Trust accounting under UPIA — the principal-and-income split that wealth-tech keeps getting wrong

The Uniform Principal and Income Act (UPIA) is the body of state law (in some form, in 47 states) that governs how a trust's receipts are allocated between income (which the income beneficiary receives) and principal (which is held for the remainder beneficiary). For most trusts, this allocation is the most important ongoing administration question — it determines how much the income beneficiary actually receives, what their tax treatment is, and how the trustee reports to the beneficiaries each year.

The allocation rules are non-trivial. They differ by asset class. They have specific exceptions for certain investment types. They interact with the Internal Revenue Code's distinct definitions of "Distributable Net Income" (DNI) and "Trust Accounting Income" (TAI). Modern UPIA versions allow the trustee to convert between income and unitrust models, adding a strategic decision surface.

Wealth-tech platforms — even those serving HNW family offices and trust departments — frequently treat this allocation as a black-box trustee responsibility and skip the modeling. The result: platforms that claim to track trusts can't actually answer "how much income is owed to the income beneficiary this year?" without a manual override from the trust accountant.

This article is the working note on what the data model needs to look like for proper UPIA-aware trust administration, and what the platform ships without when the model is shallow.

The principal-vs-income allocation, by asset class

UPIA's basic rule: receipts are income unless specifically allocated to principal, and disbursements are charged to income unless specifically allocated to principal. The "specifically" carve-outs are extensive.

Interest income: allocated to income. Bond coupons, savings-account interest, money-market interest — all income.

Dividend income: allocated to income. Ordinary dividends from stock holdings are income.

Capital gains and losses: allocated to principal. When a trust sells a security with a $50K gain, that gain is principal — it stays in the trust for the remainder beneficiary. The income beneficiary receives nothing from the gain directly, even though the trust's economic position improved by $50K.

Mutual fund capital-gain distributions: this is where the rules get nuanced. The distribution looks like income to a casual reader (it's a regular distribution), but the underlying economic event is a capital gain. UPIA Section 401 generally allocates these to principal, with exceptions.

Property received as principal vs. income: property received as a gift or inheritance is principal. Property received as a trust contribution is principal. Property received as a stock dividend (vs. cash dividend) may be principal depending on the corporate-action mechanics.

Liquidating distributions: often allocated to principal because they represent return of capital rather than a current-period economic gain.

Real estate income: rental income is income; sale proceeds are principal. Depreciation can be allocated either way under modern UPIA (a "delegation" decision in the trust document or under default UPIA rules).

Royalty income: typically split — a portion to income, a portion to principal — under §410 of UPIA. The allocation reflects the depleting nature of the underlying asset.

The asset-class detail matters because the platform must track, for each receipt, which category it falls into and apply the correct allocation rule.

interface TrustReceipt {
  receipt_id: string;
  receipt_date: ISODate;
  amount: number;
  source_asset_id: string;
  
  // Receipt classification
  receipt_type: 
    | 'interest'           // → income
    | 'dividend_ordinary'  // → income
    | 'dividend_qualified' // → income (UPIA basis; tax bracket may differ)
    | 'capital_gain_distribution' // → principal (with exceptions)
    | 'return_of_capital'  // → principal
    | 'rental_income'      // → income (gross)
    | 'royalty'            // → split
    | 'liquidating_distribution' // → principal
    | 'in-kind_distribution' // → principal usually
    | 'sale_proceeds'      // → principal
    | 'other';
  
  // Allocation
  allocated_to_income: number;
  allocated_to_principal: number;
  allocation_rationale: string;
  
  // Trustee discretion
  trustee_adjustment_used: boolean; // §104 power to adjust
  adjustment_documentation: string;
}

The 65-day rule and DNI flow-through

A separate-but-related concept is the trust's Distributable Net Income (DNI) as defined in IRC §661 and §662. DNI is the conduit through which the trust passes income out to the beneficiary for federal income tax purposes. The amount of DNI a trust distributes in a tax year is deductible at the trust level and includable at the beneficiary level.

The 65-day rule (IRC §663(b)) lets the trustee elect to treat distributions made within the first 65 days of a tax year as if they were made in the prior tax year. This is a powerful planning tool — by mid-March, the trustee can see the trust's DNI for the prior year and make discretionary distributions to push DNI out to beneficiaries (who may have lower marginal rates than the trust's compressed brackets).

The compressed bracket structure for trusts hits the top federal rate at very low income levels — roughly $14,500 of 2024 taxable income. By contrast, an individual hits the top bracket above $625,000 (single). For a trust with $100K of DNI, distributing it to the beneficiary instead of retaining it can save the difference between the trust's 37% rate and the beneficiary's potentially-lower rate.

A wealth-tech platform that models the trust must surface this planning surface. By early March, the platform should show: prior-year DNI, beneficiary's projected marginal rate, recommended distribution amount under the 65-day election to optimize total family tax burden.

The §104 power to adjust

Modern UPIA versions include §104, which empowers the trustee to adjust between income and principal "to the extent the trustee considers necessary if the trustee invests and manages trust assets as a prudent investor." The §104 power is the bridge between modern total-return investment philosophy and the traditional UPIA income/principal categories.

A common application: a trust holds a mix of growth equities (low dividend yield) and bonds (high coupon). Without §104 adjustment, the income beneficiary receives only the bond coupons — the equity returns are entirely principal. The income beneficiary's "income" is artificially low relative to the trust's total return, conflicting with the modern total-return investment management.

Under §104, the trustee can shift a portion of the equity total return to income, equalizing the beneficiary's economic position. The adjustment is bounded by reasonableness and must be documented.

A useful platform surfaces the §104 decision: shows the income beneficiary's mechanically-computed income, computes the equivalent total-return-based income, and presents the trustee with the recommended adjustment range (and rationale). Without the platform's surfacing, trustees often default to mechanical allocation that under-serves the income beneficiary in equity-heavy portfolios.

The unitrust election alternative

Some states allow the trust to convert from traditional income/principal accounting to a unitrust model — typically defined as 3-5% of the trust's annual fair market value paid to the income beneficiary, regardless of the underlying receipts.

The unitrust election simplifies administration substantially. No more per-receipt allocation; the income beneficiary receives a fixed percentage. It also smooths beneficiary income across years (high-yield year vs. low-yield year produce similar payments under unitrust).

The election is generally one-way (can convert from traditional to unitrust, not back) and requires either trust-document authority or court approval depending on state. The platform should surface the unitrust analysis for trusts where it's beneficial and the donor / trustee should be aware of the option.

What the test corpus needs

A test corpus for UPIA-aware trust administration needs:

• Trusts with high-yield bond-heavy portfolios (mechanical income flow)
• Trusts with low-yield equity-heavy portfolios (§104 adjustment surface)
• Trusts with mutual-fund capital-gain distributions (allocation-rule edge case)
• Trusts with rental real estate (depreciation allocation surface)
• Trusts with concentrated stock positions and corporate actions (allocation complexity)
• Trusts with multiple income beneficiaries (per-beneficiary allocation)
• Trusts approaching the 65-day window (DNI distribution decision)
• Trusts considering unitrust conversion (election analysis)
• Multi-state trust scenarios (different state UPIA versions)

Each scenario produces a different right answer from the trustee. A platform that doesn't exercise the variations produces shallow advice limited to the modal trust.

State variation

Although UPIA is "uniform," its state-by-state adoption has produced variations. Some states have enacted modified versions; some have ancient versions; some have not adopted UPIA and use older common-law rules.

For a multi-state trust portfolio (a family office holding trusts under multiple state laws), the platform must apply the correct state's UPIA rules per trust. The trust's situs (governing law) is a structured field, and the platform's allocation engine must dispatch to the right rule set.

The synthetic-corpus value

Realistic trust-accounting data is notoriously hard to obtain — the data lives in trust departments, family-office systems, and accountant's spreadsheets. Production data feeds rarely include the receipt-level classification needed for UPIA modeling. A purpose-built synthetic corpus with documented per-receipt classifications is the practical route to testing.

The corpus also templates the production data capture. If the corpus has receipt-level type classification, the platform's data ingestion must support it. Without the field, the platform's UPIA logic falls back to "all receipts are income" or "all receipts are principal" — both of which are wrong.

UPIA is the prototypical case of a wealth-tech domain where the surface looks simple ("track trust receipts") but the underlying logic is structurally rich. Platforms that handle UPIA correctly can credibly serve the family-office and trust-department market; platforms that don't can't, and default to "the trust's accountant handles that" — which is true but signals to the buyer that the platform is incomplete.