The 10-year rule, applicable to most trusts that will receive retirement benefits after passage of the Setting Every Community Up for Retirement Enhancement Act (SECURE Act), presents a dilemma for non-Roth individual retirement accounts. For the vast majority of beneficiaries whose income level doesn’t rise to the top income tax bracket, leaving such assets in trust may double or triple the income tax due to trapping the income into highly compressed trust income tax brackets. Tax advisors might recommend trust distributions of all the income to the beneficiaries to save income taxes (if possible under the terms of the trust). While this may often save income taxes, it destroys the asset protection benefits of the trust and wastes any generation- skipping transfer (GST) tax exemption as to the distributions made. Thus, even an accumulation trust will often function for all practical purposes as a conduit trust to save income tax, limiting the estate and GST tax sheltering and asset protection for traditional retirement assets to only 10 years.
An alternative trust structure can achieve the best of both worlds. A beneficiary deemed owner trust (BDOT) can shift the income tax to beneficiaries who are usually in much lower tax brackets without destroying the trust’s estate/ gift/GST tax and asset protection benefits.
Edwin Morrow is an attorney with the law firm of Kelleher + Holland, LLC in North Barrington, Ill. and co-chair of its Estate Planning Group
Flaws of See-Through Trusts
See-through trusts, designed to qualify as designated beneficiaries for determining distribution periods after the owner’s death, come in two basic varieties: conduit trusts and accumulation trusts.1 The conduit trust passes out any distributions from the retirement plan to the trust to the beneficiary; the accumulation trust may keep such distributions in the trust.
Practitioners should re-examine the types of trusts in light of whether the settlor’s goals lean towards asset protection or income tax reduction. This article won’t discuss eligible designated beneficiaries (EDBs) who may still come under the old stretch rules that allow payments over the beneficiary’s life expectancy. EDBs include surviving spouses, those who are disabled or chronically ill, minor children of the owner (until age 21, at which point the 10-year rule kicks in), and beneficiaries not more than 10 years younger.2 Although the concepts and benefits of using a BDOT structure could equally apply to them, the income tax problem associated with compressed trust tax brackets is more acute when the trust is for designated beneficiaries subject to the 10-year rule.
The popular wisdom is that conduit trusts will always be a disaster after the passage of the SECURE Act. This isn’t always true, though some may be. Some accumulation trusts might also now be a disaster. It depends on the clients’ goals, the beneficiaries’ financial situation and the trusts’ flexibility.
Conduit trusts and trusteed IRAs (which is a type of IRA with trust-like distribution restrictions built into the IRA agreement itself) drafted only to pay required minimum distributions (RMDs) (and no more) could be an income tax and estate-planning disaster. If the owner dies before their required beginning date, all the retirement funds would come out in the 10th year into the trust and then out to the beneficiaries, allowing no use of the beneficiaries’ lower tax brackets by distributing funds more gradually before then. Unless a beneficiary’s income is in the top tax bracket,3 this is a tax disaster for larger amounts because the beneficiary’s lower tax brackets weren’t efficiently used.
Most conduit trusts and trusteed IRAs, however, are drafted with much more f lexibility (for example, the trustee must pay RMDs and may distribute more at the trustee’s discretion or under ascertainable standards). Thus, conduit trusts and trusteed IRAs will rarely be an income tax disaster. The problem with conduit trusts usually isn’t the income tax ramifications but that all the funds leave the trust over 10 years, severely limiting the trusts’ estate/gift/GST tax and other asset protection benefits. Is it worth bothering with a trust for only 10 years of protection? Many taxpayers would be tempted to leave such funds outright if that were the only alternative.
Accumulation trusts are much more protective but may cause an income tax disaster if they have conservative distribution standards and lack any lifetime limited powers of appointments (POAs) or spray powers. In many such cases, retirement plan distributions may be trapped in trust at the highest trust compressed tax bracket (currently 37%, but Congress may always increase this again, and the top tax bracket is due to rise to 39.6% in 2026).
For example, if the trust instrument provides that distributions be made for “health, education and support,” but nothing further, does this permit the trustee to pay out the entire IRA? Distribution standards are restrictions, even if the beneficiary is a trustee or co-trustee. If a $2 million IRA would have paid $60,000 annual RMDs under the pre-SECURE Act rules and the trustee would have been justified in paying that much to beneficiaries, is the trustee now equally justified to distribute more than $200,000/ year over only 10 years, even when this distribution schedule leaves no principal from those accounts for the remaindermen? Is the higher distribution within the current distribution standards designed by the settlor under a quite different income tax environment? Some trust instruments instruct a trustee to be conservative with distributions of principal—or even forbid them. Retirement plan distributions over those 10 years are often likely to be 100% principal (or maybe 90% in the final year).4 Are such distributions of principal to beneficiaries justified under the trust instrument?
This issue doesn’t exist in every trust. Some trusts have an independent trustee with absolute discretion or liberal distribution standards that would justify higher payouts that are now encouraged from an income tax standpoint under the compressed 10-year stretch.
What’s the settlor’s wish? To minimize income taxes or protect the corpus and preserve principal for remaindermen? Practitioners should strongly consider giving the trustee more guidance in making distributions in light of this dilemma.
This brings us back full circle to the central tension between conduit and accumulation trusts. If a trustee of an accumulation trust distributes most or all of its taxable income received from retirement funds to be more income tax efficient, it’s effectively no different from a conduit trust as a practical matter and loses its asset protection and estate tax sheltering advantage!
The ideal scenario for many will be to enable the shifting of income tax to exploit the lower marginal tax brackets applicable to most beneficiaries without destroying the asset protection benefits of the trust. Families often want their beneficiaries to have the protections of a trust, but not if the cost of this structure is to increase the tax on millions of dollars of taxable income from 12%, 22% or 24% to 37%! This could easily cost the family hundreds of thousands of dollars in additional income tax.
How BDOTs Achieve Results
BDOTs were introduced and updated in two white papers from 2011-2018, which may be consulted for more detail.5 There are two ways to shift income to beneficiaries without making a distribution. Here’s how the most common one works. Internal Revenue Code Section 678 provides:
- General rule
A person other than the grantor shall be treated as the owner of any portion of a trust with respect to which:
- such person has a power exercisable solely by himself to vest the corpus or the income therefrom in himself, or…
(Emphasis added.)
Treasury regulations are crystal clear that “income” in IRC Section 678(a)(1) refers to taxable income, not accounting income.6 This is unlike other areas of nongrantor trust income taxation in which the term refers to accounting income.7
It is, therefore, crucial to define any withdrawal right. If the document refers to “income” with no further elaboration, this, by default, will only cover accounting income. If, however, a trust instrument provides that a beneficiary of a trust has the power solely exercisable in themselves to withdraw all the taxable income from a trust (regardless of whether it’s principal or income under state law), the beneficiary must pay the income tax on this income under Section 678(a)(1).
Importantly, whether the beneficiary actually withdraws the income (or how much or where it comes from) is completely irrelevant.8 The statute doesn’t require that the beneficiary hold any power over the principal beyond the trust’s taxable income. It applies whether the power enables vesting of the “corpus or the income” (emphasis added). Any trust in which the beneficiary must be taxed on all the taxable income yet doesn’t necessarily have any power over the corpus causing creditor access or estate taxation beyond this is referred to as a BDOT. Here’s how a BDOT may operate: John, age 71, has a $7 million estate, $2 million of which is in a traditional IRA, $1 million in a Roth IRA and $4 million of other assets. He splits his trust between his son and daughter into two accumulation trusts with $3.5 million each, $1 million of which is comprised of an inherited IRA payable to the trust, $500,000 in the inherited Roth IRA and $2 million of other assets. The 10-year rule applies as neither child is an EDB. The son and daughter otherwise have taxable income, placing them in 22% and 24% tax brackets, respectively.
In Year 1 to Year 10 after John’s death, the trustee of each subtrust takes $100,000 (not an RMD, just a distribution) for 10 taxable years from the IRA (ignoring growth to simplify the math and highlight the issue). The trust makes 3% taxable interest and dividends on the other $2 million non-Roth IRA taxable assets every year—$60,000 (again, ignoring growth).
Each year, the trust grants each child the right to withdraw the taxable income, which is $100,000 (from the traditional IRA) plus $60,000 (interest/ dividends) = $160,000. Each child is taxed on all $160,000 of income regardless of how much they actually take. Let’s assume that the child withdraws 40% of this amount to pay their taxes, spends a bit ($64,000) and allows the $96,000 to lapse and remain in the trust. Had a conduit or accumulation trust with liberal distribution standards been used, $160,000 would have left this trust over 10 years instead of only $64,000. Over 10 years of distributions, this is a huge difference: $96,000 x 10 years = nearly $1 million more that’s protected in trust at the end of 10 years, with the same income tax minimization benefits. Considering growth, this difference could easily be protecting twice as much.
Wealthier beneficiaries who have the wherewithal to pay the income tax burden, just like those who voluntarily establish irrevocable grantor trusts, may leave even more in trust to let it effectively grow tax-free for their descendants like any other grantor trust, which means that more than an additional half a million dollars would be in a protected GST tax- exempt trust outside of the beneficiary’s estate.
Needier beneficiaries may withdraw more but at least there’s the option of keeping these retirement funds protected in the trust that wouldn’t be there under other trust designs without trapping the income in the trust at more confiscatory brackets. A trust protector might have authority under the trust to later eliminate this withdrawal right if warranted for non-tax reasons, but this would turn off the more advantageous BDOT tax status.
To the extent that such a power isn’t exercised and is allowed to lapse, only the greater of $5,000 or 5% is protected from being considered a taxable transfer, and any amounts allowed to lapse above this threshold will be considered an additional contribution to the trust by the beneficiary for estate/gift/GST tax purposes.9 In many states, this 5% rule also carries over into state debtor/creditor protection, but a surprising number of states follow the old common law rule that doesn’t regard such a lapse as creating a self-settled trust.10
In our example above, the entire $160,000 of taxable income in each trust was slightly under 5% of the trust corpus ($3.5 million x 5% = $175,000), and the amounts allowed to lapse ($96,000) were well under this amount. What if the trust earned $200,000 and the beneficiary had allowed it all to lapse without taking any income? In such case, $25,000 of the $200,000 total would be over the 5% lapse protection of $175,000 and would be considered a contribution to the trust for gift/estate tax purposes, which would cause a fraction to be considered self- settled under some state debtor/creditor laws and cause such amount to potentially be included in the power holder’s estate.
To avoid this, the beneficiary can withdraw $25,000 or more, and the withdrawal power over less than the 5% protected amount would safely lapse. That amount could go into any number of estate and asset protection vehicles. Beneficiaries are also known to, on rare occasions, spend money, so the likelihood is that lapses will rarely exceed 5% of the trust corpus as a practical matter, and there are various solutions to reduce this risk if it does.
For example, the BDOT should borrow from the “hanging power” concept so often used in Crummey trusts and account for any excess amount separately and provide that this amount “hang” and remain withdrawable (and likely lapse) in future years. If the amount is only withdrawable with the consent of a non-adverse trustee or the amount is instead subject to a testamentary general POA, in many states, such amounts would still be protected from creditors, yet under such conditions, this wouldn’t cause a taxable release for federal estate/gift/GST tax purposes because it would still be a general POA.11
In short, the 5% limitation on lapse protection will rarely be a major problem, especially when other assets comprise a portion of the trust (and may be invested to produce more growth and less taxable income if desired).
Crediting Distributions
The second way a trust might shift the tax burden without actually distributing assets out of the trust is for the trustee to credit a beneficiary with having received the income, which can still be considered a distribution that carries out distributable net income (DNI) to the beneficiary.12 This method is less commonly recommended because of various uncertainties in whether such a power can lapse and what’s required to credit such assets to a beneficiary.13
Top Tax Brackets
Wealthier beneficiaries whose income is in the top tax bracket may not experience such tension. Usually, those taxpayers are concerned with asset protection and estate and GST tax. A BDOT structure exploits the GST tax exemption and protection much more effectively than an ordinary nongrantor trust because a BDOT can grow income tax-free by shifting the income tax bill to the beneficiary.
BDOTs for Roth IRAs
Roth IRAs don’t have a problem trapping income at the top tax bracket. Trustees will want to wait until the last year possible to withdraw such funds because they usually come out tax-free. However, for those concerned with asset protection and estate/GST tax (and the number of taxpayers in this category may greatly increase in 2026 with the expiration of the doubled exclusion amount), the Roth IRA payable to a GST tax-exempt BDOT is a match made in heaven for long-term estate and GST tax savings. There are 10 years of completely tax-free growth due to the Roth IRA tax shelter, followed by tax-free growth after the Roth IRA is fully distributed and reinvested due to the BDOT shifting the taxable income burden to the beneficiary.
Building Flexibility
A BDOT won’t fit every situation—no one trust design ever does. Granting an unfettered withdrawal power to substance abusing or special needs beneficiaries doesn’t make any sense.
That said, it will fit many situations. An independent trustee or trust protector may even have the power to prospectively add or eliminate these withdrawal rights.
The old paradigm in which settlors rigidly distribute accounting income only so that principal grows for the next generation is a relatively rare request these days. It’s much more common for clients to want to minimize dead hand control and maximize primary beneficiary access, while still providing the most protection they can for creditor and divorce situations and being able to shelter assets from potential estate tax exposure. A BDOT design fits this desire perfectly and can largely solve the inherent tension between asset protection and income tax minimization in light of compressed tax brackets and compressed distribution schedules that the SECURE Act now foists on unsuspecting taxpayers.
Endnotes
- Treasury Regulations Section 1.401(a)(9)-5, Q&A7(c)(3) contains two The proposed regulations issued in 2022 have definitions of the two terms. See Prop. Regs. Section 1.401(a)(9)-4(f)(1)(ii).
- See Internal Revenue Code Sections 401(a)(9)(E)(ii) and 401(a)(9)(H)(ii).
- For tax year 2024, single taxpayers don’t reach the top tax bracket until taxable income exceeds $609,350, and married couples filing jointly don’t reach this until it exceeds $731,200. Trusts and estates, however, start at only $15,200. Revenue Procedure 2023-34.
- See Uniform Principal and Income Act (UPIA) Section The general rule that required minimum distributions (RMDs) are 90% principal and 10% accounting income and non-RMDs are 100% principal has been changed by the Uniform Fiduciary Income and Principal Act (UFIPA), but at the time of this article has thus far only been passed in seven states. See www.uniformlaws.org for updates in this area. Some states have modified this section of the UPIA without passing UFIPA. Recall that when the owner dies before their required beginning date, there are no RMDs until the 10th year, so any distributions from the retirement account during this time would likely be treated as principal.
- The concept was introduced in 2011 in Edwin Morrow, “The Optimal Basis Increase and Income Tax Efficiency Trust,” https://papers.ssrn.com/sol3/ cfm?abstract_id=2436964 and more specifically and thoroughly in Edwin Morrow, “IRC §678 and the Beneficiary Deemed Owner Trust (BDOT),” https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3165592.
- Treas. Regs. Section 1.671-2(b).
- Treas. Regs. Section 1.643(b)-1.
- See, g., Mallinckrodt v. Nunan, 146 F.2d 1 (8th Cir. 1945), Pre-IRC Section 678 Clifford regulations under Treas. Regs. Section 39.22(a)-22, IRC Section 678(a)(1) and Treas. Regs. Section 1.678-1, Campbell v. Commissioner, T.C. Memo. 1979-495, Private Letter Ruling 201633021 (Aug. 12, 2016).
- IRC Section 2514(e); Regs. Section 26.2652-1(a)(5).
- See Edwin Morrow, supra note 5, Appendix, 50-state chart on state protections for such powers and
- See Uniform Trust Code Section 103(11) for how a power of withdrawal that causes self-settled trust/creditor access is defined. See IRC Section 2514 for presently exercisable general powers and lapses; IRC Section 2041 for testamentary general powers; and PLR 9309023 (Dec. 3, 1992) for an example of a presently exercisable power converting to a testamentary power without causing a taxable lapse/release.
- IRC Section 661(a)(2).
- For more extensive discussion and comparison of this method to using IRC 678, see supra note 10, Part nn.
To download PDF, click here.