Tax Loss Harvesting vs Roth Conversion: Which Should You Do First in 2026?
Both moves save you money on taxes. Only one of them can go first, and doing them in the wrong order in the same calendar year can cost you the entire benefit of the cheaper one. That is the practical problem with tax loss harvesting vs Roth conversion: they both consume the same scarce resource — room inside your current-year taxable income — and they consume it in opposite directions.
This post lays out how the two strategies actually interact under 2026 tax rules, which one deserves priority in five common situations, and the specific income thresholds that should decide it for you. By the end you will know which lever to pull this year and which to defer.
What each strategy actually does to your tax bill
These get lumped together as “tax optimization,” which hides the fact that they work on completely different parts of the return.
Tax loss harvesting sells a taxable-account holding that is worth less than you paid for it, banks the loss, and immediately buys something similar but not “substantially identical.” The realized loss first offsets capital gains. If losses exceed gains, up to $3,000 of the excess offsets ordinary income, and anything left over carries forward indefinitely. That $3,000 cap has not moved since 1978 and is not indexed to inflation, which is the single most under-appreciated fact in this whole discussion. The IRS wash sale rule voids the loss if you buy a substantially identical security within 30 days before or after the sale — a 61-day blackout window, in both directions.
A Roth conversion moves money from a traditional IRA or 401(k) into a Roth account and adds the converted amount to your ordinary income this year. You pay tax now at your current marginal rate to buy tax-free growth and tax-free withdrawals later. Since the Tax Cuts and Jobs Act eliminated recharacterization in 2018, a conversion is permanent. You cannot undo it in April when you discover you misjudged your income.
Harvesting lowers this year’s taxable income. Conversion raises it. That is the whole conflict.
Tax loss harvesting vs Roth conversion: the side-by-side
| Tax loss harvesting | Roth conversion | |
|---|---|---|
| Effect on 2026 income | Reduces it (capped at $3,000 against ordinary income) | Increases it (uncapped — you choose the amount) |
| Reversible? | Yes, in effect — you can harvest again next year | No. Permanent since 2018 |
| Requires | A taxable brokerage account holding a loss | Pre-tax IRA/401(k) balance + outside cash for the tax |
| Best window | Any market drawdown; opportunistic | Low-income years — gap years, sabbaticals, early retirement |
| Typical annual value | $3,000 × marginal rate — roughly $360 at 12%, $720 at 24% | Rate arbitrage on the full converted amount — often thousands |
| Main trap | Wash sale rule; basis reduction just defers the gain | Bracket, IRMAA and subsidy cliffs; no do-overs |
| Deadline | Dec 31 (trade must settle in-year) | Dec 31 — no extension, unlike IRA contributions |
Look at the “typical annual value” row and the asymmetry jumps out. Harvesting has a hard ceiling on the ordinary-income side. Conversion does not. When the two compete for the same income room, the uncapped strategy usually wins on raw dollars — but only when your rate math supports it.
Which one goes first: five scenarios with the actual thresholds
The 2026 numbers that decide this, all from IRS Rev. Proc. 2025-32: the standard deduction is $16,100 single and $32,200 married filing jointly. The 12% bracket runs to $50,400 of taxable income single, $100,800 joint. The 24% bracket tops out at $201,775 single. And long-term capital gains stay at 0% up to $49,450 of taxable income single, $98,900 joint.
1. You have a big unrealized loss and no gains to offset. Harvest first, and harvest aggressively. You are not competing with anything. The loss offsets $3,000 of ordinary income and the remainder carries forward with no expiration — per IRS Publication 550, carryforwards last until used up. That carryforward is a standing asset. Then convert whatever the year allows.
2. You are in a genuine low-income year and sitting near the top of the 12% bracket. Convert first. Filling the 12% bracket with converted dollars is a rate arbitrage you may never see again, and it is worth far more than $360 of harvested-loss savings. Harvest afterward only if losses remain that offset realized gains, not ordinary income.
3. Your taxable income sits just under the 0% capital gains threshold. Neither, at first. Consider tax gain harvesting instead — realizing gains at 0% to reset your cost basis upward. Harvesting losses here is close to worthless, because you would be spending a loss to shelter income that is already taxed at zero. A conversion in this spot fills the bracket with ordinary income and pushes gains out of the 0% band, so run the numbers carefully before you do it.
4. You are within two years of Medicare enrollment. Harvest first, convert cautiously. IRMAA uses your modified AGI from two years prior, and 2026 surcharges start at $109,000 single and $218,000 joint. It is a cliff, not a phase-in — one dollar over triggers the full surcharge for the year. A conversion that clears the threshold by $500 can cost more in premiums than it saves in taxes.
5. You are buying health insurance on the ACA marketplace. Harvest first. Marketplace subsidies phase out against MAGI, and a Roth conversion raises MAGI dollar for dollar. Harvested losses reduce it, which is one of the rare cases where the $3,000 cap punches above its weight — it can preserve a subsidy worth several times the direct tax savings.
The pattern: harvesting is the safe, small, always-available move. Conversion is the large, irreversible move that needs a specific window. If you are unsure where either fits in the broader sequence, our order of operations for tax-advantaged accounts maps out where every extra dollar should go before you start optimizing at the margins.
The mistake that costs people the most: doing both badly in the same year
The genuinely expensive error is not picking wrong. It is converting in December, discovering in February that a harvested loss carryforward would have been more valuable applied against a gain you had not yet realized, and having no way to unwind either decision.
Two specifics worth internalizing.
First, harvested losses offset capital gains before they touch ordinary income, and they cannot offset the ordinary income created by a Roth conversion beyond the $3,000 cap. People routinely assume a $40,000 harvested loss can shelter a $40,000 conversion. It cannot. It shelters $3,000 of it, and the other $37,000 waits for future capital gains.
Second, tax loss harvesting on a buy-and-hold index position is usually a deferral, not a permanent saving. Selling at a loss and buying a similar fund lowers your cost basis, which means a larger gain later. You are borrowing from your future self at 0% interest, which is worth something — but it is not the same as the permanent rate arbitrage a well-timed conversion delivers. We ran the full math on when the deferral is actually worth the effort in our breakdown of whether tax loss harvesting is worth it for small portfolios, and the honest answer for accounts under roughly $50,000 is often no.
Wondering what tax-free compounding on a converted balance is actually worth over 20 years?
How I sequence tax loss harvesting vs Roth conversion in my own accounts
I write software for a living, and my instinct with anything like this is to turn it into a decision function with clear inputs rather than a judgment call I re-litigate every December. So a few years ago I wrote down my own rule and have mostly stuck to it: harvest opportunistically whenever a position is meaningfully underwater and the replacement fund is close enough that I do not care about the swap, then look at conversions exactly once, in the second week of December, when I finally know what my income for the year actually was.
What surprised me was how rarely the conversion side triggers. In a normal earning year the marginal rate is high enough that converting is just prepaying tax at a bad rate — the arithmetic says wait. The years it fires are the unusual ones. Meanwhile the harvesting side runs almost automatically and produces a modest, boring, reliable benefit that shows up as a smaller number on line 7 of the 1040.
The behavioral trap I had to design around is that conversions feel productive in a way harvesting does not. You are doing something big and permanent. That feeling is exactly why the December-only rule exists. I do not trust myself to evaluate an irreversible tax decision in October on a hunch, and the research on how we overweight vivid, effortful actions over quiet ones is not flattering to anybody.
The other thing I do not do is complicate my account structure to chase either strategy. Both work best against a portfolio simple enough that you can see all of it at once — which is roughly the argument we make in our look at asset location versus asset allocation, where the location tinkering matters far less than most guides claim.
When neither one is your best move
If you are still filling tax-advantaged space, stop and do that instead. A dollar into a 401(k) match or an HSA beats a harvested loss by an order of magnitude, and it beats a conversion by definition — you cannot convert money you have not contributed. Both strategies discussed here are optimizations that sit on top of a funded plan, not substitutes for one.
If your traditional IRA balance is zero and you are a high earner, the relevant move is not a conversion at all but the annual contribute-then-convert routine described in our backdoor Roth IRA step-by-step guide, where the pro-rata rule does most of the damage people run into. And if you are early in your career and still deciding which account type to fund at all, Roth versus traditional in your 20s is the more useful question than either strategy here.
Frequently asked questions
Can I use harvested losses to offset the tax on a Roth conversion?
Only up to $3,000 per year. A Roth conversion generates ordinary income, and capital losses offset ordinary income at the statutory $3,000 annual limit ($1,500 married filing separately). Any excess carries forward against future capital gains. A large harvested loss will not shelter a large conversion.
What is the deadline for each?
December 31 for both. Unlike IRA contributions, which you can make until the April filing deadline, a Roth conversion and a harvesting trade must be completed within the calendar year to count for that year. Give yourself a settlement buffer in late December rather than trading on the 31st.
Does the wash sale rule apply to Roth conversions?
Not directly, but there is a related trap. If you sell a security at a loss in your taxable account and buy a substantially identical security in your IRA — including a Roth IRA you just converted into — within the 61-day window, the loss is permanently disallowed. Unlike an ordinary wash sale, you do not even get a basis adjustment, because the replacement shares sit in a tax-sheltered account.
Should I convert during a market drawdown?
A drawdown is a reasonable time to convert, because you move more shares for the same tax bill and the recovery happens inside the Roth. But it is also when harvesting opportunities are most abundant, so the two compete. If both are available, size the conversion to your bracket target first, then harvest what remains.
Is there a scenario where I should skip both?
Yes — several. If you have unused 401(k), IRA or HSA capacity, fund that first. If your taxable income is low enough to sit in the 0% long-term capital gains band, harvesting losses is close to pointless and gain harvesting may be better. And if a conversion would push you over an IRMAA or ACA subsidy threshold, skipping it entirely is often the higher-return decision.
Photo by Kelly Sikkema on
Unsplash
