IV first stage regression

[Catherine Lee]

Hi,
I was asked to do the first stage (only) in IV regression and then get a table about it. Here are the issues I faced which also made me confused:

1. I regressed y (which is a dummy variable) on all the control variables to get one of the coefficients which was useful for my task. And I think I need to use all the control variables in the regression otherwise I got wrong coefficient which I want.

2. However, after regression the F-stat is missing. I guess the reason is that there are too many variables/clusters/predictors. But if I delete other variables, the result will be wrong.

3. Since the F-stat is ".", I used code test to check the instrument variable. Nevertheless, I got the wrong number of F-stat.


I cannot post data or the table here so sorry for the inconvenient. And I just want some suggestions about what can I do to solve this problem. Hope anyone can help me.


Thanks a lot.
C.L.

[Carlo Lazzaro]

Catherine:
if what above relates to a home/class assignment, please see https://www.statalist.org/forums/help#adviceextras #4.
That said, see -ivregress- entry in Stata .pdf manual and -help j_robustsingular-.

[Christos Chrysanthakopoulos]

Hi, i would like to know, if is it possible to see the first stage after an IV estimator using the new "locproj" command. By typping just "first" after the following command (as in a typical xtivreg) doesn't work.

Here is an example of my code:

locproj Y, shock(shock1) controls(x1 x2 x3 i.year) ylags(2) hor(6) conf(99) transf(cmlt) zero ttitle(Year) met(xtivreg) instr(l(1/3).shock2) noisily

Thanks!
CC