[GNC] Kind of a Philosophical Question about Transaction Dates

[Patrick James]

Every journal entry has separate components. One is date. Another is the debit(s)/credit(s). Thus the date of a journal entry and the fact that debits and credits must equal are separate issues. If you go back to the days of pen and paper, there really was no way to enter a past/future date for a journal entry.

This discussion about dates is only possible as a result of modern information systems. I mean, yeah, sure accruals and adjusting entries existed, and for good reason.

If I make a donation on December 31, at 23:59, I mean, like, who is keeping track of time zones and whatnot. If I declare that for the December year, then that's probably fine, but I very much doubt anyone would take issue if I record the transaction in January of the following year.

Overall this doesn't make much difference unless there is some significant materiality to the transaction(s).



> On 03/03/2026 9:25 AM PST Hartmut W Sager <hwsager at marityme.net> wrote:
> 
>  
> I've now read this whole thread up-to-the-minute (fine granularity :) ), and I picked your response for my reply because you gave the best explanation.
> 
> But I'm replying because the whole thread didn't mention one particular point:  In double-entry accounting, the books "have to balance" at every instant in time, and thus all the legs of a transaction have to be dated/timed in the same instant.  Otherwise there would be points in time where the books aren't balanced.
> 
> I gave this issue much thought when I wrote my own accounting software (first in DOS Basic in 1982 and second in FoxPro in 1992), and I have two fields, TransDate and PostedDate.  The TransDate is the same for all the legs of a transaction, and the PostedDate is you-know-what.  The PostedDate can be used for summaries/reports where we really need the posted date, but of course, such summaries/reports must not be used where they conflict with "always balanced".
> 
> Hartmut W Sager
>