This article is educational and not individualized tax advice; consult a licensed CPA or tax attorney before acting. All tax thresholds reflect the 2026 tax year unless a figure’s year is labeled inline.
TL;DR — Quick Verdict
- Harvested losses offset capital gains dollar-for-dollar with no cap, but only $3,000 of net loss ($1,500 married filing separately) can hit ordinary income per year — a limit unchanged since 1978 (IRC §1211(b)).
- At a 37% short-term bracket, that $3,000 ordinary-income deduction is worth $1,110 in cash; the real money is in wiping out gains taxed at up to 23.8%.
- Excess losses carry forward indefinitely (IRC §1212(b)) — they never expire while you’re alive.
- The wash-sale rule (IRC §1091) disallows the loss if you rebuy a substantially identical security inside a 61-day window; it also catches your IRA and your spouse’s account.
- DIY costs $0 in fees; Wealthfront and Betterment automate it for a 0.25% advisory fee — worth it above roughly $50,000–$100,000 of taxable assets.
- Recommendation: Harvest deliberately in volatile years, respect the 61-day clock, and treat the $3,000 line as a bonus, not the point.
Roughly 96% of Wealthfront’s taxable clients have had their entire advisory fee covered by harvested losses, according to the firm’s own 2024 performance disclosure — a striking figure that also reveals how badly most investors misread the strategy. Tax-loss harvesting is not a way to “write off $3,000.” That cap, set in Internal Revenue Code §1211(b), is the smallest part of the machine. The larger engine is offsetting capital gains, which carries no annual limit at all. This article shows the real 2026 math: what a harvested loss is worth against gains versus against your salary, how the wash-sale rule quietly voids losses you thought you’d banked, and when paying a robo-advisor like Wealthfront or Betterment beats doing it yourself in a Fidelity or Schwab brokerage account. Every dollar figure below is modeled against the IRS’s 2026 brackets from Revenue Procedure 2025-32, so you can map the mechanics onto your own bracket rather than a generic example.
What Tax-Loss Harvesting Actually Does to Your Tax Bill
Selling a losing investment converts a paper loss into a realized capital loss — a usable tax asset. The IRS makes you apply it in a fixed order. First, net short-term gains and losses against each other, then net long-term against long-term, then net the two results. Only the final net loss reaches ordinary income, and only up to $3,000 per year.
Here is the distinction most guides blur. Against capital gains, losses have no ceiling: a $200,000 harvested loss can erase a $200,000 gain and drop that tax to zero. Against ordinary income — your wages, interest, business profit — the deductible slice is capped at $3,000. That asymmetry drives every smart harvesting decision.
Consider a concrete case. You sell a fund for a $50,000 loss. In the same year you also sold appreciated stock for a $45,000 long-term gain. The loss cancels the entire gain, leaving a $5,000 net loss. Of that, $3,000 comes off your salary this year; the remaining $2,000 carries forward. The gain that would have been taxed at up to 23.8% (the 20% top long-term rate plus the 3.8% net investment income tax) simply vanishes from your return. Understanding how your capital gains rates by holding period interact with harvested losses is the difference between a modest deduction and a five-figure deferral.
The Real 2026 Savings Math, Bracket by Bracket
A loss is only worth the tax rate it offsets. That single principle produces very different dollar outcomes depending on what the loss cancels and what bracket you’re in. The table below models the $3,000 ordinary-income deduction across 2026 marginal brackets, then contrasts it with offsetting a $3,000 long-term gain.
Modeled by Real Cost Report using 2026 marginal rates from IRS Revenue Procedure 2025-32 and the 3.8% NIIT under IRC §1411. Verify at irs.gov.
Notice the counterintuitive result: a loss thrown against ordinary income at 37% is worth more per dollar ($1,110) than the same loss thrown against a long-term gain at 15% ($450). This is why practitioners prefer to realize long-term losses and route the $3,000 against salary — long-term gains are already taxed cheaply, so spending a loss there wastes its highest-value use. The 3.8% NIIT surtax kicks in once modified adjusted gross income tops $200,000 single or $250,000 married filing jointly, thresholds frozen by statute and never indexed, which quietly pulls more households into its reach each year. If your income hovers near that line, review the NIIT surtax rules and avoidance strategies before you harvest.
What Determines Whether a Harvest Is Worth the Effort
Three variables decide the payoff: the size of the loss available, the rate it offsets, and whether you trigger a wash sale. A retiree in the 0% long-term bracket — taxable income under $49,450 single or $98,900 married filing jointly in 2026 — gains almost nothing from harvesting against gains, because those gains were already tax-free. For that person, the strategy flips: harvesting gains to reset basis often beats harvesting losses.
Timing compounds the effect. Markets that drop sharply mid-year create harvesting windows that a flat market never offers. Wealthfront’s disclosed 10-year average “harvesting yield” of 4.23% of portfolio value shows how much raw loss volatility can surface, though that figure reflects continuous automated selling most DIY investors won’t match. A single manual harvest in a down quarter can still bank thousands in carryforward.
The trap is transaction friction. Every sale you make to harvest must avoid rebuying the same position for 31 days, which means sitting in a correlated-but-different fund and accepting slight tracking drift. For most diversified investors that drift is trivial; for a concentrated single-stock holder it can be material. Weighing harvesting alongside your asset location strategy across account types keeps you from harvesting in a taxable account while an identical fund quietly triggers a wash sale in your IRA.
The Wash-Sale Rule: How Losses Silently Disappear
The single most expensive mistake in harvesting is the wash sale. Under IRC §1091, if you sell a security at a loss and buy a “substantially identical” security within 30 days before or after the sale — a 61-day window including the sale day — the IRS disallows the loss. The disallowed amount isn’t lost forever; it’s added to the cost basis of the replacement shares, deferring the benefit until you sell those.
Where investors get burned is the reach of the rule. Per IRS Publication 550, a wash sale is triggered by a purchase in your spouse’s account or a corporation you control, and — critically — by a repurchase inside your IRA or Roth IRA, where the disallowed loss vanishes permanently because there’s no taxable basis to adjust. Automatic dividend reinvestment is the most common accidental trigger: a scheduled reinvestment three weeks after your harvest quietly voids part of the loss.
Source: IRC §1091 and IRS Publication 550, Investment Income and Expenses. Verify at irs.gov.
One notable gap: cryptocurrency is not currently subject to §1091, which applies to “stock or securities,” so crypto investors can harvest losses and immediately rebuy — though legislation to close this has been proposed repeatedly. Wash-sale interactions with retirement accounts overlap heavily with how a backdoor Roth IRA process for high earners routes money between account types, so coordinate the two.
DIY Brokerage vs. Robo-Advisor: Which Is Better for Automated Harvesting?
Two paths exist. Do it yourself in a Fidelity, Schwab, or Vanguard brokerage account for zero fees, or let Wealthfront or Betterment run daily automated harvesting for a 0.25% annual advisory fee. The choice hinges on portfolio size and how much loss volatility your holdings actually generate.
DIY costs nothing but demands attention: you must spot the losses, avoid wash sales across every account, and file Form 8949 with the correct code “W” adjustments. A robo-advisor scans daily, swaps into pre-vetted alternate ETFs, and coordinates across your linked accounts to sidestep wash sales — Wealthfront reports wash sales affect under 0.01% of daily dollars traded on its platform.
The fee is the catch. On a $150,000 taxable account, 0.25% is $375 per year, every year, whether or not the market cooperates. In a flat or rising year with few losses to harvest, you pay the fee for little benefit. Direct indexing — Wealthfront and Betterment both offer it, generally above a $100,000 balance — multiplies harvesting opportunities by holding individual stocks instead of a single fund, which is where the fee earns its keep.
Verdict
Below roughly $50,000 in taxable assets, DIY wins — the harvestable losses are too small to justify $125+ per year in fees. Between $50,000 and $100,000, it’s a toss-up that favors automation if you won’t reliably do it yourself. Above $100,000, especially with direct indexing and a high marginal rate, a 0.25% robo fee is usually worth it: the harvesting yield on a stock-level portfolio typically clears the fee several times over in volatile years, though not in calm ones. Retirees in the 0% long-term bracket should skip automated harvesting entirely — there’s little gain to offset.
What Most People Get Wrong About Harvesting
Even careful investors repeat the same errors. Each one has a specific consequence and a specific fix.
Mistake 1: Treating the $3,000 cap as the goal
Consequence: They harvest just enough to hit $3,000 against income and stop, leaving large unrealized losses that could have erased six-figure gains. Correct action: Harvest all available losses against gains first — that offset is uncapped — and treat the $3,000 ordinary-income deduction as leftover.
Mistake 2: Rebuying too soon
Consequence: A repurchase on day 25 voids the loss and, if inside an IRA, destroys it permanently. Correct action: Mark day 31 on the calendar, pause dividend reinvestment on the sold security, and check your spouse’s accounts.
Mistake 3: Wasting long-term losses on long-term gains
Consequence: A long-term loss cancels a gain taxed at only 15%, when it could have offset ordinary income at 37%. Correct action: Where you have a choice, route long-term losses to the $3,000 ordinary-income deduction and use short-term losses against short-term gains.
Mistake 4: Ignoring the reset-basis problem
Consequence: Harvesting lowers your cost basis, so you’ve deferred tax, not eliminated it — a larger gain waits when you eventually sell. Correct action: Harvest with an exit plan, and know that under current law a step-up in basis for inherited assets can erase that deferred gain entirely at death, making late-life harvesting especially powerful.
Who Should Harvest — and Who Shouldn’t
The strategy pays off for specific profiles and does nothing for others. High earners in the 32%–37% brackets with sizable taxable accounts get the most: they have gains worth offsetting, income worth deducting against, and enough loss volatility to surface real losses. Anyone facing a large one-time gain — a business sale, a concentrated stock windfall, real estate — can bank carryforward losses in advance to cushion it.
Harvesting does little for three groups. Retirees in the 0% long-term bracket have no expensive gains to offset. Investors holding everything in tax-advantaged accounts — 401(k)s, IRAs — can’t harvest at all, since those accounts have no taxable capital gains. And anyone whose entire taxable portfolio sits in a single broad-market fund with steady appreciation simply has no losses to harvest in most years.
The break-even logic is conditional: if your marginal rate is high, your taxable balance is substantial, and you’ll face gains within a few years, harvesting is clearly worth the effort or the fee. If any one of those is missing, the benefit shrinks fast. For investors weighing harvesting against other deferral tools, compare it to a 1031 exchange for deferring real estate gains or, for concentrated appreciated stock, a donor-advised fund’s costs and tax benefits — each solves a different piece of the gains puzzle. Those planning large IRA-to-Roth moves should also check how harvesting interacts with a Roth conversion’s effect on Medicare premiums, since both shift taxable income in the same year.
Frequently Asked Questions
Can I harvest more than $3,000 in losses in one year?
Yes — the $3,000 limit (IRC §1211(b)) applies only to net losses deducted against ordinary income. Losses offset capital gains with no annual cap. If you have $80,000 in gains and $80,000 in harvested losses, all $80,000 offsets, and you owe zero on those gains. Only leftover net losses beyond your gains are subject to the $3,000 ceiling, with the rest carried forward.
How long do carried-forward losses last?
Indefinitely, under IRC §1212(b). Per IRS Publication 550, capital-loss carryforwards for individuals last until fully used or until death — there’s no expiration date. A $30,000 net loss could take a decade to absorb at $3,000 per year against income, or be consumed instantly by a single large gain. The character (short- or long-term) is preserved as it carries.
Does the wash-sale rule apply to cryptocurrency?
Not currently. IRC §1091 applies to “stock or securities,” and the IRS has not classified crypto as a security for this purpose, so an investor can sell Bitcoin at a loss and rebuy it immediately. Congress has proposed closing this gap several times without success as of mid-2026, so treat it as a rule that could change and confirm current guidance before relying on it.
Is a robo-advisor’s 0.25% fee worth it for harvesting?
It depends on your balance and volatility. On $150,000, 0.25% is $375 yearly. Wealthfront reports its average client’s harvesting benefit has run several times the fee over an account’s life, but that reflects continuous automation across many holdings. Below about $50,000 taxable, DIY in a free brokerage usually wins; above $100,000 with direct indexing and a high tax rate, automation typically justifies the cost.
How We Researched This Article
Every tax threshold, rate, and rule in this article was verified against primary federal sources before publication. The $3,000 net-loss deduction limit and its indefinite carryforward were confirmed against Internal Revenue Code sections 1211(b) and 1212(b), and cross-checked against IRS Publication 550, Investment Income and Expenses. The wash-sale mechanics — the 61-day window, the “substantially identical” standard, and the IRA and spousal traps — were drawn from IRC §1091 and Publication 550, with the crypto exclusion confirmed against the statute’s “stock or securities” language and corroborated by the U.S. Securities and Exchange Commission’s investor guidance at investor.gov.
The 2026 long-term capital gains brackets (0%, 15%, 20%), their taxable-income thresholds, and the 37% top ordinary rate applied to short-term gains come from IRS Revenue Procedure 2025-32, the official annual inflation-adjustment guidance published by the Internal Revenue Service. The 3.8% net investment income tax and its frozen $200,000/$250,000 thresholds are set by IRC §1411. Robo-advisor pricing and harvesting-benefit estimates were taken from Wealthfront’s and Betterment’s published disclosures, treated as trade sources that contextualize but do not substitute for the statutory figures; those benefit multiples reflect each firm’s own methodology and self-reported client data, not measured universal outcomes.
All bracket-based dollar savings in the tables are modeled calculations, not measured averages — they multiply the $3,000 cap or a $3,000 gain by the stated 2026 rate to show marginal value. Individual results vary with state taxes, filing status, and the specific character of gains and losses, none of which this article can model for a particular reader. Where firm-reported harvesting yields appear, they describe historical platform performance and do not predict future results. This analysis was last conducted in July 2026 against then-current IRS guidance; because thresholds adjust annually, verify the current year’s figures before filing. All figures were verified against named primary sources before publication.