The Greenhouse Gas (GHG) Protocol Corporate Standard — first published in 2001, revised in 2004, with the Scope 3 Standard added in 2011 — divides corporate emissions into three nested scopes. This architecture was designed to prevent double-counting within a company's boundary, though as we will see, it creates significant double-counting across company boundaries.
flowchart TB
subgraph Company["Company Operational Boundary"]
direction TB
S1["🏭 SCOPE 1\nDirect Emissions\n\n• Fuel combustion (boilers, furnaces, vehicles)\n• Process emissions (cement calcination, steel smelting)\n• Fugitive emissions (refrigerant leaks, methane from pipelines)\n• Owned/controlled sources only"]
S2["⚡ SCOPE 2\nIndirect — Energy\n\n• Purchased electricity\n• Purchased steam, heat, cooling\n• Location-based vs market-based methods\n• Typically 10–30% of industrial footprint"]
end
subgraph VC["Value Chain (Scope 3)"]
direction LR
UP["⬆️ UPSTREAM\nCat 1–8\n\nWhat you buy\nand how you buy it"]
DOWN["⬇️ DOWNSTREAM\nCat 9–15\n\nWhat happens after\nyou sell it"]
end
Company --> VC
UP -.->|"Typically\n70–90%\nof total footprint"| DOWN
style S1 fill:#ff6b6b,color:#fff
style S2 fill:#ffa94d,color:#fff
style UP fill:#4dabf7,color:#fff
style DOWN fill:#69db7c,color:#000
| Scope | Rationale | Control Level | Typical % of Total |
|---|---|---|---|
| 1 | Direct operational control | Full | 5–15% |
| 2 | Purchased energy — contractually controllable | Partial | 5–20% |
| 3 | Value chain — influenceable but not controlled | Indirect | 70–90% |
The Protocol's logic: companies should fully account for what they directly control (Scope 1), should account for the carbon embedded in energy they choose to buy (Scope 2), and should also disclose the broader value chain impact (Scope 3) — even though they cannot directly control it.
This last requirement is where the complexity explodes.
The GHG Protocol's Scope 3 Standard (2011) defines 15 distinct categories, split into upstream (1–8) and downstream (9–15).
flowchart LR
subgraph US["UPSTREAM SCOPE 3"]
C1["Cat 1\nPurchased Goods\n& Services"]
C2["Cat 2\nCapital Goods"]
C3["Cat 3\nFuel & Energy\nRelated Activities"]
C4["Cat 4\nUpstream\nTransportation\n& Distribution"]
C5["Cat 5\nWaste Generated\nin Operations"]
C6["Cat 6\nBusiness Travel"]
C7["Cat 7\nEmployee\nCommuting"]
C8["Cat 8\nUpstream\nLeased Assets"]
end
subgraph DS["DOWNSTREAM SCOPE 3"]
C9["Cat 9\nDownstream T&D"]
C10["Cat 10\nProcessing of\nSold Products"]
C11["Cat 11\nUse of\nSold Products"]
C12["Cat 12\nEnd-of-Life\nTreatment"]
C13["Cat 13\nDownstream\nLeased Assets"]
C14["Cat 14\nFranchises"]
C15["Cat 15\nInvestments\n(Financed Emissions)"]
end
style C1 fill:#339af0,color:#fff
style C2 fill:#339af0,color:#fff
style C3 fill:#339af0,color:#fff
style C4 fill:#339af0,color:#fff
style C5 fill:#339af0,color:#fff
style C6 fill:#74c0fc,color:#000
style C7 fill:#74c0fc,color:#000
style C8 fill:#74c0fc,color:#000
style C9 fill:#51cf66,color:#000
style C10 fill:#51cf66,color:#000
style C11 fill:#2f9e44,color:#fff
style C12 fill:#51cf66,color:#000
style C13 fill:#74c0fc,color:#000
style C14 fill:#74c0fc,color:#000
style C15 fill:#2f9e44,color:#fff
The single largest Scope 3 category for most companies. All emissions associated with producing the goods and services a company buys — raw materials, components, professional services, IT, travel, consumables.
Why it's hard:
- Requires emissions intensity data from every supplier
- Most companies have thousands of suppliers across dozens of countries
- Emission factors vary enormously by production geography, technology vintage, and energy mix
- Services (consulting, legal, IT) have poorly defined emission factors
Calculation methods (in decreasing order of accuracy):
flowchart TD
M1["Supplier-Specific Method\n(Primary Data)\n★★★★★ Accuracy\nDirect emissions data from suppliers\nrequires supplier capacity & trust"]
M2["Hybrid Method\n★★★★☆ Accuracy\nCombines primary supplier data\nwith secondary emission factors\nfor remaining gaps"]
M3["Average-Data Method\n★★★☆☆ Accuracy\nIndustry-average emission factors\nper unit of product (kg CO₂e/kg steel)"]
M4["Spend-Based Method\n★★☆☆☆ Accuracy\nEmission factors per $ of spend\n(kg CO₂e/$ in sector X)\nHighest error, easiest to compute"]
M1 --> M2 --> M3 --> M4
style M1 fill:#2f9e44,color:#fff
style M2 fill:#94d82d,color:#000
style M3 fill:#ffa94d,color:#000
style M4 fill:#ff6b6b,color:#fff
Spend-based error range: A spend-based factor for "basic iron and steel manufacturing" might be 0.8–2.4 kg CO₂e per dollar, depending on geography and the emission factor database used. Actual intensity can vary 3–10x between a coal-powered Chinese steel mill and a hydrogen-reduced Swedish steel plant — both of which look identical in a spend-based model.
Emissions from producing capital equipment — machinery, buildings, vehicles, IT infrastructure. Often overlooked because capex is episodic, but can be material for capital-intensive industries.
Key complexity: How do you allocate the lifetime emissions of a piece of equipment to a single reporting year? The GHG Protocol says to attribute all capital good emissions to the year of acquisition — which creates lumpy, potentially misleading annual disclosures.
Upstream emissions from fuel production (extraction, refining, transport) not already captured in Scope 1 or 2. Includes:
- Well-to-gate emissions for purchased fuels
- Transmission and distribution losses for electricity
- Extraction and transport of fuels used to generate purchased electricity
This is sometimes called "Scope 2.5" — it fills gaps between Scope 2 (combustion of purchased energy) and the full lifecycle of that energy.
Emissions from transporting and distributing purchased goods — from supplier to the company's facility. Includes third-party logistics, rail, sea freight, air freight.
Key tension: The company controls which logistics providers it uses, but has limited data on their actual fuel consumption and load factors. Shipping companies vary enormously in carbon intensity per tonne-kilometer.
Emissions from waste disposal — landfill, incineration, wastewater treatment — for waste generated at company-operated facilities. Often a small category, but can be significant in food processing, chemical manufacturing, and construction.
Employee travel by air, rail, car (non-company-owned vehicles), and hotels. Typically easier to measure (travel management systems hold booking data), but radiative forcing from aviation contrails is a significant area of scientific debate and methodological inconsistency.
Radiative forcing factor: Aviation emissions at altitude have an additional warming effect beyond CO₂ from fuel combustion — the IPCC estimates a radiative forcing index (RFI) of 2–4x. Some companies apply this; most do not. The GHG Protocol does not require it.
Emissions from employees traveling to and from work. Requires survey-based estimates of commute distance, mode, and frequency. Work-from-home policies complicate this significantly — does remote work move emissions from Cat 7 to Cat 11 (home energy use)?
Emissions from assets leased by the company (as lessee) not captured in Scope 1/2. Relevant primarily for companies with large leased fleets, data centers in colocation facilities, or leased retail space.
Emissions from transporting sold products from the point of sale to end customers and between intermediaries. Particularly material for e-commerce companies, consumer goods, and distributors.
Emissions from processing intermediate products sold to other businesses — e.g., a chemical company selling feedstocks that customers then process further. Requires understanding customers' manufacturing processes.
The category that defines energy-transition-era Scope 3.
For any company selling products that consume energy during their lifetime — cars, appliances, electronics, HVAC systems, industrial equipment — Category 11 captures lifetime operational emissions from those products in customers' hands.
Why this is uniquely difficult:
flowchart LR
subgraph Assumptions["Assumptions Required for Cat 11 Calculation"]
A1["Product\nLifetime\n(years)"]
A2["Annual\nEnergy Use\n(kWh/year)"]
A3["Grid Carbon\nIntensity\n(current & future)"]
A4["Usage\nPatterns\n(actual vs rated)"]
A5["Fleet\nMix\n(product variants)"]
A6["Geographic\nDistribution\nof sales"]
end
Assumptions --> R["Cat 11\nEmissions\nEstimate\n\n±50–200%\nuncertainty range"]
For an automotive OEM, Cat 11 is typically 85–95% of total Scope 3 and 75–85% of the company's entire GHG footprint (Scopes 1+2+3). Reducing Cat 11 requires transitioning to EVs — but even EVs' Cat 11 depends on grid decarbonization that is outside the manufacturer's control.
Emissions from disposal or recycling of sold products at end of life. Companies must estimate product retirement rates, disposal methods by geography, and emissions from various waste treatment processes.
Emissions from assets owned by the company (as lessor) and leased to others. Real estate investment trusts (REITs) and commercial property companies face this as a major category.
Emissions from operations of franchisees. Material for companies like fast food chains (McDonald's, Subway), hotel chains, and retail franchises — where the corporate entity owns the brand but franchisees own the operations.
Key problem: Corporate cannot mandate emissions reporting standards on independent franchisees without creating legal risk around control/employment classification.
Emissions attributable to a company's equity investments, debt financing, and other financial services. This is the Scope 3 category that applies to banks, asset managers, insurance companies, and institutional investors.
Governed primarily by the Partnership for Carbon Accounting Financials (PCAF) standard, which extends the GHG Protocol for financial institutions.
flowchart TD
Bank["Commercial Bank\n$500B loan portfolio"]
subgraph Portfolio["Financed Emissions Calculation (PCAF)"]
P1["Corporate Loans\nAttribution = (Outstanding Loan / Enterprise Value) × Borrower Emissions"]
P2["Project Finance\nAttribution = (Outstanding Loan / Total Project Costs) × Project Emissions"]
P3["Mortgages\nAttribution = (Outstanding Mortgage / Property Value) × Property Emissions"]
P4["Listed Equity & Bonds\nAttribution = (Investment Value / Enterprise Value) × Investee Emissions"]
P5["Business Loans < $1M\nAttribution = Sector emission factor × Revenue"]
end
Bank --> Portfolio
Portfolio --> Total["Bank's Scope 3 Cat 15\n(Financed Emissions)\n\nTypically 100–1000x\nlarger than Scope 1+2"]
style Total fill:#e03131,color:#fff
For a major commercial bank, financed emissions are typically 100–1,000 times larger than the bank's own operational emissions (Scopes 1 and 2). This means a bank's net-zero commitment is essentially meaningless without a credible plan to decarbonize its loan book.
Not every category is material for every company. The GHG Protocol requires companies to assess and disclose all material categories. What "material" means is frustratingly vague — the Protocol says a category is material if "it comprises a significant portion of total Scope 3 emissions, contributes to [company's] GHG risks, is important for reporting purposes, or is considered material by stakeholders."
Note: The table below shows dominant category and share of total Scope 3 — not a single-axis comparison, as each sector's profile is multi-dimensional.
| Sector | Dominant Category | % of Total Scope 3 | Secondary |
|---|---|---|---|
| Automotive OEM | Cat 11 (Use of sold products) | 85–95% | Cat 1 (steel, aluminum) |
| Apparel / Fashion | Cat 1 (Purchased goods) | 60–80% | Cat 4 (shipping) |
| Commercial Bank | Cat 15 (Financed emissions) | 95–99% | Cat 6 (business travel) |
| Food & Beverage | Cat 1 (Ag raw materials) | 50–80% | Cat 11 (cooking) |
| Software / Tech | Cat 1 (Hardware supply chain) | 40–70% | Cat 3 (data center energy) |
| Oil & Gas | Cat 11 (Combustion of sold fuels) | 70–90% | Cat 1, Cat 4 |
| Cement | Cat 1 + Cat 3 | 30–50% | Cat 11 |
| Real Estate | Cat 13 (Downstream leased) | 70–90% | Cat 2 (construction) |
Scope 3 is designed to be a company-level accounting tool, not a system-level one. When you sum all companies' Scope 3 disclosures, you count most emissions multiple times.
flowchart LR
subgraph Chain["Value Chain: Iron Ore → Steel → Car"]
Mine["Iron Ore Mine\n\nScope 1: 100t CO₂e\n(mining operations)"]
Steel["Steel Mill\n\nScope 1: 800t CO₂e\n(smelting)\nScope 3 Cat 1: 100t\n(purchased ore — same\natoms as Mine Scope 1)"]
Auto["Auto OEM\n\nScope 1: 50t CO₂e\n(assembly)\nScope 3 Cat 1: 900t\n(steel+components — overlaps\nMine S1 + Mill S1)\nScope 3 Cat 11: 5,000t\n(lifetime vehicle use — same\nemissions as Driver S1)"]
Driver["Driver\n\nScope 1: 200t CO₂e\n(fuel combustion\nover vehicle life)"]
end
Mine --> Steel --> Auto --> Driver
Note["Physical system total (actual CO₂e emitted):\nMine 100t + Mill 800t + Assembly 50t + Driving 200t = 1,150t\n\nSum of all companies' Scope 3 disclosures:\n~6,100t — because the same physical\nemissions appear in multiple companies' accounts.\n\nThis is by design: Scope 3 is a company-level\nrisk tool, not a system-level inventory."]
style Note fill:#fff3bf,color:#000
This is not a bug in the Protocol — it is a deliberate design choice. Scope 3 is intended to give each company visibility into its value chain risk, not to produce system-level accounting. But it creates enormous confusion when aggregating disclosures, setting sector-level targets, or comparing company disclosures.
The implication for policy and markets: you cannot simply add up Scope 3 disclosures to understand economy-wide emissions. National GHG inventories (using the IPCC methodology) are the authoritative system-level account — and they do not map cleanly onto corporate Scope 3 boundaries.