Guimaraes et al. (Rev Econ Stat, 2003, 85/1) describe the conditions under which the results from poisson regression models and conditional logit models are equivalent. I am trying to find out whether this result also holds i) for fixed effects panel data; ii) for negative binomial models; and iii) (for completeness' sake) fixed effect negative binomial models. It would be extremely helpful if you could point me to some relevant literature, if it exists.

  • 1
    $\begingroup$ Can you give a full citation? Off-the-cuff I doubt iii holds, as the estimate of the dispersion parameter makes the negative binomial panel model not really "fixed effects" - see Fixed–Effects Negative Binomial Regression Models (Allison & Waterman, 2002), PDF here, for discussion. $\endgroup$ – Andy W Aug 7 '14 at 12:47
  • $\begingroup$ @AndyW The full citation is: Guimaraes, P., O. Figueirdo, and D. Woodward, 'A tractable approach to the firm location decision problem', the Review of Economics and Statistics, 2003, 85(1): 201-204. doi:10.1162/003465303762687811 $\endgroup$ – Matthijs Aug 7 '14 at 13:12
  • $\begingroup$ @AndyW And if I understand Allison&Waterman correctly, I could estimate unconditional fixed effects models by including dummy variables in the specification of the mean. The incidental parameter problem might exist, but seems to be small in this case. $\endgroup$ – Matthijs Aug 7 '14 at 13:20
  • $\begingroup$ Well that is the recommendation by Allison in that particular paper, supported by a simulation study with a fairly large cross section (100) and a small number of repeated observations (2). It is unclear to me if the same advice would hold in different circumstances, and/or with more interesting unobserved fixed effects. $\endgroup$ – Andy W Aug 7 '14 at 14:25

Check out:

I think it shows the connection to the negative binomial.

  • $\begingroup$ Welcome to the site, @user61841. Would you mind expanding on the connection shown there a little? The citation is helpful, but a little more information would be nice. At present, this is more of a comment than an answer. $\endgroup$ – gung - Reinstate Monica Dec 1 '14 at 15:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.