Under the constraints of critical core technologies and the outsourcing of low-end industries, accurately measuring the level of autonomy and control in the manufacturing sector, as well as conducting an in-depth analysis of its underlying causes are essential requirements for China in constructing a modernized industrial system. Based on the industrial linkage perspective, this article utilizes the world input-output table to measure the level of autonomy and controllability in the manufacturing industry from the dimensions of production technology and process. Research shows: At the overall manufacturing level, the level of technological autonomy of developing countries represented by China is lower than that of developed countries, but its increase is more pronounced, whereas developed countries show a “U” trend with a threshold around 2008; The technological controllability level of developing countries is generally higher than that of developed countries. The production process has the characteristics of a large country with a high level of autonomy and controllability. The heterogeneity of China’s subdivided industries shows that labor-intensive industries exhibit dual high characteristics in the autonomy and controllability of production technology and processes. While some resource-intensive and high-tech industries have improved their autonomy in production technology and processes, they still rely heavily on foreign sources, resulting in a situation where “strong manufacturing” and “weak research and development” coexist. Further examination of country heterogeneity reveals that in high-end industries, China’s autonomy and controllability of production technology and processes face challenges not only from developed countries like the United States and Germany due to the“high-end return”, but also from developing countries such as Brazil and India, although this “low-end diversion” mainly occurs in the lower segments of the mid-to-high-end industries rather than traditional low-end industries. In low-end industries, comparing with developed countries where the level of autonomous production technology is high but the level of autonomous and controllable processes is low, leading to industrial “hollowing-out” issues, China’s industrial chain is relatively secure.
Achieving age-appropriate and moderate fertility among the reproductive-age women is one of the key pathways to addressing China’s current low birth rate. Based on the phenomenon of delayed mean age at first birth in China, this paper seeks to answer whether motherhood delay can alleviate the wage penalty faced by mothers with different numbers of children. By constructing a four-period overlapping generations model with a breakdown of reproductive period for theoretical analysis, and using the data of the China General Social Survey (CGSS) from 2006 to 2021 for empirical analysis, this paper finds that delayed age at first birth leads a wage premium, but this premium will fade and ultimately vanish with an increase in the number of children, thereby further intensifying the wage penalty. This conclusion is still hold after a series of robustness analyses. Heterogeneity analysis indicates that the heterogeneous wage effects of delayed motherhood are more significant in households where women have lower years of education than their spouses, and that delaying first childbirth is not a good choice for women from relatively poorer economic backgrounds. Mechanism analysis reveals that delaying the age at first birth increases the likelihood of women will currently be engaged in non-agricultural work and choose to work in large organizations or enterprises. However, an increase in the number of children prompts mothers to seek more stable and flexible careers. This paper provides evidence for accelerating the construction of a fertility support policy system, and emphasizing the need to address the employment difficulties and wage challenges faced by women of different childbearing ages, especially those who may choose to have multiple children, so as to facilitate the implementation of the three-child policy.
Strengthening supervision coordination is an important direction for our current reform of financial supervision system. Using the establishment of the China Banking and Insurance Regulatory Commission as a plausible exogenous shock to financial supervision coordination, this article examines the effect of financial supervision coordination on bank default risk using Chinese A-share listed banking and capital market service companies from 2014Q1–2022Q4. We find that financial supervision coordination significantly reduces the bank default risk. After several robustness tests, the conclusion remains robust. Further research finds that improving the external regulatory environment and regulating institutions’ internal behavior are mechanisms that work. In addition, the risk suppression effect of financial regulatory coordination is more significant for banks with lower regional financial regulatory intensity and stronger political connections.
Facing the increasing repayment pressure of local governments, the central government has advocated establishing a government debt service reserve fund system, which has been implemented in most provinces. Utilizing the quasi-natural experimental environment created by different implementation timings across regions, this study employs the difference-in-differences (DID) method to examine the system’s impact on local government debt risks. The findings reveal that implementing the debt service reserve fund system significantly reduces local government debt ratios. The potential mechanisms include alleviating debt repayment pressure, curbing debt rollover behaviors, reasonably controlling debt scales, and incentivizing local governments to enhance fiscal capacity. The policy effects are more pronounced in county-level governments, economically underdeveloped regions. However, land finance reliance and implicit debt scales may counteract the policy effectiveness. These conclusions provide theoretical support and policy recommendations for further preventing local government debt risks and optimizing the debt service reserve fund system.
This paper analyzes the impact and mechanisms of central fiscal transfers on the economic resilience of resource-exhausted cities based on path dependence theory. A difference-in-differences model and city panel data are employed for validation. It reveals that central fiscal transfer payments weaken the economic resilience of resource-exhausted cities in some degree, and the conclusion still valid after overcoming the endogeneity problem by using the instrumental variable approach and a series of robustness tests. Mechanism analysis indicates that the suppression of technological innovation, industrial upgrading, and resource optimization are key factors by which central fiscal transfers weaken the economic resilience of resource-exhausted cities. Further research shows that the strict environmental regulatory policies further exacerbate the weakening effect of central fiscal transfers on the economic resilience of resource-exhausted cities. Finally, the results of the spatial effect test show that the central fiscal transfer not only weakens the economic resilience of resource-exhausted cities, but also exerts a significant weakening effect on the economic resilience of neighboring cities. The conclusions of this paper are of great theoretical and practical significance for China to further improve the fiscal transfer mechanism as well as optimize the policy combinations to promote the transformation and development of resource-exhausted cities.
The gig economy, characterized by its flexibility, has altered traditional labor market dynamics, including employment patterns, work styles, and job structures. This study represents worker welfare as the expected surplus, defined as the difference between expected wages and reservation wages. The ability of workers to adjust their labor supply in response to fluctuations in expected wages and reservation wages is crucial for the enhancement of welfare through flexibility. Utilizing micro-level big data from the Shanghai ride-hailing market, this paper classifies different types of ride-hailing drivers and establishes models for reservation wages and high-dimensional labor supply for each type of drivers. By employing high-dimensional MCMC (Markov chain Monte Carlo) methods to estimate reservation wages, the impact of flexibility on worker welfare is quantitatively evaluated. The findings indicate that flexibility significantly enhances labor supply and worker welfare, nearly doubling the welfare compared to fixed working hours. This study represents a cutting-edge practice in labor supply research using big data, providing an empirical framework applicable to the study of other new employment forms in the gig economy.
As urbanization enters a new stage, the strategic value of county areas surrounding major cities is increasingly prominent, with high-level security serving as the foundation and bottom line for achieving high-quality development. This paper studies 143 counties in the hinterland of three major urban agglomerations in eastern China. Using Landscan population spatial distribution data, industrial and commercial enterprise registration data, and micro-level land transaction data, it explores the economic resilience effect of urban spatial structure from the perspectives of core area agglomeration and peripheral area expansion. The study finds: 1) Under the premise of endogeneity analysis and robustness testing, spatial structure decentralization significantly reduces economic resilience. During the study period, spatial structure decentralization suppressed approximately23.6% of the potential for economic resilience improvement. 2) The negative marginal effect of spatial structure decentralization is stronger, especially in counties closer to large cities, with flatter terrain, and during external shocks. A further breakdown reveals that core area agglomeration can enhance economic resilience, while peripheral area expansion has the opposite effect, with the overall effect being the result of the positive and negative offsets. 3) Intrinsically, spatial structure decentralization negatively impacts economic resilience through three channels: inhibiting economic activity diversification, reinforcing dependence on land sales, and reducing fiscal expenditure efficiency. 4) Further discussions reveal that although online virtual agglomeration has a bidirectional impact on economic resilience, it somewhat offsets the agglomeration effect losses of the dispersed physical space, thereby overall exhibiting a positive influence. In conclusion, this study provides decision-making references for optimizing spatial structures to empower economic resilience in the counties surrounding large cities, aiding in the organic unification of “development” and “security” in the urbanization process.
Enhancing the resilience of industrial chains is a fundamental pillar for balancing national security and economic development. This study aims to uncover the systemic formation mechanisms of industrial chain resilience. On one hand, industry-level resilience serves as the foundation of overall industrial chain resilience, with significant heterogeneity across industries due to differences in structural characteristics. On the other hand, the interconnections of resilience between upstream and downstream industries are a key manifestation of systemic resilience, determining the degree of coordination in resistance and recovery during external shocks. Given that electricity consumption data can objectively and promptly reflect changes in industrial production capacity, this paper uses daily electricity consumption data from 28 industries in Fujian Province as an empirical case to measure both resistance resilience and recovery resilience. It further investigates the influencing factors and network spillover effects of industrial resilience. The main findings are as follows: 1) There are significant differences in resistance and recovery resilience across industries, and characteristics such as upstreamness, network centrality, and export dependence can effectively explain these differences. 2) Industrial resilience exhibits strong network-based spillover effects, with core industries in the production network showing more pronounced resilience transmission. 3) The upstream-downstream input-output structure is an important mechanism driving resilience linkages, and the resilience of downstream industries exerts stronger spillover effects on upstream sectors. This highlights the pivotal role of demand-side resilience in maintaining the stability and security of industrial chains. Based on these findings, this paper offers policy recommendations for building a highly resilient industrial and supply chain system.
Social media promotes the dissemination of stock market crisis information, especially the text with emotional color, which is an important driving factor for users to generate information behavior. In order to more comprehensively understand the impact of investor sentiment on their information behavior and decision making, this paper subdivides investor sentiment from the cross-dimension of emotional valence and arousal, introduces the investor concern theory, and uses 429,582 microblog data and 4964 financial data. The PVAR method is used to explore the dynamic relationship between user sentiment in social media and stock market returns through investor information behavior in the context of stock market crisis. It is found that in the stock market crisis, positive emotions have a positive impact on stock market returns when it is delayed for 2 days (dominated by high arousal emotion), and a negative impact on stock market returns when it is delayed for 4 days (dominated by low arousal emotion), and play a negative intermediary role in the forwarding behavior of positive information. Negative emotions negatively affect stock market returns, in which “negative-low arousal” emotions have a direct negative impact on stock market returns, while “negative-high arousal” emotions have a negative impact on stock market returns by promoting the dissemination of negative information. This paper reveals the dynamic influence mechanism of “emotion-information forwarding behavior-stock market” under the stock market crisis situation, which provides the decision support for the stock market participants to deal with the crisis.
Since 2018, a series of such Black Swan events as trade friction, the COVID-19 pandemic, and the Russia-Ukraine conflict have occurred frequently, leading to an unprecedented transformation in China’s economic development. In response to the increasingly complex external environment, the Chinese economy is transitioning towards a new development pattern characterized by the dominance of the domestic circulation and the mutual promotion of domestic and international circulations. In this paper, we utilize data from 73 sub-markets within the Chinese financial market system and employ the block aggregation method to construct a cross-market risk contagion network. Based on risk spillover indicators, we investigate the static and dynamic perspectives of the financial risk contagion effects under the new development pattern in China, and analyze the evolutionary paths of cross-market financial risk contagion in different stages. The empirical results indicate that: 1) Both before and after the proposal of the new development pattern, the stock market is a net transmitter of risk contagion, with the highest risk spillover to the commodity market; the bond market serves as an intermediary for risk contagion, while the foreign exchange market, commodity market, and currency market are net receivers of risk contagion; 2) before and after the proposal of the new development pattern, the internal risk contagion effects within the financial market are significantly higher than the cross-market contagion effects; under the impact of sudden crisis events, the Chinese financial market system transforms internal risk spillover among similar markets into cross-market spillovers, thereby deepening the cross-market financial risk contagion effects; 3) under the new development pattern, new characteristics emerge in China’s cross-market financial risk contagion effects; the net contagion effect from the stock market to the currency market becomes the second highest in the entire net contagion network. Additionally, with the formation of the dual circulation new development pattern, the cross-market risks in China’s financial system are effectively reduced, and economic resilience is significantly enhanced.
The factors affecting the price spread of asset with same intrinsic value among different markets have led to extensive and ongoing discussion in the financial sector. Taking the companies listed in both A-share market and Hong Kong share market at the same time as an example, this paper compares the interpretation ability of four mainstream theories (information asymmetry hypothesis, liquidity difference hypothesis, heterogeneous beliefs hypothesis and risk difference hypothesis) to the stock price spread of asset with same intrinsic value among different markets. By constructing a set of panel data models, this paper separately and jointly tests the interpretation ability of different theoretical hypotheses on the stock price spread of AH shares, tests the comprehensive interpretation ability of different theoretical hypotheses on the AH stock price spread, and forecasts the regression model. The regression results of both single factor model and joint factor model indicate that the information asymmetry, liquidity difference, heterogeneous beliefs and risk difference hypotheses all have interpretation ability to AH stock price spread. Moreover, the institutional reforms of Chinese stock market, represented by the opening of QFII, moderate the relationships between above-mentioned four factors and AH stock price spread, which indicates that the institutional factors also affect the AH stock price spread to a certain extent. With variance decomposition method based on VAR model, this paper verifies that the interpretation ability of different hypotheses to asset price spread is: information asymmetry > liquidity difference > heterogeneous beliefs > risk difference. The results of this study provide new evidence for better understanding the factors affecting the price spread of asset with same intrinsic value among different markets, and is a useful complement to the literatures on price spread of asset with same intrinsic value among different markets.
The impacts of climate change are exerting increasingly pronounced effects on economic activities and production processes. Faced with escalating climate risks, enterprises demonstrate stronger managerial incentives for earnings manipulation. This study aims to examine the effect of climate risk on corporate real earnings management and elucidate the underlying transmission mechanisms. Based on the data from Chinese listed companies covering the period 2011 – 2022, our empirical analysis reveals that climate risk significantly intensifies the degree of real earnings management. We identify three primary channels through which climate risk operates:financing constraints, environmental uncertainty, and market supervision. The heterogeneity analysis further demonstrates that non-state-owned enterprises, firms operating in heavily polluting industries, and those exhibiting high fixed asset growth rates exhibit more pronounced financial manipulation behaviors when they confronted with climate risks. This research deepens our understanding of how climate risk shapes accounting strategies among Chinese firms and offers important policy implications for refining China’s climate change response framework.
Listing live hog futures provides a hedging tool to resist and manage market risks directly. This paper investigates the hedging function of live hog futures by starting with the conversion of hedging mode, analyzing the performance of cross hedging before listing live hog futures, and direct hedging after the listing, aiming to answer the question of are live hog futures an effective hedging tool. According to the results, before the listing of live hog futures, using corn or soybean meal futures to carry out cross-hedging reduces the risk of hog price by 2%, with an excess return of 0.02%; after the listing of hog futures, using live hog futures to carry out direct hedging reduces the risk of hog price by 13%, with an excess return of 0.04%. Overall, the live hog futures are an effective hedging tool, which plays an important role in reducing risk, preserving profit and stabilising production. It is recommended that the government enhance the connectivity between futures and the spot market, the Dalian Commodity Exchange optimize contract rules, and hedgers trade rationally to promote the performance of the hog futures and serve the national strategy of securing supply and stabilizing prices.
The media has a positive effect on the active carbon market, but the media climate sentiment is also very likely to affect the carbon market volatility. This paper constructs a media climate sentiment index for the first time based on news media coverage data using text analysis, and empirically tests the impact and prediction of this index on China’s carbon market volatility using a generalised autoregressive conditional heteroskedastic mixed frequency data (GARCH-MIDAS) model. The empirical results reveal that, media climate sentiment is an influential factor of China’s carbon market volatility and can well explain the long-term component of carbon market volatility. The inclusion of media climate sentiment can improve the prediction accuracy of carbon market volatility, and its prediction ability is effective in both the long and short term. The findings of this paper provide a new theoretical perspective for understanding the mechanism of China’s carbon market volatility, and also provide a reference basis for accurately predicting carbon market volatility.
The problem of cost allocation arises in situations where multiple individuals decide to work together and distribute the common costs in a reasonable manner. However, all individuals expect to pay a small cost to participate in this work, at this time it is easy to free ride, which may lead to the failure of the work. Therefore, this paper proposes a binding third-party arbitration method to solve the cost allocation problem where individuals prefer to pay a small cost. In this paper, an arbitration game model of cost allocation is established to discuss the form of pure strategy Nash equilibrium and its necessary and sufficient conditions. Furthermore, the sufficient and necessary conditions for the equilibrium costs shared by the participants to be equal are obtained. Finally, two methods to improve the arbitration game and then reduce the conflict degree of participants are analyzed. This study shows that adding proposal cost and expanding the number of arbitrators in arbitration can effectively reduce the degree of conflict between participants. The research of this paper provides a theoretical basis for using arbitration to solve the cost allocation problem.
The private label introduction has become an effective way for e-commerce platforms to break through the bottleneck of profit growth. The selection of which product category to introduce is a key strategic decision. This paper establishes a platform-based supply chain composed of the e-commerce platform and the brand manufacturer. Using the Stackelberg-Nash game method, we explore the strategy of private label introduction and product category selection under the product competition. The study finds that regardless of whether the category is centralized or decentralized, the base demand level of private labels and the commission rate are key factors impacting the e-commerce platform’s decision to introduce private labels. Under the centralized category, the introduction of private labels by e-commerce platforms harms brand manufacturers. However, under the decentralized category, introducing private labels can benefit both the platform and the brand manufacturer when the base demand level is high, or when the base demand level of private label is relatively low and the cross-price sensitivity with brand product is low. The e-commerce platforms achieve higher profits when introducing the private label under the centralized category than that under the decentralized category. Conversely, introducing the private label under the decentralized category is more effective in expanding the overall demand of category. Moreover, as the number of products in the initial category increases, both the e-commerce platform and the brand manufacturer can achieve a win-win situation from the introduction of private labels under the centralized category. This work provides constructive managerial insights for e-commerce platforms in selecting the appropriate product categories for introducing private labels.
Supply chain network disruption risk can be propagated to upstream and downstream enterprises through the supply and demand relationship, consequently endangering the overall performance of the supply chain network. In the process of disruption risk propagation, it is often accompanied by risk information dissemination, risk information disclosure, and risk control measures adoption. The propagation scale and outbreak threshold are either directly or indirectly related to these factors. Therefore, studying the dynamic coupling propagation characteristics of risk information dissemination, risk information disclosure, risk control measures adoption and disruption risk propagation is crucial to understanding the underlying mechanisms of disruption risk propagation in supply chain network. In this paper, a three-layer complex network model is constructed to explore the risk disclosure awareness, risk control measures adoption and the direction of risk propagation on disruption risk propagation in supply chain network. Different from previous studies, this paper divides enterprises into enterprises that disclose risk information and enterprises that do not disclose risk information, and the disclosure of risk information affects the information dissemination process; moreover, this paper, unlike the assumption of risk dissemination non-directionality in previous studies, considers the influence of both forward and backward disruption risk propagation on the disruption propagation process. By building a state probability transmission tree, the state transmission equation is established using the microscopic Markov chain method, and the disruption risk outbreak threshold is derived. The results of simulation experiments show that enhanced awareness of risk disclosure can increase the disruption risk outbreak threshold and reduce the scale of disruption risk propagation. In addition, it is essential to distinguish between forward and backward disruption risk propagation because of the differences in their influence on disruption risk propagation.
High quality development and rural revitalization are the dual national strategic goals of China. In this context, the orderly promotion of village prefabricated buildings also needs to be supported by financial subsidies from local governments. In this paper, we find that there exists an interactive strategic between the demand side of the policy (prefabricated building manufacturers and residents) and the supply side of the policy (local government). Therefore, we adopt an evolutionary game approach, set the subsidy rate describing the degree of subsidy, and conduct simulations using actual project data. The results of the study show that: 1) There exists an inverted U-shape to realize the dynamic change of subsidies that enhance firstly and weaken secondly. 2) There exists a policy accessing (exiting) rate that realizes the optimal performance of subsidies.
Metropolitan transportation systems are overwhelmed not only by congestion but also by increasing concerns about the emission problem, especially after the carbon peaking and carbon neutrality strategy was established in China. In this paper, we employ a fuel consumption cost strategy in a bi-modal transportation system with customized bus (CB) and private car modes. We build a bi-modal network equilibrium model under this strategy to analyze the relationship between congestion and emission and study the equilibrium state of a low-carbon and low-congestion urban road transportation system. According to bilateral choices for the CB platform and passengers, we derive the matching conditions for CB mode to screen potential riding requests. Moreover, we make an equivalence transformation and relaxation to the proposed model to facilitate the solutions of Lagrangian relaxation algorithm. Numerical results verify the validity of the proposed model, and reveal the effects of settings of CB mode and the fuel consumption cost strategy on congestion and emission, and analyze effects on the phenomenon of inconsistency between congestion and emission, and examine the impacts of changes in demands on the performance of CB mode in terms of system performance improvement. This study provides the valuable theoretical support for congestion alleviation and emission reduction in urban transportation systems.
Civil aircraft development has developed rapidly under the guidance of the model based systems engineering (MBSE) method. As an important link in the development process of civil aircraft, the complexity of civil aircraft system manufacturing has been increasing. In the process of civil aircraft system manufacturing process highlights the long manufacturing cycle, resource consumption and other practical problems. Aiming at the above problems, firstly, the complex relationship between the manufacturing processes of MBSE civil aircraft system is analysed, and a mathematical optimization model of the manufacturing process of civil aircraft system is established. Secondly, in order to take into account the convergence of the algorithm and the diversity of the solution set, an improved SPEA2 algorithm (strength pareto evolutionary algorithm 2, SPEA2) is designed, and a neighbourhood search method based on the density of the solution region and a fuzzy adaptive genetic operator dynamic adjustment method are proposed. Finally, the airborne flat view display system is analysed as an example to verify the effectiveness of the improved algorithm in terms of convergence and diversity. The validation case shows that the proposed method can effectively improve the efficiency and reduce the resource consumption of civil aircraft system manufacturing, and achieve the optimization of the manufacturing process of MBSE civil aircraft system.
To improve the efficiency of container terminal operations and reduce operating costs, we propose an operational method by using collector trucks and barges to transport containers between different terminals. Firstly, considering the influence of truck queuing and speed regulation on fuel consumption, a mixed-integer second-order conical planning model is established with the goal of minimizing its transportation cost to obtain a land-side truck transportation scheme between distant terminals that meets multiple constraints. Secondly, considering the limitation of barge berth coordination and berthing safety time interval and other factors, a closed-loop scheduling model is established for barge transportation container on the water side of the terminal, and a low-cost waterway transportation scheme is obtained in line with that between closer terminals. Finally, a reinforcement learning algorithm incorporating chaotic search and two-layer transformer encoder is proposed to solve the model characteristics. Simulation examples show that the proposed method and algorithm can effectively obtain a two-stage land-sea cooperative transportation scheme.
Urban rail transit short-term passenger flow prediction is crucial for enhancing service quality. However, existing models often overlook the correlation between passenger flows at different time periods, leading to decreased prediction accuracy. This paper first divides station passenger flows into real-time, daily, and weekly temporal patterns to extract temporal information. Furthermore, it proposes a deep learning-based composite prediction model (Graph-GAN), which integrates graph convolutional neural networks (GCN), causal convolutional layers, attention mechanisms, and generative adversarial networks (GAN). This model effectively extracts spatiotemporal interaction information between different patterns and utilizes GAN for adversarial training to enhance prediction accuracy. Tested on real datasets from the Beijing subway in 2016 and 2018, results demonstrate that Graph-GAN outperforms baseline models, with RMSE values of 34.73 and 32.87, MAE values of 20.47 and 16.85, and WMAPE values of 7.72% and 8.62% respectively for the two datasets. Multi-step prediction experiments show that Graph-GAN accurately forecasts future station passenger flows for multiple time steps, validating the model’s practicality. Sensitivity analyses further confirm the importance of GAN, interaction information between different pattern flows, and spatiotemporal features in improving prediction accuracy. Moreover, GCN effectively captures the network topology of urban rail transit.
Under the complex and variable operating environment and the uncertain external disturbances, guaranteeing the safe and reliable operation of the system has emerged as a core issue in engineering. The majority of existing studies have primarily concentrated on the static environment or the consistent damage effects of shocks, resulting in notable limitations in capturing the specific influences of random shocks on the system in the dynamic environment, and subsequently influencing the accuracy of reliability assessment. Hence, this paper conducts a comprehensive analysis of the influences of both on reliability, considers the differences in the degradation rates of the system under various environments, and categorizes the damage caused by shocks to the system into the safety zone, the random damage zone, the damage zone, and the fatal zone based on the magnitude of the shock. On this basis, a reliability assessment method for the competing failure system considering the effect of zoned shocks under dynamic environment is proposed, and the analytical expression of the system reliability index is provided. Moreover, the correctness of the method is verified based on the Monte-Carlo simulation algorithm. Finally, taking the rotor system as an example, the effectiveness of the proposed method and its potential engineering application scenarios are demonstrated. The result reveals that the proposed method can effectively enhance the accuracy of reliability assessment. Furthermore, as the probability of the shock falling into the random damage zone rises from10% to 50%, the accuracy ascends from 10% improvement to 27.5% improvement, offering substantial support for engineering practice.
The propensity score plays an important role in both treatment effect studies and missing data analyses. Recently, a novel estimator for the propensity score, the covariate balancing propensity score (CBPS), was developed and has been widely adopted by researchers across various disciplines. In this article, we investigate the computation method of CBPS, especially the efficient CBPS (E-CBPS), which is the most popular CBPS method among all CBPS variants. We demonstrate that under a certain set of conditions, the weighting matrix of the E-CBPS is singular; hence, the estimation result is unstable. The set of conditions includes three factors: A propensity score model in the form of the generalized linear model (GLM), a linear transformation function for covariate balancing, and a degenerate true propensity score. The simulations and empirical examples indicate that the singularity problem of the E-CBPS is not just a theoretical possibility but a practical issue in empirical studies. Several strategies to circumvent the singularity problem are investigated.
This paper addresses the challenge of systematic prediction modeling for seasonal time series by integrating multiple system behavior sequences within a unified, bottom-up, nested modeling framework. Initially, we expand the grey action quantity by introducing seasonal time-varying effects and shock-inducing dummy variables, thereby fully exploiting the grey information embedded in complex systems. Subsequently, we construct a discrete systematic grey model with seasonal time-varying effects (DSTSGM
