The Election Nobody Read.

This note concerns the other kind. On July 10, 2026, the Internal Revenue Service released Private Letter Ruling 202628005, granting a grantor an extension of time under §2642(g) to allocate generation-skipping transfer (“GST”) tax exemption to gifts made to three irrevocable trusts across three separate years (I.R.S. Priv. Ltr. Rul. 202628005 (PLR-112074-25), issued Apr. 14, 2026, released July 10, 2026, Index No. 2642.07-00). There is no creditor in the story. There is no fraud, no concealment, no adverse party at all. There is only a family, three well-drafted trusts, two sets of professionals, and a small box on a gift tax return that was checked the wrong way — twice — and then not discovered for years.

The trusts in that ruling are, structurally, exactly what we would counsel a client to build. And they were nearly broken anyway. That is the lesson worth spending fifteen hundred words on: a structure can be seasoned, irrevocable, discretionary, and independently administered — and still be silently defective, because protection and taxation are two different disciplines and a plan has to survive both.

The facts, stripped of the redactions the Service applies to every letter ruling, are almost aggressively ordinary.

In “Year 1,” the grantor engaged an attorney and, on that attorney’s advice, created three irrevocable trusts for the benefit of the grantor’s children and further descendants. The dispositive terms of all three were substantially similar. While the grantor and spouse are living, the trustees hold sole discretion to distribute income or principal to the grantor’s descendants for maintenance, support, or education. On the death of the survivor of the grantor and spouse, each trust divides into a separate share for each surviving child, again administered on a discretionary standard, with each child holding a limited testamentary power of appointment in favor of that child’s own descendants; in default, the share passes per stirpes.

Read that paragraph again with a planner’s eye. Irrevocable. Discretionary. Multi-generational. Trustees — plural, and not identified as the grantor. A limited power of appointment that adds flexibility without pulling the property back into a beneficiary’s taxable estate. This is not an improvisation. This is the architecture.

Then came the paperwork.

The grantor and spouse relied on the drafting attorney to prepare their Year 1 Forms 709, and elected under §2513 to treat all gifts as made one-half by each of them — ordinary gift-splitting. But the attorney, in the ruling’s words, “did not advise Grantor or Spouse regarding the consequences of failing to allocate GST exemption to the Year 1 transfers,” and so no GST exemption was allocated to any of the three trusts.

It got worse before anyone noticed. In Year 2, the grantor transferred additional property to Trust 1. The same attorney prepared the returns and affirmatively elected out of automatic GST allocation under §2632(c)(5) — again without advising the clients of the consequences. In Year 3, a different firm — an accounting firm — prepared the returns and, reasonably enough, made them consistent with the prior year. It elected out again.

That is the mechanism worth naming, because it is the one that repeats in practice: an unexamined election becomes a precedent inside the client’s own file. The second professional did not make an independent error. It copied the first one forward, on the entirely sensible assumption that a deliberate-looking election on a signed return was deliberate.

Not until Year 4, when the family hired a second attorney to review their estate planning, did anyone discover that the Year 1 gifts carried no allocation and the Year 2 and Year 3 gifts had been affirmatively opted out of the automatic rules.

To a client, “we forgot to allocate an exemption” sounds like a filing nuisance. It is not. It goes to whether the trust does the one thing a multi-generational trust exists to do.

The GST tax is imposed by §2601 on every generation-skipping transfer — taxable distributions, taxable terminations, and direct skips (I.R.C. §§2601, 2611(a)). The tax is the taxable amount multiplied by the “applicable rate,” and the applicable rate turns on the trust’s inclusion ratio: one minus the applicable fraction, whose numerator is the GST exemption allocated to the trust and whose denominator is the value of the property transferred (I.R.C. §§2602, 2642(a)(1)). Allocate exemption equal to the value of the gift, and the fraction is one, the inclusion ratio is zero, and distributions to grandchildren and remoter descendants — decades from now — pass free of GST tax. Allocate nothing, and the inclusion ratio is one. Every future skip is fully exposed.

Congress anticipated that people would forget, which is why §2632(c) automatically allocates a transferor’s unused exemption to an “indirect skip” — a transfer to a “GST trust” — to the extent needed to drive the inclusion ratio to zero (I.R.C. §2632(c)(1), (c)(3)(A), (c)(3)(B)(iv)). These three trusts appear to fall squarely inside that safety net. But §2632(c)(5)(A)(i) lets a transferor elect out of the automatic rules, either for a particular indirect skip or for all transfers to a particular trust. The safety net exists; the returns cut a hole in it.

And here is the part that turns a paperwork problem into a money problem. Under §2642(b)(1)(A), an allocation made on a timely gift tax return — or deemed made under the automatic rules — values the property as of the date of the transfer. A late allocation, absent relief, does not. It is valued when it is made. For a trust funded years earlier with appreciating assets, that difference is the whole ballgame: the same dollar of exemption buys progressively less protection the longer the error goes undetected. The delay is not neutral. It compounds against the family.

Section 2642(g)(1)(A) directs the Secretary to prescribe by regulation the circumstances and procedures for extending the time to make an allocation or a §2632(c)(5) election, and §2642(g)(1)(B) instructs that all relevant circumstances be considered — including evidence of intent in the trust instrument — treating the deadline “as if not expressly prescribed by statute.” Those regulations arrived, after a sixteen-year gestation from proposal to finality, as Treas. Reg. §26.2642-7, effective for relief requests filed on or after May 6, 2024 (Relief Provisions Respecting Timely Allocation of GST Exemption and Certain GST Elections, T.D. 9996, 89 Fed. Reg. 37,116 (May 6, 2024), finalizing regulations proposed Apr. 17, 2008; relief under §2642(g)(1) is now granted under Treas. Reg. §26.2642-7 rather than under Treas. Reg. §301.9100-3).

The standard is two-pronged and conjunctive: relief is granted when the transferor establishes, to the Service’s satisfaction, that the transferor acted reasonably and in good faith and that granting relief will not prejudice the interests of the government (Treas. Reg. §26.2642-7(d)(1)).

The regulation lists non-exclusive factors on each side. On reasonableness and good faith: the transferor’s intent to allocate timely; intervening events beyond the transferor’s control; lack of awareness despite reasonable diligence, judged against the transferor’s own experience and the complexity of the GST rules; consistency in the transferor’s prior allocation practice; and — decisively here — reasonable reliance on the advice of a qualified tax professional (Treas. Reg. §26.2642-7(d)(2)(i)–(v)). On prejudice: any attempt to benefit from hindsight; delay calculated to run out the Service’s time to challenge the transfer; intervening taxable terminations or distributions; and circumstances involving expiration of limitations periods on transfer taxes (Treas. Reg. §26.2642-7(d)(3)(i)–(iv)).

The Service concluded the requirements were satisfied and granted the grantor 120 days from the date of the letter to allocate available GST exemption to the Year 1 transfers to all three trusts and the Year 2 and Year 3 transfers to Trust 1. Critically, the allocations “will be effective as of the date of the transfers and will be based on the fair market value for federal gift tax purposes of the property transferred on such dates” (Priv. Ltr. Rul. 202628005) — the appreciation problem, undone. The mechanics: amended Forms 709 for each year, filed with a copy of the ruling attached.

So the story ends well. It is worth being clear-eyed about what “well” cost. This was a private letter ruling — a request submitted in June 2025, ruled on in April 2026, released in July 2026. That is roughly ten months of professional time, user fees, and supporting affidavits under penalty of perjury, to repair an error that a single hour of review would have prevented. And the letter ends where every letter ends: it is directed only to the taxpayer who requested it, and under §6110(k)(3) it may not be used or cited as precedent. It tells us how the Service is thinking. It does not tell anyone else they will get the same answer.

Three things, and none of them are about the GST tax specifically.

First: the failure mode of a good structure is administrative, not architectural. Nobody in this ruling designed the trusts badly. The protection features we care about most — irrevocability, genuine trustee discretion, a real separation between the settlor and the property — were all present from day one and were never in question. The defect lived in the reporting layer, which is precisely the layer clients assume is clerical. A structure is not the deed. It is the deed plus every return, election, resolution, and valuation that follows it, for as long as it stands.

Second: relief for this kind of error is discretionary, and it decays. The regulation’s prejudice factors are pointed. Hindsight is disqualifying, and so is delay that eats the government’s time to examine the transfer. An intervening taxable distribution or termination between the missed deadline and the request is a listed problem. The practical translation: the window in which this is fixable is real but finite, and it narrows every year the error sits undiscovered. That is an argument for periodic independent review of a trust’s tax file, on a schedule, by someone who did not prepare it — which is exactly what finally caught the problem here, in Year 4, when a second attorney was brought in to look at the whole plan.

Third: reliance on professionals is a defense only if it was genuinely reasonable. The regulation credits reliance on a qualified tax professional, and the Service credited it here on facts where the attorney simply never raised the issue. But that factor sits inside a broader inquiry that also weighs the transferor’s own experience and diligence. It is not a blanket indemnity for whatever a return preparer did, and it should never be treated as one by a sophisticated client.

There is a fourth point, quieter and more uncomfortable, that we will state as a question rather than an answer. The grantor and spouse elected under §2513 to split every year’s gifts, which makes each of them the transferor of one half for GST purposes. The ruling’s operative grant runs to the grantor and the grantor’s “available GST exemptions.” Whether and how a spouse’s half is separately addressed in a given case is exactly the kind of detail that turns on the full record and the ruling request as submitted — and it is the kind of detail that repays close attention from counsel rather than assumption.

We have written before that protection must be seasoned, irrevocable and discretionary, and independently administered. PLR 202628005 adds a fourth, less glamorous requirement, and it belongs on the same list: protection must be maintained.

The trusts in this ruling were built correctly on a clear day, exactly as we would counsel. They spent several years quietly failing at their tax purpose anyway, because the returns that reported them were prepared without anyone asking what the trusts were for. The family got the fix. They got it because they eventually paid a second professional to read the file with fresh eyes — and because the law contains a relief provision that, on the right facts, is generous.

Neither of those is a plan. The plan is to read the return before it is signed, and to read the file again every year afterward.