Simon Kissane

@StephenPAdams The final draft of 8601:2000 declares as a reference point that 1875 is the year of the Convention du mètre, but it doesn't actually refer to its date. I can't find any reference to the year 1875 in ISO 8601:1988. And, as I said, I think it likely this system predates 1988, and hence if it really does use 1875 as an epoch, its doing so is completely independent of ISO 8601.

@StephenPAdams: I'd question the relevance of ISO 8601. Yes, 8601:2004 mentions 1875-05-20 as a reference date (not an epoch date), but that's just an editorial decision, implementations aren't required or recommended to treat that date specially. But, I don't believe the original ISO 8601:1988 revision, or the 8601:2000 update, did. And, this system is very likely a lot older than 2004, or even 1988. So, if US SSA really does use 1875 as an epoch (unproven), it seems unlikely that actually has anything to do with ISO 8601

The problem with this – I don't think anyone actually knows for a fact that US Social Security uses 1875 as some kind of "epoch date". The claim has been circulating on social media, but I haven't seen any confirmation from an official source. If the claim is actually true, then you provide a plausible explanation as to why it is true – but I still don't think we know whether it is. And, if it is true, and your explanation is correct, then COBOL itself has nothing to do with it, it was an application level design decision