The double-counting problem in Scope 4: why most avoided CO₂ emissions figures don't hold up to investor scrutiny

By Daniel Lobo, VP, Investors at Upright

In brief:

• The problem: No mandatory standard means every company uses a different methodology. Figures are incomparable across holdings and impossible to aggregate meaningfully.
• The consequence: Double-counting is endemic. For example, component manufacturers and final assemblers routinely both claim full credit for the same avoided tonne of CO₂. 
• The solution: Product-level modelling with value-chain attribution is the only approach that produces figures rigorous enough to use.

Investors want to measure Scope 4 – but data hasn’t kept up

Avoided emissions have moved from nice-to-have to a core portfolio metric. The question is no longer whether to measure them – it's how to do it robustly.

The momentum is visible:

• In April 2025, Robeco, Mirova, Amundi and nine other asset managers launched the Avoided Emissions Platform to harmonise calculations across 65 climate solutions.
• Schroders and GIC published a joint framework integrating avoided emissions directly into portfolio analysis.
• Janus Henderson and UC Berkeley have argued that Scope 1–3 fundamentally fails to capture what innovative companies contribute.

The direction of travel is settled. What isn't is the data quality problem underneath it. For most investors right now, the figures they can access are not defensible.

Three reasons most avoided emissions data breaks down

Every avoided emissions figure rests on a counterfactual: what would have been emitted if this product didn't exist? Getting that right is hard.

Investors face three compounding problems that make portfolio-level analysis nearly impossible.

• The disclosure gap and incompatibility problem. Most companies still don't report avoided emissions at all. Those that do apply their own methodology. You cannot compare figures across holdings or aggregate them at the portfolio level without assumptions that undermine the whole exercise.
• High-level sector proxies that miss the point. Most mainstream ESG data uses top-down sector classifications. A high-efficiency manufacturer and a standard incumbent end up with the same number. For identifying who is actually driving decarbonisation, this is close to useless.
• Double-counting baked in. Without value-chain attribution, a component manufacturer and a final assembler each claim 100% of the same avoided tonne. Portfolio-level figures inflate accordingly – and "has a bit of a reputation," as one Nordic PE fund manager put it to us recently, is now the polite way to describe the problem.

To move from marketing claims to decision-useful data, investors need a unit of analysis that is granular, revenue-linked, and benchmarked against leading scientific frameworks like the WBCSD Guidance on Avoided Emissions and Project Frame.

Why the unit of analysis has to be the product, not the company

To achieve objective comparability, the analysis must shift from the company as a whole to its granular revenue-generating components. Product-level modelling resolves this because it has four properties that company-level figures structurally cannot achieve:

• Scientific traceability. Instead of accepting a company's self-reported claim, product-level data is cross-referenced with independent, public sources like Life Cycle Assessments (LCAs) and Environmental product declarations (EPDs). This anchors the impact in physical reality.
• Pathway-specific accuracy. Avoided emissions occur through specific mechanisms, such as substitution (replacing a dirtier alternative), efficiency improvements, or end-of-life avoidance. Analysing these at the product level allows for precise counterfactual modeling.
• Scaling with economic reality. By multiplying a product’s revenue by its specific avoided intensity (tons of CO2 avoided per dollar of revenue), the model reflects the actual real-world scale of the solution.
• Objective baselines: Utilising a consistent, external database of CO2 emissions intensities ensures that all companies are being measured against the same standardised benchmarks, creating a truly comparable portfolio view.

How Upright avoids the double-counting trap

Upright's avoided emissions module applies all four properties above from the bottom up – product by product, pathway by pathway – benchmarked against the WBCSD Guidance on Avoided Emissions and Project Frame. 

A key piece that most data providers do not get right is avoiding double-counting – a core principle built into Upright’s model since 2017.

To solve the double-counting problem, Upright utilises a two-step value chain allocation process to ensure realistic outcomes:

• Independent product: If a product achieves the vast majority of the impact on its own, it receives a higher attribution.
• Contributing product: If a product works alongside other solutions to enable the impact, it receives a standard 50% allocation or a specific percentage based on its relative value-add.

This consistency is what makes portfolio-level comparison possible – and what gives LPs a genuinely independent lens on GP impact claims.

What Upright's approach gives you

Upright’s approach gives you an objective view of a portfolio's climate contribution. You get:

• The same methodology applied across any company universe – listed, private, unlisted – so figures are comparable without reconciliation.
• A consistent, externally produced view that gives LPs an independent lens on GP impact claims.
• Every calculation benchmarked against the WBCSD Guidance on Avoided Emissions, Project Frame, and the Carbon Handprint Guide.

If you're working through this for your portfolio, we're happy to show you how it works in practice.