The policy is not different from that for duplicate questions in principle: any duplicates found get closed, and only one of them stays open.
Why we still close duplicates when there is no answer?
It may seem harsh to go that way, but if you go through the possible scenarios, it actually should lead to the best possible outcome available. To see it, consider two cases: we either have somebody who knows the answer (and wants to write it up), or we have no such person.
In case nobody knows the answer, then the OP doesn't get to know what they wanted. This is indeed "useless" as stated in a comment under a question - but note that under this scenario, no matter what we do, there is no way for the OP to get an answer anyway.
In the other case, we hope that the knowledgeable person will write an answer. And while it may seem like closing the question prevents it, in fact, if we assume that the questions were correctly recognized as duplicates, then the knowledgeable person still can write an answer - under the one question that stays open. From their point of view, there should be no difference which of the questions they answer, seeing that they should be getting the same answer (else they wouldn't be duplicates). Even if they landed on the closed question first, they can follow the duplicate link and write it up on the open question.
So there is no downside to closing in the sense of preventing an answer from being written. On the other hand, there are all the usual upsides of closing - specifically, once answers start coming in, they are all in the same place and readers can compare them and decide which one to pick, and answerers-to-be can save themselves writing up the same thing in multiple places. So I am quite happy with the idea of duplicates still being closed even when there is no answer.
How to pick which question to close?
Now this is the hard part. When there are answers on the old question, it makes perfect sense to close the new one and keep the one with the answers. (There is even a merge function available to moderators for when the new question has already gained valuable answers before closure). But when there are no answers?
The ideal case would be to keep open the question which is most useful in the sense of 1) attracting most searchers and 2) inviting the best answer. The problem is, the people closing usually cannot predict which question will be easier to find (the first point). For the second point, if one question is very badly written or very confusing, it is the one that gets closed. But typically, there is no obvious quality difference in the questions. In which case, we usually default to closing the newest one, mostly out of inertia.
Now that you have brought up the question, I can think of a few arguments for preferring to close the older question, mostly for UX reasons. But I could also imagine this opening a whole new can of worms - in cases where the older question has answers, but they are bad/wrong ones, it gets complicated. In any case, the question of changing the way we pick which question to close is a matter for separate discussion.
