2:14 p.m.
Hi,
I have count data with what appears to be poisson distribution of the dependent variable
in panel data. There is no automatic method for doing this in Gretl as I have understood
it, so I am using a count, poisson, model with cross-sectional dummies. Gretl deletes the
last dummy so it becomes a fixed effects model, so intuitively, it looks as if the model
should be ok. Do you agree? One problem is that Gretl fails to include an overdispersion
(Chi-square) test, so I am not sure the poisson model is valid. Can you help me with this
too?
Best regards,
Lars Ahnöand
2:55 p.m.
On Mon, 5 Dec 2022, lars.ahnland(a)outlook.com wrote:
Hi,
I have count data with what appears to be poisson distribution of the
dependent variable in panel data. There is no automatic method for doing
this in Gretl as I have understood it, so I am using a count, poisson,
model with cross-sectional dummies.
You may want to use the FEPoisson package, for estimating Poisson models
in panel datasets with fixed effects. Fortunately, the Poisson model is
one of those rare cases in which you can concentrate out the individual
effects and get ML estimates straightforwardly.
One problem is that Gretl fails to include an overdispersion
(Chi-square) test, so I am not sure the poisson model is valid. Can you
help me with this too?
The overdispersion test should be printed automatically with the Poisson
command; I'm not sure if the same test would be valid with the
fixed-effects estimates. Can anyone provide an informed opinion and/or
bibliographic references?
-------------------------------------------------------
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
-------------------------------------------------------
11 a.m.
Big thanks! The offset variable should be "population" as I have understood it.
But I don´t really understand how to implement it, do you?
Best,
Lars
10:35 p.m.
On Wed, 21 Dec 2022, lars.ahnland(a)outlook.com wrote:
Big thanks! The offset variable should be "population" as I
have understood it.
But I don´t really understand how to implement it, do you?
What the offset variable should be depends on how your model is specified
and what aims to describe; in fact, it may be the case that you don't need
an offset variable at all.
If you need a literature pointer to clarify the matter, the concept is
briefly but clearly explained in Greene(2012, 7th ed.) in Section 18.4.5,
at the start of page 808.
-------------------------------------------------------
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
-------------------------------------------------------
1:19 a.m.
On Tue, Jan 10, 2023 at 5:35 PM Riccardo (Jack) Lucchetti
<p002264(a)staff.univpm.it> wrote:
On Wed, 21 Dec 2022, lars.ahnland(a)outlook.com wrote:
> Big thanks! The offset variable should be "population" as I have
understood it.
> But I don´t really understand how to implement it, do you?
What the offset variable should be depends on how your model is specified
and what aims to describe; in fact, it may be the case that you don't need
an offset variable at all.
If you need a literature pointer to clarify the matter, the concept is
briefly but clearly explained in Greene(2012, 7th ed.) in Section 18.4.5,
at the start of page 808.
A little follow-up to Jack's post: (1) "Greene" = William H. Green's
"Econometric Analysis", and (2) there's a brief discussion of the
poisson offset variable in Section 38.9 ("Count data") of the Gretl
User's Guide (available under the Help menu in gretl).
Allin Cottrell
762
days inactive
799
days old
4 comments
3 participants
participants (3)
-
Cottrell, Allin -
lars.ahnland＠outlook.com -
Riccardo (Jack) Lucchetti