Ditto advice from SoCal and masteff. This may not a "wash" as you imply, so do each step separately and completely.
As implied in Post #1, this situation will recur. If so, Andy correctly advises a plan termination, before the next RMD is due.
QDROphile is correct.
If the NQ plan does not have any existing language, it may be prudent to consider if there is any precedent.
At any rate, perhaps the plan sponsor should take this opportunity to amend the plan to incorporate some provisions for beneficiary identification. This might emulate the qualified plan, but is not required to do so.
Your facts imply a mistake by the trustee. Is that correct? Is the bank the plan sponsor or the trustee?
If the bank is the trustee (and not also the sponsor), does it have the same rights as the the plan or the plan administrator? Perhaps the PA has standing to adjust/recoup future payments to correct its mistake, but does the trustee have the same standing?
(I don't know, just asking.)
Was the participant a VT before you took over as TPA?
- If no, then you would have found (or not found) a distribution in the assets, thus confirming the current status.
- If yes, then this sounds like a data question for the plan sponsor to answer.
Just a guess.
Really, you!
The comments by AtA and SoCal are exactly on point. Averaging removes some volatility but adds some complexity. You decide which is best for you (which is characteristic of more than one design decision). In my experience, using a one-year stability period is very common. A one-month lookback is probably not enough, but the sponsor has reasonable flexibility in choosing that definition.