Property rights system reform is a crucial means to stimulate rural development vitality and facilitate the strategy of rural revitalization. However, there is still insufficient empirical evidence regarding whether property rights delineation can reduce farming households’ precautionary savings and thereby boost their consumption. To address this, this study takes China’s new round of rural land rights confirmation (RLRC) as an example. It constructs a quasi-natural experiment based on data from the 2013 and 2015 China Household Finance Survey (CHFS) and the 2018 China Labor-force Dynamics Survey (CLDS). By comprehensively employing the Difference-in-Differences (DID) method, the Instrumental variable (IV) method, and the Mediation effect model, this study empirically investigates the causal relationship between RLRC and farming households’ consumption, with a particular focus on the mediating effect of precautionary savings. This study finds that the RLRC promotes the growth of farming households’ consumption, and this conclusion remains valid after undergoing various robustness tests. Mechanism analysis reveals that RLRC boosts farming households’ consumption by reducing their precautionary savings, and this mechanism remains robust and effective even after the implementation of the “separation of three rights” reform of rural land and the pilot programs for the mortgage loans of rural land management rights. Heterogeneity analysis indicates that for households with younger household heads and those possessing richer human capital and social capital, the stimulating effect of RLRC on their consumption is more pronounced. Further analysis shows that RLRC primarily promotes subsistence consumption among farming households, rather than developmental consumption, while also having a positive effect on their income improvement. This study not only provides a new explanation for the paradoxical phenomenon of “farming households having money but being reluctant to spend it” but also offers new empirical evidence for government departments to optimize or formulate consumption stimulation policies.
Global climate change and extreme precipitation events pose increasing threats to agricultural systems worldwide. However, existing studies often fail to integrate the synergistic effects of precipitation, neglect the role of transportation infrastructure, and rely on data with coarse spatial resolution. To address these gaps, this study develops a “climate–cropland–transportation” framework. By integrating high-resolution hourly precipitation data (0.1˚) and 30 m cropland distribution data, we quantified the thresholds, frequency, and intensity of extreme precipitation to construct a Cropland Extreme Precipitation Exposure (CEPE) index using the entropy weighting method. Four transportation accessibility scenarios—considering the presence or absence of expressways connecting prefecture- and county-level cities—were simulated to examine their spatial coupling with the CEPE. The results reveal pronounced spatial heterogeneity in CEPE across the study area. Rural population density and industrial structure exert nonlinear regulatory effects on cropland exposure to extreme precipitation. Extreme precipitation intensity peaks at moderate population levels, while areas where the primary industry accounts for 30–40% of the economy constitute a critical buffering zone. Geographically, croplands in southwest China remain highly threatened by extreme precipitation, whereas the exposure fluctuates considerably across the Huang–Huai–Hai Plain and the Northeast China Plain. Uneven transportation accessibility further amplifies these exposures, creating vulnerability traps in regions with high exposure but low accessibility, such as the Northeast China Plain. These findings offer valuable insights for developing targeted agricultural and infrastructure strategies to strengthen agricultural resilience under climate extremes.
Cross-border supply chain disruptions are increasingly posing significant challenges to global operations. However, their impact on firm resilience, particularly for small and medium-sized enterprises (SMEs), has been insufficiently investigated. This study examines how SMEs react to cross-border supply chain disruptions. Drawing on a panel sample of 10,501 firm-quarter observations of SMEs listed on China’s ChiNext market, we find that cross-border supply chain disruptions negatively affect SMEs’ resilience. Furthermore, this effect is significantly mitigated when SMEs effectively orchestrate three key resources: alternative suppliers (relational resources), governmental financial support (financial resources), and digital technology (technological resources). Our study advances logistics and transportation literature by uncovering the specific effects of cross-border supply chain disruptions on SMEs’ resilience. Moreover, our study extends resource orchestration theory into the context of cross-border supply chain disruption by showing that SMEs build resilience through orchestrating relational, financial, and technological resources.
Graph contrastive learning has been widely deployed in graph anomaly detection (GAD) due to its powerful ability to extract discriminative information. However, existing methods fail to capture deviated consistencies of anomaly characteristics from graph topologies and overlook the influences of the quality of contrastive views, obstructing effective consistency learning. To this end, we propose a Multi-scale Asymmetric Contrastive Graph Anomaly Detector (MAC-GAD). MAC-GAD overcomes the drawbacks of traditional symmetric contrastive learning and develops a multi-level dominant view generation mechanism to enhance the contrastive learning performance. To capture comprehensive consistency of anomaly nodes, based on the dominant view, MAC-GAD introduces an asymmetric contrastive learning paradigm for local and global consistency. Theoretical analysis further demonstrates the superiority of asymmetric contrastive learning. Comparison experiments on six public datasets (Cora, Citeseer, Weibo, T-Social, T-Finance, and Pubmed) demonstrate the superiority of MAC-GAD over state-of-the-art baselines, showing average improvement of 1.81% in AUROC and 4.21% in AUPRC across five out of the six datasets except for Pubmed. Extensive validations, such as ablation, sensitivity, and robustness, further verify the advancement of MAC-GAD. Code is available in https://github.com/shaieesss/MAC-GAD.
Accurate and consistent evaluation is crucial for decision-making across numerous fields, yet it remains challenging due to inherent subjectivity, variability, and scale. Large language models (LLMs) have achieved remarkable success, leading to "LLM-as-a-judge," where LLMs serve as evaluators for complex tasks. With their ability to process diverse data types and provide scalable assessments, LLMs present a compelling alternative to traditional expert-driven evaluations. However, ensuring the reliability of LLM-as-a-judge systems remains a significant challenge requiring careful design and standardization. This paper provides a comprehensive survey of LLM-as-a-judge, offering a formal definition and detailed classification while addressing the core question of how to build reliable LLM-as-a-judge systems. We explore strategies to enhance reliability, including improving consistency, mitigating biases, and adapting to diverse scenarios. We propose methodologies for evaluating reliability, supported by a novel benchmark. To advance development and deployment, we discuss practical applications, challenges, and future directions. Our contributions span multiple levels: we establish conceptual boundaries, reorganize fragmented literature into a unified framework, and propose a reliability-oriented benchmark. We articulate a forward-looking research agenda, offering theoretical foundations and practical guidance for constructing reliable and trustworthy LLM-as-a-judge systems.
