基于DAG方法的经济变量因果关系研究

A Causal Analysis on Economic Variables Based on the Directed Acyclic Graphs

经济变量因果关系研究一直是学术界讨论的焦点问题,以往的研究主要是通过Granger理论和方差分解分析变量间的动态因果关系,尽管Granger因果理论能够表明变量间短期因果关系动态,但却不能解释经济变量间的同期因果关系。在介绍并利用近年来新发展的DAG方法基本理论和算法的基础上,对浙江省宏观经济变量的因果关系进行研究。实证研究表明:投资和出口既是浙江经济增长的同期原因,又是经济增长的短期和长期原因,而消费只是经济增长的长期原因。

The basic theory and algorithm of recently developed method directed acyclic graphs is introduced and applied to the analysis on the causal relationship among macro economic variables in Zhejiang province. In contrast to prior analyses, DAG techniques can be used to provide over- identifying restrictions on innovations from structural vector autoregressive models and to examine both contemporaneous and dynamic causal structures of above - mentioned variables. Empirical studie show that capital information and exports cause economic growth, and consumption is only the long reason for economic growth.