Angrist-Pischke F-Statistic Missing in xtivreg2?

[Leon Schmidt]

Dear all,

I have two endogeneous regressors and two instruments. I read that the Kleibergen-Paap rk Wald F statistic is in this case invalid and one should use the Angrist-Pischke F-Statistic. This statistic should be included in xtivreg2. (This is all according to http://www.mostlyharmlesseconometric...ivreg2-update/ )

However, I cannot find the Angrist-Pischke F-Statistic in my standard xtivreg2 output (i.e., I just typed something like

xtivreg2 var1 var2 (endo1 endo2 = iv1 iv2), fd

). Where can I find this statistic?

Thank you very much!

All the best
Leon

[Eric de Souza]

Angrist Pschke has been replaced by Sanderson-Windmeijer, but you need to add the "first" option.

xtivreg2 var1 var2 (endo1 endo2 = iv1 iv2), fd first

If you want the AP test type

matrix list e(first)

after the above command

[Leon Schmidt]

Thank you so much for the quick and very helpful answer Eric!