我国行业投入产出网络结构与资产收益率关系研究

Research on the Relationship Between Inter-industry Input-output Network Structure and Industry Capital Return

基于复杂网络视角,本文利用2012年、2017年、2018年中国投入产出表数据,以完全消耗矩阵、完全分配矩阵、关联度矩阵,分别构建行业间消耗、分配以及综合关联网络,分别计算各行业消耗入度、分配出度、关联入度、关联出度等网络中心度指标,定量分析我国国民经济各行业在经济复杂系统的重要性。同时,利用资本资产定价模型对申万二级行业收益率逐时段回归,以统计指标R方测度各行业在资本市场中承担的系统性风险。基于多指标的多年份动态分析,结果显示,关联入度作为行业中心度表征指标,对于衡量行业在经济系统的重要性更合理,且对于区分行业的系统性风险暴露程度更有效。R方与行业中心度的相关系数显著为正,分年份看数值分别为0.38、0.41、0.28,说明中心度越高的行业在资本市场中暴露的系统性风险更高。

Based on the perspective of complex network,this research uses the data of China’s input-output table in 2012,2017,and 2018 to construct the demand,supply and correlation networks among industries with complete consumption matrix,complete distribution matrix,and correlation matrix,respectively.Calculate the network centrality indicators such as consumption in-degree,distribution out-degree,correlation in-degree,and correlation out-degree of various industries,and quantitatively analyze the importance of various industries in China’s national economic system.At the same time,the capital asset pricing model is used to regress the return rate of Shenwan’s secondary industry period by period,and the statistical index R square is introduced to measure the systematic risks assumed by various industries in the capital market.The multi-indicator-based dynamic analysis results show that,as an indicator of industry centrality,correlation in-degree is more reasonable for measuring the importance of industries in the economic system,and more effective for distinguishing the degree of systemic risk exposure of industries.The Pearson correlation coefficient between R square and industry centrality is significantly positive,and the values in three period are 0.38,0.41,and 0.28,indicating that industries with higher centrality are exposed to higher systematic risks in the capital market.