Dear StataList users,
I have the following problem. I want to estimate a linear regression
model imposing some restrictions and clustering on person identification
number. For example:
Y= alpha+beta1* X1+beta1
*X2+epsilon
I would like to impose the following restrictions:
beta1=beta2 but I would also
like to cluster on the person identification number
I can do the following:
constraint define 1 X1=X2
cnsreg Y X1 X2, constr(1)
but I do not know how to use the cluster option since cnsreg does
not support it.
I also tried to use STATA 9 and I computed the following
cnsreg Y X1 X2, constr(1)
vce(robust)
but I get an error message which says that robust option is not
allowed
Finally I have seen that in STATA 8 there is the possibility to use the
command "linest":
reg Y X1 X2, cluster(pid)
linest , c(1)
Does it compute exactly what I want?
Does anybody can help me?
