IV Regression with multiple instruments

[Patrick Johannes]

Hey there,

I need to perform an iv regression using ivreg2 an i have multiple interaction terms. I need to instrument the esg score with president congress and state_government. Further I need to implement an interaction term between the instrumented esg and advertising
this is my code

 ivreg2 firm_beta_w (esg_single c.esg_single#c.advertising_lag1 = president congress state_government c.president#c.advertising_lag1 c.congress#c.advertising_lag1 c.state_government#c.advertising_lag1) diversification_lag1 RnD_lag1 advertising_lag1 leverage_lag1 capex_lag1 cash_lag1 size_lag1 earnings_variability_lag1 state_tax_lag1 i.industry_key i.year, cluster (company_key year) first

I do not think this is correct, since the results really suck.
Can anyone give me some advice?
Appreciate it
Thanks, Patrick

[Joro Kolev]

I do not see anything wrong with what you are doing.

If you have an endogenous regressor X, an exogenous regressor W, and an instrument Z, we typically by analogy instrument for X and X*W with Z and Z*W. Which as far as I can see is what you are doing.