Four advisor-tier pieces landed in two weeks. Two of them argued the vehicle is worse than it looked in June. What the coordination-layer math and the kiddie-tax comparison actually show for the high-income parent — and how the seven non-conforming states change the reading of both critiques.
The advisor-tier discourse just sharpened
Across a two-week window from early July 2026 through the July 18–19 Kitces Weekend Reading digest, four separate advisor-tier publications addressed Trump Accounts, and an earlier June 22 Wall Street Journal opinion was actively re-surfaced within the same discourse. Michael Kitces's Weekend Reading for Financial Planners on July 4–5 led with the new Rev. Proc. 2026-25 gift-tax safe harbor (announced via IR-2026-80 on June 29); Hall / TRECPA followed in early July with a walkthrough of the gift-splitting trap the Rev. Proc. left in place; Morningstar published a wealth-trajectory piece framing outcomes as "$5.5 million or just $39,000"; the Wall Street Journal ran an opinion by Adam Michel of the Cato Institute calling the vehicle a double-taxation trap; and Ben Henry-Moreland at Kitces's Nerd's Eye View resurfaced (via the July 18–19 Weekend Reading digest, originally published December 24, 2025) a piece arguing that taxable custodial accounts beat Trump Accounts under the kiddie-tax mechanic.
For the high-income parent who read the four-vehicle comparison in June and concluded Trump Accounts were at least on the table, the July discourse is not asking whether the vehicle is any good. It is asking two follow-on questions: how does the funding decision coordinate against the annual gift-tax exclusion, the 529, and everything else already in the household's plan; and how should the operator-tier reader receive two credible advisor-tier arguments that the vehicle is suboptimal.
The coordination layer: what stacks against the annual exclusion, the 529, and everything else
The $19,000 per-donor per-donee annual gift-tax exclusion (IRC §2503(b); Rev. Proc. 2025-32 §4.42(1)) is the frame reader households already operate within. Rev. Proc. 2026-25's safe harbor, announced via IR-2026-80 on June 29, 2026, was designed to keep a $5,000 Trump Account contribution from independently triggering a Form 709 gift-tax return — but the safe harbor is all-or-nothing, and collapses if any other reportable gift to the same donee in the same year would independently require Form 709.
For a household making annual 529 contributions, using exclusion room for direct gifts, and now considering a $5,000 Trump Account contribution, the stack needs walking with specificity — and for an S-corp owner whose funding dollars come from a mix of W-2, K-1 pass-through, and quarterly distributions, the timing of when funding dollars are available layers on top.
| Contribution stream | 2026 annual cap / exclusion | Interaction with Rev. Proc. 2026-25 safe harbor | Interaction with existing 529 / direct gifts |
|---|---|---|---|
| Trump Account contribution | $5,000 per child aggregate across contributors (IRC §530A(c)(2)(A); COLA indexing at §530A(c)(2)(C) begins for taxable years after 2027) | Safe harbor holds if no other reportable gifts to the same donee that year, per Rev. Proc. 2026-25 §4.02 | Independent of the 529 contribution mechanic; both may be made in the same year without doubling the exclusion |
| 529 contribution — annual | $19,000 per donor per donee (§2503(b)); or 5-year forward-election up to $95,000 per donor under IRC §529(c)(2)(B) | A 529 contribution is a completed gift for §2503(b) purposes; if the donor exceeds the $19,000 annual exclusion via 529, the "no other reportable gifts" condition in Rev. Proc. 2026-25 §4.02 fails and the safe harbor collapses for that year | The load-bearing interaction. A 529 contribution above the annual exclusion — or a 5-year forward-election — is the specific move that costs the household its Trump Account safe harbor |
| Direct gift (cash, appreciated stock) | $19,000 per donor per donee (§2503(b)) — shared with 529 above | Any direct gift plus 529 contribution stacking above $19,000 per donor collapses the safe harbor | If direct gifts + 529 remain at or below $19,000 per donor, the Trump Account contribution rides the safe harbor cleanly |
| Gift-splitting between spouses | Doubles all above ceilings when properly elected on Form 709 (IRC §2513) | Election of gift-splitting itself requires filing a Form 709 for the year — which by definition defeats the safe harbor whose purpose is to avoid the filing | The married-couple coordination trap Hall named at TRECPA. Using §2513 doubles capacity but costs the paperwork-avoidance benefit; the trade is deliberate, not accidental |
The multi-child stack compounds non-trivially. Each additional child has an independent $19,000 exclusion per parent and its own Trump Account safe harbor. At two-to-four kids, the $5,000 × N + $19,000 × N + $19,000 × N funding-side arithmetic starts to interact with actual household cash flow. For the S-corp owner, quarterly distribution timing becomes the load-bearing constraint: funding capacity exists only when distributions land, and the Rev. Proc. 2026-25 conditions apply annually.
The Rev. Proc. 2026-25 safe harbor: five conditions and one cliff
The safe harbor lives in Rev. Proc. 2026-25 §4.02 and applies only when all five conditions are met for the calendar year. Paraphrased from the operative language: (1) the taxpayer is an individual; (2) the only taxable gifts made by the taxpayer during the calendar year are cash contributions — cash, check, money order, or electronic funds transfer — to one or more Trump Accounts; (3) the taxpayer's total gifts during the year to each individual who is an account beneficiary do not exceed the §2503(b) annual exclusion amount ($19,000 for 2026); (4) contributions to Trump Accounts made during the calendar year do not independently generate a gift-tax or GST-tax liability for that year; and (5) no gift-tax return is required to be filed for that year by or on behalf of the taxpayer.
When the five conditions hold, qualifying Trump Account contributions are treated as completed gifts of present interests — eligible for the §2503(b) annual exclusion — and no Form 709 filing is required. Without the safe harbor, contributions during the account's growth period (before the beneficiary turns 18) default to future-interest characterization, which is not annual-exclusion-eligible and requires Form 709 filing regardless of amount. The Rev. Proc. eliminates the filing burden for the ordinary case.
The cliff is what makes it operator-relevant. If total gifts to a single beneficiary exceed $19,000 in a calendar year by any amount, the safe harbor is lost entirely for that year and for that beneficiary. Form 709 filing is required; the Trump Account contributions themselves get reclassified as future-interest gifts (not annual-exclusion-eligible) as the penalty. The interaction with 529 forward-elections is the sharpest example: a household that elects the 5-year forward-election under §529(c)(2)(B) has front-loaded $95,000 to a single beneficiary, which trips the cliff and knocks the Trump Account contribution out of safe-harbor treatment. That is not an argument against forward-elections — it is an argument for coordinating them.
The double-taxation critique: what Michel's argument actually says
Adam Michel's Wall Street Journal opinion on June 22, 2026 (republished at Cato) argued that Trump Accounts subject parental savings to double taxation by combining (i) after-tax contributions with (ii) ordinary-income tax on withdrawal under IRC §530A(a)'s pass-through to IRC §408(a). Michel's framing: "Tax the money going in or tax it coming out, but not both." Traditional IRAs tax coming out; Roth vehicles and 529s tax going in for qualified use; taxable brokerage taxes gains at capital-gains rates. Trump Accounts, in Michel's argument, do both.
The framing is technically accurate on the §530A mechanic — contributions are non-deductible; growth and the Treasury seed distribute as ordinary income at the recipient's marginal rate. Where the argument narrows is the reference class. "Double taxation" implicitly compares Trump Accounts to a Roth-style vehicle where after-tax contributions produce tax-free growth and tax-free qualified distributions. Nothing in the U.S. tax code does that for the class of high-income parents who cannot fund a custodial Roth for want of a §219(f)(1) earned-income pathway. The correct reference class is (i) a non-deductible traditional IRA at high income — same after-tax in, same ordinary-income out; (ii) a UTMA — after-tax in, kiddie-tax attribution at $400K+ MFJ; and (iii) the $1,000 Treasury seed under IRC §6434, net-new capital every eligible household receives and no comparison captures.
| Critique element | Technically accurate? | Reference class the framing implies | Where the argument holds and where it narrows |
|---|---|---|---|
| Contributions non-deductible federally | Yes — §530A provides no federal deduction | Implies Roth or traditional-IRA-with-deduction comparison | Holds against Roth and deductible traditional IRA. Narrows against non-deductible traditional IRA at high income, which is mechanically identical |
| Growth + Treasury seed taxed as ordinary income at distribution | Yes — §530A(a) pass-through to §408(a) | Implies Roth or tax-free-qualified 529 comparison | Holds against Roth and qualified 529. Narrows against UTMA (kiddie-tax ordinary at parent-attribution) and taxable brokerage held to majority |
| Recipient's marginal rate at distribution | Uncertain; depends on the recipient's earned income in the distribution year | Framing presumes a high recipient marginal | The F5 comparison modeled 12% and 22% recipient marginals as the realistic range. Below 22%, the ordinary-rate hit is more modest than the "double tax" label suggests |
| Treasury $1,000 seed as net-new capital | Yes — §6434 pilot-program contribution for births 2025–2028 | Not captured by the Michel framing at all | Every eligible household's $1,000 seed, net of the expected ordinary-income tax on that layer at distribution, is present-value positive regardless of household AGI |
Michel's "pick a lane" reform — either let contributions deduct or exempt gains at withdrawal — is a serious policy argument, not an actionable planning move. For the reader deciding what to do with the vehicle as it currently exists, the right question is not "does the Trump Account beat a hypothetical Roth" — nothing does at that comparison — but "does the Trump Account plus the Treasury seed plus the safe-harbor coordination beat the next-best available vehicle in the specific slot." That requires the coordination-layer analysis above, not the double-taxation label.
The custodial-taxable argument: Henry-Moreland at Kitces, and where the ranking flips
The more technically consequential critique is Ben Henry-Moreland's piece on Nerd's Eye View / Kitces, item 6 in the July 18–19 Weekend Reading digest. As Henry-Moreland frames the argument, taxable custodial accounts (UTMA structure) can produce better after-tax outcomes for kids' savings than Trump Accounts because of the kiddie-tax mechanic. In 2026, a dependent child with no earned income can realize up to $2,700 of qualified dividends and long-term capital gains per year at an effective federal rate of zero — the first $1,350 offset by the dependent's standard deduction, the next $1,350 falling in the child's own 0% LTCG bracket (IRC §1(g); §63(c)(5)(A); Rev. Proc. 2025-32). That annual harvest compounds; whatever gains remain distribute at long-term capital-gains rates, not at the ordinary rates Trump Accounts convert into under §530A(a).
Henry-Moreland's implicit ranking is that the custodial-taxable route wins because the LTCG-versus-ordinary-rate spread on eventual gains is a larger effect than the tax-deferred-growth wrapper the Trump Account provides. He also flags a Roth-conversion friction: converting a Trump Account to Roth at age 18-plus requires external funds to pay the conversion tax; funding the tax from the account itself triggers ordinary-income tax plus, if under 59½, the IRC §72(t) 10% additional tax on the growth-and-seed layers (subject to §72(t)(2) statutory exceptions). Henry-Moreland also flags a third friction: if the child is still a dependent in the conversion year, kiddie-tax reapplies to conversion income above $2,700 — the excess is taxed at the parent's marginal rate, not the child's — which for a $400K+ MFJ household closes the "just convert at low child bracket" escape hatch.
The argument deserves an operator-tier engagement, not a dismissal.
| Line | Custodial-Taxable (UTMA) | Trump Account |
|---|---|---|
| Annual contribution | $5,000 (after-tax at parent) | $5,000 (after-tax at parent) + $1,000 Treasury seed at year 0 |
| Growth mechanic | Taxable annually — dividends, realized gains, interest | Tax-deferred inside the account (§530A(a) pass-through to §408(a)) |
| Kiddie-tax mechanic (§1(g)) | First $1,350 untaxed; next $1,350 at child's rate (0% LTCG for the qualified-dividend / LTCG layer); above $2,700 at parent's marginal rate | Does not apply during growth phase — the tax-deferred wrapper suspends kiddie-tax exposure |
| Parent-attribution NIIT layer | $400K+ MFJ MAGI exceeds the $250K NIIT threshold (IRC §1411(b)) — where the parent elects Form 8814 inclusion, child unearned income above $2,700 picks up the additional 3.8% NIIT at the parent's return; under the Form 8615 default, the layer stays on the child's return and generally does not attach. Practical effect at this tier is captured in the cumulative drag figure below | Not applicable during growth |
| Cumulative annual tax drag over 15 years | ~$5,300 (F5 modeled at ~1.2% of average balance; threshold-crossing year ~8) | Zero during growth phase |
| Gross accumulation at year 15 | ~$116,400 (before annual drag) | ~$118,800 (including $1,000 seed) |
| Distribution treatment | Balance already net of hold-period drag; gains remaining at distribution taxed at LTCG rates | Parent-contribution basis returns tax-free under §72 basis-recovery; growth + seed (~$43,800) taxed as ordinary income to the recipient at the recipient's marginal rate |
| Net at year 15, recipient 12% marginal | ~$111,100 | ~$113,500 |
| Net at year 15, recipient 22% marginal | ~$111,100 | ~$109,200 |
| Operator-tier reading | At recipient-12%-marginal, the Trump Account edges the UTMA by ~$2,400. At recipient-22%-marginal, the UTMA edges the Trump Account by ~$1,900. Henry-Moreland's ranking holds at 22%; it flips at 12%. The tiebreaker is the recipient's projected marginal at distribution — a household choice, not a universal answer | |
Where Kitces's conclusion holds and where it doesn't
Henry-Moreland's conclusion holds when three assumptions align: the recipient's marginal at distribution lands at 22% or higher; the household is comfortable with the unrestricted control transfer at the state's UTMA age of majority (18–21 depending on state statute) rather than the §408(a) traditional-IRA framework the Trump Account transitions into at age 18; and the annual kiddie-tax-plus-NIIT drag — $5,300 cumulative over 15 years, not zero — is viewed as an acceptable cost of the LTCG-rate treatment on eventual gains.
The conclusion does not hold when any of those assumptions moves. If the recipient's marginal is modeled at 12% — a lower-earning young adult, a graduate-school year, a gap year — the Trump Account moves ahead. If the household prefers the Trump Account's post-18 traditional-IRA framework — which builds in a §72(t) behavioral gate against early withdrawal — over immediate unrestricted access, the ranking is a preference. And the $1,000 Treasury seed adds present value the UTMA cannot reproduce. The choice is genuinely close at the $400K+ MFJ tier; the tiebreaker turns on assumptions the household gets to make. Trump Accounts are neither dominated nor dominant.
The Morningstar wealth-trajectory range — briefly, for context
Spencer Look's Morningstar piece — item 4 in the same Kitces Weekend Reading digest — projected outcomes ranging from "$5.5 million" at 90th-percentile 10.29% annualized returns over 55 years to "just $39,000" for a seed-only account at 10th-percentile 2.17% annualized. Morningstar's forward-looking equity-return assumption sits at roughly 6.3%, materially below the ~10% historical U.S. equity return. The $5.5 million requires two behaviors: annual contributions near the $5,000 maximum for 17 years, and sustained equity exposure across a 55-year horizon.
The range is intellectually honest; it discloses the compounding sensitivity single-projection pieces smooth over. It is also generic-audience framing. The operator-tier decision-tree — coordination-layer stack, critique-response mechanics, household-specific tiebreaker on recipient-marginal projection — is where the reader's actual decision lives.
State conformity: the seven non-conforming states sharpen both critiques
Seven states currently do not conform to the federal Trump Account tax treatment (as reported by Washington Post and Daily Signal in early 2026): California, Hawaii, Kentucky, Massachusetts, Pennsylvania, South Carolina, and Wisconsin. In these states, annual investment gains inside the Trump Account are subject to state income tax each year during the growth period, even though federally deferred under §530A. (Employer §128 contributions — the employer-side layer, not the household contribution stream this piece addresses — may also be state-taxable in states with fixed-date IRC conformity dated before July 4, 2025.) California's AB/SB 180 has been introduced to conform but has not been enacted as of this piece.
State non-conformity sharpens rather than rebuts Henry-Moreland's argument — but the mechanic is loss-of-federal-deferral, not a state LTCG rate advantage. In a non-conforming state, the Trump Account loses its federal §530A deferral at the state level: growth is state-taxable each year during accumulation, then federally taxable as ordinary income at distribution. The custodial-taxable account has always been state-taxable during accumulation, so state non-conformity adds no new drag to the UTMA. State-rate treatment of capital gains varies across the seven (PA's flat 3.07% does not distinguish ordinary from LTCG; CA and MA apply ordinary-rate state to LTCG; WI and SC offer partial state LTCG exclusions; KY and HI vary by category). Net effect: state non-conformity blunts the Trump Account's tax-deferred-growth wrapper without adding equivalent drag to the UTMA, moving the operator-tier tiebreaker modestly toward the UTMA in these seven states. For readers in the conforming majority, the federal analysis above stands as written.
The decision the reader executes
The piece walks the sequence; the reader fills in the numbers.
- Confirm the vehicle is on the table. The June four-vehicle comparison is the entry point. Recipient-marginal projection at distribution, control-transfer preference, $1,000 seed eligibility for a child in the 2025–2028 birth window, and state of residence (conforming or non-conforming) are the four inputs that decide whether the vehicle belongs in the stack at all.
- Run the coordination-layer stack. Annual gift-tax exclusion capacity ($19,000 per donor per donee); planned 529 contribution level and whether a 5-year forward-election is in play; direct gifts the household expects to make in the year; the Trump Account $5,000 contribution; the gift-splitting decision under §2513 (understanding it triggers Form 709 filing and defeats the Rev. Proc. 2026-25 safe harbor); and the S-corp distribution timing if funding dollars are K-1-sourced.
- Read the critiques on their merits. Michel's double-taxation framing narrows against the reference-class question at high income. Henry-Moreland's kiddie-tax argument is real math whose ranking flips between 12% and 22% recipient-marginal at the $400K+ MFJ tier.
- Decide. Which streams to fund, at what levels, in what order, with what safe-harbor posture, and — for households in the seven non-conforming states — how the state-tax wedge shifts the tiebreaker. The vehicle choice is contingent, not categorical. The operator-tier work is the choice, not the label.
Regulators are watching for a technical-corrections vehicle; the qualified-use categories at Prop. Reg. §§1.530A-2 through §1.530A-6 (March 9, 2026) remain reserved for future guidance. The framing above is stable to plan against as of this piece's publication date; a subsequent revenue procedure or final regulations may sharpen or shift specific mechanics.