Scope 3 emissions data out by ‘up to 2,480%’ warns research [updated]

New analysis suggests investors’ greenhouse gas (GHG) emissions data could deviate from reality by more than 2,000%.

A comparison of estimated and verified Scope 3 emissions data for FTSE100 companies found the former was out by “up to 2,480%”.

The estimates were calculated using a popular technique known as an Environmentally Extended Input Output (EEIO) model.

EEIO models are promoted by influential bodies, including the Partnership for Carbon Accounting Financials (PCAF), whose partners include the Net Zero Asset Owner Alliance, for use in instances where primary data is not available.

In its online guidance, the GHG Protocol describes them as a way “to estimate cradle-to-gate GHG emissions for a given industry or product category”.

But London-based data house Carbon Responsible says the technique is outdated and leads to inaccurate information about Scope 3 emissions – those generated by an entity’s value chain, rather than from the assets it directly owns and operates.

“It tends to grossly overestimate the size of the emissions,” said Matt Paver, Carbon Responsible’s chief operating officer.

“When we compared it with verified, directly-measured emissions, we found the EEIO model inflated them by more than 2,000% in certain cases.”

Paver told IPE that some mining, minerals and fossil-fuel companies saw even higher overestimations in the study, but were removed from the results because they were outliers.

“From a financial risk perspective, it shows investors are making decisions using inaccurate data,” Paver continued.

“So they might end up with something that doesn’t actually have the sustainability characteristics they’ve been told.”

In addition, he said, a lot of financial institutions in Europe have made official Scope 3 reduction commitments, “and it’s likely some of those targets are based on these estimation models”.

Not only does the potential inaccuracy of the underlying figures present reputational and regulatory risks, Paver warned, but it also allows the entity to meet their targets by simply “waving a magic wand” and updating their data models, bringing their emissions down significantly without any decarbonisation efforts.

“EEIO models were developed six or seven years ago, when the understanding of Scope 3 emissions was limited, and companies didn’t have relationships with their suppliers that allowed them to acquire the data directly,” he said, adding that the approach should be rethought.

EEIO models ‘help fill gaps’

Angélica Afanador, executive director at PCAF, told IPE it was “fully aware that EEIO models can be imprecise”.

“However, it is absolutely incorrect to state that PCAF promotes them,” she said. “EEIO models have the lowest data quality score and are often used as a starting point in the greenhouse gas accounting journey.

“They also help fill gaps when higher-quality data is unavailable and are useful for conducting hotspot analyses. This allows financial institutions to identify asset classes and sectors where they should focus their efforts on obtaining better quality data and managing emissions more effectively.”

This article was updated to incorporate comments from PCAF that were received after initial publication.

The latest digital edition of IPE’s magazine is now available