Question regarding Varmax Causal Statement (Group1&Group2)

[junesalee]

Hi all,

I was wondering if I could get some clarification on the Causal Statement. It seems like the explanation given on the manual seem contradictory to each other.

SAS/ETS(R) 9.3 User's Guide

The two statements are:

1) The null hypothesis of the Granger causality test is that GROUP1 is influenced only by itself, and not by GROUP2. If the hypothesis test fails to reject the null, then the variables listed in GROUP1 might be considered as independent variables.

2) The first CAUSAL statement fits the VAR(1) model by using the variables , , , and and tests the null hypothesis that causes the other variables, , , and , but the other variables do not cause . The second CAUSAL statement fits the VAR(1) model by using the variables , , and and tests the null hypothesis that causes the other variables, and , but the other variables do not cause .

When we use the CAUSAL statement, does rejecting the null hypothesis mean that Group1 is caused by Group2 or is it vice versa?

Thank you in advance.

[SteveDenham · Accepted solution]

Rejecting the null in the CAUSAL statement means that there is evidence that GROUP1 can be predicted by, or are closely associated with, GROUP2 variables.  Since we put variables in GROUP2 that we believe are causal based on other knowledge, then rejecting the null supports the assumption of causality.

Steve Denham

[junesalee]

Thank you Steve,

I wasn't sure about it as the example shown on the user guide uses this command:

proc varmax data=one;

  model y1-y3 = x1 / p=1;

  causal group1=(x1) group2=(y1-y3);

  causal group1=(y2) group2=(y1 y3);

  run;

Just to confirm, is statement 1 in the original post the correct way to infer the results from the test?