Calculating probabilities for Multinomial Logit with Fixed Effects (femlogit)

[Sam McCaw]

Dear Stata Gurus,

I am using using the Multinomial Logit with Fixed Effects (femlogit) program to estimate a model to predict policy choices of different states.

My policy outcome variable has 5 choices, coded from 0 - 4. I am using 0 as the base outcome.

Once I run the femlogit command, my model converges but I am not sure how to get the probabilities for each outcome (as one would do in a simple mlogit model using the predict command and specifying each outcome).

Grateful if you could help with regards to the command code for calculating the probabilities of each outcome for the femlogit.

Thanks.

Sam

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Knowing exactly what you ran and being able to replicate the issue will make it easier for us to help you.

With a user written program, it is often not clear how to do such things that are easy in normal Stata procedures. This estimator does use GSEM so you might be able to program it directly in GSEM. It also saves something called e(out) - I'm not sure what is in this. Alternatively, contact the author.

[Sam McCaw]

Many thanks for the info, Phil. I reached out to Dr. Klaus regarding the femlogit model and he confirmed my suspicion that we cannot estimate the effects on outcome probabilities with this program, similar to the problem with xtlogit, fe. Dr. Klaus discusses this problem in more detail in his Stata journal article, “femlogit—Implementation of the multinomial logit model with fixed effects”.