wrong SE in dsregress with cluster() option

[Adrien Bouguen]

Hi statalisters,

it looks like when you run the following model:

dsregress Y T, controls(X1 X2...) cluster (clustervar)

dsregress does not generate an error and simply provide the double lasso estimator without correcting for clustering.

The correct code should be (as you probably know):

dsregress Y T, controls(X1 X2...) vce(cluster clustervar) 

Yet, why does stata authorize the first coding? In a classical regression reg Y T, cluster(clustervar) and regression reg Y T, vce(cluster clustervar) provide the same SE.

I think this should not be allowed since it is a source of error.

What do you think?

[Jeff Wooldridge]

This looks like an oversight. Since dsregress is relatively new, I would expect cluster (clustervar) to given an error -- as with something like xtivreg, which only recognizes vce(cluster clustervar). At a minimum, if cluster (clustervar) is to be allowed, it should return exact the same estimates and SEs as vce(cluster clustervar).

[Adrien Bouguen]

Ok that my opinion as well. I will flag the issue to stata. Thanks

[Di Liu (StataCorp)]

Hi Adrien,

Thank you for pointing this out. -dsregress- should error out in the first case because -cluster()- is an undocumented option.