Capital Gains Rate Planning (Section 1(h))

How it works

Long-term capital gains and qualified dividends are not taxed under the ordinary bracket schedule. Section 1(h) puts them on a separate ladder with three rates, 0%, 15%, and 20%, and where a given dollar of gain lands on that ladder depends on where it stacks on top of everything else on the return. Ordinary income fills the lower brackets first, and the gain is treated as sitting on top of that.

That stacking behavior is what makes the 0% tier a planning lever rather than a fixed outcome. In a year where ordinary income is genuinely low, whether from retirement or a gap between selling one business and starting the next, there is headroom above that ordinary income and below the 0% ceiling for gain to land at a zero federal rate. Recognizing gain into that headroom is a permanent result, not a deferral. Basis resets to the sale price, and nothing about that gain is ever recaptured later the way a Roth conversion's deferred tax treatment can still be undone by choices made in a later year.

A second, separate lever sits inside the same mechanism. Section 1091, the wash-sale rule, does not apply to a position sold at a gain, only to one sold at a loss. A long-term position can be sold to realize gain and bought back immediately, in the same account, with no wash-sale consequence at all. What changes is the basis. It resets higher, to the repurchase price, which is the entire point of doing this inside the 0% tier.

The remaining levers arrange when recognition happens rather than change the rate itself: holding a position past the one-year line so it qualifies as long-term at all, splitting a large gain across two tax years at the calendar boundary, or pairing gain recognition with losses harvested elsewhere so the net figure lands where it is wanted. None of this changes what an asset is worth. All of it changes when a taxable event happens and how large it is in a given year.

What this is worth in Florida

This is a case where Florida residency does not add a separate benefit so much as it removes a layer of noise. There is no Florida individual income tax, so a client here was never going to owe a state-level tax on this income regardless of timing. The federal computation described above is not one input among several here. It is the whole picture. In a state that also taxes capital gain, the same recognition decision has to satisfy two rate schedules at once, and the two do not always agree on what a good year looks like. Here there is only one schedule to model, which makes the bracket targeting cleaner, not larger.

Who this applies to

Anyone who realizes a long-term capital gain sits inside this rate schedule. What varies is how much the timing actually matters, which depends on two things: whether the holding-period test is met at all, and how much room a given year has around the gain.

• The holding-period gate. A gain only reaches this rate schedule if the position was held for more than one year before the sale. Exactly one year is not enough. The statute treats a holding period of one year or less as short-term and more than one year as long-term, so a position sold on the one-year anniversary of its purchase is still short-term. Short-term gain gets no preferential rate. It is taxed as ordinary income, in full, at whatever bracket the taxpayer is already in.
• Whoever has room between ordinary income and the 0% ceiling. Because the gain stacks on top of ordinary income, the value of this planning concentrates in a year where ordinary income is genuinely lower than usual: a retirement year before other income sources start, or a gap year between selling one business and starting the next. A taxpayer whose ordinary income alone already fills or exceeds the top of the schedule has little such room.
• Whoever is carrying losses forward from an earlier year. Carryover losses net against gain without the ordinary annual limit that caps how much loss can offset other income; that limit governs losses used against ordinary income, not losses netted against a capital gain.

Two categories of gain sit outside the ordinary rate ladder and are not reached by 0% harvesting the same way. Unrecaptured gain attributable to depreciation on real property is capped at 25%, and gain from collectibles or the taxable portion of qualifying small-business stock is capped at 28%. A taxpayer whose gain falls into either category needs a different rate conversation than the one described here.

What it requires

Recognizing the mechanism is the easy part. Getting the number right means running several thresholds against the same projected income at once, because more than one of them reacts to the same dollars.

• A holding period confirmed lot by lot, not position by position. The same security bought at different times can have some lots long-term and others not. The more-than-one-year test applies separately to each lot. Under Revenue Ruling 66-7 the holding period begins the day after the date of acquisition. The separate point that the trade date rather than the settlement date is the one that counts comes from a different line of guidance, not from that ruling.
• A stacking order that puts ordinary income first. Projected ordinary income for the year, wages, pension income, required distributions, business income, has to be estimated before the room available for 0% or 15% gain can be known. Treating a discretionary gain as filling the lower brackets before ordinary income overstates the room actually available.
• Net investment income tax exposure checked against a threshold that never moves. The 3.8% net investment income tax applies to the lesser of net investment income or the amount by which modified adjusted gross income exceeds a fixed statutory floor: $250,000 for a married couple filing jointly or a surviving spouse, $125,000 filing separately, $200,000 for everyone else. Unlike the rate brackets, this floor carries no inflation adjustment. Capital gain counts as investment income for this purpose, so a large enough gain can turn an effective 15% rate into 18.8%, and an effective 20% rate into 23.8%.
• Every MAGI-driven cliff modeled at the same time as the rate, not after it. The same dollars that determine the capital gains rate also raise modified adjusted gross income for programs that do not share this rate schedule's patience: a Medicare surcharge keyed to income reported two years earlier, and a marketplace premium tax credit keyed to the current year's income. Neither program cares whether the gain was taxed at 0%.
• The qualified business income deduction run in the same model. For a pass-through business owner, the Section 199A deduction is computed on taxable income minus net capital gain, so recognizing more gain shrinks the deduction's base directly, and pushing total taxable income higher in the process can move an owner further into that deduction's own phase-out range, a topic I cover in my QBI deduction strategy page.

What you need to document

The paperwork here is about being able to show, lot by lot, exactly what happened and when, not about proving a position was reasonable.

Trade-date records for every lot

Purchase confirmations showing the trade date, not the settlement date, for each lot of a position that is only partly long-term. The one-year clock runs from the day after the trade date.

A specific-lot identification, made before settlement

An instruction to the broker identifying which lot is being sold, given before the trade settles. Without it, the default is first-in-first-out, which can sell the wrong lot for the intended holding period or gain amount.

The acquisition history behind a gifted or inherited position

For gifted property, the donor's original acquisition date, since the recipient's holding period tacks onto the donor's. For inherited property, confirmation that the position passed through inheritance, since it is automatically long-term regardless of how long the decedent or the heir actually held it.

A running total of carryover losses

The amount of capital losses carried forward from prior years, so they can be netted against a harvested gain with an accurate number rather than an estimate.

The income projection behind the cliff calculations

Whatever supports the projection of ordinary income and MAGI, so the harvest decision can be reconstructed later rather than taken on faith.

Where it goes wrong

None of this is a listed or reportable transaction, and recognizing a gain in a chosen year is not, by itself, an aggressive position. The exposure sits in getting the mechanics wrong, or in looking at the rate without looking at what else moves with it.

• Miscounting the holding period. The most common error is treating exactly one year as long enough, or measuring from the settlement date instead of the trade date. Either mistake can convert gain that was modeled at a preferential rate into ordinary income taxed at up to 37%.
• Letting the broker's default lot selection decide the outcome. Without a specific identification made before settlement, first-in-first-out applies automatically, and the lot it picks is not necessarily the one with the longest holding period or the smallest gain.
• Stacking the gain in the wrong order. The gain sits on top of ordinary income, not underneath it. A projection that treats the 0% bracket as available before ordinary income is subtracted out overstates the room that exists.
• Harvesting gain in a state that was left too soon. A taxpayer who realizes gain before completing a move to Florida, while still domiciled in a state that taxes capital gain, gets none of the benefit described here. What changes is whether a state also takes a share, and that turns on when domicile actually changed, not on when the paperwork says it did.

The two-year echo in Medicare and marketplace coverage

A harvest that lands at 0% federal tax can still cost money later, through a door the rate schedule never mentions. The Medicare Part B and D surcharge for a given year is set from income reported two years earlier, and it behaves as a cliff rather than a slope: crossing into the next tier by even a small amount applies that tier's full surcharge for the whole year. A gain that looked free when it was recognized can raise the income figure a Medicare enrollment decision two years later is based on, well after the return that caused it is filed. The same gain also raises the income figure a pre-Medicare taxpayer's marketplace premium tax credit depends on, in the same year it is recognized, which can claw back subsidy already advanced. Both effects run independently of whether the gain itself was taxed at 0%, 15%, or 20%, and a plan that checks only the rate will miss both of them.

Erosion on the pass-through side

For a business owner who also has pass-through income, a capital gain does not stay contained to the capital gains rate schedule. The qualified business income deduction is computed on taxable income after net capital gain is subtracted out, so a larger gain narrows that base directly, and it also raises total taxable income, the figure that determines whether the owner sits inside or past that deduction's own phase-out range. A harvesting decision made without checking the QBI computation can quietly shrink a deduction on income the gain had nothing to do with.

A situation where this comes up

The version I see most often is a taxpayer between two income states: retired but not yet drawing on every income source, or between selling one business and starting the next, with a year where ordinary income is genuinely, and for once verifiably, lower than what surrounds it. The question is whether to use that year to recognize gain that would otherwise be recognized later, once ordinary income has returned to its usual level and the headroom is gone.

What decides the answer is never the capital gains rate alone. A client on a marketplace health plan needs the premium tax credit modeled against the same gain first, since the subsidy given up can be larger than the tax avoided. A client two years from Medicare enrollment needs the same gain checked against the income figure that enrollment will eventually be priced from. A business owner needs the qualified business income deduction run against the same number. None of these checks changes the mechanism above; they decide whether using it in a given year actually helps.

The version that worries me is the harvest done on autopilot: a rule of thumb that says harvest gain whenever income looks low, applied without checking whether this is one of the years a nearby program is watching the same number. The mechanism itself is not the risk. Treating it as the only number that matters is.