As implicit government guarantees disappear and the rigidity of bond market payments is broken, the efficiency of the corporate bond market is gradually improving, highlighting the significance of studying bond market pricing patterns. Therefore, this paper proposes a parametric pricing method based on machine learning and bond characteristics. By leveraging machine learning technology to utilize high-dimensional bond characteristics, we estimate the stochastic discount factor for corporate bond pricing, fully extracting both linear and non-linear pricing information. This method can obtain analytical solutions and has economic interpretability. Theoretically, it is demonstrated that this method is equivalent to the parametric portfolio approach, enriching the economic connotation and estimation method of the stochastic discount factor for corporate bonds. Research on Chinese corporate bond market shows that: 1) The parametric pricing model extracts more corporate bond pricing information than the classical factor model by capturing complex pricing relationships and weak factors from high-dimensional bond characteristics. 2) Return-related and liquidity-related bond characteristics are most important for corporate bond pricing, and the importance and predictive direction of these characteristics exhibit strong time-varying properties. The fundamental characteristics of the issuer cannot provide additional pricing information beyond bond characteristics. 3) The parametric pricing model has stronger pricing capabilities for bonds with high duration, high volatility, low credit ratings, low liquidity, and those issued by non-state-owned enterprises and non-listed companies; its pricing ability weakens in an expanding macroeconomic state and strengthens otherwise. This paper expands the research framework of corporate bond pricing theory, helps to understand corporate bond pricing patterns, improves market pricing efficiency, and prevents bond risks.
Motivated by physics-informed neural networks, this paper introduces a new framework called finance-informed neural networks (FINN), which integrates financial theory with deep learning technology. The goal is to improve the transparency and accuracy of deep learning methods used in empirical financial research. FINN’s network structure is based on arbitrage pricing theory and empirical portfolio construction techniques. Market efficiency information is integrated into the loss function in the training of FINN. An empirical study of China’s A-share market demonstrates that FINN has several advantages. It outperforms conventional fully connected neural networks and traditional empirical APT factor models in terms of out-of-sample
As a significant force driving the leap in productivity and industrial transformation, the new generation of artificial intelligence (AI) has been deeply integrated into economic and social operations as well as industrial development. In this context, data is gradually emerging as a core element differentiating competition in the AI field, with high-quality datasets significantly enhancing the efficiency and proficiency of model training for AI service providers. This paper focuses on the collaboration between data providers and data trading platforms to improve the quality of data granules, which aims to foster the development of the AI industry. Additionally, considering the non-cooperative game in the data supply chain, this paper constructs a noncooperative-cooperative biform game model to tackle this challenge, and studies the equilibrium strategies and optimal profits of all parties involved, as well as analyzing social welfare. Finally, a numerical analysis has been performed to examine how key factors influence data supply chain strategies, offering a theoretical foundation for decision-making by data providers, data trading platforms, and government. The findings indicate that: 1) The optimal selling prices and maximum profits for data providers and data trading platforms, the quality level of data granules, and overall social welfare are negatively related to the price sensitivity coefficient. 2) As AI service providers’ preference for data granules’ quality increases, the data supply chain actively enhances the level of data granules’ quality, leading to an increase in selling prices for data providers and data trading platforms, an improvement in social welfare, while a decrease in data providers’ dependence on data trading platforms.
In the digital age, B2C (business-to-consumer) product-sharing as a new sharing economy business model emerges as the times require. A B2C product-sharing platform usually has two channel structure models to choose from, namely model R (only establishing product-sharing channel) and model RS (establishing a product sales channel in addition to the product-sharing channel). Considering the two unique characteristics of B2C product-sharing: Usage value discount (a consumer perceives a low usage value due to the inconvenience of going to a designated location or waiting for some time to use a shared product) and pooling effect (a same shared product can be used by multiple consumers at different times over an entire period), this paper investigates whether and under what conditions a B2C product-sharing platform should add a sales channel, that is, adopting model RS, and further analyzes the impact of the channel structure model change on consumer surplus. The results show that when the usage value discount factor is not small, the platform should add a sales channel if the pooling effect is small, at this time, adopting model RS will maximize both platform’s profit and consumer surplus, achieving a win-win outcome; Otherwise, the platform should not add the sales channel. Under certain conditions, abandoning the sharing channel to transform into a retail platform that only establishes a single sales channel may increase platform’s profit and consumer surplus.
As the platform-as-a-service (PaaS) industry rapidly rises, the nascent nature of the industry has led to sharp conflicts and challenges between market expansion and profit maintenance. Addressing these issues, this study incorporates the heterogeneous characteristics of user innovation, and from the perspectives of market-scale optimization and profit maximization, proposes two pricing strategies: Uniform pricing and personalized pricing, and explores the pricing strategy selection and service quality upgrade design of PaaS service providers. The findings indicate that personalized pricing generally exhibits broader advantages. When service quality requires frequent upgrades or user preferences are moderate, the PaaS provider should prioritize market-scale optimization. Conversely, profit maximization should be the primary goal when service upgrades are less frequent and user preferences are weaker. Profit-maximization strategies help mitigate conflicts between initial service quality, service quality upgrades, and personalized pricing. Additionally, a critical threshold in service upgrade frequency can reverse the negative impact of quality improvement on personalized pricing, while increasing upgrade frequency alleviates early adopters’ aversion to personalized pricing.
To dispel consumers’ doubts about counterfeit goods sold through online channels, many companies are adopting blockchain to build product traceability systems for providing tracing information. Given the differences in product traceability scope and brand value enhancement offered by consortium blockchains, it is crucial to explore how companies adopt blockchain traceability systems. By constructing a supply chain analysis model that includes one manufacturer and two retailers, this paper explores the optimal adoption strategies of manufacturers and their official channel contracts, which compares three adoption strategies of manufacturers: Adopting blockchain alone, joining a consortium blockchain, and leading consortium blockchains. The results indicate that manufacturers are always motivated to adopt the consortium blockchain strategy. For manufacturers, if consumers have a moderate level of suspicion about channel sales and high degree of perception of product brand value, it is always advantageous to join the consortium blockchain under wholesale contracts; otherwise, it is advantageous for manufacturers to lead the consortium blockchain when consumers have a low degree of perception of product brand value. Additionally, the impact of different channel contracts on channel profits, consumer surplus, and social welfare varies. This article provides a theoretical basis and decision-making suggestions for manufacturers on whether and how to adopt consortium blockchains, thereby improving the brand value and consumer trust.
Time is one of the main factors for each decision maker to evaluate military action in military game. In order to describe the influence of time factors on decision makers in military conflicts, a military game graph model based on decision time is established firstly, and a series of modeling factors such as the two kinds of time required for unilateral movement and unilateral improvement based on decision time are defined. Secondly, the stability definitions of graph model based on decision time are given. Finally, the model is applied to a potential military conflict case to solve the stable states and analyze the evolution path of the game. The analysis results of the model show that the model and method constructed in this paper can effectively simulate the behaviors of the decision makers, directly reflect the game results, and have certain guiding significance for the actual combat.
In the era of the digital economy, safeguarding corporate information assets has emerged as a crucial research issue. This paper constructs an optimal security decision model for firms, considering their risk preferences (aggressive and conservative) and categorizing information assets (critical and non-critical) alongside hacker attack types (indiscriminate and targeted). Research indicates that strategic hackers may not always focus exclusively on critical assets; instead, when hackers have a conservative risk preference and employ targeted attacks, they are likely to target non-critical assets. As non-critical assets become increasingly insignificant, firms may intensify protection of critical assets. However, when firms are relatively conservative and face indiscriminate attacks, they also bolster protection of non-critical assets. Further investigation reveals that when a firm’s risk preference shifts from conservative to aggressive, and if there are no constraints on the security budget, efforts to protect non-critical assets under indiscriminate attacks exhibit a positive “U” shape. Conversely, with budget constraints, efforts to protect non-critical assets decline continuously. Additionally, without budget constraints, regardless of whether the firm’s risk preference is conservative or aggressive, the probability of attack on information assets is equal under both attack types. With budget constraints, if the risk preference is either low or high, the probability of attack on information assets is lower under targeted attacks compared to indiscriminate attacks; however, this conclusion is reversed when the risk preference is moderate. By delineating asset types, attack types, and risk preferences, and considering budget constraints, firms can devise scientifically sound information security management strategies and prioritize defensive measures effectively.
Photovoltaic (PV) modules have been ushered in the “retirement tide”, and recycling retired PV modules is an important means to use resources and environmental protection efficiently. This paper discusses the PV module recycling supply chain composed of PV manufacturers and PV recyclers. It explores the optimal recycling strategy of retired PV modules using the noncooperative-cooperative biform game models. On this basis, with the help of the numerical simulation method, we analyze the influence of consumer recycling price sensitivity coefficient, competition intensity, and other factors on the optimal strategy. The study shows that: 1) The recycling price, recycling volume, enterprise recycling profit, and processing technology level all show a positive relationship with consumer recycling price sensitivity coefficient, government subsidy strength, and recycling value, and they all show an inverse relationship with recycling competition intensity, commissioned processing cost and processing cost. 2) The proportion of R&D costs borne by manufacturers shows an inverse relationship with the consumer recycling price sensitivity factor, the strength of government subsidies, and the recycling value, and a positive relationship with the intensity of recycling competition, commissioned treatment costs, and treatment costs. This study aims to provide a reference for the selection of recycling strategies for retired PV modules and the formulation of recycling subsidy policies.
Influential users in online social networks are those who significantly impact information dissemination and opinion interactions, and who can greatly influence public opinion on some topics. Reasonably and effectively identifying such users and their influence contributes to managing information dissemination and network governance. To identify users influence in social networks and analyze its heterogeneity among different groups, this paper proposes a user heterogeneity and re-weighting network (HRNet), considering the causal relationship between user attributes and user behaviors. Under the causal inference model for user online discussion behavior, HRNet identifies the impact of user attributes on behavior variables, and adjusts the combined latent representation of covariates to obtain precise sample reweighing and accurate estimation of causal effects. In the case study, the dose-response functions of reply activity and post activity on content participation influence are shown. Using individual causal effects, this paper estimates each user’s influence, identifies high-influence users, and analyzes the heterogeneity of reply and post activity among different groups, thus distinguishing specific user groups.
Improving intergenerational mobility in education is crucial for alleviating inequality in educational opportunities and accelerating the urban-rural integration. This article analyzes the urban-rural differences in intergenerational upward mobility of education in China based on CFPS2018 and CMDS2017 data, and decomposes it into urban exposure effects and urban-rural sorting effects through group comparison method. It identifies the net effect of urban-rural educational environment on intergenerational mobility of education. Furthermore, the endogeneity is alleviated by displacement shock method. The study finds that among the urban-rural outcome differences of intergenerational upward mobility of education across the entire sample, the urban exposure effect accounts for one-third, and sorting effect accounts for two-thirds; the urban-rural difference in the intergenerational upward mobility of education shows an overall downward trend, among which the offspring born in the 1980s and 1990s experienced a sharp decline in the urban exposure effect and a significant increase in the sorting effect. The main reason for the urban-rural outcome difference is gradually shifting from the gap in the educational environment to the ability endowments differences of individuals. Further tests based on the intensity DID show that the reform of the “new mechanism” of compulsory education is the key channel for the decline of the urban exposure effect during the period.
Based on the data of A-share listed companies from 2010 to 2022, the article examines the impact of digital infrastructure construction on supply chain resilience using an asymptotic double-difference model by considering the Broadband China pilot policy as an exogenous shock to digital infrastructure construction. The study finds that digital infrastructure development significantly enhances supply chain resilience, with reduced credit mismatch, increased risk-taking level and improved inventory turnover efficiency being the channels through which digital infrastructure development enhances supply chain resilience. Heterogeneity analysis suggests that the promotion effect of digital infrastructure construction on supply chain resilience is more significant in growing and maturing firms and competitive industries, and when the city is far away from neighboring prefectures and provincial capitals, the promotion effect of digital infrastructure construction on its supply chain resilience is more significant. Further research finds that digital infrastructure construction only has a positive spillover effect on supply chain resilience in cities with neighboring cities as pilot cities, and the spillover effect decreases with increasing geographic distance.
The high-tech industry serves as a crucial engine for China’s transformational development, with its innovation serving as a pivotal driver for economic growth and enhanced competitiveness. Scientifically and accurately assessing the innovation efficiency of the high-tech industry and analyzing the pathways to enhance efficiency hold significant positive implications for promoting high-quality development within this sector. This study constructs a dynamic network DEA with global weight and explores the paths that drive the improvement of innovation efficiency in high-tech industries using fuzzy set qualitative comparative analysis (fsQCA). This study finds that: 1) The overall innovation efficiency of high-tech industries in China is relatively low, with imbalanced innovation development among provincial regions; 2) there are two configuration paths for improving the innovation efficiency of high-tech industries; that is, digital-driven type, industry-university-research-market-environment synergy type. Finally, according to the efficiency evaluation results and the efficiency improvement path, some policy suggestions are given: 1) Changing the orientation of innovation policy from supply-oriented innovation policy to demand-oriented innovation policy; 2) enhancing technological development, promoting subject cooperation, and optimizing the innovation environment.
Transnational patent licensing has become an important part of international trade of intellectual property rights, and the licensing contract is the key guarantee for transnational patent licensing cooperation. The optimal timing and contracts design of transnational patent licensing are studied through modeling analysis. Two co-opetition games models are constructed for the pre-licensing and lag-licensing scenarios, in which foreign firm engages in Bertrand game and bargaining-based transnational patent licensing cooperation with domestic firm. The results of this paper show that, both pre-licensing and lag-licensing may be chosen by the foreign firm, depending on the degree of competitive product differentiation. However, regardless of the timing of the licensing, the two-part tariff licensing contract is optimal strategy, which facilitates a larger share of profits for foreign firm. The timing of the licensing of foreign firm is also favorable to domestic firm, where pre-licensing can lead to price competition of differentiated products and increase the inequality of profit distribution among firms. In addition, regardless of other conditions, the stronger bargaining power benefits both foreign and domestic firms. When the tariff rate is low, the lag-licensing is conducive to reducing the tax burden of foreign firm and weakening the social welfare effect of transnational patent licensing. Finally, we expand the game model of transnational patent licensing cooperation, and reveal the value of foreign firm’s market entry on licensing contract design and licensing timing.
In the context of China’s transition to a low-carbon economy, the “dual credit” policy stands as the most significant regulatory measure for the NEV industry. Technological innovation is crucial for NEV enterprises to achieve low-carbon upgrades and transition. It remains to be explored whether the implementation of the dual credit policy can effectively incentivize technological innovation among NEV enterprises, which vary in supply chain location, region, asset size, and ownership structure. To address this, the paper combines differences-in-differences model and propensity score matching analysis to thoroughly investigate the impact of the dual credit policy on the technological innovation of heterogeneous enterprises. The results show that the policy can effectively incentivize the technological innovation of all types of NEV enterprises, with significant variations in its effects on heterogeneous enterprises. Specifically, the “dual credit” policy is more likely to stimulate innovation output in downstream enterprises, central region enterprises, small and medium-sized enterprises and state-owned enterprises. Additionally, it favors innovation input of upstream enterprises, western region enterprises, large enterprises and state-owned enterprises. However, the policy’s influence has not significantly enhanced innovation resource input for downstream enterprises. Moreover, enterprises in less economically developed regions exhibit a “high input, low output” phenomenon, characterized by increased competitive pressure and market risk.
This paper investigates the optimal pricing strategies for asymmetric and competitive manufacturers (an innovative manufacturer and a regular manufacturer). Each manufacturer has two pricing strategies, uniform pricing and differentiated pricing, thus constituting four pricing combination models: Uniform pricing for both manufacturers, differentiated pricing for one manufacturer, and differentiated pricing for both manufacturers. By analyzing the prices, innovation levels and the profits under optimality, the following conclusions are obtained. First of all, when manufacturers adopt differentiated pricing strategies, they charge higher prices to old consumers, i.e., “killing the regulars”, to poach market. This strategy further incentivizes manufacturers to improve the level of innovation, resulting in the highest level of innovation. In terms of equilibrium, there are two Nash equilibriums, i.e., both manufacturers adopt uniform pricing strategy and both manufacturers adopt differentiated pricing strategy, in which the former is Pareto-optimal and risk-advantageous. This illustrates that the differentiated pricing based on consumer purchase history does not benefit either side. Therefore, the better strategy for competing manufacturers is to adopt “treating all equally” rather than “killing the regulars”. Furthermore, differentiated pricing strategy increases consumer surplus, but its impact on social welfare is related to the innovation ability.
The persistent issue of carbon emissions from enterprises seriously impacts the effective implementation of the “carbon neutrality” target and hinders the upgrading of industries. Against the backdrop of digital empowerment and government support, the potential effect of public environmental participation in promoting the collaborative governance for pollution and carbon emission of enterprises has become increasingly obvious. Utilizing the data from A-share listed companies between 2011 and 2021, we measured the synergy index for pollution and carbon reduction of enterprises; based on this, through multivariate econometric models to empirically test the synergy of public environmental participation. The results demonstrate that there are still 51.02% of enterprises with different degrees of maladjustment, but the coordination between pollution reduction and carbon reduction has been significantly improved by public environmental participation. This stems from four internal responding mechanisms: The disclosure effect, the source reduction effect, the innovation effect and the investment effect. Moreover, the synergy in emission reduction resulting from public environmental participation is not only strengthened by the empowerment of digital technology, but it can create a powerful combination with governmental ecological regulation that synergistically promotes the collaborative governance in pollution and carbon emission. This effect is particularly pronounced when public participation by the approach of environmental organizations. Further analyses reveal the asynchronous between dust removal, carbon and oxygen control and carbon reduction are the primary factors of the maladjustment of pollution and carbon reduction. However, public environmental participation has been demonstrated to possess notable advantages in enhancing the synergy between dust removal and carbon reduction, which is a pivotal force in promoting the collaborative governance for pollution and carbon emission. And the advantageous impact of public environmental participation is more significant for companies with high media attention, strong political connections, those belonging to highly polluting and energy-intensive industries, and those located in regions facing high carbon emission pressures. Delving into the motivational dynamics, we find that under the influence of public environmental participation, enterprises may sacrifice short-term profits to pursue long-term value growth, thereby actively engaging in collaborative governance for pollution and carbon emission. Fundamentally, this research not only provides new empirical evidence supporting the incentive effect of public environmental participation, but also offers unique insights into maximizing this effect, which holds substantial meaning for promoting the green, low-carbon, high-quality development of industries.
This paper uses extreme industry returns in an unrelated industry of institutional investors’ portfolios as the exogenous shock that shifts institutional investors’ attention away from a firm, and investigates the effect of institutional investors’ distraction on market feedback. Analyzing data from A-share listed companies spanning from 2007 to 2023, this study finds that institutional investors’ distraction reduces the firm’s investment sensitivity to stock prices, indicating distracted institutional investors weaken market feedback effect and curtail managers learning from stock prices. The results show the following mechanisms through which institutional investors’distraction affects managerial learning: Reduced incorporation of traders’ private information into stock prices and increased difficulty in interpreting stock price movements. Additional evidence suggests the distraction of pressure-resistant institutional investors is more likely to lead to the decline of investment-price sensitivity. Lower sensitivity is concentrated in firms whose managers have a stronger willingness and demand to glean external information from stock prices. Institutional investors’ distraction decreases the learning of information about growth opportunities, industry factors, and macroeconomic factors. Besides, the results demonstrate that the distraction of institutional investors has a substantial negative impact on firms’ investment efficiency.
Promoting the synergistic effect of the financial systemic liquidity spiral in the financial system and the development of the real economy is of great significance for improving the financial efficiency, innovation ability, and stability of the financial system. This paper constructs a quadrilateral evolutionary game model “real enterprise-commercial bank-central bank-government”, and applies simulation methods to study the path mechanisms and enhancement strategies of financial systemic liquidity spiral enpowering real economic development. We employ computational and experimental methods with real data and has achieved following conclusions: 1)The quadrilateral evolutionary game system eventually reaches an equilibrium state (self-management, active supervision, provision of liquidity, government supervision) after the quadrilateral game players continuously adjust their own strategies, and the system evolution speed is significantly increased when the initial probability of choosing this strategy combination is relatively high. 2) Quadrilateral game players will have an important influence on the financial systemic liquidity spiral. Among them, the operation strategy and investment decision of the real enterprise will determine whether it abuses liquidity, and they will play a decisive role in the direction of the financial systemic liquidity spiral. Commercial banks play a dual role in the systemic liquidity spiral, which can drive the financial systemic liquidity spiral up or down in terms of the liquidity of real enterprises, and they are the recipients of central bank liquidity and subject to its macro-control. If they choose an active supervision strategy, they can effectively avoid the liquidity loss caused by the risk management activities of real enterprises, which benefits financial systemic liquidity spiral up. The central bank maintains the stability of the financial system and effectively control the speed and direction of the financial systemic liquidity spiral. Among them, the central bank will tend to provide more liquidity to commercial banks with active supervision to maximize social welfare, which ensures the financial systemic liquidity spiral up and promotes the stable and healthy development of real enterprises, forming a kind of principal-agent relationship. 3) The government plays an indispensable role in the quadrilateral evolutionary game system. The government can exert a decisive influence on the operation strategy and investment decisions of real enterprises management, promote the choice of real enterprises self-management strategy, and is an essential link in the formation of a financial systemic liquidity spiral up that can ensure the healthy and rapid development of the real economy.
To explore the impact of managerial affectivity on corporate low-carbon strategy, we take 216 US-listed companies reported by the Carbon Disclosure Project (CDP) from 2011 to 2019 as a sample, theoretically explore and empirically identify the causal relationship between managerial affectivity and corporate low-carbon strategy. The results show that there is an inverted U-shaped relationship between managerial affectivity and corporate low-carbon strategy. The findings remain valid after a series of robustness tests. Operational efficiency moderates the inverted U-shaped by weakening the negative mechanism of affectivity and moving the inflection point to the right and upwards. Artificial intelligence moves the inflection point downward to the right and flattens the inverted U-shaped relationship by weakening the positive and negative mechanisms of emotion. Our findings provide important managerial insights for governments and firms.
The increasing frequency of extreme weather events caused by global warming poses a serious challenge to the high-quality development of the economy. As a key technological vehicle for intelligent manufacturing systems, the rise of industrial robots presents new opportunities for firms to address climate challenges. Drawing on the dynamic capabilities theory, this study examines the impact of climate risk on industrial robot application, based on data from listed manufacturing firms on the Shanghai and Shenzhen A-shares from 2016 to 2023. The findings reveal that: 1) Climate risk promotes industrial robot application, and this conclusion remains valid after a series of robustness tests. 2) In terms of the mechanism of influence, climate risk drives industrial robot application through the factor substitution effect and the strategic option effect. 3) Heterogeneity analysis indicates that the positive impact of climate risk on industrial robot application is more pronounced in private firms, large-scale enterprises, enterprises in coastal regions, enterprises with high carbon emission intensity, and highly competitive industries. This research provides micro-level evidence for effectively responding to climate risks and promoting the intelligent transformation of production processes, while also offering valuable insights for advancing the strategic layout of intelligent manufacturing and achieving high-quality economic development.
Complex systems in the real world often exhibit multiplicity, where a set of entities is connected through various types of relationships. Therefore, modeling with multiplex networks aligns more closely with reality. Multiplex network embedding aims to map the nodes within the network to low-dimensional dense vector representations while preserving the multilayer structural information as much as possible. However, existing methods ignore the structural role information, resulting in the learned node representation being unable to effectively preserve the original structural characteristics. To address this issue, this paper proposes three structural role-based multiplex network embedding methods (named RMNE) from different perspectives, and systematically investigates the enhancing effect of structural role information on multiplex network embedding methods. Specifically, different from the traditional random walk based on node IDs, the first method (RMNE-1) proposes a role-based random walk in multiplex networks, which performs the walking strategy at the level of structural roles of nodes, and then learns node embeddings based on structural roles. The second method (RMNE-2) incorporates structural role information into a unified optimization framework based on existing research, to learn node embeddings enhanced by structural roles. Furthermore, the semi-supervised variant(RMNE-2
