Statalist


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

st: xttobit not converging log likelihood "not concave"


From   "Anthony A. Pezzola" <apezzola@uc.cl>
To   statalist@hsphsun2.harvard.edu
Subject   st: xttobit not converging log likelihood "not concave"
Date   Thu, 19 Jun 2008 19:05:01 -0400 (CLT)

I am attempting to fit a model using xttobit, however, I cannot get xttobit to
fit with even the most basic model: log likelihood is "not concave."

I have 9040 observations and 89 groups, with a minimum of 1, a maximum of
1252, and an average of 101 observations per group.

There is no time variable in the model

The dependent variable is truncated at 0 and 20, with 5% at zero and about 8%
at 20.

It should be noted that for some of the groups there is little or no variation
on the dependent variable.  There are at least 5 groups that have no variation
on the DV.

I have tried the following code:
       xttobit aec rca, i(cnae) ll(0) ul(20)

       xttobit - aec imp_pen p_m srca rca d_rca itt wage wt_unid polcon prod
polcon_prod, i(cnae) ll(0) ul(20)

and neither will converge after more than 400 iterations.

I have run -  xtreg aec imp_pen p_m srca rca d_rca itt wage wt_unid polcon
prod polcon_prod, i(cnae)

and received what logical estimations for the coefficients and a rho of 0.48.

Any help would be greatly appreciated.

-- 
Anthony A. Pezzola
(02) 354-7823
Profesor de Ciencia Política
Instituto de Ciencia Política
Universidad Pontifica Católica de Chile
Santiago de Chile

*
*   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