2c:null 2d:["$","title",null,{"children":"National Economies and Climate Change | Yash | Observable"}] 32:T769,{ const labelled = ["Qatar", "United States", "China", "Germany", "United Kingdom", "India", "Nigeria", "Switzerland", "Saudi Arabia", "Democratic Republic of Congo", "Brazil"] const pc2023 = countries .filter(d => d.year === 2023 && d.co2_per_capita > 0 && d.population > 0) .select("country", "co2_per_capita", "population") .objects() .map((d, i) => ({...d, jitter: (i * 37) % 100})) const worldPC2023 = data.filter(d => d.country === "World" && d.year === 2023).get("co2_per_capita", 0) const x = vl.x().fieldQ("co2_per_capita") .scale({type: "log"}) .title("CO₂ per capita in 2023 (tonnes, log scale)"); const y = vl.y().fieldQ("jitter") .axis(null) .title(null); return vl.layer( vl.markCircle({opacity: 0.55, stroke: "white", strokeWidth: 0.5}).data(pc2023).encode( x, y, vl.size().fieldQ("population") .scale({range: [10, 1200]}) .legend({ title: "Population", values: [1e7, 1e8, 1e9], format: "~s", labelExpr: "replace(datum.label, 'G', ' billion')", symbolOpacity: 0.55, symbolStrokeColor: "white", symbolStrokeWidth: 0.5 }), vl.tooltip(["country", "co2_per_capita", "population"]) ), vl.markText({dy: -10, fontSize: 10}).data(pc2023) .transform(vl.filter(`indexof(${JSON.stringify(labelled)}, datum.country) >= 0`)) .encode(x, y, vl.text().fieldN("country")), vl.markRule({color: "red", strokeDash: [4, 3]}) .data([{co2_per_capita: worldPC2023}]) .encode(x), vl.markText({color: "red", align: "left", dx: 4, dy: 148, fontSize: 10, baseline: "top"}) .data([{co2_per_capita: worldPC2023, t: `World average: ${worldPC2023.toFixed(1)} t`}]) .encode(x, vl.y().value(0), vl.text().fieldN("t")) ) .title({text: "CO₂ Per Capita (2023)"}) .width(640).height(160) .render(); }33:T7de,{ const t = countries .filter(d => d.year === 2023 && d.co2 != null && d.population > 0 && d.co2_per_capita != null) .orderby("co2_per_capita") .derive({cumPop: aq.rolling(d => aq.op.sum(d.population)), cumCO2: aq.rolling(d => aq.op.sum(d.co2))}); const P = t.rollup({p: aq.op.max("cumPop")}).get("p", 0); const C = t.rollup({c: aq.op.max("cumCO2")}).get("c", 0); const lorenz = [{country: "(start)", pop: 0, emis: 0}].concat( t.objects().map(d => ({country: d.country, pop: d.cumPop / P, emis: d.cumCO2 / C}))); const interp = x => { const i = lorenz.findIndex(d => d.pop >= x); const a = lorenz[i - 1], b = lorenz[i]; return a.emis + (b.emis - a.emis) * (x - a.pop) / (b.pop - a.pop); }; const b50 = interp(0.5), b90 = interp(0.9); const lorenzMarks = [ {pop: 0.5, emis: b50, label: `Lowest-emitting 50% of people: ${Math.round(b50 * 100)}% of CO₂`}, {pop: 0.9, emis: b90, label: `Highest-emitting 10%: ${Math.round((1 - b90) * 100)}% of CO₂`} ]; return vl.layer( vl.markLine({color: "grey", strokeDash: [4, 4]}).data([{pop: 0, emis: 0}, {pop: 1, emis: 1}]) .encode(vl.x().fieldQ("pop"), vl.y().fieldQ("emis")), vl.markArea({color: "blue", opacity: 0.25, line: {color: "blue"}}).data(lorenz) .encode(vl.x().fieldQ("pop").axis({format: "%"}).title("Cumulative share of country population, ordered from lowest to highest CO₂ per capita"), vl.y().fieldQ("emis").axis({format: "%"}).title("Cumulative share of CO₂"), vl.tooltip(["country"])), vl.markPoint({filled: true, color: "red", size: 60}).data(lorenzMarks) .encode(vl.x().fieldQ("pop"), vl.y().fieldQ("emis")), vl.markText({align: "right", dx: -8, dy: -8, fontSize: 11, fontWeight: "bold"}).data(lorenzMarks) .encode(vl.x().fieldQ("pop"), vl.y().fieldQ("emis"), vl.text().fieldN("label")) ) .title({text: "Cumulative Share of CO₂ vs. Population (2023)"}) .width(460).height(360) .render() }34:T63d,{ const wealth2022 = countries .filter(d => d.year === 2022 && d.gdp > 0 && d.population > 0 && d.co2_per_capita > 0) .derive({gdp_pc: d => d.gdp / d.population}) .select("country", "gdp_pc", "co2_per_capita", "consumption_co2_per_capita", "population") .objects() const wealthLabels = ["Qatar", "Switzerland", "Sweden", "United States", "China", "India", "North Korea", "Trinidad and Tobago", "Mongolia", "South Africa", "Ethiopia"] return vl.layer( vl.markCircle({opacity: 0.6}).encode( vl.size().fieldQ("population").scale({range: [15, 1500]}).legend(null), vl.tooltip(["country", {field: "gdp_pc", format: ",.0f"}, "co2_per_capita"]) ), vl.markLine({color: "red"}).transform( vl.regression("co2_per_capita").on("gdp_pc").method("pow") ), vl.markText({dy: -9, fontSize: 10}).transform( vl.filter(`indexof(${JSON.stringify(wealthLabels)}, datum.country) >= 0 && datum.country !== "Sweden"`) ).encode(vl.text().fieldN("country")), vl.markText({dy: 11, fontSize: 10}).transform(vl.filter("datum.country === 'Sweden'")) .encode(vl.text().fieldN("country")) ) .data(wealth2022) .encode( vl.x().fieldQ("gdp_pc").scale({type: "log", domain: [500, 200000], nice: false}).title("GDP per person (2022) (int'l-$, log scale)"), vl.y().fieldQ("co2_per_capita").scale({type: "log", domain: [0.04, 60], nice: false}).title("CO₂ per person (2022) (t, log scale)") ) .title({text: "GDP per Person vs. CO₂ per Person (2022)"}) .width(600).height(420) .render() }35:Tbe9,{ const focus = ["United States", "United Kingdom", "Germany", "Sweden", "China", "India"] const series = ["GDP", "CO₂ (territorial)", "CO₂ (consumption)"] const data = countries .params({focus}) .filter((d, $) => aq.op.includes($.focus, d.country) && d.year >= 1990 && d.year <= 2022) .groupby("country") .orderby("year") .derive({ "GDP": d => (d.gdp - aq.op.first_value(d.gdp)) / 1e12, "CO₂ (territorial)": d => d.co2 - aq.op.first_value(d.co2), "CO₂ (consumption)": d => d.consumption_co2 - aq.op.first_value(d.consumption_co2) }) .ungroup() .select("country", "year", ...series) .fold(series, {as: ["series", "value"]}) .objects() // Pick each axis's range so that 0 sits at the same height on both sides const alignedDomains = rows => { const range = vals => [Math.min(0, d3.min(vals)), Math.max(0, d3.max(vals))] const g = range(rows.filter(d => d.series === "GDP").map(d => d.value)) const c = range(rows.filter(d => d.series !== "GDP" && d.value != null).map(d => d.value)) const zeroAt = ([lo, hi]) => -lo / (hi - lo) // fraction of the axis below zero const p = (zeroAt(g) + zeroAt(c)) / 2 // shared zero height, halfway between the two const fit = ([lo, hi]) => zeroAt([lo, hi]) < p ? [-p * hi / (1 - p), hi] // stretch the bottom down : [lo, p === 0 ? hi : -lo * (1 - p) / p] // stretch the top up return [fit(g), fit(c)] } const x = vl.x().fieldQ("year").axis({format: "d", tickCount: 4}).title(null) const color = vl.color().fieldN("series").scale({domain: series, range: ["green", "red", "pink"]}).title(null) const dash = vl.strokeDash().fieldN("series").scale({domain: series, range: [[1, 0], [1, 0], [4, 3]]}).title(null) const panel = country => { const rows = data.filter(d => d.country === country) const [gDomain, cDomain] = alignedDomains(rows) return vl.layer( vl.markRule({color: "#999", strokeDash: [2, 2]}).data([{y: 0}]) .encode(vl.y().fieldQ("y").scale({domain: gDomain, nice: false}).axis(null)), vl.markLine({strokeWidth: 2}) .transform(vl.filter("datum.series === 'GDP'")) .encode(x, color, dash, vl.y().fieldQ("value").scale({domain: gDomain, nice: false}) .axis({titleColor: "green"}).title("Δ GDP (trillion int'l-$)")), vl.markLine({strokeWidth: 2}) .transform(vl.filter("datum.series !== 'GDP'")) .encode(x, color, dash, vl.y().fieldQ("value").scale({domain: cDomain, nice: false}) .axis({orient: "right", titleColor: "red"}).title("Δ CO₂ (million tonnes)")) ) .data(rows) .resolve({scale: {y: "independent"}}) .title({text: country, fontSize: 11, fontWeight: "normal"}) .width(200).height(150) } return vl.concat(...focus.map(panel)) .columns(3) .title({text: "Change in GDP and CO₂ Emissions Since 1990"}) .render() }36:T999,{ const pick = (yr, sfx) => countries .params({yr}) .filter((d, $) => d.year === $.yr) .select({country: "country", gdp: "gdp" + sfx, co2: "co2" + sfx, consumption_co2: "cons" + sfx, population: "pop" + sfx}); const decoupling = pick(2005, "05") .join(pick(2022, "22"), "country") .filter(d => d.gdp05 > 0 && d.gdp22 > 0 && d.co205 > 0 && d.co222 > 0) .derive({ gdpRatio: d => d.gdp22 / d.gdp05, co2Ratio: d => d.co222 / d.co205, consRatio: d => d.cons05 > 0 && d.cons22 > 0 ? d.cons22 / d.cons05 : null }) .derive({ pattern: d => d.gdpRatio > 1 && d.co2Ratio < 1 ? "Absolute decoupling (GDP ↑, CO₂ ↓)" : d.gdpRatio > 1 ? "Emissions growing (GDP ↑, CO₂ ↑)" : "Economy shrank (GDP ↓)" }) .objects(); const x = vl.x().fieldQ("gdpRatio").scale({type: "log", domain: [0.35, 6], nice: false}).title("GDP in 2022 ÷ GDP in 2005 (log)"); const y = vl.y().fieldQ("co2Ratio").scale({type: "log", domain: [0.35, 20], nice: false}).title("Territorial CO₂ in 2022 ÷ 2005 (log)"); const named = ["United Kingdom", "United States", "China", "India", "Laos", "Ethiopia", "Qatar", "Brazil"]; return vl.layer( vl.markRule({color: "#888"}).data([{co2Ratio: 1}]).encode(y), vl.markRule({color: "#888"}).data([{gdpRatio: 1}]).encode(x), vl.markCircle({opacity: 0.7, stroke: "white", strokeWidth: 0.5}).data(decoupling).encode( x, y, vl.color().fieldN("pattern").scale({domain: ["Absolute decoupling (GDP ↑, CO₂ ↓)", "Emissions growing (GDP ↑, CO₂ ↑)", "Economy shrank (GDP ↓)"], range: ["#2ca02c", "#d62728", "#9467bd"]}).legend({orient: "bottom", labelLimit: 300, title: null}), vl.size().fieldQ("pop22").scale({range: [15, 1500]}).legend(null), vl.tooltip(["country", {field: "gdpRatio", format: ".2f"}, {field: "co2Ratio", format: ".2f"}]) ), vl.markText({dy: -10, fontSize: 10}).data(decoupling) .transform(vl.filter(`indexof(${JSON.stringify(named)}, datum.country) >= 0`)) .encode(x, y, vl.text().fieldN("country")), vl.markText({align: "right", fontSize: 12, fontWeight: "bold", color: "#2ca02c"}) .data([{gdpRatio: 5.5, co2Ratio: 0.4, t: "absolute decoupling zone"}]).encode(x, y, vl.text().fieldN("t")) ) .title({text: "Change in GDP vs. Change in CO₂, 2005–2022"}) .width(600).height(400) .render(); }37:T47b,{ const incomeGroups = ["High-income countries", "Upper-middle-income countries", "Lower-middle-income countries", "Low-income countries"]; const groupTrade = aggregates .params({incomeGroups}) .filter((d, $) => aq.op.includes($.incomeGroups, d.country) && d.year >= 1990 && d.year <= 2023) .derive({"Territorial (production)": d => d.co2 / 1000, "Consumption-based": d => d.consumption_co2 / 1000}) .select("country", "year", "Territorial (production)", "Consumption-based") .fold(["Territorial (production)", "Consumption-based"], {as: ["accounting", "gt"]}) .objects(); return vl.markLine({strokeWidth: 2.5}) .data(groupTrade) .encode( vl.x().fieldQ("year").axis({format: "d"}).title(null), vl.y().fieldQ("gt").title("CO₂ (billion tonnes)"), vl.color().fieldN("accounting").scale({range: ["#ff7f0e", "#1f77b4"]}) .legend({orient: "top", direction: "horizontal", title: null}), vl.facet().fieldN("country").sort(incomeGroups).columns(4).title(null) ) .width(230).height(200) .title({text: "Territorial vs. Consumption CO₂ by Income Group"}) .render(); }38:Tadb,{ const pick = (yr, sfx) => countries .params({yr}) .filter((d, $) => d.year === $.yr) .select({country: "country", gdp: "gdp" + sfx, co2: "co2" + sfx, consumption_co2: "cons" + sfx}); const decoupledWithTrade = pick(2005, "05") .join(pick(2022, "22"), "country") .filter(d => d.gdp05 > 0 && d.gdp22 > 0 && d.co205 > 0 && d.co222 > 0 && d.cons05 > 0 && d.cons22 > 0) .filter(d => d.gdp22 > d.gdp05 && d.co222 < d.co205) .objects() .map(d => { const consRatio = d.cons22 / d.cons05; return { country: d.country, territorial: (d.co222 / d.co205 - 1) * 100, consumption: (consRatio - 1) * 100, survives: consRatio < 1 ? "Still decoupled after trade adjustment" : "Not decoupled once imports are counted" }; }) .sort((a, b) => a.territorial - b.territorial); // one row per point, so shape and color can come from a field (which creates a legend) const points = decoupledWithTrade.flatMap(d => [ {country: d.country, measure: "Territorial", value: d.territorial}, {country: d.country, measure: "Consumption-based", value: d.consumption} ]).filter(d => d.value <= 60); const y = vl.y().fieldN("country").sort(decoupledWithTrade.map(d => d.country)).title(null); const xT = vl.x().fieldQ("territorial").scale({domain: [-70, 60], clamp: true}).title("% change in CO₂, 2005–2022"); const measures = ["Territorial", "Consumption-based"]; return vl.layer( vl.markRule({color: "#888"}).data([{territorial: 0}]).encode(xT), vl.markRule({strokeWidth: 2}).data(decoupledWithTrade).encode( y, xT, vl.x2().field("consumption"), vl.color().fieldN("survives").scale({domain: ["Still decoupled after trade adjustment", "Not decoupled once imports are counted"], range: ["#2ca02c", "#bbbbbb"]}) .legend({orient: "bottom", labelLimit: 300, title: "Line", columns: 1}) ), vl.markPoint({filled: true, opacity: 1, size: 60}).data(points).encode( y, vl.x().fieldQ("value"), vl.shape().fieldN("measure").scale({domain: measures, range: ["circle", "diamond"]}).legend({orient: "bottom", title: "Change in CO₂", columns: 1}), vl.color().fieldN("measure").scale({domain: measures, range: ["#1f77b4", "#ff7f0e"]}).legend({orient: "bottom", title: "Change in CO₂", columns: 1}) ), vl.markText({align: "right", dx: -4, dy: -7, fontSize: 9, color: "#777"}).data(decoupledWithTrade) .transform(vl.filter("datum.consumption > 60"), vl.calculate("60").as("edge")) .encode(y, vl.x().fieldQ("edge"), vl.text().value("+343% →")) ) .resolve({scale: {color: "independent", shape: "independent"}}) .title({text: "% Change in CO₂ for Decoupled Countries, 2005–2022"}) .width(460).height(580) .render(); }39:T9dd,{ const withCons = wealth2022.filter(d => d.consumption_co2_per_capita > 0); const fitProd = logFit(withCons, "gdp_pc", "co2_per_capita"); const fitCons = logFit(withCons, "gdp_pc", "consumption_co2_per_capita"); const pcTrade = countries .filter(d => d.year === 2023 && d.co2_per_capita > 0 && d.consumption_co2_per_capita > 0) .derive({ importShare: d => d.consumption_co2_per_capita / d.co2_per_capita, logShare: d => Math.log2(d.consumption_co2_per_capita / d.co2_per_capita) }) .select("country", "co2_per_capita", "consumption_co2_per_capita", "population", "importShare", "logShare") .objects(); const x = vl.x().fieldQ("co2_per_capita").scale({type: "log", domain: [0.08, 60], nice: false}).title("Territorial CO₂ per person, 2023 (t, log)"); const y = vl.y().fieldQ("consumption_co2_per_capita").scale({type: "log", domain: [0.08, 60], nice: false}).title("Consumption-based CO₂ per person, 2023 (t, log)"); const named = ["Switzerland", "Sweden", "Singapore", "Hong Kong", "United States", "China", "India", "Qatar", "South Africa", "Russia", "United Kingdom", "Malta", "Vietnam", "Togo", "Nepal"]; return vl.layer( vl.markLine({color: "#999", strokeDash: [4, 4]}) .data([{co2_per_capita: 0.08, consumption_co2_per_capita: 0.08}, {co2_per_capita: 60, consumption_co2_per_capita: 60}]) .encode(x, y), vl.markCircle({opacity: 0.75}).data(pcTrade).encode( x, y, vl.color().fieldQ("logShare") .scale({scheme: "redblue", reverse: true, domain: [-1, 3], domainMid: 0, clamp: true}) .legend({ title: "consumption / territorial", values: [-1, 0, 1, 2, 3], labelExpr: "datum.value < 0 ? '1/' + pow(2, -datum.value) + '×' : pow(2, datum.value) + '×'", gradientLength: 200 }), vl.size().fieldQ("population").scale({range: [15, 1200]}).legend(null), vl.tooltip(["country", "co2_per_capita", "consumption_co2_per_capita", {field: "importShare", format: ".2f"}]) ), vl.markText({dy: -10, fontSize: 10}).data(pcTrade) .transform(vl.filter(`indexof(${JSON.stringify(named)}, datum.country) >= 0`)) .encode(x, y, vl.text().fieldN("country")) ) .title({ text: "Territorial vs. Consumption CO₂ per Person, 2023", subtitle: `Correlation of CO₂ per person with GDP per person (${fitCons.n} countries): r = ${fitProd.r.toFixed(2)} territorial, r = ${fitCons.r.toFixed(2)} consumption-based` }) .width(460).height(460) .render(); }3a:T51a,{ const measures = ["Share of all emissions since 1750", "Share of 2023 emissions", "Share of world population"]; const worldPop = data.filter(d => d.country === "World" && d.year === 2023).get("population", 0); const responsibility = countries .filter(d => d.year === 2023 && d.share_global_cumulative_co2 != null) .params({worldPop}) .derive({"Share of world population": (d, $) => 100 * d.population / $.worldPop, "Share of 2023 emissions": d => d.share_global_co2, "Share of all emissions since 1750": d => d.share_global_cumulative_co2}) .orderby(aq.desc("Share of all emissions since 1750")) .slice(0, 10) .select("country", ...measures) .fold(measures, {as: ["measure", "pct"]}) .objects(); return vl.markBar() .data(responsibility) .encode( vl.y().fieldN("country").sort(null).title(null), vl.yOffset().fieldN("measure").sort(measures), vl.x().fieldQ("pct").title("% of world total"), vl.color().fieldN("measure").scale({domain: measures, range: ["maroon", "red", "silver"]}).title(null), vl.tooltip(["country", "measure", {field: "pct", format: ".1f"}]) ) .width(480).height(420) .title({text: "Share of Historical Emissions, Current Emissions, and Population"}) .render(); }3b:[{"id":"be58c2fd703991b1947cf587789d9286bbda13d0b36e8c94397f3ea8c2f24d30123c2f8d38374a9a9aac4cf428626fba1de38ae689e7d885ae4562e53c8fb378","href":"https://static.observableusercontent.com/files/be58c2fd703991b1947cf587789d9286bbda13d0b36e8c94397f3ea8c2f24d30123c2f8d38374a9a9aac4cf428626fba1de38ae689e7d885ae4562e53c8fb378","name":"owid-co2-data.csv","type":"text/csv","lastModified":1791752619000,"size":14377942}] 30:{"id":"288d88260db53db1","slug":"national-economies-and-climate-change","title":"National Economies and Climate Change","thumbnail":"71e039807cf545f377eb34a4967874896a20397e8b2a763dd32b61b4ec29c9e2","access_level":"public","create_time":"$D2026-10-11T19:08:22.414Z","update_time":"$D2026-10-12T04:35:52.593Z","view_time":null,"first_public_time":"$D2026-10-12T04:19:55.023Z","has_unpublished_edits":null,"fork":false,"stars":1,"starred":false,"owner":{"type":"team","id":"1c9a1ce810efacf0","login":"yashuap","name":"Yash","avatar_url":"https://avatars-next.observableusercontent.com/eef5cb95caa7a977e323189831f2e9c03a177613e0a569ac75c06dd1ed4825b2","tier":"pro_2026","bio":""},"version":507,"publish_version":507,"role":null,"authored":false,"authors":[{"type":"individual","id":"636ac52ec479c507","login":"yashuap","name":"Yash","avatar_url":"https://avatars-next.observableusercontent.com/eef5cb95caa7a977e323189831f2e9c03a177613e0a569ac75c06dd1ed4825b2","tier":null,"bio":""}],"members":[{"type":"team","id":"1c9a1ce810efacf0","login":"yashuap","name":"Yash","avatar_url":"https://avatars-next.observableusercontent.com/eef5cb95caa7a977e323189831f2e9c03a177613e0a569ac75c06dd1ed4825b2","tier":"pro_2026","role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"titling":"auto","title":"National Economies and Climate Change","cells":[{"id":5,"value":"# National Economies and Climate Change","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":12,"value":"Climate change is often talked about as a physical science and engineering problem, but the emissions that cause it come from economies and people. Countries differ hugely in how much CO₂ they emit, and those differences are tied with wealth, industry, and trade. One idea in climate policy discourse is that wealthy countries have or are in the process of \"decoupling,\" meaning their economies keep growing while their emissions fall. A counterargument to this is that these counties have just moved their factories overseas, so the emissions still happen, just somewhere else.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":13,"value":"

\n\n## Dataset: Our World in Data's CO₂ and Greenhouse Gas Emissions Data\n\nThis dataset combines emissions estimates from the Global Carbon Project, GDP from the Maddison Project, and population from the UN. Each row is (almost) one country in one year, from 1750 to 2024. There are a few rows that are for larger regions and wealth strata (e.g. Africa and high-income countries), which will be removed for the analyses. It has about 50,000 rows and 79 columns. Most importantly for my questions, it has two different measures of emissions: territorial emissions, which count the CO₂ released inside a country's borders, and consumption-based emissions, which reassign the CO₂ emissions of traded goods to the country that ends up consuming them.\n\nThe data is available from the [owid/co2-data](https://github.com/owid/co2-data) repository.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":8,"value":"const data = aq.fromCSV(await FileAttachment(\"owid-co2-data.csv\").text())\ndisplay(Inputs.table(data));","pinned":true,"hidden":false,"mode":"ts","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":18,"value":"

\n\n## Analysis Questions\n\n**Q1.** How unequal are CO₂ emissions across countries, and how much of that is explained by how wealthy a country is?\n\n**Q2.** Have wealthy countries actually managed to grow their economies while cutting emissions?\n\n**Q3.** If so, is that real, or have they just outsourced their emissions to the countries that make their imports?","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":21,"value":"

\n\n## Data Cleaning\n\nThe country column is poorly labeled, since there are rows that belong to non-countries. Rows without an ISO code are things like continents, income groups, the EU, and \"World,\" plus a few odd entries like international aviation, international shipping, and the 1991 Kuwaiti oil fires. (Kosovo is the one real country without an ISO code, so it will explicitly be included). That leaves 219 countries.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":22,"value":"const countries = data.filter(d => d.iso_code != null || d.country === \"Kosovo\")\nconst aggregates = data.filter(d => d.iso_code == null && d.country !== \"Kosovo\") // everything that's not a country\n\ndisplay(countries.rollup({n: aq.op.distinct(\"country\")}).get(\"n\", 0))","pinned":true,"hidden":false,"mode":"ts","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":24,"value":"

\n\n## Emissions Per Capita by Country (2023)\n\nStarting from individuals, this plot shows the CO₂ emitted per person for every country in 2023 (log scaled).","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":26,"value":"$32","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":27,"value":"The spread is large. The average person in Qatar emitted about 40 tonnes, in the US about 14, in China 8.6, in India 2.1, and the Democratic Republic of Congo under 0.1, which is the lowest. The median country (3.1 tonnes) is below the world average (4.7 tonnes), because a small number of very high emitters pull the average up. The very top is mostly oil states, while the countries with the greatest populations, China and India, are near the middle.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":28,"value":"

\n\n## Share of Emissions by Share of Population\n\nPer-person numbers don't show how much of the world's total each group is responsible for. Sorting countries from lowest to highest per-person emissions gives a Lorenz curve, where the dashed diagonal means everyone emits the same amount.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":29,"value":"$33","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":31,"value":"The half of the world's population living in the lowest-emitting countries produced only about 14% of 2023 emissions, while the tenth living in the highest-emitting countries produced about 30%. One thing to keep in mind is that this data only has national averages, so it treats every American as emitting 14 tonnes. The real inequality between individual people is bigger than this shows, since it ignores the differences between rich and poor people within the same country.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":32,"value":"

\n\n## Emissions vs. Wealth, 2022\n\nTo see if these differences are due to wealth, GDP per person is plotted against CO₂ per person. GDP ends in 2022, so 2022 data for emissions is used here. Both axes are on log scales, which uncovers a fairly linear relationship.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":33,"value":"$34","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":35,"value":"const logFit = (data, xKey, yKey) => {\n const pts = data.filter(d => d[xKey] > 0 && d[yKey] > 0).map(d => [Math.log10(d[xKey]), Math.log10(d[yKey])]);\n const n = pts.length, mx = d3.mean(pts, p => p[0]), my = d3.mean(pts, p => p[1]);\n const sxy = d3.sum(pts, p => (p[0] - mx) * (p[1] - my)), sxx = d3.sum(pts, p => (p[0] - mx) ** 2), syy = d3.sum(pts, p => (p[1] - my) ** 2);\n const slope = sxy / sxx;\n return {n, slope, intercept: my - slope * mx, r: sxy / Math.sqrt(sxx * syy)}\n}\n\nconst wealth2022 = countries\n .filter(d => d.year === 2022 && d.gdp > 0 && d.population > 0 && d.co2_per_capita > 0)\n .derive({gdp_pc: d => d.gdp / d.population})\n .select(\"country\", \"gdp_pc\", \"co2_per_capita\", \"consumption_co2_per_capita\", \"population\")\n .objects()\n\nconst fit2022 = logFit(wealth2022, \"gdp_pc\", \"co2_per_capita\")\n\ndisplay(fit2022)","pinned":true,"hidden":false,"mode":"ts","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":34,"value":"Across the 164 countries with both values, the correlation on the log scales is about 0.91, and the slope is about 1.04, which means a country that is ten times richer emits roughly ten times more per person. Many countries above the center line are oil and gas producers and coal-heavy economies like Trinidad and Tobago, Mongolia, South Africa, and North Korea. Below it are very poor farming economies like Ethiopia, and two wealthy countries known for their green initiatives, Sweden and Switzerland. Switzerland is further below the line than any other country, emitting about 3.7 times less than its income would predict.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":36,"value":"

\n\n## GDP and Emissions Since 1990\n\nThe change in GDP and emissions since 1990 is plotted for a few large economies, with GDP on the left axis and emissions on the right. Both axes start from zero in 1990, so lines above the dotted baseline have grown since then and lines below it have shrunk. Consumption-based emissions are plotted as a dashed line.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":38,"value":"$35","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":39,"value":"There are two different profiles. In the US, UK, Germany, and Sweden, GDP roughly doubled while territorial emissions fell. US emissions kept rising until around 2005 before dropping, and the UK's fell to about half of their 1990 level. In China and India, emissions rose along with GDP, though a bit more slowly; China's GDP grew about eight times and its emissions about five times. In the UK and US, consumption-based emissions grew faster than territorial emissions through the 2000s, so some of the decline may have been imported. Sweden was the opposite, with consumption emissions falling even faster.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":40,"value":"

\n\n## Change in GDP vs. Change in Emissions, 2005–2022\n\nEach country's GDP and emissions in 2005, around when wealthy-country emissions peaked, are compared with 2022. Both axes show growth ratios (2022 value ÷ 2005 value) on log scales, since a few countries' emissions grew so much that a percent-change axis compressed everything else. Countries in the lower right grew their economies while cutting emissions.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":42,"value":"$36","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":43,"value":"45 of the 164 countries are in the lower right. Almost all are in Europe, along with the US, Japan, Canada, New Zealand, and Mexico. The UK is one of the clearest cases, with GDP up 21% and emissions down 45%. Most of the world is in the upper right, where emissions still rise with growth. Some poorer countries had large jumps from small starting points; Laos's emissions grew about seventeen times as it built coal power plants. Decoupling does happen for a meaningful number of countries, but only when counting territorial emissions.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":44,"value":"

\n\n## Territorial vs. Consumption Emissions by Income Group\n\nTo test the outsourcing argument in general, territorial and consumption-based emissions are compared for the World Bank income groups. This is the only analysis that uses the aggregate rows.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":45,"value":"$37","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":47,"value":"For high-income countries, consumption-based emissions exceed territorial emissions, and the gap grew from about 2% in 1990 to about 13% in 2023, roughly 1.8 billion tonnes a year embedded in imports. Middle-income countries show the opposite, with consumption emissions about 12% below production, mostly because China and India manufacture for export. This supports the outsourcing argument, though wealthy-country consumption emissions have also been falling since around 2007.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":48,"value":"

\n\n## Does Decoupling Hold Up After Counting Imports?\n\nOf the 45 decoupled countries, 39 also have consumption-based data. For each, the plot shows the percent change in territorial emissions (circle) and consumption-based emissions (diamond) from 2005 to 2022. Malta's consumption emissions rose 343%, so the axis is cut off at +60%.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":49,"value":"$38","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":50,"value":"27 of the 39 countries remain decoupled after counting imports, including the UK, US, Germany, France, Japan, and Spain, which cut consumption-based emissions by 16–34% while growing. The other 12 are mostly smaller, import-dependent countries such as Switzerland (down 28% territorial, up 13% consumption), Hong Kong, Malta, Jamaica, and the Baltic states. Poland, Mexico, and Australia also fall in this group, since their small territorial cuts (2-3%) were outweighed by rising imports. Outsourcing is real and growing, but it doesn't explain away decoupling for most large wealthy economies.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":51,"value":"

\n\n## Territorial vs. Consumption Emissions per Person, 2023\n\nThis plot compares each country's territorial and consumption-based emissions per person. Countries above the diagonal consume more CO₂ than they produce; countries below it produce more than they consume.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":54,"value":"$39","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":55,"value":"Switzerland emits 3.6 tonnes per person at home but about 13.3 tonnes once imports are counted, around 3.7 times as much. Only Malta (about 8 times) and Hong Kong (about 4.4 times) are more extreme. Sweden and the UK are also clearly above the line, while major exporters like China, India, Vietnam, Russia, and South Africa are below it.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":52,"value":"

\n\n## Historical vs. Current Emissions\n\nWarming depends on the total CO₂ accumulated in the atmosphere, not a single year's emissions. For the ten countries with the largest cumulative emissions, this plot compares their share of all emissions since 1750, their share of 2023 emissions, and their share of world population.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":57,"value":"$3a","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":56,"value":"The US has produced 23.7% of all CO₂ since 1750 with only 4.2% of the world's population. China emits 32% of current emissions but accounts for 15% of the historical total, slightly below its 17.6% population share. India is the starkest contrast: 17.8% of the population, 8% of current emissions, and 3.5% of historical emissions. Germany and the UK, early industrializers, have historical shares (5.2% and 4.4%) well above both their current emissions and population. These figures exclude deforestation and other land use, which would raise countries like Brazil and Indonesia, and use present-day borders.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":53,"value":"

\n\n## Conclusions\n\n**Q1.** Emissions per person differ by a factor of several hundred between countries, and the lowest-emitting half of the world produces about 14% of emissions. Most of this variation is explained by wealth, as emissions rise almost one-for-one with income.\n\n**Q2.** 45 of 164 countries grew their economies while cutting emissions between 2005 and 2022, almost all of them wealthy. For most of the world, emissions still rise with growth.\n\n**Q3.** Wealthy countries consume about 13% more CO₂ than they produce, and the gap has grown since 1990. However, decoupling holds after counting imports for 27 of 39 countries with data, including all large wealthy economies.\n","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W3b","stdlib":"2"}} 31:"$30:body"