issues with clustering in ivreghdfe - insufficient observations

[Nikhil Kumar]

I needed some advice on clustering while using ivreghdfe.

My code is -

ivreghdfe yvar (ever_win = true_treat), absorb(strata gender#batch city#batch yrs_school#batch age#batch mult_apply_hh#batch) vce(cluster unique_id)

I have 78,670 observations with about 42000 distinct unique_ids. The FEs in absorb() amount to about 3000 dummies.

The error I get is

 insufficient observations

.

When I remove the vce(cluster) option, the code runs properly. Could you help out with what I must be doing wrong here?

[Joro Kolev]

Try instead of
vce(cluster unique_id) cluster(unique_id)

[Nikhil Kumar]

I get the same output -

insufficient observations

[William Lisowski]

You have 78,670 observations and "about 42,000" clusters (distinct unique_id's). So you will about 5,300 or more clusters with a single observation. I suspect that is the source of the problem.

One way of testing this is

by unique_id: drop if _N<2

and then see if what remains can be fit using your command. If not, try "a larger value of 2" as it were.

[Nikhil Kumar]

Thanks William Lisowski . So, is there no way to do IV estimation with large number of FE but few clusters only?