Insight impact ending after tax estate delta

Estimated change in ending after-tax estate from an insight action.

Field: endingAfterTaxEstateDelta Unit: US dollars Basis: nominal dollars Computed by: Engine Status: Complete chain

Appears in: insights (InsightCardView "Ending estate delta" after Preview impact; rough "≈ ... estate delta" when no qualitative text)

Calculation records

• Insight impact: ending after-tax estate and lifetime tax deltas Formula record, justified by derivation

  Purpose: What an insight's modeled action does to the estate and to lifetime taxes, as candidate minus baseline.

  Statement: decisions/evaluateCandidate.ts#evaluateCandidate prices the candidate on its own full year-by-year projection and subtracts the shared baseline in the same direction for both fields: endingAfterTaxEstateDelta = candidateSummary.endingAfterTaxEstate - baselineSummary.endingAfterTaxEstate, and lifetimeTaxDelta = candidateSummary.lifetimeTaxesAndPenalties - baselineSummary.lifetimeTaxesAndPenalties. The estate field's comment states that order outright and the lifetime-tax field defines a negative value as savings, which is the same convention: an estate improvement is positive and a tax saving is negative. Units: nominal dollars, not today's: the estate delta is in dollars of the plan's last year (both runs end in that year), and the lifetime tax delta sums each year's own dollars. Rounding: none.

  Formula: estateDelta = E_candidate - E_baseline; taxDelta = T_candidate - T_baseline

  • E: Ending after-tax estate of a run's summary (unit: usd) (domain: finite)
  • T: Lifetime taxes and penalties of a run's summary (unit: usd) (domain: nonnegative)

  Timing: once per evaluated candidate, against the context's single shared baseline

  Rounding: none

  Limits:

  • Beyond the worksheet's inputs the evidence plan fixes two real one-year projections that land on the worksheet's four summary figures exactly: a 1963-born single filer (over 59.5 so no early-distribution rule, under 65 so no Medicare premium) with one $1,800,000 traditional account, a flat 25% tax double, heirTaxRatePct 50, zero return and zero inflation, and base spending of $600,000 in the baseline against $555,000 in the candidate scenario. At a flat rate r the withdrawal funding spending S is S / (1 - r) and the tax is r x that, and the estate is the remaining balance net of the heir rate
  • The flat calculator is a test double, never the shipped stack: it exists so the test can name exact dollars without recomputing federal law, and it must not be read as a tax result
  • Both deltas are differences of SUMMARIES, so they inherit every convention of summarizeProjection: the heir-tax haircut, charity carve-outs, and the fact that property, debts and ladder face ride through net worth without per-account estate rows
  • The irmaa-tier-edge Insight writes its annual premium cliff into endingAfterTaxEstateDelta as an avoidance signal rather than an estate change; that exception belongs to that Insight’s own record and the evidence here does not exercise it

  Provenance: derived by codex, implemented by claude.

  Independent review: grok

  Gates (6 of 6 pass): fixture registers a test: pass; pins resolve: pass; families exist: pass; worksheet exists: pass; mutation receipt exists: pass; provenance independent: pass

  Worksheet: insight-impact-estate-and-lifetime-tax-deltas.md Mutation receipt: insight-impact-estate-and-lifetime-tax-deltas.mutation.md

  Evidence tests: evaluateCandidate.evidence.test.ts

  Implemented in: evaluateCandidate.ts; types.ts; insightsAdapter.ts

  Is the number of: insight-impact-ending-after-tax-estate-delta, insight-impact-lifetime-tax-delta

Records that feed this number

• IRMAA tier-edge household Medicare premium cliff Model record, justified by derivation

  Purpose: The extra annual Medicare premium a MAGI just over an IRMAA threshold triggers two years later, versus MAGI one dollar under it.

  Statement: For a selected IRMAA edge with n Medicare enrollees in the charge year (lookback MAGI year + 2): cliff = n × ((Part B + Part D surcharge)_just-above − (Part B + Part D surcharge)_one-dollar-below), using the healthcare-premium scale for the charge year. Units: nominal dollars per year. Rounding: none; the card floors a negative difference at 0.

  Formula: cliff = n (P_above − P_below); P = partBAnnual + partDSurchargeAnnual

  • n: Medicare enrollees (alive, age 65+) in the charge year (unit: people) (domain: integer n >= 1)
  • P_above: Per-person annual Part B + Part D premium just above the MAGI threshold, already resolved (unit: usd/year) (domain: finite)
  • P_below: Per-person annual Part B + Part D premium one dollar below the threshold, already resolved (unit: usd/year) (domain: finite)

  Timing: charge year = MAGI year + 2 (the statutory lookback)

  Rounding: a negative (P_above − P_below) is floored at 0 before multiplying by n; the card publishes whole-dollar strings (formatWholeUsd, Math.round) and the evidence asserts the published figure

  Limits:

  • A rough avoidance signal, not a claim that reducing current-year income changes a premium already charged
  • The worksheet supplies the already-resolved per-person premiums; the evidence passes those through rather than recomputing medicareAnnualPremiumPerPerson
  • The Insight only fires when MAGI is at most $5,000 over the next threshold; that proximity window is a screen, not part of the cliff identity

  Provenance: derived by codex, implemented by grok.

  Independent review: cursor

  Gates (6 of 6 pass): fixture registers a test: pass; pins resolve: pass; families exist: pass; worksheet exists: pass; mutation receipt exists: pass; provenance independent: pass

  Worksheet: insight-irmaa-tier-edge-premium-cliff.md Mutation receipt: insight-irmaa-tier-edge-premium-cliff.mutation.md

  Evidence tests: irmaaTierEdge.evidence.test.ts

  Implemented in: irmaaTierEdge.ts

  Is the number of: insight-irmaa-tier-edge-premium-cliff

  Feeds: insight-impact-ending-after-tax-estate-delta

Engine source: engine/src/insights/types.ts (InsightImpact.endingAfterTaxEstateDelta)