Testing if coefficients differ across subsamples in IV regression

[Lucy Ty]

Good Morning,

I am trying to determine whether my endogenous variable’s coefficient differs across two subsamples of my population. I have used previous forum answers to try and solve the problem myself.

1^st attempt to solve I received an error after the first line of code:

ivreg2 Y a c d (e = z1 z2 z3) if t==0, pscore(est1)
ERROR “option pscore() not allowed”
est store est1
est change est1, scorevars(est1)
ivreg2 Y a c d (e = z1 z2 z3) if t==1, pscore(est2)
est store est2
est change est2, scorevars(est2)
suest est1 est2

2^nd attempt to solve (change ivreg2 to ivreg) I received an error after the final line of code”

ivreg Y a c d (e = z1 z2 z3) if t==0, pscore(est1)
est store est1
est change est1, scorevars(est1)
ivreg Y a c d (e = z1 z2 z3) if t==1, pscore(est2)
est store est2
est change est2, scorevars(est2)
suest est1 est2
ERROR “ivreg is not supported by suest”

3^rd attempt to solve:

ivreg2 Y a c d (e = z1 z2 z3) if t==0
est store est1
ivreg2 Y a c d (e = z1 z2 z3) if t==1
est store est2
gmm est1: Y-{b1}a -{b2}c -{b3}d -{b4}e-{b0} ///
ERROR “must specify parameters() or nparameters()”
gmm est2: Y -{c1}a -{c2}c -{c3}d -{c4}e-{c0} ///
instruments(est1: a c d z1 z2 z3) ///
instruments(est2: a c d z1 z2 z3) ///
onestep winitial(unadjusted, indep)
test ([b4]_cons = [b4]_cons)

Please may you help me to understand why my previous codes have not worked and, if possible, suggest a way forward. Thank you so much.

[William Lisowski]

Reposted and responded to at

https://www.statalist.org/forums/for...ies-on-my-post