The Roth Conversion Proposal That Costs More Than It Saves
Ken, 61, retired in January. He and Lisa, 58, need $85,000 a year. Assets: $980,000 in traditional IRAs and 401(k)s, $260,000 in a taxable brokerage account (mostly index funds with $90,000 of unrealized gain), $120,000 in Roth IRAs, and $50,000 in cash. Their only other income is $12,000 of interest and dividends. They buy health coverage on the federal marketplace; at a MAGI of about $60,000 they receive an advance premium tax credit worth roughly $14,000 a year. A salesperson proposes converting $150,000 a year to Roth for four years, which would push them into the 22% bracket and eliminate the credit. Ken turns 63 in two years and 65 in four. Neither has claimed Social Security; Ken's benefit at 67 would be $2,900.
1. Run a marginal-cost analysis of the proposed conversion program and judge whether to adopt it.
Read the worked answer
Cost of a $150,000 conversion year:
- With only $12,000 of other income, taxable income lands near $130,000 - taxed in layers at 10%, 12%, and 22%, producing roughly $18,500–$19,000 of federal tax. The salesperson's '22% bracket' framing overstates the rate but understates the cost.
- Lost premium tax credit: about $14,000, which adds roughly 9 percentage points to the effective rate.
- All-in: about $33,000 on $150,000, or 22%, before state tax.
The timing argument is stronger than the rate argument. Ken's RMD age is 75, so he has ten years - 65 to 74 - of low-income, pre-RMD, pre-Social Security years in which the credit is gone anyway and Medicare replaces the marketplace. At $100,000+ a year of conversions, that window alone can absorb the entire $980,000 at 12%–22% marginal rates with no subsidy sacrificed. Paying 22% now to accelerate conversions that could be done at the same or lower rate later is value destruction with extra steps.
The complication that should drive the recommendation: after 2025 the enhanced credits are scheduled to expire and the 400%-of-poverty cliff returns. A cliff means one dollar over roughly $85,000 of MAGI forfeits the entire credit. So the correct program during ages 61–64 is not $150,000 - it is a conversion sized to land just under the cliff each year, perhaps $60,000–$70,000, preserving the credit while still filling the 10% and 12% brackets. Starting at 63, the same MAGI controls IRMAA at 65 through the two-year lookback, which reinforces keeping it low rather than high.
2. Where should the conversion taxes come from, and in what order should accounts be spent?
Read the worked answer
Pay the tax from the taxable account or cash, never by withholding from the IRA. Withholding shrinks the amount that actually reaches the Roth, and Ken is past the penalty window but not past the arithmetic: every dollar withheld is a dollar that never compounds tax-free.
The $260,000 taxable account is doing double duty - it funds $85,000 of spending and pays the conversion taxes - so its size, not the IRA balance, sets the ceiling on the program.
Spending order:
- Cash, then taxable, harvesting long-term gains where they fall in the 0% capital gains bracket. Watch the interaction: realized gains count in ACA MAGI, so gain harvesting and premium-credit headroom compete for the same dollars. In the 61–64 window, the credit usually wins.
- Traditional IRA withdrawals only after the conversion years or when essentials require them - ideally never in the same year as a large conversion.
- Roth last. It is the asset that protects Lisa, who will likely file single with roughly half the bracket room.
Revisit if Ken's health changes: a materially shorter life expectancy argues for fewer conversions, keeping the traditional IRA for QCDs and a lower-bracket heir under the 10-year rule.
3. What if the couple's spending need rises to $120,000 a year? How does the strategy change?
Read the worked answer
At $120,000 the household no longer has a conversion problem; it has a cash-flow problem, and the two must not be confused.
The draw is roughly $120,000 plus taxes, less $12,000 of interest and dividends - call it $115,000, or 8.4% of $1.41 million of investable assets, sustained for four years before Medicare and ten before Social Security. That is not sustainable, and it must be fixed before any conversion is sized.
Consequences:
- Conversions shrink to near zero, not to zero. Spending from the taxable account keeps MAGI low enough to preserve the credit, but it also burns the account that pays conversion taxes. Once the taxable account is drawn down, conversions must be funded from the IRA itself, which is materially less efficient.
- The conversion program should be deferred wholesale, not scaled down, until either spending falls or Medicare begins. The 65–74 window can still absorb several hundred thousand dollars at 12%–22%.
- A $120,000 need argues for re-examining the retirement date and the Social Security claiming plan - Ken delaying to 70 would raise the floor and reduce the portfolio draw by roughly $45,000 a year from age 70 onward, worth more than any conversion in this scenario.
The judgement to demonstrate: conversions are a use for surplus low-bracket capacity, never a substitute for a funding plan that doesn't work.