Logit in 2 steps

[Umair KKhan]

Hi,

I want to model the following.

In the internationalization of the firm, it can decide to enter foreign market (Yes/No) and conditional on Yes, it can choose Acquisition or Greenfield investment.

Which model would be appropriate to handel this?

Thanks a lot for help.

Kind regards,
Umair

[George Ford]

Control Function may be sensible start.

https://www.stata.com/meeting/sandiego12/materials/sd12_baum.pdf

[Leonardo Guizzetti]

Whether you need to model this in two steps is not clear because you don't tell us your research question. It may be that you only care about modelling type of investment, in which case, you need only the subset that enters a foreign market.

Presuming you really do need to model this conditionally, a simple approach can be a so-called hurdle model. In its essence, it is precisely a 2 stage model. First, model the probability of entering a foreign market. Second, the likelihood function is augmented by then modelling investment type conditional on (on the subset defined by) the firm entering foreign markets.

[Carlo Lazzaro]

Umair:
welcome to this forum.
Just to top-off George and Leonardo's excellent advice , you may find what follow interesting: https://journals.sagepub.com/doi/pdf...867X0300300207.
Like all the papers dealing with hurdle models methodology, it heavily draws upon John Mullahy 's pivotal contribution: Mullahy, J. 1986. Specification and testing of some modified count data models. Journal of Econometrics 3: 341–365.

[George Ford]

I'm not sure Stata's hurdle model algorithms permit a dichotomous second stage (or hurdle models generally)?