3:35 p.m.
Hello all,
Why does this model generate different outcomes, even if the same variables
and specification are been used? There is just a change in the position of
the variables.
Many thanks,
Bruno
3:42 p.m.
Am 04.10.2012 17:35, schrieb Bruno Thiago Tomio:
Hello all,
Why does this model generate different outcomes, even if the same
variables and specification are been used? There is just a change in
the position of the variables.
yhat1 <> logarea -> different results
Many thanks,
Bruno
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
6:53 p.m.
New subject: RES: Same model, different position of variables, different results
Sorry, I did not get it, Pindar. Could you please be more specific?
De: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Em nome de Pindar
Enviada em: quinta-feira, 4 de outubro de 2012 12:42
Para: Gretl list
Assunto: Re: [Gretl-users] Same model, different position of variables,
different results
Am 04.10.2012 17:35, schrieb Bruno Thiago Tomio:
Hello all,
Why does this model generate different outcomes, even if the same variables
and specification are been used? There is just a change in the position of
the variables.
yhat1 <> logarea -> different results
Many thanks,
Bruno
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
7:01 p.m.
New subject: RES: Same model, different position of variables, different results
Am 04.10.2012 20:53, schrieb Bruno Thiago Tomio:
Sorry, I did not get it, Pindar. Could you please be more specific?
I know just enough Portuguese (and also know gretl's messages...) to see
that you have an exact collinearity problem, so gretl drops some
arbitrary variables, and this arbitrary choice depends on the ordering.
Or is there more to it? If so, please show us something without the rank
deficiency.
cheers,
sven
7:05 p.m.
New subject: RES: Same model, different position of variables, different results
Probably gretl is discarding the first variable found to be colinear.
So yhat1 is colinear to logarea. When you exchange the position, you
get a diferent set.
Hélio
On Thu, Oct 4, 2012 at 7:53 PM, Bruno Thiago Tomio
<brunottomio(a)yahoo.com.br> wrote:
Sorry, I did not get it, Pindar. Could you please be more specific?
De: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Em nome de Pindar
Enviada em: quinta-feira, 4 de outubro de 2012 12:42
Para: Gretl list
Assunto: Re: [Gretl-users] Same model, different position of variables,
different results
Am 04.10.2012 17:35, schrieb Bruno Thiago Tomio:
Hello all,
Why does this model generate different outcomes, even if the same variables
and specification are been used? There is just a change in the position of
the variables.
yhat1 <> logarea -> different results
Many thanks,
Bruno
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
10:30 p.m.
New subject: RES: Same model, different position of variables, different results
Dear Bruno,
The problem is simple: the variable "logarea" is not equal to
"yhat1".
logarea yhat1
1 13.73169 17.32861
2 13.88510 16.59191
3 14.90278 9.24835
4 10.38514 43.42590
5 5.48064 80.59877
6 10.92590 37.84531
...
Um abraço,
Henrique
1:12 p.m.
New subject: RES: Same model, different position of variables, different results
Dear Bruno,
It seems that you are running a regresion with boht logarea and yhat1,
however, gretl drops one of the variable because they collinear. That is
right given that correlation is 99.93%
abertura logrendpc logarea yhat1
1.0000 0.1402 -0.6693 0.6698 abertura
1.0000 -0.1715 0.2093 logrendpc
1.0000 -0.9993 logarea
1.0000 yhat1
Now, your question is why you have a different dropped variable if you
change the order of the regressors, right? I am not sure why that is the
case, but it happens also in other packages. I remember similar issue in
Stata, when you have a different dropped variable if you use a different
command, but forcing to apply the same procedure. For example comparing OLS
con IV (but with trivial instruments).
Although, I don't have the question I think is not a serious problem. If 2
variables are highly correlated then any of those can be used. Indeed, many
times you could cut down your regressor into principal components to avoid
collinearity. Moreover, yhat1 is predicted variable from a previous model
and it is likely that logarea (or some similar variable) is a relevant in
that model.
Best, Rodrigo.
2012/10/4 Henrique Andrade <henrique.coelho(a)gmail.com>
Dear Bruno,
The problem is simple: the variable "logarea" is not equal to
"yhat1".
logarea yhat1
1 13.73169 17.32861
2 13.88510 16.59191
3 14.90278 9.24835
4 10.38514 43.42590
5 5.48064 80.59877
6 10.92590 37.84531
...
Um abraço,
Henrique
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
1:40 p.m.
New subject: RES: RES: Same model, different position of variables, different results
Dear all,
Thank you very much for your replies (Rodrigo, Sven and Henrique).
This was part of an exogeneity test. When the variable yhat1 was excluded,
the test does not work. But when one rearrange the model in a way that yhat1
is not placed at the end, the test is possible (i.e. gretl excludes logarea
instead of yhat1, due to the collinearity problem).
Best,
Bruno
De: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Em nome de Rodrigo Alfaro
Arancibia
Enviada em: sexta-feira, 5 de outubro de 2012 10:12
Para: Gretl list
Assunto: Re: [Gretl-users] RES: Same model, different position of variables,
different results
Dear Bruno,
It seems that you are running a regresion with boht logarea and yhat1,
however, gretl drops one of the variable because they collinear. That is
right given that correlation is 99.93%
abertura logrendpc logarea yhat1
1.0000 0.1402 -0.6693 0.6698 abertura
1.0000 -0.1715 0.2093 logrendpc
1.0000 -0.9993 logarea
1.0000 yhat1
Now, your question is why you have a different dropped variable if you
change the order of the regressors, right? I am not sure why that is the
case, but it happens also in other packages. I remember similar issue in
Stata, when you have a different dropped variable if you use a different
command, but forcing to apply the same procedure. For example comparing OLS
con IV (but with trivial instruments).
Although, I don't have the question I think is not a serious problem. If 2
variables are highly correlated then any of those can be used. Indeed, many
times you could cut down your regressor into principal components to avoid
collinearity. Moreover, yhat1 is predicted variable from a previous model
and it is likely that logarea (or some similar variable) is a relevant in
that model.
Best, Rodrigo.
11:31 a.m.
New subject: Signs and Interval of Error Corrections Terms in within a VECM framework
Hi All,
I understand within a bivariate VECM framework, the Error Correction
Terms(ECTs) have to be negative and situated within the interval: 0 and -1. Does this
principle on sign and interval apply to a multivariate VECM framework?
Cheers
6:34 p.m.
New subject: Signs and Interval of Error Corrections Terms in within a VECM framework
Dear Asongu,
In Vinod, H.D. (2008)'s "Hands-on Intermediate Econometrics Using R", in
sections 3.4.2 to 3.4.5 I've found some hints for your question. The author
uses a bivariate ECM. If you consider no a priori knowledge for the
relationship between, say, x and y, so we have a system of two equations,
both with ECM's.
In this case, says the author: "If the equilibrium error experienced by
economic agent at time t-1 is positive, in inequality (yt-1 > bxt-1) must
hold. During the current period t decreasing the left-hand side (yt < yt-1
or deltayt < 0) and increasing the right-hand side (bxt > bxt-1, deltaxt >
0, since b >0) of the inequality reduces the equilibrium error. If the
agent learns from past errors in predictable ways, we have seen that this
implications on the signs of coefficients in [number of equation] implying
nonrejection of two hypothesis, gama1 > 0 and gama2 <0".
Gama1 and 2 are, respectively, the coefficients of the long-run
relationship in t-1 (as usual in VECM). The gama 1 is for delta xt's
equation and the gama 2 for the delta yt's one).
So, in this case, you should expect a positive coefficient for the
coefficient. Is that what you asked? Hope to have helped you.
Best Wishes,
Claudio D. Shikida
http://www.shikida.net and http://works.bepress.com/claudio_shikida/
Esta mensagem pode conter informação confidencial e/ou privilegiada. Se
você não for o destinatário ou a pessoa autorizada a receber esta mensagem,
não poderá usar, copiar ou divulgar as informações nela contidas ou tomar
qualquer ação baseada nessas informações. Se você recebeu esta mensagem por
engano, por favor avise imediatamente o remetente, respondendo o presente
e-mail e apague-o em seguida.
This message may contain confidential and/or privileged information. If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose or take any action based on this message or
any information herein. If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message.
On Sat, Oct 13, 2012 at 8:31 AM, Anutechia Asongu
<simplice_peace(a)yahoo.com>wrote:
Hi All,
I understand within a bivariate VECM framework, the Error
Correction Terms(ECTs) have to be negative and situated within the
interval: 0 and -1. Does this principle on sign and interval apply to a
multivariate VECM framework?
Cheers
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
11:35 a.m.
New subject: Signs and Interval of Error Corrections Terms in within a VECM framework
Hi
I would not say that the ECT terms have to be between 0 and -1. it was not
clear to me whether you are referring to the cointegrating parameters (the
betas) or the speed adjustment (alphas)? For the betas, the values from
these vectors can take any value, but need to be normalised for
identification purpose.
For the alphas, the expected sign depends precisely on the corresponding
value in the cointegrating vector. The reasoning in the bi-variate case can
be simply extended to the k-variate. Taking a vector: y= b1 x - b2z, (with
b1 and b2 positive):
With a positive shocks, so that y > b1 x - b2z, if variables are "error
correcting" you expect:
-y to have negative alpha coefficient (should decrease the y to "reduce"
the inequality by reducing the left)
-x to have a positive coefficient (should increase the x in order to
"reduce" the inequality by increasing the right)
-z to have a negative coefficient (should increase the z in order to
"reduce" the inequality by increasing the right)
So I do not think one can say that the values should be bezween 0 and -1.
Best
Matthieu
2012/10/15 Claudio Shikida (敷田治誠 クラウジオ) <cdshikida(a)gmail.com>
Dear Asongu,
In Vinod, H.D. (2008)'s "Hands-on Intermediate Econometrics Using R", in
sections 3.4.2 to 3.4.5 I've found some hints for your question. The author
uses a bivariate ECM. If you consider no a priori knowledge for the
relationship between, say, x and y, so we have a system of two equations,
both with ECM's.
In this case, says the author: "If the equilibrium error experienced by
economic agent at time t-1 is positive, in inequality (yt-1 > bxt-1) must
hold. During the current period t decreasing the left-hand side (yt < yt-1
or deltayt < 0) and increasing the right-hand side (bxt > bxt-1, deltaxt >
0, since b >0) of the inequality reduces the equilibrium error. If the
agent learns from past errors in predictable ways, we have seen that this
implications on the signs of coefficients in [number of equation] implying
nonrejection of two hypothesis, gama1 > 0 and gama2 <0".
Gama1 and 2 are, respectively, the coefficients of the long-run
relationship in t-1 (as usual in VECM). The gama 1 is for delta xt's
equation and the gama 2 for the delta yt's one).
So, in this case, you should expect a positive coefficient for the
coefficient. Is that what you asked? Hope to have helped you.
Best Wishes,
Claudio D. Shikida
http://www.shikida.net and http://works.bepress.com/claudio_shikida/
Esta mensagem pode conter informação confidencial e/ou privilegiada. Se
você não for o destinatário ou a pessoa autorizada a receber esta mensagem,
não poderá usar, copiar ou divulgar as informações nela contidas ou tomar
qualquer ação baseada nessas informações. Se você recebeu esta mensagem por
engano, por favor avise imediatamente o remetente, respondendo o presente
e-mail e apague-o em seguida.
This message may contain confidential and/or privileged information. If
you are not the addressee or authorized to receive this for the addressee,
you must not use, copy, disclose or take any action based on this message
or any information herein. If you have received this message in error,
please advise the sender immediately by reply e-mail and delete this
message.
On Sat, Oct 13, 2012 at 8:31 AM, Anutechia Asongu <
simplice_peace(a)yahoo.com> wrote:
>
> Hi All,
> I understand within a bivariate VECM framework, the Error
> Correction Terms(ECTs) have to be negative and situated within the
> interval: 0 and -1. Does this principle on sign and interval apply to a
> multivariate VECM framework?
> Cheers
>
>
> _______________________________________________
> Gretl-users mailing list
> Gretl-users(a)lists.wfu.edu
> http://lists.wfu.edu/mailman/listinfo/gretl-users
>
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
12:25 p.m.
New subject: Signs and Interval of Error Corrections Terms in within a VECM framework
Am 16.10.2012 13:35, schrieb Matthieu Stigler:
So I do not think one can say that the values should be bezween 0 and -1.
This statement is correct. But please forgive me that I'm not going to
contribute to this non gretl-specific discusson apart from this
reaction. (Which doesn't mean that I necessarily want you to stop the
discussion, but I would recommend that the interested people study the
textbook literature as well/instead/first.)
cheers,
sven
4213
days inactive
4225
days old
11 comments
9 participants
participants (9)
-
Anutechia Asongu -
Bruno Thiago Tomio -
Claudio Shikida (敷田治誠 クラウジオ) -
Henrique Andrade -
Hélio Guilherme -
Matthieu Stigler -
Pindar -
Rodrigo Alfaro Arancibia -
Sven Schreiber