Dear all,

Greetings. I meet the below problems and hope someone could give me hints. Thank you very much for your replies.

I had two endogenous variables (Var1 Var2) and two instruments(IV1 IV2). The two-stage instrument regression is:
ivreg2 y (var1 var2= iv1 iv2) $control, vce

The first-stage regressions for each endogenous variable are:
(a) reg var1 iv1 iv2 $control, vce
(b) reg var2 iv1 iv2 $control, vce

In the first stage regressions, both instruments are significant and F stats are greater than 10. But in the second stage, the Kleibergen-Paap rk Wald F statistic is smaller than 10. Why would this happen? and I wish to know whether weak instrument problems bias my results.

Regards,
Haiyan