基于混合Copula函数的价格指数实证研究

Empirical study on the price indices based on mixed Copula functions

消费者价格指数(CPI)和生产者价格指数(PPI)是衡量物价水平的两大重要指标.为研究CPI和PPI间的联系,引入混合Copula函数的概念,结合三种常用的阿基米德Copula函数的加权和分析两者之间的相关性.选取1994年1月1日到2016年12月1日的CPI和PPI数据,求出其对应的对数环比增长率.利用核密度估计模型建立变量的边缘密度函数,并利用EM算法与BFGS算法对模型进行参数估计,最后利用最小欧式距离法对模型进行检验.结果表明,CPI与PPI在经济衰退时期的相关性表现更为紧密.

Consumer price index(CPI)and producer price index(PPI)are two important indices of price level.By introducing the concept of the mixed Copula function model and combining it with the weighted sums of three Archimedes Copula functions,the correlation between CPI and PPI is analyzed.The data are drawn from recorded CPI and PPI from 01/01/1994 to 01/12/2016,the corresponding logarithmic growth rate is calculated.The kernel density estimation model is used to derive the edge density function for the variables,and the parameters of the model are estimated by EM algorithm and BFGS algorithm.Finally,the model is tested by using the minimum Euclidean distance method.The results show that CPI and PPI are more closely related in the economic recession period.