IVREG - First-stage marginal effects

[Chiara Piazzo]

Dear all,

I am running a IV regression using ivreg;

ivregress total i.year i.month i.HHID c.income##c.income (intake = count) [pw=mult_new].

I can obtain the marginal-effects after the second stage with;

margins, at(income = (0(100)3000)) vsquish

.

Is it possible to obtain the marginal-effects for the first-stage? Any suggestions will be highly appreciate.

Sincerely,

Chiara

[Andrew Musau]

The first stage in ivregress 2SLS is the usual regression with the endogenous variable as the outcome, although your standard errors will be slightly off if you are using probability weights.

regress intake i.year i.month i.HHID c.income##c.income count [pw=mult_new]

[Chiara Piazzo]

Thank you so much! As you said, standard errors are only slightly different.

Sincerely,

Chiara