10.1 Granger causality test
Prior to Granger causality test it is cruical to select the optimal lag length, because the choice of lags p can influence the results (using too few lags may omit relevant information, while too many can introduce noise)
Lag selection is usually based on information criteria that penalize model complexity, such as Akaike Information Criterion (AIC) or Bayesian Information Criterion (BIC) in (6.14)
Sometimes, AIC and BIC do not indicate the same lag selection and additional criteria can be used: Hannan-Quinn Criterion (HQC) and Final Prediction Error (FPE)
Variable xt is said to “Granger-cause” yt if the past values of xt significantly influence the current value of yt. Assuming that the optimal lag selection p=1, Granger causality test is performed simultaneously by testing two null hypothesis
H0:β1,2=0 (xt does not cause yt)H0:β2,1=0 (yt does not cause xt)
Granger causality test reduces to testing the significance off-diagonal terms of the matrix A1 (so called companion matrix)
If both null hypothesis (10.4) are rejected then causality in both directions exist
F−statistic or Wald statistic (χ2 version) can be used to check both null hypothesis
Exercise 44. Conclude about Granger causality direction(s) in the following cases:
I. null hypothesis H0:β1,2=0 is not rejected, while null hypothesis H0:β2,1=0 is rejected
II. null hypothesis H0:β1,2=0 is rejected, while null hypothesis H0:β2,1=0 is not rejected
III. both null hypothesis H0:β1,2=0 and H0:β2,1=0 are rejected
IV. both null hypothesis H0:β1,2=0 and H0:β2,1=0 are not rejected
Solution
I. yt causes xt onlyII. xt causes yt only
III. there is no any causality between xt and yt
IV. there is causality in both directions