In his acceptance speech in 2003, Granger himself emphasized the limitations of his citation and the appearance of "many absurd papers". Because its statistics is essentially the prediction of stationary time series data, it is only applicable to the variable prediction of econometrics and cannot be used as a standard to test the true causal relationship.

In the case of time series, the Granger causality between two economic variables X and Y is defined as: if only based on the past information of variables X and Y, the prediction effect of variable Y is better than that of variable Y, that is, variable X helps to explain the future changes of variable Y, then variable X is considered as the Granger cause of variable Y.

A prerequisite for Granger causality test is that the time series must be stationary, otherwise the problem of false regression may occur. Therefore, before Granger causality test, we should first test the stationarity of each index time series. Extended Dickey-Fuller test (ADF test) is often used to test the stationarity of each index sequence separately.