ReThink History: The Payor of Last Resort

Let's ReThink History
Let's ReThink History

A federal statute requires any hospital with an emergency department to examine and stabilise anyone who walks through the door, regardless of ability to pay. It appropriates no money. The physician on call is obliged to come, and obliged to work for nothing if the patient does not pay. That statute is the plainest statement in American law of a bargain the country made and has never revoked: when no one else pays, society does. Twelve episodes walk that bargain with dates and sources — the poor rate of 1601 and the workhouse test of 1834; the Delaware hospital that turned away a four-month-old in 1959 and owed him nothing; the 467 patients transferred to a Chicago public hospital where lack of insurance was the stated reason in most cases; the federal court that wrote that society picks the person up off the highway, and the footnote in which it refused to go further; and Buck v. Bell, where the same reasoning about what people cost society, with nothing to bound it, produced a sterilisation order that has never been overruled. Then the harder half. Clean water, childhood vaccination and taking the lead out of gasoline were bought upstream and repaid many times over. Most prevention does not: fewer than one preventive measure in five is cost-saving, a population that quits smoking costs more over a lifetime because it lives longer, and thirty-four Medicare programmes produced no net savings between them. The series proves that against its own thesis, then follows the same structure into the present — the tobacco settlement that won outright and delivered hundreds of billions, of which three and a half percent now reaches prevention; the anti-obesity drugs scored in the red; the California wildfire pool holding seven hundred and sixty-eight billion dollars of exposure; a forty-thousand-dollar retrofit that earns a two-hundred-and-sixteen-dollar discount; and the smoke that kills tens of thousands a year and has no owner to bill. The verdict is not that prevention pays for itself. It is that we already decided we would catch people, and the only thing left to choose is when we pay. Compiled by a practising anesthesiologist bound by law and conscience to treat anyone who arrives in an emergency department needing care, and to do it for free if they refuse to pay. No patient, case or clinical detail appears anywhere in this series, and nothing in it is medical advice. Every cost-offset claim is presented with its rebuttal and its date; pending litigation is described in its procedural posture only. AI-narrated from a fully cited research dossier for every episode.

  1. Episode 1

    Ep 1 — The Man Who Cannot Say No

    On April 7, 1986, a budget reconciliation bill was signed with a short provision buried at section 9121. It gave anyone who comes through an emergency department door a legal right to be examined and stabilized — with no reference to money, insurance, citizenship or desert — and it appropriated nothing. EMTALA, 42 U.S.C. § 1395dd, is the only universal, insurance-independent entitlement to care in United States law, and in forty years Congress has never gone back and written down who pays for it. Episode 1 walks the statute clause by clause — the screening duty, the duty to stabilize, the transfer rules, the obligation of a hospital with specialized capability to accept a transfer, and the sentence forbidding any delay in screening in order to ask about method of payment. Then it follows how a law aimed at institutions reaches one person by name: the on-call list a hospital must maintain as a condition of its Medicare provider agreement (42 C.F.R. § 489.20(r)(2)), and a civil money penalty assessable against a responsible physician personally — set at $136,886 per violation by the HHS inflation adjustment published January 28, 2026, and not covered by malpractice insurance. CMS's rule of September 9, 2003 (68 Fed. Reg. 53222, effective November 10, 2003) confirmed that the hospital decides how deep its call panel goes; the exposure begins the moment a name goes on the list. The problem Congress was answering was counted, not just asserted. Himmelstein and colleagues tracked 458 consecutive transfers out of 14 private hospitals into one public emergency room in Alameda County over six months in 1981, pulled 103 high-risk charts, and found substandard care in 33 of them; Schiff and colleagues published 467 consecutive transfers to Cook County Hospital in the New England Journal of Medicine on February 27, 1986. The episode then gives the strongest case against. David Hyman's 1998 critique argues the legislative record rested on anecdote — "the plural of anecdote is not data" — that the famous quarter-million-a-year figure was an extrapolation from three cities, and that the emblem case circulated in the hearings did not survive sworn testimony. A senator said on the record before passage that the evidence he had seen was primarily anecdotal. A good law can rest on a bad evidentiary record. Where it lands: EMTALA is a sequencing rule, not a payment rule. It moves the money question from before the care to after it, and the patient still owes the bill. Enforcement has always been thin — GAO-01-747 (2001) found violations, fines and terminations all rare, and the pattern has held at roughly twenty penalty settlements a year. The only federal money ever earmarked against the obligation was $250 million a year for fiscal 2005 through 2008 under § 1011 of the Medicare Modernization Act; it expired and was not renewed. Current preemption litigation is described here in procedural posture only: as of late August 2026 there is no Supreme Court merits ruling. Episode 2 goes back to where the obligation to catch people was first written as a tax. Sources include 42 U.S.C. § 1395dd and 42 C.F.R. §§ 489.20 and 489.24; the CMS final rule at 68 Fed. Reg. 53222 and the HHS penalty adjustment of January 28, 2026; GAO-01-747 and GAO-04-472; Wilmington General Hospital v. Manlove (Del. 1961) and Roberts v. Galen of Virginia (1999); and the published literature on both sides — Himmelstein et al. (AJPH, 1984), Schiff et al. (NEJM, 1986), Hyman (Health Matrix, 1998) and Sawyer (West J Emerg Med, 2017). ▶ Watch on YouTube: https://www.youtube.com/watch?v=HvrdFkkM_JA 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  2. Episode 2

    Ep 2 — The Parish and the Poor Rate

    In 1601 an English parliament did something no state had done before: it made relief of the poor a compulsory tax. The Poor Relief Act (43 Eliz. 1 c. 2) ordered churchwardens and "substantial householders" to raise a weekly rate on every inhabitant and occupier of lands, houses and tithes — a levy assessed on property, collected by neighbours, spent on strangers. That is the founding document of the payor of last resort, and the same parliaments wrote the other half of the pattern in the same breath. Episode 2 follows both halves for three and a half centuries. The Vagabonds Act of 1597 punished the wandering poor and conveyed them back to their parish of origin. The Settlement Act of 1662 turned that into a fiscal instrument: within forty days of a newcomer settling in a tenement under £10 yearly value, two justices could remove anyone "likely to be chargeable" — exclusion as a forecasting exercise, before any relief was ever claimed. The Royal Commission appointed on 1 February 1832, chaired by Blomfield with Nassau Senior a member and Edwin Chadwick its secretary, produced the principle of less eligibility. Note what the Poor Law Amendment Act of 14 August 1834 actually did: it created unions, elected guardians and a central commission — it never enacted less eligibility or the workhouse test at all. Those were commission policy. Did austerity work? On its own terms, yes, and the companion figure is the one never quoted beside it. Real relief per head fell about 43% between 1831 and 1841 (index 107.9 to 61.8), from roughly 2.0% to 1.12% of national income — while the recorded number of persons relieved rose from about 1.04 million to about 1.30 million. More people caught, for much less money each. The system crossed the Atlantic intact: local fiscal responsibility, settlement and removal, and the moral triage of the worthy and unworthy poor. On 3 May 1854 President Pierce vetoed the 10,000,000-acre land grant for the indigent insane, writing that he could find no authority to make the federal government "the great almoner of public charity" — and naming the sorting problem exactly: if Congress can provide for the indigent insane, it has the same power to provide for the indigent who are not. Then the federal turn, in instalments — Title VI of the Social Security Act (1935), Hill-Burton (1946) with its twenty-year uncompensated-services clock and its separate-but-equal clause, struck down in Simkins v. Moses H. Cone (4th Cir., en banc, 1 November 1963), and Medicare and Medicaid on 30 July 1965 — alongside the quieter bargain of Revenue Ruling 69-545, whose first community-benefit factor is an emergency room open to all regardless of ability to pay. Meanwhile the courts killed the removal power: Edwards v. California (1941), Shapiro v. Thompson (1969), which traced residence tests by name to the 1662 statute, and Memorial Hospital v. Maricopa County (1974). But in 1961 a Delaware court still held that a private hospital "owes the public no duty to accept any patient not desired by it," expressly refusing to convert public grants and tax exemption into a duty to admit. In 1986 the duty finally attached with no settlement test and nowhere to send anybody. The sorting did not disappear. It moved into the ledger — and Episode 3 asks what that ledger is allowed to justify. Sources include the Poor Relief Act 1601, the Vagabonds Act 1597, the Settlement Act 1662 and the Poor Law Amendment Act 1834; the Poor Law Commissioners' Report of 1834, Part II § 1; Pierce's veto message of 3 May 1854; Wilmington General Hospital v. Manlove, Simkins v. Moses H. Cone, Edwards, Shapiro and Memorial Hospital v. Maricopa County; Revenue Rulings 56-185, 69-545 and 83-157 with IRC § 501(r); and George Boyer's EH.net relief series with Mark Blaug on the commission's methods. ▶ Watch on YouTube: https://www.youtube.com/watch?v=qo-EWnubSTM 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  3. Episode 3

    Ep 3 — Society Picks Him Up Off the Highway

    On June 28, 1972, a three-judge federal court in Boston upheld the Massachusetts motorcycle helmet law and wrote the payor-of-last-resort argument into constitutional law in one unbroken sentence: from the moment of the injury, society picks the person up off the highway, delivers him to a municipal hospital and municipal doctors, provides unemployment compensation, and may assume responsibility for his family's continued subsistence. It is a chain of custody, and every link is a real institution with a real budget line. Simon v. Sargent, 346 F. Supp. 277, was summarily affirmed at 409 U.S. 1020, which is why federal helmet challenges have gone nowhere for fifty years. Episode 3 is about the part nobody quotes. In a footnote, the same court refused the broader version of its own argument: its decision did "not rest broadly, as have some others, on the state's generalized assertion of an interest in the continued productivity of its citizenry." It rested on direct outlays only — ambulance, hospital, doctors, unemployment, disability. That is the first limiting principle anyone in this tradition drew and wrote down. The episode also fixes the other side of the docket: American Motorcycle Ass'n v. Davids (Mich. 1968) rejected the public-charge rationale as one that "could lead to unlimited paternalism," while People v. Fries (Ill. 1969) is weaker counter-authority than it is usually made — it rejects self-protection as a police-power basis and never reaches public cost at all. Take the affirmative case at full strength. NHTSA's own tally (Kahane, DOT HS 812 069, January 2015) credits vehicle safety technologies and the standards behind them with 613,501 lives saved from 1960 through 2012 — 329,715 from seat belts alone — while the occupant fatality rate per vehicle mile fell 81 percent. CDC's MMWR of June 15, 2012 put helmet use at $3 billion in costs saved and 1,544 lives in 2010, with $1.4 billion and 709 lives still available. GAO's 1991 audit found helmet laws cut fatality rates 20 to 40 percent. Then set two federal ledgers side by side: that $3 billion is a comprehensive societal figure, dominated by lost productivity, while NHTSA's state-level estimates of what a treasury actually saves from a primary seat belt law run to $2.3 million in Florida, $1.2 million in Colorado, $374,000 net in Rhode Island. Three orders of magnitude apart — and the productivity component that makes the big number big is precisely the interest the Simon footnote refused. Then the crack. Joel Feinberg's objection is that the cost is a public cost only because we already decided to insure one another; the externality is constructed by our own prior commitment, not discovered. Peter de Marneffe (Philosophy & Public Affairs, 2006) presses it into a reductio: if externalized cost is really the reason, the consistent and far less intrusive remedy is to stop paying. No legislature will discuss it. Devin Christensen (Political Research Quarterly, 2022) adds that framing accidents by their cost to everyone else encourages victim-blaming and corrodes the solidarity the argument depends on. The consistency test is on the record: California requires helmets of every motorcycle rider (§ 27803, effective January 1, 1992) and of class 3 e-bike riders (§ 21213), but § 21212 reaches cyclists under 18 only — and as of late August 2026 no American state requires adult cyclists to wear helmets statewide, though 17 states and D.C. have universal motorcycle laws. Same head, same trauma bay, same claim. What survives is narrower and steadier: given a promise we are not going to break, buying upstream is the cheaper way to honor it — a claim that does not require prevention to pay for itself, and that comes with a rule the fiscal version lacks. Act on the hazard, not on the person. Episode 4 is what happens when nobody applies it. Sources include Simon v. Sargent (D. Mass. 1972), aff'd 409 U.S. 1020, with American Motorcycle Ass'n v. Davids, People v. Fries, State v. Fetterly and Picou v. Gillum; Kahane's DOT HS 812 069 (2015), CDC MMWR 61(23):425, NHTSA's state seat-belt Medicaid assessments and GAO RCED-91-170; California Vehicle Code §§ 21212, 21213 and 27803; and Feinberg's Harm to Self with de Marneffe (2006) and Christensen (2022). ▶ Watch on YouTube: https://www.youtube.com/watch?v=jKzxaJMYQTs 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  4. Episode 4

    Ep 4 — Three Generations

    On February 20, 1905, the Supreme Court upheld a five-dollar fine for refusing a smallpox vaccination, and Justice Harlan wrote four limits into the opinion that upheld it. On May 2, 1927, the same Court cited that case — in nine words, in an opinion barely a thousand words long — to uphold the surgical sterilization of a teenage girl in Virginia. None of the four limits came with it. Buck v. Bell, 274 U.S. 200, is the American law of what people cost society argued with no limiting principle at all, and it has never been overruled. Episode 4 sets Jacobson v. Massachusetts against Buck v. Bell and reads what each opinion actually says. Virginia's Eugenical Sterilization Act and its Racial Integrity Act were signed on the same day, March 20, 1924. Carrie Buck entered the Virginia State Colony that June; the trial in Amherst County was held on November 18, 1924, judgment entered April 13, 1925, affirmance by the Supreme Court of Appeals of Virginia on November 12, 1925. Holmes's opinion accepted three generations of findings in three sentences, without examining how any of them had been reached, and reasoned from conscription: the public welfare may call upon the best citizens for their lives, so it may call upon "those who already sap the strength of the State" for lesser sacrifices. That is a fiscal characterization of a person — not of a hazard, not of a condition. Judgment affirmed, eight to one; Justice Pierce Butler dissented without writing a word. What the record shows now was largely absent from the courtroom then. Her court-appointed guardian's lawyer had sat on the institution's board and called no witnesses. When K. Ray Nelson went looking for surviving patients from 1979, he found Carrie Buck alive near Charlottesville — clinicians who examined her found no mental illness and no intellectual disability — and her sister Doris, sterilized at sixteen in 1928 and told it was an appendectomy. Vivian Buck's school records, published by Stephen Jay Gould in 1984, show her on the honor roll in April 1931; she died of enterocolitis on July 3, 1932, aged eight. She is the third generation. Carrie Buck died on January 28, 1983. Virginia repealed the Act in 1974, passed a resolution of regret on February 7, 2002, apologized through Governor Mark Warner on the seventy-fifth anniversary of the decision, and in February 2015 established compensation of up to $25,000 for each living victim. What the decision authorized was counted by the movement itself: the Human Betterment Foundation's own published tally stood at 27,869 sterilizations in the United States as of January 1, 1938, cumulative since Indiana's first statute in 1907; roughly 8,300 people were sterilized in Virginia between 1927 and 1972. Where it lands: Skinner v. Oklahoma (June 1, 1942) struck a sterilization statute, but on a classification — grand larceny in, embezzlement out — and distinguished Buck v. Bell rather than disturbing it. Roe v. Wade cited it at 410 U.S. 154 on January 22, 1973, without qualification, as ordinary authority that privacy is not unlimited. The only doctrinal movement has been on the other case: in Roman Catholic Diocese of Brooklyn v. Cuomo (5–4 per curiam, November 25, 2020) Justice Gorsuch read Jacobson back down to its own facts. The tools to bound the argument existed before it failed — Jacobson's four limits, and the unequal-hand standard of Yick Wo applied in Jew Ho v. Williamson in 1900 — which means the failure was chosen. This episode does not argue the country should stop buying upstream; Episode 5 is the record of what upstream purchases bought. It sets a burden of proof on everything after it: an argument from what people cost society is only as good as its limiting principle. Sources include Jacobson v. Massachusetts (1905), Buck v. Bell (1927), Skinner v. Oklahoma (1942), Roe v. Wade at 410 U.S. 154, Roman Catholic Diocese of Brooklyn v. Cuomo (2020), Jew Ho v. Williamson (1900) and Yick Wo v. Hopkins (1886); Virginia's 1924 statutes and the Amherst County record; Paul Lombardo's "Three Generations, No Imbeciles" (NYU Law Review, 1985) and his 2008 book; Stephen Jay Gould's "Carrie Buck's Daughter" (1984); and the Human Betterment Foundation's Human Sterilization Today (1938). ▶ Watch on YouTube: https://www.youtube.com/watch?v=6N0J894LYkk 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  5. Episode 5

    Ep 5 — The Great Purchases

    On September 26, 1908, a New Jersey physician named John L. Leal began feeding chloride of lime into forty million gallons a day of Jersey City's drinking water — while a court was still deciding whether he was allowed to. That is where the American case for buying upstream begins, and this episode is the series making its own best case: clean water, childhood vaccination and unleaded gasoline, the three times public health bought the hazard and the arithmetic came back overwhelming. Episode 5 walks each purchase with its dates. Hiram Mills's slow sand filter at Lawrence, Massachusetts in September 1893; the Boonton chlorination plant built in ninety-nine days and ruled lawful by special master William Magie on May 9, 1910. Cutler and Miller's 2005 Demography paper on thirteen cities, 1900–1936 — roughly half the total mortality decline, and a social rate of return above 23 to 1. The measles resurgence of 1989–91 — about 55,000 cases and 123 deaths, mostly among children whose families could not pay for a dose — and the Vaccines for Children program, operational October 1, 1994, which bought not vaccination but its price. And EPA's February 1985 regulatory impact analysis of the leaded-gasoline phasedown: $7,821 million in monetized 1986 benefits against $608 million in refining costs. Then every headline number gets the correction published against it, on the record and with its date. In April 2022 Anderson, Charles and Rees called the total-mortality result "extremely fragile"; Cutler and Miller conceded "unambiguous data mistakes" and revised filtration's share of the total mortality decline from 43% to 38% — the intervention still worked, the number moved, and the 23-to-1 return has never been recomputed at the revised effect sizes. CDC's August 8, 2024 accounting of 117 million children — 508 million illnesses and 1,129,000 deaths prevented, 3.3 to 1 direct and 10.9 to 1 societal — rests on modelled counterfactuals, and the 10.9 counts productivity that arrives in no treasury. The largest monetized benefit of taking lead out of gasoline was not children's brains but blood pressure in men aged 40–59, at $5,897 million. And Milwaukee, spring 1993 — roughly 403,000 people ill and at least 69 dead from a filtration plant that was operating and failing at once — shows that a great purchase can be un-bought by deferred maintenance. What survives is not a slogan but a diagnostic: these worked because the intervention was cheap against the harm, the hazard could be counted, there was a single decision point, and nobody had to be persuaded one household at a time. Every one acted on the water, the fuel or the price of a dose — on the hazard, never on a person. And where the fourth condition is weakest, vaccination, the purchase is visibly eroding: kindergarten MMR coverage at 92.5% for 2024–25 and 2,777 confirmed measles cases as of August 20, 2026, with nothing repealed. These are the exceptions, not the rule, and the episode says so — Episode 6 breaks the claim that prevention pays for itself, and nothing here leans on it. Sources include Cutler and Miller in Demography 42(1) (2005) and the full 2022 exchange in AEJ: Applied 14(2); Zhou et al., MMWR 73(31), August 8, 2024; CDC coverage and measles surveillance data; EPA's Costs and Benefits of Reducing Lead in Gasoline (EPA-230-05-85-006, February 1985) read through Newell and Rogers (RFF, 2003); and the Jersey City chlorination record of 1906–1910. ▶ Watch on YouTube: https://www.youtube.com/watch?v=T3vxQ_mKUSA 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  6. Episode 6

    Ep 6 — The Ledger That Doesn't Balance

    On October 9, 1997 the New England Journal of Medicine published a Dutch finding that has never been withdrawn: a population in which nobody smoked would cost about 7% more for men and 4% more for women over a lifetime, because people who do not die at sixty-two live to seventy-nine and acquire something else. This is the episode where the series breaks the most popular sentence in American health policy — that preventive care pays for itself — using the health economists who measure prevention for a living and mostly believe in it. Episode 6 puts the fiscal claim on trial with the record. Cohen, Neumann and Weinstein, NEJM, February 14, 2008: drawing on roughly six hundred cost-effectiveness studies published 2000–2005, fewer than 20% of preventive interventions reduced medical spending — about the same share as treatments. The Congressional Budget Office's issue brief of January 18, 2012 reviewed ten Medicare demonstrations comprising 34 programs and found that in nearly every one, spending was unchanged or increased once the fees paid to participating organizations were counted; the one clear success was a bundled payment for coronary bypass, at about 10% saved. And two beloved American prevention programs came back at zero and below zero: D.A.R.E., weighted mean effect on drug-use behaviour of 0.06 (AJPH, September 1994), and the prison-visit deterrence programs, pooled odds of subsequent offending 1.68 to 1.72 across nine trials (Campbell Systematic Reviews, 2013). The counterweight is given at full strength. The death-offset result is real and model-dependent at once — Barendregt's team counted a short list of smoking-attributable diseases on a 1997 Dutch cost structure, and broader modelling such as the German EstSmoke work in Addiction (2017) moves it materially. Barendregt himself said the finding "ought to be of very limited importance." Braveman and Gottlieb's 2014 statement of the social determinants of health is careful and never claims savings; Chris Pope's July 2024 critique lands hardest on those who cited them, on the misread "clinical care is 20% of health" variance decomposition, and on the January 9, 2020 NEJM randomization of Camden's hotspotting model, where 180-day readmissions were 62.3% against 61.7%. Camden tested one program in one city on one very ill population — a null on readmissions is not a null on quality of life. And in the international check, the preventive share sits in low single digits under every financing architecture: 2.9% of US health spending in 2018 against a 2.4% peer average, and 3.7% of EU current health expenditure in 2023. What survives is narrower and harder to dismiss. Cost-effectiveness is untouched — the right question is which specific measures are worth buying at what price, and a great deal of prevention is an excellent purchase that saves nobody a nickel. The commitment is untouched too: the emergency-care duty appropriates nothing and was never justified by a spending projection. So the series relocates its argument for good — not whether prevention turns a profit, but at which point in the sequence a promise we are going to keep anyway is cheapest to honour — and it takes from this hour the limiting principle the rest of the series runs on: the frame that asks what a hazard costs survives; the frame that asks what a person costs produced Buck v. Bell. Episode 7 watches what happens when that argument wins in court and the money lands in a general fund. Sources include Cohen, Neumann and Weinstein, NEJM 358:661–663 (2008); Barendregt, Bonneux and van der Maas, NEJM 337:1052 (1997); van Baal et al., PLoS Medicine 5(2):e29 (2008); CBO's Lessons from Medicare's Demonstration Projects (January 18, 2012); Ennett et al., AJPH 84(9) (1994); Petrosino et al., Campbell Systematic Reviews 9(1) (2013); Finkelstein et al., NEJM 382:152–162 (2020); Braveman and Gottlieb, Public Health Reports 129 (2014); and expenditure series from Peterson-KFF and Eurostat. ▶ Watch on YouTube: https://www.youtube.com/watch?v=8a3M4hLo1a8 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  7. Episode 7

    Ep 7 — What the Settlement Bought

    On November 23, 1998, forty-six states, the District of Columbia and five territories signed the Master Settlement Agreement with four tobacco manufacturers: roughly $206 billion over twenty-five years, with payments continuing in perpetuity. The agreement contains no provision requiring a single dollar of it to be spent on tobacco. For fiscal 2025 the states will collect $22.1 billion from settlement payments and tobacco taxes and spend $764.8 million — 3.5% — on prevention and cessation. Episode 7 traces the theory that made the case winnable and what happened to the money. Mike Lewis's 1993 idea — sue for the state, not the smoker, because the state cannot be blamed for its own disease — filed by Mississippi Attorney General Mike Moore on May 23, 1994 for roughly $940 million. Four states settled alone before the national deal: Mississippi at $3.4 billion, Florida at $11.3 billion, Texas at roughly fifteen billion, and Minnesota at $6.1 billion after the trial that forced millions of pages of industry documents into public view — about $40 billion on top of the MSA's $206 billion, and neither figure fixed, because volume, inflation and non-participating-manufacturer adjustments move the payments every year. Read the MSA's recitals and the Medicaid theory is not in them: they rest on consumer-protection and antitrust law, youth smoking and public-health funding. The suits were cost recovery; the agreement was not. Then GAO's own surveys: 7% to tobacco control in the first two years, and by fiscal 2004, health 39.6%, debt service on securitized funds 17.9%, general purposes 13.3%, tobacco control 3.9%. ProPublica's August 7, 2014 investigation found more than a hundred securitization deals, $36 billion of bonds outstanding, and roughly $3 billion borrowed through capital appreciation bonds against a promise to repay about $64 billion. The states have a real answer and it gets full weight here. The money was damages, not a grant, and nobody supervises a plaintiff's judgment; no legislature can bind its successors; the largest single category was health care, which is exactly what the suits were brought to recover; and earmarking inside a general fund substitutes as often as it adds. Two genuine purchases belong on the ledger as well: the truth campaign, funded by the one slice the agreement actually directed, estimated to have prevented about 450,000 adolescents from starting to smoke between 2000 and 2004; and the broader tobacco-control record — about 8 million premature deaths avoided and 157 million life-years gained, 1964–2012 — most of which predates the settlement entirely and cannot be put on its tab. And the boundary Episode 6 set holds here: none of this shows prevention pays for itself; the Farrelly figure is a cost per averted initiation and the Holford figure is lives, not dollars. Where it lands is a design lesson, not cynicism. The theory won, comprehensively, twice; the advertising restrictions are permanent; the money is real and arrives every year. But winning an argument produces an appropriation, and an appropriation is not a purchase. The opioid settlements were drafted by people who had read the first draft — at least 85% required to go to abatement, more than $54 billion in total — and KFF Health News' national accounting of December 16, 2024 found roughly a third of the first $6 billion spent, a third reserved, a third untrackable for want of any public report, with more than $240 million spent on purposes that do not qualify as remediation. Three findings carry forward: the earmark is not a formality, reporting may matter more than restriction, and do not sell the stream. Episode 8 takes the same question to a drug on the market right now. Sources include the Master Settlement Agreement itself (recitals, §§ VI, VIII and IX) via the Public Health Law Center; GAO-01-851, GAO-05-312 and GAO-07-534T; Holford et al., JAMA (January 8, 2014) and the 2014 Surgeon General's report; Farrelly et al., AJPM (February 2009); Campaign for Tobacco-Free Kids, Broken Promises to Our Children (December 18, 2024); Cezary Podkul's ProPublica/Marketplace securitization investigation (August 7, 2014); and Aneri Pattani's KFF Health News opioid settlement database. ▶ Watch on YouTube: https://www.youtube.com/watch?v=ENemyU2Hj_4 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  8. Episode 8

    Ep 8 — The Drug That Might Pay For Itself

    On October 8, 2024 the Congressional Budget Office scored the strongest case anyone has that a downstream payer should buy something upstream: Medicare covering anti-obesity medications would add about $35.5 billion to deficits between 2026 and 2034 — $38.8 billion in drug spending against $3.4 billion in offsetting health savings, an offset of roughly nine cents on the dollar. This episode is about who pays. It gives no medical advice, takes no view on whether anyone should take these drugs, and names no brand. Episode 8 keeps two ledgers apart, because most public argument collapses them. The federal cash ledger asks whether the Treasury gets its money back inside a budget window; CBO found savings per user rising twenty-fold across nine years and still an order of magnitude short of cost per user at every point. An independent University of Chicago model published in JAMA Health Forum in 2025 reached a threefold-higher offset rate — $65.9 billion in drug spending against $18.2 billion in savings over ten years — and still had costs exceeding savings at twenty and thirty years. The value ledger asks what a year of healthy life costs: the same group's March 14, 2025 analysis put semaglutide at $467,676 and tirzepatide at $197,023 per quality-adjusted life year against a conventional $100,000 bar, and calculated the price cuts that would clear it — 81.9% and 30.5% off net, not list, prices. Then the price moved, and not because of anything that happened in a laboratory. On November 6, 2025 arrangements put injectables at $245 a month for Medicare and Medicaid with a $50 beneficiary copay; on November 25, CMS published a negotiated 2027 price of $274 per 30-day supply against a 2024 list of $959 — 71% off. A temporary Medicare GLP-1 Bridge began operating on July 1, 2026 and was extended to December 31, 2027; KFF counted 3.8 million eligible beneficiaries and a programme cost anywhere from $1.3 billion to $10 billion, and CMS said on July 29, 2026 that 250,000 had signed up. And no re-score exists. The October 2024 report remains the operative federal number, priced off assumptions a single afternoon in November made obsolete; the one published re-estimate at the new prices, Sullivan et al. in Value in Health (online February 5, 2026), answers a narrower question about spending on one drug. Run the arithmetic in the open and label it as arithmetic: at $2,940 a year, semaglutide plausibly still misses and tirzepatide plausibly clears — and rescaling a model's output is not re-running it. Where it lands is the point. Every other case in this series has a seam — the payer who buys is not the payer who saves. This one does not: Medicare buys the drug and Medicare pays for the heart attack, through machinery that already exists. The seamless case still does not clear on cash, and the adjacent claims about drinking and smoking rest on trials of 48 and 108 people, short and selected, whose own authors warn against reading them as licence to prescribe. What changed the answer was price, not evidence — and an announced arrangement is not a statute. So the argument the series has left is the one nobody can score: we decided we catch people, we wrote it down, we have never taken it back, and the only question is when we pay. Episode 9 asks the same question of the insurer of last resort. Sources include CBO's How Would Authorizing Medicare to Cover Anti-Obesity Medications Affect the Federal Budget? (October 8, 2024); Hwang et al., JAMA Health Forum 6(3) and 6(4) (2025); Sullivan et al., Value in Health 29(7) (2026); CMS's negotiated 2027 price fact sheet (November 25, 2025) and Bridge programme documents; KFF's June 29, 2026 eligibility analysis; the USC Schaeffer Center critique of October 10, 2024; and the alcohol and tobacco trial literature in JAMA Psychiatry, The Lancet and Annals of Internal Medicine. ▶ Watch on YouTube: https://www.youtube.com/watch?v=rUzqkWt4SzI 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  9. Episode 9

    Ep 9 — The Insurer of Last Resort

    In fifty-seven years the California FAIR Plan had assessed its member insurers three times, for $260 million in total. On February 11, 2025, after the Palisades and Eaton fires ran $4.039 billion of incurred loss through a wildfire insurance pool holding $510 million of retained earnings, Commissioner Ricardo Lara signed Order No. 2025-1 and billed the industry $1 billion, payable in thirty days — and the same afternoon issued Bulletin 2025-4, permitting those insurers to recover half of it from their own policyholders. Episode 9 moves this series' structure from bodies to property, where it is written down. The FAIR Plan is not a climate-change invention: it comes out of Watts in 1965 and Newark and Detroit in 1967, the Hughes Panel's 1968 report on insurance in riot-affected areas, and the Urban Property Protection and Reinsurance Act of 1968, which offered states federal riot reinsurance if they built a plan of last resort. California enacted Chapter 9 of its Insurance Code that year; the federal reinsurance was discontinued in 1982 and the state pools were not. Section 10095(c) makes every admitted property insurer a member as a condition of its licence, sharing writings, expenses, profits and losses by market share. Policy counts sat flat near 140,000 through 2018, hit 189,790 in 2019, and reached 668,609 policies and $724 billion of exposure by December 2025 — over 675,000 and $768 billion by June 2026. The pass-through has a serious defence, and the episode gives it: the median charge was about $28 a year, capped at two years and separately itemized, where an absorbed billion would have returned through rates invisibly and permanently. Consumer Watchdog, which wrote Proposition 103, filed a verified petition in Los Angeles Superior Court (No. 25STCP01367) challenging two bulletins and seeking more than $400 million back; on June 30, 2026 Judge Tiana J. Murillo denied the writ, holding that § 10095(c) governs apportionment among insurers and not an insurer's later transactions with its policyholders. That is a trial-level ruling; the petitioner said on July 1 it was reviewing appellate options, and no appeal has been verified. The economics are contested too — charity-hazard work by Raschky and Weck-Hannemann and by Boomhower and colleagues against Kousky and Cooke on fat tails and tail dependence, with Keys and Mulder finding reinsurance repricing explains 45 percent of the rise in disaster-risk pass-through. Where it lands: no second assessment has been reported, the plan arranged a $600 million revolving credit line instead under Stipulation and Order No. 2026-1 on February 27, 2026, and a 29.1 percent average rate increase was announced on August 11, 2026, effective October 15. Three federal programmes have the same shape and never use the name — flood since 1968, with $22.525 billion owed to the Treasury and $16 billion cancelled outright in October 2017; crop insurance since 1938 and 1980; and terrorism risk since 2002. An open incidence always resolves downhill, to the party with no one left to bill. Episode 10 asks why nobody buys the thing that would shrink the bill. Sources include California Insurance Code §§ 10090 et seq., CDI Order No. 2025-1 and Bulletin 2025-4, Stipulation and Order No. 2026-1, the ruling in Consumer Watchdog v. Lara, the Assembly Insurance Committee's January 28, 2026 oversight record, CRS reporting on the flood programme, and work by Kousky and Cooke, Keys and Mulder, and Boomhower and colleagues. ▶ Watch on YouTube: https://www.youtube.com/watch?v=zXec9FdqbZw 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  10. Episode 10

    Ep 10 — The Discount That Isn't Big Enough

    A Class A fire-rated roof costs about $6,300, or roughly $360 a year annualized over a 25-year life. The average wildfire mitigation discount California's rate filings actually grant for it is $31.38 a year — about one-eleventh of its cost. In March 2026 the state finally measured what home hardening is worth, and the number was excellent: rebuild the nearly 30,000 homes inside the Palisades and Eaton perimeters to the wildfire-prepared standard and modelled average annual loss falls 31 percent, 35 percent at the enhanced standard, for about 3 percent more in construction cost. Episode 10 does the arithmetic between those two figures. California moved first in the nation: Safer from Wildfires, announced October 17, 2022, is the first regulation anywhere requiring insurers to reflect wildfire mitigation in their rates. Reading the resulting filings in Working Paper 25-30 (December 2025), Resources for the Future found the maximum available discount is about $215.78 a year measured against the whole premium, or about $101.52 where an insurer credits only the wildfire-attributed portion — against a comprehensive retrofit bundle of roughly $40,000, which annualizes to just under $2,300. Alabama's FORTIFIED wind credits, by contrast, run 20 to 60 percent of the wind premium, worth $366 to $915 a year. The reflex is to call the wildfire number stingy. This episode argues it is worse than that: it is correct. Under Proposition 103 a credit an insurer cannot substantiate is not a discount but a cross-subsidy, and the evidence base is genuinely young. More importantly, the 31 percent is a community-wide model result. Knapp, Valachovic, Quarles and Johnson's Camp Fire survival analysis in Fire Ecology (2021) found the strongest single predictor of a home surviving was distance to the nearest destroyed structure — 35.4 percent survival beyond about 18 metres against 5.8 percent inside it, with 73.6 percent of Paradise homes inside that distance. Syphard and Keeley find building characteristics outranking defensible space. Much of the benefit your retrofit buys is delivered to your neighbours, and no contract written on one parcel can capture it. The counterweight is real too: the Insurance Institute's December 10, 2025 Palisades and Eaton findings show four hardening features together yielding 54 percent likelihood of avoiding damage against 36 percent for one action, and fuel coverage above 25 percent in the zero-to-five-foot zone driving damage probability toward 90 percent. In a built-out neighbourhood the parcel is the only lever left. That leaves a precise and unpleasant symmetry: the work has to happen house by house, and the payoff cannot be delivered house by house. Widen out and there is one intervention measured at the scale the hazard operates on — Strabo, Bryan and Reimer in Science, May 7, 2026, across 285 western wildfires from 2017 to 2023, reporting $3.73 in benefits per dollar of fuel treatment, $5.59 on projects over 2,400 acres, with health inside the number. Note what that is and is not: it is a social ratio, spread across insurers, households, a carbon budget and a health system. No budget line falls, and nothing here claims prevention pays for itself — this series broke that claim in Episode 6. It is bought, where it is bought at all, by an appropriation. Episode 11 takes the case where the loss has no owner at all. Sources include the California Department of Insurance's March 27, 2026 mitigation study and its October 17, 2022 Safer from Wildfires release, RFF Working Paper 25-30 by Ludington, Liao and Walls, Strabo, Bryan and Reimer in Science, Knapp and colleagues in Fire Ecology, Syphard and Keeley, and the Insurance Institute's 2020 and December 2025 post-event investigations. ▶ Watch on YouTube: https://www.youtube.com/watch?v=JVkly3Q28Fk 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  11. Episode 11

    Ep 11 — Smoke With No Owner

    Wildfire smoke is associated with roughly 41,000 excess deaths a year in the contiguous United States across 2011 to 2020, on work published in Nature on September 18, 2025 — projected to 71,420 a year by 2050 and 1.9 million cumulatively between 2026 and 2055, with monetized smoke mortality exceeding all other estimated US climate damages combined. It crosses every jurisdictional line it meets, and it arrives with no defendant, no invoice and no payer. That absence, not the mortality number, is what this episode is about. Episode 11 first establishes the method and its rival: the 41,000 figure is built on long-term exposure, while acute-exposure work — Ma and colleagues in PNAS, October 2024 — runs about an order of magnitude lower, near 11,415 a year, and every benefit-cost conclusion downstream flips on which you use. Connolly and colleagues in Science Advances (June 18, 2024) attribute 52,500 to 55,700 California deaths across 2008 to 2018, monetized at $432 to $456 billion. The clearest trace of the bill landing on an institution is not a health ledger at all but a bond desk: the Dallas Fed's "Up in Smoke" finds hospital borrowing costs about 7.1 basis points higher and nursing homes 12.1 after smoke exposure — roughly $270 million in incremental interest across 2010 to 2019, with effects 66 and 75 percent larger in high-uninsurance counties, and smoke from one state raising borrowing costs in another. Then the remedy, and it is not a free win. For: Kelp and colleagues in AGU Advances (2025) find burn severity down about 16 percent and net smoke particulate down about 14 percent where prescribed fire had been applied, with prescribed fire emitting roughly 17 percent of what the displaced wildfire would; Higuera-Mendieta and Burke in Science (June 11, 2026) find severe-fire risk cut about 92 percent for roughly a decade — after an initial rise in smoke of about 50 percent, and with the authors stating plainly that they do not compute health outcomes. Against: Dennin and colleagues in Communications Earth & Environment (March 2025) price 2017 US fire-smoke health damages at about $200 billion and 20,000 premature deaths, and attribute roughly half of that damage to prescribed burning rather than wildfire. Jones and Berrens in Contemporary Economic Policy (April 2021) find every $1 of Georgia prescribed burning creating $0.43 to $2.46 in low-birth-weight and prematurity hospitalization costs. Schollaert and colleagues in Nature Sustainability (2024) show the benefit is not monotonic — past an inflection point the treatment smoke offsets the wildfire smoke avoided, and under the most aggressive scenario Butte County got worse. And four fire economists, writing in the International Journal of Wildland Fire on September 2, 2022, say of the public-health case that "we simply do not know this to be true." The episode's finding is a negative one. No operating US programme has been located in which a health payer, a hospital community-benefit fund, a Medicaid entity or a health impact bond finances fuel treatment; smoke and health appear in none of the stated criteria of the land-agency money that does fund it. The two near misses are instructive — in Flint, on March 3, 2016, the federal Medicaid agency extended coverage to the exposed population and declined the lead-abatement request, and CalAIM's asthma remediation support has been optional and slow — and the health-related-social-needs framework was rescinded on March 4, 2025. Filtration, priced honestly by Adibi and colleagues at $40,509 to $89,206 per quality-adjusted life year, is not cost-saving and acts on neither the hazard nor the person, only the interface. Nothing here argues prevention pays for itself. What is left is the question Episode 12 answers: when do we pay, and who is the "we" that could? Sources include Qiu and colleagues in Nature 647, Connolly and colleagues in Science Advances, Dennin and colleagues in Communications Earth & Environment, Kelp and colleagues in AGU Advances, Higuera-Mendieta and Burke in Science, Jones and Berrens and the 2022 Jones, McDermott, Champ and Berrens review, Schollaert and colleagues in Nature Sustainability, Adibi and colleagues in AJRCCM, Dallas Fed Working Paper 2503, and the CMS Flint demonstration record. ▶ Watch on YouTube: https://www.youtube.com/watch?v=kTI5heQvoS4 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  12. Episode 12

    Ep 12 — When We Pay

    In 1601 an English parliament made the relief of the poor a compulsory tax, and in 1986 an American Congress ordered every emergency department to screen and stabilize whoever arrives without asking about money — and appropriated nothing. Four hundred years, never repealed, no serious repeal movement in four decades. This is the verdict episode, and it is not the comfortable one: the fiscal case for buying upstream usually fails, and this series broke it against itself rather than wait for someone else to do it. Episode 12 states the four findings the record supports. The commitment is real and ancient — and every society that decides to catch people invents, in the same breath, a way to sort them: the 1601 statute sorted in its own text, the Settlement Act of 1662 made removal the first fiscal instrument of the English state, and the 1834 workhouse test ran on less eligibility. The American courts dismantled the removal power by 1974, so the sorting relocated to the ledger. Second, prevention is a category and not a discount: Cohen, Neumann and Weinstein found fewer than one preventive measure in five reduces medical spending; Barendregt and van Baal found lifetime costs highest in the healthier cohorts; CBO's January 18, 2012 brief found spending unchanged or higher in nearly every one of thirty-four demonstration programs; and CBO's October 8, 2024 score of Medicare anti-obesity coverage put deficits up roughly $35 billion. No hour of this show argues that prevention pays for itself, and this one does not. What survives needs no budget line to fall: we have already decided that we catch people, and given that, buying upstream is the cheaper way to honour a promise we are not going to break. That argument carries a debt, incurred in Episode 4 and payable here. On February 20, 1905 Justice Harlan upheld a five-dollar vaccination fine in Jacobson v. Massachusetts with four limits written into the opinion. On May 2, 1927, in Buck v. Bell, Justice Holmes carried the public-cost premise across and none of the four limits made the trip — and Buck v. Bell has never been overruled, a fact about the reports, stated in the present tense. That is not a caricature of this series' reasoning; it is this series' reasoning without a boundary. So the boundary is stated plainly, in one sentence: the intervention must act on the hazard, not on the person. Filter the water, take the lead out of the fuel, thin the stand, put the fire back, harden the roof, price the smoke — none of that sorts anybody, and the benefit reaches whoever is standing there. Then the principle is tested on the hard cases rather than the easy ones — compulsory vaccination, helmet laws, and a drug prescribed to an identified person, which stays on the right side of the line only because it is offered rather than imposed. Episode 5's diagnostic explains why the frontier is hard: fuel treatment clears on cost and fails on the single decision point; home hardening fails because the benefit belongs to the block; smoke fails institutionally, appearing on no payer's ledger at all. Which is why the answer keeps arriving as an appropriation rather than a market — and Episode 7's warning applies to every dollar of it: an appropriation is not a purchase. The frank residue is that preventive care is roughly three percent of health spending under every financing system on earth, so no country has a working model to sell. The compiler closes in his own voice on the obligation and the room: obliged to treat, permitted to be unpaid, standing in the most expensive room in the building, at the end of a sequence that began years earlier with somebody deciding not to spend. The country has already agreed to pay. All it has ever decided is how late. Sources include the Poor Relief Act 1601, the Settlement Act 1662 and the Poor Law Amendment Act 1834; Jew Ho v. Williamson, Jacobson v. Massachusetts, Buck v. Bell, Skinner and Simon v. Sargent; EMTALA at 42 U.S.C. § 1395dd; CBO's 2012 demonstrations brief and its October 2024 anti-obesity score; Cohen, Neumann and Weinstein, Barendregt and van Baal; Strabo, Bryan and Reimer in Science; Qiu and colleagues in Nature; RFF Working Paper 25-30; and Eurostat's 2023 preventive-care share. ▶ Watch on YouTube: https://www.youtube.com/watch?v=2AZbV3LOR50 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

See All (12)

About

A federal statute requires any hospital with an emergency department to examine and stabilise anyone who walks through the door, regardless of ability to pay. It appropriates no money. The physician on call is obliged to come, and obliged to work for nothing if the patient does not pay. That statute is the plainest statement in American law of a bargain the country made and has never revoked: when no one else pays, society does. Twelve episodes walk that bargain with dates and sources — the poor rate of 1601 and the workhouse test of 1834; the Delaware hospital that turned away a four-month-old in 1959 and owed him nothing; the 467 patients transferred to a Chicago public hospital where lack of insurance was the stated reason in most cases; the federal court that wrote that society picks the person up off the highway, and the footnote in which it refused to go further; and Buck v. Bell, where the same reasoning about what people cost society, with nothing to bound it, produced a sterilisation order that has never been overruled. Then the harder half. Clean water, childhood vaccination and taking the lead out of gasoline were bought upstream and repaid many times over. Most prevention does not: fewer than one preventive measure in five is cost-saving, a population that quits smoking costs more over a lifetime because it lives longer, and thirty-four Medicare programmes produced no net savings between them. The series proves that against its own thesis, then follows the same structure into the present — the tobacco settlement that won outright and delivered hundreds of billions, of which three and a half percent now reaches prevention; the anti-obesity drugs scored in the red; the California wildfire pool holding seven hundred and sixty-eight billion dollars of exposure; a forty-thousand-dollar retrofit that earns a two-hundred-and-sixteen-dollar discount; and the smoke that kills tens of thousands a year and has no owner to bill. The verdict is not that prevention pays for itself. It is that we already decided we would catch people, and the only thing left to choose is when we pay. Compiled by a practising anesthesiologist bound by law and conscience to treat anyone who arrives in an emergency department needing care, and to do it for free if they refuse to pay. No patient, case or clinical detail appears anywhere in this series, and nothing in it is medical advice. Every cost-offset claim is presented with its rebuttal and its date; pending litigation is described in its procedural posture only. AI-narrated from a fully cited research dossier for every episode.

Information

More From Let's ReThink History