Browse by subject:
Curriculum/Chapter 16
CHAPTER 16 OF 18

Data & methodology notes

~27 min full text
EDITORIAL REVIEW IN PROGRESS
This chapter is public working text. Its sequence and numerical framework have been reconciled, while wording, citations, and study-guide material remain under editorial review. For the learning sequence, return to the curriculum.
CORE LESSON

Data & Methodology Notes

~13 min

The computation chain from revenue to welfare destruction

A welfare-destruction figure is only as trustworthy as the steps that produce it, so the System Asset Pricing Model (SAPM) keeps every step in the open. It connects an observable quantity, annual industry revenue, to an estimated one, system welfare destruction, through a short chain of three steps. First, each domain is assigned a welfare-beta coefficient, βW (beta-W), defined as the average ratio βW = ΔW/Π, where Π is annual industry revenue. This is the framework's Iron Law: the denominator is revenue, never profit. (A distinct marginal quantity, −dW/dΠ, is a separate tool for a decision at the margin and is not used here.) Second, the annual welfare destruction for a domain is the product ΔW = βW × Π. Third, the canonical reported aggregate is the sum of those ΔW figures across the 58-domain ranked revenue-ratio portfolio1; the wider working-paper program retains a separate 61-row coverage set. Auditability is the design goal at every step. The revenue estimate, the βW coefficient, and the resulting ΔW stay separable, each traceable to a specific source rather than folded into a single opaque welfare number. A reader who disputes one input can change that input and recompute the rest, because the arithmetic is exposed rather than asserted.

Measurement requires a mitigation plan

A number is not a plan. Once a βW estimate or a Field 17 system-welfare record puts a loss on the table, the next question is what to do about it, and that question has a structure. The mitigation plan asks what game is producing the Hollow Win, what rule change would reduce the loss, which Conflictoring lanes can act lawfully, what Reform Pathfinder route applies, what Policy Lab milestones would show progress, and what evidence would show the plan is failing. For a company, that plan spans internal changes, board and CEO responsibility, disclosure discipline, and lawful industry-wide reform where no single firm can reduce the harm alone. It names the board and CEO's role, the relevant employees or whistleblowers, regulators, policymakers, shareholders, plaintiff litigators with lawful access in the relevant setting, and communities and the affected public, together with the lawful path for collaboration — including regulator-convened or regulator-sanctioned industry reform where antitrust risk matters. Measurement creates responsibility; the mitigation plan is how that responsibility becomes something an organization can act on.

Official valuation inputs: specific values with source vintage

A welfare figure hides its assumptions unless the inputs are named, so the model states each default value and the vintage of the source it came from. The value of a statistical life is $11.5 million in the current model.2 That is a single input, not a range, and the paper can show how it relates to federal-method values3, which shift by agency, base year, inflation update, income adjustment, and guidance vintage. The carbon value is $190 per metric ton of CO2, matching the EPA 2023 central social-cost-of-carbon estimate4 at the 2% near-Ramsey discount rate — the low rate used to value harms to future generations5. That, too, is a modeling choice tied to a cited vintage. Where the work departs from federal defaults, chiefly in how it discounts harms to future generations, the departure is stated as a choice, given a reason, and checked with sensitivity tests. Naming the inputs this way is what lets a later reader see exactly which assumption is doing the work behind any figure.

The headline aggregate and its publication status

The current working-paper estimate for annual system welfare destruction across the 58-domain ranked revenue-ratio SAPM portfolio is $69.0 trillion — the sum of ΔW across domains whose annual welfare loss and annual revenue share the same activity boundary. The source behind that aggregate is the System-Welfare-Adjusted GDP aggregate source. Internal 86/89 manuscript variants are provenance aliases of the same work, not separate public publications, and literature or citation claims stay unevaluated until an SSRN (the Social Science Research Network, an online repository for working papers) or other public posting exists. The status of the number matters as much as the number itself. The System-Welfare-Adjusted GDP source presents the aggregate as a capture-aware, preliminary measurement: the total and the domain ranking are both conditional on the stated price set, and its practical use is directional — a figure that points, not a precise dollar count.

The eight welfare-cost channels

A single trillion-dollar total tells a reader little about what kind of harm it stands for, so it helps to see the loss broken into channels. The current public headline is the reconciled revenue-ratio portfolio: about $69.0 trillion in annual system-welfare loss across 58 ranked domains. Earlier source materials broke that loss into four teaching channels — mortality and morbidity, cleanup, productivity, and future damages. That decomposition is still useful for explaining the kinds of harm SAPM measures, but it is not the arithmetic bridge to the $69.0 trillion headline, because it mixes annual flows with a net-present-value future-damage stock (net present value is the worth today of future costs after discounting). The live measurement rule is simpler: for each ranked domain, take the source-traced annual welfare loss and the annual industry revenue on the same activity boundary, then sum the annual losses across the ranked revenue-ratio portfolio. Keep the channels for teaching; use the same-boundary annual sum for the number.

The PFAS case as an illustration of the computation chain

The chain is easiest to see running on a single traceable case, and per- and polyfluoroalkyl substances (PFAS) supply one. The Environmental Protection Agency estimates that meeting its PFAS drinking-water rule will cost public water systems on the order of $1.5 billion a year13 — one entry in a PFAS cleanup and health-liability burden running to tens of billions of dollars a year. In the same period, 3M, one of the primary PFAS manufacturers, reported $32.6 billion in total revenue6 and settled a $10.3 billion water-contamination lawsuit while continuing to produce replacement chemistries. The sequence is the lesson: the revenue was booked, the system cost was deferred, and the gap between them never entered the original product-approval decision record. That is why the chain has to carry both sides at once — the revenue term Π that the firm books, and the welfare-destruction term ΔW that the system absorbs. Measure only the first, and you reproduce the exact blind spot that let the decision look acceptable when it was made.

Scope classifications and domain coverage

The portfolio covers 58 ranked revenue-ratio domains, with the broader study retaining 61 studied rows. The portfolio-level βW is 2.61; βW remains the average ratio ΔW/Π, with Π defined as annual industry revenue on the same boundary. One reporting choice is worth pausing on: the 58-domain ranked panel keeps every domain in view, including those whose βW is at or below 1. Dropping the low-coefficient domains would be tempting — it would leave only the most damaging cases — but it would also overstate the typical result, so the framework keeps them in view. The scope follows the source manuscripts, which run case studies across sectors. The supporting provenance source confirms that the studied market failures span finance, healthcare, technology, environment, and other sectors, so the aggregate rests on a broad evidence base rather than a single illustrative industry.

Source discrepancies and measurement limits

A measurement is only as honest as its stated limits, and this one carries five a careful reader should hold in mind. First, the aggregate level is planner-relative and capture-sensitive: it depends on whose welfare standard is applied and on how far the measurement itself can be captured by the interests it measures. Second, the domain βW table is conditional on the stated price set; change the prices used to value harm and the levels move with them. Third, no double-counting adjustment has been applied at this stage, so the market-failure domains are treated as independent even where some overlap exists. Fourth, every ΔW figure is an annual flow cost, and cumulative stock damages are not folded into that annual number. Fifth, the geographic allocation of global-scope ΔW is an approximation built on nominal GDP share. None of these limits invalidates the measurement; together they set the conditions under which the numbers should be read. The System-Welfare-Adjusted GDP source is explicit that the whole estimate is conditional: repricing one channel, such as the value of a life, can reorder domains that rely on different harm channels, so the domain ranking moves with the price set just as the total does, and every figure is preliminary — subject to peer review, replication, and revision.

Field evidence: the cost-of-capital test

The cleanest way to know whether Decision Accounting (DA) creates value is to let the market price it. Measure a firm's cost of debt, cost of equity, and weighted average cost of capital (the blended rate a firm pays to raise debt and equity) before DA adoption, then track whether and how fast those measures decline after implementation across many adopters. A large, repeated post-adoption decline — for instance near the hypothesized 200-basis-point range, where a basis point is one hundredth of a percentage point — would be strong evidence that DA is creating market value. Everything else is mechanism evidence: record completeness, weak-answer rates, prediction accuracy, governance-state recovery time, audit results, board-review quality, litigation outcomes, regulatory outcomes, and decision-quality scores. These help explain why the cost of capital moved, but they are weaker causal proof, because the same firm cannot show what its lawsuits, regulatory outcomes, or internal governance quality would have been had it never adopted DA.

Propositions 8a–8b and its implications for measurement

It is tempting to think a thorough enough audit of a firm's financial data could certify that its portfolio did no system harm. Propositions 8a–8b, reported in the $69.0T aggregate source, prove that it cannot: no rule, checklist, or audit computed from transaction data can deliver that certificate. The reason is structural, not empirical. System welfare is the omitted coordinate, and it is not derivable from the parties' payoffs, so no function of the standard observation set can recover it. The consequence for measurement is direct: assess a portfolio's welfare impact using only conventional financial or regulatory data, however detailed, and you will always miss the system-welfare coordinate. This is the formal reason the framework cannot reuse existing disclosure as-is. It needs a dedicated measurement apparatus — the βW computation chain on one side, and Decision Accounting Field 17, SYSTEM WELFARE, on the other.

The Missing System Theory as the structural mechanism

MST (the three-coordinate payoff space and the Hollow Win, C=0/A=1/B=1, taught in full earlier)7 is the mechanism the aggregate traces to.
The beta-W denominator: why revenue and not profit · ~2 min
A firm that wanted to look less harmful could not do it by refinancing — and that is the point of putting revenue, not profit, under the βW ratio. The Iron Law fixes it: βW = ΔW/Π, where Π is revenue, never profit. The theoretical case rests on what generates the externality. Revenue is the gross measure of the economic activity that produces system harm; profit is a residual left after costs, financing, and tax structure are applied. Measure βW per dollar of profit and the coefficient starts responding to accounting and capital-structure choices that have nothing to do with the underlying harm, letting a firm lower its apparent damage ratio by restructuring its financing rather than by reducing what it destroys. The practical case is observability: revenue is more standardized across industries, more consistently reported, and harder to manipulate than profit, which is what makes cross-domain comparison meaningful. The choice does draw an objection — that profit better captures the net benefit driving the destructive activity. The framework's answer is that it measures welfare destroyed per dollar of economic activity, not per dollar of profit, so revenue is the correct scale variable. The pricing-harm-and-distributive-justice source states βW as ΔW relative to revenue8, and the pricing-the-incommensurable-sign-and-rank source confirms that revenue is used specifically to prevent manipulation and ensure comparability.9
  • The βW denominator is annual industry revenue, never profit.
  • Revenue is more observable, more standardized, and harder to manipulate than profit.
  • The choice is deliberate and defended on both theoretical and practical grounds, and it answers the profit-based objection directly.
The welfare ledger as an accounting architecture · ~2 min
Arguments about welfare numbers usually collapse two different questions into one: is the bookkeeping sound, and are the plugged-in values right? The System-Welfare-Adjusted GDP source pulls them apart, treating the task as an accounting-architecture problem rather than an empirical-calibration problem, and turning the program's welfare bill into an auditable ledger. The accounting architecture is the set of rules: how ΔW is computed as βW × Π, how domain results are aggregated, and how the totals are presented as ledger entries. The empirical calibration is the separate question of which values get assigned to βW and Π for each domain. Because the source keeps the architecture distinct from the still-unexecuted calibration, the architecture holds even as individual values change: a revised price set or a corrected revenue figure moves entries within the ledger without breaking its structure. The format also carries an auditability that black-box welfare estimates lack — each entry traces to a source, and the arithmetic can be re-run independently. The reader is not asked to trust a headline; the reader is handed the ledger and can check it line by line.
  • The welfare ledger separates accounting architecture from empirical calibration.
  • The architecture stays durable even when calibration values are revised.
  • The ledger format makes the measurement auditable, decomposable, and checkable line by line.
The Public-data asset-pricing study and its measurement implications · ~4 min
A theory earns more trust when its authors report the tests it fails alongside the ones it passes. The public-data asset-pricing study reruns the SAPM asset-pricing predictions on public data10, and the result is mixed in an informative way rather than a confirming one. Four predictions are tested. Prediction 1 (that the most system-destructive firms earn no unusual return in normal times, before markets learn to price system welfare) survives: the high-minus-low βW portfolio (long the highest-βW firms, short the lowest) shows no detected normal-times alpha. The source reads this as a structural result, not a costly-information result, because the market should not be expected to price a W coordinate before it has a theory that defines W as a priced coordinate at all. Prediction 2 (that welfare-beta predicts how sharply a firm's stock reacts when a restoration event forces the hidden system harm into view) survives strongly: βW predicts the magnitude of restoration-event returns even after controlling for the Fama-French-Carhart four-factor model11, sin stocks (tobacco, alcohol, gambling), litigation risk, and a green-minus-brown climate factor (the return gap between environmentally clean and polluting firms)12. Prediction 3 (that the most system-destructive firms should have more downside, negative-skew returns) fails: high-βW skew turns out to be more positive, not more negative. Prediction 4 (that firms sharing the same damaged system should have returns that move together) fails: shared-system membership does not produce a positive within-system residual covariance load. The honest conclusion the source draws is bounded. The data support SAPM and MST as a structural no-premium and restoration-loading story, while leaving the broad hidden-tail-risk or non-diversifiable system-covariance asset-pricing theory for later evidence. The next falsifiable test the source names is diffusion: W should become priced only when βW-style disclosures, activist reports, litigation models, or shareholder adoption make the coordinate legible. The measurement implication is direct. Current βW estimates are not yet reflected in asset prices, which is consistent with the MST prediction that the system-welfare coordinate sits outside the payoff space, and it sets the empirical condition under which that would change.
  • Prediction 1 (that the most system-destructive firms earn no unusual return in normal times, before markets learn to price system welfare) survives: no detected normal-times alpha for the high-minus-low βW portfolio.
  • Prediction 2 (that welfare-beta predicts how sharply a firm's stock reacts when a restoration event forces the hidden system harm into view) survives strongly: βW predicts restoration-event return magnitude after standard controls.
  • Prediction 3 and Prediction 4 fail: the data do not support the hidden-tail-risk or system-covariance predictions.
  • The next test is diffusion: W should become priced only once disclosures and adoption make the coordinate legible.

Sources supporting the Chapter 16 measurement framework

Source linkRole in measurement frameworkContribution to Chapter 16Publication status
book-$72T welfare-destruction estimateHeadline aggregate and channel decompositionProvides the $69.0T aggregate, the eight welfare-cost channels, the portfolio-level revenue-weighted βW, the PFAS case, and the MST and Propositions 8a–8b statements.Manuscript/source; SSRN/public posting needs confirmation
book-$72T welfare-destruction estimateConfirmation and evidence scopeConfirms the $69.0T aggregate and reports the evidence base: per-domain welfare-beta estimates across finance, healthcare, technology, and environment.Manuscript/source; SSRN/public posting needs confirmation
System-Welfare-Adjusted GDP and Welfare LedgerAccounting architecture and measurement philosophyEstablishes the ledger format, separates architecture from calibration, and presents the numbers — total and domain ranking alike — as conditional on the price set and preliminary.Status requires source verification before public citation
excluded-coordinate-return-signatureEmpirical test of measurement predictionsProvides the public-data rerun: Prediction 1 and Prediction 2 survive, Prediction 3 and Prediction 4 fail, and the structural interpretation with diffusion as the next test.Status requires source verification before public citation
APPLIED EXERCISE

Tracing the computation chain for a single domain

~3 min
Select one domain from the 58-domain ranked revenue-ratio SAPM portfolio. Trace the computation chain for that domain and reconstruct the measurement process from revenue to welfare destruction. Specifically: (1) Identify the annual industry revenue for your chosen domain from the sources. If the exact revenue figure is missing, write NEEDS_SOURCE and explain what kind of source would supply it. (2) Identify the βW coefficient for your chosen domain from the sources. If the exact βW is not present, write NEEDS_SOURCE. (3) Compute ΔW as βW × Π. (4) Identify which of the eight welfare-cost channels (for example, mortality, health and disease, climate and carbon, or lost productivity) the ΔW for your domain falls into, and explain your reasoning. (5) State one limit of your computation, referencing the measurement limits from the chapter core. (6) Write a brief note on what would have to change in the sources for your computation to be considered auditable by an independent reviewer.
Answer key
  1. A strong answer identifies a specific domain and provides exact revenue and βW figures from the public table, or clearly states NEEDS_SOURCE where the figures are absent. Because the public table in this chapter does not include every domain-level revenue and βW figure, NEEDS_SOURCE is the correct response for most domains, and the exercise rewards recognizing that boundary rather than inventing values.
  2. The ΔW computation should be arithmetically correct and clearly presented, or correctly deferred when an input is NEEDS_SOURCE.
  3. The channel classification should be justified with reference to the eight welfare-cost channels described in the $69.0T aggregate source.
  4. The limit should reference one of the measurement limits from the chapter core: planner-relativity, capture-sensitivity, the domain-independence assumption, the flow-versus-stock distinction, or the geographic approximation.
  5. The auditability note should reference the ledger format from the System-Welfare-Adjusted GDP source and explain which sources an independent reviewer would need for verification.
READING PATH
  1. Primary source for the $69.0T aggregate, the eight welfare-cost channels, the source-reported portfolio βW, and the PFAS case. It also states Propositions 8a–8b and the Missing System Theory.
    Extract the headline aggregate, the eight channel values, the βW value, and the PFAS case details. Note the two theorem statements and their implications for measurement.
  2. Confirms the $69.0T aggregate and reports the evidence scope: per-domain welfare-beta estimates with cited shadow prices. It also confirms that the studied market failures span finance, healthcare, technology, and environment.
    Extract the confirmation of the aggregate and the evidence scope. Note the sector coverage and the manuscript status.
  3. System-Welfare-Adjusted GDP and the Welfare Ledger
    Establishes the accounting architecture and measurement philosophy. It separates architecture from calibration and presents every figure — total and domain ranking alike — as conditional on the price set and preliminary.
    Extract the ledger format, the architecture-versus-calibration distinction, and the conditional sign-and-rank claim. Understand why the aggregate level is planner-relative and capture-sensitive.
  4. Provides the empirical test of the measurement predictions through a public-data rerun, showing which predictions survive and which fail and establishing the structural interpretation.
    Extract the Prediction 1 through Prediction 4 results and understand why Prediction 1 and Prediction 2 survive while Prediction 3 and Prediction 4 fail. Note the structural interpretation and the diffusion test that comes next.
CHAPTER SYNTHESIS
QUESTION
What is the computation chain from revenue to welfare destruction in the SAPM framework?
ANSWER
It is three steps: βW = ΔW/Π, where Π is annual industry revenue; then ΔW = βW × Π for each domain; then the aggregate is the sum of ΔW across the 58-domain ranked revenue-ratio panel. Each step is auditable and traceable to a source.
QUESTION
What are the eight welfare-cost channels into which the $69.0T aggregate is decomposed?
ANSWER
Mortality valued at the value of a statistical life ($15.8T/year), health and disease ($12.7T/year), social and informational harm ($10.1T/year), climate and carbon ($8.4T/year), environmental and ecosystem loss ($6.5T/year), economic extraction and deadweight loss ($6.8T/year), governance and systemic failure ($5.4T/year), and lost productivity ($3.3T/year). All eight are annual flows, and together they account for the $69.0 trillion per year aggregate.
QUESTION
Why does Propositions 8a–8b matter for measurement?
ANSWER
It proves that no rule, checklist, or audit computed from transaction data can certify that a portfolio did no system harm, because the omitted system-welfare coordinate is not derivable from the parties' payoffs. A dedicated measurement apparatus is therefore required.
QUESTION
What is the difference between accounting architecture and empirical calibration in the welfare ledger?
ANSWER
The architecture is the set of rules for computing ΔW, aggregating across domains, and presenting results as ledger entries. The calibration is the actual values assigned to βW and Π. The architecture stays durable even when calibration values are revised.
QUESTION
What does the public-data asset-pricing test tell us about the current state of βW measurement?
ANSWER
Prediction 1 and Prediction 2 survive (no normal-times alpha, restoration-event loading) while Prediction 3 and Prediction 4 fail (no skew or covariance prediction). This supports a structural reading: the market does not yet price W because it lacks a theory that defines W as a priced coordinate, and diffusion of disclosures is the next falsifiable test.
QUESTION
What is the PFAS case and what does it illustrate about the computation chain?
ANSWER
3M reported $32.6 billion in revenue and settled a $10.3 billion lawsuit while the EPA estimated $1.5 billion a year for U.S. water systems (up to €80 billion a year in the EU). The recorded revenue was booked, the system cost was deferred, and the gap never entered the original product-approval record. It shows why the chain must carry both the revenue term Π and the welfare-destruction term ΔW.
QUESTION
Why is the βW denominator annual industry revenue rather than profit?
ANSWER
Revenue is the scale of the activity that generates the externality and is more observable and harder to manipulate than profit, which is a residual sensitive to accounting and financing choices. Measuring per dollar of profit would let a firm lower its apparent damage ratio by restructuring financing rather than reducing harm.
SOURCE
book-$72T welfare-destruction estimate
SOURCE
System-Welfare-Adjusted GDP and Welfare Ledger
SOURCE
book-$72T welfare-destruction estimate
NOTES & REFERENCES
  1. System-Welfare-Adjusted GDP and the Welfare Ledger: sets out the ΔW = βW × Π computation and the domain aggregation used here. summary. ↩
  2. Program calibration paper documenting the model's default value-of-statistical-life and social-cost-of-carbon inputs and their source vintage. summary. ↩
  3. W. Kip Viscusi and Joseph E. Aldy, "The Value of a Statistical Life: A Critical Review of Market Estimates Throughout the World," Journal of Risk and Uncertainty 27, no. 1 (2003): 5–76. link. ↩
  4. U.S. Environmental Protection Agency, "EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent Scientific Advances" (December 2023). link. ↩
  5. William D. Nordhaus, "Revisiting the Social Cost of Carbon," Proceedings of the National Academy of Sciences 114, no. 7 (2017): 1518–1523. link. ↩
  6. PFAS (program domain paper): the source for the manufacturer-revenue and cleanup-cost figures cited here. summary. ↩
  7. The Missing System Theory: formalizes the three-coordinate payoff space and the Hollow Win result the aggregate traces to. summary. ↩
  8. Pricing Harm and Distributive Justice (program paper): derives βW as ΔW measured relative to revenue. summary. ↩
  9. *Pricing the Incommensurable: Sign and Rank∗ (program paper): establishes that βW uses revenue rather than profit as the denominator. summary. ↩
  10. The Excluded-Coordinate Return Signature (program paper): re-runs the SAPM asset-pricing predictions on public market data. summary. ↩
  11. Eugene F. Fama and Kenneth R. French, "Common Risk Factors in the Returns on Stocks and Bonds," Journal of Financial Economics 33, no. 1 (1993): 3–56; Mark M. Carhart, "On Persistence in Mutual Fund Performance," Journal of Finance 52, no. 1 (1997): 57–82. link. ↩
  12. Harrison Hong and Marcin Kacperczyk, "The Price of Sin: The Effects of Social Norms on Markets," Journal of Financial Economics 93, no. 1 (2009): 15–36. link. ↩
  13. U.S. Environmental Protection Agency, PFAS National Primary Drinking Water Regulation (2024): EPA estimates annual compliance costs to public water systems on the order of $1.5 billion. link. ↩
DIAGRAM NOTES
These notes describe diagrams planned for this chapter. The diagrams are not published yet.
DIAGRAM NOTE
The computation chain from revenue to welfare destruction
flow diagram with four nodes and three edges
Show the sequential computation from observable industry revenue to estimated system welfare destruction, with each step auditable and traceable to a source.
DIAGRAM INPUTS
Node 1: Industry revenue (Pi) for each domain
Node 2: Beta-W coefficient for each domain
Node 3: Annual welfare destruction (Delta-W = beta-W x Pi) for each domain
Node 4: Aggregate welfare destruction (sum of Delta-W across all domains)
READER CAPTION
The computation chain connects observable revenue to estimated welfare destruction through the beta-W coefficient. Each step is auditable: the revenue estimate, the beta-W coefficient, and the resulting Delta-W can all be traced to specific sources. The chain is designed to be transparent, decomposable, and checkable line by line.
TEXT FALLBACK
If the flow diagram is not implemented, present the computation chain as a four-row table: (1) Industry revenue per domain, (2) Beta-W coefficient per domain, (3) Delta-W = beta-W x Pi per domain, (4) Aggregate Delta-W across all domains.
book-$72T welfare-destruction estimateSystem-Welfare-Adjusted GDP and Welfare Ledger
WHAT TO DO NEXT
Restate the chapter claim. For policy triage, open Policy Lab; for measurement, open Domain Tables.
← PREVIOUS
Ch. 15: k* and finite coalitions in repair-regular cases
NEXT CHAPTER →
Ch. 17: Repair-regular game-change theorem (bounded existence result)
© 2026 Erik Postnieks · Independent Researcher · Salt Lake City