ESG Values Alignment Auditor

Why this Data Set exists

ESG and values-based platforms have a data problem that's distinct from the modal wealth platform: they need to reason about preferences, not just balances. A client whose values include climate concern and biblically-responsible exclusions is different from a client with the same wealth and same risk tolerance who has no values constraints. The portfolio construction differs structurally — different equity universe, different fund selection, different sector allocations. Software supporting this needs test fixtures where the values data is structured alongside the financial data.

The second problem is privacy. Religion is often the single most important preference driver for faith-based investors, but it's also the field most carefully governed by privacy contracts. The WealthSchema v4 privacy contract addresses this with a conditional overlay: religion (and race/ethnicity) appear only in households tagged for the bundles where it's relevant to the planning. B08 is one of these bundles; B26 is the other. For other bundles, these fields are not populated by default.

This Data Set provides 25 households where the values-overlay is fully populated: ESG preference scores, sector exclusion lists, impact allocation percentages, DAF holdings (donor-advised funds are heavily used in values-based giving), and the demographic values that drive faith-based screening. The conditional privacy overlay is intentionally clear — these fields are populated only where the bundle scope makes them relevant.

Use Cases

Who uses this Data Set

What's inside

The 25 households cluster around two archetypes: ESG / Values-Based Investors (N-02, the broader preference-driven cohort) and ESG / Faith-Based / Impact Investors (ES-01, the more specifically faith-based cohort). The mix is intentional — about 60% are general values-based investors with broad ESG preferences; about 40% are specifically faith-based with religious-screening overlay.

Every household has a structured ESG / values profile: ESG preference scores across the canonical preference dimensions (climate, social, governance) on a 0-100 scale; sector exclusion lists with the specific sectors flagged (firearms, tobacco, alcohol, gambling, fossil fuels, weapons manufacture, adult entertainment); impact allocation percentages where the household has a deliberate carve-out for impact investments; and DAF holdings with grant-making history. The conditional demographic overlay populates religion and race/ethnicity fields where the bundle scope makes them relevant — for these 25 households, both fields are populated; for households in other bundles, these fields are absent per the privacy contract.

The Data Set ships as JSON and CSV. The WealthSynth Methodology PDF documents the values-preference taxonomy (with mapping to MSCI ESG categories), the sector-exclusion taxonomy, the impact-allocation methodology, the conditional privacy overlay specification (with detailed treatment of when religion/race fields should and shouldn't be populated), and the DAF grant-making framework.

Preview a sample household

A redacted summary of one household from this Data Set — names, employers, exact balances, and metro area are stripped. Ages are bucketed, income and net worth are reported as bands. The full record (and all 25 like it) ships in the ZIP.

Technical Highlights

Sample Schema Fields

{
  "esg.preference_scores": <value>,
  "esg.exclusion_sectors": <value>,
  "esg.impact_allocation_pct": <value>,
  "demographics_overlay.religion": <value>,
  "demographics_overlay.values": <value>
}

Sample queries

households.filter(h =>
  h.esg.preference_scores.climate >= 80 &&
  h.assets.equity.fossil_fuel_exposure_pct >
    h.esg.fossil_fuel_tolerance_pct
)

households.filter(h =>
  h.demographics_overlay.religion_practice ===
    'biblically_responsible'
).map(h => ({
  id: h.id,
  exclusions: h.esg.exclusion_sectors,
  violations: h.assets.equity.holdings.filter(holding =>
    h.esg.exclusion_sectors.includes(holding.sector))
}))

households.filter(h =>
  h.investments.daf_grants?.length > 0
).map(h => ({
  id: h.id,
  daf_balance: h.investments.daf_balance,
  ytd_grants: h.investments.daf_grants
    .filter(g => g.year === currentYear)
    .reduce((s, g) => s + g.amount, 0),
  prior_year_grants: h.investments.daf_grants
    .filter(g => g.year === currentYear - 1)
    .reduce((s, g) => s + g.amount, 0)
}))

households.filter(h =>
  h.esg.target_impact_allocation_pct >
    h.esg.current_impact_allocation_pct &&
  h.esg.target_impact_allocation_pct -
    h.esg.current_impact_allocation_pct > 0.05
)

Methodology

Each household's values profile is generated against archetype-specific patterns. The 25-household corpus is intentionally focused — values-based investing concentrates in specific demographic and life-stage cohorts, so a small targeted corpus serves the use cases better than a larger generic one. ESG preference scores are calibrated against US SIF Foundation data on values-based investor priorities. Sector exclusion lists use the canonical taxonomy spanning the major exclusion categories (firearms, tobacco, alcohol, gambling, fossil fuels, weapons, adult entertainment). DAF grant-making patterns are calibrated against Schwab Charitable, Fidelity Charitable, and NPT-published donor-advised-fund industry data. The conditional demographic overlay is populated for these 25 households with full transparency about the privacy implications — the Methodology PDF documents the consent / disclosure framework. The corpus passes the WealthSynth consistency validator (preference scores reconcile with sector exclusions; impact allocations are mathematically consistent; DAF math is correct) and the LLM-as-judge gate. Annual refresh tracks evolving values-investing taxonomies and any privacy-regulation updates.

Included Archetypes (2)

Frequently asked questions