EViews.com

Dear Trubador or Startz or any of the all great moderators,
Again thanks so much for the continuing support in learning how to use eviews to the full extent. I am using some of Wooldridge 2e problems and excercises to understand certain parts that I need to do to prepare and finish my paper as I mentioned above. (I have told our ph.d. director how helpful you guys are. He is a former vp at the fed. He also teaches econometrics and requires use of eviews for his course, so he was really pleased to hear about the great support that you guys give for your product. I personally have the student edition, but we have the full eviews in all our labs. Good work guys! :D)

Question:
I am trying to understand how to enter the correct parameters to perform the granger causality test so I can run it for my papers results. I am using the data that I copied below with the results.

I went to quick/group stats/granger causality test and a window pops up asking for a series, equation, etc. I entered only pcip (percentage change in indust. prod) and pcsp (percentage change in S&P 500) and got the results:

Null Hypothesis: Obs F-Statistic Prob.

PCIP does not Granger Cause PCSP 555 0.88103 0.4149
PCSP does not Granger Cause PCIP 6.43434 0.0017

Question: Is the 6.43 of probability saying that it is strongly significant that PCSP does not granger cause PCIP or is the figure/strong result reject the null?

Question:
I know this must not be how to do it. I am trying to enter in the correct parameters from the following problem to get what they got but first am trying to understand how to enter the parameters correctly and then Interpret the results in proper fashion.

Part I: In the model
pcipt = δ0 + α1pcipt-1 + α2pcipt-2 + α3pcipt-3 + γ1pcspt-1 + γ2pcspt-2 + γ3pcspt-3 + ut,
The null hypothesis is that pcsp does not Granger cause pcip. This is stated as H0: γ1 = γ2 = γ3 = 0. The F statistic for joint significance of the three lags of pcspt, with 3 and 547 df, is F = 5.37 and p-value = .0012. Therefore, we strongly reject H0 and conclude that pcsp does Granger cause pcip.

Here below is the model I started with:
The estimated AR(3) model for pcipt is
pcipt = 1.80 + .349 pcipt-1 + .071 pcipt-2 + .067 pcipt-2
(0.55) (.043) (.045) (.043)
n = 554, R2 = .166, ˆσ = 12.15.
When pcipt-4 is added, its coefficient is .0043 with a t statistic of about .10.

Question: I would like to be able to perform the granger causality test and enter the correct parameters to arrive at the same estimates as part I but I know I am entering it the wrong way. Any help would be much appreciated in terms of what to enter into the granger box that asks for series, equations, variable and how many lags to answer. If you could help me with this I would be so appreciative and would then be able to conduct the test on my paper.

This is the second granger test that they ran from wooldridge 2e.
(iii) When we add Δi3t-1, Δi3t-2, and Δi3t-3 to the regression from part (ii), and now test the joint significance of pcspt-1, pcspt-2, and pcspt-3, the F statistic is 5.08. With 3 and 544 df in the F distribution, this gives p-value = .0018, and so pcsp Granger causes pcip even conditional on past Δi3.

Question: I again I don't know what to enter into the dialogue box that pops up once you do quick/group estimates/granger test, the the amount of lags, and then to interpret the results? Sorry to be such a pain but i have looked through the manual and the forum and did not find anything that would help. God Bless you guys. Thanks very much. guero303030

hi, i can help you regarding your first question, cause i'm also using granger causality test these days.

Null Hypothesis: Obs F-Statistic Prob.

PCIP does not Granger Cause PCSP 555 0.88103 0.4149
PCSP does not Granger Cause PCIP 6.43434 0.0017

here the probilities are .4149 and .0017. basically you are looking for value below 5%. the first hypothesis cannot be rejected because of higher value of .4149, but for second hypothesis, you can reject it based on the probabilty figure of .0017 which implies that PCSP does granger cause PCIP but it is like you have to try with various lags to come to a min. prob figure to find suitable lagging period. the other figures (6.43434 and .88103) are f-test statistices. hopefully, it helped you.