This attempt at a self-answer may be completely wrong. I originally wrote it as an edit to the question, but it seems more sensible to post it as a self-answer and see what people think.
Stuart F suggests in comments that the subjunctive may be used here to describe a condition that is "probable." If that applies here, then I feel that a satisfying answer should explain when we would not use the subjunctive for a condition. Actually the quote from Romans doesn't suggest to me at all that the hypothesis is only probable -- the speaker is certain of the hypothesis. I don't know, maybe it suggests hypothetical uncertainty on an impersonal basis, or implicitly attributes that uncertainty to someone else.
For example, I could write this: "Given President Smith's conduct, impeachment follows as the night follows the day. If two plus two be four, then she deserves impeachment." This seems correct to my ear, and yet there is no uncertainty about whether 2+2=4, only a rhetorical strategy of supposing that some unnamed person might have all sorts of absurd doubts.
How about this? "My dog barks at the mailman. If she bark at him today, I will give her a time-out in the laundry room." This sounds off to me. I'm not sure if it's because the apodosis is personal rather than impersonal. Maybe the subjunctive sounds wrong here because doubt about the condition isn't necessary to the sense of the sentence -- there is no hypothetical person saying, "Aw, maybe she won't bark."
This sounds much more natural to me: "My dog barks at the mailman. If she bark at him today, then may she suffer a time-out in the laundry room." I'm not sure if it sounds right because the archaic tone is more uniform. I think the reason it sounds right may rather be that the apodosis is impersonal.
Maybe the formula here is that (1) the subjunctive introduces hypothetical doubt about the condition (but not necessarily doubt by the speaker); (2) the doubt is necessary to the sense; but (3) such a construction simultaneously requires the depersonalization the apodosis.