Page 1 of 2
cointegration test problem
Posted: Wed Aug 29, 2012 7:39 pm
by leolii
Hi
I have EViews 7.2 version. I was doing cointegration test for my panel data, when I selected Fisher(combined Johansen) and run the test, a error message says "invalid or duplicate specification or insufficient data for estimations." But If I select the other method which is Kao(Engle-Granger based), the test would be fine, it gave me result to reject the H0.
Thanks for any suggestion
Re: cointegration test problem
Posted: Wed Aug 29, 2012 8:23 pm
by EViews Gareth
You should update to EViews 7.2.
Re: cointegration test problem
Posted: Thu Aug 30, 2012 8:07 am
by leolii
You should update to EViews 7.2.
sorry, it is 7.2 version.
Re: cointegration test problem
Posted: Thu Aug 30, 2012 8:24 am
by EViews Gareth
What's the build date? (You can check from Help->About EViews)
Re: cointegration test problem
Posted: Thu Aug 30, 2012 8:49 am
by leolii
What's the build date? (You can check from Help->About EViews)
Enterprise Edition -May 4 2011
Re: cointegration test problem
Posted: Thu Aug 30, 2012 9:02 am
by EViews Gareth
Please update to the latest version (Help->EViews Update).
Re: cointegration test problem
Posted: Fri Aug 31, 2012 8:11 am
by leolii
Please update to the latest version (Help->EViews Update).
I updated to Aug 24 2012 build, but it still the same, shows the error message.
Re: cointegration test problem
Posted: Fri Aug 31, 2012 8:14 am
by EViews Gareth
Could you post your workfile?
Re: cointegration test problem
Posted: Fri Aug 31, 2012 8:48 am
by leolii
Could you post your workfile?
Sure
Re: cointegration test problem
Posted: Fri Aug 31, 2012 9:59 am
by EViews Gareth
You don't have enough time periods to do an 11 variable Fisher cointegration test. Our error message could be a little more descriptive.
Re: cointegration test problem
Posted: Fri Aug 31, 2012 10:04 am
by leolii
You don't have enough time periods to do an 11 variable Fisher cointegration test. Our error message could be a little more descriptive.
i c, i only got 10 years, so i need at least 11 years data do to the fisher test? Thank you very much.
Re: cointegration test problem
Posted: Fri Aug 31, 2012 10:09 am
by EViews Glenn
You'll need more because of the lags in the VAR.
Re: cointegration test problem
Posted: Fri Aug 31, 2012 10:13 am
by EViews Gareth
No, you'll need way more than that.
Re: cointegration test problem
Posted: Sat Sep 01, 2012 7:21 am
by leolii
No, you'll need way more than that.
Thank you very for helping
Re: cointegration test problem
Posted: Sat Sep 29, 2012 7:50 am
by amirpooya
Hi, I read your post in Eviews I have same question about panel cointegration. I used fisher ADF unit root test and then I couldn't use pedroni or fisher . I can use just Kao test . so I want to know what did you do after that? you used Kao or what?
Regards,
amirpooya