Carbon risk is becoming financial before it becomes measurable
An audience poll found most still unconvinced carbon is primarily a financial risk, even as European regulators fine banks for carbon shortcomings.
Measurement of Scope 3 emissions remains too imprecise for chief financial officers to price with confidence, so capital is shifting toward business-specific projects instead.
Data quality remains an issue in carbon pricing tout court, but this should not be treated by companies as an excuse for inaction.
A chief financial officer (CFO) who cannot quantify a liability will not pay to reduce it. That is the quiet problem sitting underneath every corporate carbon strategy built in the last five years. Companies are being asked to treat emissions as a financial risk before most of them can measure those emissions with anything close to financial precision. The result is a strange kind of progress, where the accounting language has become more sophisticated while the underlying numbers remain, by several honest accounts, barely more than a guesstimate.
That gap between financial language and financial precision framed the second panel at Carbon Credits 2026, titled From Footprints to Financials, which set out to test whether carbon pricing belongs with the CFO or still sits mainly with the chief sustainability officer (CSO). Registrants polled beforehand were unconvinced: 16 of 53 respondents called carbon primarily a technical compliance challenge, 15 called it a major financial risk comparable with liquidity or interest-rate exposure, and 12 called it primarily a reputational risk (see Chart 1). Overall, more than half the audience indicated in the pool that carbon pricing is a technical rather than a financial problem.
Chart 1
This discussion did change views. By the end, a live show of hands in the room found nobody still willing to call carbon a purely technical issue.
Regulation is certainly imparting momentum to those who see carbon pricing as a financial risk. The International Financial Reporting Standard S2 (IFRS S2) and the EU's Corporate Sustainability Reporting Directive (CSRD), have given carbon pricing a considerably harder financial edge. Even after the CSRD's scope was narrowed, its core double-materiality requirement, assessing both a company's impact on climate and climate's impact on the company, survived intact.
The European Central Bank (ECB) illustrated the financial consequences for companies of getting the risk wrong. It fined a French agricultural lender roughly €4 million for failing to adequately address climate-related risk across its loan portfolio on the timeline the ECB expected. The bank in question had already pledged not to finance oil and gas, and had no fossil-fuel exposure to manage anyway. Its actual vulnerability, and the ECB's actual concern, was the climate resilience of the farmers it lends to. It was a lending portfolio risk that had almost nothing to do with pledges made for reputational reasons.
That distinction matters more than it sounds. A bank’s public decarbonisation pledge and a regulator's assessment of its underlying portfolio risk are not the same test, and passing one does not guarantee passing the other. Scope 3 emissions - the emissions a company does not directly control but which occur across its suppliers, borrowers and customers - sit at the centre of that mismatch of expectation. For banks, and indeed many other businesses, Scope 3 emissions are usually the largest slice of the total carbon footprint. Yet they are also the hardest to measure with any confidence.
One internal review at a major oil and gas company reportedly found more than 40 different methods of carbon accounting in use across its own business, depending on which internal team was asked. That is not a data-collection failure so much as a data-governance one: the raw numbers exist in multiple places, and nobody has forced them into a single, comparable format. Construction projects, likewise, have been found to overestimate embodied emissions by 15 to 20 per cent on average, a comparable failure in a different sector.
For companies whose emissions are overwhelmingly Scope 3, banks among them, this uncertainty creates a genuine financial dilemma, not just a reporting inconvenience. If between 80 and 90 per cent of a company’s carbon footprint sits with suppliers and customers rather than its own operations, a CFO weighing whether to pay for that liability is being asked to price something nobody in the business can measure to within an order of magnitude, let alone a percentage point.
A question is what counts as a legitimate carbon strategy when a company's biggest emissions source is structurally outside its control. One view held that a company should reduce its own emissions directly wherever it physically can and rely on cross-industry collaboration to eliminate double counting in shared value chains before reaching for unrelated offsets. A competing view argues that this standard is unrealistic since a company's largest emissions source might be a national energy supply or industrial policy. The carbon intensity of the national electricity grid is not something an individual company can offset through better supply-chain management.
One answer for companies is to shift offsetting strategies from broad contribution towards focused and specific projects. One consumer brand, for example, has moved from estimating its entire coffee supply chain's footprint to identifying which specific growing regions face the greatest emissions risk, then directing investment toward agroforestry projects in exactly those regions rather than buying generic credits. A second example involved a bank financing a hydropower plant whose revenue depended on water supplied from a tropical rain forest. Buying carbon credits to protect that specific rain forest was treated as project risk management rather than philanthropy.
Whether that kind of targeted purchase should be called insetting (a company investing in carbon reduction in its own supply chain), offsetting, or simply investment in corporate resilience, matters less than the logic behind it. If a company's core physical assets depend on an ecosystem it does not own, protecting that ecosystem is a financial decision about business continuity, not a climate gesture, whatever accounting category it eventually gets filed under.
None of this settles whether Scope 3 accounting will ever reach the precision CFOs expect from other line items on a balance sheet. What the discussion suggested instead is that precision may not be the actual prerequisite for action it appears to be. Companies are already directing real capital toward specific, quantifiable risks, water security, supply-chain resilience and regulatory exposure without waiting for a perfect set of global carbon pricing numbers to become available. The CFO's carbon problem, on this evidence, is being solved piecemeal, risk by risk, well before it is solved in the aggregate.
That seems prudent. The large overhang of unretired voluntary carbon credits was seen by the panel as a reminder of how past valuation methodologies went wrong. There was no consensus on what should happen to them though a variety of solutions (Should companies retire them anyway and claim a tax write-off? Can they be swapped for higher integrity credits? Is there a market for them? Can they be rescued with better management? Can remedial actions work?) are raised. Most unretired credits were issued using methodologies since shown to be imprecise. A generous view is that the unwanted credits represent a genuine, if imperfect, application of the economic principle that whoever creates a cost should pay for it. On this view even flawed pricing of carbon risk is better than not pricing pollution at all, and retroactively shaming early buyers punishes exactly the behaviour regulators are now trying to encourage.
This does not mitigate the problem. Roughly 1.8 billion tonnes of carbon credits have been issued historically under forest-protection methodologies whose calculations were later challenged. This represents a weight on the voluntary carbon credit market, bifurcating it between a large overhang of unretired credits that are effectively worthless and a small quantity of valuable credits that trade at excessively high prices as a result. This over-supply of unretired, low-quality credits sitting alongside a scarcity of high-quality removal credits looks increasingly like a structural feature of the market rather than a temporary imbalance regulators can simply price away.
It is leading to apparently strange buying decisions. A technology company, for example, has been retiring voluntary carbon removal credits at prices reported as high as US$500 a tonne, a high cost some assumed reflected genuine environmental conviction. An alternative explanation is that the company is spreading the cost across an enormous customer base of Cloud and AI users that makes the per-unit cost invisible. On that view, what looks like a premium climate commitment is actually a straightforwardly profitable line item.
The Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA) launched by the International Civil Aviation Organisation (ICAO) is not yet established as a model standard for carbon pricing. Physically decarbonising aviation fuel is cited at costs approaching US$1,500 a tonne, against compliance credits available at roughly US$15 a tonne. Airlines, keep finding reasons not to pay even that lower figure, which suggests the problem is not that compliance credits are too expensive but that the entire pricing structure has not yet been made painful enough to change behaviour.
