use IV but without Instrumented

[Yao Zhao]

I run ivregress 2sls y x1 x2 x3, robust

What this outcome is?

It's different than reg y x1 x2 x3, robust

[Andrew Musau]

You need to adjust the degrees of freedom in ivregress to obtain the same standard errors. Otherwise, the estimated coefficients should be the same.

ivregress 2sls y x1 x2 x3, robust small
reg y x1 x2 x3, robust