The Roth conversion window

The conversion window is not symmetric, and it does not start or end at the same age for everyone. The right edge is defined by the Required Beginning Date for RMDs: age 73 for those born 1951 through 1959, and age 75 for those born 1960 or later, per SECURE 2.0. The left edge is not a statutory age at all. It is the year you actually stop drawing a paycheck, because wages consume bracket space the same way conversions do. A person still earning $200,000 at age 63 has no room to fill the 12% bracket; the window opens the year wages stop.

The practical result is that two people with identical balances and identical birthday months can have windows that differ by two full years depending solely on which side of 1960 they were born. That difference is not trivial. A couple born in 1958 who retired at 65 is currently 68 in 2026 with five years of runway before the RBD at 73. A couple born in 1962 who retired at 64 is currently 64 in 2026 with eleven years of runway before the RBD at 75. On the same $1.2M starting balance, those six extra conversion years produce a substantially different lifetime outcome.

Anatomy of the window

The left edge is entirely within your control. Earlier retirement extends the window; continued work compresses it. Social Security claiming age also matters, but indirectly: once SS starts, each conversion dollar drags additional SS into taxable income through the provisional income formula in Pub. 915, which raises the effective marginal rate on conversions. Delaying SS to 70 therefore preserves several additional years of unencumbered bracket space inside the window. The window's effective length for low-rate conversions is not simply (RBD age) minus (retirement age); it is (SS start age) minus (retirement age) for clean bracket access, and then (RBD age) minus (SS start age) for conversions that must navigate the 1.85x SS amplification band.

The right edge is fixed by statute and birth year. Under Pub. 590-B Appendix B Table III, the Uniform Lifetime divisor at age 73 is 26.5, requiring roughly 3.8% of the account to be distributed annually whether you need the income or not. At age 75 the divisor is 24.6, forcing about 4.1%. Once RMDs begin, those mandatory distributions consume bracket headroom that conversions could have used, and the arithmetic of paying tax now at a chosen rate to avoid paying more later becomes progressively weaker. This is why the right edge matters: every year before the RBD is a year where the bracket headroom is yours to allocate; every year after is a year where the IRS allocates it for you.

Worked example: the cohort difference

Two MFJ couples with identical balances, Social Security, and return assumptions. The only differences are birth year, current age, and RBD.

Filer A: born 1958, currently 68 in 2026. Retired at 65. RBD at 73, meaning five years of conversion runway remain in the current year. Filer B: born 1962, currently 64 in 2026. Retired at 64. RBD at 75, meaning eleven years of conversion runway remain. Both couples: $1.2M traditional IRA, $300k Roth, MFJ, $4,500 per month Social Security at FRA 67, claiming at 67, no other income, no state tax, 6% nominal return, 2.5% inflation, life expectancy 92. Running both through the four canonical scenarios:

The absolute dollar difference is striking: Filer B's longer window produces $128,972 in lifetime savings against do-nothing, versus $41,315 for Filer A. Both couples start the projection with identical balances. The additional six years of pre-RBD conversion runway account for essentially all of the $87,657 difference in expected benefit. Window length is not a rounding error in the analysis; it is the primary variable for the early-to-mid-retirement cohort.

Notice also that fill-22 and fill-24 both underperform fill-12 for these two couples. With $4,500/month in Social Security beginning at 67, the 1.85x amplification band in Pub. 915 makes conversions above the 12% threshold expensive during the SS years. For these filers, the right strategy is to convert aggressively at 12% in the pre-SS years and let the inflation-adjusted brackets absorb the relatively modest RMDs that remain.

Strategy implications of window length

A longer window argues for higher annual conversions earlier, not for spreading the same total amount over more years. The reason is compounding: each dollar converted and moved to Roth in year one grows tax-free for the full remaining horizon; a dollar converted in year ten gets only two or three years of tax-free growth before RMDs start. For Filer B, the calculator converts approximately $100,800 per year in the fill-12 scenario starting at age 64. After eleven years of this, the traditional IRA balance at age 75 has shrunk substantially, and first-year RMDs are a fraction of what the do-nothing balance would have generated. The bracket-filling math works because the tax paid today is at a lower rate than the tax that would have been owed on the compounded untaxed balance later.

Delaying Social Security to 70 has a specific effect on the window's structure. Instead of a four-year stretch of pre-SS clarity before the 1.85x band starts, you get seven or eight years. For Filer B who retired at 64 and delays SS to 70, that means six years of clean bracket-filling before the provisional income formula starts amplifying each conversion dollar. Whether delaying SS is right depends on longevity assumptions and whether the couple has adequate non-IRA income during the wait years; but mechanically, delayed SS widens the clean conversion window inside the overall pre-RBD period.

The late-retiree failure case

The window analysis above assumes retirement before the mid-60s. For someone who works until 70 because they enjoy their job or need the income, the arithmetic changes materially. A person born in 1958 who retires at 70 in 2028 has a conversion window of roughly three years before the RBD at 73. Social Security almost certainly starts at 70 as well, so those three years carry the full 1.85x SS amplification. In practice, the calculator produces very modest fill-22 or fill-24 conversions in these years before RMDs take over at 73.

For late retirees, the marginal math of bracket filling weakens substantially. Other strategies often produce larger results. Qualified Charitable Distributions from an IRA under IRC §408(d)(8) allow direct transfers to charity of up to $108,000 per year (2026 limit, indexed to inflation) that count toward the RMD requirement and are excluded from AGI entirely. A late retiree with charitable intent can reduce effective RMD tax cost through QCDs at 70.5 even while the conversion window has largely closed. Similarly, naming a qualified charity as a beneficiary of an IRA results in the charity receiving the full untaxed balance, which may compare favorably to converting, paying income tax, and then leaving a Roth account to heirs.

Related reading

The mechanics of how bracket filling works and what savings are available in the default MFJ scenario are covered in What is bracket filling? How Social Security's provisional income formula amplifies the marginal rate inside the window is covered in Roth conversion and Social Security. The IRMAA two-year lookback, which affects Medicare premiums during the window years, is covered in Roth conversion and IRMAA. For the four scenarios where conversion does not help, see When not to do bracket filling.

Run your own numbers.