More particularly: Presentation variant is not a justification for inclusion in Unicode BUT prior encoding in another character set is.
Unicode sets a high priority on roundtripping. The idea is that if you take some data in any one character set X and convert it to Unicode, you should preserve all the meaning by doing this, such that you could losslessly convert it back to encoding X.
It's like the wordprocessor problem where users say they only want 10% of the features of a popular wordprocessor but it turns out each user wants a different 10% and so the only way to deliver what they all want is to deliver 100% of the features. Likewise, Unicode has all the weird features of every legacy character set which was embraced BUT it doesn't arbitrarily add new weird features, although you could argue that some of the work done for Unicode has that effect e.g. the way flags work or the Fitzpatrick modifiers.
If Unicode had insisted upon never encoding anything that might be a presentation feature, it'd be a long forgotten academic project that never went anywhere and we'd all be using some (probably Microsoft designed) 16-bit ASCII superset today.
In the case of U+1D42 Modifier Letter Capital W I was wrong about the cause, it was in fact specifically added on the rationale that for this purpose (phonetics) the presentation was semantic in nature, and so the plain text (thus Unicode) needed to preserve these symbols which could otherwise be handled by a presentation layer.
U+1D42 Modifier Letter Capital W was added in Unicode 4.0 as part of the Phonetic Extensions and Wikipedia provides a long list of Unicode committee paperwork regarding this: https://en.wikipedia.org/wiki/Phonetic_Extensions
You can see that initially it would have been numbered differently and then over the course of several drafts the proposal evolved until it was assigned U+1D42
