ENGLE-GRANGER

[super_nova]

how i will test ENGLE-GRANGER for cointegration in eviews?
1. i estimate the long run equation.
2. i save the resedual.
3. i test the ADF test.
4. i got resedual is non stationary.
5. next, what should be my equation in equation estimator to estimate short-run dynamic model with ECM and test significance of ECM.

if my function is y=f(a1, a2, a3, a4, a5)

[EViews Glenn]

Simply regress the difference of Y on the lagged difference of Y and A1 to A5, as well as the error correction term (lagged residual from the first equation).

One note on the Engle-Granger test that you've already performed. The test that you've described is only approximate since the critical values used in the default EViews ADF test don't account for the estimation of the cointegrating equation. You should probably dig up the appropriate critical values, or use the Johansen test procedure.

[super_nova]

how i will get the appropiate critical value in eviews for engle-granger.

as my matrix knowledge is poor! it's very difficult for me to understand the explanation of johanson test result. do u have any suggestion for me regarding this?

[EViews Glenn]

The standard reference for the Engle-Granger critical values is

MacKinnon, James G., (1991) "Critical Values for Cointegration Tests", in Long-Run Economic Relationships, R.F. Engle and C.W.J. Granger (eds.), London, Oxford, pp 267-276.


As to the Johansen test. To be honest, getting a description of the Johansen test from the manual for statistical software may not be the best approach. You are best off reading a description of the Johansen procedure in a textbook like Hamilton or better yet, Johansen's book.

[nannigia]

well,,, yes it is complicated,
I tried to explain for you here, but it's really taking time ... so how can i help you, I can send you some audios teaching cointegration tests ... you can listen to them and learn better. reading the books also would be good...

please let me know if you are interested and give me your email address that i send you the audio files.

Cheers,

Yeganeh

[super_nova]

thanks
super_nova1982@yahoo.com

[super_nova]

The standard reference for the Engle-Granger critical values is

MacKinnon, James G., (1991) "Critical Values for Cointegration Tests", in Long-Run Economic Relationships, R.F. Engle and C.W.J. Granger (eds.), London, Oxford, pp 267-276.

can we get the same cointegration test critical value from the mackinnon software? the link is:
http://qed.econ.queensu.ca/pub/faculty/mackinnon/jbes/

but when i put the test statistics the program automatically closed...
well i am trying to get the p value for tau statistics, for 6 variables, and my test statistics is -8.230088
can anyone help me regarding this?