RMDs, account withdrawals, 4% simplified- MAYBE?

We’ve been using our RMDs as our primary withdrawal criteria for almost seven years now (spouse having started at age 70). Even before, I used the 4% rule as a rough estimate of a safe withdrawal/spending level. In reality, we spent what we wanted to spend (within reason) and we checked the withdrawal rate afterwards (although we did keep tabs on our guardrail). This process was made more palatable since we partially annuitized our portfolio to create our own self-funded pension. That annuity income plus my SS benefit (not counting spouse’s SS benefit) covered our essential expenses. That would then leave her SS benefit plus 4% or the RMD to be used for discretionary spending. Additionally, I used a Monte Carlo simulation from 2016 (the third full year after I retired and reached my FRA) that projected our annual portfolio balances to the end of our planning period. Each year I update those projections with current data (portfolio balance and expense actuals) to verify if the original 2016 projections are still accurate. The variations over the past 10 years have been generally less than 10% so I continue to use those projected balances as my guardrail. If I stick to those projections, it will leave a minimum of five years of LTC coverage (in the T-IRA) or a substantial legacy if not used.  Additionally, this still leaves the Roth assets and home equity as a further backup.

To address potential market volatility when withdrawing RMDs, we have each been using a rolling five-year fixed-income ladder where each ladder rung satisfies our projected RMDs. We previously went through a six-year Roth conversion plan when TCJA became effective. That plan, which was completed in 2022, reduced our projected RMD by 40%. Today, roughly 50% of our portfolio is tax-deferred. As such, we can take twice our RMD (the effect of having only a tax-deferred IRA) and still see our portfolio balance continue to increase over the long term.

For a surviving spouse, we can also use the revised RMD calculation (allowed by Secure Act 2.0) to further reduce the RMDs by applying the “excess” annuity payout against the (traditional) RMD as computed using the end-of-year non-annuitized T-IRA balance.  The nominal 4% RMD, for us, then drops to roughly 2.4% (due to applying the “excess” annuity payout).  That further translates to roughly withdrawing 1.2% of the portfolio (since half is T-IRA). Additional income needs can then come from the Roth assets with no tax consequences. This strategy will mitigate the widower’s tax penalty and avoid IRMAA premiums for the surviving spouse.

Another interesting observation (using Boldin’s tool) was that without any further Roth conversions, the Roth allocation grows from the current 50% to 64% over the next 15 years. I believe this is happening since our Roth holds are growth equities, while the T-IRA is mostly fixed-income. Almost all withdrawal are from the T-IRA as well. Alternatively, the T-IRA decreases from 50% to 36%. While not quite in step with the Roth percentage increase as one ages, the reduction in T-IRA helps to offset some of that percentage increase.