ReThink History: The Price of a Tonne

A tonne of carbon dioxide is the strangest commodity ever traded. Nobody wants it, its value is the value of its absence, and the whole market rests on two things no one can see: the tonne itself, and the world in which it was emitted. Since 1997 we have built, broken and rebuilt markets for that tonne — Kyoto's flexible mechanisms, Europe's emissions trading system and the five-billion-euro tax carousel that looted it, California's cap-and-trade and its forest offsets, the United Nations scheme to pay tropical forests to stand, and a voluntary offset market that touched two billion dollars a year and then lost most of it when journalists and scientists found that most of what it sold did not exist. Fourteen episodes walk that history with dates and sources: the Chinese and Indian refrigerant plants paid more to destroy their waste gas than to make their product; the phantom rainforest credits Disney, Shell and Gucci bought, and the registry rebuttals that followed within hours; the hundred-year wildfire insurance fund that burned through in a decade; the World Bank carbon pioneer indicted over cookstove data, who denies the allegations and has not been tried; the satellites now trying to weigh the world's forests from orbit, and the one that cannot legally switch on over America; and the carbon border charge that went live on Europe's imports in January 2026. A coda then takes the restoration forestry almost everyone agrees on — thinning, cultural and prescribed fire, a native understory — and asks whether any carbon market on earth can pay for it. Compiled by a physician who spent eighteen months in 1976 and 1977 at a Forest Service laboratory in Moscow, Idaho, cutting Douglas-firs into disks and counting their rings for the growth model that still projects every forest-carbon baseline in California, and who now builds software for this market. Every over-crediting finding is presented with the registry's or regulator's rebuttal and its date; every pending criminal matter is described in the present tense with the presumption of innocence. AI-narrated from a fully cited research dossier for every episode.

  1. Episode 1

    Ep 1 — The Price of a Tonne

    In 1968 a University of Toronto economist named John Dales published a 111-page essay arguing that a government should decide how much pollution it would tolerate, print exactly that many rights, and sell them — "no person, or agency, has to set the price." That is where carbon markets, cap and trade, and carbon offsets begin: with a commodity nobody can see, whose entire value is the value of its absence. Episode 1 lays out the three answers economics gave to the same question — Pigou's tax (1920), Coase's bargain (1960), and the tradable permit that Thomas Crocker and John Dales arrived at independently in 1966 and 1968 — and then corrects the origin story most people are told. The first real trading programme was not acid rain. It was lead: the EPA's 1982–87 gasoline lead phasedown moved more than sixty percent of the lead in American gasoline through traded credits, at roughly a fifth less cost than the rules alone would have imposed. Montreal Protocol allowances followed in 1987; sulfur dioxide in 1990. Then comes the leap that made everything afterwards hard. Carbon dioxide is a global, well-mixed, stock pollutant, and a carbon market has to do three things a sulfur market never had to: count a tonne nobody can see at a stack, count a tonne that was *avoided* somewhere else, and count a tonne stored in a living thing. The first is engineering. The second and third are the rest of this series. The episode closes on the question the next thirteen will test, and hands off to Kyoto, December 1997: you can price a tonne — but can you price a tree? Sources include Dales (1968), Crocker (1966), Coase (1960) and Pigou (1920); EPA's lead-content rulemakings and its 2001 report on economic incentives; Schmalensee and Stavins' Belfer Center discussion paper on three decades of cap-and-trade; and Calel's LSE Grantham history of carbon markets. ▶ Watch on YouTube: https://www.youtube.com/watch?v=yVUBKnZ50oU 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  2. Episode 2

    Ep 2 — The Day of the Atmosphere

    On December 8, 1997, Al Gore landed at a deadlocked climate conference in Kyoto and instructed the American delegation to show flexibility — on one condition: the treaty had to contain market mechanisms. Three days later Raúl Estrada-Oyuela gavelled the Kyoto Protocol through after an all-night session and called it "the Day of the Atmosphere." The world's first global carbon market had been born as an afterthought — three sentences saying trading could happen and the rules would be written later. This episode tells how the United States switched from opposing binding targets to demanding a carbon market, and how the Senate killed the treaty before it existed: the Byrd–Hagel Resolution passed 95–0 on July 25, 1997. Clinton signed in November 1998 and never sent it to the Senate; Bush repudiated it in March 2001; the Marrakesh Accords wrote the rulebook that autumn; and Russia — which had never wanted the treaty — held the key to its life and traded it for European backing on WTO entry, ratifying in October 2004 and bringing the Protocol into force on February 16, 2005. Then the accounting. Kyoto's targets were measured against 1990, a year that a third of the signatories' economies had already collapsed far below. The UNFCCC's own December 2012 table put the surplus at roughly 13.4 billion tonnes of assigned amount sitting in Russia and Ukraine — "hot air." Joint Implementation turned out worse than the Clean Development Mechanism: a 2015 Stockholm Environment Institute study judged about three-quarters of its credits doubtful. Canada walked out in December 2011. The Doha Amendment entered into force on December 31, 2020 — the last day of the period it governed. The episode gives the results a fair hearing: all thirty-six Parties complied, at about a tenth of forecast cost. It also says what that compliance rode on. The first lesson of the series lands here — a cap is only as real as its baseline. Sources include the Kyoto Protocol text and UNFCCC decisions and press records; the Clinton White House archive and the Bush letter of March 2001; S.Res.98; the UNFCCC carry-over table of December 2012; the *Earth Negotiations Bulletin* from COP3; Shishlov, Morel and Bellassen in *Climate Policy* (2016); and the Stockholm Environment Institute's Joint Implementation working paper. ▶ Watch on YouTube: https://www.youtube.com/watch?v=YZJwI98xLEs 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  3. Episode 3

    Ep 3 — The Factory That Was Paid to Pollute

    In 2012 a chemical company in Gujarat, India, earned about $187 million from carbon credits — not for the refrigerant it manufactured, but for incinerating the waste gas the refrigerant plant produced. Nineteen plants like it, in China, India, South Korea, Argentina and Mexico, generated more than half the credits issued in the first six years of the world's first global offset market: the Kyoto Protocol's Clean Development Mechanism. The arithmetic is the story. Destroying HFC-23 cost under a dollar a tonne of CO2 equivalent; the gas has a global warming potential around 14,800; the credits sold for ten or twenty euros. UNEP's own technical panel had warned that destruction revenue could easily exceed the revenue from the refrigerant itself. The World Bank bought roughly a billion dollars' worth. China taxed the windfall at 65 percent and collected about $2 billion. When watchdogs showed plants running their waste ratio exactly at the methodology's 3% cutoff — and, in some cases, producing refrigerant only while credits were flowing — the CDM Executive Board froze issuance, cut the methodology to 1% in 2011, and the EU banned the credits outright from May 1, 2013. Then the money stopped, and some plants threatened to vent the gas again. Underneath the scandal sits the design flaw the rest of the series keeps meeting: the additionality test. The 2016 study by Cames and colleagues for the European Commission's climate directorate found that 85% of CDM projects, and 73% of the credit supply, had a low likelihood of being additional — because credit revenue is small next to the uncertainty in a project's other numbers, so the test has almost no signal. Note the two-sided record: those findings are contested, and the mechanism's defenders point to projects that plainly did deliver. The CDM is now dying on a schedule — issuance requests closed June 30, 2026, the registry closes July 2, 2027 — while its projects migrate to the Paris Agreement's Article 6.4 mechanism. The lesson goes to Episode 10: a market that pays for the gap between reality and a hypothetical rewards the most persuasive hypothetical. Sources include the Kyoto Protocol's Article 12 and the Marrakesh decisions; CDM Executive Board methodologies and UNFCCC registry statistics read in August 2026; the 2010 CDM Watch evaluation for the European Commission; Cames et al. (2016) for DG CLIMA; Wara in *Nature* (2007) and Wara and Victor's Stanford working paper; Haya's Berkeley working paper; and the Environmental Investigation Agency's *Unchecked HFC-23* briefing. ▶ Watch on YouTube: https://www.youtube.com/watch?v=XQp_Mb_jOUM 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  4. Episode 4

    Ep 4 — The Carousel

    On June 2, 2009 the Paris exchange BlueNext traded 19.8 million tonnes of European emission allowances in a single day — a record that had nothing to do with anyone's smokestack. It was the peak of a value-added-tax carousel: shell companies buying allowances tax-free across a border, selling them with VAT inside France, and vanishing with the tax. Europol priced the loss at roughly €5 billion in eighteen months and said that in some countries "up to 90% of the whole market volume" was fraudulent. The EU emissions trading system had already crashed once. Caps in Phase 1 were set from estimates, almost all allowances were given away free, and when the first verified emissions data leaked in April 2006 the price fell from €30.65 to €11.40 in two weeks and to zero by 2007. Then came the theft: German phishing in January 2010, Hungary's recycled credits, 1.6 million allowances taken from Holcim Romania, and a January 2011 sweep that pushed the Commission to suspend every national registry on January 19, 2011. Episode 4 follows the criminal record where it actually goes. In France, the "carbon connection" ended in convictions on July 7, 2016 — and in four linked killings; Arnaud Mimran was ordered in April 2025 to stand trial on homicide counts that remain pending, and he is presumed innocent of them. In Britain, the Dosanjh case took £39 million in 69 days. Frankfurt prosecutors raided 230 premises tied to Deutsche Bank in April 2010. The fix was one line in the tax code — reverse-charge VAT — and a rebuilt Union Registry with delays, trusted accounts and know-your-customer rules. And yet: as of late August 2026 the system has cut the emissions of the plants it covers by 51% since 2005, trades a tonne at about €80–84, and remains the reference design for every carbon market since. The lesson the episode hands forward is precise — a compliance market can be looted through its plumbing and still work, because the tonnes at the stack were always real. Every fraud here was financial, not physical. Sources include Directive 2003/87/EC and the VAT and registry regulations that followed; *Dosanjh v R* [2013] EWCA Crim 2366; the Europol release of December 9, 2009 and Commission registry memos; European Environment Agency emissions indicators; and contemporaneous French, British and German press coverage of the trials. ▶ Watch on YouTube: https://www.youtube.com/watch?v=g5gAgi03_iU 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  5. Episode 5

    Ep 5 — The Golden State's Market

    On November 14, 2012 California sold the right to emit 23 million tonnes of carbon dioxide at $10.09 a tonne, and the largest carbon market in the Western Hemisphere opened for business. Cap and trade in California has since survived a ballot initiative bankrolled by Texas refiners, a five-year lawsuit over whether an auction is a tax, a buyers' strike in 2016, a federal suit from the Trump administration, and a provincial partner that joined and quit inside seven months. It also worked, on its own terms. The state hit its 2020 target four years early. The auctions have moved about $36.2 billion into the Greenhouse Gas Reduction Fund, most of it earmarked for disadvantaged communities — and a quarter of it, now a fixed $1 billion a year through 2045, to high-speed rail. Episode 5 walks the politics: AB 32 in 2006, the defeat of Proposition 23 in 2010, the *CalChamber v. CARB* ruling that allowances are "valuable, tradable commodities," and the two-thirds vote for AB 398 in July 2017 that extended the programme and cut offset use to 4% and 6%. Then the two critiques, both given their strongest form. On environmental justice, Cushing and colleagues in *PLOS Medicine* (2018) found local emissions rose at 52% of covered facilities; Hernandez-Cortes and Meng found the exposure gap between communities narrowed by 21–30%. On the bank: the Legislative Analyst warned in 2017 that 100–300 million allowances could pile up unused; by 2022 it was roughly 321 million, and 2023 covered emissions were already above the 2026 cap. In September 2025 Newsom signed the extension to 2045 and the rename to "cap-and-invest"; on May 29, 2026 CARB removed 118 million allowances from 2027–30 — and, in the same rule, created a manufacturing incentive reserve of up to 118.3 million that EDF argues neutralises the cut. CARB disputes that reading. The May 2026 auction settled at $28.81, against a 2026 floor of $27.94 and a ceiling of $102.52. Washington, California and Quebec signed a three-way linkage agreement on June 25, 2026, with joint auctions planned for 2027. The lesson: a cap with a bank is a promise about the future, and California has been borrowing against it for a decade. Sources include CARB auction records and its May 2026 rulemaking; *California Chamber of Commerce v. CARB*, 10 Cal.App.5th 604; AB 32, AB 398, AB 1207 and SB 840; Legislative Analyst's Office reports; Cushing et al. in *PLOS Medicine*; the NBER working paper by Hernandez-Cortes and Meng; and EDF and ICAP analyses of the 2026 rule. ▶ Watch on YouTube: https://www.youtube.com/watch?v=8ns8SV6J9TY 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  6. Episode 6

    Ep 6 — Weighing a Forest

    In 1976, in Moscow, Idaho, a young biologist spent eighteen months at the Forest Service's Forestry Sciences Laboratory felling Douglas-firs and cutting them into disks — one from the base, then up the stem every ten feet or ten years of height growth, measuring the diameter and counting the rings at each — so that a computer model could learn how much tree hides behind a diameter, and how fast it had got there. Nobody measured carbon; that took two more conversions, applied by other people to other data, decades later. The model was called Prognosis — Albert Stage had published it three years earlier as Research Paper INT-137. Renamed the Forest Vegetation Simulator, it is now the nationally applicable model on California's approved list for the hundred-year projections behind every forest carbon credit the state has issued. This is the measurement episode, told first-person from that lab and then outward: German yield tables, American timber cruising, Bitterlich's angle gauge, the 1928 McSweeney-McNary Act and the Forest Survey that became today's Forest Inventory and Analysis programme — roughly one plot per 6,000 acres, annual panels since 1999, about 125,000 forested plots, all public. Then allometry: the equations that turn a diameter into a tonne. Jenkins and colleagues in 2003, Chojnacky's 2014 revision drawn from 3,464 published equations, and the 2024 step change — the National Scale Volume and Biomass estimators, built on 234,823 felled trees across 339 species, with species-specific carbon fractions averaging 0.477 rather than the old flat 0.5. The result: identical field data, and America's forests came out **14.6% heavier**. Indiana rose 37%; Washington fell 1.5%. The national forest sink moved from 13% of emissions to 16.1% to 14.6% across three successive inventories. The same trees, three numbers. The episode ends on the thing that should worry anyone buying a forest credit: three different objects get called "the forest carbon number" — a tree prediction with roughly 25% error, a state-level estimate with a few percent error that says nothing about any particular 2,000-acre project, and a wall-to-wall map with large correlated errors. Projects are credited at a precision the tools do not have. And in 2026 the institution that holds the scale is being cut back, with research facilities closing and a $0 FY2027 request for Forest Service R&D. Sources include Stage's INT-137 and the Forest Service research-paper lineage through Monserud, Wykoff and Crookston; FIA field methods and the public database; Jenkins et al. (2003), Chojnacky et al. (2014) and Westfall et al., GTR-WO-104 (2024); the EPA inventory data released to EDF under FOIA; the University of Maryland's independent 2026 inventory; and 2026 reporting on the Forest Service reorganisation. ▶ Watch on YouTube: https://www.youtube.com/watch?v=NUs9usJ0d_g 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  7. Episode 7

    Ep 7 — Common Practice

    In July 2021 the Bootleg Fire burned through the Klamath East project near Bly, Oregon — one of the forest carbon offset projects whose credits California's regulated polluters had bought to cover their emissions. Within two years the programme's hundred-year wildfire insurance reserve, meant to last until 2115, was in the words of the scientists who audited it "almost surely depleted." California's 2011 forest protocol works like this: measure the carbon in your trees, compare it to the regional average for similar forests — "common practice" — and get paid up front for the difference, on a promise to keep the carbon there for a hundred years, insured by a shared buffer pool of withheld credits. By January 2025 the programme had issued more than 265 million offset credits, most from forests, including more than 90 million to tribal projects worth roughly $1.5 billion. That includes the Yurok Tribe's project, whose revenue helped fund the 47,097-acre Blue Creek reacquisition completed in 2025 — the case for. Then the case against, with the rebuttal attached. Bloomberg in December 2020 and ProPublica with MIT Technology Review in April 2021 found projects credited against clear-cuts that were never going to happen. Badgley, Haya and colleagues, publishing in *Global Change Biology* in 2021–22, examined 65 projects and 102 million credits and found roughly 29.4% over-credited — about 30 million tonnes, some $410 million — because supersection averages pool very different forest types together. **CARB rejected the finding in October 2021**, arguing the paper addressed only one element of additionality and ignored legal and financial constraints and the protocol's minimum-baseline backstop. The regulator has not disputed the second finding: the same team's August 2022 buffer-pool paper, which showed wildfire losses — Trinity, Eddie Ranch, Montesol, Colville, Warm Springs, Klamath East — consuming 95 to 114% of a century's wildfire reserve in under ten years. As of late August 2026 the 2015 protocol is still the operative rule, a rewrite is mandated by 2029, an offsets study is due December 31, 2026, and the buffer keeps burning. The state priced a standing tree by comparing it to its neighbours and insured it for a century — and the century arrived in a decade. Sources include CARB's U.S. Forest Projects protocol and its October 2021 response; Climate Action Reserve protocol history and its June 2026 permanence decision; Badgley et al. in *Global Change Biology* and *Frontiers in Forests and Global Change*; CarbonPlan's project and buffer-pool monitoring; Legislative Analyst's Office report 5097; and the Bloomberg and ProPublica investigations. ▶ Watch on YouTube: https://www.youtube.com/watch?v=Qk4dFtX0O3M 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

  8. Episode 8

    Ep 8 — Paying the Forest to Stand

    In December 2007, at the UN climate conference in Bali, the world accepted an idea Kyoto had refused: pay tropical countries for the carbon in forests they choose not to cut. Norway put up the money first — up to three billion kroner a year — and REDD+ was written into the Bali Action Plan. Eighteen years later, in Belém, Brazil launched a $125-billion fund to do the same job with no carbon credits at all, because the credits had become the problem. Episode 8 follows the money channel by channel: the Amazon Fund, which took more than $1.2 billion from Norway before Bolsonaro froze it in 2019 and Lula restarted it in April 2023; Indonesia's on-again partnership; Guyana; the Green Climate Fund's results-based payments; and the LEAF Coalition's $1.5 billion in pledges against agreements whose disbursements are not public. It also takes the hardest case seriously. In December 2022 Guyana sold Hess 37.5 million credits under the ART-TREES standard, for a minimum of $750 million over ten years. Rainforest Foundation UK calculated that about 84% of those credits came from a "high forest, low deforestation" adjustment rather than measured reductions; five UN Special Rapporteurs wrote to Hess in November 2024. Guyana defends the HFLD approach, and the episode gives that defence its due. Brazil's carbon cowboys — Operation Greenwashing, 538,000 hectares of grabbed public land, and the 31 people **charged in November 2025 and not convicted** — get a brief hearing here and the full treatment in Episode 11. Then the twist. The Trees, Forests and Climate Fund launched in November 2025 with about $6.7 billion pledged against a $25 billion sponsor tranche, and by April 2026 had paid no forest country anything. Meanwhile the forest itself is doing better than the market: Brazil's PRODES figure for 2025 was 5,796 km², the lowest since 2014, and first-half 2026 alerts fell 38% to a decade low — driven by enforcement and weather, not credit revenue. The lesson: paying for a forest to stand works; paying for the *difference* between what stands and what would have been cut is where the counting breaks. Sources include UNFCCC decisions from Bali through the Warsaw Framework; Norway's climate and forest initiative records and Amazon Fund donation data; Green Climate Fund documents; Guyana's low-carbon strategy and ART-TREES records with the Rainforest Foundation UK and Forest Peoples Programme case studies; the UN Special Rapporteurs' communication; Brazilian federal police and prosecutorial filings; and INPE deforestation data. ▶ Watch on YouTube: https://www.youtube.com/watch?v=i5UXkO7lY-E 🎙 AI-narrated (Google NotebookLM), researched from primary sources.

About

A tonne of carbon dioxide is the strangest commodity ever traded. Nobody wants it, its value is the value of its absence, and the whole market rests on two things no one can see: the tonne itself, and the world in which it was emitted. Since 1997 we have built, broken and rebuilt markets for that tonne — Kyoto's flexible mechanisms, Europe's emissions trading system and the five-billion-euro tax carousel that looted it, California's cap-and-trade and its forest offsets, the United Nations scheme to pay tropical forests to stand, and a voluntary offset market that touched two billion dollars a year and then lost most of it when journalists and scientists found that most of what it sold did not exist. Fourteen episodes walk that history with dates and sources: the Chinese and Indian refrigerant plants paid more to destroy their waste gas than to make their product; the phantom rainforest credits Disney, Shell and Gucci bought, and the registry rebuttals that followed within hours; the hundred-year wildfire insurance fund that burned through in a decade; the World Bank carbon pioneer indicted over cookstove data, who denies the allegations and has not been tried; the satellites now trying to weigh the world's forests from orbit, and the one that cannot legally switch on over America; and the carbon border charge that went live on Europe's imports in January 2026. A coda then takes the restoration forestry almost everyone agrees on — thinning, cultural and prescribed fire, a native understory — and asks whether any carbon market on earth can pay for it. Compiled by a physician who spent eighteen months in 1976 and 1977 at a Forest Service laboratory in Moscow, Idaho, cutting Douglas-firs into disks and counting their rings for the growth model that still projects every forest-carbon baseline in California, and who now builds software for this market. Every over-crediting finding is presented with the registry's or regulator's rebuttal and its date; every pending criminal matter is described in the present tense with the presumption of innocence. AI-narrated from a fully cited research dossier for every episode.

More From Let's ReThink History