QTIP Trust: Defer the Estate Tax, Keep Control of Who Inherits

Both qualify for the unlimited marital deduction. Both give the surviving spouse all the income for life. The difference is one word: control. A general power of appointment marital trust hands the survivor the power to decide where the property goes. A QTIP trust does not, and that is the entire reason it exists.

The general power of appointment trust lives in IRC section 2056(b)(5). To qualify, the surviving spouse must be entitled to all the income for life, payable at least annually, and must hold a power exercisable alone and in all events to appoint the entire interest to herself or to her estate. Practically, she can rewrite the remainder. She can leave it to her children, her second husband, a charity, or nobody in particular. The first spouse’s wishes are a suggestion.

The QTIP under section 2056(b)(7) flips that. The surviving spouse still gets all the income, at least annually, for life. She may get principal if the document allows it. What she cannot do is redirect the remainder, no person, including her, may hold a power to appoint any part of the property to anyone other than her during her lifetime, and at her death the property passes exactly where the first spouse’s document says it goes. In exchange for that constraint, Congress requires an affirmative election by the executor.

A useful middle ground exists and gets used constantly: a QTIP with a limited power of appointment exercisable by the surviving spouse only by her will, and only among a class the first spouse defined, say, “my descendants.” She can adjust shares among the children based on how the next twenty years go. She cannot hand the money to a new spouse or a new family. That structure preserves QTIP status because the power is not exercisable during her lifetime and cannot benefit anyone outside the permitted class.

A worked example. Robert dies with $9,000,000. He has two children from his first marriage, ages 34 and 31, and a second wife, Diane, who is 58 and has two children of her own. His will leaves everything to a marital trust paying Diane all income quarterly for life.

Version one: a general power of appointment trust. Diane has the power to appoint the property to herself or her estate. Robert’s estate takes a full $9,000,000 marital deduction and pays no federal estate tax. Diane lives 27 more years. In year nine she remarries. In year twelve she signs a new will leaving the trust property, now worth $17,400,000, to her own two children. That is entirely within her power, and Robert’s children receive nothing. Their only remedy is a lawsuit they will lose.

Version two: a QTIP. Same income to Diane, same $9,000,000 marital deduction on Robert’s Form 706, same zero tax at his death. But the document names Robert’s two children as remainder beneficiaries, and Diane cannot change it. At her death 27 years later the $17,400,000 is included in her gross estate under IRC section 2044, her executor computes the incremental estate tax attributable to it, and IRC section 2207A lets her estate recover that tax from the trust, so Robert’s children bear the tax on their own inheritance rather than Diane’s children subsidizing it. Robert’s children receive the property with a fresh basis under section 1014(b)(10). Everyone ends up where Robert intended.

The tax outcome is identical in both versions. Only the human outcome differs, and the human outcome is why families pay for the drafting.

The common mistake: using a general power of appointment trust in a second marriage because the attorney’s form file defaults to it, or because “she’d never do that.” Twenty-seven years is a long time and people remarry. The mirror-image mistake is using a QTIP in a long first marriage with shared children and no blended-family risk, where the added rigidity buys nothing and an outright bequest with a portability election is simpler, cheaper, and gets the same tax answer.

A second technical mistake shows up in the drafting. A QTIP that permits the trustee to distribute principal to anyone other than the surviving spouse during her life is disqualified, and the marital deduction fails for the whole trust. A well-meaning clause letting the trustee help a grandchild with tuition destroys the election. So does a document that gives the trustee discretion over income rather than mandating annual distribution. Both errors are usually invisible until the estate tax return is being prepared nine months after death, and by then the document is what it is.

New York adds a wrinkle that catches out-of-state drafters. Under EPTL 5-1.1-A a surviving spouse has a right of election against the will, generally the greater of $50,000 or one-third of the net estate, counting a long list of will substitutes such as joint accounts and retirement plans. For deaths after August 1992 that elective share has to be satisfied with property passing outright. An income interest in a QTIP does not count toward it. A surviving spouse who dislikes the arrangement can elect against the will, take her elective share free of trust, and unwind a large part of the plan. Prenuptial or postnuptial waivers are the usual answer, and in a second marriage that waiver is not optional paperwork.

One more distinction worth holding onto. The general power of appointment trust is included in the surviving spouse’s estate under IRC section 2041 because she holds the power. The QTIP is included under section 2044 because of the election. Same inclusion, different statute, and the section 2207A recovery right applies only to the QTIP. That asymmetry decides who actually writes the check at the second death, which is why the apportionment language in the survivor’s own will has to be reviewed against the first spouse’s trust rather than drafted in isolation.

Going forward, the questions to bring to the drafting meeting are these: does the surviving spouse need the ability to redirect the remainder, or is that the exact risk we’re trying to eliminate? Are there children from a prior marriage? Is a limited power exercisable only by her will, among a defined class, the right compromise? And who serves as trustee, given that whoever it is will be balancing an income beneficiary against remainder beneficiaries who may not be her children? Our guide to the trustee role covers that last question in detail. This is general information, not legal advice for your family; work through the specifics with an estate attorney and a licensed CPA.

The surviving spouse pays tax on the income she receives, because she has to receive all of it. The trust pays tax on whatever is left, which in practice means capital gains allocated to principal. Two taxpayers, one trust, and the split follows the principal-and-income rules rather than the tax code.

Start with the mandate. To qualify under section 2056(b)(7), the spouse must be entitled to all the income from the property, payable annually or at more frequent intervals. That is a floor set by federal tax law and it overrides any contrary language in the document. A trustee who accumulates income has not merely breached a fiduciary duty. He has jeopardized the QTIP status of the entire trust.

What counts as income is a different question, answered by state law and the trust instrument. Under most states’ versions of the Uniform Principal and Income Act, interest, dividends, and net rental income are income. Realized capital gains are principal. Depreciation reserves, capital improvements, and debt principal payments come out of principal. Trustee commissions are typically split between the two. New York’s default rules sit in EPTL Article 11-A.

The regulations add a protection people forget: the surviving spouse must have the power to compel the trustee to make unproductive property productive, or to convert it to productive property within a reasonable time. A QTIP funded with vacant land, a non-dividend-paying stock, or a vacation house the survivor doesn’t use is a qualification risk unless the document grants that power. Auditors have raised it.

On the income tax mechanics, the trust files Form 1041. Because it must distribute all income annually, it takes a distribution deduction under IRC section 661 limited by distributable net income, and issues the surviving spouse a Schedule K-1 showing her share by character, ordinary dividends, qualified dividends, taxable interest, tax-exempt interest, rental income. She reports each piece on her own Form 1040 in the corresponding place, at her rates. Get the K-1 to her late and you have handed her a reason to extend her personal return every year.

A worked example. A QTIP holds $6,400,000: a $4,100,000 balanced portfolio, a $1,900,000 net-leased retail building, and $400,000 of cash. For the year it produces $71,000 of taxable dividends and interest, $22,000 of tax-exempt municipal interest, $138,000 of net rent after operating expenses, and a $260,000 long-term capital gain from rebalancing. Trustee commissions run $38,000 and tax preparation $4,500.

Fiduciary accounting income is $231,000, the dividends, the muni interest, and the net rent, reduced by the income share of the commissions, call it roughly $212,000 payable to the surviving spouse. The trust distributes it quarterly. Her Schedule K-1 reports it by character, so the $22,000 of municipal interest passes through tax-exempt and only about $190,000 is taxable to her. At a 32% marginal rate that’s roughly $61,000 of federal tax on her 1040.

The $260,000 capital gain is a different story. Allocated to principal and not distributed, it stays in the trust and is taxed at trust rates, the 20% long-term capital gain rate plus the 3.8% net investment income tax under IRC section 1411, reported on Form 8960. That’s about $62,000 of tax, paid by the trust, reducing the principal that eventually goes to the remainder beneficiaries. Same trust, same year, two taxpayers, and the second one is paying at compressed rates on money nobody received.

Now the planning point. Had the trust instrument allowed the trustee to allocate capital gains to income, or to include them in distributable net income consistently, that $260,000 could have been pushed out to the surviving spouse and taxed in her brackets. Whether that helps depends entirely on her other income, if she is already at the top rate it saves almost nothing, and if she is in the 15% capital gain bracket it saves real money. The regulations require consistency: you cannot include gains in DNI in a year it helps and exclude them in a year it doesn’t. Trustees who set that policy in the first year of administration capture the benefit for the life of the trust. Trustees who discover it in year seven usually cannot fix the earlier years.

The common mistake: a trustee who reads the Form 1041 and distributes “the income” shown on the tax return. Taxable income on the 1041 in the example above is well over $400,000, and distributing that number would push roughly $200,000 of principal to the surviving spouse that belonged to the remainder beneficiaries. The reverse error, distributing only what the K-1 reports and shorting a spouse entitled to more under the document, is equally common and equally uncorrectable after a few years. Run fiduciary accounting income and taxable income as two separate computations every year, in two separate columns, from the first transaction.

The second common mistake is timing. Income earned but undistributed at the surviving spouse’s death, stub income, does not have to be payable to her estate for the trust to qualify, and most documents provide that it isn’t. Trustees who reflexively cut a final check to the estate are giving remainder beneficiaries’ money away.

One timing tool is worth knowing. IRC section 663(b) lets a trustee elect to treat distributions made within the first 65 days after year end as though they were made on the last day of the prior tax year. For a QTIP that has to distribute all income annually, the 65-day election is how a trustee cleans up a fourth-quarter shortfall discovered while the Form 1041 is being prepared, a genuinely useful backstop, but it has to be affirmatively elected on a timely filed return and it cannot rescue a distribution made in month four of the following year.

Going forward: distribute at least quarterly and document each distribution; set the capital gain allocation policy in year one and write down the reasoning; issue K-1s by the statutory deadline rather than on extension; and give the surviving spouse an annual accounting whether or not anyone asks, because the accounting she never received is the first exhibit in every trust dispute. Our bookkeeping team maintains principal-and-income ledgers for exactly this reason. None of this is advice for your trust, have a licensed CPA review the instrument and the numbers before you set an allocation policy you’ll live with for decades.

The honest answer is that portability handles the tax and a QTIP handles everything portability cannot. For a first marriage with shared children, modest assets, and no state estate tax, portability alone is usually enough and far cheaper to administer. For a second marriage, a state with its own estate tax, a family with generation-skipping goals, or a surviving spouse with creditor exposure, portability alone is a mistake.

Portability arrived permanently in 2012. When the first spouse dies, the executor can elect to transfer the deceased spousal unused exclusion amount to the survivor by making the election on a timely filed Form 706. The survivor then has her own exclusion plus the DSUE. Simple, and it eliminated the old requirement to split assets into two trusts purely to use both exclusions.

Here is what portability does not do, item by item.

It does not control the remainder. Property left outright to a surviving spouse is hers. She can leave it to a new husband, to her own children, to a charity, or to a person nobody in the family has met. A QTIP is the only structure that gets the marital deduction while nailing down the remainder.

It does not survive her remarriage cleanly. DSUE comes from the survivor’s last deceased spouse. If she remarries and her second husband predeceases her, the DSUE from husband number one is replaced by whatever husband number two’s estate provides, which may be nothing.

It does not carry the GST exemption. The generation-skipping transfer tax exemption is not portable. Full stop. A couple who wants to fund a dynasty trust for grandchildren must use the first spouse’s GST exemption at the first death, and the tool for doing that inside a marital trust is the reverse QTIP election under IRC section 2652(a)(3). Rely on portability and that exemption evaporates.

It does not protect assets. Property held in a properly drafted QTIP is generally beyond the reach of the surviving spouse’s creditors, a future divorce, or a bad second marriage. Property received outright is not.

It does not help a non-citizen spouse at all. The unlimited marital deduction is unavailable for property passing to a surviving spouse who is not a U.S. citizen unless it passes to a qualified domestic trust under IRC section 2056A, which requires at least one U.S. trustee with the power to withhold estate tax on principal distributions. DSUE does not fix that. A QTIP for a non-citizen spouse has to satisfy both the section 2056(b)(7) requirements and the QDOT requirements simultaneously, and green-card status is not citizenship for this purpose.

It does not work for New York. New York has no portability. Neither do most states with their own estate tax. A couple relying entirely on the federal DSUE wastes the first spouse’s New York exclusion completely, and New York’s cliff makes that expensive: exceed 105% of the state exclusion and the exclusion vanishes rather than phasing down.

A worked example. A New York couple, both 71, holds $14,000,000 of combined assets, roughly evenly titled. Assume a New York basic exclusion of about $7,200,000 per person. The figure is indexed annually, so confirm the current one on the Department of Taxation and Finance estate tax page before planning.

Plan A, everything outright with a federal portability election. Husband dies first leaving $7,000,000 to his wife. No federal tax, no New York tax, DSUE elected. She now owns $14,000,000. At her death, assume it has grown to $18,000,000. Federally she is fine, her own exclusion plus his DSUE covers it. In New York she has one exclusion of roughly $7,200,000, and $18,000,000 is far past 105% of it, so the exclusion is gone entirely and New York taxes the full $18,000,000 on a graduated scale topping at 16%. The New York bill lands well over $2,000,000.

Plan B, the same estate with a QTIP and a New York state-only QTIP election on Form ET-706, combined with a credit shelter trust funded to the New York exclusion. At the husband’s death, roughly $7,200,000 goes into a credit shelter trust using his New York exclusion, and the balance goes into a QTIP with the New York election made. No tax at the first death either way. At the second death, the credit shelter trust, now perhaps $9,300,000 after growth, is outside her New York estate entirely. Her own New York exclusion shelters another $7,200,000 or so. The New York tax drops by well over a million dollars.

The catch, and it is a real one: the credit shelter portion gets no second basis step-up. The children inherit that $9,300,000 with the husband’s original basis. If his basis was $2,000,000, they are carrying $7,300,000 of unrealized gain that would have disappeared under Plan A. At combined federal and New York capital gain rates, that embedded tax is worth roughly $2,000,000 if they ever sell. The QTIP portion, by contrast, is included under section 2044 and does get a fresh basis under section 1014(b)(10).

The common mistake: treating this as a rule rather than a computation. The right split between credit shelter and QTIP depends on the state exclusion, the basis of the specific assets, the survivor’s life expectancy, and whether the family intends to sell or hold. A Clayton-style formula that lets the executor decide nine months after death, with actual numbers, beats any allocation locked in years earlier. The second common mistake is failing to file Form 706 at all in a modest estate, which forfeits portability outright, Rev. Proc. 2022-32 gives a five-year rescue for estates not otherwise required to file, and families routinely find out about it in year six.

Going forward, if you live in a state with its own estate tax, assume portability alone is not the plan. Model both structures with real basis numbers. Build in executor flexibility rather than a fixed formula. And file the estate tax return at the first death even when nothing is owed. Our capital gains guide covers the basis side of this tradeoff. This is general information rather than advice for your estate; the numbers above are illustrative and the exclusions change annually, so work through your own with a licensed CPA and an estate attorney.

Yes to the first, and the second is a separate election that does something entirely different despite the similar name. Both are made on Form 706, both are executor decisions rather than drafting decisions, and both are irrevocable once the filing deadline including extensions has run.

The partial election. An executor may elect QTIP treatment for part of a trust rather than all of it. The critical rule is the form the election takes: it must be expressed as a fraction or percentage of the entire property, so that the elected and non-elected portions share proportionately in subsequent income, appreciation, and depreciation. An executor cannot elect QTIP treatment for “the first $4,000,000” of a trust and leave the rest, a pecuniary carve-out fails. Elect 62% of the trust and the elected share is 62% of everything, forever.

Why bother? Because the exclusion amount and the size of the estate are both known nine months after death, and neither was known when the document was signed. The non-elected portion does not qualify for the marital deduction, so it uses the first spouse’s exclusion, which is exactly what a credit shelter trust does. Rather than dividing assets into two trusts on day one, a single trust with a partial election accomplishes the same split with better information.

The Clayton election takes this a step further. Regulations permit the disposition of property to depend on the executor’s QTIP election: the portion elected stays in the marital trust with mandatory income to the spouse, and the portion not elected pours into a bypass trust with different terms, discretionary distributions, different beneficiaries, no requirement to pay all income to the survivor. It is the most flexible drafting device in this area, and it requires an executor who is not the surviving spouse, since letting her choose creates a taxable gift problem.

The reverse QTIP election is unrelated to any of that. Under IRC section 2652(a)(3), an executor may elect to treat the first spouse as the transferor of QTIP property for generation-skipping transfer tax purposes only. Without it, the surviving spouse becomes the GST transferor when the property is included in her estate under section 2044, and the first spouse’s GST exemption, which is not portable, is simply wasted. The reverse QTIP election is how a couple uses both GST exemptions when the first spouse’s assets pass through a marital trust.

Two mechanical points. The reverse QTIP election must be made for the entire trust for which the QTIP election was made. You cannot elect reverse QTIP on part of a QTIP trust. If you need a partial result, split the trust into two separate trusts and make the reverse election on one of them. And the reverse election has no effect on estate tax at all; the property is still included in the surviving spouse’s gross estate under section 2044 exactly as before.

A worked example. Margaret dies with $19,000,000. Her will funds a single marital trust for her husband Paul, remainder to their descendants, with grandchildren as ultimate takers. Assume a federal exclusion and GST exemption of $15,000,000 each for the year of death, confirm the current figures on the IRS estate tax page, since both are indexed.

Her executor makes a partial QTIP election. He elects QTIP treatment for a fraction equal to $4,000,000 divided by $19,000,000, about 21%, so that roughly $4,000,000 qualifies for the marital deduction and defers, while the remaining $15,000,000 does not qualify and absorbs Margaret’s full exclusion. Federal estate tax at her death: zero. If instead he had elected 100%, the entire $19,000,000 would have deferred to Paul’s estate, where it would compete with Paul’s own $15,000,000 exclusion and his own assets, and Margaret’s exclusion would have gone unused except through portability, which does not carry GST exemption.

Now the GST layer. The executor splits the trust and makes a reverse QTIP election on the elected marital portion, then allocates Margaret’s GST exemption to it. Result: for GST purposes Margaret is the transferor, her exemption shelters that trust, and distributions to grandchildren decades later avoid the generation-skipping transfer tax entirely, a 40% tax on the value transferred. Skip the reverse election and Paul becomes the transferor at his death, Margaret’s GST exemption is gone, and Paul’s own exemption has to stretch across everything.

Assume the trusts grow to $34,000,000 by Paul’s death and roughly a third eventually reaches grandchildren. The reverse QTIP election on those facts is worth several million dollars of GST tax. It is one checkbox and one allocation schedule on a return filed nine months after death.

The common mistake: making a 100% QTIP election reflexively because it produces a zero tax bill at the first death and the return is easy. That is optimizing for the wrong year. Over-electing pushes everything into the survivor’s estate and wastes the first spouse’s exclusion and GST exemption. Under-electing generates tax at the first death that did not need to be paid. The right fraction is a computation involving the exclusion amount, projected growth, the survivor’s life expectancy, the survivor’s own assets, the basis of the underlying property, and any state estate tax, and it is genuinely different for every family.

The second common mistake is a protective or defective election. The regulations are specific about form, and an election expressed in dollars rather than as a fraction, or a reverse QTIP election made on part of a QTIP trust, may not hold up. This is not the return to hand to a preparer who files two estate tax returns a year.

Going forward, the executor’s checklist at month six looks like this: value everything, compute the exclusion actually available after lifetime gifts, model the second death at several growth rates, decide the QTIP fraction, decide whether to split trusts for GST purposes, make the reverse election on the right one, and allocate GST exemption on the schedule. Then file by month nine, or extend on Form 4768 and use the extra six months to get it right. Our tax strategy consulting practice runs those models. This page is general information and not advice for your estate; an estate attorney and a licensed CPA should make these elections together, on your facts.

Three things happen at once, and the executor of the surviving spouse’s estate has to handle all of them. The property is pulled into her taxable estate, someone has to pay the resulting tax, and the assets get a fresh income tax basis. Miss any one and a family loses money.

Inclusion. IRC section 2044 includes the value of QTIP property in the surviving spouse’s gross estate at fair market value on her date of death. She never owned it, could not sell it, and may have had no idea it existed as a separate pot. It goes on her Form 706 anyway, reported on Schedule F. This is the deferral coming due, and the top federal estate tax rate is 40%. Her executor needs the first spouse’s estate tax return to do this correctly, which means someone has to find a document filed potentially decades earlier, a recurring practical nightmare and a good reason to store the first spouse’s Form 706 with the survivor’s will rather than in a lawyer’s dead files.

Who pays. IRC section 2207A gives the surviving spouse’s estate a statutory right to recover from the QTIP property the incremental estate tax attributable to including it. The difference between the tax actually due and the tax that would have been due without the QTIP inclusion. The remainder beneficiaries bear the tax on their own inheritance. That is the default and it is usually what everyone intends.

The right can be waived, and waiving it by accident is one of the most expensive drafting errors in estate practice. A boilerplate clause in the survivor’s will directing that “all estate taxes be paid from my residuary estate” can override section 2207A, shifting the entire QTIP tax onto her residuary beneficiaries, typically her own children, while the first spouse’s children take the QTIP free of tax. Families have litigated for years over one sentence copied from a form. If a QTIP exists anywhere in the picture, the survivor’s tax apportionment clause has to be drafted with section 2207A open on the desk.

Basis. Because the property is included under section 2044, it gets a basis adjustment to date-of-death value under IRC section 1014(b)(10). Remainder beneficiaries take the assets with a fresh basis, and decades of unrealized appreciation disappear for income tax purposes. This is the underappreciated half of the QTIP bargain, and with federal exclusions where they are it is frequently worth more than any estate tax the structure defers.

A worked example. Robert died in 2009 leaving $4,800,000 to a QTIP for Diane, remainder to his two children. His executor made the QTIP election and took a full marital deduction. Diane dies today. The trust is now worth $11,600,000, and the underlying assets have a carryover basis of about $2,900,000. Diane’s own assets total $9,400,000.

Her gross estate is $21,000,000, her $9,400,000 plus the $11,600,000 of QTIP property under section 2044. Assume a federal exclusion of $15,000,000; the figure is indexed, so verify the current one before relying on it. Her taxable estate is roughly $6,000,000, taxed at 40%, producing federal estate tax of about $2,400,000. Without the QTIP inclusion, her taxable estate would have been zero. The full $2,400,000 is therefore attributable to the QTIP property, and under section 2207A her executor recovers it from the trust. Robert’s children net roughly $9,200,000. Diane’s own beneficiaries receive her $9,400,000 intact.

Now flip one clause. Suppose Diane’s will says all death taxes are paid from her residuary estate and does not preserve the section 2207A right. Robert’s children take the full $11,600,000, Diane’s children absorb the $2,400,000, and they receive $7,000,000 instead of $9,400,000. Same statute, same assets, one sentence, $2,400,000 moved between two families.

The basis piece runs alongside. That $11,600,000 of QTIP property carried a $2,900,000 basis and now steps up to $11,600,000. If Robert’s children sell, they avoid tax on $8,700,000 of gain, worth roughly $2,600,000 at combined federal and New York rates on a long-term gain including the net investment income tax. That step-up would not have happened had Robert used a credit shelter trust instead, which is the tradeoff every couple with a taxable estate has to price.

New York and the states. If the survivor was a New York resident, her executor also files Form ET-706, and the New York taxable estate includes QTIP property for which a New York election was made. New York’s cliff bites here, exceed 105% of the state exclusion and the exclusion disappears entirely rather than phasing out. New York estate tax rates top out at 16%. The federal and state elections were separate, so the answers can differ, and the surviving spouse’s executor has to know what was elected at both levels years earlier.

The common mistake: the survivor’s executor who does not know the QTIP exists. Nothing in her own records shows it. The trustee may not volunteer it. A Form 706 filed without section 2044 property is a substantially understated return, and the estate tax statute of limitations does not start running on property that was never disclosed. The second common mistake is failing to claim the section 2207A recovery, either because the will inadvertently waived it or because the executor never raised it with the trustee. The third is forgetting the trust’s own final Form 1041, the QTIP terminates, files a final fiduciary return, and passes out any excess deductions and unused loss carryovers to the remainder beneficiaries on their final K-1s. Those carryovers are worth real money and get abandoned constantly.

Going forward, keep the first spouse’s Form 706 with the survivor’s estate documents, permanently. Tell the survivor’s executor the QTIP exists before it matters. Review her tax apportionment clause against section 2207A whenever her will is updated. Get date-of-death appraisals on the trust assets, not just her own. And run the trust’s final fiduciary return with the excess deduction pass-through in mind. Our guide to taxes on inheritance covers what beneficiaries face on the receiving end. This is general information rather than tax or legal advice for your situation; a licensed CPA and an estate attorney should review the documents and the numbers before anything is filed.