Researchers Unveil Carbon at Risk Framework for CDR Permanence

A group of researchers has released a framework called Carbon at Risk (CaR), a probabilistic measure designed to standardize how permanence risk is quantified across carbon dioxide removal (CDR) technologies. The metric defines, at a given confidence level and time horizon, the expected shortfall between contracted removals and realised storage.

The name and the logic arrive directly from financial markets. Just as banks use Value at Risk to bound potential trading losses, CaR expresses the downside risk of promised storage: the probability, magnitude and timing with which a project's delivered tonnage falls short of what a buyer contracted. The authors present the framework as a new kind of asset for CDR buyers, investors and regulators, one that gives risk a common unit in which it can be compared and priced.

Why did researchers build it?

The team, drawn from established organizations and developed through a series of workshops between academics and industry stakeholders, identified a gap at the centre of carbon markets. No standardized, quantitative measure of permanence risk currently exists, and no continuous probabilistic metric anchors risk management across regulators, buyers, insurers and investors.

Instead, CDR projects are conventionally sorted using qualitative labels. Permanent versus nonpermanent. Nature-based versus engineered. Some programs supplement these categories with buffer pools, but the sizing of those pools is only weakly linked to quantified risk.

The CaR creators do not dismiss these heuristics entirely. "These heuristics can be useful for first-pass screening, but they do not quantify the probability, magnitude, or timing of under-performance," they write. "The result is distorted capital allocation."

That distortion is the practical target. When a forestry project and a mineralization project both carry a "buffer" sized by rule of thumb rather than measured risk, capital flows to the wrong places, and insurers price the wrong exposures.

What does the framework actually change?

The methodology provides a technology-agnostic basis for three concrete applications, according to the authors:

• Risk-calibrated buffer sizing for project developers and registries
• Insurance pricing for underwriters covering delivery and reversal risk
• Reserve-fund design for buyers and portfolio managers

The framework works over both policy-relevant and climate-relevant time scopes, which matters because durability claims diverge sharply depending on whether the horizon is a decade or a century.

The authors acknowledge a limitation. Early applications of CaR will rely on assumptions and abstractions until the underlying distributions can be estimated empirically from observed data. CDR is a young field, and reversal records are thin.

Their answer to that criticism is pointed. The alternative, they argue, is to continue using heuristics that rely on unexamined assumptions of their own.

"Even with current data limitations, CaR serves two functions that heuristic labels cannot: it makes assumptions explicit and auditable, and it provides a common unit in which the risks of removal technologies can be compared and priced," the researchers stated.

What happens next?

The milestone to watch is empirical validation. The framework moves from abstraction to market infrastructure only when registries, insurers or large buyers adopt it and when observed reversal and underperformance data replace the placeholder distributions. Until an accreditation body or a major CDR purchaser embeds CaR in contracting, it remains a methodology paper with auditable assumptions rather than a pricing standard.

The arrival of the framework nonetheless signals a shift in how the sector talks about durability. Regulators drafting removal accounting rules, and buyers weighing delivery risk in procurement, now have a quantitative template that labels like "durable" were never designed to supply.