CO2 Emissions Cost Society Far More Than Previously Thought

The $185 Question: Why Your Carbon Footprint Just Got a Lot More Expensive

Imagine you walk into a store, buy a gallon of gasoline, and the cashier says, “That’ll be $1.85.” You pay, you leave. But here’s the catch: The real price of that gas isn’t $1.85. It’s nearly seven dollars. The difference? Society picks up the tab for the damage that burning that fuel will cause to the climate. The cashier just doesn’t tell you.

For years, the U.S. government has used a number called the “social cost of carbon” to estimate that hidden tab. It’s a single dollar figure that represents the damage caused by emitting one metric ton of carbon dioxide. Policymakers use it to decide everything from fuel economy standards to power plant regulations. The current official number, set in 2021, is $51 per ton.

That number is wrong. Not just a little wrong. Wrong by a factor of 3.6.

In a 2022 paper published in Nature, a team led by Kevin Rennert of Resources for the Future, along with Frank Errickson, Brian Prest, and Lisa Rennels, built a new, open-source model called the Greenhouse Gas Impact Value Estimator (GIVE). Their central finding: The social cost of carbon is actually $185 per ton. That’s $185 for every metric ton of CO2 pumped into the atmosphere. For context, the average American household emits about 48 tons of CO2 per year from all sources. At $185 per ton, that’s nearly $9,000 in damages per household per year that nobody pays at the pump.

The paper, “Comprehensive evidence implies a higher social cost of CO2,” is not a radical rethinking of climate economics. It is an update. A careful, methodical, and long-overdue update that incorporates better climate models, better economic projections, and a better understanding of how risk compounds over time. The result changes the math on nearly every climate policy decision the government makes.

Why the Old Number Was So Low

The $51 figure came from a model developed by the U.S. government’s Interagency Working Group in 2010, with updates in 2013 and 2016. That model was built on climate science and economics from the early 2000s. A 2017 report from the National Academies of Sciences, Engineering, and Medicine (NASEM) essentially said: This thing is outdated. The science has moved on. The economics has moved on. You need to rebuild it.

Rennert and his team took that advice seriously. They rebuilt the model from scratch, using four major improvements that each pushed the number higher.

Better Climate Models

The old model used a simplified climate model that did not capture the full range of possible warming outcomes. The new GIVE model uses a more sophisticated climate module that better represents the physics of how CO2 traps heat. It also accounts for the fact that the Earth’s climate system has more inertia than previously assumed. Once you heat the planet, it stays hot for a very long time. The authors found that the updated climate science alone increased the damage estimate by about 20 percent.

Better Economic Projections

The old model assumed that future economies would look essentially like today’s economies, only bigger. The new model uses probabilistic socioeconomic projections from the Shared Socioeconomic Pathways (SSPs), which account for different possible futures: one where the world cooperates on climate, one where it doesn’t, one where inequality grows, and so on. This matters because damages from climate change hit poor countries harder, and the distribution of future wealth affects how much damage a ton of CO2 causes today. The authors found that incorporating these projections raised the estimate by another 15 percent.

Better Damage Functions

This is where things get painful. The old model used damage functions that were essentially linear: a 1°C temperature rise causes a certain percentage loss in GDP, and a 2°C rise causes double that loss. But the real world is not linear. The GIVE model uses updated damage functions that incorporate empirical evidence from climate impacts on agriculture, labor productivity, energy demand, and mortality. These functions show that damages accelerate as temperatures rise. A 2°C world is not just twice as bad as a 1°C world. It might be three or four times as bad. This change alone pushed the estimate up by roughly 30 percent.

Better Discounting

This is the most technical change, but also the most philosophically important. The discount rate is the rate at which you convert future damages into present-day dollars. A high discount rate means you care less about the future; a low rate means you care more. The old model used a constant discount rate of 3 percent. The new model uses a declining discount rate that starts at 2 percent and falls over time. This reflects the fact that as the future becomes more uncertain, you should value it more, not less. The authors also incorporated “risk aversion,” meaning they account for the fact that society hates catastrophic outcomes. When you combine a declining discount rate with risk aversion, the present value of future damages skyrockets. This change alone roughly doubled the estimate.

When you add all these improvements together, you get $185 per ton. The 5 percent to 95 percent range is $44 to $413 per ton, meaning there is a 1 in 20 chance the true number is below $44 or above $413. That is a wide range, but even the low end is close to the old central estimate. The high end is nearly ten times higher.

What $185 Actually Means

A number like $185 per ton is abstract. Let me make it concrete.

The average U.S. car emits about 4.6 tons of CO2 per year. At $185 per ton, that is $851 in damages per car per year. The average household emits about 48 tons per year from all sources (transportation, heating, electricity, food, everything). That is $8,880 in damages per household per year.

Now, you do not pay these costs directly. They are externalities. They show up as property damage from floods, crop failures from droughts, lost work hours from heat waves, and premature deaths from air pollution. But they are real costs. The fact that nobody charges you for them at the pump or on your utility bill does not mean they do not exist. It just means someone else pays them.

The implications for policy are straightforward. Any regulation that costs less than $185 per ton of CO2 reduced is now economically justified. The old threshold was $51. This means that stricter fuel economy standards, tighter power plant emissions rules, and more aggressive renewable energy mandates all pass a cost-benefit test that they previously failed.

Consider the EPA’s proposed rules for existing power plants. The agency estimated that the rules would cost about $60 per ton of CO2 reduced. Under the old social cost of carbon, that was borderline. Under the new $185 figure, it is a no-brainer. The benefits of the rule are three times the costs.

What This Does Not Prove

Before we get carried away, let me be clear about what this paper does not claim.

It does not claim that $185 is the final, definitive number. The authors are explicit that their estimate is a “preferred mean” based on the best available science. But the science will keep improving. Future models might push the number up or down. The 5 percent to 95 percent range of $44 to $413 per ton is a reminder that there is enormous uncertainty baked into these estimates. We are not dealing with a precise measurement. We are dealing with a well-informed guess.

It does not claim that all climate damages can be monetized. Some things are hard to put a price on: the loss of a species, the disappearance of a cultural tradition, the psychological trauma of losing your home to a wildfire. The GIVE model includes what it can measure, but it does not include everything. If anything, the true social cost of carbon is probably higher than $185.

It does not claim that carbon pricing is the only solution. A social cost of carbon is a tool for cost-benefit analysis, not a prescription for policy. It tells you how much damage a ton of CO2 causes. It does not tell you the best way to reduce emissions. That could be a carbon tax, a cap-and-trade system, direct regulation, or something else entirely.

It does not claim that the United States should unilaterally adopt $185 as its official number. The authors note that their estimate is based on a global damage function, meaning it includes damages to everyone in the world, not just Americans. Some argue that the U.S. government should only count damages to the United States. That would produce a much lower number, around $20 per ton. But that argument assumes that the U.S. economy is insulated from the global economy. It is not. A flood in Bangladesh disrupts supply chains. A drought in Brazil drives up food prices. A war in the Middle East triggered by climate refugees... you get the idea.

The Open Question: How Do We Use This?

The biggest open question is not whether $185 is the right number. It is whether the U.S. government will actually use it.

The social cost of carbon has become a political football. The Trump administration slashed it to $1 per ton. The Biden administration raised it back to $51. A federal court recently blocked the Biden administration from using the higher figure in regulatory decisions, ruling that the government had not followed proper procedures. The case is ongoing.

Meanwhile, the GIVE model is open source. Anyone can download it, run it, and see exactly how the number is calculated. That transparency is a double-edged sword. It makes the model more credible, but it also makes it easier for critics to pick apart. Expect legal challenges if the government tries to adopt $185.

There is also the question of international coordination. The European Union uses a social cost of carbon around $100 per ton. The United Kingdom uses a figure around $80. If the U.S. adopts $185, it would be the highest in the world. That could create friction in trade negotiations, but it could also set a new global standard. Other countries might follow suit.

What This Actually Means

• ▸Every ton of CO2 matters more than we thought. The old $51 figure made it easy to dismiss small emissions reductions as not worth the effort. The new $185 figure changes that calculus. A 10 percent reduction in household emissions now saves society nearly $900 per year. That is real money.

• ▸Regulations that looked marginal now look essential. The EPA’s power plant rules, the Department of Transportation’s fuel economy standards, and the Department of Energy’s appliance efficiency standards all relied on the old social cost of carbon. Under the new figure, the benefits of these rules are three to four times larger than the costs. The case for regulation just got a lot stronger.

• ▸Carbon pricing is still the most efficient policy, but the price needs to be much higher. If you want to use a carbon tax to internalize the externality, the tax should be $185 per ton, not $51. That would translate to roughly $1.65 per gallon of gasoline. That is a politically toxic number. But it is the economically honest one.

• ▸The uncertainty is huge, and that is a feature, not a bug. The range of $44 to $413 per ton means that we are not dealing with a precise number. We are dealing with a distribution of possibilities. Policymakers should use the central estimate for cost-benefit analysis, but they should also acknowledge that the true number could be much higher. That argues for acting sooner rather than later. It is insurance against the worst-case scenario.

• ▸The social cost of carbon is a tool, not a solution. It tells you the price of damage. It does not tell you how to reduce emissions. That requires political will, technological innovation, and behavioral change. But at least now we know the true cost of doing nothing. It is $185 per ton. And rising.