Which Year-End Move Cuts a Retiree's 2026 Tax Most?
Which year-end tax move cuts a retiree's 2026 tax the most?
It depends on the household. Two moves led most often, tied: QCD instead of cash (median change −$1,717 where it applied) and Harvest losses (median change −$607 where it applied) cut the 2026 total most for 5 of the 12 households each. No single move won everywhere: 3 different moves came first. A Roth conversion never lowers 2026 tax; $25,000 converted cost 16.0% to 98.7% of the amount in 2026, depending on the household.
| Households x moves x sizes | 12 x 7 x 3 |
|---|---|
| Harvest losses: largest cut for | 5 of 12 households (median change −$607, 9 eligible) |
| Harvest gains at 0%: largest cut for | 0 of 12 households (median change $2,240, 8 eligible) |
| Bunch charitable giving: largest cut for | 2 of 12 households (median change $0, 10 eligible) |
| QCD instead of cash: largest cut for | 5 of 12 households (median change −$1,717, 6 eligible) |
| Take the rest of the RMD: largest cut for | 0 of 12 households (median change $2,482, 4 eligible) |
| Q4 estimate or December withholding: largest cut for | 0 of 12 households (median change −$34, 10 eligible) |
| $25,000 Roth conversion, 2026 cost | 16.0% to 98.7% of the amount |
| Joint run differs from the sum of singles | 9 of 9 households (largest gap $2,552) |
Twelve retiree households, each run through the seven moves a retiree can still make for 2026, at three sizes each, by the module behind the Year-end Tax Moves tool. Every number is the change in the household's 2026 total: federal income tax (with any RMD excise and underpayment penalty), state income tax, the IRMAA surcharge 2026 income sets for 2028 Medicare premiums, and the ACA premium tax credit.
The twelve households
| Id | Household | 2026 total before |
|---|---|---|
| H01 | Single, 74, Ohio: Social Security, IRA withdrawals, gives to charity | $1,458 |
| H02 | Couple, 73 and 71, California: pension, Social Security, an RMD not yet taken | $6,498 |
| H03 | Single, 66, Texas: large IRA, near the first IRMAA tier | $15,203 |
| H04 | Couple, both 62, Ohio: on an ACA plan, living on savings | −$18,836 |
| H05 | Couple, 76 and 74, Florida: an RMD short by $20,000, generous givers | $10,643 |
| H06 | Single, 70, Pennsylvania: Social Security and pension, a loss carried in | $863 |
| H07 | Couple, both 68, New York: a big capital gain this year | $17,467 |
| H08 | Single, 72, Georgia: pension and IRA, no investment income | $8,448 |
| H09 | Couple, 75 and 70, Arizona: underpaid estimated tax | $10,606 |
| H10 | Single, 64, Colorado: on an ACA plan, near 400% of the poverty line | −$5,318 |
| H11 | Head of household, 67, Illinois: Social Security and IRA withdrawals | $1,320 |
| H12 | Couple, both 73, Massachusetts: the first RMD year, no investment income | $3,888 |
Which move cuts each household's 2026 total most?
| Household | Harvest losses | Harvest gains at 0% | Bunch charitable giving | QCD instead of cash | Take the rest of the RMD | Q4 estimate or December withholding | Largest cut |
|---|---|---|---|---|---|---|---|
| H01 | −$598 | +$2,659 | $0 | −$1,076 | −$76 | −$43 | QCD instead of cash |
| H02 | −$516 | +$1,557 | −$1,794 | −$2,210 | +$5,040 | $0 | QCD instead of cash |
| H03 | −$2,880 | $0 | – | – | – | −$276 | Harvest losses |
| H04 | −$1,939 | +$23,290 | $0 | – | – | – | Harvest losses |
| H05 | – | – | −$3,312 | −$2,246 | −$1,512 | −$26 | Bunch charitable giving |
| H06 | −$370 | +$2,699 | $0 | – | – | $0 | Harvest losses |
| H07 | −$7,929 | – | −$8,003 | – | – | $0 | Bunch charitable giving |
| H08 | – | $0 | $0 | −$2,185 | – | −$321 | QCD instead of cash |
| H09 | −$450 | – | $0 | −$1,250 | – | −$298 | QCD instead of cash |
| H10 | −$1,222 | +$10,676 | – | – | – | – | Harvest losses |
| H11 | −$607 | +$1,820 | $0 | – | – | −$52 | Harvest losses |
| H12 | – | – | −$115 | −$975 | +$6,501 | $0 | QCD instead of cash |
"Take the rest of the RMD" cuts the total by removing the 25% excise on an RMD not taken by December 31; "Q4 estimate or December withholding" by lowering the underpayment penalty. Bunching moves 2027–2028 deductions into 2026: the 2026 cut shown here comes back as higher tax in those years (the tool shows the three-year comparison).
The Roth conversion: a 2026 cost and a marginal rate
| Household | Federal | State | IRMAA (2028) | ACA credit lost | 2026 cost | Rate on the $25,000 | Next $1,000 costs |
|---|---|---|---|---|---|---|---|
| H01 | +$5,299 | +$682 | $0 | $0 | $5,980 | 23.9% | 25.9% |
| H02 | +$3,000 | +$998 | $0 | $0 | $3,998 | 16.0% | 16.0% |
| H03 | +$5,900 | $0 | $0 | $0 | $5,900 | 23.6% | 25.4% |
| H04 | +$2,760 | +$701 | $0 | +$21,204 | $24,665 | 98.7% | 14.8% |
| H05 | +$4,392 | $0 | $0 | $0 | $4,392 | 17.6% | 12.5% |
| H06 | +$4,868 | $0 | $0 | $0 | $4,868 | 19.5% | 12.0% |
| H07 | +$7,897 | +$1,479 | +$2,297 | $0 | $11,673 | 46.7% | 38.0% |
| H08 | +$5,830 | +$250 | +$1,148 | $0 | $7,228 | 28.9% | 28.1% |
| H09 | +$3,913 | +$647 | $0 | $0 | $4,560 | 18.2% | 27.4% |
| H10 | +$4,503 | +$1,100 | $0 | +$10,212 | $15,815 | 63.3% | 26.4% |
| H11 | +$4,021 | $0 | $0 | $0 | $4,021 | 16.1% | 13.2% |
| H12 | +$5,223 | +$1,250 | $0 | $0 | $6,473 | 25.9% | 17.5% |
The highest rates are the ACA households: their conversion pushed household income over 400% of the poverty line, where the whole 2026 premium credit ends. A conversion is tax paid in 2026 at a known rate in place of tax on the same dollars later. Whether it pays depends on the rate those dollars would face later, which this study does not assume.
Moves together are not the sum of moves alone
| Household | Moves | Together | Sum of singles | Interaction | Thresholds crossed |
|---|---|---|---|---|---|
| H01 | M6 M5 M2 M7 | −$1,458 | −$1,792 | +$334 | – |
| H02 | M5 M4 M2 | −$3,511 | −$4,520 | +$1,009 | – |
| H03 | M2 M7 | −$3,105 | −$3,156 | +$51 | irmaa |
| H05 | M6 M5 M4 M7 | −$7,206 | −$7,096 | −$111 | – |
| H07 | M4 M2 | −$13,381 | −$15,932 | +$2,552 | – |
| H08 | M5 M7 | −$2,433 | −$2,506 | +$73 | – |
| H09 | M5 M2 M7 | −$1,943 | −$1,998 | +$55 | – |
| H11 | M2 M7 | −$607 | −$660 | +$52 | – |
| H12 | M5 M4 | −$1,175 | −$1,090 | −$85 | – |
Where the two differ, a threshold sits between them: a bracket edge, the Social Security taxation formula, the senior-deduction phase-out, a state exclusion limit, or a cliff (an IRMAA tier, 400% of the poverty line).
Every move at three sizes
| Household | Small | Medium | Large |
|---|---|---|---|
| H01 | −$538 (realise $2,667 of losses) | −$598 (realise $5,333 of losses) | −$598 (realise $8,000 of losses) |
| H02 | −$237 (realise $1,667 of losses) | −$516 (realise $3,333 of losses) | −$516 (realise $5,000 of losses) |
| H03 | −$2,308 (realise $6,667 of losses) | −$2,880 (realise $13,333 of losses) | −$2,880 (realise $20,000 of losses) |
| H04 | −$933 (realise $5,000 of losses) | −$1,939 (realise $10,000 of losses) | −$2,822 (realise $15,000 of losses) |
| H06 | −$370 (realise $2,000 of losses) | −$370 (realise $4,000 of losses) | −$370 (realise $6,000 of losses) |
| H07 | −$3,964 (realise $16,666 of losses) | −$7,929 (realise $33,334 of losses) | −$11,060 (realise $50,000 of losses) |
| H09 | −$450 (realise $4,000 of losses) | −$450 (realise $8,000 of losses) | −$450 (realise $12,000 of losses) |
| H10 | −$521 (realise $3,333 of losses) | −$1,222 (realise $6,667 of losses) | −$1,222 (realise $10,000 of losses) |
| H11 | −$362 (realise $1,667 of losses) | −$607 (realise $3,333 of losses) | −$607 (realise $5,000 of losses) |
| Household | Small | Medium | Large |
|---|---|---|---|
| H01 | +$1,439 (realise $11,200 of gains) | +$2,659 (realise $22,400 of gains) | +$4,173 (realise $30,000 of gains) |
| H02 | +$728 (realise $18,250 of gains) | +$1,557 (realise $36,500 of gains) | +$12,522 (realise $80,000 of gains) |
| H03 | $0 (realise $0 of gains) | $0 (realise $0 of gains) | +$11,528 (realise $60,000 of gains) |
| H04 | +$22,154 (realise $34,050 of gains) | +$23,290 (realise $68,100 of gains) | +$32,503 (realise $120,000 of gains) |
| H06 | +$1,759 (realise $12,500 of gains) | +$2,699 (realise $25,000 of gains) | +$2,699 (realise $25,000 of gains) |
| H08 | $0 (realise $0 of gains) | $0 (realise $0 of gains) | +$8,674 (realise $40,000 of gains) |
| H10 | +$757 (realise $5,275 of gains) | +$10,676 (realise $10,550 of gains) | +$18,330 (realise $50,000 of gains) |
| H11 | +$1,325 (realise $10,000 of gains) | +$1,820 (realise $20,000 of gains) | +$1,820 (realise $20,000 of gains) |
| Household | Small | Medium | Large |
|---|---|---|---|
| H01 | $0 (2027 giving now) | $0 (2027 and 2028 giving now) | $0 (2027 and 2028 giving now, as stock) |
| H02 | −$354 (2027 giving now) | −$1,794 (2027 and 2028 giving now) | −$1,794 (2027 and 2028 giving now, as stock) |
| H04 | $0 (2027 giving now) | $0 (2027 and 2028 giving now) | $0 (2027 and 2028 giving now, as stock) |
| H05 | −$2,552 (2027 giving now) | −$3,312 (2027 and 2028 giving now) | −$2,300 (2027 and 2028 giving now, as stock) |
| H06 | $0 (2027 giving now) | $0 (2027 and 2028 giving now) | $0 (2027 and 2028 giving now, as stock) |
| H07 | −$4,050 (2027 giving now) | −$8,003 (2027 and 2028 giving now) | −$8,003 (2027 and 2028 giving now, as stock) |
| H08 | $0 (2027 giving now) | $0 (2027 and 2028 giving now) | $0 (2027 and 2028 giving now, as stock) |
| H09 | $0 (2027 giving now) | $0 (2027 and 2028 giving now) | $0 (2027 and 2028 giving now, as stock) |
| H11 | $0 (2027 giving now) | $0 (2027 and 2028 giving now) | $0 (2027 and 2028 giving now, as stock) |
| H12 | $0 (2027 giving now) | −$115 (2027 and 2028 giving now) | −$115 (2027 and 2028 giving now, as stock) |
| Household | Small | Medium | Large |
|---|---|---|---|
| H01 | −$970 (QCD $5,000) | −$1,076 (QCD $10,000) | −$1,076 (QCD $20,000) |
| H02 | −$1,000 (QCD $5,000) | −$2,210 (QCD $10,000) | −$1,970 (QCD $20,000) |
| H05 | −$1,136 (QCD $5,000) | −$2,246 (QCD $10,000) | −$4,157 (QCD $20,000) |
| H08 | −$971 (QCD $5,000) | −$2,185 (QCD $10,000) | −$3,930 (QCD $20,000) |
| H09 | −$750 (QCD $5,000) | −$1,250 (QCD $10,000) | −$3,062 (QCD $20,000) |
| H12 | −$1,175 (QCD $5,000) | −$975 (QCD $10,000) | −$975 (QCD $20,000) |
| Household | Small | Medium | Large |
|---|---|---|---|
| H01 | −$382 (leave it) | −$76 (take it) | −$382 (take it as a QCD) |
| H02 | $0 (leave it) | +$5,040 (take it) | $0 (take it as a QCD) |
| H05 | −$5,549 (leave it) | −$1,512 (take it) | −$5,549 (take it as a QCD) |
| H12 | $0 (leave it) | +$6,501 (take it) | $0 (take it as a QCD) |
| Household | Small | Medium | Large |
|---|---|---|---|
| H01 | −$16 (January estimate $930) | −$43 (December withholding $930) | −$43 (December withholding $465) |
| H02 | $0 (January estimate $0) | $0 (December withholding $0) | $0 (December withholding $0) |
| H03 | −$104 (January estimate $6,000) | −$276 (December withholding $6,000) | −$138 (December withholding $3,000) |
| H05 | −$10 (January estimate $563) | −$26 (December withholding $563) | −$26 (December withholding $282) |
| H06 | $0 (January estimate $777) | $0 (December withholding $777) | $0 (December withholding $389) |
| H07 | $0 (January estimate $0) | $0 (December withholding $0) | $0 (December withholding $0) |
| H08 | −$121 (January estimate $7,000) | −$321 (December withholding $7,000) | −$161 (December withholding $3,500) |
| H09 | −$112 (January estimate $6,499) | −$298 (December withholding $6,499) | −$149 (December withholding $3,250) |
| H11 | −$20 (January estimate $1,142) | −$52 (December withholding $1,142) | −$52 (December withholding $571) |
| H12 | $0 (January estimate $0) | $0 (December withholding $0) | $0 (December withholding $0) |
Run the seven moves on your own 2026 numbers
Free: your 2026 baseline, one move in full and your combined total.
Download the data
Every row on this page, with its inputs: year-end-tax-moves-2026.csv · year-end-tax-moves-2026.json (CC0).
Questions
- What year-end moves can a retiree still make for 2026?
- A Roth conversion, harvesting losses, harvesting long-term gains in the 0% band, bunching charitable giving, a QCD instead of a cash gift, taking the rest of the RMD, and a fourth-quarter estimated payment or extra December withholding. All but the estimate (due January 15, 2027) must be done by December 31, 2026; a first RMD can wait until April 1, 2027.
- Does a Roth conversion lower 2026 tax?
- No. It adds ordinary income to 2026, so 2026 tax rises; it can also raise 2028 Medicare premiums and lower an ACA credit. It is a choice to pay at today's rate instead of a later one.
- Why is doing several moves not the sum of doing each one?
- Because the moves share the same thresholds. A QCD that lowers income can open room for a conversion; a loss harvest changes how much gain fits in the 0% band; together two moves can cross a bracket edge or an IRMAA tier that neither crosses alone.
- Are these my numbers?
- No. They are twelve illustrative households. The Year-end Tax Moves tool runs the same calculation on your own numbers.
Sources and method
Every figure was computed by the module behind Year-end Tax Moves, which runs QCD, conversion, harvest, giving, RMD and payment changes through QuantCalc's open tax engines — federal 2.0.0 and state 1.12.0, the same engines behind the federal and state tax APIs — for the 2026 tax year. Each move's effect is the engine's result with the move minus the engine's result without it. The engine documentation lists the official source of every rule; the rules this study leans on:
- Rev. Proc. 2025-32: 2026 inflation-adjusted items (incl. One Big Beautiful Bill amendments)
- 26 U.S. Code § 86
- 26 U.S. Code § 151
- Rev. Proc. 2025-32: 2026 inflation-adjusted items (incl. One Big Beautiful Bill amendments)
- Schedule D (Form 1040) 2025
- 26 U.S. Code § 170
- 26 U.S. Code § 170
- Notice 2025-67: 2026 retirement plan limitations
- 26 U.S. Code § 4974
- 26 U.S. Code § 6654
- 2026 Medicare Parts A & B Premiums and Deductibles
- Rev. Proc. 2025-25: 2026 applicable percentage table (§36B) and required contribution percentage
Labelled assumptions. Federal tax is the return as filed (the Tax Table step below $100,000 of taxable income). IRMAA: 2026 MAGI sets 2028 Medicare premiums; the 2028 tiers are not published, so the 2026 tiers and premiums are used (indexed tiers will be higher, so this can overstate an IRMAA effect but not miss one). ACA: the 2025 poverty line governs 2026 coverage; the benchmark premium is the engine's state-level approximation for the household's ages, all 12 months. State: a joint return is figured on the household's combined income at the filer's age (the state API's household path). Households are illustrations, not averages; the full inputs and the exact API request of every row are in the data files.
An estimate, not tax advice. These are computed examples for the 2026 tax year from QuantCalc's open tax engine (federal 2.0.0, state 1.12.0). Your own return depends on facts these examples do not include. Check any move with your tax preparer or account custodian before acting.
Cite this research study
QuantCalc Research (2026). Which Year-End Move Cuts a Retiree's 2026 Tax Most? 7 Moves x 12 Households, Federal + State + IRMAA + ACA. https://quantcalc.app/research/year-end-tax-moves-2026/ (accessed <date>).
BibTeX
@misc{quantcalc2026whichyearendmovecutsaretirees2026taxmost,
title = {Which Year-End Move Cuts a Retiree's 2026 Tax Most? 7 Moves x 12 Households, Federal + State + IRMAA + ACA},
author = {{QuantCalc Research}},
year = {2026},
url = {https://quantcalc.app/research/year-end-tax-moves-2026/},
note = {Accessed <date>}
}
Machine-readable citation metadata (schema.org identifier and citation fields) is embedded in this page's JSON-LD, at the stable identifier https://quantcalc.app/research/year-end-tax-moves-2026/.