Implement interaction term in ivregress command

[Marc Pelow]

Hi.

consider the following simplified OLS baseline regression:

y = ß₀ + ß₁X₁+ß₂X₂+ß₃X₃ + e

In a second model, a dummy variable X₄ is included which interacts with X₁:

y = ß₀+ß₁X₁++ß₂X₂+ß₃X₃+ß₄X₄+ß₅(X₁*X₄) + e

The robustness of the findings of both models should be checked using a 2SLS model with an instrument Z. The potentially endogenous variable is X₁.

The command for the baseline model, i.e., without the interaction term is:

ivregress 2sls Y X2 X3 (X1 = Z), first vce(robust)

I am not sure how to implement the interaction term in the -ivregress- command though. Do I have to build the interaction manually like so?

gen Interaction  = X1*X4
gen Interaction_IV = X4*Z

2sls with interaction:

ivregress 2sls Y X2 X3 (X1 Interaction = Z Interaction_IV), first vce(robust)

I read through this thread which gives me the feeling that the ivregress command is correct, neverless my questions would be:
1.) Is there any mistake I made in the ivregress command?
2.) I was wondering if there is a way to directly implement the Interaction term in the ivregress command without the need to create the interaction manually.

Many thanks

[Andrew Musau]

On #1, x4 is missing from your command. On #2, see

help fvvarlist

[Andrew Musau]

By the way, this statement

The robustness of the findings of both models should be checked using a 2SLS model with an instrument Z. The potentially endogenous variable is X₁.

makes me think that this is homework. If so, read our policy on that in #4: https://www.statalist.org/forums/help#adviceextras