For students, some textbooks publicize no solutions, some just half (either odd or even numbered exercises), and some all. How's this decided? Does the author have the final say? Can publishers override an author? A debate between both?
I'm not asking about instructors' solution manual that divulges solutions just to instructors.
Usually the author or authors has the final say. Presumably it is the authors or associates of the authors who will have the burden (at least in part) of providing solutions, and this is NOT indeed a very gratifying task.
In some cases solutions are available but have to be translated to English.
Some authors actually believe that producing solutions will inevitably lead to these being made public, and will eventually bring down the value of the work. (I disagree: nobody suggests that you need to take problems from same the textbook as the one used in class.)
The need to produce solutions also varies with the level of the textbook. Publishers may already have staff or freelance contributors to write end-of-chapters questions on calculus or basic physics, but that’s unlikely to be the case in an upper-level text.
I would be surprised if any publisher required a particular outcome or, especially, overrode an author's decision. It would make for an awkward relationship with the author.
They might, however, make suggestions, since the commercial value of a book might be affected by the decision. But the commercial value and the educational value might be at odds. I've never had a discussion with an editor about this subject and have published with several.
I do know of one case where a grad student was paid to produce solutions. I believe the funding came from the publisher, but it has been too long ago to remember details. But this is something an author could request of the editor. However, such payments usually come out of future royalties to the author, I'd expect.
