The social cost of carbon is an estimate, in dollars, of the total economic damage caused by releasing one additional ton of carbon dioxide into the atmosphere. The current U.S. Environmental Protection Agency estimate is $190 per ton, though that number has swung dramatically depending on the assumptions behind it. Governments use this figure to weigh the benefits of climate regulations against their costs: if a policy prevents a ton of emissions for less than the social cost of carbon, the math says it’s worth doing.
What the Number Actually Captures
That dollar figure is meant to represent everything society loses when one more ton of CO2 enters the air. The main categories of damage include changes in agricultural productivity as growing conditions shift, health effects from heat exposure and worsened air quality, property damage from rising sea levels, increased energy costs (more air conditioning, for instance), and declines in labor productivity as outdoor work becomes harder in hotter conditions. These effects ripple across decades because CO2 stays in the atmosphere for centuries, so a ton emitted today causes damage well into the 2100s and beyond.
The costs hit different economies in different ways depending on their structure. A country heavily dependent on agriculture faces different risks than one with thousands of miles of vulnerable coastline. But the core insight is the same: carbon emissions impose real, quantifiable costs on people who had no say in creating them.
How Economists Calculate It
The social cost of carbon is estimated using what are called integrated assessment models, which combine climate science with economic projections. Three models dominate the field: DICE, FUND, and PAGE. Each one takes a scenario of future emissions, projects how much warming that causes, then translates that warming into economic damages using what’s known as a “damage function,” essentially a formula that converts degrees of temperature increase into dollars lost.
These three models often produce very different numbers, even when fed similar inputs. Research published in the journal One Earth found that the core reason for these disparities lies in how each model simulates the climate system itself, not just how it estimates economic damage. When all three models were paired with the same simplified climate module, their estimates became much more consistent. In other words, disagreements about the physics of warming matter as much as disagreements about its economic consequences.
Why the Discount Rate Changes Everything
One of the most powerful levers in the calculation is the discount rate, which determines how much weight you give to damages that happen in the future compared to costs paid today. A high discount rate treats future harms as less important. A low one treats a dollar of damage in 2080 as nearly as significant as a dollar today. This is not just a technical choice; it’s a philosophical one about our obligations to future generations.
The impact on the final number is enormous. The EPA’s current $190 estimate uses a 2% discount rate. Bump that up to 2.5%, and the figure drops to $120. Lower it to 1.5%, and it jumps to $340 per ton. That’s nearly a threefold difference driven by a single assumption. This is why debates over the social cost of carbon often come down to arguments about the discount rate rather than disagreements about climate science.
The Political Rollercoaster in the U.S.
No number in U.S. climate policy has been more politically volatile. Under the Obama administration, an interagency working group set the social cost of carbon at $52 per ton (in 2020 terms), based on global damages and a 3% discount rate. The Trump administration then disbanded that working group and slashed the figure to between $1 and $8 per ton, primarily by counting only domestic damages to the United States and using assumptions that were, as the University of Chicago’s Energy Policy Institute put it, “difficult to justify based on science and economics.”
The Biden administration reversed course, initially restoring the Obama-era figure and later updating it to $190 per ton using newer science and a lower discount rate. That represents roughly a 50-fold increase over the Trump-era value. Each shift had direct regulatory consequences: a higher social cost of carbon justifies stricter emissions rules, while a lower one makes it harder to defend the cost of regulation.
Global Versus Domestic Damages
One of the biggest policy questions is whether to count damages worldwide or only within the country doing the calculation. CO2 mixes uniformly in the atmosphere, so a ton emitted in Texas causes damage in Bangladesh, Brazil, and everywhere else. The global social cost of carbon is four to 14 times higher than a purely domestic estimate. For context, when the EPA estimated $30 billion in global climate benefits from proposed power plant regulations in 2030, the domestic-only climate benefits came to just $2 to $7 billion, less than the $7.3 billion compliance cost of the rule itself. Using a domestic-only number can make climate regulations appear to fail a cost-benefit test that they would pass under a global framework.
The United States, Germany, Canada, and several U.S. states all maintain official social cost of carbon estimates for policymaking. The choice of scope (global or domestic) often determines whether a regulation looks economically justified.
What the Models Leave Out
Even the highest official estimates likely undercount the true cost. Standard models use smooth damage functions, meaning they assume economic losses increase gradually and predictably as temperatures rise. But the climate system contains tipping points: thresholds beyond which changes become abrupt and irreversible. The collapse of major ice sheets, dieback of the Amazon rainforest, disruption of Atlantic Ocean circulation patterns, and shifts in monsoon systems are all possibilities that current models handle poorly or ignore entirely.
A study in the Proceedings of the National Academy of Sciences found that most cost-benefit analyses of climate policy fail to account for these tipping points, which involve both market losses (destroyed infrastructure, failed crops) and nonmarket losses (ecosystem collapse, species extinction) that are harder to price. The models also tend to undervalue environmental goods as they become scarcer. A forest’s cooling and water-filtering services become more valuable, not less, as forests disappear, but standard economic models don’t always reflect that increasing scarcity.
These omissions help explain why most cost-benefit analyses produce “optimal” climate policies that fall well short of the internationally agreed target of limiting warming to 2°C. The models suggest less aggressive action partly because they don’t fully capture what’s at stake.
How the Number Is Expected to Rise
The social cost of carbon increases over time because each additional ton of CO2 causes more damage in a warmer world than it would in a cooler one. Using a 3% discount rate, earlier U.S. government estimates projected the cost rising from $50 per ton in 2030 to $69 per ton by 2050 (in 2007 dollars). Under a high-impact scenario using the 95th percentile of possible damages, those figures climb to $152 in 2030 and $212 by 2050. With the EPA now using a 2% discount rate and updated science, current and future estimates run considerably higher than those earlier projections.
This upward trajectory matters for long-lived infrastructure decisions being made today. A power plant or highway built now will still be operating in 2050, when the cost of each ton of CO2 it’s responsible for will be significantly greater than it is today. Using only today’s social cost of carbon to evaluate those investments systematically undervalues the future damage they’ll cause.