格兰杰因果检验和向量自回归（VAR）模型问题 我已经做完格兰杰因果检验,结果如下.Pairwise Granger Causality TestsDate:05/07/12 Time:12:55Sample:1990 2010Lags:2Null Hypothesis:\x05Obs\x05F-Statistic\x05ProbabilitySER02 do

格兰杰因果检验和向量自回归（VAR）模型问题 我已经做完格兰杰因果检验,结果如下.Pairwise Granger Causality TestsDate:05/07/12 Time:12:55Sample:1990 2010Lags:2Null Hypothesis:\x05Obs\x05F-Statistic\x05ProbabilitySER02 do
格兰杰因果检验和向量自回归（VAR）模型问题
我已经做完格兰杰因果检验,结果如下.Pairwise Granger Causality Tests
Date:05/07/12 Time:12:55
Sample:1990 2010
Lags:2
Null Hypothesis:\x05Obs\x05F-Statistic\x05Probability
SER02 does not Granger Cause SER01\x0519\x05 0.33516\x05 0.72081
SER01 does not Granger Cause SER02\x05 2.03279\x05 0.16785
SER03 does not Granger Cause SER01\x0519\x05 8.32044\x05 0.00416
SER01 does not Granger Cause SER03\x05 3.83285\x05 0.04704
SER04 does not Granger Cause SER01\x0519\x05 4.76461\x05 0.02640
SER01 does not Granger Cause SER04\x05 0.15004\x05 0.86204
SER05 does not Granger Cause SER01\x0519\x05 4.81833\x05 0.02557
SER01 does not Granger Cause SER05\x05 1.76083\x05 0.20791
SER03 does not Granger Cause SER02\x0519\x05 2.84640\x05 0.09178
SER02 does not Granger Cause SER03\x05 4.73175\x05 0.02692
SER04 does not Granger Cause SER02\x0519\x05 1.83852\x05 0.19545
SER02 does not Granger Cause SER04\x05 0.10371\x05 0.90218
SER05 does not Granger Cause SER02\x0519\x05 5.76340\x05 0.01493
SER02 does not Granger Cause SER05\x05 3.80248\x05 0.04798
SER04 does not Granger Cause SER03\x0519\x05 4.71858\x05 0.02714
SER03 does not Granger Cause SER04\x05 0.51736\x05 0.60706
SER05 does not Granger Cause SER03\x0519\x05 0.27672\x05 0.76232
SER03 does not Granger Cause SER05\x05 2.79020\x05 0.09553
SER05 does not Granger Cause SER04\x0519\x05 0.21012\x05 0.81299
SER04 does not Granger Cause SER05\x05 4.59718\x05 0.02919
他们有的相互之间存在因果关系,有的不存在.
那我下一步做VAR模型的时候如何选择外生变量啊.难道有的变量直接被抛弃?财富过后还可追加.

格兰杰因果检验和向量自回归（VAR）模型问题 我已经做完格兰杰因果检验,结果如下.Pairwise Granger Causality TestsDate:05/07/12 Time:12:55Sample:1990 2010Lags:2Null Hypothesis:\x05Obs\x05F-Statistic\x05ProbabilitySER02 do
逻辑思路好像有问题.
VAR模型建立之后,再用granger 因果部分检验其合理性.
而VAR模型的建立（即滞后阶数的选取）,不是依据granger因果关系是否成立的.
对于VAR模型,一般不选择外生变量.因为外生、内生很难界定（sims的观点）.当然若确实有理论基础或数据不够长时,也可采用外生变量,这时可以参考granger因果关系的结果.