Tax-Loss Harvesting 2026: How to Turn Losing Positions Into a Real Tax Cut

The short version: what tax-loss harvesting actually does for you

Tax-loss harvesting is the one part of investing where a losing position can hand you money back. You sell something trading below what you paid, the loss becomes “realized,” and the IRS lets you use it to cancel out taxable gains elsewhere. Wipe out your gains and you can still deduct up to $3,000 against your salary this year, then carry any leftover loss forward forever. On a $3,000 deduction in the 32% federal bracket, that is a little under $1,000 back in your pocket, before state tax even enters the picture.

Here is the catch, and it is the part most articles gloss over: harvesting usually defers tax, it does not erase it. When you sell the loser and buy something similar, your new cost basis is lower, so a bigger gain is waiting for you down the road. The benefit is real, but it lives in the details, the time value of money you did not hand the government today, the conversion of a short-term gain into a long-term one, and the possibility that you never sell the replacement at all. Get the wash-sale rule wrong and you can even end up with no benefit and a worse basis. So the mechanics matter more than the slogan.

If you are still building the mental model for how gains and losses get taxed in the first place, the broader framing in our capital gains tax guide for 2026 pairs well with everything below.

How losses offset gains: the netting order that decides your bill

You do not just subtract total losses from total gains in one lump. The IRS makes you net them in a specific order, and that order changes how much you save, because short-term gains are taxed at your ordinary income rate (up to 37%) while long-term gains top out at 20%. A loss that knocks out a short-term gain is worth more than one that only offsets a long-term gain.

The sequence runs like this. First, short-term losses cancel short-term gains, and long-term losses cancel long-term gains, within each bucket. Then, if one bucket still has a net loss, it spills over to offset the other bucket’s net gain. Whatever loss survives all of that offsets up to $3,000 of ordinary income, and the remainder carries forward.

Step	What gets netted	Why it matters
1	Short-term losses vs short-term gains	Short-term gains are taxed highest, so canceling them is the most valuable
2	Long-term losses vs long-term gains	Offsets gains taxed at 0/15/20%
3	Net loss in one bucket offsets net gain in the other	A leftover short-term loss can shield a long-term gain, and vice versa
4	Remaining net loss offsets ordinary income (max $3,000/yr)	Excess carries forward indefinitely

The practical takeaway: if you have a choice of which losing lots to sell, a short-term loss is generally the more powerful tool because it first attacks your most heavily taxed income. This is also why holding a winner one extra day past the one-year mark, to convert it from short-term to long-term, and pairing harvests thoughtfully, can matter as much as the harvest itself.

The wash-sale rule: the tripwire that can cancel your loss

This is the rule that turns a clean harvest into a mess. Under Section 1091, if you sell a security at a loss and buy the same or a “substantially identical” security within 30 days before or after the sale, the IRS disallows the loss. Note the window is symmetric: 30 days before, the day of, and 30 days after, a 61-day danger zone in total. People forget the “before” half all the time.

Disallowed does not mean destroyed. The loss you cannot claim gets added to the cost basis of the replacement shares, so you recover it when you finally sell those. But you have lost the timing benefit, which was the whole point, and if you are not tracking basis carefully you can lose the loss entirely to sloppy records.

Two words do the heavy lifting: “substantially identical.” Selling Apple and buying Microsoft is fine, they are different companies. Selling one S&P 500 index fund and buying a different provider’s S&P 500 fund is a gray area the IRS has never fully clarified, though many practitioners treat two funds tracking the same index as risky and prefer a genuinely different index. Selling a fund and rebuying the exact same fund is unambiguously a wash sale.

Do this	Avoid this
Sell Fund A (S&P 500) and buy a total-market fund instead	Sell Fund A and rebuy the identical Fund A within 31 days
Swap one company’s stock for a competitor in the same sector	Buy the same security in your IRA during the window
Wait 31+ days before rebuying the exact same security	Let auto-dividend-reinvestment repurchase the fund mid-window
Turn off automatic reinvestment on the harvested security	Assume the rule ignores your spouse’s accounts (it does not)

The nastiest version is the IRA wash-sale trap. If you harvest a loss in your taxable brokerage account and buy the same security in your IRA or Roth within the window, the loss is disallowed and, because IRA basis works differently, you get no basis adjustment to recover it later. It is simply gone. The rule spans every account you and your spouse control, taxable and retirement alike, so coordinate before you click sell.

Correlated-but-not-identical: staying invested while you harvest

The reason harvesting works at all is that you rarely want to be out of the market for 31 days. Sit in cash and a rebound can cost you more than the tax you saved. The fix is to sell the loser and immediately buy something that moves with it but is not substantially identical, so your market exposure barely changes while the wash-sale rule stays satisfied.

Classic swaps look like this: sell an S&P 500 fund and buy a total US market fund; sell a growth ETF and buy a broad-market ETF from a different index family; sell one energy company and buy a diversified energy sector fund. You keep almost identical exposure, you bank the loss, and after 31 days you can swap back to your original holding if you prefer it, or simply keep the replacement.

The trade-off to respect is tracking error. The replacement will not move in perfect lockstep, so over the 31-day window you take on a small amount of “the two things diverged” risk. For broad index swaps this is usually trivial. For concentrated single-stock swaps it can be significant, which is a good reason to swap into a sector fund rather than a single competitor when you can.

Deferral, not elimination: what harvesting really buys you

Be honest about the economics or you will over-harvest. When you sell at a loss and rebuy, your basis resets lower. That lower basis means a larger taxable gain later. So in the simplest case you have not eliminated tax, you have moved it into the future. Why is that still worth doing? Three reasons.

The first is the time value of money. A dollar of tax you defer for ten years is a dollar you can invest in the meantime; that compounding is the core benefit. The second is rate arbitrage: a loss can offset a short-term gain taxed at 37% now, while the future gain on your replacement may be long-term and taxed at 15%. You paid back a cheaper tax later than the expensive one you avoided today. The third is the exit you might never take. If you hold the replacement until death, your heirs get a stepped-up basis and the deferred gain can disappear entirely; if you eventually donate appreciated shares to charity, similar logic applies.

The mirror image is the risk. If you expect to be in a much higher bracket when you finally sell, deferral can backfire, you shielded a gain at today’s low rate and will pay tomorrow’s high rate on a larger gain. Harvesting is a bet that your future self faces an equal-or-lower rate, or never sells at all.

When harvesting helps and when it is just noise

Not every red position is worth harvesting. The benefit can shrink to nothing or go negative once you account for basis reset, transaction friction, and your bracket.

Situation	Harvest?	Reasoning
You have large short-term gains this year	Yes	Losses cancel your most heavily taxed income first
You are in the 0% long-term capital gains bracket	Usually no	You would trade a low basis for a deduction worth ~$0
High income, expecting a lower bracket in retirement	Yes	Defer now at a high rate, realize later at a low one
You expect a much higher future bracket	Be careful	Deferral can cost more than it saves
Tiny loss, wide bid-ask spread or commissions	No	Friction eats the benefit
Big volatile taxable account, automated platform	Yes	Continuous harvesting compounds small wins

The 0% bracket case deserves emphasis because it is a common unforced error. For 2026, a married couple with taxable income under roughly the low six figures can pay 0% on long-term gains. Harvesting a loss to offset a gain that was already going to be taxed at 0% gives up basis for no benefit, you would have been better off harvesting gains, not losses. Know your bracket before you act.

Automation: robo-advisors and direct indexing

If harvesting by hand across dozens of lots and a 61-day calendar sounds tedious, that is precisely the problem software solves. Most major robo-advisors scan your accounts daily and harvest losses automatically, immediately buying a correlated replacement so you stay invested and the wash-sale rule stays satisfied. Because they check every day rather than once in December, they catch dips that would have healed by year-end.

Direct indexing takes it further. Instead of owning one S&P 500 fund, you own the underlying stocks directly. Even in an up year, some individual names are down, so there are far more harvestable lots than a single fund could ever offer, the index can be green while dozens of its members are red. The costs are a management fee (often around 0.2%–0.4%), a larger account minimum, and a more complicated tax return. Automation is not free, and a mediocre platform can generate lots of tiny harvests whose benefit barely clears the fee, so weigh the fee against your actual marginal rate.

For investors who lean toward simple, low-turnover holdings, the harvesting upside is naturally smaller; a buy-and-hold dividend approach like the one in our SCHD dividend ETF guide generates fewer harvestable events by design, which is a feature, not a flaw. And if your taxable account is heavy in the volatile names covered in our AI stocks investment guide, the swings that make those positions nerve-wracking are exactly what create harvesting opportunities.

The mistakes that quietly erase the benefit

Most harvesting failures are not exotic, they are the same handful of errors repeated. The first is the accidental wash sale via dividend reinvestment. You harvest a fund on the 5th, forget that automatic reinvestment buys a few more shares of it on the 15th in your Roth, and part of your loss vanishes. Turn off auto-reinvest on the harvested security for the full window.

The second is harvesting in a 0% bracket, discussed above, giving up basis for a deduction worth nothing. The third is ignoring state tax, which cuts both ways: in a high-tax state your harvest is worth more than the federal number suggests, while in a no-income-tax state part of the benefit simply is not there, so a marginal harvest may not clear the friction. The fourth is over-trading a small account where commissions and spreads swallow the savings. The fifth is losing track of basis and carryforwards, if you do not record a disallowed loss added to replacement basis, or a loss carried forward on Schedule D, you can pay tax you already offset.

A last one worth naming: harvesting a loss you will need. If you sell a quality long-term holding purely for the tax break and the correlated replacement outperforms during your 31-day wait, the market can hand you a bigger opportunity cost than the tax you saved. The tax tail should not wag the investment dog.

Putting it together: a simple year-round process

Keep it boring and repeatable. Watch your taxable positions for meaningful unrealized losses, not just in December but whenever markets drop. Before selling, check the 61-day window across every account you and your spouse hold, including IRAs and any automatic reinvestment. Sell the loss, immediately buy a correlated-but-not-identical replacement to hold your market exposure, and record both the realized loss and the replacement’s new basis. At tax time, net short against short and long against long, apply the leftover to $3,000 of ordinary income, and carry the rest forward on Schedule D. Then repeat next year, because carryforwards can shelter gains for years to come.

Done with discipline, harvesting is one of the few reliably positive-expected-value moves available to a taxable investor. Done carelessly, it is a great way to disallow your own losses and complicate your return for nothing. The difference is entirely in the details covered above.

Read more

• Capital gains tax guide 2026: how investment gains and losses are taxed
• SCHD dividend ETF guide 2026: building a low-turnover income core
• AI stocks investment guide 2026: volatile names and how to manage them

---

This article is for general informational purposes only and is not tax, legal, or investment advice. It describes US federal tax rules in effect around the 2026 tax year; dollar thresholds, brackets, and rules change, and state treatment varies. Tax rules also differ significantly from one country to another. Consult a qualified tax professional or financial advisor about your specific situation before acting.