A Constitutional Rebuttal to IRS AM 2023-006: Defending

Trust Corpus Distributions Under the 16th Amendment

By Scott McGrath, MFP©, CWM©, CTEP©

Introduction

The constitutional power of taxation is not boundless, even when exercised by the Internal Revenue Service (IRS). Over the years, a complex interplay of constitutional language, judicial precedent, and tax code provisions has carefully delineated what constitutes taxable income. The Sixteenth Amendment serves as the cornerstone of this framework, permitting Congress to tax "incomes, from whatever source derived," while leaving intact important restrictions on the taxation of capital, gifts, and unrealized gains.

In recent years, the boundaries of that authority have been tested by administrative interpretation. One such instance is IRS Chief Counsel Memorandum AM 2023-006, which offers an expansive view of the IRS's ability to impose income tax on distributions made by irrevocable, non-grantor, complex discretionary trusts—even when those distributions consist solely of corpus that was lawfully allocated and accumulated. The IRS argues that if a trust receives gross income—such as capital gains or dividends—those funds eventually must be taxed, even if lawfully allocated to principal and distributed without ever forming part of the trust's Distributable Net Income (DNI).

This position, however, not only stretches statutory interpretation—it contradicts the constitutional limits imposed by the Sixteenth Amendment and misrepresents the legal framework under Subchapter J of the Internal Revenue Code. More importantly, it threatens to unravel over a century of trust law and fiduciary principles, particularly those affirming the trustee's authority over income allocation.

This rebuttal addresses the constitutional, statutory, and regulatory failures embedded in AM 2023-006 and explains how a properly structured irrevocable, non-grantor, complex discretionary trust—not only complies with federal tax law but is also expressly supported by longstanding legal precedent.

I. Constitutional Foundations: Defining Income

The 16th Amendment, ratified in 1913, reads:

“The Congress shall have power to lay and collect taxes on incomes, from whatever source derived, without apportionment among the several States, and without regard to any census or enumeration.”

This provision was meant to overturn Pollock v. Farmers’ Loan & Trust Co. (1895), which had severely restricted Congress’s ability to tax income derived from property without apportionment. While the 16th Amendment expanded Congress's taxing power, it retained a critical boundary: the power to tax "income"—not capital, not corpus, not unrealized appreciation.

In Eisner v. Macomber , 252 U.S. 189 (1920), the Supreme Court emphasized this boundary. The Court held that a stock dividend was not income, because it involved no realized gain. The ruling clarified:

“Income may be defined as the gain derived from capital, from labor, or from both combined... provided it be understood to include profit gained through sale or conversion of capital assets.”

Accordingly, without realization of gain, there is no constitutional income. Mere changes in form, or transfers of property within the same ownership structure (e.g., trust corpus), do not trigger taxable events.

II. Subchapter J and Fiduciary Law: Corpus Is Not Income

This statutory framework reflects the delicate balance between federal tax rules and state fiduciary obligations. Under the Uniform Principal and Income Act (UPIA), trustees are granted specific authority—often a fiduciary duty—to determine whether particular receipts should be allocated to income or principal. This discretion is grounded in the need to fairly balance the interests of income beneficiaries and remainder beneficiaries, a task that cannot be dictated by blanket federal assumptions.

Treasury Regulation §1.643(a)-3 validates this state-based approach. It not only allows but anticipates that capital gains will typically be allocated to corpus unless otherwise directed by the trust instrument or required by local law. In practice, most trustees, particularly those of non-grantor complex discretionary trusts, will follow this model to preserve principal for long-term planning and distribution flexibility.

The distinction between income and corpus in this context is not merely technical—it is fundamental to how trusts are taxed under Subchapter J. IRC §661 permits deductions to the trust for distributions of DNI, and IRC §662 includes such DNI in the beneficiary’s gross income. However, when there is no DNI—as in the case of capital gains allocated to corpus and retained—there is no deduction and no passthrough. Consequently, when corpus is eventually distributed, it does not carry out income and remains non-taxable.

The IRS’s failure in AM 2023-006 to acknowledge or apply this layered fiduciary-tax framework leads to a mischaracterization of the income mechanics within a non-grantor trust. Rather than recognizing the trust as a distinct taxable entity operating under a bifurcated income/corpus structure, the memo appears to treat the trust as a transparent conduit where all gross receipts are eventually taxed, regardless of fiduciary allocation. This fundamentally misrepresents both the statute and the nature of complex trust taxation.

By ignoring the legal mechanisms that separate income from corpus—and the statutory permissions that validate this separation—AM 2023-006 undermines not only tax policy but also the trust law framework that has guided estate and financial planning for generations.

III. The Extraordinary Dividends Exception

The characterization of dividends as "extraordinary" is not a novel concept, nor is it merely a technical footnote in fiduciary accounting. Extraordinary dividends have long been recognized as distributions that exceed the usual or expected yield from an investment, often resulting from unusual corporate events such as the liquidation of assets, sale of a subsidiary, or a special one-time distribution of retained earnings. Courts and fiduciaries alike have consistently viewed such distributions as capital in nature, rather than income, due to their non-recurring and substantive impact on the underlying investment.

This distinction has been affirmed in numerous legal and academic sources, including trust law treatises and case law interpreting the duties of trustees. The key element in determining whether a dividend is extraordinary often hinges on both its magnitude and its source—whether it represents accumulated profits or a reallocation of the corporation's capital structure. Trustees are expected to assess these characteristics in good faith and in accordance with their fiduciary duty to maintain equity between current income beneficiaries and future remainder beneficiaries.

Furthermore, the Internal Revenue Manual and prior training materials provided to IRS agents have historically included references to these principles, acknowledging that the characterization of receipts as principal or income is inherently governed by trust law and not subject to arbitrary federal reclassification. Thus, when the IRS itself recognizes the validity of treating extraordinary dividends as corpus—as it does in AM 2023-006—it is affirming a deeply entrenched legal standard.

In the context of a non-grantor, complex, discretionary trust, the allocation of extraordinary dividends to corpus serves both a legal and practical function. It enables the trustee to preserve the trust’s capital for future distributions or reinvestment while respecting the income expectations of beneficiaries. More importantly, it ensures that the trust’s accounting practices remain in harmony with state law and Subchapter J, thereby avoiding unintended tax consequences.

IV. Realization Requirement Reinforced by Case Law

This interpretation has been reinforced by the courts in several landmark cases. In Commissioner v. Glenshaw Glass Co. , 348 U.S. 426 (1955), the U.S. Supreme Court clarified that income, to be taxable, must constitute "undeniable accessions to wealth, clearly realized, and over which the taxpayers have complete dominion." This case established a three-part test that remains the standard for evaluating what qualifies as taxable income. Corpus distributions that do not involve a realization event—and which transfer previously classified capital—fail all three prongs of this test.

Similarly, in Helvering v. Bruun , 309 U.S. 461 (1940), the Court addressed whether property improvements made by a lessee created taxable income to the lessor. The Court acknowledged the concept of economic benefit but still emphasized that taxation required a separation from capital and an element of realization. This principle supports the conclusion that capital appreciation or accumulation alone, without conversion or distribution as income, does not trigger tax liability.

In Freuler v. Helvering , 291 U.S. 35 (1934), the Supreme Court further clarified the integrity of fiduciary accounting by holding that a trustee’s good-faith classification of trust receipts under state law could not be arbitrarily recharacterized by the IRS for tax purposes. This case affirms that the IRS must respect lawful allocations made by fiduciaries in accordance with the trust instrument and applicable fiduciary law.

Finally, in Commissioner v. Estate of Bosch , 387 U.S. 456 (1967), the Court reaffirmed that state law governs the classification of property interests and fiduciary actions in trust and estate contexts. The IRS cannot substitute its own judgment for that of a state court or a fiduciary acting under state authority. This deference to state law extends to the classification of receipts as income or principal, further undermining the IRS’s reasoning in AM 2023-006.

Together, these cases form a solid constitutional and jurisprudential foundation for rejecting any attempt by the IRS to impose tax on corpus distributions that do not carry out DNI and are not grounded in a realization event.

V. Mischaracterization of Trust Type: "Self-Styled" vs. Non-Grantor

The IRS memo further undermines its own authority by referring to the subject trust as a “self-styled irrevocable, non-grantor, complex trust.” This terminology is problematic. The phrase "self-styled" implies the possibility of a self-settled trust. In fiduciary and tax law, a self-settled trust—where the grantor retains a beneficial interest—is treated as a grantor trust under IRC §§671–679.

Grantor trusts are fundamentally different from non-grantor trusts. Income, deductions, and credits in a grantor trust flow through to the grantor. In contrast, non-grantor trusts are separate taxable entities. If the IRS memo in fact evaluated a self-settled arrangement, it cannot simultaneously present its findings as applicable to true non-grantor trusts.

This ambiguity undercuts the entire premise of the memo. If AM 2023-006 conflates two mutually exclusive trust categories, then its conclusions cannot be relied upon in evaluating the taxation of third- party irrevocable, non-grantor, complex discretionary trusts.

VI. Practical Implications and Structural Integrity

In a now-deleted PowerPoint presentation once hosted on its own website, the IRS clearly acknowledged that it does not determine Trust Accounting Income (TAI). This admission is critically important, as it affirms the principle that the classification of trust income versus corpus falls within the purview of the trustee, governed by the trust instrument and applicable state fiduciary law—not by IRS administrative fiat. They deleted this powerpoint about a month after coming out with this memorandum. While the presentation is no longer publicly available, its prior existence is documented, and the assertion it made remains consistent with long-standing judicial precedent and Treasury regulations that the IRS does not determine trust “accounting income” (TAI).

By admitting that it does not—and indeed cannot—define or override fiduciary accounting decisions regarding TAI, the IRS implicitly supports the authority of trustees to allocate receipts such as capital gains or extraordinary dividends to principal. This reinforces the argument that allocations made in compliance with fiduciary standards and local law are legally controlling for federal tax purposes. It also aligns with the holdings in cases such as Freuler v. Helvering and Commissioner v. Estate of Bosch , which recognized the binding nature of fiduciary determinations in trust administration.

The trust structure described in this paper—one that receives passive income, allocates extraordinary receipts to corpus, never creates DNI, and distributes only principal—complies fully with both statutory and constitutional mandates. It does not issue K-1s. It does not deduct distributions. It does not pass income through to beneficiaries.

The purpose of this design is not tax avoidance, but tax compliance. It is constructed to prevent the occurrence of taxable events by design. Such structures reflect legitimate tax planning and fiduciary responsibility—not loopholes. Any attempt by the IRS to recharacterize corpus as income undermines fiduciary law, violates Subchapter J, and contravenes constitutional tax limits.

VII. Conclusion

IRS AM 2023-006 suffers from multiple, compounding errors:

1. It contradicts constitutional principles by collapsing the legal distinction between capital and income.

2. It ignores the required element of realization before taxation is permissible.

3. It acknowledges—yet discounts—the extraordinary dividends exception, which supports valid trust corpus allocations.

4. It confuses trust classifications by introducing the term “self-styled,” suggesting the memo may apply to self-settled (grantor) trusts instead of the non-grantor trusts it purports to address.

Properly structured irrevocable, non-grantor, complex discretionary trusts that allocate capital receipts to corpus, refrain from creating DNI, and distribute only corpus remain outside the scope of income taxation under both statutory law and constitutional doctrine. AM 2023-006 misapplies law and misrepresents facts. For these reasons, it should not be relied upon in evaluating the tax treatment of trust corpus distributions.