ivregress with a categorical endogenous variable

[Samer Matta]

Hi all,

I am have model where my dependent variable is the sales growth of firms (i.e. a continuous variable) and one of my independent is possibly endogenous (firm perception towards political instability). However, this endogenous variable is a categorical variable that takes the value from 0 to 4.

I was wondering, whether the “ivregress” command still works properly given that I have an endogenous variable that is categorical and not continuous?

I would appreciate any thoughts/guidance on that issue.

Best regards,
Samer

[Ariel Karlinsky]

I see several probably relevant results when googling this question, have you had a look there?

https://www.google.co.il/webhp?sourc...e%20endogenous

[Samer Matta]

Dear Ariel,

Thank you for the reply, that might help.

Best,
Samer

[Vanessa Owusu Ansah]

Samer Matta, do you mind sharing how you went about your problem? I am facing a similar problem

[Jeff Wooldridge]

Vanessa: How will you enter your endogenous variable, say y2, into the model? If you just put in y2 then nothing really changes. You can use ivregress 2sls provided you have at least one instrument, z2, for y2. But the coefficient might be hard to interpret because y2 might only have ordinal meaning. If you want to put in separate indicators for each value of y2, that's harder. I have a suggestion, but I'll wait for your response.