With the maturation of artificial intelligence (AI) technologies,many companies have begun to innovate their business models using AI.This trend accelerated significantly following the successful release of ChatGPT 4.0,which has led numerous companies to integrate AI into their business strategies.Consequently,the innovation and evolution of business models have become focal points of discussion in both academia and the industry globally.

Despite the growing body of research on AI and business models in recent years,the findings on this topic are fragmented and there is a lack of a unified research framework for systematically understanding it.It is therefore challenging to identify the core themes in current research on AI and business models and key directions for future research.In response to this shortcoming in the literature,this study systematically reviews 70 key articles on AI and business models from leading international journals.By classifying and organizing this literature,the paper identifies four main research themes:① The impact of AI on business model innovation,including its influence on overall business models and their components;② Archetypes of business models based on AI;③ The evolution of business models enabled by AI;④ The co-evolution of AI capabilities and business models.Through summarising and analyzing current research,the paper proposes key areas for future research,including the impact of AI on business model innovation at firm and industry levels;the classification of AI-enabled business models into archetypes and the drivers and outcomes of each archetype;the interactive factors involved in,drivers of,and barriers to the evolution of AI-enabled business models;and the relationship between AI capabilities and business models from a co-evolutionary perspective.

This paper contributes to the field in three main ways.First,it synthesizes the themes and gaps in research on AI and business models,thereby advancing the study of business models in the age of AI.Second,it reveals that AI-enabled business models exhibit characteristics of complex adaptive systems,including self-iteration and adaptability.Finally,by systematically reviewing current research,it highlights future research directions,guiding subsequent studies on AI and business models.Overall,the paper provides new theoretical and practical insights into how AI is reshaping business models and outlines potential pathways for future research.

The paper is structured as follows.The first section discusses the theoretical and practical background of the study.The second section reviews and defines relevant concepts in AI and business models.The third section details the literature review process followed and analyses the 70 articles.The fourth section elaborates on the four research themes related to AI and business models identified in the study and discusses the gaps in the literature regarding these themes.The final section presents the conclusions of the study and the prospects for future research on this topic.

The number of registered firms is widely used as an indicator of economic vitality and entrepreneurship,ignoring the fact that registered firms may be unproductive and destructive.This paper is among the first to explore unproductive firms by analyzing the visit records of nearly 50000 randomly selected registered firms.Normal business activities require firms to be fully exposed to the market so that they can be contacted by their upstream,downstream,and peer parties.However,there are always firms engaged in unproductive activities that choose to hide their contact information and avoid exposure to the market.Due to the lack of data,little is known about unproductive firms,and even some basic facts are not clear,such as the percentage of unproductive firms and their causes.Using the number of registered firms as a measure of economic vitality or entrepreneurship can be biased if the difference between the number of registered firms and the number of firms engaged in productive activities is ignored.

The Enterprise Survey for Innovation and Entrepreneurship in China (ESIEC) 2023 sampled nearly 50000 registered firms from the China Business Registration Database.The field study shows that 35.2% of registered firms cannot be contacted by both address and telephone number.By matching the surveyed firms with other big data on economic activities,we find that the probability of an out-of-contact firm appearing in the blacklist of firms is 0.3% higher than that of contacted firms,and the proportion of those engaged in normal economic activities,such as posting online job vacancies,bidding for government projects,applying for trademarks and patents,is even lower.Finally,out-of-contact firms are more likely to be shell firms; their registered capital (in log form) is 14.1% higher than that of contacted firms,but their number of employees in social security (adding one and taking the log form) is 15.1% lower than that of others.Further analysis shows that out-of-contact firms are more likely to be in high-tech service industries with more industrial policy (as measured by the share of government consumption) and in regions with poorer business environments.

This paper contributes to the literature in two ways.First,this paper combines field study data with other big data to profile productive and unproductive registered firms in China.Since Baumol (1996) proposed a distinction between productive,unproductive,and destructive entrepreneurial activities,few studies have been conducted to analyze unproductive firms.Among the smaller body of literature,Desai et al.(2013) constructs a model of destructive firms; Sobel (2008) measures unproductive activities with the number of political and lobbying organizations in capital cities in the United States,and productive activities with patents; it finds that excellent institutional environments lead to a greater influx of entrepreneurial talent into productive activities. First, this paper enriches this strand of literature by offering stylized facts in the Chinese setting.Second,the findings of this paper could help scholars and officials to understand and properly utilize the registration database.Third,our finding indicates that industries with more policies tend to have a higher proportion of out-of-contact firms,highlighting the need for improved policy implementation.

Digitalization is a critical factor in the modernization of traditional industries,but there is still a lack of quantitative assessment of the current level of digitalization in China's universal enterprises in academia.Specifically,it is unclear whether Small and Medium Enterprises (SMEs),which account for 90% of registered firms and 80% of urban employees,have adopted any digital technologies.The paper titled ‘Stylized Facts about Digitalization Processes of Chinese Enterprises—Evidence from the Enterprise Survey for Innovation and Entrepreneurship in China' aims to answer the question of which digital technologies Chinese enterprises adopt and how these technologies impact their business.It also aims to explore why some firms choose to conduct digital transformation while others do not.The study is based on the Enterprise Survey for Innovation and Entrepreneurship in China (ESIEC) conducted by the Peking University's Center for Enterprise Research.The Center for Enterprise Research designed the ESIEC 2023 questionnaire to collect information on the use of various digital technologies by sample enterprises.The questionnaire was informed by academic literature and surveys conducted in Europe and the United States.The collected results can be categorized into the following four types of digital technologies:① The term ‘Internet’ refers to various online platforms such as e-commerce,self-owned websites,and apps that are primarily used for external operations.② ‘Digital management’pertains to electronic data storage and professional business software that are primarily used for internal management. ③ The‘Online business division’ includes cloud services and internet outsourcing.④ ‘Digital machines’refer to CNC machine tools and industrial robots that are primarily used in manufacturing enterprises.ESIEC 2023 collects micro-survey data at the enterprise level through stratified sampling and interviewer field surveys.The sample includes all registered companies and individual households in the past six years in six provinces and cities:Liaoning,Shanghai,Zhejiang,Henan,Guangdong,and Gansu.

In this paper,we first estimate the digitalization process of new private enterprises of different industries and sizes by calculating the enterprise adoption ratios and employee penetration of four digital technologies.We summarize the stylized facts of the current digitalization process of Chinese enterprises from multiple perspectives,by comparing the differences in terms of enterprise type,regional distribution,and industry distribution.We comprehensively examine the level of digital transformation of Chinese enterprises through international comparisons using surveys conducted in Europe and the United States.Additionally,we explore the factors that influence enterprise digital transformation.

The paper's findings reveal that the digitalization of Chinese enterprises is rapidly developing and has a profound impact on the labor market.Specifically,61% of Chinese firms have adopted at least one of the digital technologies in Internet use,digital management,Internet outsourcing,and digital machines,affecting 85% of the employed population.However,there is significant heterogeneity among the different digital technologies.Among corporate companies,approximately 64% have adopted electronic data preservation.The adoption rate of digital management software,Internet platform operation and e-commerce sales is about 35%,and the adoption rate of cloud services,self-hosted websites and applications,and industrial robots is about 10%.Third,there are regional differences in the digitization process of enterprises.Zhejiang Province has the highest digital technology penetration among the six provinces,affecting 93% of the employed population,while Liaoning and Henan Province has the lowest digital technology penetration among the six provinces,but still reaches over 80%.Fourth,there are differences in the digitalization process of enterprises across industries.Business services have the highest digital technology penetration,affecting 91% of the employed population,followed by industry at 86%,residential services,and agriculture with the lowest.Fifth,in general,Chinese enterprises are still lagging behind European and American countries in terms of digital transformation.Chinese enterprises have a lower adoption ratio (39%) in Internet platform operation compared to European enterprises (59%).However,they have a developmental advantage in the adoption of e-commerce and short-video platforms.Additionally,the adoption ratio of Chinese enterprises in digital management is only half of that of U.S.and European enterprises.Sixth,this paper highlights that enterprise digital transformation is influenced by sizes,ages,and wage.Specifically,companies that have larger sizes,smaller ages,and higher wages are more likely to implement digital transformation.

This paper has the following outstanding features:First,objectivity,this paper's metrics for enterprise digital transformation are based on specific digital technology adoption in micro-survey data.The constructed indicators objectively reflect the real situation of enterprise digitization.Second,multidimensionality,this paper utilizes research questionnaires from Europe and the United States to refine four specific technologies:Internet use,digital management,online business division of labor,and digital machines and other digital technologies.Third,representativeness,this paper uses ESIEC 2023 data sampled based on the business registration database of enterprises,and the research object covers enterprises of different industries and sizes.Fourth,timeliness,this paper analyzes the digital transformation of the newly built enterprises in the last six years in 2022,which can reflect the latest situation of the digitalization of China's industries.

Innovation is a key driver of productivity,enhancing enterprise competitiveness,performance,and national economic growth.Therefore,technological upgrading and innovation have become crucial for national development and transformation.

For Micro,Small,and Medium Enterprises (MSMEs),innovation holds increasing importance for survival and development.However,the research on the innovation activities of MSMEs remains more limited compared to larger firms in large due to lack of data.

The Enterprise Survey for Innovation and Entrepreneurship in China (ESIEC) addresses this data gap by conducting surveys specifically on MSMEs.This article first reviews the theories and literature related to innovation measurement and innovation surveys.It then introduces the ESIEC questionnaire design,analyzes the basic facts of enterprise innovation data in ESIEC surveys,and finally employs the CDM model to analyze the relationship between R & D,innovation,and productivity in MSMEs included in the ESIEC surveys.The results confirm consistency with existing literature on the relationship between these three factors.Compared to existing micro-level innovation data in China,the ESIEC survey offers four key advantages:①It covers multidimensional innovation measurements and comprehensive information on innovation,including not only standard international innovation survey data but also incorporating indicators relevant to China's unique economic characteristics and high-quality development goals; ② It includes small and micro-enterprises as survey subjects; ③ It encompasses a variety of service industries alongside the manufacturing sector; ④ Unlike many online or telephone surveys,ESIEC utilizes face-to-face interviews conducted by rigorously trained interviewers with enterprise leaders or executives,ensuring data quality.In conclusion,ESIEC data provides a valuable data source for analyzing innovation activities of MSMEs in China.Preliminary analysis based on ESIEC data aligns with the existing literature,suggesting reliable data and the possibility of comparative analysis with well-known surveys,such as  the  Community Innovation Survey (CIS) and the Annual Business Survey (ABS) in the United States.

Small and medium-sized enterprises (SMEs) have made important contributions to China's economic development.However,SMEs generally exhibit weak resilience and are constrained by limited capacities of financing and risk sharing.When facing negative economic shocks,they often lack the resources and capabilities to adjust and alleviate difficulties.Therefore,what challenges will SMEs face under negative economic shocks?How will these challenges hinder SMEs from recovery and growth?What adjustments will they make to cope with these challenges?How will negative economic shocks affect the development of SMEs in the medium and long term?Answering these questions not only holds clear theoretical significance but also practical significance and policy implications,especially today when many SMEs have not fully recovered from the negative economic impacts of the COVID-19 pandemic.

Based on data from two rounds of the Enterprise Survey for Innovation and Entrepreneurship in China (ESIEC) conducted in 2018 and 2023,we attempt to answer the series of questions raised above using the COVID-19 pandemic as a typical negative economic shock.The ESIEC is a field survey conducted by Peking University on private enterprises in China.In 2018,the research team conducted the first baseline survey,covering 6198 firms and accumulating quite detailed enterprise data.In 2023,the research team conducted the second large-scale field survey,completing on-site surveys for 6117 enterprises and investigating the challenges encountered by enterprises during the pandemic and their responses.The two rounds of surveys happened to span the COVID-19 pandemic,providing us with an opportunity to systematically evaluate the impacts of negative economic shocks on SMEs,the underlying mechanisms,and to detail the adjustments made by SMEs to cope with these shocks.
In this paper,we employ a “quasi-difference-in-differences” model to examine the challenges faced by SMEs under the impact of the COVID-19 pandemic,as well as the strategies they adopted in response and adjustment,taking advantage of the variation in strength of pandemic control measures across regions and over time.We find that:① The pandemic influences the supply side of SMEs mainly due to the “shutdown” effect caused by supply chain disruptions.②  To cope with the impact of the pandemic on the supply side and strengthen the resilience of the supply chain,SMEs need to systematically adjust the size and spatial layout of  the supplier network,but at the same time bearing higher transportation and transaction costs.This effect is more pronounced in regions with lower industrial agglomeration.③ The pandemic affects the demand side of SMEs mainly due to the “contraction” effect brought about by market shrinkage.④ To cope with the impact of the pandemic on the demand side,SMEs need to further explore customer resources,strengthen innovation efforts,and adjust business models to ensure market size and increase profitability.Whether these adjustments can be made in a timely and effective manner is highly correlated with the resources and backgrounds of the enterprises and entrepreneurs.⑤ The pandemic has significantly negative effects on revenue and gross profit,and may lead to a “scarring effect” delaying the recovery and development of SMEs.

These findings contribute to our understanding of the key mechanisms for the recovery of SMEs in the post-pandemic era.They also provide theoretical insights and empirical evidence for the targeted design of policies to assist SMEs,facilitating an efficient and balanced economic recovery.Specifically,the  “scarring effect” of the negative economic shock like the COVID-19 pandemic may hinder the development of SMEs in the medium and long term not only due to the persistent impacts of the  “shutdown”and  “contraction”effects but also largely due to a series of adjustments made by enterprises in response to the shock.Policymakers should consider aligning relief policies with the adjustments made by enterprises themselves to amplify policy impacts,thereby effectively promoting and accelerating the recovery of enterprises and the economy.At the same time,negative economic shocks often lead to a  “reshuffling”of the market landscape.Enterprises that can adjust promptly and effectively,and even capitalize on opportunities to upgrade,often gain greater competitive advantages and market share after the shock,potentially leading to changes in market structure and even widening regional disparities.Therefore,in policy design for the post-pandemic era,it is also necessary to fully consider the balance between enterprises and between regions,thus achieving economic recovery efficiently and equitably.

Entrepreneurial networks play a pivotal role in the viability and advancement of startups.The introduction of the network agency perspective has directed research attention toward entrepreneurial networking,investigating the question of “where does the entrepreneurial network come from”,i.e.,how entrepreneurs establish and develop networks.Entrepreneurial networking,in essence,encompasses a series of behaviors,activities,and means undertaken by entrepreneurs to form and develop networks with external stakeholders.

However,previous studies on entrepreneurial networking remain in a fragmented stage of strategy exploration,failing to elucidate the underlying theoretical assumptions of different entrepreneurial networking strategies comprehensively.Besides,the conceptual connotations of entrepreneurial networking and the decision logic behind it have not been clearly identified.Specifically,following different theoretical assumptions such as “design-precedes-execution”,“execution-precedes-design” and “convergence of design and execution”,entrepreneurial networking exhibits differentiated modes and strategies.Meanwhile,the exploration of antecedents of entrepreneurial networking in existing research is still in its nascent stage,making it difficult to fully explain the reasons why entrepreneurs adopt distinctive networking modes.Hence,this paper coordinates various underlying theoretical assumptions of entrepreneurial networking and systematically analyzes its conceptual connotation.Then,we take the decision logic of entrepreneurs into account and summarize three paths concerning the “decision logic-networking mode” relationship.

Based on systematic collection and analysis of existing literature,this paper demonstrates a path framework of entrepreneurial network formation.First,according to its theoretical assumptions,entrepreneurial networking is divided into three modes,namely “design-execution” mode,“execution-design” mode,and “design×execution” mode.Specifically,the “design-execution” mode regards entrepreneurial networking as a goal-oriented activity and a strategic and instrumental resource-seeking action,while the “execution-design” mode takes entrepreneurial networking as an inspiring entrepreneurial network-building action,which provides basis for the formation of goals of entrepreneurial action.Combining the two assumptions,the “design×execution” mode treats entrepreneurial networking as actions based on a vague and dynamically adjusted plan,with an orchestration of people,resources,and ideas.

Second,grounded in the cognitive perspective,this paper initiates a comprehensive framework for entrepreneurial networking,discerning three source paths of entrepreneurial networks.First,in network collecting path,entrepreneurs adopt the causation logic and are driven by goals,strategically selecting valuable network relationships,making plans,and establishing network relationships with target partners.During such process,entrepreneurs face a risky context,and entrepreneurial networking is essentially about finding the best means to achieve the established goal.And the final structure and form of the entrepreneurial networks depend on the goals set by the entrepreneurs in advance.Second,in the network convergence path,entrepreneurs adopt the effectuation logic,combining internal factors such as personal experience and ability with external conditions such as accidents to take different behaviors to develop the entrepreneurial network.During such process,entrepreneurs will give full play to their subjective initiative,utilize contingency to control new situations,and even create new means to strive for the best possible results.The structure and form of the entrepreneurial networks are jointly created and defined by entrepreneurs and their partners without possibility to be predicted in advance.

Third,in the network “collecting+convergence” path,entrepreneurs adjust their decision logics and networking strategies according to the level of situational uncertainty throughout the life cycle of the enterprise development,taking analytical and planned networking actions to integrate the entrepreneurial network while remaining open to  contingency. They usually connect with partners in an inspiring way and gradually refine entrepreneurial networks while focusing on achieving their long-term goals.Then,taking regional institutional environment,stage of entrepreneurial enterprises,pre-existing network characteristics and entrepreneurs' previous experience as examples,this paper discusses the moderating effects of uncertainty on the above paths of entrepreneurial network formation.

Eventually,future research directions are proposed from three aspects:refining the concept of entrepreneurial networking,exploring new issues in digital context,and expanding research methods.First,future research is encouraged to conduct in-depth exploration of the process mechanism of dynamic evolution of the “collecting+convergence”path.For example,dynamically characterizing the behaviors of entrepreneurial networking driven by various decision logics.Second,based on the characteristics of the digital organizations,the unique concept,strategies and mechanisms of platform-based entrepreneurial networking are under exploration.Third,it is recommended to apply new research methods such as machine learning and large models to effectively utilize digital platform big data and capture entrepreneurial' networking strategy and its mechanisms.

This paper contributes to entrepreneurial network research by summarizing various modes of entrepreneurial networking and proposing a path framework concerning “cognition-behavior” framework.Hence,this research helps to enrich the understanding of entrepreneurial networking and explore the antecedents of entrepreneurial networking strategy to understand why and how entrepreneurs adopt distinctive networking strategies from a more integrated perspective.Moreover,this review helps entrepreneurs understand the conceptual connotation and strategies of entrepreneurial networking more comprehensively and guides them to adopt appropriate strategies based on various decision logics,so as to construct their entrepreneurial networks effectively.

China's automotive industry is undergoing a significant transformation towards electrification and intelligent mobility,emphasizing the need for effective marketing strategies to enhance brand perception and navigate the challenges in traditional sales channels.Within this evolving landscape,telesales play a crucial role in screening customer intentions,a process vital for boosting sales efficiency and reducing costs.However,the reliance on the subjective judgment of sales representatives in this multi-step process often leads to inefficiencies,including missed opportunities and resource wastage on low-intent customers,underscoring the need for a more precise approach in aligning marketing efforts with customer intentions in the automotive sector's dynamic environment.

Our research introduces an innovative model aimed at identifying customers' purchase intentions through the analysis of voice data from telesales conversations.Utilizing advanced voice analysis technology,this model scrutinizes various aspects of customer responses,including emotional,attitudinal,and cognitive dimensions,employing logistic regression techniques for accurate classification.The validation of this model involved an extensive dataset of 542 authentic customer voice recordings,which unveiled a significant link between factors such as emotional arousal,patience,and the likelihood of purchase intentions.Impressively,the model's precision,determined by AUC metrics in tests beyond the initial sample,surpassed the 92% mark.This level of accuracy underscores the model's effectiveness in pinpointing true purchasing intentions,setting it apart in the realm of customer intention analysis.The outcomes of our study underscore the model's practical value,especially in the context of telesales where distinguishing genuine buyers from a vast pool of contacts is crucial.By homing in on pivotal indicators like emotional arousal and patience,our model adeptly filters through the noise to identify those customers most inclined towards making a purchase.This capability not only streamlines marketing efforts but also significantly boosts sales efficiency by allocating resources more judiciously and increasing the focus on high-potential leads.Moreover,the real-world applicability of our model was further evidenced by its remarkable performance metrics,achieving a 90% success rate in capturing positive examples at a coverage rate of 23.1%.Such compelling results highlight the model's robustness and its suitability for broad implementation across the telesales operations within the automotive sector.The research not only paves the way for a more focused and effective marketing strategy but also heralds a new era in customer relationship management,where understanding and meeting the nuanced needs of potential buyers through sophisticated voice analysis becomes a cornerstone of success.

The automotive industry's evolution towards electrification,digitization,and sustainability marks a pivotal era demanding innovative marketing strategies.This model represents a significant advancement in utilizing voice analysis for customer intention identification,offering a more objective,scientific approach to understanding and catering to potential buyers' needs.Through this model,automotive companies can refine their telesales strategies,prioritizing high-intent customers,and thereby maximizing the efficiency of their marketing efforts.

Banks can provide illiquid loans to enterprises,which are financed by the demand deposit which allows depositors to withdraw funds at any time,that is,liquidity creation.However,the potential mismatch of liquidity in the banks' balance sheet caused by liquidity creation may lead to bank runs.It is obvious that households will tend to withdraw funds when they hold negative expectations about the banks.Therefore,bank runs have always been a major obstacle faced by banks and received significant attention from academia and governments despite the continuous improvement of financial regulation and supervision.An effective means to deal with bank runs is deposit insurance.However,it has been proven to have many limitations such as moral hazard in recent literatures,thus we try to find some macroprudential policies to address bank runs in this article.

This article aims to extend the Diamond-Dybvig model(DD model) and establish a bank run model that includes a household sector,a production sector,a banking sector,and a government sector.In period 0,financial intermediaries absorb savings from the household sector and make their investment decision at the end of this period; due to the news shock,the withdrawal decisions of the heterogeneous households determine whether a bank run occur,and the impatient household sector begins to consume in period 1; then all deposits are returned,and the patient household sector begins to consume in period 2.Compared with the DD model,banks can freely choose investment portfolios (liquid assets or loans) to maximize their profits and face the risk of early withdrawals by patient households in our model.In addition,this model introduces a government sector to analyze the impact of liquidity coverage ratio on bank run risk and social welfare loss when bank runs happen due to irrational expectations of household sector.Thus,this paper can provide theoretical and practical significance for the establishment of our country's macroprudential framework.

We find that banks can help form the optimal allocation of resources in the economy through liquidity creation.But when they choose risky investment decisions according to profit maximization,bank runs often occur due to insufficient liquidity,thereby destroying the optimal allocation of resources and forming a run equilibrium.As a result,the economy suffers an efficiency loss,and the welfare of households decreases sharply.The liquidity coverage ratio could prevent bank run risks from two aspects.On the one hand,liquidity coverage ratio could force banks to hold more liquid assets,thereby improving the bank's asset structure and directly reducing welfare losses caused by bank runs; on the other hand,liquidity coverage ratio could ensure that banks can demonstrate to all households that they can“provide household deposits which they commit at any time” when faced with early withdrawals by some households in period 1,which can help patient households form good expectations towards the bank,thereby reducing the impact of news shocks on early withdrawals by patient households.However,liquidity coverage ratio regulation is not cost-free.This article explores the optimal liquidity coverage ratio and finds that while liquidity coverage ratio helps increase banks' liquid assets,it also leads to the production sector receiving less funds,thus a loss in the efficiency of the economy.In addition,this paper further expands the model to explore the optimal liquidity coverage ratio when there is uncertainty in household heterogeneous shock and news shock.We find that when uncertainty exists,the liquidity coverage ratio set by the government to prevent bank runs needs to cover the size of households with early withdrawals under the upper bound of the shocks.Therefore,the bank's profitability would further decline in this case,which implies that the funds towards production will further decrease,and the efficiency of the economy further decline.

This article provides a general model framework based on the DD model that can be used for policy and welfare analysis and enriches the theoretical model of bank runs and macroprudential policy in China.Meanwhile,this article introduces a macroprudential policy,that is,liquidity coverage ratio,and explores the optimal ratio and transmission channel of this policy,providing theoretical enlightenment for the government to formulate and implement related macroprudential policies,which is of great significance in the future policy design.

Based on the theoretical mechanism of bank run risk and the macroprudential policy described by the DD model,this paper provides a potential model framework for future theoretical analysis of bank runs,macroprudential policies,and other economic policies to explore the effects of different policy combinations.In addition,this article does not introduce more complex information friction situations in the model.According to the existing literature,information friction is often one of the important factors that lead to bank runs.Therefore,how to introduce information friction into the theoretical models of bank run is also an important future research direction.

The key to sustainable growth in banking lies in the Communist Party of China's (CPC) leadership and the refinement of internal governance mechanisms.Effectively preventing financial risks will significantly propel the banking sector towards high-quality development.Hence,it is crucial to actively explore how organizations of the CPC can engage in the governance structure of banks,and establish a governance model tailored to China's unique characteristics.Commercial banks face substantial information asymmetry and high regulatory costs,making refined internal governance mechanisms pivotal for risk management within these institutions.As the statutory procedures and mechanisms for CPC committee members to join corporate boards and senior management continue to improve,a “two-way entry and cross-serving” model emerges,forming a dual governance framework that blends “CPC-led governance” with “modern corporate governance”.

This paper summarizes the involvement of CPC organizations in the governance of 215 commercial banks from 2002 to 2020.In the context of economic policy uncertainty,this paper examines how the integration of CPC into governance through “two-way entry and cross-serving” affects risk prevention in banks.The findings suggest:Firstly,CPC's involvement significantly reduces banks' risk exposure by mitigating the adverse effects of economic policy uncertainty.Particularly,when the CPC committee's Secretary concurrently serves as the Chairman,this combination has a more pronounced impact on reducing bank risk compared to CPC committee members in supervisory and managerial roles.Secondly,CPC's “Three Importance and One Greatness” system,covering major issues,important appointments and removals,major projects,and the use of large amounts of money,displays proactive risk prevention features.Thus,during increased economic policy uncertainty,banks with CPC involvement in governance mitigate operational volatility,reducing risk exposure.This is evident in banks avoiding excessive liquidity hoarding and diversifying loans to decrease concentration,enhancing their capacity to serve entities.Thirdly,even after accounting for inherent influences by organizing and assessing martyrs' cemeteries and the number of martyrs in the banks' regions,the results remain significant.

The above results demonstrate that integrating CPC with bank governance effectively enhances the capability to serve financial entities.Continuing to refine the “two-way entry and cross-serving” leadership model proves crucial in bolstering risk prevention in the banking sector.It also adeptly counters the adverse impacts of rising global economic policy uncertainties on the financial system.

Therefore,this paper provides China's experience in preventing and resolving significant financial risks by integrating CPC with bank governance.In contrast to Western traditional corporate governance studies,this article combines China's institutional system,cultural background,and the centralized leadership of CPC in economic affairs to demonstrate that CPC involvement in bank governance aligns with the interests of the CPC and the people.Simultaneously,it effectively reduces principal-agent conflicts.Combining data and empirical analysis,the dual governance system integrating Party with corporate governance is scientifically and significantly essential for refining internal governance mechanisms within banks.It offers a new research perspective on bank governance and risk control.

CPC organizations embody a collectivist culture.Through the principle of “the Party assuming the responsibility for cardres' affairs”,these organizations can engage more deeply in supervising and managing processes within banks and even the entire financial sector,aligning well with China's reality.In the future,exploring the specific pathways of CPC organization involvement in governance and finding more scientific means to leverage this role will be crucial for better utilizing this unique system and enhancing internal governance mechanisms.

The digital economy has developed rapidly in recent years and has become an important  economic development engine  in the new era.Data,as a strategic resource for the development of the digital economy,in which enterprises can release massive amounts of data information, are playing an important role in data sharing, openness, and development and utilization.However, in recent years, frequent data leakage incidents have exposed a series of issues such as inadequate supervision and imperfect sharing mechanisms in data sharing.Given that data has become a strategic resource driving enterprise value creation,governments around the world are paying more and more attention to the governance of the enterprise data market.In May 2018,the EU enforced the General Data Protection Regulation (GDPR),which sets out the standards for personal data protection and requirements for enterprise data sharing.In addition,the EU is actively exploring a data pooling mechanism to facilitate data sharing among enterprises.China's Development and Reform Commission (DRC) and many other departments have also encouraged all types of subjects to voluntarily participate in data element sharing by issuing relevant documents and establishing laws and regulations to promote data sharing.The aim is to mitigate privacy risks and hazards and to enable openness and sharing of data.

This study considers big data alliances, members of big data alliances, and relevant government departments, and conducts a theoretical analysis of the effective sharing of decision-making impact data among various entities based on the theory of tripartite evolutionary games.In view of the problems of free-riding by big data alliance members and data leakage caused by big data alliance malfeasance in the process of data sharing,this study constructed a three-party game model to explore the decision-making choices of the big data alliance,big data alliance members,and relevant government departments in data governance,analyzed the stability of the three-party evolution game,judged the stability of equilibriums,combined with the simulation analysis of the data to test the impact of the factors on the stability of strategies,and verified the robustness and effectiveness of the evolutionary results from the theoretical level.factors on the stability of the strategy,verified the robustness and validity of the evolutionary results,and analyzed the mechanism and potential impact of these speculative behaviours from the theoretical level.

The research results indicate that:① Although the probability of big data alliance leakage increases as the value of data information among members of the big data alliance increases, the government's regulatory approach of allocating margin benefits still has a certain degree of governance effectiveness.② The decision-making of big data alliances is difficult to be influenced by social gains or losses, and the diverse governance at the social level lacks influence on the choice of game strategies.③ Members of the Big Data Alliance are highly sensitive to the margin system, and this measure as compensation for data governance can effectively alleviate the phenomenon of free riding among members of the Big Data Alliance.④While evaluating governance measures to improve the data market, the government also needs to measure whether the distribution of profits from margin is fair and reasonable.The tripartite evolutionary game model constructed in this article can provide a certain reference basis for improving the data governance mechanism and promoting the sustainable development of the data-sharing market.These results we got show that the government's adoption of the margin system has a certain constraint on the decision-making choices of big data alliances,and the government should formulate the margin system reasonably and ensure that the allocation scheme of the margin is scientific and appropriate.The members of big data alliances are more sensitive to the security deposit levied by big data alliances,so big data alliances should optimze the security deposit system to cope with the free-riding behaviour of their members.It is difficult for big data alliance to be influenced by external gains and losses brought by the social level in its decision-making process,so it is difficult to influence the choices of big data alliance by influencing its corporate reputation through social opinion,etc.In addition to the government's deposit system,strengthening the construction of internal self-governance mechanism is also one of the effective paths.If a member of a big data alliance leaks information about its members' data,it will quickly attract the attention of the relevant government departments and be governed,while the free-riding behaviour of the big data alliance members is difficult to be detected by the government in a timely manner.Therefore,the internal regulatory mechanism can be strengthened to give full play to the governance effectiveness of big data alliances.

The research in this paper treats the data shared by big data alliance members as homogeneous and does not further distinguish between the difference in data costs of members upstream and downstream of the supply chain as well as the cooperative relationship between members.Therefore,it is the next research direction to consider including both upstream and downstream enterprises of big data alliance members in the model,constructing an evolutionary game model under the participation of upstream and downstream members,and investigating the mechanism of bargaining power in the negotiation of cost-benefit distribution in the co-operative governance among the members on data sharing,so as to put forward constructive suggestions for the realization of effective data governance.