Hi all,
I am running an IV regression and i have the following issue:
- Say i have an endogenous variable X and an instrument Z. Both are continuous.
- I am sure to have non-linear effects and my strategy was to bin the endogenous variable X (capital returns). I was thinking about creating bins for my instrument as well (potential capital returns):
Say i have X divided into X1 (bin 1), X2, X3 and X4 and Z into Z1 (bin 1), Z2, Z3, and Z4.
I will have some bins which are positive and some which are negative.
I went through the IV chapter in Angrist's Mostly Harmless Econometrics but i'm still a bit lost. Has anyone there tried to do something alike?
Thank you all!
Best
Joao
I am running an IV regression and i have the following issue:
- Say i have an endogenous variable X and an instrument Z. Both are continuous.
- I am sure to have non-linear effects and my strategy was to bin the endogenous variable X (capital returns). I was thinking about creating bins for my instrument as well (potential capital returns):
Say i have X divided into X1 (bin 1), X2, X3 and X4 and Z into Z1 (bin 1), Z2, Z3, and Z4.
I will have some bins which are positive and some which are negative.
I went through the IV chapter in Angrist's Mostly Harmless Econometrics but i'm still a bit lost. Has anyone there tried to do something alike?
Thank you all!
Best
Joao
Comment