Skip to content
Research
KnowledgeVisuals
Pricing↗Open the workspace↗Get a free API key↗
.orgResearchResearch — current site.netWorkspace↗Workspace — opens riskmodels.net.appAPI↗API — opens riskmodels.app
Next filing · Form 10-Q · Q3 2026 · 20 days
Next filing · Form 10-Q · Q3 2026 · 20 daysAPI Update · Point-in-Time (PIT) historical commit tracking — securities and funds unifiedAPI Update · ERM3 L3 variance partition — institutional transparency releaseFactor Research · L-Star adaptive hedge depth — out-of-sample validation across ten allocator universesPart 1 · One Position, Four BetsPart 2 · Risk Structure in 13F FilingsNext filing · Form 10-Q · Q3 2026 · 20 daysAPI Update · Point-in-Time (PIT) historical commit tracking — securities and funds unifiedAPI Update · ERM3 L3 variance partition — institutional transparency releaseFactor Research · L-Star adaptive hedge depth — out-of-sample validation across ten allocator universesPart 1 · One Position, Four BetsPart 2 · Risk Structure in 13F Filings
Ledger

Concept note · Personal Financial Record · 2026-10-19

A private 10-Q for personal finance.

Intelligent financial applications need permissioned data with a stable home and shared meaning. A Personal Financial Record combines user-controlled storage with a taxonomy for assets, liabilities, policies, cash flows and measured risk.

Storage preserves the record. Taxonomy makes it computable. Tools measure what the record means.

The investment portfolio is the first working module. It is one schedule inside the broader personal balance sheet.

01 · The operating model

A useful record needs a home, a language and measurement.

01

Storage

A dated record under the owner’s control, with permission, revocation, history and source evidence.

02

Taxonomy

Defined objects and terms so rate, benchmark, maturity, identity, value and provenance mean the same thing to every app.

03

Measurement

Specialist tools compute panels from the record and return dated, attributed results instead of generated estimates.

02 · The health precedent

Medical data already separates storage, structure and terminology.

HL7 states that automated clinical decision support requires structured and standardized data. FHIR supplies exchangeable resource structures. LOINC identifies observations and documents. SNOMED CT gives clinical concepts computable meaning. The application can reason more reliably because the record carries defined objects, dates, units, relationships and provenance.

Personal finance has mature feeds, payment standards and identifiers. The missing layer is a portable record assembled around the individual, with a shared taxonomy spanning the entire balance sheet and a consistent place for derived measurements.

LayerHealthPersonal finance
Structured recordFHIR resources give applications a consistent way to represent and exchange health information.PFR resources describe people, accounts, assets, liabilities, policies, cash flows and relationships.
Shared terminologyLOINC identifies observations and documents; SNOMED CT supplies computable clinical meaning.A financial taxonomy identifies instrument type, rate type, benchmark, term, security, valuation basis and provenance.
Measured panelA lab result carries a test identity, value, unit, reference, date and source.A financial panel carries a metric identity, value, unit, benchmark, as-of date, method and provider.
Permissioned applicationsAuthorized applications read the same record and can reason from the same defined concepts.Gemini, Google Finance and professional applications read the same owner-controlled record.

Structured record

Health

FHIR resources give applications a consistent way to represent and exchange health information.

Personal finance

PFR resources describe people, accounts, assets, liabilities, policies, cash flows and relationships.

Shared terminology

Health

LOINC identifies observations and documents; SNOMED CT supplies computable clinical meaning.

Personal finance

A financial taxonomy identifies instrument type, rate type, benchmark, term, security, valuation basis and provenance.

Measured panel

Health

A lab result carries a test identity, value, unit, reference, date and source.

Personal finance

A financial panel carries a metric identity, value, unit, benchmark, as-of date, method and provider.

Permissioned applications

Health

Authorized applications read the same record and can reason from the same defined concepts.

Personal finance

Gemini, Google Finance and professional applications read the same owner-controlled record.

03 · The personal 10-Q

One dated balance sheet, with a typed schedule behind every line.

The “personal 10-Q” is a mental model rather than a regulatory filing. It records assets, liabilities, protection, recurring commitments and changes over time. Each total reconciles to instrument-level schedules and source evidence. Quarterly snapshots make changes inspectable while preserving the original facts and valuation dates.

Illustrative storage

personal-finance/
  manifest.pfr.json
  snapshots/2026-Q3/
    balance-sheet.json
    assets/
      investments.json
      real-estate.json
    liabilities/
      mortgages.json
      credit.json
    protection/
      life-insurance.json
    panels/
      liquidity.json
      portfolio-risk.json
    evidence/
      source-index.json

Source documents can remain separate and access-controlled. The manifest carries references, hashes, permissions and provenance rather than copying unnecessary sensitive data.

Manifest

Owner, as-of date, permissions, data-use terms, schema and taxonomy versions

Balance sheet

Typed assets, liabilities, policies, commitments and ownership relationships

Schedules

Instrument-specific terms such as rate, benchmark, maturity, premium and reset rules

Evidence

Statements, policies and source references with extraction date and provenance

Panels

Derived liquidity, debt, insurance and investment measurements, each labeled and versioned

History

Periodic snapshots and changes, preserving what was known and when it became known

04 · Taxonomy examples

“Rate” only becomes useful when the record defines its role.

A value such as 6.25 has no operational meaning by itself. The record must state whether the value is fixed or floating, its unit, effective date, benchmark, spread, reset rule and term. The same rule applies to an insurance crediting rate, a portfolio benchmark and every valuation in the balance sheet.

ScheduleStored positionTaxonomy and termsComputable panel
MortgagePrincipal, collateral, lien priority, paymentRate type, rate, index or benchmark, spread, reset cadence, amortization term, maturityPayment schedule, rate-reset exposure, maturity ladder
Life insuranceCarrier, policy type, face amount, cash value, policy loansPremium schedule, term or maturity, guarantees, credited-rate or index method, ridersCoverage, premium commitments, cash-value and guarantee schedule
Investment portfolioAccounts, positions, lots, cash, stable security identityTax status, currency, benchmark, corporate-action chain, valuation and model datesAllocation, factor exposure, risk decomposition, dated performance
Real estateProperty interest, ownership share, cost basis, current valueValuation date and source, encumbrances, taxes, insuranceEquity, leverage, carrying cost, concentration
Cash and creditAccount type, currency, balance, credit limitAPR, index or benchmark, spread, payment date, statement periodLiquidity, borrowing cost, near-term obligations

Mortgage

Stored position
Principal, collateral, lien priority, payment
Taxonomy and terms
Rate type, rate, index or benchmark, spread, reset cadence, amortization term, maturity
Computable panel
Payment schedule, rate-reset exposure, maturity ladder

Life insurance

Stored position
Carrier, policy type, face amount, cash value, policy loans
Taxonomy and terms
Premium schedule, term or maturity, guarantees, credited-rate or index method, riders
Computable panel
Coverage, premium commitments, cash-value and guarantee schedule

Investment portfolio

Stored position
Accounts, positions, lots, cash, stable security identity
Taxonomy and terms
Tax status, currency, benchmark, corporate-action chain, valuation and model dates
Computable panel
Allocation, factor exposure, risk decomposition, dated performance

Real estate

Stored position
Property interest, ownership share, cost basis, current value
Taxonomy and terms
Valuation date and source, encumbrances, taxes, insurance
Computable panel
Equity, leverage, carrying cost, concentration

Cash and credit

Stored position
Account type, currency, balance, credit limit
Taxonomy and terms
APR, index or benchmark, spread, payment date, statement period
Computable panel
Liquidity, borrowing cost, near-term obligations

05 · Why applications become more accurate

The record constrains the model before the model writes an answer.

A language model can summarize source documents. It should not invent the identity of a security, infer whether a mortgage rate resets, or guess which benchmark governs a policy. Typed fields and controlled terms make those facts explicit. Specialist tools can then calculate from the same dated object and return results with units, method and provenance.

  1. 01

    Resolve

    Map each account, policy, liability, property and security to a stable identity.

  2. 02

    Normalize

    Attach dates, units, currencies, benchmarks, terms and controlled classifications.

  3. 03

    Measure

    Call the appropriate risk, debt, insurance or cash-flow tool using defined inputs.

  4. 04

    Explain

    Generate prose from measured outputs while citing the record, method and data date.

06 · Working module · investments

The first module is a portfolio schedule with a measured risk panel.

The draft v0.1 record contains accounts, holdings, stable identifiers, valuation dates, provenance and labeled derived results. The illustrative $250,000 portfolio holds 12 stocks across seven sectors. RiskModels measures a 0.84 market beta and attributes 66.5% of total risk to market and sector movements as of 2026-10-16.

12
stocks
7
sectors
0.84
market beta
66.5%
systematic share

This module establishes the storage, provenance and measurement pattern. Mortgage, insurance, real-estate, cash-flow and liability schedules remain proposed extensions.

Your browser does not support the video tag.
Gemini reads a sample portfolio record and calls RiskModels through MCP. Caption-first, 60 seconds.

07 · Product path

One owner-controlled record can support several authorized applications.

Gemini

Read the PFR in Drive, call specialist tools and explain the measured financial position.

Google Finance

Import and update investment schedules while writing dated portfolio facts back to the record.

Professional workflows

Share selected schedules with an advisor, accountant, lender or insurer under explicit permission.

The proposed data-use fields express the owner’s requested terms. Each platform must independently agree to enforce those terms. The record is descriptive infrastructure and does not authorize an app to move money, change an account or provide regulated advice.

08 · Draft and evidence

Review the working investment module.

The current bundle includes the draft specification, JSON Schema, illustrative portfolio record, summary PDF and validator. It is a v0.1 investment module. The broader personal-10-Q taxonomy is the next design step.

Download draft bundleOpen full-resolution videoContact Conrad

Sources for the analogy

  • HL7 FHIR overview — structured and standardized health data for exchange and machine processing.
  • LOINC — a shared standard for health observations, measurements and documents.
  • SNOMED CT — a terminology designed for meaning-based retrieval and processing.
  • Financial Data Exchange — existing financial transport concepts and JSON exchange patterns.
RiskModels · BW Macro · Concept for discussionPersonal Financial Record · architecture draft
← All demonstrations

RiskModels.org

A research surface for hierarchical orthogonal decomposition, variance attribution, and allocator-grade risk measurement. Operational APIs and developer workflows live at riskmodels.app.

Subscribe to the Quarterly Attribution Review.

Research notes on risk decomposition, fund attribution, 13F filings, and benchmark structure — a few times a quarter.

By registering, you agree to receive technical factor research and API deployment logs. RM-Registry-2026. Privacy Policy.

Sign inHomePrimerWorkspaceResearchVideosDemosKnowledgeConceptsReviewsLedgerReferencesAboutSubscribeMethodology noteOne-pagerAPI docsWeb appContactPrivacyStatusRSS
RiskModelsResearch/Workspace/API