Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: poisson will not converge


From   "Rodrigo A. Alfaro" <ralfaro76@hotmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: poisson will not converge
Date   Wed, 24 May 2006 11:54:24 -0400

Scott,

What is "fip" variable? state? where did you get 55? It seems this variable 
is highly correlated with others in the model. A quick inspection could be 
helpful: -xtsum rp srfm age2 hgc hhd1 fip-. Moreover, is this panel 
balanced?

Rodrigo.

----- Original Message ----- 
From: "Scott Cunningham" <scunning@gmail.com>
To: <statalist@hsphsun2.harvard.edu>
Sent: Wednesday, May 24, 2006 10:27 AM
Subject: st: poisson will not converge


I am estimating a poisson with fixed effects at the individual level,
year effect, state effects and state-year effects.  There are
approximately 3 groups of 800 observations in the sample.  It will
not converge unfortunately.

I've posted this problem before, but previously I thought the problem
could be rectified if I were to acquire Stata SE (9.2) so that I
could use the -poisson- rather than the -xtpoisson- model with fixed
effects.  With Stata SE, I could expand the matrix size to 5000 and
thus use i.id to control for fixed effects, rather than -fe i(id)- in
the xtpoisson. My hope was that this would work better, but
unfortunately it is no closer.  Here's my output so far.  Any
suggestions?  I am sending this before letting it run for thousands
of iterations because in my experience, once it locks up and says
"not concave" then it won't improve with thousands of iterations.
What is causing this problem?  Poisson models are straightforward, so
there has to be something wrong here, right?  I will note that the
problem appears to be the extensive dummies included.

 xi:poisson rp srfm age2 hgc hhd1 i.year i.fip i.year*i.fip i.id,
cluster(id)
 > robust
i.year            _Iyear_1998-2002    (naturally coded; _Iyear_1998
omitted)
i.fip             _Ifip_1-55          (naturally coded; _Ifip_1 omitted)
i.year*i.fip      _IyeaXfip_#_#       (coded as above)
i.id              _Iid_6-7612         (naturally coded; _Iid_6 omitted)
note: _Ifip_2 dropped due to collinearity
note: _Ifip_35 dropped due to collinearity
note: _Iyear_2000 dropped due to collinearity
note: _Iyear_2002 dropped due to collinearity
note: _Ifip_4 dropped due to collinearity
note: _Ifip_5 dropped due to collinearity
note: _Ifip_6 dropped due to collinearity
note: _Ifip_8 dropped due to collinearity
note: _Ifip_9 dropped due to collinearity
note: _Ifip_10 dropped due to collinearity
note: _Ifip_11 dropped due to collinearity
note: _Ifip_12 dropped due to collinearity
note: _Ifip_13 dropped due to collinearity
note: _Ifip_17 dropped due to collinearity
note: _Ifip_18 dropped due to collinearity
note: _Ifip_20 dropped due to collinearity
note: _Ifip_21 dropped due to collinearity
note: _Ifip_22 dropped due to collinearity
note: _Ifip_24 dropped due to collinearity
note: _Ifip_25 dropped due to collinearity
note: _Ifip_26 dropped due to collinearity
note: _Ifip_27 dropped due to collinearity
note: _Ifip_28 dropped due to collinearity
note: _Ifip_29 dropped due to collinearity
note: _Ifip_32 dropped due to collinearity
note: _Ifip_34 dropped due to collinearity
note: _Ifip_36 dropped due to collinearity
note: _Ifip_37 dropped due to collinearity
note: _Ifip_38 dropped due to collinearity
note: _Ifip_39 dropped due to collinearity
note: _Ifip_40 dropped due to collinearity
note: _Ifip_41 dropped due to collinearity
note: _Ifip_42 dropped due to collinearity
note: _Ifip_44 dropped due to collinearity
note: _Ifip_45 dropped due to collinearity
note: _Ifip_47 dropped due to collinearity
note: _Ifip_48 dropped due to collinearity
note: _Ifip_51 dropped due to collinearity
note: _Ifip_53 dropped due to collinearity
note: _Ifip_55 dropped due to collinearity
note: _IyeaXfip_2000_38 dropped due to collinearity
note: _IyeaXfip_2002_2 dropped due to collinearity
note: _IyeaXfip_2002_38 dropped due to collinearity
note: _Iid_1817 dropped due to collinearity
note: _Iid_3799 dropped due to collinearity
note: _Iid_6767 dropped due to collinearity
note: _Iid_6934 dropped due to collinearity
note: _Iid_7074 dropped due to collinearity

Iteration 0:   log pseudolikelihood = -66141.901
Iteration 1:   log pseudolikelihood = -38218.672  (backed up)
Iteration 2:   log pseudolikelihood = -29999.973  (backed up)
Iteration 3:   log pseudolikelihood = -17459.693
Iteration 4:   log pseudolikelihood =  -9438.949
Iteration 5:   log pseudolikelihood = -5553.0749
Iteration 6:   log pseudolikelihood = -4394.8428
Iteration 7:   log pseudolikelihood = -4292.1682
Iteration 8:   log pseudolikelihood = -4286.6815
Iteration 9:   log pseudolikelihood = -4286.6698
Iteration 10:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 11:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 12:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 13:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 14:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 15:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 16:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 17:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 18:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 19:  log pseudolikelihood = -4286.6698  (not concave)
Iteration 20:  log pseudolikelihood = -4286.6698  (not concave)

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index