You're mixing up two fundamentally different topics: the homework tag itself and the topic how to deal with the homework questions themselves.
The common arguments for and against the homework tag are roughly:
- Allows users to ignore all (correctly-tagged) homework questions
- Might indicate to users that not a full solution should be given
- It is a meta tag, it does not tell us what the question is about, only what type of question it is
- Such meta tags are used very often if they are tolerated, so they tend to end up in the HTML title of the question (where they are completely useless and replace other, more useful tags)
- Meta tags tend to be useless for searching, as they don't provide useful information
- Allows users to only tag questions with homework, and omit any actually useful tag
about the topic of a question
- Causes noise in comments and additional potential for conflict if the tag is applied by other users than the asker, if they disagree whether a specific question is homework or not
- Allowing users to hide a tag that contains a disproportionate amount of low-quality questions is just a way to hide the actual problem and makes it harder to actually improve the quality of the site.
Meta tags have been strongly discouraged for a while, homework was just the last major one on SO to survive for a bit longer. There are still a few sites that use the homework tag like Physics, but the SE-wide consensus seems in opposition of the tag now.
My impression is that here on MSE the number of users that filter out homework is rather high, which would explain why there is no serious effort to remove the tag here. MSE also seeems to have a greater distance to the SE network as a whole, and diverges in some aspects from the network-wide consensus. My suspicion is that how closely a site follows the general SE consensus is determined by a large part on how much previous experience with SE the initial core community on a new site has. Once some practices are established, it is much harder to change them.
The other part of this problem is the question on whether to allow homework questions at all or where the minimal requirements in quality and effort should be set. This is a controversial topic SE-wide, though here on MSE it tends to be discussed with too much hyperbole and provocative language, and that tends to make the discussions more personal and less constructive.
The position that seems to have the strongest support SE-wide in my observation is that the problem should not be framed in terms of homwork, but of low-quality and low-effort questions. SO for example does not allow "Gimmeh teh codez!" questions that just expect the answerer to do their work for them. This distinction is reasonably easy to make on SO, it is a bit fuzzier in many other topics in my opinion.
Pure "assignment dump" question that consists of nothing but a copy-pased problem are not allowed on any of the scientific sites. That is one aspect where there is a reasonable consensus SE-wide in my observation.
So I don't think MSE is that far from the SE consensus, it is pretty much going through the same discussions as SO did and the arguments are rather similar.