There are two ways I see it not being helpful:
It's not actually a duplicate. But I mean this in a pretty loose manner: if the original question covers the same ground as the new question, starting with the same set of assumptions, it's a duplicate. If the question is asking about a different problem but still has the same answer as the original, it's not a duplicate.
The original question is worse than the new question. Generally, this happens when the older question was terribly worded, contained no useful information, or was otherwise unsalvageable whereas the new question was borderline or had far more useful answers, and has a chance to be reopened.
The second one requires a little bit of a judgement call, but I think it comes down to effort expended. If you took a look at both questions side-by-side, which one has the least amount of work ahead of it to get it to meet our quality standards? If it's the older one, close the new question as a duplicate. If it's the new one, close it for one of the other reasons.
But in many cases, this is largely a pragmatic decision: in a perfect world, we wouldn't have unsalvageable questions lying around, and all questions that have a chance to be reopened with a heroic edit would, in fact, be open.
But due to questions slipping through the cracks or lack of community effort, we still do have a lot of the former and salvageable questions that simply aren't.
So it's okay, in some cases, to rely on a closed question as a resource particularly for closing as an exact duplicate. If the question is bad, but there happened to be an existing question that—while also bad—actually answers the new question, I say point it out.