And the old default rule for sole spousal beneficiaries continues to apply under which an inherited IRA from which the RMD is not distributed is deemed to have been assumed by the survivor. Therefore, the unknown is now whether that will happen after one year for failure to take the annual RMD as an EDB or at the end of 10 years for failure to drain the account under the 10 year rule. The spouse needs to know whether they have defaulted to ownership or not, so we are back to square 1 with the question of whether the spouse can elect the 10 year rule to avoid annual RMDs.
With respect to rolling over at the end of year 9, that should be OK, but they should never wait until the 10th year if they are going to take a distribution since that distribution could be deemed to be an RMD under the 10 year rule and not rollover eligible. The spouse should always do the spousal rollover by direct trustee transfer, that way there is NO distribution that the IRS could deem to be an RMD.
This is analogous to the old rule for a qualified plan with the 5 year rule as the default for deaths prior to RBD. For a QRP where direct rollovers are treated as distributions, a surviving spouse had to do any rollover before the 5th year. If they waited until the 5th year the entire balance became an RMD because it had to be reported as a distribution on the 1099R, and the stretch was lost.
Of course, we are discussing the outlier of a spouse preferring the 10 year rule to avoid annual RMDs and that should be very rare. It almost never happened with the old 5 year rule, but perhaps 10 years will be more tempting.
Am still speculating here, not knowing how the eventual IRS guidance will read. Natalie Choate for one apparently feels that the 10 year rule is OFF the table with respect to EDBs, but there are still several aspects of Secure that even she is not sure what the IRS will do. The Secure guidance has been pushed back by Covid, so there may continue to be several situations remaining in limbo as time passes.