《保险研究》20190502-《巨灾风险放大及其影响因素测度》(姚领、谢家智、车四方)

[中图分类号]F840 [文献标识码]A [文章编号]1004-3306(2019)05-0031-13 DOI:10.13497/j.cnki.is.2019.05.002

资源价格:30积分

  • 内容介绍

[摘   要]巨灾风险放大是现代风险社会的显著特征。文章基于自然灾害学与灾害社会学理论融合视角,对巨灾风险放大机理及其影响因素进行了分析,并采取随机权神经网络方法,以中国地震灾害为例,对巨灾风险放大影响因素进行了测度。研究表明:巨灾风险放大是由自然与社会双重因素耦合而成的灾情“加重性”、“脱域性”、“加快性”和“延时性”复杂现象,巨灾风险放大程度由致灾因子破坏力、承灾系统脆弱性、承灾系统抗逆力和风险沟通有效度共同决定,且越来越取决于社会性因素。在巨灾风险管理实践中,既要合理避免遭遇破坏力强的灾害,降低风险的自然放大,也要从降低承灾系统脆弱性、提高承灾系统抗逆力、做好风险沟通的维度,降低风险的社会放大。

[关键词]巨灾风险放大;影响因素测度;随机权神经网络

[基金项目]重庆市重点人文社科基地项目(18SKB021);国家社科基金项目(12AGL008)。

[作者简介]姚领,西南大学经济管理学院博士研究生、高级经济师;谢家智,西南大学经济管理学院教授、博士生导师;车四方,西南大学经济管理学院博士研究生。


Catastrophe Risk Amplification and Its Influencing Factors Measurement

YAO Ling,XIE Jia-zhi,CHE Si-fang

Abstract:Catastrophe risk amplification is a prominent feature of modern risk society. This paper analyzed the mechanism of catastrophe risk amplification and its influencing factors based on the fusion of natural disaster and disaster sociology theories. Taking the earthquake disaster of China as an example,the influencing factors of catastrophe risk amplification were measured by using the random weights neural network method. The research showed that the catastrophe risk amplification was a complex phenomenon of “Aggravating”,“Disembodying”,“Accelerating” and “Delaying” caused by the combination of natural and social factors,and the amplification of catastrophe risk was determined by the destructive forces of disaster-causing factors,the vulnerability of the disaster-bearing system,the resilience of the disaster-bearing system and the effectiveness of communication and was increasingly dependent on social factors. Catastrophe risk management practices revealed that,it was necessary to not only avoid destructive disasters and reduce the natural amplification of risks,but also reduce the social amplification of risks through reducing the vulnerability of the disaster-bearing system,improving the resilience of the disaster-bearing system,and making good communication of risks.

Key words:catastrophe risk amplification;measurement of influencing factors;Neural Networks with Random Weights