Granger因果关系检验应用中的若干问题

Some Problems in the Application of Granger Causality Test

分析了实践中应用Granger因果关系检验存在的一些问题，如信息遗漏，变量变换改变因果关系的性质、变量单整性对检验程序的影响以及检验模型的选择等，并提出在线性投影上有初步证据的因果概念。进而，应用单整变量之间Granger因果关系的一般检验程序对1978～2013年我国货币供给量与价格水平、产出之间的Granger因果性重新进行检验。

This paper analyses some problems in the application of Granger causality test, such as missing information, the transformations of variables changing causality a- mong them, the effect of integration of variables on the test procedure and the choice of ag- gression models, and proposes the concept of prima facie causality in linear projection. Further, we retest the causality among the Chinese money supply, price level and output from 1978 to 2013 u- tilizing the general test procedure.