Gain Harvesting vs Roth Conversion: Which Uses the Space?

By Tom, CPA · Updated October 25, 2026

TL;DR: Ordinary income, including Roth conversions, fills the tax brackets first, and long-term gains stack on top. Every dollar converted pushes a dollar of gain out of the 0% band. Gain harvesting saves a future 15% tax; conversion saves a future ordinary tax and shrinks RMDs. Which wins depends on your balances, future rates and ACA status.

Educational only, not tax, legal, or investment advice. Examples are illustrations, not predictions. Rules and limits change; verify current figures and talk to a qualified professional about your situation.

Two strategies want the same cheap tax seats. Somebody has to stand. Early retirees love two moves: harvesting long-term gains at 0% and converting traditional IRA money to Roth at 10% or 12%. Both use the same low-income years, and both use the same bracket space. This page shows how they collide and how to choose.

CPA Warning Label: You cannot have both in full. Converting to the top of 12% while harvesting gains turns your 0% gains into 15% gains.

What each move buys you

Gain harvesting sells appreciated shares in a taxable account and buys them back, raising your cost basis at a 0% federal rate. The benefit is avoiding a future capital gains tax, usually 15%, on that gain. If you hold the shares until death, heirs generally receive a stepped-up basis anyway, so harvesting saves nothing on assets you never sell.

Roth conversion moves money from a traditional IRA to a Roth IRA, paying ordinary income tax now. The benefit is avoiding a future ordinary income tax, potentially at 22% or more, plus smaller RMDs, lower future MAGI for IRMAA and Social Security taxation, and tax-free money for heirs.

Bracket room overlap

Ordinary income is applied first; long-term gains and qualified dividends stack on top. For 2026, the 0% rate on gains applies up to $98,900 of taxable income for married couples filing jointly and $49,450 for single filers. The 12% ordinary bracket ends at $100,800 and $50,400 (IRS). The two ranges nearly coincide, so the same space can be filled with either.

Illustration. Assumptions: married couple, 2026, $40,000 of IRA withdrawals, standard deduction $32,200.

The last row shows the collision: after converting to the top of 12%, the $30,000 harvest is taxed at 15%, adding $4,500.

When harvesting wins

• Your traditional IRA is small relative to your taxable account, so RMDs will be modest.
• You will need to sell taxable shares to fund spending in a future higher-income year.
• You rely on ACA premium tax credits and need to keep MAGI low: harvesting adds MAGI as conversions do, but you can then live off high-basis shares that create little income later.
• You are in a state with no income tax, so the harvest is fully free.
• You have a capital loss carryforward that would absorb gains anyway (see capital loss carryforwards).

When converting wins

• Your traditional balance is large and RMDs at 73 would push you into 22% or higher.
• A surviving spouse will file single, with narrower brackets.
• Heirs are in high brackets and would face the 10-year rule on inherited traditional IRAs.
• Your taxable account will mostly pass to heirs with a step-up, making harvested basis less valuable.
• You want to build Roth money accessible before 59½ through a conversion ladder.

Per dollar of bracket space, conversion avoids tax on the full amount at a future ordinary rate, while harvesting avoids tax only on the gain at 15%. For many large-IRA households, conversion is the better use of the space.

Illustration: what the space is worth

Illustration. Assumptions: the same married couple with $40,000 of IRA withdrawals and $45,550 of extra bracket space to use in 2026.

• Convert $45,550: costs $5,126 now, an average of about 11.25%. If that money would otherwise be withdrawn later at 22%, it avoids about $10,021 of future tax, a net gain of about $4,895 (ignoring growth, which affects both sides equally).
• Harvest $45,550 of gains: costs $0 now. If those shares would otherwise be sold in a year with a 15% rate on gains, it avoids $6,833 of future tax. If they would be sold at 0% anyway or held until death, it avoids nothing.

On these assumptions, harvesting looks better, but only if the future 15% sale is likely. Change the assumptions (larger IRA, single survivor, step-up likely) and conversion wins. The answer is in the assumptions, which is why you should write them down.

ACA interaction

Both moves raise MAGI for the premium tax credit. As of Oct 2026, households above 400% of the federal poverty level get no credit, and there is no repayment cap for tax years after 2025 (IRS). That makes the ACA limit, not the bracket, the binding constraint for many early retirees on marketplace coverage. In that case, compare the tax saved per dollar of MAGI used: a conversion that avoids future 22% tax uses MAGI at a 12% cost; a harvest that avoids future 15% tax uses MAGI at 0%. Remember that losing even part of the credit adds an effective cost to every dollar of MAGI, often several percentage points.

Decision grid

Other tie-breakers

• Social Security timing. Conversions before claiming reduce future IRA withdrawals that count in provisional income; harvested gains do not change future provisional income much, since they are a one-time event.
• IRMAA. After 63, both moves count; conversions are often larger, so watch tiers.
• Liquidity. Conversion tax must be paid from somewhere, usually the taxable account, which reduces what is left to harvest. Harvesting costs nothing in cash.
• Loss carryforwards. A carryforward offsets gains, letting you harvest beyond the 0% band without using it, which frees the space for conversions.
• State taxes. Most states tax both as ordinary income; the federal 0% rate does not apply at the state level.

A note on qualified dividends

Qualified dividends from a taxable account stack with long-term gains and use the same 0% space. A $500,000 taxable portfolio yielding 1.5% produces $7,500 of qualified dividends each year before you harvest anything. Those dividends also get pushed into the 15% band by conversions. Include them in every calculation, and remember that fund capital gain distributions arrive in December, after many people have already sized their conversions. Leave a buffer.

Common mistakes

• Converting first, then harvesting in December without rechecking the 0% room.
• Forgetting that the standard deduction shelters the first dollars of either move.
• Ignoring ACA effects, where the lost credit can dwarf the tax.
• Harvesting gains on shares you plan to hold forever.

Sequencing across years

You do not need to choose once. A common sequence: harvest gains in the first year or two after retiring, when the taxable account funds spending; then shift to conversions as the IRA becomes the main source and RMDs approach. Re-run the numbers each year with the Roth conversion calculator and the gain harvesting calculator.

Lavish Larry tried to do both to the max and paid 15% on gains he meant to harvest for free. Frugal Fran picked one per year and wrote it on her calendar.