It's hard to say. The relationship between uniparental and linguistic formation and distribution is not clear-cut. Then again, with that being Ethiopia and the extended passage of time, it would not be strange to have several identity turn-overs you had to uncover before getting the root. The resolution is low. For all we know, it could even be a haplogroup associated with a dead Somali sibling branch that we autosomally absorbed. There is also our assumption, important nonetheless, to associate sub-clades with linguistics. Linguistics might have unique misaligned, yet often paralleled characteristics relative to genetic markers across time and space.
Autosomal DNA signatures form stronger associations with linguistics within the reasonably stable context and contrasted delineated landscape than haplogroups. Also, haplogroups are simpler to interpret downstream, given the earlier sub-clade formations lacking complete clarity.
Oromo is likely an explanation. I prefer Somali as it appeals to convenience and makes for an engaging conversation with no illusion of the bias present. Given what I touched upon in the above text, if E-Z21175 was proto-Oromo, population incoherency created it as such that I am sure that, if Somalis had a comprehensive Y-DNA coverage, we would be more consistent with that haplogroup than other Ethiopian groups that only show in irregular ethnic-geographic distribution.
The sub-clade E-Z21175 might also indicate that the carriers were in early point agriculturalist with less uniformity. That is, however, pure speculation. But it is important to emphasize that our ancestors were quite flexible than average with specializations and, at the same time, very adaptive and economically complex, very impressive in that regard.
