As a short follow-up, I just remembered that I posted something
similar about half a year ago
(http://www.stata.com/statalist/archive/2009-06/msg00045.html) - in
fact, I was somewhat more detailed back then.
So, in a sense, apologies for double-posting,. I would, howver, still
be interested in hearing any opinions (I didn't receive an answer back
then).
Best,
Nils
On Tue, Jan 5, 2010 at 2:03 PM, Nils Braakmann
<[email protected]> wrote:
> Dear all,
>
> I have a large panel of roughly 50,000 observations over 20 years. I
> model the effects of a time-varying variable on individual employment
> probabilities using -clogit- with the individuals as groups. To be
> specific, I use the following command
>
> - clogit work degree `indepvars', group(persnr) cluster(persnr) -,
> which runs fine.
>
> The trouble starts when I add year dummies, which leads to severe
> convergence problems (nonconcave likelihood for several hundred
> iterations, etc.). If I remember correctly, I had the exact same
> problem in a different context some time ago. I was able to use a
> work-around back then, which unfortunately isn't possible with my
> current problem. The strange thing is that this only happens with year
> dummies. For instance, if I add about the same number of region
> dummies the model converges just fine. Given that I've encountered the
> exact same problem twice, I ask myself: Am I just unlucky and this is
> a strange coincidence or do I miss something fundamental about
> conditional logits on panel data and time effects?
>
> Any input is greatly appreciated.
>
> Best,
> Nils
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
