The Mechanism of Cross-Industry Contagion of Industrial Chain Risks from the Perspective of Multi-Layer Networks
The industrial chain consists of a complex network of supply and cooperation relationships,featuring not only vertical linkages across industries but also horizontal collaboration within industries.This multi-level structure propagates risks across upstream and downstream links in response to extreme shocks,affecting the entire production network.Understanding the mechanisms of risk contagion within the industrial chains and quantifying risk exposure and contribution are essential for mitigating systemic risks.This paper develops a multi-layered industrial chain network model based on transactional and equity relationships.Integrating firm heterogeneity and the SIRS contagion model,it establishes a framework for analyzing risk contagion across institutions and industries.The findings highlight corporate heterogeneity as a key factor in systemic risks:when credit risk awareness,information transparency,or risk management capabilities are low,an increase in equity and affiliated entities correlates with a rise in defaulting enterprises.Conversely,the improvement in these factors is associated with fewer defaults.The mechanism analysis further reveals that all else equal,increased weighting of equity or transactional ties among enterprises leads to more defaults.Building on this theoretical foundation,this paper applies the TENET model and the SIM-LASSO-CoVaR indicator to construct a time-varying network of risk spillovers and systemic risk metrics.By analyzing both intra-industry and inter-industry dynamics,it examines the nonlinear characteristics of risk contagion in China's industrial chain from 2013 to 2022.The empirical results demonstrate a pronounced cascading effect in inter-industry risk contagion:vertically,upstream industries serve as primary sources of risks,with downstream industries as major recipients of this risk exposure;horizontally,inter-industry risk spillovers exceed those within individual industries.This approach provides a fresh perspective on systemic risk measurement and forecasting through the lens of nonlinear network connections in industrial chains.The main contributions of this paper are as follows.Firstly,incorporating transaction and equity associations as key factors representing industrial trade and M&As,this paper constructs a multi-layered industry chain network structure segmented by industry.Integrating these associations with firm heterogeneity,it establishes a cross-agency and cross-industry dynamic risk contagion model to examine risk transmission mechanisms within the industrial chain structure.Secondly,addressing the nonlinear characteristics of risk contagion and dimensionality challenges in industrial chains,this paper introduces a novel conditional value-at-risk index using a single-index quantile model and the LASSO-CoVaR technique.Combined with the TENET model,it constructs a time-varying tail risk spillover network,providing an empirical analysis of the dynamic features of systemic risk from both intra-industry and inter-industry perspectives.
tail risk networkindustry chainsystemic risknonlinear network association
NETL
NSTL
万方数据
