wealthschemaresourcesarticlesPass-through tax modeling — QBI, reasonable comp, and the K-1 cascade
Article

Pass-through tax modeling — QBI, reasonable comp, and the K-1 cascade

Section 199A is the most-tested provision in the TCJA. The engines that handle it are still mostly handling it wrong — and the 2025 sunset is going to compound the bugs, not fix them.

WealthSchema StaffTax modelingMay 8, 20265 min read

The Section 199A QBI deduction was the single most complex tax provision introduced by the TCJA, and the most error-prone in production wealth-tech engines we audit. The deduction's two limitation regimes (W-2 wage / UBIA limitation; SSTB phase-out), the reasonable-compensation negotiation upstream of the QBI calculation, and the cascade of K-1 information through to the 1040 produce a chain of dependencies that engines routinely break. Nine in ten K-1-handling engines we audit have at least one material QBI bug.

The 2025 sunset, which schedules QBI to expire entirely at the end of 2025 (with potential legislative extension), adds a forward-looking complication. Engines have to model both pre-sunset and post-sunset QBI behavior to produce honest projections.

This article is the working note for engineering teams whose product touches K-1 income and the QBI deduction. Related: building a small-business-owner financial platform, which sits one layer up and covers the W-2-vs-K-1 reasonable-comp negotiation that drives the QBI inputs.

What the QBI deduction is supposed to do

Section 199A allows a deduction of up to 20% of "qualified business income" from pass-through entities (S-corps, partnerships, sole proprietorships, certain trusts) against taxable income. The deduction equalizes (approximately) the federal tax treatment of pass-through income with the C-corp rate reduction the TCJA provided.

The deduction is computed at the individual level (1040), not the entity level (1120-S, 1065). The K-1 reports the entity-level inputs; the individual's calculation flows them through to the 1040.

Formula
Section 199A flow
qbi_deduction = min( 20% × QBI, greater(50% × W2_wages, 25% × W2_wages + 2.5% × UBIA), 20% × (taxable_income − net_capital_gains) ) — adjusted for SSTB phase-out and aggregation elections
QBI
= Qualified business income from the pass-through, not including capital gains, dividends, interest
W2_wages
= W-2 wages paid by the entity (or allocable to the partner). The 'reasonable comp' calculation lives upstream.
UBIA
= Unadjusted basis immediately after acquisition — the original cost of qualified property
SSTB
= Specified Service Trade or Business — health, law, accounting, etc. — phase-out at high income
Each of the inputs is itself a calculation. QBI is not just net income — it excludes guaranteed payments, certain investment items, and entity-level distributions. W-2 wages depend on the reasonable-comp negotiation. UBIA depends on asset history. SSTB classification is fact-specific and disputed for many businesses.

The reasonable compensation negotiation

For an S-corporation owner-employee, the entity must pay "reasonable compensation" before the residual is treated as QBI. The IRS and case law have developed criteria for what's reasonable — comparable industry pay, hours worked, role, geographic factors. A profitable S-corp owner can reduce their reasonable compensation (and increase their QBI) by being aggressive about classification, but the IRS audits this regularly and reclassifies.

The reasonable-compensation question affects the QBI calculation in two ways:

  1. Lower comp = higher QBI. Reduced W-2 reduces the 50%-of-wages limit but increases the QBI base. Engines have to evaluate the net effect.

  2. Below-50%-of-wages QBI is capped. If reasonable comp is too low, the W-2 wage limitation binds and reduces the QBI deduction. The optimum reasonable-comp level depends on the entity's specific numbers — there's no single answer.

The W-2 wage and UBIA limitation

Above the threshold ($383,900 MFJ in 2025; $191,950 single), the QBI deduction is limited by a wage-and-UBIA test:

  • Greater of 50% of W-2 wages OR 25% of W-2 wages + 2.5% of UBIA.

The UBIA path is critical for capital-intensive businesses (real-estate partnerships, manufacturing) with low W-2 wages relative to QBI. A real-estate partnership with $1M of QBI, $40K of W-2 wages, and $5M of UBIA has:

  • 50% × W-2 = $20K (binds — caps deduction at $20K)
  • 25% × W-2 + 2.5% × UBIA = $10K + $125K = $135K (looser cap)

The greater is $135K, so the deduction is min(20% × $1M, $135K) = $135K. Without the UBIA path, the deduction would be capped at $20K — a difference of $115K of deduction (and $46K+ of federal tax at top rates).

Engines that don't model UBIA correctly miss the deduction for capital-intensive pass-throughs. This is a recurring bug class for real-estate partnership investors.

The SSTB phase-out

Specified Service Trades or Businesses face an income-based phase-out of the QBI deduction. The 2025 phase-out range:

 Filing statusPhase-out startPhase-out end (deduction = $0)
MFJ$383,900$483,900
Single$191,950$241,950

Within the phase-out range, the deduction is reduced linearly. Above the upper threshold, an SSTB has zero QBI deduction. Below the lower threshold, the SSTB qualifies normally.

SSTB classification is fact-specific. Health, law, accounting, actuarial services, performing arts, consulting, athletics, financial services, and brokerage services are explicit SSTBs. Engineering, architecture, real estate brokerage, and insurance agency work are not SSTBs (post-2018 clarification). The borderline cases produce most of the audits.

The K-1 cascade

A complete K-1 has 30+ boxes, many of which flow into specific 1040 schedules. The QBI-related boxes are 17 (S-corp K-1) or 20 (partnership K-1). Engines that handle K-1s have to:

K-1 cascade requirements

  • Parse all 30+ boxes per K-1, including Section 199A information codes (V, W, X, etc.)
  • Aggregate K-1s from multiple entities (taxpayer can elect aggregation if entities meet the SAG test)
  • Distinguish QBI-eligible entities from SSTB entities at high income
  • Track UBIA cumulative across years (acquired property + new acquisitions − dispositions)
  • Flow Section 1231 gains/losses to Schedule D, with character determined per disposition
  • Handle entity-level credits passed through (foreign tax credit, R&D credit, low-income housing credit)
  • Handle entity-level basis tracking — partner's outside basis vs S-corp shareholder's stock basis
  • Surface the at-risk and passive-activity limitations that may suspend losses

The basis tracking is the most-overlooked piece. A partner's outside basis affects the deductibility of losses, the recognition of distributions, and the eventual gain on sale of the partnership interest. Engines that don't track outside basis silently produce wrong tax computations for any partnership investor with significant losses or distributions.

What about the 2025 sunset

QBI is scheduled to expire at the end of 2025 along with other TCJA individual provisions. Without legislative extension, pass-through income would no longer get the 20% deduction starting in 2026. The implications:

  • Engines projecting future tax liability must support both scenarios — sunset occurs vs. extension occurs — and surface the sensitivity.
  • Multi-year tax planning (e.g., Roth conversion timing) interacts with QBI status. A pre-sunset year with QBI deduction available is effectively a lower-tax year for pass-through owners; conversion math should reflect this.
  • Estate planning for pass-through owners changes if QBI sunsets — the wealth held in the entity is taxed differently going forward.
  • Reasonable compensation analysis changes — without QBI, the incentive structure for low S-corp comp disappears. Many owners may want to revisit their compensation structure.

Engines that have hard-coded "20%" as the deduction rate need explicit time-series handling for the sunset. The same time-series machinery used for estate exemption sunset (in the parallel article in this series) applies here.

What this means for synthetic test data

A synthetic dataset for QBI / K-1 engines has to include:

  • S-corp owners with full reasonable-comp specification at various ratios to QBI
  • Partnership investors with multiple K-1s and aggregation elections
  • SSTB entities at and around the phase-out thresholds
  • Capital-intensive entities where the UBIA path matters
  • Multi-year basis tracking with contributions, distributions, and allocated losses
  • Entity dispositions with capital-gain bifurcation
  • Foreign-tax-credit and R&D-credit pass-throughs
  • Pre-sunset and post-sunset projection scenarios

The corpus needs to span the full phase-out range, both SSTB and non-SSTB cases, and several entity-type combinations within a household. Test data with only S-corp owners misses partnership-specific issues; test data with only profitable entities misses the loss-suspension cases.

Key takeaways

  • Section 199A QBI is the most error-prone tax provision in production engines. Nine in ten K-1-handling engines have at least one material QBI bug.
  • Reasonable compensation for S-corp owners is the upstream negotiation that drives the QBI calculation. Aggressive comp reductions trigger IRS audits.
  • Above the threshold, the QBI deduction is limited by the greater of 50% of W-2 wages OR 25% of W-2 + 2.5% of UBIA. Engines missing UBIA mis-cap the deduction for capital-intensive entities.
  • SSTB classification phase-out at $383,900 MFJ / $191,950 single. Borderline cases (financial planners, consultants) need fact-specific classification — industry codes alone are insufficient.
  • The K-1 cascade has 30+ boxes flowing into multiple 1040 schedules. Outside basis tracking is the most-overlooked piece and produces silent errors.
  • QBI sunsets end of 2025 absent legislative extension. Engines must support both scenarios in multi-year projections.
  • Test data needs to span the full phase-out range, both entity types, capital-intensive and labor-intensive entities, and pre-/post-sunset scenarios.

Frequently asked questions

How do we handle aggregation elections in test data?+
Aggregation under IRC §199A allows a taxpayer to combine multiple pass-through entities for the W-2 wage and UBIA limitation. Eligible entities must meet the same-trade-or-business test (sharing common ownership, providing same product/service, or sharing significant operational coordination). Test data should include households with multiple entities where aggregation is and is not elected, so engines can be tested on both. Most engines we audit handle aggregation poorly because the cases are rare in any given audit but common enough across portfolio that they matter.
Does QBI interact with state pass-through entity tax (PTET) elections?+
Yes, in non-obvious ways. Many states (NY, NJ, CT, MA, others) introduced PTET regimes post-TCJA that allow pass-throughs to elect entity-level tax with a federal-tax-deductible payment. The PTET payment reduces federal taxable income (and thus QBI). Engines have to model the PTET interaction explicitly: the PTET election may reduce QBI by enough to offset the federal SALT-cap benefit. The optimization is non-trivial and varies by state.
What about the QBI loss carryover?+
Negative QBI in one year can be carried forward to offset positive QBI in a future year. Engines have to track per-entity QBI loss carryovers, which can persist for many years. The interaction with at-risk and passive-activity rules is intricate — an entity-level loss may be suspended at the partner level even before reaching the QBI calculation. Engines that don't track this carryover correctly under-deduct in subsequent profitable years.
How do trusts and estates handle QBI?+
Trusts can have QBI directly (if the trust holds the pass-through interest) and can pass QBI through to beneficiaries via DNI distribution. The trust's calculation uses the trust's own taxable income for the threshold calculations, not the beneficiaries'. Engines serving trust accounts need a separate QBI calculation per trust plus distribution-aware passing-through to beneficiaries. The complexity multiplies for trusts with multiple beneficiaries and tiered structures.