Hi,

I am running two regressions using subsets of the data. I wanted to find if the coeffiecients are significantly different from each other

Code:
reghdfe dep time effect interaction_time_effect if size ==0, absorb(fe year) vce(cluster double_cluster)
est store small

reghdfe dep time effect interaction_time_effect if size ==1, absorb(fe year) vce(cluster double_cluster)
est store large

test [five_mean]interaction_time_effect = [six_mean]interaction_time_effect
I am getting the following error. "five was estimated with cluster(double_cluster). re-estimate without the cluster() option, and specify the cluster() option with suest."

I wanted to know if 1) How to estimate the differences in coefficients with clustering 2) If putting the cluster() option in suest, would it compare the standard errors similar to what it would with the use of cluster() in the regression.


Thanks