Electricity Maps has been working on marginal emission factors since 2018, when we developed our first algorithm using machine learning. We wrote an explanation of what marginal emissions are in 2019, wrote about their applicability to real-time decisions in 2022 and more recently wrote about our latest innovations in this space. However, we never wrote about its applicability to Scope 2 accounting.

A number of stakeholders are promoting a shift in global Scope 2 carbon accounting rules towards a system based on marginal emission factors. The proposal, going under the name of “impact accounting” or “emissionality” has the ambition to be an alternative to the location-based and market-based method of Scope 2 accounting. The Greenhouse Gas Protocol’s existing Scope 2 Guidance is clear on the topic, stipulating p. 53 that “companies shall not use marginal emission factors” for this type of accounting.

However, it seems worth investigating why “impact accounting” isn’t compatible with Scope 2 accounting. We’ll see that “impact accounting” can lead to claiming more emissions than physically emitted by the grid. Furthermore, we’ll see that it conflicts with the definition of Scope 1 and Scope 2. This makes marginal emission factors unsuitable for accounting for emissions of a company’s inventory both now and in the future. This article assumes the reader is familiar with marginal emission factors.

A refresher of “impact accounting”

The proposed idea is to base an accounting system on marginal emission factors by defining the footprint of a consumer as the difference between the emissions that are induced by their electricity consumption and the emissions that are avoided by their purchases of clean energy. The concept has been advocated by two white papers each written by WattTime and Resurety, originally based on work made by Aleksandr Rudkevich, with an open version available here and a condensed slide version here. More formally, the following rules are used:

1. Firstly, consumers are attributed emissions by assigning the grid’s marginal emission factor to their consumed electricity. These are called induced emissions.
2. Second, generators are attributed emissions by assigning to their generated electricity the difference between the grid’s marginal emission factor and the generator’s emission factor. These are called avoided emissions.
3. Consumers are allowed to procure electricity certificates (RECs, GOs, EACs..) from generators. When they do, they can deduct the avoided emissions of their purchase from their induced emissions.

The “impact accounting” Scope 2 footprint of a consumer is then calculated as the difference between their induced emissions, and the avoided emissions of their procured certificates:

Putting it into practice with a simple example

Let’s take a simple example of a grid which has a wind farm and a coal power plant, with two consumers. The coal power plant is the only power plant with spare capacity, and is thus the marginal power plant.

We assume there is 2 MWh of wind power production and 1 MWh of coal production during a particular hour. The marginal power plant is the coal power plant, with an emission factor of 800 kgCO2e per MWh generated. There are two consumers on the grid. One consumes 1 MWh of power and the other 2 MWh.

Let’s calculate the avoided and induced emissions of each entity on the grid, as well as the “impact accounting” Scope 2 footprint of consumers.

Consumers can purchase certificates from specific generators. In the following example, the second consumer procures all the certificates of wind generated during this particular hour. That consumer can now deduct the avoided emissions of the wind generator, thus reaching an “impact accounting” Scope 2 footprint of 0 kg.

Issue #1: avoiding more than physically emitted

In this simplified example, 800 kg of CO2 are physically emitted into the atmosphere through the process of burning coal. However, this accounting system states that the wind turbine avoided 1600 kg, which is twice that amount. That claim is then purchased by a consumer, who is claiming to have avoided more than is physically emitted on the grid. How is this possible?

The discrepancy comes from the fact that a single marginal emission factor is used across all entities of the grid. In reality, the coal power plant has a limited capacity, and therefore, once all of its physical emissions (800 kg) have been avoided, there is nothing else left to avoid. Marginal emission factors are normally only reserved for small changes on the grid, and shouldn’t be applied to all consumers at such a scale. Using a single marginal emission factor results in an accounting system that departs from physical reality by allowing consumers to claim they’ve avoided more emissions than physically possible.

Issue #2: attributing Scope 2 emissions to *generators*

Scope 1 emissions are direct emissions made by generators, whereas Scope 2 emissions are indirect emissions caused by consumers using electricity from generators. Direct Scope 1 emissions from generators always match Scope 2 emissions from consumers. However, in our simplified example without any procurement, you’ll notice that the “impact based” Scope 2 footprint of consumers (800+1600=2400 kg) is not equal to the Scope 1 emissions from generators (800 kg). To fix this discrepancy, “impact accounting” not only attributes Scope 2 emissions to consumers, it also attributes them to generators. This is confusing enough that it bears repeating: “impact accounting” assigns Scope 2 emissions to generators in order to ensure total Scope 2 emissions match total Scope 1 emissions.

Let’s add the “impact accounting” Scope 2 footprint of generators:

Assigning negative “impact based” Scope 2 emissions to the wind generator balances out the Scope 2 emissions surplus of consumers (which, as we saw previously, are larger than what is physically emitted by the grid), and ensures that total Scope 2 emissions, which now also have to include generators, sum up to Scope 1 emissions.

Introducing negative Scope 2 emissions and assigning them to generators in order to balance the equation moves us even further away from physical reality. Not only are Scope 2 footprints not meant to be negative, but the original intention with Scopes 1 and 2, as defined in the Greenhouse Gas Protocol’s Corporate Guidance, are to segment direct generator emissions (Scope 1) and indirect consumers emissions (Scope 2). Creatively redefining the boundaries of Scope 2 to include generators constitutes a significant incompatibility with modern carbon accounting practices.

Conclusion

The use of marginal emission factors for accounting emissions of a company’s inventory leads to conclusions that depart from physical reality, and conflicts with the original definition of Scopes 1 and 2. That may very well be why the Greenhouse Gas Protocol’s Scope 2 Guidance “does not support an “avoided emissions” approach for Scope 2 accounting due to several important distinctions between corporate accounting and project-level accounting” (p 28). This shouldn’t be surprising as marginal emission factors are a form of consequential accounting, whereas Scope 2 accounting is a form of attributional accounting (read more here).

Does that mean marginal emission factors should never be used for any accounting purposes? In a recent publication, the Greenhouse Gas Protocol writes that “however, companies can report avoided grid emissions from energy generation projects separately from the Scopes using a project-level accounting methodology”. It has to be noted however that marginal emission factors are still subject to debates due to difficulties in verifying their accuracy. 50 Hertz, the East German Transmission System Operator, writes that “Determining the marginal power plant is extremely difficult” and PJM, a grid operator in the United States, warns that their marginal data comes without “any guarantees as to the accuracy of the information”. This difficulty is confirmed by our own research, which showed that a slight change in model assumptions drastically changed the estimated consequence of an action on the grid.

Marginal emission factors may well have a future – but it’s probably not within the realm of Scope 2 accounting.

Further reading

• Understanding Electricity Scope 2 Attribution Rules
• Marginal signals in practice