Page 1 of 1
Gregory-Hansen Cointegration Test
Posted: Thu May 09, 2019 9:06 am
by nndzama
Greetings
When I try to run the run Gregory Hansen Cointegration Test I get this error:Near singular matrix error. Regressors may be perfectly
collinear in "DO_ GHC.LS Y C (@TREND>6-2) G"[/b].[/u] I have tried all the tricks I know but no solution. I only have 1 dependent variable (exports) and 1 independent variable (imports).
Please assist.
Kind Regards
Nwabisa
Re: Gregory-Hansen Cointegration Test
Posted: Thu May 09, 2019 10:07 am
by startz
What values do you see for the series
?
Re: Gregory-Hansen Cointegration Test
Posted: Thu May 09, 2019 10:18 pm
by nndzama
Hi Startz
I also have no idea where is that coming from (@Trend>6-2, it pops up when I run the program. All I know is that I selected REGIME SHIFT model because I want to capture both change in intercept and trend.
Kind Regards
Nwabisa
Re: Gregory-Hansen Cointegration Test
Posted: Thu May 09, 2019 11:30 pm
by EViews Gareth
You might want to post your workfile and exact instructions on what you are doing.
Re: Gregory-Hansen Cointegration Test
Posted: Thu May 09, 2019 11:53 pm
by nndzama
Hi Gareth
I have attached my workfile and G-H program. The study examines the sustainability of the current account. The aim of running this test is to find if the long-run relationship exists between exports and imports and account for any structural breaks.
Kind Regards
Nwabisa
Re: Gregory-Hansen Cointegration Test
Posted: Fri May 10, 2019 5:56 am
by startz
The problem is that LRX and LRM_ are only defined from 2009 and that @TREND>6-2 is always 1 in this period, hence collinear with the intercept.
As a guess, try setting the smpl to the period in which all your variables are valid.