Capital Gains Harvesting and the 400% FPL Boundary

Capital gains occupy a confusing place in the ACA Premium Tax Credit calculation because the federal income tax system treats long-term capital gains as a separately-rated income class under IRC § 1(h) — with 0%, 15%, and 20% brackets — but the PTC computation under IRC § 36B uses Modified Adjusted Gross Income, which is defined to include all capital gains net of capital losses. So a household whose long-term capital gains are taxed at 0% federally can still see those same gains push them across the 400% FPL cliff and forfeit thousands of dollars in PTC. The 0% LTCG rate and the PTC live in different tax universes that intersect only through MAGI.

The LTCG 0% bracket for 2026, indexed under Rev. Proc. 2025-32, applies to taxable income up to roughly $48,350 (single) and $96,700 (MFJ). The 400% FPL cliff for tax year 2026 sits at approximately $62,640 (single, HH=1) and $84,600 (MFJ, HH=2) using 2025 HHS poverty guidelines. The two thresholds are close but not identical, and the FPL is keyed to MAGI while the LTCG bracket is keyed to taxable income (MAGI minus standard or itemized deduction, with some adjustments). For an MFJ household of two taking the 2026 standard deduction (~$31,500), MAGI of $84,600 produces taxable income of roughly $53,100 — well inside the 0% LTCG bracket. So that household could realize, say, $20,000 of long-term capital gains, owe zero federal capital gains tax, AND lose its entire PTC if those gains pushed MAGI from $74,600 to $94,600.

The planning question is therefore not "what is my LTCG tax rate" but "where does each realization put my MAGI relative to the cliff." For households below the cliff with unrealized appreciation, two distinct tactics emerge. Tax-loss harvesting under IRC § 165(f) and § 1211 sells underwater positions to generate realized losses that offset realized gains (net capital losses up to $3,000/year deduct against ordinary income, with the remainder carried forward indefinitely under § 1212(b)). Reverse gain harvesting — sometimes called tax-gain harvesting — does the opposite: in a year when MAGI is unusually low, the household intentionally sells appreciated positions, realizes the gain at the 0% LTCG rate, and immediately repurchases the same security to step up basis at no federal capital gains cost. Both tactics are bounded by the same MAGI ceiling: realized gains, harvested or otherwise, count.

The wash-sale rule under IRC § 1091 is the principal constraint on loss harvesting. The rule disallows a loss deduction if the taxpayer acquires "substantially identical" stock or securities within 30 days before or after the sale. Practical implication: the harvest can sell SPY at a loss and immediately buy IVV (different issuers, both tracking the S&P 500) without triggering the rule — because IVV is not "substantially identical" to SPY under IRS Rev. Rul. 2008-5 and related guidance, even though they track the same index. The rule is asymmetric: it applies to losses only, not gains. So reverse gain harvesting can sell and immediately repurchase the exact same security with no wash-sale exposure (sometimes called the "tax-gain wash" — the basis steps up and the position is maintained).

Worked example. An MFJ couple in 2026 has baseline MAGI of $70,000 (wages plus modest dividends), is enrolled on the marketplace, and receives roughly $9,000 of annual PTC. They hold a brokerage position with $20,000 of unrealized long-term gains they would like to monetize for a home renovation. Option A: sell all $20,000 in 2026. New MAGI = $90,000, which is above the ~$84,600 cliff for HH=2; PTC = $0; clawback of any APTC received; net cost of the realization = $9,000 lost PTC + ~$0 LTCG tax (still within 0% bracket because taxable income of ~$58,500 after standard deduction is under the $96,700 0% ceiling). Option B: realize $14,000 in 2026 (MAGI = $84,000, just under cliff, PTC preserved) and $6,000 in 2027 (when MAGI will likely also be ~$76,000). LTCG tax is $0 in both years; PTC is $9,000 in 2026 and ~$9,000 in 2027. Option B is $9,000 better than Option A for the cost of waiting six months on $6,000 of proceeds.

When Roth conversions and capital gains harvesting are both contemplated in the same year, the joint MAGI ceiling becomes binding. A common error is treating each as having its own cliff budget; in reality they share one budget. The simplest sequencing rule: subtract baseline MAGI from the cliff, designate that headroom for ONE of the two (whichever has greater time-value urgency), and defer the other to a different year. A household with $14,000 of cliff headroom can convert $14,000 to Roth OR harvest $14,000 of gains — not both.

The 0% LTCG bracket and the 400% FPL cliff are independent constraints in the Internal Revenue Code, and the binding constraint for cliff-proximate marketplace enrollees is almost always the FPL cliff rather than the LTCG bracket. A household can owe zero federal capital gains tax and still pay $10,000-$15,000 in lost PTC because of those same gains. The two thresholds use different base measures (taxable income vs. MAGI), and standard deductions sit between them — for MFJ with the 2026 standard deduction of ~$31,500, MAGI of $84,600 (right at the FPL cliff) produces taxable income of ~$53,100, which is comfortably inside the $96,700 LTCG 0% ceiling.

The wash-sale rule under IRC § 1091 only attaches to losses, never to gains. This asymmetry is the foundation of tax-gain harvesting: a taxpayer can sell SPY at a $5,000 gain and immediately repurchase SPY the same day with no wash-sale consequence — basis steps up, position is maintained, and the realized gain (if within the 0% bracket and below the FPL cliff) carries no federal tax cost. Reverse direction (loss harvesting) requires either a 31-day wait or substitution with a non-substantially-identical security. The IRS has never published a bright-line list of "substantially identical" pairs, but practitioner consensus treats SPY/IVV/VOO as non-identical to each other, while SPY and an SPY-tracking index futures contract would likely be identical.

Mutual fund capital gains distributions are a perennial cliff hazard because they arrive in November and December — too late to offset with most planning moves. A taxpayer holding an actively managed equity fund in a taxable brokerage account can receive a 1099-DIV distribution of 3-8% of the fund's NAV in any given year, even if the taxpayer made no trades. For cliff-proximate households, switching from active to passively-managed funds (which distribute much smaller capital gains, often less than 1%) inside taxable accounts is a structural cliff-defense move that requires no ongoing attention.

Loss carryforwards under IRC § 1212(b) are unlimited in time and amount, surviving until used. A taxpayer with a $40,000 carryforward from the 2022 bear market can absorb $40,000 of realized gains in 2026 with zero net capital gain hitting MAGI. This is one of the few "free" cliff-management tools because the loss was already realized in a prior year — it costs nothing to use it now. Cliff-proximate households should always run Schedule D line 14 (long-term carryover) and line 6 (short-term carryover) into their year-end MAGI projection before deciding to defer additional realizations.