How We Actually Measure Financed Emissions in Retirement Plans
Financed emissions — the carbon tied to where retirement plan dollars are invested — are one of the hardest numbers in sustainability reporting to get right. Here's how we approach the calculation, and why the method matters as much as the result.
Third-party data coverage gaps of 20 to 40% are common even with the most robust sources available. What an organization does with that gap — and how transparent it is about the assumptions involved — determines whether the resulting number is a credible estimate or just a number.
What actually goes into a financed emissions calculation
Two data sources anchor most institutional approaches to this problem. The first data source provides fund-level detail: net asset value, assets under management, holdings, and reported company-level Scope 1 and Scope 2 emissions. The second data source fills part of the gap with sector and subsector industry averages, used as a proxy for the securities that don't report directly.
The distance between what these sources cover and what they don't is where methodology starts to matter — and where most off-the-shelf approaches fall short.
Four ways to close that gap
Through building the Financial Emissions Calculator (FEC), we've implemented, tested, and refined four distinct approaches to estimating total fund emissions. Each makes a different set of trade-offs between precision, scalability, and the resources required to run it. None of them is universally "correct" — the right one depends on what a plan sponsor is trying to defend, and to whom.
| Approach | What it does | Best for | Main trade-off |
|---|---|---|---|
| Relationship-based | Works directly with fund managers to source proprietary estimates for the portion of a portfolio that third-party data doesn’t cover | Concentrated fund lineups with strong existing manager relationships | Most nuanced results, but not scalable — depends on manager willingness to disclose |
| Straight-line scaling | Extrapolates a fund’s reported emissions proportionally, based on its data coverage percentage | A fast first-pass estimate across a large universe of funds | Simple and fast, but assumes uncovered holdings look like covered holdings — often not true |
| Bottoms-up (unscaled) | Builds the estimate security by security, using each holding’s ownership share and reported or industry-average emissions | Equity funds where holding-level precision matters | Captures real industry variation, but doesn’t yet reconcile against reported totals |
| Bottoms-up (scaled) | Starts from the same holding-level build, then calibrates it against the fund’s reported figures for the portion that is directly verifiable | Reporting, disclosures, or any number that needs to hold up to outside scrutiny | The most rigorous and defensible estimate we produce — and the core of how FEC calculates total emissions |
We won't walk through the exact calibration mechanics here — that's the part of FEC we consider our edge, and it's built, tested, and refined specifically so plan sponsors don't have to build it themselves. What we will say: the scaled bottoms-up approach is designed to solve the single biggest weakness of the simpler methods — the assumption that what you can't see looks like what you can.
Choosing the right approach isn't just a technical decision
For an initial, internal read on order of magnitude — is this 10,000 tonnes or 100,000? — straight-line scaling is a reasonable place to start. It's fast, and it's often enough to make the internal case that this category deserves real attention.
For anything that will face outside scrutiny — a sustainability report, a net-zero strategy, an auditor, a board — the bottoms-up scaled approach is the standard we'd recommend. It's the only one of the four that's simultaneously granular, calibrated to real reported data, and traceable back to its sources.
And for organizations with a concentrated set of fund managers and real relationships in place, the relationship-based approach can supplement either bottoms-up method, particularly for holdings in sectors where the uncovered tail is likely to look very different from the covered portion.
The honest caveat
No method here produces a measurement. Each produces an estimate, built on assumptions that are real, disclosed, and worth interrogating. The industry averages used to fill data gaps introduce a margin of error — one that should be stated plainly rather than smoothed over.
What separates a credible estimate from an unreliable one isn't the absence of assumptions. It's transparency about which assumptions were made, and why. That's the standard we hold FEC to, and it's the same standard we'd encourage any plan sponsor to apply before publishing a number.
What's next
Everything above addresses equity funds — the component of a retirement plan where third-party data coverage is most developed and the calculation is most tractable. Fixed income, target-date funds, and alternative vehicles make up a meaningful share of most plans too, and the methodologies for those asset classes are still evolving industry-wide, including through frameworks like PCAF. A complete Scope 3, Category 15 inventory will eventually need to cover all of it. Starting with equities now — and building the underlying data infrastructure in the process — is what makes that fuller picture possible later.
The question that actually matters
The right question isn't "which method gets us the lowest number." It's "which method produces the number we can defend." As Scope 3 reporting expectations continue to tighten across voluntary frameworks and emerging regulation, that distinction is only going to matter more.
The emissions are there. The question is whether and how you're counting them.
