A kilowatt-hour does not carry one permanent carbon label. Its reported emissions depend on where electricity was generated, when it was used, which power plants are included and whether the calculation describes an average grid mix or a change caused by new demand.
Key takeaways
- An annual grid-average factor is useful for inventories but can hide large hourly and seasonal changes.
- Location-based and market-based Scope 2 results answer different accounting questions.
- Average, marginal and life-cycle factors should not be substituted for one another.
- A credible claim states the factor’s year, geography, unit, source and system boundary.
What an electricity emissions factor measures
An electricity emissions factor converts electricity use into greenhouse-gas emissions. The basic calculation looks simple:
Electricity used × emissions factor = estimated emissions
If a facility uses 1,000 kWh and applies a factor of 0.50 kg CO₂ per kWh, the estimate is 500 kg CO₂. But the arithmetic is the easy part. The difficult question is whether 0.50 is the right factor for that decision.
The IEA’s Emissions Factors 2025 database includes annual country factors for electricity and heat generation, electricity-only factors, provisional recent-year values, grid-loss corrections and trade corrections. Each series has a specific purpose. Selecting a number without reading the documentation can change the result materially.
Five choices that change the number
1. Geography
A national average blends every connected generator in a country. A regional or utility factor may better represent the network serving a facility, while an island or isolated grid can have a very different mix from the national system. Cross-border electricity trade can also complicate the picture.
2. Time
An annual factor averages dry and wet seasons, daylight and night, weekday and weekend. A power system with large shares of solar may have lower operational emissions around midday than after sunset. Hydropower output can change with rainfall, while extreme heat can increase cooling demand and bring different generators online.
The IEA’s real-time emissions factor methodology combines frequently updated generation and demand information with annual emissions data to estimate hourly carbon intensity. Hourly data are useful for analysing timing, but they are not automatically the correct input for every corporate inventory.
3. Average or marginal
An average factor divides total grid emissions by total electricity supplied over a period. It describes the mix as a whole. A marginal factor estimates which generators increase or decrease output when demand changes. That makes marginal factors relevant to questions such as the short-run effect of shifting a flexible load from one hour to another.
The two measures can move differently. A system may have a low annual average because of hydropower or solar, yet rely on a fossil-fuel plant at the margin during a peak hour. Marginal estimates require dispatch data or models and should be presented with uncertainty rather than treated as a directly observed meter value.
4. Direct or life-cycle emissions
Operational factors usually count emissions released while fuel is burned at power plants. Life-cycle factors can also include fuel extraction, processing, transport, construction and equipment manufacturing. Renewable generators have no fuel-combustion emissions during operation, but they are not literally zero under a full life-cycle boundary.
This distinction is central to comparisons such as electric vehicles. Our article on EV life-cycle emissions in Southeast Asia explains why vehicle production, electricity generation and driving distance must be separated before the totals are compared.
5. Electricity delivered or generated
A factor measured at the power plant may not include electricity lost before it reaches the customer. Some datasets provide a grid-loss correction so emissions are assigned to delivered electricity. Read our explainer on electricity generated versus electricity delivered for the denominator problem behind that adjustment.
Location-based and market-based Scope 2 accounting
The GHG Protocol Scope 2 Guidance distinguishes two methods for purchased electricity.
- Location-based accounting uses average emissions from the grid where consumption occurs. It describes the physical electricity system associated with the site.
- Market-based accounting uses qualifying contractual information, such as supplier-specific factors or energy attribute certificates, where those instruments and market rules meet the guidance’s quality criteria.
A company operating in a market with product or supplier choice may need to report both totals. The results should not be blended into one number. A low market-based total can reflect procurement instruments while the local grid remains carbon-intensive. Conversely, a location-based result does not capture a specific renewable contract.
A worked example
Consider a hypothetical data centre using 10,000 MWh in one year. These figures are illustrative:
| Method | Factor | Estimated emissions | Question answered |
|---|---|---|---|
| National annual average | 0.55 t CO₂/MWh | 5,500 t CO₂ | What is the average grid association? |
| Regional annual average | 0.42 t CO₂/MWh | 4,200 t CO₂ | What is the average for the serving region? |
| Hourly matched average | Varies by hour | Calculated hour by hour | How does consumption timing align with grid intensity? |
| Market-based factor | Contract-dependent | Depends on valid instruments | What does procurement accounting report? |
None of these values alone proves that the facility caused a particular quantity of emissions to increase. A causal estimate would require a marginal or project-level analysis with assumptions about what generation changed.
Why this matters in Southeast Asia
Southeast Asian power systems differ widely in fuel mix, hydropower availability, grid connections and market structure. National averages can therefore conceal substantial local and seasonal variation. Electricity demand is also changing quickly as cooling, data centres, industry and electric transport expand.
Clean capacity growth does not translate one-for-one into low-carbon electricity at every hour. See our explanation of renewable capacity versus actual generation. For flexible loads, the time of use can matter alongside the annual total; our Thailand demand-flexibility analysis explores that operational question.
Regional interconnection will make accounting more complex. Power can cross borders, while contractual claims and physical flows follow different rules. Transparent disclosure should identify whether imported electricity, transmission losses and certificates are reflected in the factor.
How to choose the right factor
Match the factor to the question:
- Use a documented annual grid-average factor for a standard location-based inventory.
- Use qualifying supplier or contractual data for market-based Scope 2 reporting.
- Use hourly average data to examine time matching and operational patterns.
- Use marginal analysis for changes in demand, dispatch or avoided-emissions questions.
- Use life-cycle factors when comparing technologies across their full supply chains.
Every result should disclose the source, publication year, data year, geography, unit, greenhouse gases included, treatment of trade and grid losses, and whether the factor is average or marginal.
The practical conclusion
An electricity emissions factor is a model of a defined system, not a universal property of one kWh. The right number depends on the claim being made.
Clear boundaries make the calculation useful. Hidden boundaries make precise-looking carbon figures misleading. Before comparing results, compare the methods first.
Image note: The featured image is an original AI-generated editorial illustration created for ClimateCrisis.net. It depicts a generic Southeast Asian electricity control room and is not a photograph of a real facility.