The social cost of carbon is a monetary estimate of the damage caused by emitting an extra tonne of carbon dioxide, expressed in dollars per tonne. Governments use it to weigh climate policies.
This page is part of the climate change explained series and looks at how the number is built, why published values differ, and what it can and cannot tell a reader.
The definition behind the number
A peer-reviewed study in Nature defines the metric as the monetized value of the damages to society caused by an incremental metric tonne of CO2 emissions. The same paper notes that it is used by governments and other decision-makers in benefit-cost analysis. In plain terms, it answers a narrow question: if one more tonne of carbon dioxide enters the atmosphere today, what is the present-day value of all the harm that tonne will cause over the coming centuries?
The idea is an old one in environmental economics, and in the United States it entered regulatory practice through the courts. The US Environmental Protection Agency reports that it has used estimates of the social cost of greenhouse gases in analyses of actions that affect emissions since 2008, and that it began regularly including such estimates following a court ruling in which an agency was ordered to consider the value of reducing emissions in a rulemaking.
It is a marginal measure. It does not price all climate damage to date, and it is not the bill for a country or an industry. It prices the next tonne, which is what makes it useful for comparing a rule that avoids emissions with the cost of that rule.
Construction of the estimate
Calculating the figure means chaining together several disciplines. The Nature paper describes the sequence: emissions change atmospheric concentrations, concentrations change temperature and sea level, the resulting impacts are monetized and aggregated as economic damages, and discounting then combines all future damages into a single present value.
Each link in that chain carries uncertainty. Climate models give a range for warming per tonne emitted. Damage functions translate warming into economic losses, and projections of population and income decide how much there is to lose. The study itself reports its result as an average with a wide spread rather than one value.
The US agency’s report uses the same architecture, with separate modules for climate, damages and discounting, and reports a distribution of outcomes instead of a single number. Its discounting module uses a set of dynamic discount rates rather than one constant rate.
Reasons published values differ
Quoted figures vary widely, and the gap is mostly explained by method rather than disagreement over whether emissions cause harm. The Nature study reports a preferred mean of $185 per tonne of CO2, with a 5%–95% range of $44–$413 per tonne, and says this is 3.6 times higher than the US government value of $51 per tonne that was in use at the time. The authors attribute the increase to updated socioeconomic projections, climate models, damage functions and discounting methods.
The EPA’s report reached a similar level. Its central estimate for 2020 emissions is $190 per tonne, based on a 2% near-term discount rate, which the report describes as around 280% larger than the earlier interim estimate of $51 per tonne that used a 3% constant discount rate.
Discounting is the biggest single lever. A high discount rate treats damages far in the future as worth little today; a low one treats them as nearly as important as present harm. The IPCC’s Working Group III summary makes the same point about cost-benefit results in general: the global cost of limiting warming to 2°C is lower than the global economic benefit unless climate damages are towards the low end of the range, or future damages are discounted at high rates.
Gaps in the estimates
Every published value is a partial accounting. The EPA report says that data and modeling limitations restrain the ability of the estimates to include all physical, ecological and economic impacts of climate change, implicitly assigning a value of zero to the omitted damages. It concludes that the estimates are a partial accounting of climate change impacts and likely underestimate the marginal benefits of abatement, a finding reported in the same document.
Specific omissions are named. The modeling omits most of the consequences of changes in precipitation, damages from extreme weather events, and the potential for nongradual damages from passing critical thresholds such as tipping elements.
The earlier interim values from the US Interagency Working Group carried a similar caveat, stating that the interim estimates likely underestimate societal damages from greenhouse gas emissions.
Uses of the figure and its limits
The main use is in policy appraisal. When a regulator evaluates a rule that changes emissions, multiplying the tonnes avoided by the social cost gives a monetized benefit that can be set against compliance costs. Because the answer is sensitive to the discount rate, the interim US estimates were published as a range of four values instead of one headline number.
The figure is also different from a carbon price. A market price, such as the price of an allowance in a trading scheme or of a voluntary offset credit, comes from supply and demand and policy design. The social cost is an estimate of harm. The two can be compared, but they need not be equal, and neither one is a recommendation about what any person or company should pay.
The estimates also cover more than one gas. Methane and nitrous oxide have their own social cost estimates alongside carbon dioxide, so a guide to the main greenhouse gases is a useful companion to this page.
Policy targets give the figure a practical context. The temperature goals agreed internationally are set out in the explainer on Paris Agreement goals, and the social cost of carbon is one of the tools used to compare the cost of climate action with the damage it avoids.
A business that quotes a social cost figure in public statements about its climate performance should check the advertising and consumer-protection rules where it operates, since a modeled estimate is not a measured result and values vary by method.
Frequently asked questions
Is the social cost of carbon the same as a carbon tax?
No. A carbon tax is a policy that sets a price on emissions. The social cost of carbon is an estimate of the damage caused by an extra tonne, and it may be used to help decide how high a tax or standard should be, but the two are separate concepts.
Why does the estimate change over time?
New climate science, updated economic projections and different discounting choices all move the result. The Nature study and the EPA report both found higher values than the earlier interim estimate for these reasons.
Does a higher value mean climate change is worse than thought?
Not necessarily. A higher estimate can reflect a lower discount rate, broader coverage of damages or updated projections, so the method matters as much as the headline number.
Can the figure be used to price an individual purchase?
It is designed for policy analysis, not as a price list for individual decisions. Anyone using it for a specific purpose should state the method, the year and the discount rate behind the value.
