Hello. I'm wondering if anyone has insights on the following. I'm trying
to obtain robust clustered standard errors after the "sureg" command and
apparently I'm not using the "_robust" programming command correctly. I say
this because I get unbelievable robust standard errors (many z-values are
greater than 100). The commands I'm using are:
sureg (s_off p2 p3 d97 age educ hhsize fsize) (s_for p2 p3 d97 age educ
hhsize fsize)
mat D = e(V)
mat b = e(b)
predict double e1, equation(#1) residuals
predict double e2, equation(#2) residuals
_robust e1 e2, cluster(panelid) minus(8)
ereturn display
