I suggest the mathoverflow policy: no obvious homework questions, ever.
I clarify: the question must always be formulated or refer to something that is not already answered in an undergraduate or first year graduate level textbook.* A suggestion on how to generalize even an undergraduate result is fine; the undergraduate result itself, in general, is not fine, but clutter in a very harmful way.
Consider this: if obvious homework questions are allowed, most of the site will be homework.
The problem is that first year macro- or microeconomics (like first semester calculus) are general education curriculum courses all over the world. Every student takes them. Later they take their phone, search, and post questions that bore them here, so that we too can be bored.
Which working scientist or any individual at all, who gave serious thought to a question and already looked at the literature, would spend time to post interesting questions and discuss answers in a place where these questions and answers would be typically bumped down off the front page, within half an hour, by a dozen homework questions?
*TO CLARIFY: What I am thinking of by "anything in an undergraduate textbook" is any "question" (since that is what we are talking about allowing or removing) found in it. Because this is generally an object from the homework questions section inevitably supplied by each textbook author.
I certainly don't mean to suggest any concepts mentioned in a textbook are therefore subjects of inquiry excluded from discussion on the site.