What an 8-Year Head Start on Roth Conversions Is Worth

Call them Robert and Diane. Robert is 70 and has been retired for five years. Diane is 66 and is retiring this year. Between them they hold roughly $3 million in traditional IRAs and 401(k)s, built over three decades of steady contributions and market growth. Almost none of it is in Roth accounts. Robert's portfolio is allocated 70% stocks and 30% bonds. Diane's is more aggressive at 90% stocks and 10% bonds. Neither holds bonds in a taxable account.

Robert's first Required Minimum Distribution is coming up in a couple of years. Diane's is close to a decade out. Robert has already started Social Security. Diane plans to wait until 70 to claim, which is the right call for maximizing her benefit, but it also means her income (and taxable Social Security) picture gets more complicated later, not less. Both of them are already sitting in the third Medicare IRMAA tier, meaning their Medicare Part B and Part D premiums are already elevated, and that tier only gets harder to escape once RMDs begin stacking on top of Social Security.

Their instinct, a very common one, is to shrug and accept it. The tax bracket is what it is, the IRMAA tier is what it is, and starting Roth conversions now feels like it would only make this year's tax bill worse without changing much of anything. Some in this position even start wondering whether they should intentionally slow down the growth of their traditional accounts, since every extra dollar of growth is eventually a bigger forced withdrawal, not more spendable wealth.

That instinct is understandable. It is also, in most cases, backwards. The better question is not whether the account grew too much. It's whether the growth is sitting in the right bucket. To make that concrete, here's the same couple, two different paths, starting eight years earlier at age 62.

Run the math backward from today's roughly $3.0 million balance at a flat, illustrative 6% annual return, and Robert's household was sitting on approximately $1.88 million at age 62, the year he first became eligible to think seriously about a conversion window. In this scenario, he does nothing different. The traditional balance simply compounds at 6% a year for eight years and arrives at $3.0 million today, all of it still fully taxable on the way out, every single dollar.

Nothing about this path is a mistake in the way people usually think about investing mistakes. The portfolio grew. Nobody panicked and sold in a downturn. Nobody chased a bad annuity or paid excessive fees. It is, in a lot of ways, a retirement savings success story. The problem is entirely about structure, not performance: all $3.0 million sits in one tax bucket, and that bucket is ordinary income waiting to happen.

Now rewind the same eight years, starting from the same $1.88 million, and add one change: at 62, this household begins a deliberate Roth conversion plan. From 62 to 64, before Medicare enrollment and before the two-year IRMAA lookback window starts to matter, they convert more aggressively, about $120,000 a year, using cash flow and taxable savings to cover the resulting tax bill so the full converted amount lands in the Roth account. From 65 to 69, once Medicare is in the picture, the pace drops to a more IRMAA-aware $70,000 a year, still meaningful, but sized to avoid tripping into a materially higher premium tier.

Over those eight years, that's $710,000 converted from traditional to Roth, at an estimated blended tax cost of roughly $163,000, paid from outside the IRA, not from the conversion itself. Both the remaining traditional balance and the growing Roth balance compound at the same illustrative 6%. Run it forward to today, age 70, and the total portfolio is still almost exactly $3.0 million. The growth assumption never changed. What changed is the mix: about $2.04 million traditional and $960,000 Roth, instead of $3.0 million traditional and zero Roth.

That chart is the entire thesis in one picture. Same total balance. Same assumed rate of return. The only variable that moved is which government has a future claim on the money. In Scenario A, the IRS has a claim on all of it. In Scenario B, roughly a third of it is already settled, tax-free, forever.

The real payoff of Scenario B doesn't show up at 70. It shows up once Required Minimum Distributions begin at 73 and keep compounding for the next two decades. In Scenario A, the traditional balance keeps growing untouched until 73, reaching roughly $3.57 million, and the IRS Uniform Lifetime Table then forces out a first distribution of about $135,000, a number that only grows every year after that as the life-expectancy divisor shrinks. In Scenario B, the smaller traditional balance of roughly $2.43 million at 73 produces a first RMD of about $92,000, and the gap between the two paths only widens with age.

Add up every forced distribution from 73 through 90 and Scenario A pushes out an estimated $3.84 million of taxable income the household never asked for and can't opt out of. Scenario B pushes out about $2.61 million over the same stretch. That's roughly $1.23 million less lifetime taxable income landing on the return in years when it stacks on top of Social Security, keeps IRMAA elevated, and, for whichever spouse outlives the other, gets taxed at single-filer brackets that are meaningfully less forgiving than married-filing-jointly brackets.

A few things fall out of this illustration that are worth separating from the specific dollar figures:

• The account total barely changes. This isn't a story about a better investment return. Both paths assume the identical 6% growth rate. The entire difference comes from moving dollars from one tax bucket to another while income is relatively low, not from picking better funds.
• The cost is real, and it's supposed to be. Roughly $163,000 of upfront tax over eight years is not a rounding error. It's the price of converting future uncertainty into a known, controlled outcome today. The comparison that matters isn't zero tax versus some tax. It's some tax now, at a rate you can see and choose, versus more tax later, at a rate the IRS sets for you once RMDs and IRMAA tiers are locked in.
• The IRMAA-aware pacing matters. The plan doesn't convert as aggressively after Medicare enrollment as it does before. That's deliberate. A conversion that saves income tax but trips a higher IRMAA tier for two years can erase a meaningful chunk of the benefit.
• The Roth bucket keeps paying dividends after age 90 too. Left untouched, that $960,000 Roth balance is projected to grow past $3 million by age 90, entirely tax-free, with no RMDs ever forced on the original owner, and available as a flexible, tax-free lever for whichever spouse is managing things alone later.

So what about a household that's actually standing where Robert and Diane are today, at 70 and 66, RMDs already close, IRMAA tier already crossed, wondering if the window already closed? Two things are true at once. First, the eight-year head start illustrated above is gone and can't be recovered. Second, the window isn't fully shut, because it never closes all at once for a married couple.

Diane's RMDs are still close to a decade away. That gap, between now and when her distributions become mandatory, is still a real conversion window, just a shorter and more tightly calibrated one than starting at 62 would have been. The household can still convert modest, IRMAA-aware amounts each year, still use Robert's Social Security-only income years to fill lower brackets before Diane's RMDs stack on top of everything else, and still coordinate whether Diane delaying Social Security to 70 changes the ideal conversion amount in any given year. None of that requires slowing down portfolio growth on purpose, which is generally the wrong lever to pull. Slowing growth doesn't reduce the tax problem, it just makes the household poorer while leaving the same structural issue in place.

The honest version of the answer is this: the biggest win was available at 62 and it's not available anymore. That doesn't mean there's nothing left to do. It means the remaining moves need to be sized correctly for where the household actually stands today, which is exactly the kind of year-by-year modeling a real plan requires, not a rule of thumb.

• Robert and Diane are a composite, hypothetical household, not an actual Singh PWM client or any specific individual.
• A flat 6% annual growth rate is assumed in every scenario and every year. Real portfolios do not grow in a straight line; actual sequencing of returns will change every dollar figure in this illustration.
• The starting balance of approximately $1.88 million at age 62 is back-solved from an assumed current balance of $3.0 million at age 70, since the source scenario did not specify an actual balance.
• Required Minimum Distributions are projected using the IRS Uniform Lifetime Table divisors currently in effect, applied to a balance compounding at 6% net of each year's distribution. Actual divisors and RMD start ages are set by current law and can change.
• Tax cost is estimated using blended effective rate assumptions (approximately 23% on converted dollars, 31.5% on Scenario A's RMD-driven income, and 24.5% on Scenario B's smaller RMD-driven income) meant to reflect a realistic mix of federal bracket, Arizona's flat state rate, and IRMAA and Net Investment Income Tax effects. These are illustrative blended rates, not a bracket-by-bracket tax return calculation for any specific filer.
• This is a planning illustration, not tax or legal advice. Any Roth conversion decision should be modeled against your actual balances, brackets, and IRMAA thresholds before you act.