Qualified Charitable Distributions: the AGI Exclusion vs the Charitable Deduction (2026)

IRC §408(d)(8)(A) provides that the aggregate amount of qualified charitable distributions “shall not be includible in gross income” up to an annual cap — $111,000 per person for 2026, since IRS Notice 2025-67 states the limit “is increased from $108,000 to $111,000” (the once-per-lifetime split-interest election rises to $55,000). The transfer must go directly from the IRA to the charity, the owner must have “attained age 70½” on the date of the distribution (§408(d)(8)(B)), and — per IRS Pub 590-B — “a qualified charitable distribution will count towards your required minimum distribution.” No charitable deduction is allowed for it (§408(d)(8)(E)): none is needed, because exclusion from AGI is worth more than a deduction of the same size. AGI is the number that feeds the §86 Social Security taxability test, the IRMAA Medicare surcharge (whose statutory MAGI is “adjusted gross income … increased by … interest … exempt from tax”, 42 U.S.C. §1395r(i)(4)), and the OBBBA senior deduction's phase-out. A cash gift touches none of those: under 2026 law an itemizer deducts cash gifts only above a new 0.5%-of-AGI floor (§170(b)(1)(I), first effective this year), and the majority who take the standard deduction get at most the new §170(p) non-itemizer deduction — $1,000 single / $2,000 joint. Every number below is computed through the same parity-tested engine as the live QCD calculator.

Three routes, one mechanism

The table below is the whole study in qualitative form. Everything follows from where each route touches the return — above the AGI line or below it:

Worked example 1 — the deduction routes at full strength (single, 73)

$2,000,000 IRA, RMD $75,472, Social Security $24,000, gift $25,000. Note the itemized route is computed with everything 2026 law gives it — the 60%-of-AGI cash ceiling is slack here, the 0.5% floor costs only $479.36, and the itemizer keeps the senior deduction — and the QCD still beats it by $4,006.14, because only the QCD moves AGI:

The QCD's edge decomposes exactly (an identity the generator asserts at every one of the 792 grid rows): AGI falls by the excluded $25,000 of IRA income plus the taxable Social Security that exclusion pulls out of the §86 test — here the benefit stays at its 85% ceiling, so the whole edge is bracket relief plus the restored senior deduction ($1,252.30 more deduction at the lower MAGI).

Worked example 2 — the Social Security torpedo in reverse (MFJ, 73)

Married couple, both 73; $2,000,000 IRA, RMD $75,472; $60,000 of combined Social Security; gift $25,000. At the couple's income the §86 worksheet has their benefits at the 85% ceiling — $51,000.00 taxable. The QCD pulls provisional income back down the phase-in ramp, and $13,999.06 of Social Security drops out of income on top of the excluded distribution itself — the tax-torpedo mechanism running in the taxpayer's favor:

The QCD cuts this couple's federal tax by 52.1% ($8,980.60 → $4,300.72); the identical cash gift under the standard deduction saves $240.00.

Worked example 3 — IRMAA and the cap

The IRMAA flip (single, 80, $2M IRA, $40,000 Social Security, $25,000 gift): the baseline MAGI of $133,009.90 sits in the single tier-1 band (above $109,000), costing $1,148.40 a year in Medicare surcharges two years later. A $25,000 QCD lands MAGI at $108,009.90 — $990 under the floor — and the surcharge is $0. The same $25,000 in cash, standard or itemized, leaves MAGI exactly where it was: a below-the-line deduction cannot move an IRMAA tier, because the statutory MAGI is AGI plus tax-exempt interest.

Giving at the cap (single, 80, $500,000 IRA, $40,000 Social Security, $111,000 gift): the cap comfortably exceeds this IRA's $24,752 RMD, so a cap-sized QCD wipes the entire taxable distribution out, provisional income falls below the $25,000 threshold, taxable Social Security goes to $0, and the return owes $0.00. The identical cash gift — even itemized, with the deduction running into its 60%-of-AGI ceiling ($22,843.29 deductible after the floor) — leaves $954.88 due, because $13,639.60 of Social Security is already locked into AGI before any deduction applies:

When can the numbers cross?

One boundary worth stating plainly: the QCD's exclusion can shelter at most the IRA income actually being distributed. When the gift exceeds the RMD (and nothing beyond the RMD would have been withdrawn anyway), the extra QCD dollars come out of the IRA untaxed but shelter no further income this year — while a very large itemized cash gift keeps deducting (up to its ceiling). In this grid's sweep that crossover never materializes: the QCD route’s federal tax is at or below the best cash route in every one of the gifted scenarios, because the AGI exclusion keeps working on Social Security taxability and the senior deduction even after taxable IRA income hits zero. The generator proves the dominance claim only where it is provable — gift ≤ RMD — and counts the rest rather than assuming them (count: 0).

The full grid (792 scenarios)

Single and MFJ; ages 73, 75, 80 (Uniform Lifetime divisors 26.5, 24.6, 20.2); IRA balances $500,000, $1,000,000, $2,000,000; Social Security $24,000/$40,000 (single) and $40,000/$60,000 (MFJ); gifts $0–$50,000 in $2,500 steps plus the $111,000 cap. The table shows gifts of $5,000, $25,000, $111,000 (108 rows); the CSV/JSON carry the full sweep (792 rows) with per-strategy detail. Cash-itemized cells marked * are scenarios where itemizing never beats the standard route, so the best cash figure IS the standard one. “QCD saved /$1,000” includes the IRMAA change.

CC-BY-4.0 — free for any use including republication and journalism, with attribution to QuantCalc Research.

Assumptions

• Income model: ordinary income is the IRA's required minimum distribution (balance ÷ the Uniform Lifetime divisor from the RMD table) plus Social Security — no pension, no capital gains, no tax-exempt interest. On MFJ rows both spouses are the stated age, the IRA belongs to one spouse, and the gift comes from that one IRA (each spouse with an IRA has their own $111,000 cap; a two-IRA couple could give $222,000).
• QCD mechanics: the gift transfers directly from the IRA; it counts toward the RMD; taxable IRA income is max(RMD − gift, 0); amounts above the RMD still leave the IRA tax-free (up to the cap) but shelter no additional current-year income. Hybrid strategies (QCD to the RMD, cash for the rest) are out of scope.
• Deduction convention: the standard route is base + IRC §63(f) aged addition + OBBBA senior deduction, the same convention as the live calculators; the §170(p) non-itemizer deduction is added for cash gifts (min(gift, $1,000/$2,000); cash to public charities assumed — DAFs and supporting organizations do not qualify). Itemizers keep the senior deduction, lose the aged addition and §170(p), and have NO other itemized deductions in this model — a filer with large state-tax or mortgage deductions reaches the itemizing threshold sooner than shown.
• Charitable-law basis (2026, enacted): itemized cash gifts are capped at 60% of AGI (§170(b)(1)(G)) and reduced by the 0.5%-of-contribution-base floor (§170(b)(1)(I), P.L. 119-21 §70426, first effective for taxable years beginning after December 31, 2025); the contribution base is modeled as AGI. The §68 35% cap on the value of itemized deductions binds only in the 37% bracket, which no scenario here reaches (gate-enforced). Single-year view: carryforwards of ceiling-limited or floor-disallowed amounts are out of scope.
• IRMAA: the 2026 tier schedule applied to each scenario's MAGI at that filing status, per enrolled person — the surcharge that MAGI produces two years later. A married couple with both spouses enrolled pays any surcharge twice.
• The gift itself is not a tax strategy: in every scenario the charity receives the same amount and the giver is out of pocket the same amount before tax; the study compares only the tax consequences of the route. Giving appreciated stock — deducting fair value while never realizing the gain — is a fourth route with its own (30%-of-AGI) ceiling, out of scope here; its deduction is still below the AGI line, so everything this study says about AGI-keyed surfaces applies to it too.
• No state tax. Several states follow federal AGI, which would compound the QCD's exclusion advantage; out of scope.

Methodology

Related

• QCD Calculator — run your own balance, benefit and gift through the same engine
• How much can I give to charity from my IRA? — the cap, the age rule and the mechanics in plain language
• RMD Table 2026 — the Uniform Lifetime divisors used to size every scenario's distribution
• IRMAA Brackets 2026 — every tier, premium and the two-year lookback
• Tax Torpedo Calculator — the §86 phase-in this study runs in reverse
• 2027 vs 2026 key numbers — the delta dataset carrying the QCD cap rows cross-checked here
• All QuantCalc research

Changelog

• v2026.1 (2026-08-15) — initial release. 792 scenarios (2 filings × 3 ages × 3 balances × 2 benefit levels × 22 gift levels) × 3 giving strategies + baseline; 7 sanity-gate families + 3 mutation runs + 18 live-page cross-checks green at generation time.

Dataset license: CC-BY-4.0. QuantCalc is an independent retirement-planning research project. Not affiliated with, endorsed by, or sponsored by the IRS, SSA, or CMS. Educational research, not financial, tax, or legal advice.

@misc{quantcalc2026qualifiedcharitabledistributionstheagiex,
  title  = {Qualified Charitable Distributions: the AGI Exclusion vs the Charitable Deduction (2026)},
  author = {{QuantCalc Research}},
  year   = {2026},
  url    = {https://quantcalc.app/research/qcd-lifetime-tax-2026/},
  note   = {Accessed <date>}
}