基于Copula函数的CPI与PPI相关性分析

Dependence Analysis of CPI and PPI Based on Copula Function

摘要: CPI与PPI是价格体系中的两大核心指标，两者之间相关关系的度量及分析，有利于定量刻画经济发展趋势，为政府制定经济政策提供依据和建议。考虑到CPI与PPI数据的本质非线性关系，选取表现能力更强的Copula函数来研究CPI与PPI之间的相依性。采用2011年1月至2016年12月全国月度同比CPI与PPI数据进行相关性研究，结果表明Gaussian Copula在刻画秩相关性上效果较好，而在尾部相关性以及与原始数据的拟合程度上，Gumble Copula要优于其他Copula。

Abstract: CPI and PPI are core indexes in price systems, measure and analysis of the dependence between CPI and PPI can help us to better understand the economic development trend, which provides a reliable basis and suggestion for the government in making economic policy. Considering the essential nonlinearity of CPI and PPI, copula function with perfect performance is selected to study the dependence between CPI and PPI in this paper. Using the national monthly CPI and PPI data from January 2015 to December 2016 to research dependence, the results show that Gaussian Copula is better in depiction of rank correlation, and Gumble Copula is superior to other Copula functions in tail correlation and degree of fit with original data.