Page 1 of 1
Fully Modified OLS (Phillips-Hansen)
Posted: Wed Mar 17, 2010 7:46 am
by Bigbrotherjx
From which version does Eviews begin to support this estimation method?
Re: Fully Modified OLS (Phillips-Hansen)
Posted: Wed Mar 17, 2010 9:29 am
by EViews Gareth
EV7
Re: Fully Modified OLS (Phillips-Hansen)
Posted: Wed Mar 17, 2010 1:05 pm
by Bigbrotherjx
Does this procedure require all the variables i.e. both explanatory and dependent variables to be I(1)? I have a dependent variable which is I(0) and most of the explanatory variables are I(1).
Re: Fully Modified OLS (Phillips-Hansen)
Posted: Wed Mar 17, 2010 1:38 pm
by EViews Glenn
You'll need to rearrange so the dependent is I(1).
Re: Fully Modified OLS (Phillips-Hansen)
Posted: Wed Mar 17, 2010 2:05 pm
by Bigbrotherjx
If my dependent variable is naturally I(0) but some of my independent variables are I(1), does this mean I cannot use this method then?
Re: Fully Modified OLS (Phillips-Hansen)
Posted: Wed Mar 17, 2010 3:48 pm
by EViews Glenn
There really is no meaning to dependent and independent variable in the FMOLS setting. Indeed one of the criticisms of FMOLS as opposed to Johansen cointegration is that the results can differ as you choose different I(1) variables to be the "dependent" variable in the first-stage cointegrating regression.