摘要
金融状况指数(FCI)是金融指标的线性组合,它不但能够反映当前一个国家的金融现状,同时可以作为货币政策参考指标。近年来,随着金融市场的发展,FCI的编制受到了学者和相关机构的广泛关注。本文将支持向量回归算法作为权重设定方法,使用2006年到2011年的月度数据,选取大量金融指标进行中国金融状况指数构建。结果表明,本文所构建的金融状况指数对通货膨胀指标的变化具有一定的领先作用,在通过样本内检验同时,也在样本外检验中获得了比传统计量方法(VAR脉冲响应分析)更为准确和稳定的结果。
Financial condition index is a linear combination of financial indicators,and it is not only to reflect the current financial situation of a country,but also can be used as the reference index of monetary policy. In recent years,with the development of financial markets,FCI has received widespread attention from scholars and relevant institutions. This paper takes use of support vector regression algorithm innovatively as the weight setting method firstly,by using monthly data from 2006 to 2011 and selecting a large number of financial indicators to establish China financial condition index. The results show that,the constructed financial conditions index has a leading period on the inflation index,not only pass through the in sample test,but also get more accurate and stable results in the out of sample test than traditional method( VAR impulse response analysis).
出处
《管理评论》
CSSCI
北大核心
2014年第5期3-11,共9页
Management Review
基金
国家自然科学基金项目(71071151
70921061
71110107026)
中国科学院研究生科技创新与社会实践资助专项资助广义虚拟经济专项(GX2011-1001(Z))