I think the point was that there are ways to have nobody in shelters without it being a well spent budget, and practically we don't know that there is a way to

I think the point was that there are ways to have nobody in shelters without it being a well spent budget, and practically we don't know that there is a way to spend realistic levels of funding such that no one is in shelters, so while it may be true of some ideal policy it might well not be true of the best real policy. I think it's a (marginal) nitpick over your wording, not intended as a defense of the metric you were criticizing, though of course dragonwriter can correct me on that if needed.