Emissions Intensity
Global emissions intensity fell from 0.45 kgCO₂/$GDP (2000) to 0.28 kg (2026)—38% improvement, proving GDP growth can decouple from emissions. China improved 60% (0.80 → 0.32), India 35% (0.50 → 0.33), USA 42% (0.40 → 0.23), EU 50% (0.30 → 0.15). Driven by renewables, efficiency, service economy shift. But need 8%/year decline for 1.5°C vs current 2%/year—progress insufficient.
Key Emissions Intensity Insights
Decoupling GDP from Emissions
Global emissions intensity 0.28 kgCO₂/$GDP (2026), down from 0.45 (2000)—38% improvement. Proof GDP can grow while emissions fall. Global GDP +85% (2000-2026), emissions +52%—relative decoupling. Absolute decoupling (GDP up, emissions down): EU, USA, UK since 2007. Drivers: renewables displacing coal (power sector -120 gCO₂/kWh intensity), energy efficiency (LED, insulation, heat pumps), service economy shift (finance, software less carbon-intensive than steel, cement). But progress too slow: -2%/year vs -8%/year needed for 1.5°C.
China's Dramatic Improvement
China intensity 0.32 kgCO₂/$GDP (2026), down from 0.80 (2000)—60% improvement, fastest among major economies. Coal dominance declining: was 80% electricity (2010), now 60% (2026). Solar/wind additions 230 GW/year offsetting coal. Energy efficiency mandate: close inefficient factories, modernize steel/cement. Economy shifting from heavy industry (exports) to services, consumption. But still 2× global average (0.28)—reflects manufacturing base. If China reaches EU levels (0.15), global emissions -15% even with GDP growth. Per capita rising: 9.0 tCO₂/person (exceeds EU 5.1).
EU Leads Efficiency
EU intensity 0.15 kgCO₂/$GDP (2026), down from 0.30 (2000)—50% improvement, lowest among major economies. Absolute decoupling: GDP +40% (2000-2026), emissions -35%. Drivers: coal phaseout (was 30% electricity in 2000, now 10%), renewables 45% of power, energy efficiency standards (buildings, appliances, cars), carbon pricing €90/tonne ETS. Service economy (75% of GDP)—finance, tech, tourism low-carbon. Manufacturing offshored to China—consumption emissions 20% higher than territorial. Target: climate neutral 2050, on track. Per capita 5.1 tCO₂ (below global 5.0).
Developing World Convergence
India intensity 0.33 kgCO₂/$GDP (2026), down from 0.50 (2000)—35% improvement despite coal dependence (70% of power). Efficiency gains from modernization—older factories retired. But still 2× EU (0.15). Indonesia 0.42, Vietnam 0.38, Nigeria 0.65 (high due to oil flaring, gas leaks). Convergence thesis: as countries develop, intensity falls (shift from agriculture to industry to services). Middle-income trap: countries stuck at 0.30-0.40 range—need leapfrogging (skip coal, go renewables). Africa opportunity: build low-carbon infrastructure from scratch (solar, EVs)—avoid developed world lock-in.
Emissions Intensity Trends by Country (2000-2026)
kg CO₂ per dollar GDP (2020 PPP constant dollars)
Key Finding: All major economies improved. China 0.80 → 0.32 (-60%, fastest decline), India 0.50 → 0.33 (-35%), USA 0.40 → 0.23 (-42%), EU 0.30 → 0.15 (-50%), Japan 0.28 → 0.18 (-36%), Russia 1.20 → 0.45 (-63%, post-Soviet inefficiency → modernization). Global average 0.45 → 0.28 (-38%). Convergence: developed economies 0.15-0.25 range, developing 0.30-0.40, oil states 0.50+ (Saudi 0.55, Russia 0.45). Drivers: renewables (cheaper than coal), efficiency (LED, insulation), deindustrialization (offshoring manufacturing). Need -8%/year for 1.5°C vs current -2%/year.
Decoupling GDP Growth from Emissions (2000-2026)
Indexed to 2000 = 100, showing GDP vs CO₂ trajectories
Key Finding: Global GDP +85% (2000-2026), emissions +52%—relative decoupling. EU GDP +40%, emissions -35%—absolute decoupling (gold standard). USA GDP +60%, emissions -12%—absolute. UK GDP +50%, emissions -40%—star performer. China GDP +520%, emissions +190%—relative decoupling but high absolute growth. India GDP +240%, emissions +165%—weak decoupling. Relative = emissions grow slower than GDP. Absolute = emissions fall while GDP grows. Drivers: renewables (solar, wind), efficiency (buildings, transport), structural shift (services replace manufacturing). Need: absolute decoupling globally for climate goals—emissions must fall -7%/year while GDP grows +3%/year.
Sectoral Emissions Intensity (2026)
kg CO₂ per dollar of sectoral output
Key Finding: Power & heat 1.2 kgCO₂/$output (highest, but declining -3%/year as coal displaced), cement 2.8 kg (process emissions unavoidable), steel 1.5 kg (blast furnace coal-intensive), chemicals 0.9 kg, transport 0.7 kg, agriculture 0.5 kg, buildings 0.3 kg, services 0.05 kg (lowest—finance, software, healthcare). Sectoral intensity drives overall: economies shifting to services (EU, USA) see intensity fall. Manufacturing-heavy (China, Vietnam) higher intensity. Within sectors: power improving fastest (renewables), cement/steel stuck (need hydrogen, CCS). Transport stagnant (EVs small share). Services economy decarbonization easy—finance, IT, education inherently low-carbon.
Energy Efficiency Improvements by Sector (2000-2026)
Energy consumption per unit output, indexed to 2000 = 100
Key Finding: Buildings -35% (insulation, LED lighting, heat pumps, smart thermostats), transport -25% (fuel efficiency standards, EVs), industry -30% (motors, recycling, waste heat recovery), power generation -20% (combined cycle gas, renewables). Residential: LED 85% less energy than incandescent (widespread adoption 2010-2026). Transport: cars 35 mpg (2026) vs 25 mpg (2000). Industry: electric motors efficiency standards saved 15% electricity. But efficiency gains offset by demand growth—rebound effect. Global energy consumption +50% (2000-2026) despite -30% efficiency. Jevons Paradox: cheaper energy increases use. Need: efficiency + absolute energy reduction (behavior change, sufficiency).
Per Capita vs Per GDP Emissions (2026)
Scatter plot showing tCO₂/person vs kgCO₂/$GDP for major economies
Key Finding: EU low on both axes: 5.1 tCO₂/person, 0.15 kgCO₂/$GDP—efficient and low lifestyle emissions. USA high per capita (15.0), low intensity (0.23)—inefficient lifestyles but efficient economy. China moderate per capita (9.0), moderate intensity (0.32)—improving rapidly. India low per capita (2.2), moderate intensity (0.33)—development stage. Russia high on both (12.5, 0.45)—inefficient economy and lifestyles. Gulf states extreme: Qatar 35 tCO₂/person, 0.65 kgCO₂/$GDP—oil wealth, AC, desalination. Correlation: richer countries lower intensity (technology, services) but higher per capita (consumption). Ideal: bottom-left (low on both)—EU trajectory.
Progress Toward Intensity Targets (2000-2026 vs 2030 Goals)
Actual improvement vs Paris-compatible pathways
Key Finding: Global actual -2%/year (2000-2026), need -8%/year (2026-2030) for 1.5°C—massive gap. EU actual -2.5%/year, need -6%/year (closer but still short). USA actual -2.2%/year, need -7%/year. China actual -3.5%/year (best), need -10%/year (industrialization momentum). India actual -1.8%/year, need -5%/year. Gap reflects: insufficient policy (carbon price too low $30 vs $100 needed), fossil fuel subsidies ($7T/year globally), slow infrastructure turnover (coal plants, cars 15-40 year lifespan), rebound effects (efficiency gains offset by demand). Need: carbon pricing, fossil fuel subsidy removal, rapid renewable deployment, efficiency standards, behavior change.
Understanding Emissions Intensity
Key Concepts
What is Emissions Intensity?: CO₂ emitted per unit economic output—typically kgCO₂ per dollar GDP (PPP, constant 2020 dollars). Measures carbon efficiency of economy. Lower = more efficient (less carbon per $ produced). Global 0.28 kgCO₂/$GDP (2026). Alternative metrics: emissions per capita (lifestyle), emissions per kWh (power sector), emissions per tonne-km (transport). Intensity preferred for economic analysis—shows decoupling potential. Per capita better for equity (lifestyle footprint). Both needed: intensity (economic efficiency), per capita (individual responsibility).
Relative vs Absolute Decoupling: Relative = emissions grow slower than GDP (intensity falls, both rise). Absolute = emissions fall while GDP grows (intensity falls faster). Global: relative decoupling (GDP +85%, emissions +52% since 2000). EU/USA/UK: absolute decoupling (GDP up, emissions down since 2007). China: relative only (both rising but intensity improving). Absolute decoupling necessary for climate goals—can't just slow growth, must reverse. Driven by: renewables (zero-carbon GDP), efficiency (more output per energy), structural shift (services replace industry). Critics note: offshoring (manufacturing moves to China—territorial vs consumption emissions).
GDP Measurement (PPP vs MER): PPP (purchasing power parity) adjusts for price differences—$1 buys more in India than USA. MER (market exchange rate) uses currency conversion. IEA, World Bank use PPP for emissions intensity—fairer comparison. China GDP $25T (PPP) vs $18T (MER). Matters for intensity: China 0.32 kgCO₂/$GDP (PPP) vs 0.45 (MER). PPP preferred—reflects real economic activity not currency fluctuations. Constant 2020 dollars eliminate inflation—track real efficiency not price changes. Intensity falling 2%/year real terms (2000-2026).
Territorial vs Consumption Intensity: Territorial = emissions within borders / GDP. Consumption = emissions embedded in goods consumed / GDP. Matters for trade: China exports steel, USA imports—territorial overstates China intensity, understates USA. Consumption-based: China 15% lower intensity, USA 10-20% higher. EU offshored manufacturing to China 1990-2020—territorial emissions -35%, consumption -15%. Carbon border adjustment (CBAM) addresses—tariffs on imports by carbon intensity. Consumption intensity fairer but harder to measure (supply chain data complex). This page uses territorial (IEA standard)—data availability, policy relevance.
Drivers of Intensity Improvement
- Renewables displacement: Solar, wind zero emissions per kWh vs coal 1 kgCO₂/kWh, gas 0.45 kg. Power sector intensity 435 gCO₂/kWh (2026) vs 550 (2010)—21% improvement. Renewables 35% of global electricity (2026) vs 20% (2010). Every % renewable = -0.8% power emissions intensity. Cheapest source now: solar $0.02/kWh, wind $0.03, coal $0.05, gas $0.06. Economics favor transition—no subsidies needed in most markets.
- Energy efficiency: Buildings (insulation, LED, heat pumps), transport (fuel standards, EVs), industry (motors, recycling). Global energy intensity (MJ/$GDP) fell 30% (2000-2026). LED lighting: 85% less energy than incandescent—widespread adoption 2010-2026. Heat pumps: 3× more efficient than gas furnace (COP 3.0). EV 90 mpge vs ICE 35 mpg. But rebound effects: cheaper energy increases use (Jevons Paradox). Absolute energy consumption still rising +1%/year despite efficiency gains.
- Structural economic change: Services (finance, software, healthcare) 10× lower emissions intensity than manufacturing (steel, cement). EU, USA 75% services economy—intensity falls as manufacturing offshored. China shifting: 45% services (2010) → 55% (2026)—helps explain rapid intensity decline. Within sectors: high-tech manufacturing (chips, software) less intensive than heavy (steel, cement). Digital economy inherently low-carbon—data centers 1% of electricity but power GDP growth.
- Fuel switching: Gas replacing coal (power, industry)—50% less CO₂ per kWh. USA coal generation -55% (2010-2026), gas +20%. But methane leaks (3-8% of production) negate climate benefit if >2%. Biomass, waste-to-energy (counted zero emissions—regrowth offsets). Controversial: lifecycle emissions often exceed coal (forests take decades to regrow, "carbon debt"). Hydrogen potential: green H₂ (electrolysis) zero emissions but expensive €50/kg vs gray H₂ (gas reforming) €15/kg but emits 10 kgCO₂/kg.
Sectoral Intensity Differences
High-intensity sectors: Cement 2.8 kgCO₂/$output (process emissions—calcination unavoidable without CCS), steel 1.5 kg (blast furnace coal-intensive), chemicals 0.9 kg (high heat, feedstocks), power 1.2 kg (coal dominance in many countries). Low-intensity: Services 0.05 kg (finance, software, healthcare, education), agriculture 0.5 kg (methane-heavy but dispersed), buildings 0.3 kg (mostly electricity, improving as grids clean). Middle: Transport 0.7 kg (oil dependence, EVs small share). Sectoral decarbonization pathways: power easiest (renewables), buildings medium (heat pumps, efficiency), transport hard (aviation, shipping), industry hardest (cement, steel need hydrogen/CCS). Economy-wide intensity falls as service share rises—but can't offshore all manufacturing (need steel, cement for infrastructure).
Country Intensity Profiles
Low intensity (0.10-0.20 kg): EU 0.15, UK 0.12, Japan 0.18, France 0.10 (nuclear-heavy). Driven by: renewables, efficiency, services economy, historical deindustrialization. Medium (0.20-0.35 kg): USA 0.23, China 0.32, India 0.33, global average 0.28. Balanced: some heavy industry, growing services, improving efficiency. High (0.35-0.65 kg): Russia 0.45 (oil/gas economy, inefficient Soviet legacy), Saudi 0.55 (oil, desalination, AC), Indonesia 0.42 (coal power, deforestation), Nigeria 0.65 (gas flaring, oil leaks). High reflects: fossil fuel dependence, industrial focus, inefficiency. Convergence: as countries develop, intensity falls (technology diffusion, structural change). But also divergence: oil states stuck high (Dutch disease—oil wealth delays diversification).
Limitations & Criticisms
GDP quality: Not all GDP equal—building weapons increases GDP but no welfare. Services GDP (haircut, legal advice) low-carbon but doesn't reflect "real" economy. Manufacturing GDP creates physical goods, emissions. Intensity metric rewards financialization, penalizes making things. Alternative: emissions per capita (lifestyle), emissions per material output (tonnes steel). Offshoring: EU, USA intensity fell partly by moving manufacturing to China—territorial vs consumption. Consumption-based EU emissions only -15% vs -35% territorial. Globalization lets rich countries "launder" emissions. Carbon border adjustment (CBAM) addresses but controversial (trade war risk). Rebound effects: Efficiency makes energy cheaper, increases use—Jevons Paradox. Cars more efficient but bigger (SUVs), more miles driven. Homes insulated but larger, more heated. Intensity falls but absolute emissions flat. Need: efficiency + sufficiency (behavior change, reduced consumption). Inequality: Country average hides distribution—top 10% emit 5-10× bottom 50%. Intensity metric doesn't capture: billionaire private jets, mansions vs median household. Per capita metric also flawed—need distributional analysis.
Pathways to Low Intensity
1.5°C pathways (IPCC) require -8%/year intensity decline (2026-2030) vs current -2%/year—4× acceleration. Mechanisms: (1) Renewables scale-up: Solar/wind 80% of electricity by 2030 (currently 35%), coal phaseout by 2040. Reduces power intensity to 100 gCO₂/kWh (from 435). (2) Electrification: EVs 60% sales by 2030 (currently 20%), heat pumps replace gas furnaces, industrial heat (where possible). Shifts emissions to power sector, then decarbonizes. (3) Efficiency: +3%/year energy productivity (currently +1%). Buildings retrofits, appliance standards, transport fuel economy. (4) Hydrogen, CCS: Green H₂ for steel, cement, chemicals—eliminates process emissions. CCS 5 Gt/year by 2050 (currently 0.05 Gt). (5) Structural change: Service economy growth, circular economy (recycling, reuse—60% less material/emissions). But limits: can't offshore all industry, services need physical goods. Need: absolute emissions cuts, not just intensity improvements.
Data Sources
Primary: IEA CO₂ Emissions from Fuel Combustion (territorial emissions by country), World Bank GDP (PPP, constant 2020 dollars), Global Carbon Project (validation). Sectoral: EXIOBASE (input-output tables, sectoral intensity), IEA Sectoral Approach (power, transport, industry detail), EDGAR (monthly emissions updates). Consumption-based: GTAP (Global Trade Analysis Project—trade flows), Eora MRIO (multi-region input-output), Peters et al. (consumption emissions research). Efficiency: IEA Energy Efficiency Report (annual), ACEEE (buildings, appliances), ICCT (transport). Discrepancies: territorial vs consumption ±20%, GDP PPP vs MER ±30%, sectoral allocation methods differ. Cross-validation essential.