Roth Conversion at 63 With a $900,000 IRA: Fill the 12% Bracket ($56,500) or Go to 22% ($80,000)? SECURE 2.0 RMD Age 75 Math
You are 63, single, with $1.2M: $900,000 in a traditional IRA, $50,000 in a Roth, and $250,000 in a taxable account. This week's headlines are not calming. CNBC reports that the Fed is widely expected to raise rates by a quarter point at its September meeting ("The Fed is likely to raise interest rates as inflation persists. What that means for consumers"). Social Security reform is a live issue in the Senate races this November. Former Treasury Secretary Jack Lew is telling lawmakers to "keep their options open."
So you may be asking whether you should convert now, wait, or do nothing until Congress settles things.
My answer, after 20 years of doing this for a living, is that most of the decision doesn't depend on Congress at all. Below I build one worked example with real brackets and real RMD divisors, then stress-test it against the policy risks in the news. The example is illustrative. Your numbers will differ, and the differences can flip the answer.
What SECURE 2.0 Already Settled (So You Can Stop Worrying About It)
Some rules are already law and don't hinge on the November elections:
| Birth year | RMD start age | You turn that age in |
|---|---|---|
| 1951–1959 | 73 | 2024–2032 |
| 1960 or later | 75 | 2035 or later |
If you are 63 in 2026, you were born in 1963, so your first RMD is at 75 (in 2038), not 73. Plenty of articles still say 73. That's true for someone born in 1955, not for you. (If you are closer to 73, see SECURE 2.0 RMD Age 73 and a $1.3M Traditional IRA.)
Three other SECURE 2.0 provisions matter here:
- Missed-RMD penalty: now 25% of the shortfall, or 10% if you correct it promptly. It used to be 50%.
- Super catch-up at ages 60–63: if you are still working, your 401(k) catch-up limit is $11,250 for 2026, on top of the standard $24,500 deferral. Above roughly $150,000 in prior-year FICA wages, the catch-up must go in as Roth. Confirm the exact figures with your plan administrator.
- Roth 401(k) accounts: no lifetime RMDs since 2024. This is a quiet reason to prefer Roth dollars.
The 2025 tax law also made today's 10/12/22/24% bracket structure permanent. "Permanent" in Washington means "until changed," but nothing is scheduled to sunset and push your rates up.
The Worked Example: Three Ways to Handle a $900,000 IRA
Assumptions, kept simple so you can check my math:
- Single filer, 2026 brackets and standard deduction ($16,100). I ignore the extra deduction for age 65+ and any state tax.
- 2.5% real (after-inflation) return, so everything stays in today's dollars.
- $10,000 a year of interest and dividends is your only other income. Living costs come from a cash reserve I'm not modeling.
- Social Security of $40,000 a year starts at 70, so 63–69 is your low-income window.
- Conversion taxes are paid from the taxable account.
- Tax brackets for a single filer: 10% to $12,400, 12% to $50,400, 22% to $105,700.
- The Uniform Lifetime Table divisor at 75 is 24.6.
Strategy A: Wait. The IRA grows for 12 years: $900,000 × 1.025¹² ≈ $1,210,400. First RMD at 75 = $1,210,400 ÷ 24.6 ≈ $49,200.
Strategy B: Fill the 12% bracket. Convert $56,500 a year for ages 63–69. That puts taxable income right at $50,400 ($56,500 + $10,000 − $16,100). Tax per year: $1,240 + $4,560 = $5,800 (a 10.3% effective rate on the conversion).
Strategy C: Fill into the 22% bracket. Convert $80,000 a year for ages 63–69. Taxable income is $73,900. Tax: $1,240 + $4,560 + 22% × $23,500 = $10,970 a year (13.7% effective, 22% on the top slice).
Here is what each does to your age-75 picture, with Social Security at $40,000 (up to 85% taxable):
| A: Wait | B: $56,500/yr | C: $80,000/yr | |
|---|---|---|---|
| Total conversion tax (7 yrs) | $0 | $40,600 | $76,790 |
| IRA balance at 75 | $1,210,400 | ≈ $715,900 | ≈ $510,200 |
| First RMD | ≈ $49,200 | ≈ $29,100 | ≈ $20,700 |
| Federal tax that year | ≈ $11,674 | ≈ $5,612 | ≈ $3,747 |
| Annual savings vs. waiting | n/a | ≈ $6,062 | ≈ $7,927 |
| Simple payback | n/a | ≈ 6.7 years | ≈ 9.7 years |
The Finding: The 22% Slice Barely Pays for Itself
Going from B to C costs you an extra $36,190 in conversion tax. It buys only about $1,865 a year of additional tax savings at 75. On that narrow measure, the extra 22% slice takes roughly 19 years to pay back.
The first $56,500 a year is a very different story. It pays for itself in under seven years, because you are paying 10–12% now to avoid 22% later. Filling the 12% bracket is a strong default. Pushing into the 22% bracket needs a reason beyond the RMD tax comparison, and I'll cover three such reasons below.
This is the kind of bracket-by-bracket comparison Lontevis runs for you, so you don't have to build the spreadsheet yourself.
What This Table Leaves Out
The payback numbers understate the benefit of converting in these ways:
- Tax-free growth on the converted dollars. Strategy B moves about $395,500 into Roth and Strategy C moves $560,000. Those dollars never generate RMDs, and any growth on them is tax-free.
- Survivor and heir taxes. If you leave the account to a non-spouse beneficiary, the 10-year rule forces the balance out on their tax schedule. My post on the inherited IRA 10-year rule and SECURE 2.0 shows how a $350,000 inheritance can produce a $58,000 tax spike.
- IRMAA. Your MAGI in the conversion years here is $66,500 (B) or $90,000 (C). Both stay below the first Medicare surcharge tier for single filers, roughly $109,000 for 2026. A larger conversion, or a capital gain in the same year, can cross that line.
And in these ways it overstates the benefit:
- A big taxable-account draw. Strategy C spends $76,790 of your $250,000 taxable account on taxes. That is 31% of the account. If markets fall right after you retire, you'll wish that cash were still there. (See sequence risk at 63 with $1.3M.)
- Deferral value. Money kept in the IRA longer is money you aren't paying tax on today.
Stress Test 1: What If Social Security Gets Cut?
The CNBC pieces "Social Security reform plans could sway voters in battleground Senate races, survey finds" and the Jack Lew interview both point to the same fact. The senators elected this November will be in office when the retirement trust fund is projected to run dry. The 2025 Trustees Report says that, absent legislation, incoming revenue could cover roughly 77% of scheduled retirement benefits after depletion. That is a cut, not a zero, and Congress has historically acted before deadlines.
Cut the $40,000 benefit to $30,800 and rerun the age-75 tax:
| A: Wait | B: $56,500/yr | C: $80,000/yr | |
|---|---|---|---|
| Federal tax at 75 | ≈ $9,954 | ≈ $5,143 | ≈ $3,278 |
| Savings vs. waiting | n/a | ≈ $4,811 | ≈ $6,676 |
The conversion advantage shrinks by about 16% (from $7,927 to $6,676 for Strategy C), but it doesn't disappear. The RMD drives the case, not the benefit. The bigger effect is on your budget: you have $9,200 a year less income. At a 4% withdrawal rate, replacing that takes about $230,000 of extra portfolio.
For most people, then, a Social Security cut argues for holding more accessible cash and Roth money. It does not argue against converting.
Here's how I would handle it: don't let uncertainty about a cut change your conversion decision. Let it inform your claiming age instead. For the break-even math on 62, 67, and 70, see Social Security at 62 vs 67 vs 70 on $1.3M Saved. If you can model your own benefit and a haircut scenario side by side at Lontevis, you'll see which decisions hold up under both.
Stress Test 2: The Fed Hike and Your Student Loans
The rate hike. A quarter-point increase raises the yield on the cash and CDs you would use to pay conversion taxes. It also pushes down the price of bonds held inside your IRA. That second effect has a silver lining: converting shares that are temporarily down moves more future growth into the Roth per dollar of tax paid. I would not time a conversion around a Fed meeting, though. The bracket math matters far more than a quarter point.
The student loan wrinkle. CNBC's "Student loan borrowers exiting SAVE may face sharply higher payments if they don't take action soon" is a reminder that many people now enter their 60s with federal loans. They may be their own loans or Parent PLUS loans taken for a child. Income-driven repayment plans generally key off your adjusted gross income, and a Roth conversion counts as income. An $80,000 conversion could raise your monthly loan payment for that year. If you have federal loans on an income-driven plan, price that in before choosing between B and C.
When the Bigger Conversion Does Win
I've shown that the $80,000 conversion has a long payback. Here are three situations where it could still be the right call:
- You expect a higher bracket later. If you have $2M in the IRA instead of $900,000, your RMD tax at 75 might land in the 24% bracket, not 22%. Then the 22% slice looks better.
- You want to leave Roth dollars to heirs. Non-spouse heirs get 10 years to empty an inherited IRA. Roth money they inherit comes out tax-free.
- You have low-income years that won't repeat. A business sale, a sabbatical, or a gap before Social Security starts can leave room in the bracket now that will never be there again.
The reverse also applies. If you plan to give a large share of your IRA to charity, or you expect to move to a state with no income tax, the case for converting weakens. So does converting if your income at 75 will be lower than you think.
For a longer look at conversions running before RMDs start, see Roth Conversion at 63 vs RMDs at 75.
Run the Numbers on Your Version of This
The example above has clean assumptions. Yours won't:
- Filing status. A married couple has a 12% bracket that runs to $100,800 (2026) instead of $50,400. That leaves far more room to convert cheaply, and it makes the survivor's tax jump when one spouse dies a real risk.
- Portfolio size and mix. Change the pre-tax balance from $900,000 to $1.5M and the payback picture changes.
- Social Security timing. Claim at 62 instead of 70 and your low-income window disappears.
- Health. A shorter expected life shortens the payback window. A longer one lengthens it.
- Taxable income. Add $30,000 of part-time work and the whole bracket math moves.
Each of these can change the answer by tens of thousands of dollars. That is why I wouldn't take my $56,500 figure or anyone else's as a rule. It is a starting point you should test against your own inputs.
If you want to see where your own 12% and 22% bracket edges fall, your first-RMD age, and the payback on each conversion size, you can model this for your situation at Lontevis. Free your decisions from the headlines. The Fed will do what it does, and Congress will act when it acts. Meanwhile your bracket space this year is yours to use or lose.
This post is educational and uses an illustrative example. It is not tax, legal, or investment advice. Tax law, brackets, and IRS tables change, so verify current-year figures with the IRS or a qualified professional before acting.
Sources
- Citi Adds Japan Airlines as Its Newest Transfer Partner — NerdWallet Retirement
- The Fed is likely to raise interest rates as inflation persists. What that means for consumers — CNBC Personal Finance
- Student loan borrowers exiting SAVE may face sharply higher payments if they don't take action soon — CNBC Personal Finance
- Social Security reform plans could sway voters in battleground Senate races, survey finds — CNBC Personal Finance
- Lawmakers must 'keep their options open' on Social Security reform, former Treasury Secretary Jack Lew says — CNBC Personal Finance