Exactly why are we unable/unwilling to implement this feature?
There's no "we" when it comes to implementing features. The entirety of the software that runs Stack Exchange sites is written by Stack Exchange, and it is essentially the same software that's in use for over 150 different sites. We, as the user community of one of them, can request features, but it is up to SE to decide whether those features are worth investing developer time into.
For a large chunk of the history of SE, the best way to request features for the software which cut across the full spectrum of sites has been as a feature request on Meta Stack Exchange (previously Meta Stack Overflow, until MSE was unfolded out of MSO in 2014). However, as you may be aware if you followed this recent fracas, it appears that SE feels that the feature-request-on-meta is not scaling well enough with the growth of the site base, user base, and code base - see in particular this thread.
For the feature request you're interested in, there are already MSE posts here and here with substantial support. So, with that in mind, it is perfectly legitimate to ask
Exactly why are Stack Exchange developers unable/unwilling to implement this feature?
and the short answer is that we don't know. As Tim Post writes in the thread linked above, the communication channels to and from SE on this type of request have not been working very well lately. There's good reason to hope that they're working on ways to improve those types of communication channels, but for the time being it's hard to know.
That specific example, for instance, is certainly a worthwhile goal, and if you get ahold of Adam Lear, Tim Post, or Shog9 (say) on the MSE chat, I suspect that they'll tell you that it's something that they would like to implement. However, it's not an easy thing to roll out: there are significant design questions and challenges (how should it work? how should it look like? are users going to understand it?) and there are significant performance challenges (the feature would represent a whole lot of extra calculations regarding who needs to get notified, over a very large userbase×questionbase, and that has a significant impact on server load) so it could be that it's just not feasible enough. Or it could be that it's potentially feasible but they don't feel it'd be useful enough to merit the investment. Or it could be that they disagree that it would be ultimately useful or in line with the site design goals.
Which of those is the case, though? at present: we don't know.