Worked research note
One household, four accounts
A worked asset-location case separates the modeled placement benefit from the path required to reach it.
By ArthaPilot Research
Published 2026-10-11
Research question
A fictional household's pinned 2026 snapshot compares current tax drag, constrained placement, transition paths, and income and cost-basis sensitivities.
Scope: One fictional household cannot establish a general placement benefit or asset ordering.
The account map defines the opportunity
The sample holds $1,274,100.00 across four accounts, including $254,100.00 in taxable assets. The account balances come from the generated household fixture.
The Household profile has a separate target allocation. This placement calculation instead holds the current mix fixed at AGG 13.35%, BND 9.34%, SCHH 3.14%, VNQ 1.60%, VTI 64.36%, VXUS 8.21%; this isolates account location from a rebalance to target.
Exhibit 01
Content-pinned result
Current balances across four accounts
Account balances and asset mix used for placement.
How much does each account hold before placement changes?
Construction. Read each account's marketValue without reclassifying or combining account types. Check that the four balances sum to the pinned $1,274,100 household market value. Plot the balances on a common dollar axis; no tax-drag result enters this first exhibit.
| Category | Market value |
|---|---|
| Joint taxable brokerage | $254,100 |
| Traditional IRA (Jordan) | $450,100 |
| Roth IRA (Riley) | $119,950 |
| 401(k) (Riley) | $449,950 |
Source and construction: ArthaPilot generated sample household, input fingerprint e4880095…, anchored to May 1, 2026.
How this figure was calculated
Inputs
- May 1, 2026 account balances from the canonical request payload.
- Four account containers: taxable, traditional IRA, Roth IRA, and workplace 401(k).
- Placement keeps the actual holdings mix fixed, rather than the separate household target.
Calculation
- Read each account's marketValue without reclassifying or combining account types.
- Check that the four balances sum to the pinned $1,274,100 household market value.
- Plot the balances on a common dollar axis; no tax-drag result enters this first exhibit.
What this step cannot establish
These balances describe one fixture. They do not estimate placement value or represent other households.
Exact values behind the figure
| Series | Account | Market value (USD) |
|---|---|---|
| Market value | Joint taxable brokerage | $254,100 |
| Market value | Traditional IRA (Jordan) | $450,100 |
| Market value | Roth IRA (Riley) | $119,950 |
| Market value | 401(k) (Riley) | $449,950 |
This case covers only the stated accounts, holdings mix, tax context, and asset-efficiency classes.
The annualized drag estimate does not generalize to households with other accounts, brackets, assets, or policies.
Placement changes the model by $1,045.58 per year
On one marginal-rate bundle, modeled annual drag is $12,487.29 for the current placement, $12,031.20 for the naive placement, and $11,441.71 for the lowest-drag feasible placement.
The $1,045.58 annual difference is a placement estimate. It is not historical excess return, a forecast, or an after-tax account balance.
Exhibit 02
Content-pinned result
Current, naive, and lowest modeled annual drag
Current household-wide weights stay fixed under the stated tax schedules and asset-efficiency assumptions.
How does modeled annual drag change with holdings weights and tax assumptions fixed?
Construction. Read the three engine-produced drag estimates on one common dollar basis. Subtract suggested drag from current drag: $12,487.29 − $11,441.71. Keep engine precision in chart values. Round display copy to cents.
| Category | Annual tax drag |
|---|---|
| Current | $12,487 |
| Naive | $12,031 |
| Lowest-drag feasible | $11,442 |
Source and construction: ArthaPilot Asset-Location Auto-Optimizer response projection, placement model realization_aware_v2 and tax kernel unified_v1.
How this figure was calculated
Inputs
- Current, naive, and suggested taxDragEstimateDollars from the canonical response payload.
- The same current household-wide weights, unified_v1 tax kernel, and realization_aware_v2 placement model for every state.
- 2026 federal marginal schedule and the 2025 California resident schedule; nominal dollars at the May 1, 2026 price anchor.
Calculation
- Read the three engine-produced drag estimates on one common dollar basis.
- Subtract suggested drag from current drag: $12,487.29 − $11,441.71.
- Keep engine precision in chart values. Round display copy to cents.
What this step cannot establish
The $1,045.58 difference is an annualized model delta, not historical excess return, forecast wealth, or a result for households with other inputs.
Exact values behind the figure
| Series | Placement state | Modeled annual tax drag (USD) |
|---|---|---|
| Annual tax drag | Current | $12,487 |
| Annual tax drag | Naive | $12,031 |
| Annual tax drag | Lowest-drag feasible | $11,442 |
The solver used continuous_lp. Its suggested placement keeps the fixed asset mix and places each ticker within available account capacity.
Per-asset yields and appreciation come from disclosed efficiency classes rather than observed histories. The output is a model comparison, not an empirical performance record.
Transition paths and sensitivity change the interpretation
When every taxable position is frozen, exchanges inside protected accounts reduce modeled drag by $206.32 per year. That path captures 19.7% of the ideal difference without a taxable realization event.
The immediate path reaches the ideal placement in the fixture and reports $0.00 of current realization tax. It also reports $21,440.00 of disallowed relocation loss, which receives no current-year tax benefit.
Reruns changed income and taxable lot basis. Across the two California income cases, the ideal modeled difference ranges from $0.00 to $1,856.73 per year. At half the taxable lot basis, the same ideal difference has a 14.8-year immediate-sale payback, beyond the engine's 10-year proposal horizon, so the recommendation is withdrawn.
Exhibit 03
Content-pinned result
Ideal placement versus two transition paths
Transition estimates for the baseline and sensitivity reruns for income and taxable lot basis.
How much of the ideal annual drag difference can each transition path reach?
Construction. Use each scenario's engine-produced annual saving versus current placement. Compare $206.32 for protected-account exchanges with $1,045.58 for the ideal and immediate paths. Re-run the placement engine for each stated sensitivity input; compare the annual ideal difference, transition cost, payback gate, and disallowed loss. Show realization tax and disallowed loss separately from annual savings because their time bases differ.
| Category | Annual difference |
|---|---|
| Ideal benchmark | $1,046 |
| Protected exchanges | $206 |
| Immediate transition | $1,046 |
Source and construction: ArthaPilot sample-household transition scenarios at the May 1, 2026 valuation and sale anchor.
How this figure was calculated
Inputs
- savingVsCurrentDollars for the ideal, exchanges_only, and immediate_sale scenarios.
- The exchanges-only rule that freezes every taxable position.
- The immediate path's $0.00 realization tax and separately reported $21,440.00 disallowed loss.
- California income cases use $0 for both bases, or $500,000.00 federal ordinary income and $472,222.22 California taxable ordinary income. Other cases use $180,000 federal-only income or 50% and 125% taxable lot basis. Schedule years stay fixed.
Calculation
- Use each scenario's engine-produced annual saving versus current placement.
- Compare $206.32 for protected-account exchanges with $1,045.58 for the ideal and immediate paths.
- Re-run the placement engine for each stated sensitivity input; compare the annual ideal difference, transition cost, payback gate, and disallowed loss.
- Show realization tax and disallowed loss separately from annual savings because their time bases differ.
What this step cannot establish
Equal bars do not make the immediate path costless. The selected sensitivities are model inputs, not probabilities or estimates for other households.
Exact values behind the figure
| Series | Transition path | Annual drag difference versus current (USD) |
|---|---|---|
| Annual difference | Ideal benchmark | $1,046 |
| Annual difference | Protected exchanges | $206 |
| Annual difference | Immediate transition | $1,046 |
The free path and ideal benchmark answer different questions: what can change without a taxable sale, and what the model estimates after full placement.
A zero tax line does not make the immediate path costless. The response preserves the disallowed-loss amount so the tax attribute is visible beside the recurring drag estimate.
Method and constraints
Method
For this fictional household with a fixed holdings mix, how does account placement affect annual tax drag? How much can change without taxable sales?
Use the generated sample household and May 1, 2026 holdings snapshot without editing its account balances, actual asset mix, lots, or tax context.
Compare current, naive, and constrained suggested placements with the same marginal-rate bundle and realization_aware_v2 placement model.
Report the ideal benchmark, protected-account exchanges-only path, and immediate transition from the engine result projection generated alongside the sample fixture.
Commit the publication request and response projection as exact JSON bytes and verify both SHA-256 identities in tests.
Determinism anchor: Source input fingerprint e4880095133390d051bf46a8037ab42fb0e0d457eea8d6333e4f85be057d92be; May 1, 2026 snapshot prices; federal schedule year 2026; California schedule year 2025; holding periods; and transition pricing.
Dollar basis: Annual drag and transition amounts are nominal 2026 U.S. dollars at the May 1, 2026 price anchor. Marginal rates use the published 2026 federal schedule and 2025 California schedule; source pages were checked on September 25, 2026.
Verification guard: Request and disclosed response-projection hashes must match the exact bytes generated with the sample household fixture.
Verification guard: Displayed balances, drag estimates, scenario savings, and disallowed loss must match the generated sample source.
Verification guard: An independent ledger must reproduce current and suggested drag using the published tax schedules and stated efficiency assumptions.
Verification guard: Sensitivity results must come from rerunning the engine with changed income or lot-basis assumptions and must disclose payback gating.
Verification guard: Every exhibit must remain linked to the same computed artifact and May 1, 2026 valuation anchor.
Verification guard: Any broader household-grid or historical claim requires a new artifact and cannot be added under this worked-case identity.
Assumptions held fixed
- The placement target holds current aggregate weights fixed: AGG 13.35%, BND 9.34%, SCHH 3.14%, VNQ 1.60%, VTI 64.36%, VXUS 8.21%. The household's separate target allocation is not applied in this comparison.
- The tax kernel is unified_v1 and the placement model is realization_aware_v2.
- NIIT uses a lower-bound MAGI proxy: ordinary income plus modeled net investment income. No MAGI override is supplied for the baseline or sensitivities.
- The federal marginal schedule is for 2026. The California resident marginal schedule is for 2025, the latest schedule shown by the FTB source checked September 25, 2026. The valuation dollar basis is separately anchored to May 1, 2026.
- Per-asset income and appreciation inputs come from efficiency classes, with 50% deferral and realization factors as disclosed by the generated engine response.
- The exchanges-only scenario freezes every taxable position; the immediate path uses the fixture's recorded lots and the generated fixture's `optimized` account lot methods.
- Relocation losses receive no current tax benefit and are disclosed as disallowed loss rather than negative transition tax.
What this study establishes
This fictional household keeps its assets and household-wide mix fixed. Only account placement changes.
The ideal benchmark, protected-account exchanges, and taxable transition show ongoing drag separately from transition effects.
Relocation losses are reported as disallowed rather than harvested. They are a foregone tax attribute, not free transition value.
Limitations
- One fictional household cannot establish a general placement benefit or asset ordering.
- Annualized drag is not a historical backtest and does not model every future contribution, withdrawal, rebalance, or tax-law change.
- Zero realization tax depends on this fixture's loss lots. Gain lots can incur tax.
- Asset-efficiency classes are modeling assumptions rather than forecasts of distributions or appreciation.
- Legal, plan-menu, liquidity, estate, and behavioral considerations can sit outside this calculation.
Sources and prior research
- Optimal asset location and allocation with taxable and tax-deferred investing (2004)
Dammon, Spatt, and Zhang examine allocation and location across taxable and tax-deferred wealth.
- Asset location in tax-deferred and conventional savings accounts (2004)
Shoven and Sialm show how asset tax efficiency and the available investment set affect preferred location.
- IRS topic 409, capital gains and losses
Authoritative background for basis, holding-period character, capital-gain netting, and ordinary treatment of net short-term gains.
- IRS revenue procedure 2025-32, 2026 inflation-adjusted federal provisions
Primary source for the published 2026 federal tax-rate schedule and inflation-adjusted thresholds.
- California FTB 2025 resident tax-rate schedules
Primary source for the latest published California resident income-tax schedule used by this asset-location artifact.
- IRS topic 559, net investment income tax
Authoritative scope for investment income and gains that can enter NIIT, including tax-exempt municipal interest exclusions.
Evidence and reproducibility
sample-household-asset-location-2026-05-01
One fictional two-member household with $1,274,100 across taxable, traditional IRA, Roth IRA, and 401(k) accounts.
Evidence date: 2026-05-01. Nominal 2026 dollars at the May 1, 2026 valuation and sale anchor.
request=2ec0443fa96d2dd96cb6af227b2fbae817daa72f2e680cbb3ec65c2d82810e62; response=a20c10688c226c72df1d7283156e933b34b5004448b3aefa30e54f8c84b6e1d6
Source: web/src/data/research/planning/sampleHouseholdResearch.generated.ts. Versions: sourceFingerprint=e4880095133390d051bf46a8037ab42fb0e0d457eea8d6333e4f85be057d92be, placementModel=realization_aware_v2, taxKernel=unified_v1.
- Per-asset yield and appreciation assumptions come from efficiency classes rather than observed return histories.
- Zero realization tax depends on this fixture's loss lots.
- This household case does not support a population claim.
Model the same trade-off for a household
The linked tools keep household context, dates, tax assumptions, and result limitations attached to the calculation.
- Model a household placement Run the same placement and transition analysis with a household's accounts, policies, holdings, and valuation anchor.
- Review household inputs Inspect the accounts, holdings snapshot, lots, and tax context that determine the available placement set.