Page 1 of 1
Error correction Models with Cointegration vector FMOLS
Posted: Thu Sep 10, 2009 8:30 am
by Nadiarock1
Hi,
Previously, you gave me information about getting the Error Correction Terms for a FMOLS procedure.
I have 5 variables: py1 py2 raw terms and a constant. Py1 is the dependent variable.
I got the following cointegration coefficients for each of the above variables: 0.31 -0.49 -3.02 0.12 15.32.
I want to get my error correction terms. I have followed the procedure of example 10.6
Here is my syntax:
@fm(det=constant, lags=1)
#py1 py2 raw terms
equation(coeffs=||0.31 -0.49 -3.02 0.12 15.32) ect
#py1 py2 raw terms
system(model=ect)
variables py1 py2 raw terms
lags 1 to 4
det constant
ect ecteq
end(system)
estimate
It doesnt work. I think the problem is to link the ECM with my FMOLS coefficients. If you could give me help getting the ECM, it would be really appreciated.
Thank you , Nadia
Re: Error correction Models with Cointegration vector FMOLS
Posted: Thu Sep 10, 2009 11:41 am
by TomDoan
Nadiarock1 wrote:Hi,
Previously, you gave me information about getting the Error Correction Terms for a FMOLS procedure.
I have 5 variables: py1 py2 raw terms and a constant. Py1 is the dependent variable.
I got the following cointegration coefficients for each of the above variables: 0.31 -0.49 -3.02 0.12 15.32.
I want to get my error correction terms. I have followed the procedure of example 10.6
Here is my syntax:
Code: Select all
@fm(det=constant, lags=1)
#py1 py2 raw terms
equation(coeffs=||0.31 -0.49 -3.02 0.12 15.32) ect
#py1 py2 raw terms
First off, you need , separating the elements in the ||...||, and you're missing a closing ||. If you just retyped (incorrectly) your code, don't do that - paste in
exactly what you ran. Also, simply saying something "doesn't work" doesn't help. How does it not work?
The fm procedure should only be producing four numbers - the coefficients on (in order) constant, py2, raw and terms in a regression normalized with py1 (the first variable) as the dependent variable. You should be able to define this with
Code: Select all
@fm(det=constant, lags=1)
#py1 py2 raw terms
equation(coeffs=%beta) ect py1
# constant py2 raw terms
[/quote]
This defines an error correction term normalized with a 1.0 coefficient on py1.
Re: Error correction Models with Cointegration vector FMOLS
Posted: Mon Sep 14, 2009 6:45 am
by luxu1983
TomDoan wrote:Nadiarock1 wrote:Hi,
Previously, you gave me information about getting the Error Correction Terms for a FMOLS procedure.
I have 5 variables: py1 py2 raw terms and a constant. Py1 is the dependent variable.
I got the following cointegration coefficients for each of the above variables: 0.31 -0.49 -3.02 0.12 15.32.
I want to get my error correction terms. I have followed the procedure of example 10.6
Here is my syntax:
Code: Select all
@fm(det=constant, lags=1)
#py1 py2 raw terms
equation(coeffs=||0.31 -0.49 -3.02 0.12 15.32) ect
#py1 py2 raw terms
First off, you need , separating the elements in the ||...||, and you're missing a closing ||. If you just retyped (incorrectly) your code, don't do that - paste in
exactly what you ran. Also, simply saying something "doesn't work" doesn't help. How does it not work?
The fm procedure should only be producing four numbers - the coefficients on (in order) constant, py2, raw and terms in a regression normalized with py1 (the first variable) as the dependent variable. You should be able to define this with
Code: Select all
@fm(det=constant, lags=1)
#py1 py2 raw terms
equation(coeffs=%beta) ect py1
# constant py2 raw terms
This defines an error correction term normalized with a 1.0 coefficient on py1.[/quote]
this can be used for panel cointegration ' ECM