The 2025 return was filed in mid-April. The Q1 2026 estimated payment went out the same week — or didn't, and that question lives in a companion piece. Late May 2026 is the moment most readers stop thinking about estimated tax until September. That is the wrong moment to stop. The Q2 deadline is Monday, June 15, 2026, and the choice the reader has to make before that date is not "how much do I send" but "which safe harbor am I aiming at." The §6654 statute offers two paths to penalty-free status, and the standard financial-press framing flattens the choice into a single number that misleads in both directions.
The safe-harbor selection is the strategic move; the quarterly check is the operational consequence. A hybrid earner whose side business is climbing aggressively versus a hybrid earner whose side business is contracting are not running the same calculation, even though they are reading the same IRS instructions. The mid-year reforecast — pulling year-to-date numbers in late May, projecting the full year, and comparing the two §6654 paths against the projection — is the work that decides which number gets remitted on June 15 and on each quarter after.
This piece is the operator's view of that decision. The road map: the two-path structure under §6654(d)(1)(B) and why the two paths produce different answers; the year-over-year income trajectory as the variable that picks between them; a four-step mid-year reforecast workflow; and three concrete scenarios — income climbing, income dropping, income roughly flat — with the path selection worked through in each.
The two-path structure most coverage flattens
The §6654 safe harbor is satisfied by meeting either of two thresholds across the four required installments. The taxpayer does not have to meet both, and the statute does not specify which one to aim at — the choice is the taxpayer's.
Path A — Prior-year safe harbor. Pay, across the four required installments, at least 100% of the prior year's total tax liability. The 100% figure becomes 110% when the prior year's adjusted gross income exceeded $150,000 (the threshold is statutory at §6654(d)(1)(C) and is not inflation-indexed). For most readers of this publication, the operative number is the 110% one — household income over $150,000 is the threshold for being inside the target audience, not an aspirational mark.
Path B — Current-year safe harbor. Pay at least 90% of the current year's total tax liability. The figure here is unconditional — there is no upward adjustment for higher-income taxpayers, no inflation-indexed step. 90% of the current year is 90% of the current year.
The statute lets the taxpayer meet whichever is lower. Form 2210's penalty calculation at line-by-line reads against both paths and applies the lower one — but only if the four installments were sized correctly during the year. The trap is that the choice of path effectively has to be made at the time the quarterly payments are sized, not at filing time, because the per-quarter-stands-alone rule under §6654 means a quarter under-funded against both paths accrues penalty interest regardless of which path the annual reconciliation ultimately uses. The mid-year reforecast is what aligns the quarterly payment with the path the taxpayer is actually aiming at.
| Variable | Path A — Prior year | Path B — Current year |
|---|---|---|
| Threshold (annual) | 110% of prior-year total tax (100% if prior-year AGI ≤ $150,000) | 90% of current-year total tax |
| Quarterly target | Annual threshold ÷ 4 — mechanical, fixed at the start of the year | Projected current-year liability × 90% ÷ 4 — depends on the projection's accuracy |
| Information needed | Prior-year Form 1040 total-tax line — already filed, known with certainty | Full-year income projection — uncertain, sensitive to side-business volatility |
| When it wins | Current-year income is climbing relative to prior year — paying 110% of last year undercuts what 90% of this year would require | Current-year income is dropping — paying 110% of last year overpays relative to the actual liability |
| Execution risk | Low — calculation is mechanical and the number doesn't move during the year | Higher — if the year-end income projection misses high, the 90% target was set too low and the safe harbor is missed |
The two paths produce dramatically different quarterly targets in years where income is moving. A reader whose 2025 total tax was $52,000 (Path A threshold at 110% = $57,200 annual / $14,300 per quarter) and whose 2026 projected total tax is $90,000 (Path B threshold at 90% = $81,000 annual / $20,250 per quarter) has a $5,950 per-quarter gap between the two answers. Path A is cheaper through the year; Path B is the path the taxpayer might think the statute requires if they only read the "90% of this year" sentence.
Income trajectory is the decision variable
The honest framing of the path choice: pick whichever produces the lower number, subject to the constraint that the lower number actually has to be paid quarterly to count. The variable that drives which path produces the lower number is the year-over-year income trajectory — specifically, the trajectory of total tax liability, not gross income, because progressive-rate effects and deduction-mix changes can make the two numbers move at different rates.
Three structural cases sit underneath the choice, and each maps to one of the scenarios developed later in this piece.
Income climbing fast. The hybrid earner whose side business grew meaningfully between 2025 and 2026 — a consulting LLC that scaled from $40K to $120K, an S-corp distribution that jumped from $50K to $150K, a short-term-rental property that added a second unit and doubled net cash flow. For this reader, 110% of last year is substantially less than 90% of this year. Path A produces the lower number. The honest counter-argument: aim at Path A and you'll owe a meaningful balance at filing time in April 2027, but no §6654 penalty attaches because Path A's threshold was met. The "balance due in April" is a cash-management question, not a §6654 question, and most readers prefer the cash-management trade-off to writing larger checks every quarter.
Income dropping. The hybrid earner whose 2025 was unusually strong — a one-time S-corp distribution that won't recur, a windfall consulting engagement that ended, a property sold that won't be sold again — and whose 2026 is reverting to a lower run rate. For this reader, 110% of last year is substantially more than 90% of this year. Path B produces the lower number, but only if the year-end projection is accurate enough to defend if challenged. The reader trades the certainty of Path A for the lower payment of Path B; if 2026 ends higher than projected, the safe harbor wasn't met against Path A (the quarterly payments were too small) or against Path B (the 90% threshold was higher than the payments).
Roughly flat. The reader whose 2025 and 2026 are within a normal-variance band of each other — household income drifting by 5–10% in either direction, no structural change to the side business. For this reader, the two paths produce close numbers and Path A wins on execution simplicity. The calculation is mechanical, the number is fixed at the start of the year, and the reader does not have to revisit the projection each quarter.
The skeptical aside the publication writes here. The standard advice — "just pay 110% of last year and you're fine" — is calibrated for a reader whose income is stable. It is correct as far as it goes, but it leaves money on the table for the income-dropping reader (overpaid every quarter, refund at filing) and leaves the income-climbing reader exposed to a larger-than-expected April balance that can wreck cash flow if it wasn't planned for. The two-path structure exists because Congress recognized that some taxpayers' incomes move; the path that's optimal depends on which direction.
The mid-year reforecast — a four-step workflow
The mid-year reforecast is the operational discipline that turns the path-selection question into a number on a check. It runs in late May, with the Q2 deadline three to four weeks out. Four steps, each operationally simple, but the discipline of running all four is what closes the loop.
Step 1 — Pull year-to-date actuals through April 30. The numbers that matter for the projection are the income items that will appear on the 2026 return. For the W-2 side, this is the most recent pay stub's year-to-date gross wages and year-to-date federal income tax withheld. For the side-business side, this is the income earned through April 30 by category — consulting fees billed, S-corp distributions taken, rental net income after expenses, 1099 board payments received. Pull these from the bookkeeping system, the bank-deposit log, or the household tracker the reader is already running. The point is to have an actual number, not a memory of one.
Step 2 — Project the remaining eight months. Annualize the year-to-date number where the income is ratable (a steady consulting client, a salaried W-2). Apply a forward-looking adjustment where it isn't — an STR property whose summer season is the main revenue (most of the year's income will land in Q3, so the April 30 number is not 1/3 of the annual); an S-corp distribution that the reader controls timing on; a board retainer that pays annually in October. The projection is a number, not a precise forecast — being within 10% of the actual year-end number is usually enough to make the path choice correctly.
Step 3 — Compute both safe-harbor thresholds. Path A is the 2025 Form 1040 total-tax line × 110% (if 2025 AGI exceeded $150,000) or × 100% (if not). Path B is the projected 2026 total tax liability × 90%. The projected 2026 total tax can be estimated by running the projected income through the 2026 tax brackets published by the IRS and adding the side-business self-employment tax layer (§1401: 12.4% Social Security on the first $184,500 of combined wages-and-SE earnings for 2026, plus 2.9% Medicare with no cap, plus the 0.9% additional Medicare surtax above $200K single / $250K joint). The S-corp side of the picture changes this calculation — S-corp distributions are not subject to SE tax, so the salary-versus-distribution split affects the projection. For readers thinking through the S-corp side specifically, the S-corp election math for W-2 earners works that calculation in detail.
Step 4 — Pick the lower number and size the remaining quarters. Whichever path produces the lower annual threshold is the path the taxpayer aims at. Divide that annual threshold by four. Subtract from the per-quarter figure (a) the W-2 federal withholding the taxpayer's paychecks are already producing and (b) whatever Q1 payment was already remitted. The result is the Q2 supplemental payment due June 15. The same per-quarter math then applies to Q3 (September 15) and Q4 (January 15, 2027), with one re-check of the projection in late August before Q3 to confirm the year-end picture hasn't shifted materially.
The point of running all four steps is that the path choice gets made deliberately rather than by inertia. A reader who hasn't reforecasted by late May is implicitly defaulting to whatever path the previous April's payment was sized against — usually Path A on the prior-year-known mechanic, which is fine for stable-income readers and expensive for everyone else.
Scenario A — Income climbing fast (the 110% trap)
The reader. A salaried professional with $220,000 of 2025 W-2 income and a side consulting LLC that distributed $60,000 in 2025. 2025 AGI was $280,000; 2025 total tax was $48,000. The consulting LLC has had a strong start to 2026 — three new retainer clients added in February and March, year-to-date through April 30 billing is $58,000, and the reader projects $160,000 of 2026 LLC net income against the $60,000 prior year. W-2 income is steady at $220,000. Projected 2026 AGI: $380,000. Projected 2026 total tax (running through 2026 brackets with the SE tax layer): roughly $92,000.
The path comparison. Path A: $48,000 × 110% = $52,800 annual safe-harbor target, or $13,200 per quarter. Path B: $92,000 × 90% = $82,800 annual safe-harbor target, or $20,700 per quarter. Path A is $30,000 cheaper across the year and $7,500 cheaper per quarter. Path A wins decisively.
The trap most coverage misses. A reader who reads "90% of this year" and aims at Path B is overpaying by $30,000 across the year — money that sits with the IRS as a refund in April 2027 (no interest paid to the taxpayer on the float, by the way; §6611 only requires the IRS to pay interest on refunds delayed past 45 days from the return-filing date, not on the overpayment float itself). The income-climbing reader who aims at Path A pays the lower quarterly amount through 2026, owes a balance at filing time in April 2027 of approximately $39,000 (the difference between $92,000 actual liability and $52,800 plus W-2 withholding contributions), and pays no §6654 penalty because Path A's threshold was met.
The cash-management corollary. The reader who plans to use Path A in an income-climbing year should set aside the balance-due amount in a high-yield savings account through the year rather than letting it accumulate accidentally. The IRS's current Q2 2026 underpayment rate is 6% (per Internal Revenue Bulletin 2026-8); a high-yield savings rate around 4% means the float is costing the reader roughly 2% on the gap — not zero, but materially less than the alternative of overpaying $30,000 across the year and earning zero on it.
Scenario B — Income dropping (the 90% relief)
The reader. A salaried professional with $250,000 of 2025 W-2 income and an S-corp that distributed $180,000 in 2025 — an unusually strong year driven by a one-time client engagement that ended in December 2025. 2025 AGI was $430,000; 2025 total tax was $108,000. The S-corp's 2026 has reverted to its normal run rate; year-to-date through April 30 distributions are $25,000, and the reader projects $80,000 of 2026 S-corp distributions against the $180,000 prior year. W-2 income is steady at $250,000. Projected 2026 AGI: $330,000. Projected 2026 total tax: roughly $72,000.
The path comparison. Path A: $108,000 × 110% = $118,800 annual safe-harbor target, or $29,700 per quarter. Path B: $72,000 × 90% = $64,800 annual safe-harbor target, or $16,200 per quarter. Path A is $54,000 more expensive across the year and $13,500 more expensive per quarter. Path B wins decisively.
The risk Path B introduces. The 90%-current-year path is correct only if the current-year projection is correct. If 2026 ends at $95,000 of total tax rather than $72,000 (because the S-corp's projection missed high — say a Q4 engagement comes in that wasn't in the May reforecast), the Path B threshold rises to $85,500 and the per-quarter requirement was actually $21,375, not $16,200. A reader who paid $16,200 per quarter through Q1 and Q2 against the Path B target now has under-paid both quarters against the corrected number. The §6654 penalty interest attaches to each shortfall from its due date forward, at the published quarterly underpayment rate (6% for Q2 2026; 7% for the Q1 2026 period it would back-trace into).
The operational discipline for Path B. The projection has to be re-checked at each subsequent quarter, not set once in May and forgotten. The pattern that fits income-dropping readers: run the May reforecast, pick Path B, size Q2 against the lower number, and then re-run the projection in late August before Q3 to confirm the year-end picture hasn't moved. If the August reforecast lifts the projection materially, the Q3 payment scales up to make up the implied Q1/Q2 underpayment relative to the new threshold. If the August reforecast holds, Q3 and Q4 continue at the original per-quarter figure.
Scenario C — Roughly flat (default to 110%)
The reader. A salaried professional with $250,000 of 2025 W-2 income and an S-corp that distributed $80,000 in 2025 — a steady-state side business in its third year. 2025 AGI was $330,000; 2025 total tax was $66,000. Year-to-date through April 30 S-corp distributions are $22,000; the reader projects $85,000 of 2026 S-corp distributions, a modest 6% lift. W-2 income is steady at $258,000 (a normal 3% merit raise). Projected 2026 AGI: $343,000. Projected 2026 total tax: roughly $70,000.
The path comparison. Path A: $66,000 × 110% = $72,600 annual safe-harbor target, or $18,150 per quarter. Path B: $70,000 × 90% = $63,000 annual safe-harbor target, or $15,750 per quarter. Path B is $9,600 cheaper across the year and $2,400 cheaper per quarter. Path B wins on the math — but by a smaller margin than the income-dropping scenario, and the projection risk is the same.
The operational answer for flat-income years. The case for Path A in flat-income years is execution simplicity, not pure cost minimization. Path A's number is known with certainty by April 16 — pulled directly off the prior-year return — and does not require a mid-year reforecast to defend. Path B saves roughly $9,600 across this year, but requires running the four-step reforecast workflow each quarter to guard against projection drift. For a reader whose income is genuinely flat, that overhead is high relative to the savings; for a reader whose income is moving, the overhead is the work that produces the right answer.
The pattern that fits flat-income years: aim at Path A, accept the modest overpayment as the cost of operational simplicity, and reserve the Path B mechanic for years when the trajectory is genuinely moving. The reforecast still runs — but it runs to confirm Path A is still the right path, not to size Path B's quarterly payment.
| Scenario | Path A — 110% prior | Path B — 90% current | Recommended path |
|---|---|---|---|
| A — Climbing fast ($48K → $92K total tax) |
$52,800 / yr $13,200 / qtr |
$82,800 / yr $20,700 / qtr |
Path A — $30,000 cheaper |
| B — Dropping ($108K → $72K total tax) |
$118,800 / yr $29,700 / qtr |
$64,800 / yr $16,200 / qtr |
Path B — $54,000 cheaper |
| C — Roughly flat ($66K → $70K total tax) |
$72,600 / yr $18,150 / qtr |
$63,000 / yr $15,750 / qtr |
Path A — simplicity over $9,600 |
What lands on the desk by June 15
The four-step reforecast produces a single number: the supplemental Q2 estimated payment due June 15. That number is remitted through IRS Direct Pay (linked to the taxpayer's bank account, no fee), EFTPS (which requires enrollment but produces a confirmation number for audit defense), or a mailed check accompanied by the Form 1040-ES Q2 voucher. Direct Pay is the right choice for most readers — fast, fee-free, and the confirmation lands by email within minutes of submission.
The number is also the input to the same calculation in September and January. A reader who has done the May reforecast and locked the path has already done the work for Q3 and Q4 — barring a material change in the year-end projection, the per-quarter number holds. The late-August re-check (the second reforecast, before Q3 on September 15) is the discipline that catches projection drift; it takes thirty minutes if the bookkeeping is clean, longer if it isn't.
What this article does not claim. None of this avoids the underlying tax. The 2026 federal liability on the W-2, the S-corp distribution, the consulting fees, the rental net income — owed either way. The path selection affects only the timing of when that liability is paid through the year and the §6654 penalty interest on the timing mismatches. Readers shopping for tax reduction need a different conversation — the QBI deduction at high income levels piece, the Solo 401(k) coordination with a W-2 plan piece, the S-corp election math. This piece is about choosing the safe-harbor path that minimizes the cash outflow and the penalty exposure through the year, not about reducing what's ultimately owed.
The reforecast pays the reader twice. Once on the safe-harbor question — the right path, sized correctly, gets the June 15 payment to a defended number. And once on the projection itself — a reader who knows where 2026 is heading by late May has nine months of runway to make year-end moves (retirement plan contributions, charitable bunching, S-corp salary structure decisions) against a real number rather than a guess. The reforecast is upstream of every other tax-planning conversation that lands between now and December 31.