I am looking for an R package that can handle multinomial logistic regression and also supports offsets, as well as—ideally—clustered standard errors and fixed effects.

By offsets, I mean a control for differential time exposure in the dataset. In binomial logistic regression in R, I’ve included offsets like so:

`m1 <- glm(Outcome ~ Var1 + 
                  Var2 + 
                  offset(log(Months_In_Data)), 
                data = Data, 
                family = binomial)`

From my scouring online, it looks like there is a package that has this capability in Stata: femlogit (see here). Ideally, I am looking for an R equivalent.

So far, I’ve explored the following options:

**nnet::multinom()**: This function allows for multinomial logistic regression, but I am unable to figure out how to include offsets, clustered standard errors, or fixed effects.

**glmmTMB::glmmTMB()**: While this package supports offsets, clustered standard errors, and fixed effects, it does not have a multinomial logistic regression option.

feglm: This package, as per its documentation (see here), does not support multinomial logits. This is also covered in past questions, but without reference to offsets: see here.

mclogit: According to the documentation (see here), it handles random effects but not fixed effects.

Does anyone know of an R package that can fulfill these requirements for multinomial logistic regression? Or is there a statistical reason that I am not aware of for why offsets are not available for multinomial logistic regression?

Thank you in advance for your help!