Pairwise Granger Causality Tests
Date: 06/15/05 Time: 00:51
Sample: 1979 2004
Lags: 1
Null Hypothesis: Obs F-Statistic Probability
Y does not Granger Cause XM 25 0.64706 0.42977
XM does not Granger Cause Y 6.01380 0.02259
Pairwise Granger Causality Tests
Date: 06/15/05 Time: 00:51
Sample: 1979 2004
Lags: 2
Null Hypothesis: Obs F-Statistic Probability
Y does not Granger Cause XM 24 0.43590 0.65299
XM does not Granger Cause Y 4.25287 0.02976
以上结果表明,无论在1阶还是2阶条件下,XM与Y之间都是XM为因,Y为果的单向因果关系。
B,对XM(-1)有:
Pairwise Granger Causality Tests
Date: 06/15/05 Time: 00:51
Sample: 1979 2004
Lags: 1
Null Hypothesis: Obs F-Statistic Probability
Y does not Granger Cause XM(-1) 25 0.64706 0.42977
XM(-1) does not Granger Cause Y 6.01380 0.02259
Pairwise Granger Causality Tests
Date: 06/15/05 Time: 00:51
Sample: 1979 2004
Lags: 2
Null Hypothesis: Obs F-Statistic Probability
Y does not Granger Cause XM(-1) 24 0.43590 0.65299
XM(-1) does not Granger Cause Y 4.25287 0.02976
由ARCH检验的结果,在各个阶次下的F检验P值都很大,因此不能拒绝原假设,同时所有滞后项的t检验值都不显著,因此我们可以认为该短期调整模型不存在由于随机扰动项的逐期积累而产生的异方差性。
WHITE检验:
White Heteroskedasticity Test:
F-statistic 0.816420 Probability 0.572862
Obs*R-squared 5.391098 Probability 0.494715