Hello forum members,

i do have a multinomial logit model where one binary explanatory variable is endogenous (EEV). I am aware of solutions for the case of binary outcome variable and binary EEV, such as, e.g. bivariay probit models (see Wooldrige, 2010, p. 595 ff.) - which is conveniently implemented in Stata's biprobit function.

However, are you aware of any approach for the case of categorical outcomes (e.g. multinomial logit) and binary EEV? [Any help is appreciated]

Refs.
Wooldridge, J. M. (2010). Econometric Analysis of Cross Section and Panel Data. MIT press.