The Carbon at Risk measure for managing carbon removal risk and guiding portfolio construction

Journal

Nature Climate Change

Authors

Bali Lee, Tom Bearpark*, Oleksander Kit, Delia Meth-Cohn, Frederic Olbert, Frank Venmans*, Gabrielle Walker, Paul Young, Ben Groom*.

*The starred authors are contributors from the Grantham Research Institute

Abstract

Carbon removal credits sell for anywhere between $1 and $1,000 per tonne. The gap largely reflects how long the carbon stays stored and how certain that storage is, but buyers cannot easily tell how much risk they avoid by paying more. The authors adapt Value at Risk, a standard measure from finance, to create Carbon at Risk (CaR). CaR puts different types of removal on a common scale and shows how many extra credits a buyer should purchase as a buffer to meet a permanence target. The cheapest buffer depends not only on how risky each credit is, but also on whether credits tend to fail together. Where they do, adding more of the same removal type becomes an expensive form of protection, and mixing in a removal type exposed to different risks can be cheaper.

Key points for decision makers

1. Carbon at Risk gives buyers and regulators a common risk metric. It makes nature-based and technological removals comparable on the same scale, moving beyond the divide between the two.
2. Diversification can lower the cost of permanence. Buying a mix of removal types can be cheaper than relying on any single storage method.
3. Correlated failure matters. Wildfires, storms and policy shifts can hit many projects at once, and the more geographically concentrated a portfolio, the more exposed it is. In the authors’ stylised example, forests alone are cheapest over 200 years if projects fail independently. Once failures are correlated, the cheapest portfolio is a mix of forests and direct air capture with carbon storage (DACCS).
4. The time horizon shapes the optimal mix. The study uses 200 years, the storage threshold adopted for carbon removals in the UK Emissions Trading Scheme. Forest risks can be modelled reliably over centuries and geological storage over thousands of years, so the longer the horizon, the larger the role for more durable removals.
5. Fixed-share buffer pools may be inadequate. Today’s buffer pools typically set aside a fixed share of credits, which can be too small for growing climate risks, or too large for projects with less risk. Also, a mix of credits and cash, invested in assets whose value does not depend on forests burning or storage leaking, could offer more reliable and cost-effective protection, much as an endowment or pension fund sets money aside for future obligations.
6. Compliance markets will need standardized risk measures. As removals move into national climate plans and compliance markets such as the UK and EU emissions trading schemes, CaR can be a tool for governments and policy makers to manage permanence risk.

Code and data are available at github.com/BenGroom/CarbonAtRisk, and more information at carbonatrisk.org.