Dynamic panel data model for Negative binomial regression

[Ayushi Basoya]

Hi All,
I am using STATA 17 and working on a Panel data structure. My dependent variable is a count variable. Therefore, I am using negative binomial regression method as there is over dispersion in the data. I also want to include lagged dependent variable as an independent regressor in the model. Can anyone suggest the Stata command for combining dynamic panel model with negative binomial regression/distribution?

Thank you.

[Clyde Schechter]

You can incorporate a lagged independent variable into a model by using Stata's L operator. You must first -xtset- or -tsset- your data. See -help tsvarlist- for details.

For using the lagged dependent variable as an independent variable, you will need a different approach. If you have a large N small T data set, you may want to look at -xtabond-.

[Ayushi Basoya]

Hi Clyde, Thank you for your response. I have tried the -xtabond- command, but since my dependent variable is a count variable (number of migrants in a household) ranging between 0 and 6, -xtabond- does not seem to cater to that. Is there any other alternative approach I can use?

[Clyde Schechter]

Sorry, I'm afraid I don't know of one. Perhaps somebody else will respond.

[Ayushi Basoya]

No problem! Thank you for responding!