OLS vs. Lewbel instrument variable

[Maya Lani]

Dear all,

I used Lewbel instrument variable which is the command ivreg2h in Stata: ivreg2h dependent variable exogenous_variables (endogenous_varibale=)

It is a method that is best used when you cannot find an exogenous IV. However, my problem is that the estimation I got were almost the same as those of OLS...

This is puzzling because OLS estimates are expected to be biased due to reverse causality between my dependent variable and an endogenous regressor.


Can anyone please explain to me how it is possible to have the OLS and ivreg2h output almost the same.

Thank you so much,
Maya L.

[Phil Bromiley]

You didn't get a quick response. You'll increase your chances of a helpful answer by following the FAQ on asking questions - provide Stata code in code delimiters, Stata output, and sample data using dataex. Also, try to reduce the code you include to what is needed to demonstrate your problem. You should also have searched for the major terms in your question from the command window of Stata (findit majorterm) and in the subject index in the documentation (under index at the bottom of the list of documentation).

There are many possible reasons. You may not have endogeneity. Alternatively, the estimation procedure doesn't work for various reasons. There is a reason that most folks do not use heteroskedasticity-based instruments.

[Sayli Javadekar]

Hi Maya
The reason you get that is because you don't specify the endogenous variable.
You code should be

ivreg2h depvar exogvar (endogvar=)
eg: xi:ivreg2h norm_Wb_rhat hindu musl scst obc Age age2 (manedu=),small robust

Best
Sayli