An Introduction to Carbon Pricing
I was planning on covering Zipf’s law today, the rank/size distribution pattern that we see in city sizes and many other areas. It is interesting, though I am still not sure what the policy significance is. But then a friend suggested a topic related to carbon pricing. Over the years, I have commented on carbon pricing many times and why I think that this is the proper way to address carbon dioxide emissions, and so I find it puzzling that I have not done a full post on the subject until now. Zipf’s law will probably be in the near future.
As this is meant to be a broad overview of the subject, I start by describing what carbon pricing is and the basic rationale for it. I will briefly cover the current state of play for carbon pricing, how a price can be structured (carbon tax or cap and trade). Then I will touch on some of the major issues around carbon pricing: how the price is determined, the use of revenue, dealing with imports, and the use of carbon offsets. I will conclude on why I think that carbon pricing should be at the center of any credible climate change strategy.
Carbon Pricing: Theory
Carbon-based fossil fuels—coal, petroleum, and natural gas—are critical to prosperity in the modern world, and there is at present no full, suitable replacement (Shizawa (2023)). However, carbon dioxide emissions (Jones et al. (2024) via Our World in Data)) from the combustion of fossil fuels (Ritchie (2020)) comprise about three quarters of human-sourced greenhouse gas emissions, which are widely understood to cause damage to human civilization and the non-human ecosystem via global warming and ocean acidification (International Panel on Climate Change (2023)).
Greenhouse gas emissions are a classic example of an externality, which is an effect of an economic activity that is not accounted for in the good’s nominal price (Schauer (1995)). Consequently, there is a discrepancy between the price that a customer pays for fossil fuels and the actual society-wide cost. Imposing a tax on the good that is equal to the external damage in the logical way to bring the nominal and actual prices into alignment, and the theory to do so is credited to Arthur Cecil Pigou (1920); hence such taxes are often also called Pigouvian taxes. The theory was brought into a modern economic formulation by William Baumol (1972).
The theory of Pigouvian taxation is not without theoretical challenge. For example, Carlton and Loury (1980) argue that Pigouvian taxation needs to be supplemented by lump sum taxes or subsidies to prevent the tax from merely increasing the number of firms in the industry with the externality.
Ronald Coase’s (1960) famous The Problem of Social Cost discusses how the dynamic and reciprocal nature of externality pricing makes the endeavor especially difficult. On the reciprocal nature of harm, Coase states,
The traditional approach has tended to obscure the nature of the choice that has to be made. The question is commonly thought of as one in which A inflicts harm on B and what has to be decided is: how should we restrain A? But this is wrong. We are dealing with a problem of a reciprocal nature. To avoid the harm to B would inflict harm on A. The real question that has to be decided is: should A be allowed to harm B or should B be allowed to harm A? The problem is to avoid the more serious harm.
Coase gives several examples to illustrate the principle, but another good illustration relates to housing development. Most cities have restrictive zoning codes that limit the amount and nature of new housing development that can occur. These regulations exist on the grounds that development causes harm to current residents in the form of traffic congestion, increased crime, and other perceived ills. Coase’s reciprocity is that existing residents in turn inflict harm with zoning regulations, in the form of higher costs for renters and would-be homebuyers, loss of economic dynamism, and decreased profits for developers.
Who has the better case—advocates or opponents of zoning regulations—depends in part on which of the two costs detailed above is more serious. Another determinant relates to the nature of property rights, as Coase proceeds to argue in the paper. Advocates of development argue that the overriding property right is for the owner of a parcel of land to use it as (s)he sees fit, whereas advocates of zoning argue that the overriding property right is for residents’ right to avoid harms from neighbors. This question is ultimately a subjective one.
A way in which the reciprocity issue manifests itself in the context of carbon pricing is in adaptation. Deng, Wu, and Xu (2025) show that the social cost of carbon decreases in the presence of adaptation, and that the greater potential for adaptation, the lower the social cost of carbon. Examples of adaptation include building seawalls, concentrating development away from vulnerable coastal areas, and planting drought-resistant crops. As Ford (2007) documents, many climate advocates of the late 20th century and into the 2000s, such as Al Gore, opposed adaptation on the grounds that it would reduce the perceived urgency of reducing emissions, but such opposition has now largely disappeared.
Coase (1960) is a classic and highly readable paper, and I recommend reading it for anyone who is interested in the economics of externality management.
The State of Play
According to the World Bank (2025), 28% of world greenhouse emissions were subject to some kind of carbon price in 2024. That is down about a percentage point from the previous year, but up from 5% in 2005.
Coverage varies by sector. The report finds that power generation is the most heavily covered sector, with 51% of emissions subject to carbon pricing. That figure is 45% for industry and no more than 15% in any other sectors. Emissions from waste, agriculture, and land use change are virtually uncovered by carbon pricing.
As of 2025, the average price of emissions that are covered by carbon pricing is $19 per ton CO₂, down from a peak of $22/ton in 2022 and up from $10/ton in 2015, with of course large variation between jurisdictions. These prices generated $102 billion in government revenue in 2024, over half of which was in Europe.
Overall price levels have fluctuated but have remained rather steady over the last several years. They are inadequate to achieve deep decarbonization.
Carbon Tax Versus Cap and Trade
There are two mechanisms by which carbon pricing can be assessed.
A carbon tax is a tax of a fixed amount per ton of carbon emissions. As in the World Bank report, direct taxation is the less popular of the two instruments, with about 4% of world emissions covered by a carbon tax and 24% covered by cap and trade (a small amount is covered by both types of instruments). The theory of a carbon tax dates to David Gordon Wilson in the 1970s. Some of the design issues around a carbon tax, such as the problem of leakage, determination of the appropriate tax level, and how to use offsets, are addressed later.
Bertaud (2024) writes of congestion pricing, another Pigouvian tax mechanism, at Market Urbanism. He highlights Singapore’s dynamic Electronic Road Pricing system as a model for insuring the smooth flow of traffic in the downtown area. He contrasts this to congestion pricing systems in other cities whose objectives are more to raise revenue or to punish motorists, highlighting the importance of understanding the purpose of a policy in evaluating its effects. Another example is a noise tax for aviation in France, a charge on airplane takeoffs from the nine busiest airports in France.
The other mechanism is a cap and trade system, also known as an emissions trading system. Under such a system, the government sets an upper bound on the amount of emissions from the covered areas and then sells a limited number of tradable credits to emit. The carbon price is then the market price of those credits.
Newell and Rogers (2003) discuss how a similar trading system had been used successfully in the phaseout of leaded gasoline for cars in the United States from 1976 to 1996. Sweeney et al. (2012) comment on the success of the cap and trade system to reduce sulfur dioxide emissions from coal-fired power plants.
My preference is for a carbon tax rather than cap and trade. The two systems are functionally similar, but a carbon tax is conceptually simpler, offers fewer opportunities for cheating, and presents less volatile pricing, as Santabárbara and Suárez-Varela (2024) find.
Social Cost of Carbon
Perhaps the first question that arises regarding carbon pricing is what the price should be. It should not surprise anyone that this is a highly nontrivial question.
One’s first intuition is that the logical rate for a carbon tax should be equal to a reasonable estimate of the amount of damage caused by emissions, a figure known as the social cost of carbon. Efforts to estimate the social cost of carbon have been over the map. Fisher (2024) at the Cato Institute discusses some estimates used by the U.S. EPA. They include $43 per ton of carbon dioxide under the Obama administration, $3-5/ton under the first Trump administration, and $51/ton under Biden with a proposal to raise the value to $192/ton to be in line with some more recent research. Now Trump II is seeking to discontinue the social cost of carbon entirely except where required by law.
Two major factors account for these wide variations. The first is the choice of discount rate. When making any kind of financial decision, we typically value a dollar in the future less than we do at the present. If someone offered me the choice of $1000 today or a certain amount of money 10 years from now, it would need to be a larger amount of money, maybe $2000, to cause me to be willing to wait. If I value equally $1000 today or $2000 10 years from now, then I am operating with a discount rate of 7%, since if we take 7% away from $2000 10 times (with compounding), we end up with about $1000.
Climate change is a problem where the benefits of emitting activities are felt today, and the costs are spread out decades or even centuries into the future. Consequently, the lower the discount rate, the higher the social cost of carbon. According to Fisher (2024), the Obama and Biden estimates used a discount rate of 3%, the Trump I rate was 7%, and the proposed upward revision used 2%, and this is the most important factor explaining the differences between the numbers. See Prest (2022) for a more detailed discussion of discount rates, and Rennert, Kingdon, and Prest (2025) for how this applies to the social cost of carbon.
The second major factor is whether domestic or global harms should be considered. If only domestic harms, i.e. those to citizens and residents of the United States, are considered, then the social cost of carbon will be significantly less. This is the criterion that was set forth in Bush 43’s Circular A-4, setting guidelines for federal agencies in crafting regulations. A-4 remained in effect until updated in 2023, and now the Trump administration is seeking to reinstate the 2003 version. Nevertheless, the Obama administration’s social cost of carbon estimate considered a global perspective.
There are many more issues in the social cost of carbon that I have to skip over, such as the potential for adaptation to mitigate damages and uncertainty, including how to account for uncertainty, as Lemoine (2016) discusses. It is hard to imagine how the social cost of carbon is useful for setting regulation, let alone for being the basis of taxation, when the value is so subject to manipulation. One could instead set a price based on a lowest common denominator—the lowest value of the social cost of carbon that most mainstream sources agree upon—or whatever price is necessary to catalyze transformative change, however that is defined.
The problem does not go away with a cap and trade program; in fact, if anything it is worse. Then the problem becomes how to set a cap, and this question will be even more defined by subjective value judgments than the social cost of carbon.
Use of the Revenue
For many carbon pricing advocates, the use of revenue raised by the pricing mechanism is of secondary concern to the price itself, though the use of revenue is also very important.
Broadly speaking, there are two main options. The first is for activities related to greenhouse gas mitigation, such as clean energy research, development, and deployment. Most proceeds from California’s cap and trade program, for instance, to go the Greenhouse Gas Reduction Fund, which in turn funds various emissions reduction programs. According to the California Legislative Analyst Office, the largest outlay of funds from the Greenhouse Gas Reduction Fund from 2013 (when it started) to 2023 (the date of the cited source) has been for the high speed rail, which is statutorily required to receive at least 25% of the funds. Now that is a whole other blog post.
The other option is some kind of revenue recycling. That could be done through a dividend, such as in Citizens’ Climate Lobby’s fee and dividend proposal. The money raised by CCL’s proposal would be deposited in a Carbon Fees Trust Fund, 100% (minus administrative costs) of which would be returned as a dividend. Revenue recycling could also be done by reducing some other taxes, such as the income tax or the corporate tax.
Tax offsetting seems to me like the best policy, but we do have to account for political economy constraints. The CCL dividend proposal is designed specifically to be visible to the public, thereby raising public confidence that the carbon tax will not be used to increase the size of government.
Here, a digression into a semantic issue is in order. CCL refers to their proposal as “fee and dividend”, but I characterize it as a carbon tax. Moon (2009) at the Tax Foundation explains that the difference is that a fee is intended to cover the cost of a service, such as the fee you pay for a drivers’ license, while a tax is for revenue purposes. I don’t think there is enough of a direct connection between CCL’s proposal and a specific service to justify calling it a fee.
Common sense would suggest that the effect of a carbon price, or any other environmental tax, would be the loss of economic efficiency in exchange for environmental gain. Maybe there would be an eventual gain economically, but since the benefits from greenhouse gas reduction accrue over a long period, we would expect it to take a long time to see net economic gain, if that ever happens at all.
This argument assumes that the policy situation is optimal, with the exception of the unpriced externality. This is seldom the case, though. Fullerton and Metcalf (1997) consider the “double dividend” hypothesis, which is that environmental taxes have the obvious environmental benefit and, additionally, economic benefit that accrues even if one neglects the environmental effects. They find that the double dividend can occur, but it should not expected to occur in general.
The double dividend can cut the other way. Bovenberg and de Mooij (1994) find that in the presence of distortionary taxes, the socially optimal tax for a product with externalities might be less than the Pigouvian tax. The intuition for this finding is that the distortion raises the cost of public funds.
In 2013, Citizens Climate Lobby commissioned a report from Regional Economic Models, Inc. (Nystrom and Luckow (2014)) that found that their fee and dividend proposal would have net economic benefit, independently of emissions reductions. The double dividend effect is a major reason for this.
I put forward three reasons to be leery of the use of dedicated funding and instead to favor revenue recycling for carbon pricing. First, regardless of the funding source, we all know that government programs are not always well-crafted. Second, overuse of dedicated funding hampers a government’s ability to set priorities for future budgets, particularly during downturns. Third, tying a revenue source to a funding source creates an unhealthy political codependency: constituencies for those programs that get their money from carbon taxes will have an incentive to perpetuate emissions and thereby preserve their funding source.
Border Adjustment
Another common objection to carbon pricing is that it will cause emissions-heavy industry to decamp for jurisdictions with no such price. This effect is also called carbon leakage. This is a serious issue, and a solution is a border adjustment. In a jurisdiction with carbon pricing, a border adjustment is a tariff on an imported good that is equal to the carbon intensity of the good, multiplied by the difference between the jurisdiction’s carbon price and the carbon price where the good is manufactured. The purpose is to equalize the carbon prices between domestic manufacture and imports so as not to put domestic manufacture at a disadvantage.
The European Union has a Carbon Border Adjustment Mechanism to do just this. After a trial period, the CBAM is set to fully apply in 2026. In 2024, the Center for Climate and Energy Solutions reported that the EU’s CBAM was the only border adjustment in operation, though there is interest elsewhere in the world.
This is an area to be careful about. There is the Foreign Pollution Fee Act of 2025, sponsored by Senator Bill Cassidy (R-LA) and cosponsored by Lindsey Graham (R-SC). This looks like a border adjustment, but it’s not because there is no national carbon pricing in the U.S. Citizens Climate Lobby supports the bill, but I am wary that, as not being a true border adjustment, the bill is just protectionism with a green tint.
While I agree that a border adjustment is a necessary complement to carbon pricing, there are several other pitfalls and limitations. Ambec (2022) comments on the severe difficulty of calculating the carbon intensity of imported goods and notes that the EU’s CBAM is limited to a few carbon-intensive staple commodities and to scope 1 (direct) emissions, and he discusses how both of these features might create perverse incentives.
Durel (2024) finds that the EU’s CBAM is likely to be found compatible with the World Trade Organization’s rules. This might not be the case if jurisdictions adopt rules such as the Foreign Pollution Act which only look like border adjustments. We can expect that this will be a fruitful area for international trade lawyers.
Kurz and Rubínová (2024) find, rather intuitively, that carbon pricing also hurts exports in carbon-intensive industries. While a border adjustment is meant to equalize prices between domestic manufacture and imports, there is no equivalent mechanism to aid exports.
Carbon Offsets
If activities that generate greenhouse emissions are taxed, then it stands to reason that activities that reduce emissions should be subsidized. This is done through offsetting, another area that, despite its theoretical appeal, turns out to be more complicated that one imagines.
Both Washington State’s cap and invest program and California’s cap and trade feature offsetting provisions. Both programs provides sets of rules of approved programs, which feature programs related to forestry, livestock, and ozone depleting substances, and in California’s case, rice methane capture and mine methane capture.
The Ecosystem Marketplace reports that the average carbon offset in 2023 cost $6.97 per ton, quite a good deal. But there are all sorts of questions about how effective they are. Cames et al. (2016) find that 85% of forestry projects under the Clean Development Mechanism, a troubled United Nations carbon offset program, overstate emissions reductions. Potts (2023) of Carbon Direct finds than less than 10% of the offset projects they examined meet their standards of effectiveness.
Independently of taxes or cap and trade, the 45Q program, created under the Bipartisan Budget Act of 2018 and expanded under the Inflation Reduction Act of 2022, creates a tax credit for carbon capture and sequestration. Indirect though it is, 45Q is the closest thing to a carbon price at the national level in the United States. Friedmann and Potts (2024) at Carbon Direct highlight direct air capture as a solution for carbon removal. At generally $500-1000 per ton, direct air capture is still prohibitively expensive at a large scale, but unlike other offset schemes, direct air capture has the potential both for large scale usage, durability, and certainty that it is removing carbon dioxide as advertised.
The Necessity of Carbon Pricing
I have only been able to scratch the surface of the issues around carbon pricing, and I suspect that I will have more depth on some of these topics later. But I want to conclude today by outlining why carbon pricing must be at the center of any credible climate policy.
Carbon pricing makes sense. Greenhouse gas emissions have social costs that are, in an unregulated market, imposed on the public. Carbon pricing insures that producers and consumers of fossil fuels pay those costs instead. I have tried not to shy away from the real challenges of carbon pricing, such as use of the revenue, carbon leakage, offsets, and proper calculation of the price. But all of these challenges are tractable and fail to override the basic logic of carbon pricing.
The ideal policy, in my view, is a economy-wide carbon tax on the order of $100/ton of CO₂ with provisions for proven-effective offsets, a border adjustment, and with the revenue used to offset corporate taxes. I recognize that for reasons of political economy, many compromises with this formula are necessary.
Carbon pricing complements an innovation-centered approach to climate policy. Innovations such as advanced nuclear power, enhanced geothermal, and next generation solar cells will enter the market more slowly or not at all unless there is a market pull like carbon pricing. Some innovations, like direct air capture and carbon capture and sequestration, are not feasible at all without a market pull. Additionally, carbon pricing aids upstream innovation by giving research and development a strong market signal.
Carbon pricing is superior to a policy based on regulation and subsidy. Regulations such as Corporate Average Fuel Economy standards offer a limited pathway to reducing emissions, which may or may not be economically wise. Carbon pricing offers consumers the choice of more fuel efficient vehicles, carpooling, living closer to work or remote work, mass transit, eco-driving, or another way of reducing emissions.
The most puzzling objection I see is that carbon pricing is a bad approach because it is politically infeasible. Aside from being circular logic, this argument does not consider which policies are most politically feasible for their effectiveness. There are many policies which, like cheap carbon offsets, are politically popular but accomplish little.