I'll just start by saying that writing a good abstract is difficult.
Difficult and frustrating and humbling.
OK, and there you have all the reasons not to write an abstract.
In favor of writing an abstract, or at least trying to, are many factors.
I find that I'm often pushing these factors on those who resist writing the abstract.
First of all, there's the very obvious fact that you need an abstract.
Academic papers have abstracts. If you're writing an academic paper--especially a thesis or a dissertation--you need to have an abstract.
Secondly, the abstract helps the author see the overall structure of the work (the forest) instead of just the details (the trees).
Third, it only takes about an hour to write a draft of an abstract, so it's a good exercise.
Fourth, it's easy to get feedback on an abstract, even detailed feedback, in a way that you can't if you have a fifty page draft.
Fifth, the abstract helps the author see the overall structure. What? I said that already? Of course I did, but it's so important it bears repeating. Authors get lost in the details and lose sight of the general intention of their project, thereby wasting time and effort rather than moving towards the goal of completing the work.