Tax-Loss Harvesting Math: Dollar Savings Analysis

A $50,000 realized loss does not save a $150k+ investor $50,000. Against long-term capital gains (LTCG), it shelters dollars taxed at a combined federal rate of up to 23.8% — meaning the same $50,000 loss is worth $11,900 in a top-bracket portfolio and as little as $7,500 in the 15% LTCG band. The number that matters in tax-loss harvesting is not the size of the loss. It is the marginal rate the loss displaces.

That distinction gets buried under a familiar headline: harvest losses, lower your taxes. True, but uselessly vague. The dollar value of a harvested loss is entirely a function of what it offsets — short-term capital gains (STCG) taxed as ordinary income, LTCG taxed at preferential rates, or up to $3,000 of ordinary income annually. Each carries a different rate, and the gap between them is where the real money sits.

The numbers that define the harvest

Five figures determine what any harvested loss is actually worth. They are summarized below for featured-snippet reference, then unpacked in detail.

Sources: IRS Revenue Procedure 2024-40 (2025 LTCG thresholds); IRS Topic 409 and Publication 550 (loss deduction limit, wash sale rule, carryforward); IRC §1411 (NIIT). Accessed June 2026.

The 40.8% top figure assumes a short-term gain taxed at the 37% ordinary bracket plus the 3.8% net investment income tax (NIIT). The 23.8% figure is the long-term equivalent: the 20% LTCG rate plus NIIT. Note the spread. A loss applied against a short-term gain is worth nearly twice as much, per dollar, as the same loss applied against a long-term gain — a point most coverage of short-term versus long-term capital gains mentions only in passing.

Loss value by offset type and bracket

Order of operations matters. The IRS requires losses to first offset gains of the same character — short-term losses against short-term gains, long-term against long-term — before any cross-netting, and only the residual net loss flows to ordinary income, capped at $3,000 per year per IRS Topic 409. The character of the gain you are offsetting is therefore the single largest driver of the harvest’s value.

Consider a $40,000 harvested loss applied three different ways for a $150k+ household. The dollar savings diverge sharply.

Sources: IRS Revenue Procedure 2024-40; IRS Topic 409; IRC §1411 (NIIT). Rates shown are top-of-bracket; actual displaced rate depends on the household’s taxable income. Accessed June 2026.

The bottom row exposes the most common misconception. If a household has no realized gains to offset and harvests $40,000 in pure losses, only $3,000 reduces ordinary income this year — saving roughly $1,110 at the 37% bracket. The remaining $37,000 does not vanish; it carries forward indefinitely, retaining its short- or long-term character. But the immediate cash value of a loss with nothing to offset is small. Harvesting is most powerful when paired with a realized gain of matching or higher-taxed character.

The wash sale rule is the real constraint

Rates set the ceiling on harvest value. The wash sale rule sets the floor on what survives. Under IRC §1091 and IRS Publication 550, a loss is disallowed if a substantially identical security is acquired within 30 days before or 30 days after the sale — a 61-day window that includes the sale date itself. The disallowed loss is not forfeited outright in a taxable account; it is added to the cost basis of the replacement shares and the holding period tacks on. The economic effect is deferral, not loss.

The trap that catches sophisticated investors is account scope. The rule applies across all accounts a taxpayer controls, including a spouse’s accounts and IRAs. Fidelity notes that a replacement purchased inside an IRA permanently forfeits the loss under Revenue Ruling 2008-5 — the basis adjustment that normally preserves the deferral does not apply to the IRA. A December harvest in a taxable account, undone by an automatic dividend reinvestment in a spousal IRA two weeks later, converts a planned deduction into a disallowed one. This interaction with the tax cost of selling early is where harvesting strategies most often quietly fail.

The Finluxy After-Tax Gain Rate, with and without harvesting

The cluster’s proprietary metric isolates what harvesting actually changes. The Finluxy After-Tax Gain Rate is net gain after all applicable taxes — federal LTCG, NIIT, and state — divided by original cost basis, expressed as a percentage. The tax haircut is the pre-tax gain rate minus the after-tax rate. Applying a harvested loss against a realized gain shrinks the haircut directly.

Take a high-income California household — the harshest combination of capital gains tax across the 50 states. Stock bought at $100,000 (cost basis) sold for $300,000 after three years: a $200,000 long-term gain, a 200% pre-tax gain rate. The combined rate stacks the 20% LTCG rate, 3.8% NIIT, and California’s 13.3% top marginal rate for 37.1% total.

Sources: IRS Revenue Procedure 2024-40 (20% LTCG); IRC §1411 (3.8% NIIT); California Franchise Tax Board / Tax Foundation (13.3% top marginal rate). Cost basis $100,000. Net gain in harvested row reflects tax saved on the offset gain; the harvested asset’s own loss is a separate realized position. Accessed June 2026.

The harvested $50,000 long-term loss removes $50,000 from the taxable gain, cutting the tax bill from $74,200 to $55,650 — a $18,550 reduction. That is the 37.1% combined rate applied to $50,000. The pre-tax gain rate stays 200%, but the tax haircut on the offset gain narrows by the full combined rate. In a no-state-tax jurisdiction, the same $50,000 long-term offset would save only $11,900 (23.8%) — which is precisely why California’s capital gains burden makes harvesting more valuable there than anywhere else.

What most coverage overlooks

Harvesting guides obsess over the $3,000 ordinary-income deduction and the 30-day wash window. The overlooked variable is character matching. The data shows the harvest’s dollar value swings by a factor of more than two — from $0.188 per dollar in the 15% LTCG band to $0.408 per dollar against a top-bracket short-term gain — based solely on what the loss offsets. A long-term loss applied against a long-term gain captures 23.8% at most federally. The same loss, if it can be matched against a short-term gain through deliberate timing of which positions to realize, captures up to 40.8%.

This means the highest-value harvesting move is not selling the biggest loser. It is sequencing realizations so that losses land against the highest-taxed gains available — short-term gains and gains exposed to the 3.8% NIIT threshold. Most investors harvest in December against whatever gains they happen to have. The data argues for harvesting against the gains they choose to have.

The $150k+ household calculus

At $150k+ taxable income, a household is almost certainly past the NIIT threshold ($250,000 MAGI for married filing jointly, $200,000 single per IRC §1411) on at least part of its investment income, and likely in the 15% or 20% LTCG band. That changes the harvesting decision in two concrete ways.

First, the breakeven on transaction friction is low. A harvested loss offsetting a 23.8% federally taxed gain returns $238 per $1,000 — far exceeding any realistic bid-ask or advisory cost on a liquid position, so the bias should be toward harvesting whenever a meaningful loss exists alongside a gain to absorb it. Second, the carryforward is an asset, not a consolation prize. A household selling a business, exercising stock with a large after-tax gain, or liquidating a concentrated position in a future year can bank harvested losses now and deploy them against that event. Unused losses carry forward indefinitely with their character intact — a long-term loss harvested today still offsets a long-term gain a decade out.

The trade-off is opportunity cost and tracking-error risk during the 31-day exclusion window, plus the discipline to avoid the cross-account wash sale that silently voids the deduction. For a household coordinating taxable accounts, spousal IRAs, and automatic reinvestment, the harvest is worth the most precisely when it is hardest to execute cleanly — which is the argument for treating it as a year-round sequencing decision rather than a December scramble, ideally modeled against a full read of the capital gains tax framework for $150k+ investors before any position is sold.