Invalid instruments for 2SLS

[Will OLeary]

Hi guys,

I have some IVs that are stongly correlated with my endogenous variable but when I use an "estat overid" command I get a p value of 0.000 and so the null hypothesis of all instruments being valid is rejected. Can I still use these IVs in my paper and explain why this might be a problem in the results explanation part of my paper or would this be detrimental to my study?

thanks

[George Ford]

Might try fewer IVs and see if certain of them are causing the problem. More IVs is not always better.