The Reformulation Bill

1,547 Days

On 18 January 2022, an independent testing laboratory found fluorine indicators in a Lululemon Align Highrise Pant. The sample came back at 32 parts per million organic fluorine — low end of a range that ran to 284 ppm across thirty-two pairs of activewear sampled the same way.[¹] The result was published on the website of a consumer-advocacy outlet. No regulator acted. The share price did not move. A reader could buy the same pant the following weekend, unchanged.

On 13 April 2026 — 1,547 days later — the Texas Attorney General issued a Civil Investigative Demand to Lululemon USA Inc. The subject was PFAS in women's performance apparel. The garment line named in the investigation included Align.[²] The share price declined up to 4.5% intraday.

Same company. Same chemical class. Same garment. Between those two dates, a consumer standing in a Lululemon store in Birmingham or Denver with a pair of leggings in her hands could not answer, at the till, whether the thing she was buying contained what the lab had found. She still cannot.

The gap is four years and 86 days. It is what this report is about.

[¹]: Mamavation, "Non-Toxic Activewear Guide: PFAS in Workout Leggings & Yoga Pants," 18 January 2022. Testing by EPA-certified laboratory using ion-selective-electrode combustion. Thirty-two pairs sampled; eight positive (25%); range 10–284 ppm organic fluorine.

[²]: Office of the Attorney General of Texas, Press Release, "Attorney General Ken Paxton Launches Investigation into Lululemon Over Potential Presence of Toxic 'Forever Chemicals' in Activewear," 13 April 2026. Investigation covers marketing claims, Restricted Substances List, testing protocols, supply-chain practices.

Three Clocks

Stay with me, because the arithmetic matters.

A pair of performance leggings sold in 2026 carries three clocks on one piece of fabric.

The chemistry clock. The molecules used in fluorinated durable water repellents — the family generally called PFAS — do not degrade on any horizon a consumer product has to reckon with. Peer-reviewed estimates place environmental half-lives across the family between hundreds and tens of thousands of years, depending on the specific compound. For the reader's purpose, call it 10² to 10⁴ years.

The tort clock. If the chemistry proves to harm her, or the drinking water downstream of her washing machine, the recovery window under most consumer-protection statutes closes two to six years from discovery. The Texas Deceptive Trade Practices Act — the probable statutory basis for the April 2026 CID — carries a two-year discovery-rule window. Federal class actions under the consumer-fraud statutes sit in a similar band. Call it 10⁰ to 10¹ years.

The market clock. Equity prices move on information that reaches an authority format — an enforcement filing, a regulator's action, a securities-material disclosure. The reaction, when it comes, takes minutes. Call it 10⁻¹ to 10⁰ days.

The ratios are the finding. Chemistry to tort is 10² to 10⁴. Chemistry to market is 10⁵ to 10⁷. No mechanism operating on the shorter clocks can in principle settle a harm that runs on the longest one. The premium a consumer pays for a "PFAS-free" legging is not priced against the chemistry. It is priced against the shortest of the three — the brand's claim-verification exposure, which is what the market can actually see.

This is not a market failure in the usual sense. The prices are doing what markets price. They are pricing the risks the market has instruments for. The chemistry's clock is not one of those risks, because no instrument at point of sale measures it.

£118

The Receipt is the sum you pay. The decomposition is what your money bought.

Take a £118 Lululemon Align Highrise Pant — the segment anchor and the SKU the Texas investigation names.[³] Strip the £118 into the ledgers that lie behind it. Two of the ledgers surface at the till or in the utility bill. One of them does not.

On the till, on the brand's side:

Off the till, on the consumer's side, over a three-year garment life:

These numbers are reproducible.[⁴] The table sums across three evidence types: one sourced (commercial testing), one observed (category price-pair), four inferred (chemistry delta from supplier data, retooling from trade-press, legal reserve from enforcement-risk pricing events, supplier-margin from consolidation). The retail-markup line is a category premium rather than a per-SKU verified delta: no peer-reviewed corpus has tested the Align against a matched clean-verified comparator; the clean-premium is a premium over an unverified peer. The re-proofing range revises the £30 round figure used in the earlier draft. The claim is modest, and its modesty is the point. The consumer pays somewhere between £43 and £63 in the first ledger she can see — the £30 observed retail markup, plus £10–£25 sourced re-proofing, plus £3–£8 sourced utility — spread across the till and her own utility account, for a garment whose chemistry-clock outlives every institution that could send her a refund. The envelope is wider than a single-source figure and narrower than a rhetorical one.

[³]: Lululemon Align High-Rise Pant 25", lululemon.co.uk, retrieved on drafting day. Price range £88–£118 GBP; $98–$128 USD at shop.lululemon.com. Screenshot held with drafting notes.

[⁴]: EPA Method 1633 commercial pricing: mytapscore.com / SimpleLab $835 per water sample; textile matrix $800–$2,000 per SKU (industry-consensus range). GBP conversion at 0.79 GBP/USD (April 2026 spot): approx. £630–£1,580 per SKU — the table's "£600–£1,500" rounds the lower bound and slightly understates the upper. Archroma Smartrepel Hydro — 100% water-based non-fluorinated supplier data sheet, "more than 20 washing cycles" (archroma.com, retrieved drafting day). Nikwax TX.Direct 300 ml — nikwax.com, UK retail £8–£11. Non-fluorinated C0 DWR durability: Archroma Smartrepel Hydro supplier data sheet, "more than 20 washing cycles," cited above. The re-proofing cost arithmetic in the "On the till" table uses three to seven applications per year across a three-year garment life, which is conservative against a 20-wash baseline (a garment washed weekly over three years sees roughly 150 wash cycles, requiring 7 re-proofing applications at the 20-wash spec).

£0.00002 and £0.82

Here is the line the reader is looking for.

What is the present value, on one pair of leggings, of the harm the "PFAS-free" label gestures at? The honest answer is a range, and the range is what matters.

Take the conservative end. Assume the Align at 32 ppm organic fluorine corresponds to something on the order of ten to twenty milligrams of PFAS-equivalent mass on a 200-gram garment. Assume the share of drinking-water PFAS loading that originates with textile use-and-disposal is in the single digits — one to five per cent, because the dominant sources are AFFF firefighting foam, industrial emissions, and biosolids, and no peer-reviewed study has disaggregated textile origin from those. (This assumption is material. It is also the one the literature does not let us close.) Take the HM Treasury Green Book social time preference rate — 3.5% for years one through thirty, declining through 3.0% and 2.5% on the Weitzman-gamma schedule the Green Book itself encodes.[⁵] Apply the Uppsala full-scale plant's GAC cost per cubic metre treated.[⁶] Do the arithmetic line by line.

You get £0.00002 per garment on the present-value liability side, before any additional markup for infrastructure life or replacement cost.

Now take the generous end. Replace the Green Book with Nicholas Stern's 2006 pure-time-preference rate: ρ = 0.1%. Raise the textile allocation fraction to the top of its plausible band. Use the stricter GAC treatment regime, the 10 ng/L standard instead of the 85 ng/L standard. Run the same arithmetic.

You get £0.82 per garment.

These two numbers belong in the reader's eye together. Separated, they read as two calculators saying different things. Paired, they show what is actually going on: across the plausible assumption-space — allocation fractions from 0.01 to 0.05, discount rates from 0.001 to 0.035, treatment-cost regimes at either end — the present-value externality per garment stays under £1. A forty-thousand-fold range between the endpoints.

That range is what refutes the first objection a trained economist would raise. The objection says: the present value rounds to pennies at any sensible discount rate, so there is no mispricing; the market is functioning exactly as it should. The answer is that the range does not narrow when you change the discount rate. The range straddles the full defensible span of δ, η, g, and allocation, and the number stays rounding-error against a £118 garment. The discount rate is not the binding constraint. The binding constraint is that no instrument at point of sale prices the chemistry at all. The information-price is £0.00, and £0.00 is not reducible by any δ.

That is the finding. Not that the externality is large — it is small at every plausible combination of assumptions the reader cares to pick. The finding is that its size has nothing to do with whether the market prices it. The market prices what it has formats for. It has none here.

[⁵]: HM Treasury, The Green Book: Central Government Guidance on Appraisal and Evaluation, 2022 edition, Annex A6. The declining schedule — 3.5% / 3.0% / 2.5% — operationalises the gamma-discounting argument made formally in Weitzman ML, "Gamma Discounting," American Economic Review 91(1):260–271 (2001). The reader who accepts the Green Book accepts the Weitzman structure.

[⁶]: Belkouteb N et al., "Removal of per- and polyfluoroalkyl substances (PFASs) in a full-scale drinking water treatment plant: Long-term performance of granular activated carbon (GAC) and influence of flow-rate," Water Research 182:115913 (2020). DOI: 10.1016/j.watres.2020.115913. Allocation fractions for textile-origin PFAS in drinking-water loading are not disaggregated in the peer-reviewed literature; the 1–5% band used here is the analyst's assumption, disclosed at the point the numbers appear. A reader who doubles or halves it gets present values that remain under £1 per garment at every defensible discount rate.

2022-01-18 → 2026-04-13

The gap is 1,547 days. What ran inside it.

On 18 January 2022, Mamavation published the activewear investigation.[¹] Four months later, on 4 May 2022, Silent Spring Institute's Rodgers and Swartz paper appeared in Environmental Science & Technology — a peer-reviewed ninety-three-product PFAS survey with zero activewear in the sample and zero named brands.[⁷] The peer-reviewed consumer-testing literature has not caught up with Mamavation to this day.

Legislatures then moved.

California AB 1817 was signed on 29 September 2022, banning PFAS in apparel; it took effect 1 January 2025 at a 100-ppm total-fluorine threshold. New York S6291A was signed on 30 December 2022 with a similar architecture. Minnesota's Amara's Law followed in May 2023. The European Chemicals Agency published the universal PFAS restriction proposal on 7 February 2023 and received more than five thousand six hundred consultation comments over the summer. The EPA finalised the National Primary Drinking Water Regulation on 10 April 2024 — the rule that set the $1.63 billion per year compliance cost on US public water utilities.[⁸] The UK Department for Environment published the National PFAS Plan on 3 February 2026. The ECHA Socio-Economic Analysis Committee adopted its draft opinion on the universal PFAS restriction on 10 March 2026, with textiles in scope.

Every one of those events is an authority-format event. Lululemon's share price moved on none of them.

Neither, it turns out, did Lululemon's disclosure language. The company files a Form 10-K each fiscal year with the Securities and Exchange Commission. The filing contains the company's own account of the material regulatory risks to its business, in its own words, under its own signature. The FY2021 annual report — covering the fiscal year that ended 30 January 2022, twelve days after Mamavation — contained no reference to PFAS, to restricted substances, to named product-chemistry regulation, or to the word "chemicals" in its risk-factor discussion of product regulation.[¹⁷] The FY2022 annual report, covering the fiscal year ending 29 January 2023, carried the identical sentence unchanged. The California AB 1817 signing of 29 September 2022 and the New York S6291A signing of 30 December 2022 had occurred inside that fiscal year. The European Chemicals Agency's universal PFAS restriction proposal was submitted on 13 January 2023, sixteen days before the fiscal year closed. None of those events prompted a single-word addition to the trade-regulations risk factor.

The first insertion arrived in the FY2023 filing — the period ending 28 January 2024, filed with the SEC two months later. Into the sentence that had stood unchanged across two prior annual reports, Lululemon added eleven words: as well as components of our products, including chemicals. That was the extent of the upgrade. No PFAS. No named statute. No restricted-substances-list reference. No Proposition 65. No named chemistry class. One generic phrase inserted into a sentence that had previously concerned the "labelling, distribution, importation, marketing, and sale" of products and now also concerned their components. The FY2024 filing (period ending 2 February 2025) carried that insertion forward and added two sentences — one noting that "our ability to track and respond to regulations may not be sufficient to meet the increased number and complexity of regulations we are subject to globally," and one noting that "changes in consumer perceptions of the components of our products" could result in compliance costs or loss of revenue. Those were the additions. Across four consecutive annual filings spanning the 1,547-day window, the word "PFAS" did not appear. Neither did "polyfluoroalkyl," "perfluorinated," "forever chemical," "Proposition 65," "restricted substances," "AB 1817," "ECHA," or any named apparel-chemistry regulation.[¹⁷]

Read as an evidence trail, the pattern is specific. Peer-reviewed consumer-testing evidence did not move the disclosure. Individual state-legislation signings did not move the disclosure. The ECHA restriction proposal's publication did not move the disclosure. The industrial-source settlements of June 2023 did not move the disclosure. The eleven-word insertion arrived in the March 2024 filing, into a fiscal year during which California's effective date (1 January 2025) and Minnesota's Amara's Law effective date (1 January 2025) lay immediately ahead, the ECHA proposal had accumulated a year of consultation comments, and the EPA's final National Primary Drinking Water Regulation was publicly imminent. The upgrade corresponds to the moment the regulation was about to bite the business — not to the moment the evidence existed.

Read as evidence of mechanism, the pattern names itself. The 1,547 days does not measure the time the market took to read peer-reviewed consumer-journalism evidence. It measures the time between the first publishable signal and the first firm-proximate compliance-bite event that converted it into a tradable-format disclosure — and, at the end of the window, into a firm-specific enforcement-format event that named the garment line directly. The latency is specific: firm-specific enforcement-format latency combined with compliance-bite disclosure-format latency — narrower twice over than the wider claim the earlier draft suggested, and tighter for the narrowing.

The disclosure architecture confirms what the price-action implies. Peer-reviewed consumer-testing evidence does not reach the risk-factor disclosure. Individual state-legislation signings do not reach the risk-factor disclosure. An ECHA proposal submitted for consultation does not reach the risk-factor disclosure. Industrial settlements do not reach the risk-factor disclosure. Only events that are about to cost the business something — a state law's effective date inside the next fiscal year, a federal regulation's final rule about to land, a firm-specific attorney-general enforcement action — reach the disclosure. The instruments of the transaction — the 10-K's language, the share price's movement — track compliance exposure. They do not track persistence. The market prices the shortest-clock instrument it has. It does not price the instrument it does not have. The body walks through the window. The share price does not move, and the 10-K does not name the chemistry, until the window closes at a compliance-bite or firm-specific enforcement event. Four years and 86 days is how long the body waits.

Across those four years and 86 days, the molecule whose discovery in the Align in January 2022 started the clock never once appeared by name in the company's own account of its material risks. It entered the 10-K, when it finally entered, as one word — "chemicals" — inside a sentence about trade regulation.

The US federal agencies that might close this gap have not. The Consumer Product Safety Commission has no PFAS-specific apparel rule. The Environmental Protection Agency has a water rule and a chemical-manufacture rule but no apparel-at-point-of-sale rule. The Food and Drug Administration has no apparel jurisdiction. The Securities and Exchange Commission applies only above a materiality threshold that apparel-sector PFAS has not, until Texas, crossed. The jurisdictional gap is where format-latency lives. On the UK side, the Department for Environment's 2026 plan addresses sources and pathways. It does not address shelves.

The Attorney General of Texas did on 13 April 2026 what no federal agency and no UK agency had done. He named a specific firm and a specific garment line. The market re-priced within hours. This is the market working. It is also the single window through which the harm was, during the preceding four years and 86 days, not priced at all.

[⁷]: Rodgers KM, Swartz CH, Occhialini J, Bassett P, McCurdy S, Schaider LA, "How Well Do Product Labels Indicate the Presence of PFAS in Consumer Items Used by Children and Adolescents?" Environmental Science & Technology 56(10):6294–6304 (2022). DOI: 10.1021/acs.est.1c05175. Ninety-three products; twenty-six apparel items (school uniforms, casual wear, infant wear, face masks, menstrual underwear); zero activewear; zero brand names. Fifty-four of 93 positive above 10 ppm total fluorine (58%).

[⁸]: Environmental Protection Agency, "PFAS National Primary Drinking Water Regulation," final rule 10 April 2024. Agency compliance-cost estimate $1.63 billion per year across US public water systems; industry estimate (AWWA) $3.8 billion per year. Four thousand one hundred to six thousand seven hundred of sixty-six thousand public water systems affected. Five-year compliance window.

[¹⁷]: Lululemon Athletica Inc., Annual Reports on Form 10-K, as filed with the US Securities and Exchange Commission, CIK 0001397187. Filings reviewed: FY2021 (period ending 30 January 2022), accession 0001397187-22-000014; FY2022 (period ending 29 January 2023), accession 0001397187-23-000012; FY2023 (period ending 28 January 2024), accession 0001397187-24-000010; FY2024 (period ending 2 February 2025), accession 0001397187-25-000013. Diff method: full-text extraction of each filing's HTML; keyword search across the terms PFAS, polyfluoroalkyl, perfluorinated, perfluoro-, fluorinated, fluoropolymer, "restricted substances," "Proposition 65," "AB 1817," "S6291," TSCA, "toxic substance," "hazardous chemical," "forever chemical," "chemical management," CPSC, "Consumer Product Safety," ECHA, REACH, "Amara's Law," "product chemistry," "restricted substances list" or "RSL." Across all four filings, zero matches for any PFAS-specific or named-chemistry term were found. The only product-chemistry-adjacent term appearing in the risk-factor discussion of trade regulation was the single phrase "components of our products, including chemicals," first appearing in the FY2023 filing and carried forward to FY2024. Primary sources retrievable from sec.gov/Archives/edgar/data/1397187/ at the accession paths given. Retrieved 21 April 2026.

May 2000

The anchor is twenty-six years old.

On 16 May 2000, 3M announced jointly with the United States Environmental Protection Agency that it would voluntarily phase out perfluorooctanesulfonate — PFOS — from its product lines.[⁹] The products affected included Scotchgard, firefighting foams, coatings for fabrics and leather, certain paper products, and industrial mist suppressants. At the time, 3M was the only US manufacturer of PFOS. The company's own biomonitoring data had found the molecule in human blood globally. No federal regulation required action. 3M withdrew the compound voluntarily.

The announcement was the first moment at which the modern PFAS story had a regulatory clock to measure against.

Twenty-three years after that announcement, 3M settled with US public water utilities for $10.3 billion present value, $12.5 billion nominal, over thirteen years.[¹⁰] Three weeks earlier DuPont, Chemours and Corteva had settled a parallel class for $1.185 billion. The settlements are large and public. They cover industrial-source PFAS — the water systems downstream of AFFF firefighting foam and manufacturing emissions. Apparel-origin PFAS is not in the frame of either agreement.

Twenty-six years after the May 2000 announcement, the European Chemicals Agency's SEAC is finalising the socio-economic analysis for a universal PFAS restriction that would cover textiles. The UK is two to three years behind that timeline. Neither jurisdiction has instrumented testing at the point of sale.

A woman in a Lululemon store in Birmingham the day after the Texas CID was filed held in her hands a garment whose chemistry descends from the molecule 3M had agreed to withdraw twenty-six years earlier. She had no way, from the hanger, to determine whether a descendant of that chemistry was in the fabric. That is not a statement about the diligence of any retailer. It is a statement about the infrastructure at point of sale. Twenty-six years of voluntary phase-outs, settlements, state legislation, and committee opinions have built no instrument that can answer her question while she is deciding whether to buy.

[⁹]: Environmental Protection Agency, "EPA and 3M Announce Phase Out of PFOS," joint press release, 16 May 2000. Archived at epa.gov/archive/epapages. 3M data supplied to EPA indicated PFOS is "very persistent in the environment, have a strong tendency to accumulate in human and animal tissues and could potentially pose a risk to human health and the environment over the long term."

[¹⁰]: 3M Corporate News, "3M Resolves Claims by Public Water Suppliers," 22 June 2023. Final court-approval order entered 29 March 2024, US District Court for the District of South Carolina (Judge Gergel), with 3M's investor press release announcing the approval on the next business day, 1 April 2024. DuPont / Chemours / Corteva: dupont.com press release, 2 June 2023, "Chemours, DuPont, and Corteva Reach Comprehensive PFAS Settlement with U.S. Water Systems."

"Regrettable Substitution"

This is not the first time. That is the scholarly finding.

In 2019, a review in Environmental Science & Technology Letters asked a question in its title: are organophosphate flame retardants a regrettable substitution for the brominated ones they replaced?[¹¹] The answer, across nine authors including Arlene Blum, Heather Stapleton, Linda Birnbaum, and Miriam Diamond, was a qualified yes. The organophosphate esters — TDCPP, TCEP, TPP — inherited the persistence-toxicity architecture of the polybrominated diphenyl ethers they were brought in to replace. House-dust studies through 2009 and 2012 found the replacements at geometric mean concentrations comparable to, and in some cases exceeding, the PBDE levels that had prompted the substitution.[¹²] TDCPP and TCEP are now listed as Proposition 65 carcinogens in California.

The shape of the story is the same as the PFAS shape: a long-persistent chemistry, a replacement introduced without complete structural testing, a slower clock on which the replacement's persistence is discovered, a regulatory catch-up that runs years to decades behind the substitution itself. The flame-retardant timeline in the United States runs from PBDE voluntary phase-out in 2004 through the Proposition 65 listings of TCEP in 2008 and TDCPP in 2011 through the ongoing EU Flame Retardant regulatory cycle. Thirteen years from substitution to first retrospective warning label. The chemistry's clock ran the entire time.

The portability test for the Temporal Asymmetry of Forever runs across four domains.

The last row is the one that matters most for what the pattern is and is not. Where the point-of-sale format is instrumented — where a black-box warning updates under 21 CFR 201.57 when the Food and Drug Administration's surveillance system flags a signal, where adverse events are mandatorily reported through FAERS, where the Sentinel Initiative runs active post-market surveillance on electronic healthcare data — the three clocks converge. The pharmaceutical system has had instrumentation failures at scale (Vioxx, OxyContin, thalidomide under the pre-1962 regime). The failures are in enforcement and physician-level response, not in whether the format at point of sale exists. Where the format exists, the asymmetry does not hold in the 10³-order-of-magnitude form this report is describing.

The asbestos row is the one that tempts the reader to say the system eventually works. The RAND Institute for Civil Justice, working from Tillinghast-Towers Perrin projections, documented total US asbestos-liability costs through 2002 at more than $70 billion in defence-and-insurance spending, with projected eventual totals exceeding $200 billion across the trust-administration architecture of latent-injury tort.[¹³] The system compensated a fraction of the exposed. Most of the compensation reached plaintiffs decades after exposure, some posthumously. Sixty-four years passed between Wagner, Sleggs and Marchand's 1960 documentation of the mesothelioma causation[¹⁴] and the EPA's March 2024 chrysotile ban — the first comprehensive US prohibition under the strengthened Toxic Substances Control Act. The clock bent, over sixty-four years, enough to reach some of the harm. The body-clock ran the entire time.

That is the condition the Temporal Asymmetry of Forever names. It is not that compensation is impossible. It is that compensation, where it arrives, arrives on a clock many decades later than the harm it compensates, and for a minority of those harmed. A system that reaches a fraction of those exposed — estimates in the academic literature range across a third to roughly half — decades after the exposure that caused the harm, is the system the retrospective asbestos record documents. It is not a mechanism capable of pricing the harm at point of sale. It is an extraction mechanism whose temporal gearing is incompatible with the time-constant of the persistence it is trying to reach.

The discount-rate critique of long-horizon environmental harm has prior art — Weitzman's gamma-discounting formalism, the thirteen-author Science consensus, the declining schedule the UK Green Book now encodes in Annex A6.[¹⁵] That literature tells what the discount rate should be; it does not tell whether the future is reachable. δ controls how much a future harm counts; the institutional envelope controls whether the future is reachable at all. No δ, however low, rescues a harm clock that sits beyond the recovery clock's horizon. The asymmetry is not an under-pricing problem. It is an envelope-termination problem — and that is where the argument departs.

[¹¹]: Blum A, Behl M, Birnbaum LS, Diamond ML, Phillips A, Singla V, Sipes NS, Stapleton HM, Venier M, "Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers?" Environmental Science & Technology Letters 6(11):638–649 (2019). DOI: 10.1021/acs.estlett.9b00582.

[¹²]: Stapleton HM, Klosterhaus S, Eagle S et al., "Detection of Organophosphate Flame Retardants in Furniture Foam and U.S. House Dust," Environmental Science & Technology 43(19):7490–7495 (2009). Dodson RE, Perovich LJ, Covaci A et al., "After the PBDE Phase-Out: A Broad Suite of Flame Retardants in Repeat House Dust Samples from California," Environmental Science & Technology (2012).

[¹³]: Carroll SJ, Hensler D, Gross J et al., Asbestos Litigation, RAND Institute for Civil Justice MG-162-ICJ (2005), drawing on Tillinghast-Towers Perrin projections: defendants and insurers had spent more than $70 billion on asbestos litigation through 2002, with projected eventual totals exceeding $200 billion. US Government Accountability Office, "Asbestos Injury Compensation: The Role and Administration of Asbestos Trusts" GAO-11-819 (2011): over 100 asbestos trusts; aggregate trust assets approximately $36 billion; approximately $17.5 billion paid out to claimants through 2010.

[¹⁴]: Wagner JC, Sleggs CA, Marchand P, "Diffuse pleural mesothelioma and asbestos exposure in the North Western Cape Province," British Journal of Industrial Medicine 17(4):260–271 (1960).

[¹⁵]: Weitzman ML, "Gamma Discounting," American Economic Review 91(1):260–271 (2001). Arrow KJ, Cropper ML, Gollier C, Groom B, Heal GM, Newell RG, Nordhaus WD, Pindyck RS, Pizer WA, Portney PR, Sterner T, Tol RSJ, Weitzman ML, "Determining Benefits and Costs for Future Generations," Science 341(6144):349–350 (2013). Gollier C, Pricing the Planet's Future: The Economics of Discounting in an Uncertain World, Princeton University Press (2012).

Margin as Verification Sorting

One more piece of arithmetic, because it changes what "clean premium" means.

An EPA Method 1633 test — the forty-analyte panel used for PFAS characterisation in non-drinking-water matrices — costs $835 at a commercial consumer-facing laboratory. Textile-matrix analysis, which requires additional extraction, runs $800 to $2,000 per stock-keeping unit. A single SKU sold in 500,000 units at £60 retail generates gross profit of £13.5 million at a 45% margin and £18 million at a 60% margin. Testing cost, even at the high end — $2,000 for a one-time panel per SKU — is between 0.008% and 0.011% of gross profit at either tier.

This is worth saying slowly. The "clean-premium" retailer and the mid-market retailer face essentially the same testing-cost burden relative to what they earn from the SKU. A brand operating at 45% gross margin can absorb a $2,000 test as readily as a brand operating at 60% gross margin. The arithmetic does not sort by margin capacity. It sorts by something else.

The something else is capital-allocation priority in the absence of a verification requirement. No jurisdiction requires a per-SKU third-party PFAS test at point of sale. No retailer faces a disclosure rule that names the instrument it did or did not run. Every brand chooses, against other calls on capital, whether to commission testing it is not obliged to commission. The verification-absent default — the state of a retailer that has not tested — persists wherever nobody is required to test.

The consequence at the hanger is what it was: a premium segment that tests some SKUs at a frequency it discloses in general terms; a mid-market segment that tests at a lower frequency and discloses less; a mass segment that typically does not disclose. No brand need be named for the structural finding to land. The peer-reviewed consumer-testing corpus through 2022 does not disaggregate to mid-market activewear SKUs.[⁷] The Mamavation testing hit the Align because it hit the premium segment; no equivalent study has indexed the mid-market or mass tier. The silence of the corpus is not a gap to close and file. It is the format-latency argument repeated at the level of the academic literature: the evidence does not exist because no format has been instrumented to produce it.

A reader who wants to name the brands sorted by this pattern will find, on inspection, that she has more rhetorical material than she has public-record evidence. The report does not name them. The structural consequence — that the "PFAS-free" label is uniform across the segment while the investment behind the phrase tracks the rate at which disclosure is required — holds on the data that are public. It does not require additional names to hold.

The Steelman

The best version of the objection is this.

The pattern this report describes is real. The chemistry-clock does run long; the recovery-clock does run short; the ratio is not controversial. But the gap has been named, in economic scholarship, for decades under the existing vocabulary: long-tail externality theory, gamma discounting, latent-injury tort, declining social discount rates, the Green Book's own schedule. The Temporal Asymmetry of Forever is a good name for a known problem. It is not a new mechanism.

The apparel-PFAS instance is a textbook case of Bayesian efficient-markets updating. Low-probability industry-wide signals (Mamavation, Silent Spring's peer-reviewed adjacency, the ECHA consultation) remain priced into industry-wide risk premia until firm-specific enforcement crystallises (Texas, 13 April 2026), at which point the specific firm re-prices. The 1,547-day gap is not format-latency as a structural finding; it is the ordinary interval between diffuse signals and firm-specific news. The market is functioning as the event-study literature says it should.

And the premium the consumer pays for a "PFAS-free" legging is doing what premiums usually do. It is pricing claim-survivability — the brand's investment in the kind of legal-compliance verification that insures against the very state-AG enforcement the Texas CID represents. The consumer is paying for protection from a specific class of risk. The market is calibrated to that risk. The harm-clock mismatch is real but is the sort of long-horizon uncertainty no existing instrument could price better than the existing instruments do. Calling the gap "Temporal Asymmetry of Forever" re-dresses the problem. It does not re-tool the instruments.

Grant the steelman in its fullest form. Grant that the pattern is named in prior art; that Bayesian updating explains the 1,547 days; that the £30 premium is reasonable claim-survivability insurance. Then ask what the reader walks out of the shop with. She walks out with a pair of leggings whose chemistry will outlive her, having paid a premium that priced something — just not the thing the label appeared to describe. The price was honest. The label was not, at the till, falsified. But the thing the label gestured at — the "forever" chemistry — is not what the premium was ever priced against. It was priced against the short-horizon risk the market can see.

That finding — that the premium bought verification-against-enforcement-risk rather than verification-of-chemistry — is not one the discount-rate literature makes. The discount-rate literature asks what weight to give to future harms. It does not ask what the consumer is told she is buying when she cannot, at point of sale, distinguish the two. That question is downstream of the economics. It is about the format of the transaction, not the mathematics of the discount. And the answer, from the hanger, is that the consumer is sold a premium for something the price signal cannot in principle verify.

The prior-art literature does not stand at the hanger. That is the extra sentence the new coinage earns.

The Hanger

Return to the Align Highrise Pant, because the spine object is the point.

On the day the Texas Attorney General filed the CID, the pant was for sale at lululemon.co.uk at its established retail price. The search bar on the product page returned, for the query "PFAS," the corporate statement the company had used for several fiscal years: that it does not use PFAS in its products today, with a hedging clause about unintentional reintroduction. The same pant had been tested at 32 ppm organic fluorine by an EPA-certified laboratory four years and 86 days earlier. Lululemon's own Restricted Substances List, revision 9.0, set a 50-ppm total-organic-fluorine threshold.[¹⁶] The garment cleared the threshold the brand set. It carried the organic-fluorine signal the Mamavation lab reported. The information at point of sale did not let the consumer tell those facts apart.

The Temporal Asymmetry of Forever does not describe a market failure. It describes a market doing what markets do, with an information infrastructure that measures what it can measure, against a harm whose clock the infrastructure cannot reach. The £30 premium is not a deception. The consumer is not a dupe. The three clocks do not converge, and none of the actors in the transaction — brand, lab, regulator, consumer — is positioned, under current instrumentation, to make them. The reformulation bill arrives at the till. The harm bill cannot.

On the hanger, the pant still hangs. The chemistry outlives the receipt. The receipt outlives the tort window. The tort window outlives the news cycle. And the news cycle has already moved on.

[¹⁶]: Lululemon Restricted Substances List v9.0 — 50 ppm total organic fluorine threshold per Pillsbury PFAS Observer analysis (17 April 2026).

This report was produced by YAN Consumer Intelligence using AI analytical systems that cross-reference regulatory frameworks, scientific evidence, and publicly observable product information to identify information gaps. Where organisations or products are referenced, they appear as representative examples of systemic patterns -- not as subjects of allegation. Our analysis examines what standards and regulations require, not whether any individual organisation has acted wrongfully. Organisations referenced in this report are invited to provide additional context or corrections at [email protected]. See our Methodology for full details of our analytical process and corrections policy.