To emphasize, this question is about downvotes on questions, not answers. Downvotes are great, e.g., for answers that are coherent and well-written, but are factually wrong. The utility of downvotes on questions is less clear to me.
As discussed recently (e.g., here and here), we have a pretty aggressive closure policy on Chem.SE, and per my answer to one of those questions I think that this is probably for the best. An implication of this, though, is that a lot of questions that might just get downvoted on other SE sites end up closed on Chem.SE. The most prominent example of this is the "Off-Topic: Homework" closure-reason, by which questions showing insufficient research effort are scrubbed. Absent this particular closure feature of Chem.SE, I would think such questions would get downvoted and (if I understand correctly) eventually garbage collected by the system. So, the outcome would be the same, just on a different timeframe by a different method. For every other reason for downvoting I can think of, there are similar direct-voted closure mechanisms.
Similarly, I would hate to bring the SE user-lockout hammer down on someone who has lots of genuine, but misguided, chemistry questions, and who ends up with a raft of downvoted questions barring them from engagement on the site. On the other hand, I guess there's value to the site in keeping people with misinformed/misguided questions from flooding the zone with such.
So: Why bother downvoting questions, instead of just closing/flagging? Conversely, what are circumstances where one would downvote but not vote/flag to close? Is there some aspect of the inner workings of SE, such as user censure, which makes downvoting both useful in a way other than triggering closure and desirable in terms of maintenance of quality content/community?
To note: It doesn't bother me hugely that people do downvote questions; I just rarely see the point in downvoting them myself.