As China's bond market expands rapidly, the complexity of financial guarantee networks and their implications for credit risk have become critical issues in both academic research and financial practice. Utilizing micro-level data from China's credit bond market spanning 2014 to 2024, this study constructs a multi-layer network incorporating bonds, guarantors, and issuing firms to empirically examine the impact of guarantor network centrality on bond credit spreads. The results reveal a significant U-shaped relationship: moderate centrality reduces spreads by bolstering market confidence, whereas excessive centrality increases them due to heightened systemic risk. Mechanism analyses identify systemic risk and information asymmetry as key mediating channels through which centrality affects credit risk premiums. Heterogeneity tests indicate that this U-shaped pattern is more pronounced among state-owned guarantors, real estate firms, and high-risk clusters within the network. Furthermore, both cross-layer connectivity within the multi-layer structure and regional financial development levels significantly moderate the centrality-spread relationship. These findings offer a structural perspective on credit risk pricing in emerging markets and provide valuable policy insights for credit rating system design, guarantee regulation, and systemic risk prevention. International investors could also leverage these findings to better assess systemic risk in interconnected financial markets across emerging economies.

This paper analyzes the impact of risk disclosures in bond prospectuses on the credit risk premium in the Chinese corporate bond market through six textual characteristics comprehensively. In the empirical analysis, the collected 5199 bond prospectuses and structured data concerning control variables from 2006 to 2021 are used to perform the fixed effect regression analysis. The results show that fewer Words, less Boilerplate, higher Fog Index, more HardInfoMix, more Redundancy, and higher Specificity of risk disclosures in bond prospectuses will lead to a higher credit risk premium. Further tests demonstrate that ceteris paribus, the negative impact of Words and Boilerplate will be strengthened by implicit government guarantees carried by a state-owned enterprise but be weakened by better corporate business performance. However, ceteris paribus, positive effects of the Fog Index, HardInfoMix, Redundancy, and Specificity will be weakened when the bond issuer is state-owned but be strengthened by better corporate business performance.