Using IV

[hussein bataineh]

Hi,

I need some advice on using instrumental variables (IV). In my case, I have three first-stage regressions using probit models, and the second-stage regression is an OLS.

Could you please guide me on how to perform this correctly?

Thank you!

[Jeff Wooldridge]

You should show at least your model. You have three endogenous explanatory variables, each of which is binary? You can estimate a probit for each of them, where you include ALL exogenous variables in each probit. Then use the probit fitted values from the three first stages as instruments in the second stage -- not as regressors, but as instruments.

[Seungmin Lee]

You keep posting the same questions (link1) (link2) (link3), and many people answered to your questions (including mine, which is same as what Prof. Wooldridge suggested above).
If you could share with us why the earlier answers didn't work for you, it would be helpful to answer your question.

[hussein bataineh]

Thanks professor Jeff Wooldridge and Seugmin

This is my main model

asdoc reg market_adjusted_return_w hard_final_Exact_new csopresence1 FreezeXCSO Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.year i.ff_12 , robust cluster (id) nest replace drop(i.year i.ff_12 ) dec(4) save(qqqq)


actually we assume that these three binary variable are exogenous (hard_final_Exact_new csopresence1 FreezeXCSO ) . we assign Mem_w as instrumental variable for for hard_final_Exact_new . CSO_Percentage as instrumental variable for csopresence1 and second_instrumatial as instrumatial for FreezeXCSO


This is what i did . I'm not sure if this is correct


probit csopresence1 CSO_Percentage Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.year i.ff_12 , nocons robust cluster (id)
predict X1hat, xb
gen X1prob = normal(X1hat)


probit hard_final_Exact_new rate_w Mem_w Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.year i.ff_12 , nocons robust cluster (id)
predict X2hat, xb
gen X2prob = normal(X2hat)


probit FreezeXCSO second_instrumatial Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.year i.ff_12 , nocons robust cluster (id)

predict X3hat, xb
gen X3prob = normal(X3hat)



asdoc ivregress 2sls market_adjusted_return_w Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.year i.ff_12 ( csopresence1 hard_final_Exact_new FreezeXCSO = X1prob X2prob X3prob) , robust cluster (id)


estat first, forcenonrobust
estat endogenous hard_final_Exact_new csopresence1 FreezeXCSO