I would rather have seen the existing tag definiton amended rather than a new tag created, especially as it it already used in conjunction with DNA questions for just this purpose (of the 18 questions tagged with ethnicity, 50% are also tagged with a DNA tag). We should probably have edited the tag wiki some time ago.
I doubt we need the granularity that the two versions provide (and if we do, the existing combination of ethnicity and dna would provide it.
By acting hastily when few people are around we have now made a re-tagging rod for our own backs (posters will reach for whichever tag they see first and not appreciate the fine and unnecessary distinction).
However, if it is decided to go ahead with the unnecessary new tag, somebody should sort out the mistagging elsewhere.
Edited to add:
Given that the community has previously agreed addressing all the 'What % ethnicity am I if my parents/grandparents were from ethnicity X' by using a canonical question (Determining ethnicity according to ancestry?) to close them as duplicates;
...and discussed having a canonical question that can be used to address the 'I've "done my DNA" and it says my ethnicity is X. Wah! How can it be X?' aka 'Understanding Ethnicity estimates from DNA testing' (which nobody has got around to creating yet)
... what we're currently left with is a (small) set of questions -- currently 5 -- about analysing ancestral data, or more complex ethnicity estimate questions (and yes this number will grow). We don't have as far as I can tell any "real" ethnicity questions that fir the existing tag definition.
So, some questions (to which I don't have answers yet):
Do we need the experts to cast an eye of all the 'Ethnicity estimate' questions to make sure that they don't require an expert answer (not just closing as duplicates)? In which case they will have to monitor ethnicity anyway.
Can we work having ethnicity dna and ethnicity-dna, cope with the confusion it will cause to non-expert users (they'll not read the tag wikis even if they all existed, and will choose a random single tag or all of them and either approach makes the tags useless for monitoring) and live with the ongoing retagging workload.
I'm still thinking this through.