Scholarlink

<p>In this paper, we propose a method to construct a multi-layer co-jump network based on a multi-layer degree-corrected stochastic block model, aims to infer risk propagation pathways from stock co‑jump patterns. Through the high-frequency price data, we extract the co-jump matrix of stocks during the positive and negative market jumps respectively, and set thresholds to construct a two-layer co-jump network. On this basis, the multi-layer normalized spectral clustering community detection algorithm is used to group stocks and identify stock communities with systematic risk correlation. The empirical results show that, compared with the single-layer co-jump network, the multi-layer co-jump network can capture a more stable and detailed community structure. The investment portfolio constructed according to this structure shows superior risk-adjusted returns, while significantly outperforming the market benchmark. These findings verify the practical value of the multi-layer co-jump network in risk management and asset allocation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-layer Co-jump Network Under Market Asymmetric Jumps

  • 期刊论文

  • 发表日期：

    2026年7月31日

查看全文
  • Luwen Zhang,
  • Wuchen Li,
  • Shuoyan Tang,
  • Li Wang

摘要

In this paper, we propose a method to construct a multi-layer co-jump network based on a multi-layer degree-corrected stochastic block model, aims to infer risk propagation pathways from stock co‑jump patterns. Through the high-frequency price data, we extract the co-jump matrix of stocks during the positive and negative market jumps respectively, and set thresholds to construct a two-layer co-jump network. On this basis, the multi-layer normalized spectral clustering community detection algorithm is used to group stocks and identify stock communities with systematic risk correlation. The empirical results show that, compared with the single-layer co-jump network, the multi-layer co-jump network can capture a more stable and detailed community structure. The investment portfolio constructed according to this structure shows superior risk-adjusted returns, while significantly outperforming the market benchmark. These findings verify the practical value of the multi-layer co-jump network in risk management and asset allocation.