Am I right to consider /΄veikənt/ and /΄veikəns/ in those words as variants of one and the same root morpheme in Modern English. But it makes me hesitate in my morphemic division if we take for analysis preside - presid-ent - presid-ency (comp. also agent - agency, pregnant - pregnancy and also pirate - piracy, private - privacy, literate - literacy).
I think that the unit vaca could well be a morpheme in modern English.
Take the words vacant, vacancy, vacation, evacuation, vacuum, etc. They all have the vaca/vacu unit, which always pertains to emptiness in some way.
I think that since the words are related by emptiness, the original root is less important than how the morpheme is currently used in English.
It is, therefore, an allomorph since the morpheme varies phonologically, but its meaning remains the same.
I would say no; those words are not perceived as composed from two morphemes, in English.
To talk of allomorphs, the words should be each composed by two morphemes, one of which is an allomorph and maintains the same function in both the words.
Supposing that the words are composed from vacan- as root, and -t/-cy, vacan cannot be the allomorph, as
- the pronunciation of vacan in both the words is the same (/ˈveɪkən/)
- vacan is written in the same way in both the words
-t and -cy cannot be allomorphs as the existing suffixes -t and -cy have two different functions.