growth cycles [41]. Public data elements, an essential digital resource outside the company, can be absorbed and transformed in multiple ways by businesses in varying stages of development. While businesses in the decline period typically rely on internal data for decision making and are insensitive to external data and information, businesses in the growth and maturity periods must depend more on public data for pertinent decision making and analysis, which lowers exploration and transaction costs. However, it is possible to effectively alter and use public data elements, which requires businesses to exhibit strong financial standing and cutting-edge innovation skills. In the age of the digital economy, businesses in the growth phase are typically founded with venture capital and other forms of financial support, and they are naturally creative and adaptable, which allows them to make effective use of public data elements; businesses in the mature phase have a certain amount of financial strength and concentrate on ongoing innovation to keep their competitive advantages; they can set up a system that effectively transforms the value of public data elements for their own digital transformation. Businesses in the decline phase may encounter operational challenges and financial limitations, which make it challenging to absorb and transform the economic value of public data elements. Thus, the following theories are proposed in this paper:

H4. 

The impact of public data elements on enterprise digital transformation is heterogeneous regarding data quality, digital economic dynamism, and life cycle heterogeneity. The impact of public data elements on enterprise digital transformation is more significant in areas of high digital economic dynamism, more prominent in cases of high-quality data, and more important in terms of the growth and maturity of enterprises.

2.5. Public Data Elements and Enterprise Digital Transformation: Regulatory Role

Data factors exhibit a certain level of dependence and require using the Internet and other carriers to generate greater value, in contrast to traditional production elements like labor, capital, and land, which can independently engage in the production process [42]. The development of a digital infrastructure and the use of digital products can enhance the role of public data elements in advancing enterprise digital transformation. This is because in order for public data elements to fully release their economic value, they must be based on a comprehensive digital infrastructure and cutting-edge digital products. First, from the perspective of how digital infrastructure is regulated, improving the effectiveness of public data element use is a key area of focus for increasing their influence on businesses’ digital transformation. Digital infrastructure, which serves as the foundation for data element operation, can significantly increase the effectiveness of businesses’ use of public data elements. The Internet and 5G base stations are new-generation information and communication infrastructures that can lower data transmission costs and increase data transmission efficiency. Second, computing infrastructure, such as supercomputing and bright computing centers, meet the computing power requirements of enterprise digital transformation at a lower cost and assist in data analysis. The economic value of public data elements can only be effectively released and provide a greater impetus for digital transformation via sufficiently advanced digital products. From the perspective of the regulating effect of digital products, public data elements are essentially public resources. This means that the more sophisticated the digital product, the more it can utilize the information resource sharing and financing constraint alleviation effects of public data elements. Thus, the role of public data elements in promoting the digital transformation of enterprises depends on the advanced level of the digital products.

In exploring the regulatory role of public data elements in enterprise digital transformation, the enterprise resource allocation method is a key regulatory variable. Traditional enterprise investment (such as fixed asset investment) and innovative investment (such as R&D investment) represent the different development strategies and resource tendencies of enterprises, and the two play completely different regulatory roles in the use of public data elements. On the one hand, corporate investment has a strong path dependence and is often concentrated in “heavy asset” areas such as expanding production capacity and building infrastructure. This type of investment displays strong sunk costs and low flexibility, which may lead to slow response in the face of a rapidly changing digital environment, inhibiting the effective integration of emerging technologies and data resources, thereby weakening the positive role of public data elements in promoting the digital transformation of enterprises. More importantly, excessive reliance on traditional capital investment may cause enterprises to strategically stick to their original development paths and ignore the transformation logic based on data-driven alternatives, thus showing a negative regulation effect. On the other hand, innovation investment focuses on technological research and development, data capacity building, and organizational innovation, reflecting the forward-looking layout of enterprises for future uncertainty and emerging capabilities. This type of investment helps enterprises establish technical absorption and organizational adaptability that matches public data elements, allowing enterprises to more effectively explore, analyze, and apply open data resources and to realize the digital and intelligent reconstruction of business processes. Therefore, innovative investment can enhance the driving force of public data elements toward the digital transformation of enterprises and show a significant positive regulatory effect. Therefore, the different types of resource allocation strategies of enterprises display differentiated regulatory effects regarding the use of public data elements.

H5. 

Digital infrastructure and digital goods positively moderate public data elements related to business digital transformation.

H6. 

In the regulatory role of public data elements influencing the digital transformation of enterprises, corporate investment plays a negative role, while innovative investment plays a positive role.

3. Research Design

3.1. Modeling

To test the impact of public data elements on the digital transformation of enterprises, the benchmark regression model is set, as shown in Equation (1):

$${DigiT}_{it}=\alpha +\beta {PData}_{it}+\phi X_{it}+{\eta }_i+{\theta }_t+{\epsilon }_{it}$$

(1)

In Equation (1), DigiT_it is the digital transformation index of firm I at year t, PData_it is essentially a two-dimensional dummy variable constructed as an interaction term between the time dummy and individual entity dummy variables. It indicates whether the government data platform is online in the city that firm i belongs to at year t, X_it is the set of control variables, ${\eta }_i$ is the firm fixed effect, ${\theta }_t$ is the year fixed effect, and ${\epsilon }_{it}$ is the random error term.

Figure 1 shows the gradual process of government data open platform policy implementation. As shown in the figure, the government data open platform policy started in 2012 and was implemented in 10 phases by 2022. During this period, one or more new regions were added every year to implement the policy. By 2022, the regression model included a total of 10 treatment groups and one control group.

3.2. Measurement of Key Variables

3.2.1. Digital Transformation (DigiT)

The measurement of enterprise digital transformation is still in the process of continuous improvement, and various methods are proposed in the existing literature. Early research usually uses the proportion of intangible assets to total assets as agent variables, assuming that the digital capabilities of a company are reflected in its intangible capital. In recent years, scholars have generally adopted a text analysis method to measure the degree of digitalization of a company by counting the frequency of digital-related keywords in the management discussion and analysis (MD&A) part of the annual reports of listed companies. Unlike the aforementioned measurement method, this article uses the China Listed Companies Digital Transformation Index, developed by CSMAR and a research team from colleges and universities, to measure the degree of digital transformation of enterprises through the comprehensive index method. The index starts with six first-level dimensions, i.e., strategic leadership, technology-driven, organizational empowerment, environmental support, digital achievements, and digital application, and combines multiple secondary indicators for weighted calculations to fully reflect the multi-dimensional performance of enterprises in the process of digital transformation. For details of the specific indicator composition and weight allocation, please refer to

Table 1. Detailed indicators and weights of the digital transformation index.

First-Level Indicator	Weight of First-Level Indicator	Second-Level Indicator	Weight of Second-Level Indicator
Strategic Leadership	34.72%	Establishment of Digital Executive Roles	23.82%
		Forward-Looking Nature of Innovation Orientation	27.88%
		Persistence of Innovation Orientation	18.79%
		Breadth of Innovation Orientation	12.83%
		Intensity of Innovation Orientation	16.68%
Technological Advancement	16.20%	Artificial Intelligence Technology	55.04%
		Blockchain Technology	12.98%
		Cloud Computing Technology	18.32%
		Big Data Technology	13.66%
Organizational Empowerment	9.69%	Digital Capital Investment Plan	50.22%
		Digital Human Capital Investment Plan	25.53%
		Digital Infrastructure Construction	12.06%
		Construction of Science and Technology Innovation Bases	12.19%
Environmental Support	3.42%	Number of Invention Patents in the Industry	19.23%
		R&D Activities in the Industry	17.79%
		New Product Development and Sales in the Industry	14.98%
		Intensity of Digital Technologies in the Industry	11.57%
		Digital Capital Investment Intensity in the Industry	11.40%
		Human Capital Investment Intensity in the Industry	7.89%
		Optical Cable Density in the City	4.77%
		Mobile Switch Capacity in the City	4.03%
		Broadband Internet Access User Scale in the City	4.00%
		Mobile Internet User Scale in the City	4.34%
Digital Outcomes	27.13%	Digital Innovation Standards	36.68%
		Digital Innovation Papers	11.74%
		Digital Invention Patents	23.54%
		Digital Innovation Qualifications	14.73%
		National-Level Digital Innovation Awards	13.31%
Digital Application	8.84%	Technological Innovation	63.42%
		Process Innovation	23.78%
		Business Innovation	12.80%

.

On the one hand, the comprehensive index method can more fully reflect the multi-dimensional characteristics of digital transformation than can the intangible asset proportion method. As a static financial indicator, it does not adequately identity which parts of the intangible assets of an enterprise are directly related to digital transformation and as such, are susceptible to industry heterogeneity and accounting valuation methods. By integrating structured data and text information, the comprehensive index method not only captures the real investment of enterprises in dimensions such as technology, talents, and capital, but also dynamically tracks the evolution of enterprises’ digital capabilities, making this method more compatible with the essential characteristics and policy orientation of digital transformation. On the other hand, compared with the text analysis method, the comprehensive index method can effectively overcome the problem of information disclosure errors inherent in this method. Text analysis methods usually rely only on the frequency of keywords in the annual report of the company. This method is susceptible to management expression style, language ambiguity, and information disclosure strategies, and may not accurately reflect the actual digital investment and progress of the company. Therefore, the comprehensive index method introduces multi-source heterogeneous data, such as technical patents, digital capital investment, the industry innovation environment, and other objective information, and builds an evaluation index system from multiple dimensions, enhancing the objectivity and representativeness of the measurement. In addition, the comprehensive index method displays stronger scalability and timing consistency, which is suitable for horizontal and vertical comparisons of enterprises in different scales and industries and can more accurately characterize the entire process of digital transformation. Therefore, this paper selects the comprehensive index of digital transformation in the CSMAR database as the proxy variable for enterprise digital transformation and uses traditional digital word frequency measurement methods for robustness testing.

As shown in Figure 2, Since the implementation of the policy in 2012, the digital transformation index of the treatment group has increased significantly, and the gap between the treatment group and the control group has gradually widened. This trend shows that the government’s open data policy plays a significant role in promoting the digital transformation of enterprises. In the initial stage of policy implementation (2012–2014), the digital transformation level of the processing group was improved, but the growth rate was relatively moderate. With the deepening of the policy, the digital transformation index of the treatment group showed a rapid growth trend from 2015 to 2020, far exceeding that of the control group. This shows that the positive effects of policies may accumulate in the short term and play a more significant role in the medium and long term. In addition, although the index of the treatment group declined in 2021–2022, it was still higher than that of the control group as a whole, showing the sustained impact of the policy. On the whole, the time series analysis supports the positive role of the government data open platform in promoting the digital transformation of enterprises and confirms the effectiveness of the policy.

3.2.2. Public Data Elements (PData)

The Government Open Data Platform, an information resource-sharing platform run by the government with involvement from various government departments, is the primary source of public data elements. It offers structured data sets that can be downloaded or accessed through interfaces, supporting the development of public data by market participants and achieving government governance goals. Government data openness, on the other hand, is an institutional tool for the government to improve administrative efficiency in the digital age, with the goal of administrative rationality and a greater focus on development goals in the economic and social fields. Government information disclosure aims to achieve administrative legitimacy by achieving transparency and accountability.

Finding the government public data platform is the first stage in the three-phase data-gathering process used in this paper. The domain name “data.gov.cn” is used in this document to guarantee that the government is the primary entity of the retrieved data platform. This article does not include the county and district levels; only government public data platforms with an administrative level corresponding to prefecture-level cities and higher are selected. The second stage is the query for the public data platform’s online time. To find the platform, Baidu and other search engines combine manual and automated crawling for the terms “city name+data service”, “city name+data disclosure”, “city name+government data”, etc. To verify the platform’s initial launch time, keywords such as “city name data service”, “city name data disclosure”, “city name government data”, etc. were searched. Verification and validation are included in the third step. The China Local Government Data Openness Report (Cities) (released by the Digital and Mobile Governance Laboratory of Fudan University) and the Research Report on Open Data Utilization of the Chinese Government (released by the School of Information Management, Huazhong Normal University) are compared and verified with the data collection results in this paper. Finally, the pertinent data from 187 public data government platforms as of the end of 2022 are obtained.

The public data element (PData), which is defined as a dummy variable in this paper’s analysis using the progressive double-difference method, takes the value of 1 if the prefectural-level city where the enterprise is registered already has a government data platform and 0 otherwise.

3.2.3. Control Variables

To control for other factors affecting firms’ digital transformation, this paper refers to the work of Wen and Deng (2023) [6] and incorporates the following firm characteristic control variables: (1) cash management (CASH), which is measured by using the ratio of the sum of money funds and trading financial assets to total assets; (2) total liabilities rate (TL), which is measured by using the ratio of total liabilities to total assets; (3) return on total assets (ROA), measured using the ratio obtained by dividing the firm’s net profit by the total assets; (4) dummy variable, i.e., whether or not the firm is a state-owned enterprise (SOE); if the firm is a state-owned enterprise, SOE takes the value of 1; otherwise, it is 0; (5) age of the firm’s listing (LAGE), measured using the difference between the logarithmized current year and the year of the firm’s listing; (6) size of the firm (SIZE), measured using logarithmized total assets; (7) segregation of equity (SEP), measured using the difference between control and ownership; (8) concentration of equity (CONC), measured using the proportion of shares held by the first largest shareholder; and (9) book-to-market ratio (MBRATIO), measured using the ratio of total assets to market capitalization.

3.3. Data Description

In this paper, China A-share listed companies are taken as basic samples, according to the start-up time of the government data open platform policy pilot, and the time dimension of enterprise digital transformation data is comprehensively considered, so as to construct panel data with a time span of 2011–2022, with data coming from the CSMAR database. To prevent misleading results in the subsequent analysis, the raw data is cleaned as follows: (1) listed companies with listing statuses of “suspended,” “ST,” or “*ST” are excluded; (2) the data of the financial industry are excluded due to their unique asset–liability structure; (3) the samples whose registered enterprise locations are in non-city administrative units, such as a region, an autonomous state, a league, etc., are excluded; (4) the tails of all the continuous variables are shrunk to eliminate the influence of the extreme values; (5) the missing values of the main variables are filled in using the linear interpolation method. (6) the descriptive statistics of the main variables in this paper are displayed in

Table 2. Descriptive statistical characteristics.

Variable	Meaning	Obs.	Mean	Min	Max	Std. Dev.
DigiT	Digital Transformation Index	33,396	23.1454	23.1454	64.2934	10.1130
PData	Public data elements	33,396	0	0	1	0.4971
CASH	Cash management	33,396	0.0144	0.0144	0.7309	0.1559
TL	Total debt ratio	33,396	0.0075	0.0075	178.3455	1.0964
ROA	Return on total assets	33,396	−0.3281	−0.3281	0.2080	0.0699
SOE	State-owned or not	33,396	0	0	1	0.4772
LAGE	Age of Listing	33,396	0.0000	0.0000	3.3673	0.9686
SIZE	Enterprise size	33,396	19.6123	19.6123	26.4725	1.3591
SEP	Degree of shareholding separation	33,396	0.0000	0.0000	28.3182	7.3908
CONC	Shareholding concentration	33,396	8.4800	8.4800	74.9500	14.8037
MBRATIO	Book-to-market ratio	33,396	0.0835	0.0835	1.2124	0.2745

.

4. Baseline Regression Analysis

4.1. Baseline Regression

Regressing the benchmark model constructed according to Equation (1) and with standard errors clustered to the firm level, the results of the benchmark regression are shown in columns (1)–(4) of

Table 3. Benchmark regression results.

Variable	(1)	(2)	(3)	(4)
	DigiT
PData	0.6543 *** (0.1268)	0.5332 *** (0.1225)	0.2742 ** (0.1187)	0.4763 *** (0.1203)
GDP				0.0003 (0.0026)
HD				0.0009 ** (0.0005)
Control variable		Yes	Yes	Yes
Constant	30.4265 *** (0.1315)	−3.9673 (2.6884)	−2.1565 (1.6692)	−5.2219 * (2.8213)
Firm fixed effect	Yes	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes	Yes
Industry-by-year FE			Yes
Observed value	33,396	33,396	27,830	31,320
Adjusted R2	0.3527	0.3809	0.5384	0.3828

Note: *, **, and *** indicate significance levels of 10%, 5%, and 1%, respectively. Standard errors clustered to the firm level are shown in parentheses. Control variable includes: CASH, TL, ROA, SOE, LAGE, SIZE, SEP, CONC, and MBRATIO.

. Specifically, the estimated coefficient of the explanatory variable public data elements (PData) in Column (1), without adding any control variables, is 0.6543, which is significantly positive at the 1% level. Column (2) represents the regression results with the addition of control variables, and the regression coefficient decreases to 0.5332, which is still significant at the 1% level. To address potential confounding effects arising from industry-specific cyclical fluctuations and sectoral policy shocks, while simultaneously controlling for heterogeneous responses across industries to macroeconomic fluctuations, we incorporate a comprehensive set of industry–year interaction fixed effects in Column (3) of our econometric specifications. Column (4) indicates that the control variables at the city level are included in the benchmark regression, and in this paper, we refer to the work of Wen and Deng (2023) [6], selecting the GDP growth rate and urban population ratio (HD). The regression coefficients of the public data elements in Columns (3) and (4) are significantly positive, at at least 5%. The above results indicate that public data element openness significantly enhances enterprise digital transformation, and Hypothesis 1 is confirmed.

The economic intuition behind this result is that public data elements exhibit both public and high-value characteristics, which helps reduce institutional friction and information asymmetry between enterprises and public systems and market entities during circulation and use, thus promoting the digital transformation of enterprises. Chen and Zhang (2024) [30] also analyzed the impact of public data on the digital transformation of enterprises. The study proved that public data also promotes the digital transformation of enterprises, which is consistent with the conclusions of this article. The difference is that its promotion effect is mainly achieved by reducing operating costs, strengthening management, and intensifying competition in local markets.

4.2. Parallel Trend Test

Before the government data platform goes online, the double-difference model requires no significant difference between the experimental group and the control group in terms of the overall development of the enterprise. Otherwise, the test results will not be credible due to the omission of the effects of other factors. Drawing on the work of Beck (2010) [43], this paper sets up the following parallel trend test model:

$${DigiT}_{it}=\alpha +{\displaystyle {\sum }_{k\ge -5}^{k\le 3}}{\beta }_k\times {Treat}_{ik}+\phi X_{it}+{\eta }_i+{\theta }_t+{\epsilon }_{it}$$

(2)

Treat_ik in Equation (2) denotes the kth year when the government data platform goes online. The rest of the variables are set consistent with those in Equation (1). The difference in the level of enterprise digital transformation between the experimental group and the control group is manifested in the significance of ${\beta }_k.$ As shown in Figure 3, $\beta$ is not significant in the 5 years before the data platform goes online, and the estimation starts to be significantly positive from the second year onwards after the policy occurs. This indicates, on the one hand, that there is no systematic difference between the experimental and control groups in terms of time change before the data platform goes online, and the parallel trend test of the asymptotic double-difference model is validated; on the other hand, the facilitating effect of the public data elements on the digital transformation of the enterprises has a time lag, and the impact increases from the second year onwards. It is maintained until the end of the sample period.

4.3. Robustness Tests

4.3.1. Placebo Test

The model identification passes the placebo test, and the benchmark regression results are robust and reliable. To ensure that changes in enterprise digital transformation are accurately attributed to public data elements, a placebo test is necessary to rule out confounding errors from other unknown factors. The entire sample is first randomly sampled to form virtual experimental and control groups. Then, the samples of these two groups are regressed according to the model constructed in Equation (1). This process was repeated 500 times to obtain the kernel density distribution, as shown in Figure 4. The results show that none of the estimated coefficient t-values obtained from the sampling exceed 2 in absolute value, and the p-value exceeds 0.1.

4.3.2. Endogenous Test

Although the results of the benchmark regression confirm that there is a significant positive relationship between public data elements and enterprise digital transformation, the possible omission of the control variables of enterprise digital transformation leads to the continuation of the endogeneity problem, and this paper adopts the instrumental variable approach to deal with this endogeneity. Based on the work of Nunn and Qian (2014) [44] and Ling et al. (2024) [45], using the cross-multiplier term to construct instrumental variables, this paper uses the cross-multiplier term between the characteristics of the city party secretary’s place of origin in each city in 1984–2009 (before the city’s public data open platform was online), i.e., the inverse of the average distance from the coastline of the city of origin and the number of Internet access ports in the country. The cross-multiplier term of the number of access ports (Sec_Net) is used as an instrumental variable for public data elements. On the one hand, the distance of a city from the coastline is generally closely related to the degree of openness and economic development of the city, which means that the acceptance of open policies, such as the launch of data platforms by municipal party secretaries, can be reflected by the characteristic of the distance from the coastline of their place of domicile. The characteristic of the place of domicile of municipal party secretaries prior to 2009 affects the acceptance of the launch of data platforms during the sample period in terms of the continuation of the policy, etc. Therefore, the relationship between the launch of government’s data platform and the municipal party secretary’s place of origin characteristics are strongly correlated and satisfy the correlation hypothesis. On the other hand, from a theoretical perspective, the birthplace characteristics of municipal party secretaries prior to 2009 represent immutable personal historical traits that were formed before they assumed office and began policy-making. Therefore, the birthplace characteristics of municipal party secretaries does not directly affect firms’ digital transformation during the sample period, thus satisfying the exclusion restriction. The origin itself will not directly affect the digital strategy or technology investment behavior of the enterprises in the city they work for. Its possible impact on the digital transformation of enterprises can only be achieved through the indirect role it exerts on government policy formulation (such as whether or not to promote the construction of a government data platform) after the secretary takes office. Therefore, the homeland variable is used as the basic dimension of the instrumental variable, which is in line with the logical starting point of the assumption of exclusivity. In assessing the effect of the regression of the instrumental variable (Sec_Net) on the explanatory variable (PData) in Column (1) of

Table 4. Endogenous test results.

Variable	(1)	(2)	(3)	(4)
	IV = Sec_Net		IV = L2.PData
	PData	DigiT	PData	DigiT
IV	0.0001 *** (4.95 × 10−6)		0.06878 *** (0.0047)
PData_IV		2.2022 *** (0.8104)		4.2670 *** (0.2255)
Control variable	yes	yes	yes	yes
Firm/time fixed effects	yes	yes	yes	yes
WIVT	491.954 ***		387.058 ***
Identification of insufficient tests	841.741 ***		986.637 ***
Observed value	31,436	31,436	23,779	23,779
Adjusted R2	0.3399	0.1308	0.5715	0.1323

Note: *** indicate significance levels of 1%.

, the coefficient of Sec_Net is significantly positive, which indicates that the municipal party secretary’s place of origin characteristics significantly affect the possibility and time of local government data platform going online, confirming the correlation hypothesis of the instrumental variable. In addition, both the weak instrumental variable test and the under-identification test are validated, indicating that the instrumental variables are justified. The second-stage estimation results shown in Column (2) suggest that the coefficient of the public data elements after exogenous variable fitting is significantly positive at the 1% level, indicating that public data elements are still beneficial to enterprise digital transformation after using instrumental variables to solve the endogeneity problem.

Considering that unobserved historical factors may affect the exogeneity of instrumental variable selection, it is helpful to refer to the method of Bellemare et al. (2017) [46]; this paper uses explanatory variables with two lags as the second instrumental variable to conduct endogeneity tests. On the one hand, this instrument variable meets the correlation requirements; that is, there is a significant correlation between the public data elements lagging behind in the two periods and the public data elements in the current period; on the other hand, it also meets the exogenous conditions; that is, after controlling the fixed effects and time trends of the enterprise, the public data elements lagging behind in the two periods do not directly affect the current digital transformation level of the enterprise, and their impact only affects the public data elements in the current period, thereby effectively avoiding the correlation problem between the instrument variables and the error terms. The empirical results show that the estimated direction of the two instrumental variables is consistent with significance. The above analysis once again verifies Hypothesis 1.

4.3.3. Exclusion of Contemporaneous Policy Effects

The launch of the government data platform was accompanied by several concomitant policies, such as the Big Data Comprehensive Experimental Zone Policy (Bdata) and the Broadband China Pilot Policy (Bbond). Implementing these two concomitant policies implies that the effect of public data elements on enterprise digital transformation is not a net effect. In order to eliminate the interference of accompanying policies on the estimation results, this paper draws lessons from Cao (2021) [47], adding the interactive item identifying whether or not it is a pilot policy, along with the implementation time, to the control variable of Formula (1), followed by re-estimating, and the regression results are shown in

Table 5. Robustness test results.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	Exclusion of Concomitant Policies		Metrological Replacement		Machine Learning
	DigiT	DigiT	DigiT2	DigiT3	Random Forest	Lassoed Return
PData	0.5312 *** (0.1223)	0.5015 *** (0.1221)	0.0779 *** (0.0199)	0.0346 *** (0.0061)	4.2429 *** (0.1933)	0.4002 ** (0.1775)
Bdata	0.4428 ** (0.2083)
Bond		0.5909 *** (0.1993)
Control variable	yes	yes	yes	yes	yes	yes
Firm/time fixed effects	yes	yes	yes	yes	yes	yes
City/time fixed effects	no	no	no	no	yes	yes
Observed value	33,396	33,396	33,396	33,396	33,396	33,396
Adjusted R2	0.3813	0.3815	0.3210	0.6771

Note: ** and *** indicate significance levels of 5% and 1%.

. Columns (1) and (2) represent the estimation results after adding the Big Data Comprehensive Experimental Zone Policy and Broadband China Pilot Policy to the original model, in turn, and the results show that the coefficients of the PData are significantly positive, indicating that after excluding the disturbing effects of concomitant policies, the public data element still substantially enhances the level of digital transformation of the enterprises. The robustness of the baseline regression results are again verified.

4.3.4. Replacement of the Measurement of the Explanatory Variables

This paper focuses on the CSMAR digital transformation lexicon and performs a full-text search of the annual financial reports of the listed enterprises, along with a separate search of their Management Discussion and Analysis (MD&A) sections, respectively, in reference to the work of Liu and Wang (2023) [48], which uses text analysis to construct digital transformation indicators. The results in Table 5 (3) and (4) demonstrate that the regression coefficients of public data elements (PData) in DigiT2 and DigiT3 are significantly positive, and the results of replacing the explanatory variables are still robust.

4.3.5. Dual Machine Learning Models

In the era of digital economy, the non-linear relationship between variables is more common, but traditional linear regression assumes that the variables are linear, and there is bias in the model set, making the estimation insufficiently robust. However, the machine learning algorithms embedded in dual machine learning can effectively deal with the non-linear data, successfully avoiding the adverse effects caused by the bias in the model set. On the other hand, although machine learning comes with regularization, which can achieve the effect of variable selection, there is still “regularity bias.” Dual machine learning, on the other hand, can effectively eliminate the bias through the use of residual modeling to ensure the unbiased estimation of the disposal coefficients. In addition, dual machine learning can construct confidence intervals, and its convergence speed is greater than that of machine learning. Thus, the robustness of the results, based on the dual machine learning model and the partially linear dual machine learning model, is verified, as follows:

$${\mathrm{D}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{T}}_{it+1}={\theta }_0{PData}_{it}+g\left(X_{it}\right)+{\epsilon }_{it}$$

(3)

$$E\left({\epsilon }_{it}|{PData}_{it},X_{it}\right)=0$$

(4)

Among these, ${\theta }_0$ is the coefficient of disposition that is the concern of this paper, and the rest of the variables are consistent with the representation of Model 1, with the exception that the specific functional form of $g\left(X_{it}\right)$ in Equation (3), $\widehat{g}\left(X_{it}\right)$, must be estimated using machine learning algorithms, and Equation (4) indicates that the conditional mean of the random error term is 0. Columns (5) and (6) in Table 5 show the regression results obtained using the random forest algorithm and the lasso regression algorithm, respectively, and these results show that the estimated public data element coefficients are significantly positive, indicating that the underlying conclusions still hold.

5. Further Analysis

5.1. Impact Mechanism Testing

Having obtained the empirical evidence above that public data elements enhance enterprises’ digital transformation, this section reveals the impact mechanism of the effects of information resource sharing and financing constraint alleviation. In order to verify the above effects, this section follows the principle of “cleaner causal identification” in modeling and establishes the following mechanism testing model, based on the research of Hayes (2022) [49]:

$$M_{it}={\lambda }_0+{\lambda }_1{PData}_{it}+\phi X_{it}+{\eta }_i+{\theta }_t+{\epsilon }_{it}$$

(5)

where $M_{it}$ is the mechanism variable, and the rest of the variables are consistent with the baseline model. The estimated coefficient of the public data element ${\lambda }_1$ is the object of interest in this paper, and the significance of this coefficient indicates that the mechanism is valid.

5.1.1. Information Resource Sharing Effect

Mechanism analysis clarifies that public data elements can promote firms’ digital transformation by enhancing information resource sharing, and to test whether this effect is practical, we refer to the work of Lang et al. (2012) [50]. We applies the number analysts tracking the listed companies (Analyst), the degree of attention paid to the research reports (Report), and the accuracy of the analysts’ surplus forecasts (Accuracy) to measure each firm’s data sharing as a proxy variable for the data information environment faced by the firms. The number of analysts tracked is the number of analysts or teams that have tracked the firm in a logged year; similarly, the number of reports followed is the number of reports that have tracked the firm in a logged year, and the measure of analyst earnings forecasting accuracy (AFA) is computed using the following formula:${ Accuracy}_t=-|\left({Med}_t-{Actual}_t\right)/{Price}_{t-1}|$, where ${Med}_t$ is the analysts’ predicted median EPS, ${ Actual}_t$ is the actual EPS, and ${Price}_{t-1}$ is the firm’s price per share in the previous year. Columns (1)–(3) of

Table 6. Mechanism Tests.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	Analyst	Report	Accuracy	WW	KZ	SA
PData	0.0415 ** (0.0188)	0.0617 *** (0.0237)	0.0044 ** (0.0019)	−0.0177 ** (0.0086)	−0.0721 *** (0.0277)	−0.0050 ** (0.0023)
Control variable	yes	yes	yes	yes	Yes	Yes
Firm fixed effect	yes	yes	yes	yes	Yes	Yes
Time fixed effect	yes	yes	yes	yes	Yes	Yes
Observed value	22,798	22,862	24,082	21,563	28,021	30,326
Adjusted R2	0.2556	0.2528	0.3481	0.2642	0.5963	0.2759

Note: ** and *** indicate significance levels of 5% and 1%.

show the results of regressing the public data elements for the mechanism variables Analyst, Report, and Accuracy, respectively. The results show that the estimated coefficients of the public data element are all significantly positive, at at least the 5% level, which indicates that the public data element can promote the level of digital transformation of enterprises by effectively enhancing the sharing of information resources, and Hypothesis 2 can be verified.

5.1.2. Financing Constraint Mitigation Effect

Mechanism analysis indicates that public data elements can promote enterprise digital transformation through financing constraint alleviation; to test whether this effect is practical, we adopt the WW index to measure financing constraints, i.e., the mechanism variable M is the WW index. Drawing on the ideas of Whited and Wu (2006) [51], this paper adopts the following method to calculate the WW index:

$${WW}_{i,t}={-0.091\ast CF}_{it}-0.062\ast {DivPos}_{i,t}+0.021\ast {Lev}_{i,t}{-0.044\ast Size}_{it}+0.102\ast {ISG}_{i,t}-0.035{\ast SG}_{i,t}$$

(6)

Here, CF represents the ratio of cash flow to total assets, which is the ratio of net cash flow generated from operating activities to total assets; DivPos represents the dummy variable of the cash dividend payment, which is 1 if the cash dividend is distributed in the current period, and 0 otherwise; Lev stands for corporate financial leverage ratio, which is the ratio of long-term liabilities to total assets; Size represents the enterprise scale, which is the natural logarithm of total assets; ISG represents the average sales growth rate of the industry; SG stands for the growth rate of sales revenue.

In addition, this paper also uses the KZ index and the SA index to measure financing constraints to enhance robustness. Columns (4)–(6) in Table 6 are the results of the regression of the financing constraint variables WW, KZ, and SA using public data elements. The results show that the estimation coefficients of the public data elements are all significantly negative, which indicates that the public data elements are beneficial in alleviating the financing constraints faced by enterprises. The existing research fully shows that financing constraints inhibit the improvement of the digital transformation level of enterprises (Luo, 2022; Xu et al., 2023) [52,53]; therefore, public data elements are conducive to digital transformation via the alleviation of financing constraints, and Hypothesis 3 is verified.

5.2. Heterogeneity Analysis

5.2.1. Analysis Based on Open Data Quality

The quality of open data is reflected by the value it can create, and the release of value is the result of the combined effect of the strength of the policy guarantee, the quality of the open data, and the construction of the platform system [39]. High-quality public data elements reflecting accuracy, completeness, and timeliness can enhance the efficiency of the decision making and operation of the enterprise and may be more conducive to promoting enterprise digital transformation. In order to verify the above speculation, this paper adopts the China Open Forest Index, made public by the Digital and Mobile Governance Laboratory of Fudan University, to measure the quality of public data elements. this measurement is obtained by multiplying the scores of four aspects, i.e., readiness, data layer, service layer, and utilization layer, and summing them up with corresponding weights, providing a more objective and comprehensive evaluation of public data open platforms around the world. Among these, the readiness index consists of two first-level indicators regarding regulations, policies, and organizational promotion, comprising the foundation of data opening, and which can characterize the strength of the policy guarantee. The data layer index consists of four first-level indicators regarding data quantity, open scope, data quality, and security protection, comprising the core of data opening and characterizing the quality of open data. The service layer index consists of four first-level indicators regarding the platform system and function operation, which is the pivotal point of data opening; it can characterize the construction of the platform system. The utilization layer index consists of five first-level indicators, such as the number of results, the quality of results, the value of results, etc., that show the effectiveness of data opening and which can be used to characterize the construction of the platform system in regards to the level of effect. In terms of data processing, this paper divides the samples into high-quality and low-quality groups, according to the median. The regression results are shown in Columns (1) and (2) of

Table 7. Heterogeneity analysis.

Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Quality of Data Elements		Vitality of Different Digital Economies		Enterprise Life Cycle
	Higher Quality	Lower Quality	High Vitality	Low Vitality	Growth	Maturity	Decline
PData	0.3596 * (0.2145)	−0.0052 (0.1828)	0.5978 ** (0.1826)	0.2432 (0.1872)	0.5038 ** (0.2025)	0.8003 *** (0.1923)	0.3141 (0.2295)
Control variable	yes	yes	yes	yes	yes	yes	yes
Firm fixed effect	yes	yes	yes	yes	yes	yes	yes
Time fixed effect	yes	yes	yes	yes	yes	yes	yes
Observed value	11,196	11,546	16,669	16,727	10,190	10,297	10,269
Adjusted R2	0.4054	0.3842	0.3694	0.3563	0.3790	0.3689	0.3422

Note: *, **, and *** indicate significance levels of 10%, 5%, and 1%, respectively.

. The estimated coefficient of the high-quality public data elements is significantly positive at the 1% level, but the impact of the low-quality public data elements on enterprise digital transformation is insignificant. This suggests that high-quality public data elements are more significant in enhancing enterprise digital transformation. Therefore, the government should focus on improving the quality of data elements to better assist enterprise digital transformation, and Hypothesis 4 is verified. Li and Xu (2024) [54] also mention the importance of data quality in discussing the influence of digital government on the digital transformation of enterprises. It is believed that higher data quality can reduce the information search costs of enterprises and promote digital transformation. This coincides with the view of this article.

5.2.2. Analysis Based on the Dynamism of the Digital Economy

Market dynamics in the digital era are closely related to business activities and profoundly affect the digital transformation of enterprises. Under different scenarios of digital economy vitality, the role of public data elements in the digital transformation of enterprises in different regions varies. The China Digital Economy Innovation and Entrepreneurship Index focuses on the innovation and entrepreneurial behavior of enterprises and can scientifically depict the level of digital economy vitality in each region. In this paper, the sample cities are categorized into high and low digital economy vitality regions, according to the median of the index for empirical analysis. The regression results are shown in Columns (3) and (4) of Table 7, and the estimated coefficient of the public data elements is significantly positive at the 1% level in high, but not in low, digital economic vitality areas. The possible reasons for this are that first, compared to low digital economic vitality regions, high digital economic vitality regions display higher degrees of technological development and stronger technological use capabilities, making it easier to exert both the financing constraint and the information resource sharing effects, rendering these areas more conducive to the digital transformation of enterprises. Second, compared to low digital economic vitality regions, high digital economic vitality regions exhibit richer public data elements, which can fully release the inherent technological capabilities and promote enterprise digital transformation. Therefore, public data elements significantly improve the level of enterprise digital transformation in regions with high digital economic vitality, and Hypothesis 4 is verified. Li et al. (2022) [55] found that the digital economy promoted digital transformation, which means that the digital economy is more dynamic, and that public data elements play a more obvious role in promoting the digital transformation, thus further supporting this view.

5.2.3. Analysis Based on the Enterprise Life Cycle

Unlike enterprises experiencing a recessionary period, enterprises in the growth and maturity periods tend to exhibit stronger capabilities to absorb and transform public data element resources. In order to compare the differences in the role of public data elements for enterprise digital transformation in the different life cycles of enterprises, this paper draws on the categorization of Anthony and Ramesh (1992) [56]. It uses the composite score discrimination method to classify enterprises into life cycles based on the scores of four indicators—sales revenue growth rate, retained earnings rate, capital expenditure rate, and age of the firm—to classify the firm life cycle. The operation is as follows: the total sample is divided by industry type in order to score the enterprises according to the above four indicators and calculate the comprehensive score; then, each industry sample is sorted in descending order, according to the comprehensive score, and the three-part method is chosen to divide the industry sample, with the first 1/3 of the enterprises with the highest scores classified in the growth period, the middle part of the enterprises classified in the maturity period, and the last 1/3 of the enterprises, with the lowest scores, classified in the recession period. Thus, for empirical analysis, the research sample is divided into the following categories: growth, maturity, and decline. The regression results, as shown in Columns (5)–(7) of Table 7, show that the estimated coefficients of the public data elements are significantly positive, at at least the 5% level, in the growth and maturity periods, but not in the decline period. This suggests that the facilitating effect of public data elements regarding enterprise digital transformation is more significant in the growth and maturity periods, validating the theoretical expectations, and Hypothesis 4 is verified.

Considering that life cycle classification methods based on comprehensive indicator scoring may be affected by sample survival bias, and that some companies, particularly those in recession, are more likely to withdraw from the samples during the study period, this paper introduces the cash flow classification method proposed by Dickinson (2011) [57] for robustness testing. This method divides the life cycle stages of the enterprise based on the positive and negative combination of three types of cash flows in enterprise operations, investment, and financing, and can characterize the financial behavior characteristics of the enterprise, to a certain extent. This paper uses this method to re-segment the life cycle of the enterprise and re-study heterogeneity on this basis. The relevant results are shown in

Table A2. Lifecycle heterogeneity based on cash flow classification.

Variable	(1)	(2)	(3)
	Enterprise Life Cycle
	Growth	Maturity	Decline
PData	0.6124 *** (0.1753)	0.4849 *** (0.1948)	0.4422 (0.2720)
Control variable	yes	yes	yes
Firm fixed effect	yes	yes	yes
Time fixed effect	yes	yes	yes
Observed value	14,574	10,918	6,157
Adjusted R2	0.4058	0.3832	0.3376

Note: *** indicate significance levels of 1%.

in Appendix A. The empirical results show that the positive impact of public data elements on enterprise digital transformation is mainly reflected in enterprises in growth or maturity periods. This finding is consistent with the conclusions obtained based on the comprehensive score method, which further enhances the reliability of the conclusions in this paper. Hu et al. (2023) [58] found that the mismatch of enterprise maturity can inhibit the digital transformation of enterprises, and this is more obvious in growing or mature enterprises. On the contrary, this paper finds that public data elements play a more significant role in promoting the digital transformation of enterprises in the growth and maturity period.

5.3. Analysis of Moderating Effects

Compared with traditional factors such as land and capital, the release of the value of public data factors is dependent on modern information networks and other carriers. The release of the economic value of public data elements requires digital infrastructure to provide technical support and the corresponding digital products to support services. Therefore, the release of the economic value of public data elements must be built on the matching digital infrastructure and digital products and services. Digital transformation is an essential economic behavior of enterprises; in order to test how digital infrastructure and digital products affect the promotion of public data elements in regards to enterprise digital transformation, this section establishes the following model:

$${DigiT}_{it}=\alpha +{\beta }_1{PData}_{it}\times R_{it}+{\beta }_2{R}_{it}+{\beta }_3{PData}_{it}+\phi X_{it}+{\eta }_i+{\theta }_t+{\epsilon }_{it}$$

(7)

where the moderating variable is R_it, and the remaining variables are the same as those in the baseline model.

This paper uses the Internet penetration rate to measure digital infrastructure and the digital finance level to measure digital products. The National Bureau of Statistics provides data on the Internet penetration rate, and the level of digital finance comes from the Digital Inclusive Finance Index of Peking University.

Column (1) in

Table 8. Analysis of regulatory effects.

Variable	(1)	(2)	(3)	(4)
	R = Inf	R = Pro	R = Invt	R = Inno
	DigiT	DigiT	DigiT	DigiT
PData × R	0.0033 ** (0.0017)	0.4501 *** (0.1208)	−4.5435 *** (1.5519)	0.4807 *** (0.1338)
PData	0.5078 *** (0.1212)	0.0121 *** (0.0033)	0.5110 *** (0.1221)	0.1813 ** (0.0724)
R	0.0051 * (0.0029)	0.0067 *** (0.0032)	−1.6101 * (0.8944)	0.4503 *** (0.0674)
Control variable	yes	yes	yes	yes
Firm fixed effect	yes	yes	yes	yes
Time fixed effect	yes	yes	yes	yes
Observed value	33,396	33,396	33,353	26,537
Adjusted R2	0.3814	0.3816	0.3812	0.3926

Note: *, **, and *** indicate significance levels of 10%, 5%, and 1%, respectively.

reports the results of the moderating effect of digital infrastructure on public data elements. The interaction term of digital infrastructure (Inf) and public data elements is positively significant, i.e., digital infrastructure moderates the relationship between public data elements and the digital transformation of enterprises. This shows that digital infrastructure is the basis for public data elements to release economic value, and digital infrastructure can significantly strengthen the facilitating effect of public data elements on enterprise digital transformation. This is because, for enterprises, digital infrastructure can not only improve the efficiency of data transmission but also enhance the ability for data storage and calculation, realize the efficient integration of data, and then enhance the influence of public data elements on digital transformation. Column (2) in Table 8 reports the results of the moderating effect of digital products on public data elements. The regression coefficient of the interaction term between digital products (Pro) and public data elements is significantly positive at the 1% level, i.e., digital products positively regulate the relationship between public data elements and enterprise digital transformation. This shows that the release of the economic value of public data elements must be based on advanced digital product services, and the higher the penetration rate of digital products, the more pronounced the facilitating effect of public data elements on enterprise digital transformation. This is because advanced digital tools reduce the complexity of data processing, improve the ability for data analysis, and greatly reduce the barriers for enterprises in the adoption of public data elements, thus enhancing the influence of public data elements on digital transformation. In view of the above analysis, Hypothesis 5 is verified. Most existing research focuses on the public data elements themselves [19], but lacks studies are lacking regarding the supporting measures to release the value of public data elements. Based on the empirical results of the above regulatory effect, this paper proposes that advanced digital infrastructure and digital products are prerequisites for public data elements to release economic value, which is conducive to strengthening the role of public data elements in promoting the digital transformation of enterprises.

This paper uses the investment expenditure rate, that is, the proportion of cash paid for the purchase and construction of fixed assets, intangible assets, and other long-term assets to total assets to measure corporate investment, and uses R&D investment to measure innovative investment. Column (3) in Table 8 reports the results of the regulation effect of traditional enterprise investment (Invt) on public data elements. The interaction terms between traditional enterprise investment and public data elements are positively significant; that is, traditional enterprise investment positively regulates the relationship between public data elements and enterprise digital transformation. This shows that traditional enterprise investment is the basis for public data elements to release economic value, and traditional enterprise investment can significantly strengthen the role of public data elements in promoting the digital transformation of enterprises. This is because for enterprises, traditional enterprise investment may strengthen path dependence, inhibit the efficiency of data transformation, and thus inhibit the digital transformation process of enterprises. The results of the regulation effect of innovation investment (Inno) on public data elements are reported in the column (4) of Table 8. The regression coefficient of the interaction terms between innovation investment and public data elements is significantly positive at the level of 1%; that is, innovation investment positively regulates the relationship between public data elements and enterprise digital transformation. This shows that the release of the economic value of public data elements needs to be based on advanced digital products and services. The higher the penetration rate of innovation investment, the more obvious the role of public data elements in promoting the digital transformation of enterprises. The reason is that innovative investment effectively stimulates the potential of data elements by improving the data absorption capacity and transformation intention of enterprises, thereby enhancing the impact of public data elements on digital transformation. In view of the above analysis, Hypothesis 6 is verified. The existing research analyzes the mechanism and heterogeneity of public data elements affecting the digital transformation of enterprises [30]. Regarding the lack of research on the regulation effect of enterprise investment, this article is based on the analysis of the regulation effect of traditional enterprise investment and innovative investment, pointing out that in the process of public data elements promoting the digital transformation of enterprises, enterprises should give full play to the positive regulatory role of innovative investment.

6. Conclusions and Policy Recommendations

Releasing the value of public data elements to drive the digital transformation of enterprises can promote the construction of smart city planning systems, thus achieving efficient allocation and sustainable use of land resources. Theoretical analysis shows that public data elements can be effectively integrated into enterprise production and operation to promote enterprise digital transformation. Based on the quasi-natural experiment with government data platforms launched online from 2011 to 2022, this paper empirically examines the relationship between public data elements and enterprise digital transformation, along with its mechanism of influence. The main conclusions are as follows: First, public data elements facilitate the level of enterprise digital transformation. Second, public data elements mainly affect enterprise digital transformation through two mechanisms, i.e., the information sharing effect and the financing constraint alleviation effect, as the public attributes of public data elements and the high value of data elements reduce the friction between the enterprise and the public system and market players. Third, the heterogeneity analysis finds that the enhancing effect of public data elements on enterprise digital transformation mainly exists in the case of high-quality data openness, high digital economic vitality regions, and enterprises experiencing growth and maturity. Fourth, a complete digital infrastructure and advanced digital products positively moderate the facilitating effect of public data elements on a firm’s digital transformation. In addition, corporate investment plays a negative regulatory role, while innovative investment plays a positive regulatory role.

The following policy recommendations are suggested in the article, based on the study’s findings.

The government should first play a significant role in releasing public data elements to encourage the conversion of public data into factors of production and to support the digital transformation of enterprises. Currently, the primary obstacle to opening public data elements is the lack of clarity regarding who owns the data and the uniformity of the opening mode, which results in a limited degree of opening. Therefore, the government must encourage the implementation of the public data rights authorization mechanism, attempt to clarify who owns the resources of public data, encourage the vitality of data opening, increase the effective supply of public data, and adopt the three open modes, i.e., unconditional openness without compensation, conditional openness without compensation, and conditional openness with compensation.

Second, to better support the digital transformation of businesses, policymakers should implement several measures to strengthen the financing constraint-relieving effect and the information resource-sharing effect. On the one hand, the government should encourage cross-industry and cross-departmental data integration and sharing, break down data barriers between departments, and maximize the use of public data resources. On the other hand, through policy guidance, the government can encourage financial institutions to use public data elements to implement data-driven financing products and services and investigate creative financing models that utilize public data elements to continuously optimize the financing environment for businesses.

Moreover, based on the circumstances of various regions and businesses, the government should implement unique and dynamic strategies to increase the success rate of businesses’ digital transformation. Since the driving force behind the digital transformation of low digital economy vitality regions and declining businesses has not yet been fully explored, the government should offer special funds and tax exemptions to companies with weaker digital technology capabilities to encourage them to introduce and adopt advanced digital technologies. At the same time, it should make it easier for businesses to access and use public data, provide low-tech companies with easily navigable tools and platforms, and offer them specialized technical support so they can easily apply data to their actual business, get past the challenges of digital transformation, optimize the allocation of land resources, and promote sustainable development.

Finally, the government should focus on digital infrastructure construction and digital product research and development to provide support for public data elements to help enterprises in digital transformation. Current challenges include insufficient capital investment in digital infrastructure construction and digital product research and development, difficulty breaking through technical bottlenecks, and a lack of digital talents. Therefore, the relevant governments should do the following: First, guide private capital to participate in digital infrastructure construction through PPP (government–social capital cooperation) and other means and increase investment in digital infrastructure construction. Based on the national Eastern Data Western Computing project hub node, pilot PPP projects in new infrastructure fields such as computing power centers and 5G base stations should be introduced. Second, provide R&D subsidies, tax incentives, and other policies to encourage enterprises to increase R&D investment in digital technology, strengthen technological innovation, and break through the technological bottlenecks. In particular, in the fields of artificial intelligence, industrial Internet, and other choke points, enterprises with more than 8% of R&D investment should be subsidized, step by step. Third, encourage colleges, universities, and scientific research institutions to enhance digital talent training, while enabling the introduction of overseas high-level talents and promoting the emergence of high-quality digital training. They should also establish a digital talent “special zone” in the national level digital economy demonstration zone, implement the “dual mentor system” for master’s and doctoral training, and require top enterprises (such as Alibaba Cloud) to take on no less than 100 targeted trainees annually.

Compared to existing research, this article has made significant progress in revealing the mechanisms by which public data elements drive digital transformation in enterprises. The existing literature mainly focuses on the endogenous digital transformation path of enterprises, while this article, through a quasi-natural experimental design, systematically demonstrates, for the first time, the causal effect of government data openness as an exogenous policy on enterprise digital transformation. This study found that public data elements have a dual empowering effect on enterprise digital transformation through information resource channels and financing constraint channels, deepening relevant research on enterprise digital transformation.

This study exhibits two limitations. Firstly, due to the regional segmentation attribute of the data open platforms, this article failed to delve into the synergistic effect of cross-regional data element flow, possibly underestimating the multiplier effect of infrastructure interconnection on digital transformation. Secondly, the lack of the dynamic assessment of potential risks in regards to the operation of public data authorization may weaken the constraining effect of institutional deficiencies on policy effectiveness.

The future research direction lies in the following: on the one hand, it is necessary to deepen the integration research of data elements and new infrastructure, construct a “data computing algorithm” trinity analysis framework, and deeply analyze the improvement space for infrastructure upgrading in regards to enterprise digital transformation. On the other hand, with the accelerated improvement in data infrastructure, the deep integration of public data openness, the industrial Internet, and the AI big model will reshape the underlying logic of enterprise digital transformation.