I) In 99 percent of the cases, the homework tag runs smoothly and as intended. In a few cases, OP does not agree with labeling his question with the homework tag; often it's a new user that hasn't read the wiki descriptionwiki description or the site policy, but sometimes OP is right and the homework tag is removed again from the question with no further discussion.
II) In the early days of Phys.SE, whenever there was posted a homework question without the homework tag, then the moderators would politely ask OP if this was homework or not? Typically, a couple of full answers to the homework were posted and accepted before the issue of whether or not to add the homework tag was settled. Thus OP would succeed in getting a homework solution from our site, which in other respect is mainly dedicated to just conceptional physics questions.
III) In the last year or so, the tagging practice was changed, and users with tagging privileges would now often add the homework tag whenever they felt it was appropriate without first asking OP. The answers were now typically just hints, advices, outlines, etc., instead of full solutions. So the homework tag clearly had the desired effect. Moreover, this new practice didn't seem to spur repeated back-and-forth edit war over the homework tag. Typically it became clear to everybody whether or not the homework tag would apply in a given case or not.
IV) If it ain't broke, don't fix it. I fear that by banning the homework tag, we in practice loose a useful user-based moderation option, and effectively turn this site into a homework help service.
V) Another aspect is that many of our readers use the homework tag to decide whether or not to read a post. This goes both ways.