e(clust#) not working in reghdfe

[Isabel Martinez]

Hi,
I ran into a problem where I do not know whether this is a bug or whether I am doing something wrong. I would like to extract the number of clusters used in each regression and ultimately save it as a scalar. According to the help file, this number should be stored in e(clust#). But this always seems to be empty. Consider the following MWE:

clear all
sysuse auto

reghdfe price weight length, absorb(rep78) vce(cluster rep78 foreign)

di e(clust1) // should be 5, but returns .
di e(clust2) // should be 2, but returns .

di e(N_clust) // this works and correctly displays 2

di e(N_clustervars) // this works too and displays 2

What am I missing?

(I'm running Stata 16.1 on a Mac and use the most recent version of reghdfe.)

[Andrew Musau]

reghdfe is from SSC, as you are asked to explain (FAQ Advice #12). See

ereturn list

[Isabel Martinez]

The mystery has been solved (thank you Ercio Munoz for your Tweet!)

It should be e(N_clust#), not e(clust#) as stated in the help file.

[Andrew Musau]

It is no magic at all. Try #2, really! The broader point is that estimation results are stored in e().

[Isabel Martinez]

Andrew Musau : ok I see your point now, thank you. An explaining sentence re why -ereturn list- helps here would have made your original post more useful.