3:20 p.m.
Hello everyone,
I need to estimate an equation with tsls. This equation has a restriction on
parameters. The restriction has inside itself values coming from the
estimation parameters of another equation and it is the following:
b3=[a1/(1-a2)](1-b2)-b1, on which "a" indicate a parameter of the first
equation (already estimated) and "b" indicate a parameter of the second
equation (which has to be estimated with tsls) and the numbers indicate the
parameter associated to a regressor.
How can I estimated the second equation once I have estimated the first.
Thank you for the support!
3:57 p.m.
On Thu, 4 Feb 2016, Carlo Maria Petrangelo wrote:
Hello everyone,
I need to estimate an equation with tsls. This equation has a restriction on
parameters. The restriction has inside itself values coming from the
estimation parameters of another equation and it is the following:
b3=[a1/(1-a2)](1-b2)-b1, on which "a" indicate a parameter of the first
equation (already estimated) and "b" indicate a parameter of the second
equation (which has to be estimated with tsls) and the numbers indicate the
parameter associated to a regressor.
How can I estimated the second equation once I have estimated the first.
Premise: for this kind of things, your best bet is ALWAYS two estimate the
two equations jointly. Otherwise, you have to make a few adjustments to
you standard errors which are not only boring to compute, but also prone
to sizeable finite-sample bias.
That said, I see three alternatives:
* write the log-likelihood for the two equations under normality and use
mle
* re-express your problem as a suitable set of moment conditions and use
gmm
* use a lego-like approach:
- estimate 1st stage via ols and get a1 and a2, plus residuals u
- assuming your "main" equation looks like
y = x1 * b1 + x2 * b2 + x3 * b3
compute w = y - (a1/(1-a2))*x3
compute z1 = x1 - x3
compute z2 = x2 - (a1/(1-a2))*x3
- run ols on
w = z1 * b1 + z2 * b2 + gamma * u
- bootstrap the whole thing
this will give you consistent estimates of a1, a2, b1 and b2, from which
you can compute b3 via the delta method.
HTH,
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
6:11 p.m.
Am 04.02.2016 um 16:57 schrieb Riccardo (Jack) Lucchetti:
On Thu, 4 Feb 2016, Carlo Maria Petrangelo wrote:
> Hello everyone,
>
> I need to estimate an equation with tsls. This equation has a
> restriction on
> parameters. The restriction has inside itself values coming from the
> estimation parameters of another equation and it is the following:
> b3=[a1/(1-a2)](1-b2)-b1, on which "a" indicate a parameter of the first
> equation (already estimated) and "b" indicate a parameter of the second
> equation (which has to be estimated with tsls) and the numbers
> indicate the
> parameter associated to a regressor.
>
> How can I estimated the second equation once I have estimated the first.
Premise: for this kind of things, your best bet is ALWAYS two estimate
the two equations jointly. Otherwise, you have to make a few adjustments
to you standard errors which are not only boring to compute, but also
prone to sizeable finite-sample bias.
While Jack's answer is econometrically the proper one, it could be that
it's not the one you wanted to hear. If you could live without the
correct standard errors and are really only interested in the point
estimates, then technically it's easy to save (or write down) the values
a1 and a2 and plug them into the restriction concerning b1 and b2,
especially since that restriction seems to be linear, given a1 and a2.
So, please read the documentation of the "restrict" command, and report
back if anything remains unclear.
good luck,
sven
6:39 p.m.
New subject: R: Restriction on model coefficients
Thank you Sven for your advice!
Indeed I think the Riccardo solution is econometrically appropriate, but I
need an easier one.
I have already tried the restriction command, but I did not get how to
impose the restriction and then launch the estimation. For just one equation
the guide says that the restriction is applied to the estimation (already
run) of the equation.
So in theory:
1) you run the equation estimation (in my case with 2sls)
2) you put the restriction
3) it tests the coefficient restriction b3=[a1/(1-a2)](1-b2)-b1, using the
estimates coming from the "a" equation (already estimated), and the b1 and
b2 estimates coming from the estimation of the "b" equation (the one
estimated in point 1).
Instead, I need that in the estimation of the equation (the one estimated in
point 1), the coefficients would be estimated using the restriction, that is
using the restriction during the estimation.
I hope to have been enough clear.
Thank you.
-----Messaggio originale-----
Da: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Per conto di Sven Schreiber
Inviato: giovedì 4 febbraio 2016 19:12
A: gretl-users(a)lists.wfu.edu
Oggetto: Re: [Gretl-users] Restriction on model coefficients
Am 04.02.2016 um 16:57 schrieb Riccardo (Jack) Lucchetti:
On Thu, 4 Feb 2016, Carlo Maria Petrangelo wrote:
> Hello everyone,
>
> I need to estimate an equation with tsls. This equation has a
> restriction on parameters. The restriction has inside itself values
> coming from the estimation parameters of another equation and it is
> the following:
> b3=[a1/(1-a2)](1-b2)-b1, on which "a" indicate a parameter of the
> first equation (already estimated) and "b" indicate a parameter of
> the second equation (which has to be estimated with tsls) and the
> numbers indicate the parameter associated to a regressor.
>
> How can I estimated the second equation once I have estimated the first.
Premise: for this kind of things, your best bet is ALWAYS two estimate
the two equations jointly. Otherwise, you have to make a few
adjustments to you standard errors which are not only boring to
compute, but also prone to sizeable finite-sample bias.
While Jack's answer is econometrically the proper one, it could be that it's
not the one you wanted to hear. If you could live without the correct
standard errors and are really only interested in the point estimates, then
technically it's easy to save (or write down) the values
a1 and a2 and plug them into the restriction concerning b1 and b2,
especially since that restriction seems to be linear, given a1 and a2.
So, please read the documentation of the "restrict" command, and report back
if anything remains unclear.
good luck,
sven
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
8:26 p.m.
New subject: R: Restriction on model coefficients
Am 04.02.2016 um 19:39 schrieb Carlo Maria Petrangelo:
Instead, I need that in the estimation of the equation (the one
estimated in
point 1), the coefficients would be estimated using the restriction, that is
using the restriction during the estimation.
You're right that with tsls the restricted estimates aren't printed; I
wasn't aware of this and I don't know why. (It's mentioned in the docs:
"If the original model was estimated via OLS then the restricted
coefficient estimates are printed;", but that doesn't explain why. Maybe
Allin can elaborate.) Similarly with the "--full" option which only
applies to OLS (and VECMs, but that's a different story). So you also
cannot use the $coeff accessor after the restriction.
One not very satisfactory workaround would be to split the two stages of
two-stage-least-squares, so that you get the second stage mechanically
as OLS where you can retrieve the restricted results. But this is really
lame I admit.
sorry,
sven
9:02 p.m.
New subject: R: Restriction on model coefficients
On Thu, 4 Feb 2016, Sven Schreiber wrote:
Am 04.02.2016 um 19:39 schrieb Carlo Maria Petrangelo:
> Instead, I need that in the estimation of the equation (the one estimated in
> point 1), the coefficients would be estimated using the restriction, that is
> using the restriction during the estimation.
>
You're right that with tsls the restricted estimates aren't printed; I
wasn't aware of this and I don't know why. (It's mentioned in the docs:
"If the original model was estimated via OLS then the restricted
coefficient estimates are printed;", but that doesn't explain why. Maybe
Allin can elaborate.)
You can get hold of the restricted estimates by treating it as a
one-equation "system":
<hansl>
tsmod <- system
equation y 0 X ; 0 Z
end system
restrict tsmod
b2 - b3 = 0 # or whatever
end restrict
estimate tsmod method=tsls
</hansl>
Allin
6:40 p.m.
New subject: R: R: Restriction on model coefficients
Thank you so much Profesor Cottrell. It works and it is what I needed.
Kind regards,
Carlo Maria Petrangelo
-----Messaggio originale-----
Da: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Per conto di Allin Cottrell
Inviato: giovedì 4 febbraio 2016 22:02
A: Gretl list
Oggetto: Re: [Gretl-users] R: Restriction on model coefficients
On Thu, 4 Feb 2016, Sven Schreiber wrote:
Am 04.02.2016 um 19:39 schrieb Carlo Maria Petrangelo:
> Instead, I need that in the estimation of the equation (the one
> estimated in point 1), the coefficients would be estimated using the
> restriction, that is using the restriction during the estimation.
>
You're right that with tsls the restricted estimates aren't printed; I
wasn't aware of this and I don't know why. (It's mentioned in the docs:
"If the original model was estimated via OLS then the restricted
coefficient estimates are printed;", but that doesn't explain why.
Maybe Allin can elaborate.)
You can get hold of the restricted estimates by treating it as a
one-equation "system":
<hansl>
tsmod <- system
equation y 0 X ; 0 Z
end system
restrict tsmod
b2 - b3 = 0 # or whatever
end restrict
estimate tsmod method=tsls
</hansl>
Allin
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
9:50 p.m.
New subject: R: Restriction on model coefficients
Am 04.02.2016 um 22:02 schrieb Allin Cottrell:
You can get hold of the restricted estimates by treating it as a
one-equation "system":
I think this workaround should be mentioned somewhere in the tsls
documentation. (Any other estimators with the same pattern here?)
Or, given that gretl already knows how to do it, could the --full option
be somehow enabled for "restrict" after "tsls"?
thanks,
sven
10:43 a.m.
New subject: R: R: Restriction on model coefficients
I want to ask to you all for another question.
When I restrict the system, gretl puts the coefficient constrained equal to
zero
wage<-system
? equation w 0 w_1 lnGDPdeflator lnGDPdeflator_1 time;0 w_1 w_2 w_3
lnGDPdeflator_1 lnGDPdeflator_2 lnGDPdeflator_3
? end system
? restrict wage
? b4=[0,0408413/(1-4,17913)](1-b3)-b2
? end restrict
Vincolo:
b[1,4] = 0
? estimate wage method=tsls
Why???
Carlo Maria Petrangelo
-----Messaggio originale-----
Da: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Per conto di Sven Schreiber
Inviato: venerdì 5 febbraio 2016 22:50
A: gretl-users(a)lists.wfu.edu
Oggetto: Re: [Gretl-users] R: Restriction on model coefficients
Am 04.02.2016 um 22:02 schrieb Allin Cottrell:
You can get hold of the restricted estimates by treating it as a
one-equation "system":
I think this workaround should be mentioned somewhere in the tsls
documentation. (Any other estimators with the same pattern here?)
Or, given that gretl already knows how to do it, could the --full option be
somehow enabled for "restrict" after "tsls"?
thanks,
sven
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
12:12 p.m.
New subject: R: R: Restriction on model coefficients
Am 09.02.2016 um 11:43 schrieb Carlo Maria Petrangelo:
I want to ask to you all for another question.
When I restrict the system, gretl puts the coefficient constrained equal to
zero
wage<-system
? equation w 0 w_1 lnGDPdeflator lnGDPdeflator_1 time;0 w_1 w_2 w_3
lnGDPdeflator_1 lnGDPdeflator_2 lnGDPdeflator_3
? end system
? restrict wage
? b4=[0,0408413/(1-4,17913)](1-b3)-b2
? end restrict
Vincolo:
b[1,4] = 0
? estimate wage method=tsls
Please provide a working example of what you mean. For example, this
line cannot really work AFAICS: "b4=[0,0408413/(1-4,17913)](1-b3)-b2"
First there's a missing multiplication sign *, and secondly the comma as
decimal separator is not allowed in scripts.
cheers,
sven
1:21 p.m.
New subject: R: R: R: Restriction on model coefficients
In which sense a working example?
For now I apply your suggested modifications, thank you!
However, if I put "b4-b2=[0,0408413/(1-4,17913)](1-b3)" instead of "
b4=[0,0408413/(1-4,17913)](1-b3)-b2" it seems better, because it puts
"b4+b2=0" (obviously), and estimate the coefficients without a predetermined
zero value in b4 as it did before.
Carlo Maria Petrangelo
-----Messaggio originale-----
Da: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Per conto di Sven Schreiber
Inviato: martedì 9 febbraio 2016 13:12
A: gretl-users(a)lists.wfu.edu
Oggetto: Re: [Gretl-users] R: R: Restriction on model coefficients
Am 09.02.2016 um 11:43 schrieb Carlo Maria Petrangelo:
I want to ask to you all for another question.
When I restrict the system, gretl puts the coefficient constrained
equal to zero
wage<-system
? equation w 0 w_1 lnGDPdeflator lnGDPdeflator_1 time;0 w_1 w_2 w_3
lnGDPdeflator_1 lnGDPdeflator_2 lnGDPdeflator_3 ? end system ?
restrict wage ? b4=[0,0408413/(1-4,17913)](1-b3)-b2
? end restrict
Vincolo:
b[1,4] = 0
? estimate wage method=tsls
Please provide a working example of what you mean. For example, this line
cannot really work AFAICS: "b4=[0,0408413/(1-4,17913)](1-b3)-b2"
First there's a missing multiplication sign *, and secondly the comma as
decimal separator is not allowed in scripts.
cheers,
sven
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
1:43 p.m.
New subject: R: R: R: Restriction on model coefficients
Am 09.02.2016 um 14:21 schrieb Carlo Maria Petrangelo:
In which sense a working example?
It is almost totally unclear for the list readers including myself what
you are seeing. You haven't included the actual error message or
detailed output, and with "working example" I mean a script that I can
run on my gretl installation, preferably using a built-in example dataset.
For now I apply your suggested modifications, thank you!
However, if I put "b4-b2=[0,0408413/(1-4,17913)](1-b3)"
How can "[...](1-b3)" work in gretl's language hansl? AFAICS it has to
be "(...)*(1-b3)" at least. If gretl doesn't return an error at this
point, that's already strange and might be a parsing bug.
But to investigate what kind of bug it is we need an example script ideally.
thanks,
sven
2:51 p.m.
New subject: R: R: R: Restriction on model coefficients
On Tue, 9 Feb 2016, Sven Schreiber wrote:
Am 09.02.2016 um 14:21 schrieb Carlo Maria Petrangelo:
> In which sense a working example?
It is almost totally unclear for the list readers including myself what you
are seeing. You haven't included the actual error message or detailed output,
and with "working example" I mean a script that I can run on my gretl
installation, preferably using a built-in example dataset.
> For now I apply your suggested modifications, thank you!
>
> However, if I put "b4-b2=[0,0408413/(1-4,17913)](1-b3)"
How can "[...](1-b3)" work in gretl's language hansl? AFAICS it has to be
"(...)*(1-b3)" at least. If gretl doesn't return an error at this point,
that's already strange and might be a parsing bug.
There was a bug here (now fixed in git), in the sense that when the
right-hand side of a restriction was garbage (as above) this was being
replaced by zero without an error being flagged.
We were being too trusting -- assuming that users would have read the
doc on "restrict" and would therefore just supply a scalar value on
the right, as stipulated. Now we check rigorously.
Allin
3:15 p.m.
New subject: R: R: R: Restriction on model coefficients
On Tue, 9 Feb 2016, Carlo Maria Petrangelo wrote:
In which sense a working example?
For now I apply your suggested modifications, thank you!
However, if I put "b4-b2=[0,0408413/(1-4,17913)](1-b3)" instead of "
b4=[0,0408413/(1-4,17913)](1-b3)-b2" it seems better, because it puts
"b4+b2=0" (obviously), and estimate the coefficients without a predetermined
zero value in b4 as it did before.
Please read the documentation for the "restrict" command. When using
the simplest method (which you are attempting) the restriction must
take the form of a linear combination of parameters on the left and a
scalar value on the right.
Putting your restriction into something resembling this form we get
(b4-b2)/(1-b3) = -0.012846691
but since this is nonlinear you actually have to use the "rfunc" form
of the command, as in
function matrix my_restrict (const matrix b)
matrix v = (b[4]-b[2])/(1-b[3]) + 0.012846691
return v
end function
restrict wage
rfunc = my_restrict
end restrict
gretl will then use the delta method to evaluate the test.
Allin Cottrell
5:14 p.m.
New subject: R: R: R: R: Restriction on model coefficients
Thank you all for the time spending to help me, and I apologize for my
ignorance in econometrics and in gretl use (I have been using it for two
months :-) ).
Yes, no one error came out, but just a value zero for the variable as you
already said.
Now I'm going to try the solution proposed by Professor Cottrell, and in
case of further problems I'm going to send you thee gdt file.
Thank you once again,
Carlo Maria Petrangelo
-----Messaggio originale-----
Da: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Per conto di Allin Cottrell
Inviato: martedì 9 febbraio 2016 16:16
A: Gretl list
Oggetto: Re: [Gretl-users] R: R: R: Restriction on model coefficients
On Tue, 9 Feb 2016, Carlo Maria Petrangelo wrote:
In which sense a working example?
For now I apply your suggested modifications, thank you!
However, if I put "b4-b2=[0,0408413/(1-4,17913)](1-b3)" instead of "
b4=[0,0408413/(1-4,17913)](1-b3)-b2" it seems better, because it puts
"b4+b2=0" (obviously), and estimate the coefficients without a
predetermined zero value in b4 as it did before.
Please read the documentation for the "restrict" command. When using the
simplest method (which you are attempting) the restriction must take the
form of a linear combination of parameters on the left and a scalar value on
the right.
Putting your restriction into something resembling this form we get
(b4-b2)/(1-b3) = -0.012846691
but since this is nonlinear you actually have to use the "rfunc" form of the
command, as in
function matrix my_restrict (const matrix b)
matrix v = (b[4]-b[2])/(1-b[3]) + 0.012846691
return v
end function
restrict wage
rfunc = my_restrict
end restrict
gretl will then use the delta method to evaluate the test.
Allin Cottrell
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
6:30 p.m.
New subject: R: R: R: R: Restriction on model coefficients
Am 09.02.2016 um 18:14 schrieb Carlo Maria Petrangelo:
Thank you all for the time spending to help me, and I apologize for
my
ignorance in econometrics and in gretl use (I have been using it for two
months :-) ).
No worries -- in fact, questions like this can be useful for gretl, too:
In your case one bug was discovered and fixed.
cheers,
sven
6:50 p.m.
New subject: R: R: R: R: R: Restriction on model coefficients
So, I do not know if I had to write the commands in the exact way Professor
Cottrell did, but I did it since I do not know the command and I did not get
the ratio of the process (I'm sorry, but I'm not a math student and in this
case I just need the results and I can't spend time in understanding that).
So I ran the commands below and I got an error:
?wage<-system
? equation w 0 w_1 lnGDPdeflator lnGDPdeflator_1 time;0 w_1 w_2 w_3
lnGDPdeflator_1 lnGDPdeflator_2 lnGDPdeflator_3
? end system
? function matrix my_restrict (const matrix b)
matrix v = (b[4]-b[2])/(1-b[3]) + 0.012846691
return v
end function
? restrict wage
? rfunc = my_restrict
? end restrict
La statistica richiesta non è disponibile (Which in english should sounds
like "The statistic request is not available")
If you want, I send you the gdt file with the data.
Carlo Maria Petrangelo
-----Messaggio originale-----
Da: gretl-users-bounces(a)lists.wfu.edu
[mailto:gretl-users-bounces@lists.wfu.edu] Per conto di Sven Schreiber
Inviato: martedì 9 febbraio 2016 19:30
A: gretl-users(a)lists.wfu.edu
Oggetto: Re: [Gretl-users] R: R: R: R: Restriction on model coefficients
Am 09.02.2016 um 18:14 schrieb Carlo Maria Petrangelo:
Thank you all for the time spending to help me, and I apologize for
my
ignorance in econometrics and in gretl use (I have been using it for
two months :-) ).
No worries -- in fact, questions like this can be useful for gretl, too:
In your case one bug was discovered and fixed.
cheers,
sven
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu <mailto:Gretl-users@lists.wfu.edu>
http://lists.wfu.edu/mailman/listinfo/gretl-users
6:53 p.m.
New subject: R: R: R: R: R: Restriction on model coefficients
Am 09.02.2016 um 19:50 schrieb Carlo Maria Petrangelo:
So, I do not know if I had to write the commands in the exact way
Professor Cottrelldid, but I did it since I do not know the command and
I did not get theratio of the process (I'm sorry, but I'm not a math
student and in this case I just need the results and I can'tspend time
in understandingthat).
So I ran the commands below and I got an error:
?wage<-system
? equation w 0 w_1 lnGDPdeflator lnGDPdeflator_1 time;0 w_1 w_2 w_3
lnGDPdeflator_1 lnGDPdeflator_2 lnGDPdeflator_3
? end system
? function matrix my_restrict (const matrix b)
> matrix v = (b[4]-b[2])/(1-b[3]) + 0.012846691
> return v
> end function
?restrict wage
? rfunc = my_restrict
? end restrict
La statistica richiesta non è disponibile(Which in englishshould
soundslike"The statistic request is not available")
Not 100% sure if this is the cause, but try to move the function
definition to the top, so that the 'restrict' block comes right after
the system block.
hth,
sven
7:29 p.m.
New subject: R: R: R: R: R: Restriction on model coefficients
On Tue, 9 Feb 2016, Carlo Maria Petrangelo wrote:
So, I do not know if I had to write the commands in the exact way
Professor Cottrell did, but I did it since I do not know the
command and I did not get the ratio of the process (I'm sorry, but
I'm not a math student and in this case I just need the results
and I can't spend time in understanding that).
So I ran the commands below and I got an error:
?wage<-system
? equation w 0 w_1 lnGDPdeflator lnGDPdeflator_1 time;0 w_1 w_2 w_3
lnGDPdeflator_1 lnGDPdeflator_2 lnGDPdeflator_3
? end system
? function matrix my_restrict (const matrix b)
> matrix v = (b[4]-b[2])/(1-b[3]) + 0.012846691
> return v
> end function
? restrict wage
? rfunc = my_restrict
? end restrict
La statistica richiesta non è disponibile (Which in english should sounds
like "The statistic request is not available")
OK, here's what's happening. The variant of "restrict" where you
test a nonlinear restriction via the delta method does not support
the production of restricted estimates -- it wants an already
estimated system to which to refer. So something like the following
will work:
<hansl status="incomplete">
# define your restriction function
function matrix nl_restrict (const matrix b)
matrix ret = <whatever>
return ret
end function
# define your equation system, as in
wage <- system
equation <whatever>
end system
# estimate the system
estimate wage method=tsls
# then test the nonlinear restriction
restrict wage
rfunc = nl_restrict
end restrict
</hansl>
This will give you a Wald test of the nonlinear null hypothesis
embedded in your "nl_restrict" function; it will not give you the
restricted estimates. If you really want those, you will have to go
back to the suggestions made by Jack Lucchetti some days ago.
However, you might get "lucky": if the null is soundly rejected then
you won't need the restricted estimates ;-)
Allin Cottrell
9:05 p.m.
New subject: R: Restriction on model coefficients
Thank you so much Sven! The method you have proposed is too difficult for my needs
unfortunately.
I'm going to ask to my professor for an help.
Kind regards
Carlo Maria Petrangelo
Il giorno 04 feb 2016, alle ore 21:26, Sven Schreiber
<svetosch(a)gmx.net> ha scritto:
> Am 04.02.2016 um 19:39 schrieb Carlo Maria Petrangelo:
>
> Instead, I need that in the estimation of the equation (the one estimated in
> point 1), the coefficients would be estimated using the restriction, that is
> using the restriction during the estimation.
You're right that with tsls the restricted estimates aren't printed; I
wasn't aware of this and I don't know why. (It's mentioned in the docs:
"If the original model was estimated via OLS then the restricted
coefficient estimates are printed;", but that doesn't explain why. Maybe
Allin can elaborate.) Similarly with the "--full" option which only
applies to OLS (and VECMs, but that's a different story). So you also
cannot use the $coeff accessor after the restriction.
One not very satisfactory workaround would be to split the two stages of
two-stage-least-squares, so that you get the second stage mechanically
as OLS where you can retrieve the restricted results. But this is really
lame I admit.
sorry,
sven
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
3241
days inactive
3246
days old
19 comments
4 participants
participants (4)
-
Allin Cottrell -
Carlo Maria Petrangelo -
Riccardo (Jack) Lucchetti -
Sven Schreiber