Carbon Pricing: Putting a Number on the Externality
Almost every energy comparison in this library quotes costs that exclude the damage done by emissions - which is precisely the problem carbon pricing exists to fix. It is the most economically orthodox climate policy there is, supported across the political spectrum in theory, and the one whose implementation is fought over hardest in practice.
Why Put a Price on It at All
When a power station burns coal, the electricity has a cost that appears on a bill and a set of costs that do not: respiratory illness downwind, crop damage, and the warming contribution of the carbon dioxide. Those are borne by other people, often in other countries and other decades. Economists call this an externality, and its defining feature is that the market price is wrong - too low - so the market produces too much of the thing.
Carbon pricing corrects the price rather than prohibiting the activity. A power producer facing 80 euros per tonne finds that coal at 820 grams per kilowatt-hour carries an extra 65 euros per megawatt-hour, while gas at 490 grams carries 39. That difference changes the merit order directly and can move gas ahead of coal without anyone banning anything.
The appeal to economists is that it does not require the regulator to know which technology should win. Every emitter faces the same price and each finds its own cheapest response - fuel switching, efficiency, capture, or paying and continuing. The cheapest reductions happen first, which is exactly what a fixed technology mandate cannot guarantee.
The objection is distributional rather than economic. A carbon price is regressive by default: energy is a larger share of a poor household's spending, so an unmitigated price hits them hardest. Every functioning scheme has had to answer that, usually by returning the revenue, and the schemes that did not have generally been repealed.
Tax or Trade
A carbon tax sets a price per tonne and lets emitters decide how much to emit at that price. The cost is certain, which businesses like, and the resulting emissions are uncertain. Sweden has run one since 1991, now above 100 euros per tonne, alongside a 30 percent fall in emissions and continued economic growth.
An emissions trading system does the opposite. It sets a total quantity of allowances, distributes or auctions them, and lets firms trade. Emissions are certain because the cap is the cap; the price is uncertain and can be volatile. The EU system is the largest, covering power, heavy industry and aviation - roughly 40 percent of EU emissions.
The EU system took two decades to work properly, which is the most instructive part of its history. Early phases over-allocated allowances and handed most of them out free, so the price collapsed to near zero and changed nobody's behaviour. Reforms tightened the cap, moved to auctioning, and added a market stability reserve that withdraws allowances when a surplus builds. The price went from under 10 euros in 2017 to above 80 in the mid-2020s, and only at that level did it begin to influence investment decisions rather than merely appear on a balance sheet.
Hybrid designs have become common because the pure forms have symmetric weaknesses. Price floors stop a trading system collapsing when demand falls; price ceilings stop it spiking beyond what industry can absorb. Most systems designed since 2015 include both.
Leakage and the Border
The hardest problem in carbon pricing is that the atmosphere is global and jurisdictions are not. If Europe prices carbon and Turkey does not, a steel plant can move, and the emissions move with it. Europe's inventory improves, the world's does not, and Europe loses the industry as well. This is carbon leakage, and fear of it has constrained every carbon price ever set.
The original mitigation was free allocation: give the exposed industries their allowances for nothing so they face no cost disadvantage. It prevents leakage and also removes the incentive to abate, which is the point of the scheme. Europe ran this compromise for years and knew it was a compromise.
The Carbon Border Adjustment Mechanism is the alternative. An importer of steel, cement, aluminium, fertiliser, hydrogen or electricity into the EU must buy certificates matching the emissions embedded in the goods, priced at the EU allowance price, minus whatever carbon price was already paid at origin. A domestic producer and an importer then face the same carbon cost, and free allocation can be withdrawn.
The mechanism entered its reporting phase in 2023 and its financial phase from 2026. It is contested at the World Trade Organization and by several exporting countries as disguised protectionism, and it has also had the effect its designers hoped for in one respect: several trading partners have accelerated their own carbon pricing, because a price collected at home is revenue kept at home rather than paid to Brussels.
What Price Is High Enough
Two different questions get confused here. The social cost of carbon asks what damage a tonne does, and estimates range from about 50 to over 200 dollars depending on the discount rate applied to future harm and on whether low-probability catastrophic outcomes are included. The abatement cost asks what price is needed to trigger a given change, which is a much more concrete question with technology-specific answers.
The practical thresholds are fairly well established. Switching from coal to gas in power generation happens somewhere around 30 to 50 euros per tonne depending on relative fuel prices. Making carbon capture commercially attractive on industrial processes needs perhaps 100 to 150. Hydrogen-based steelmaking needs a similar range. Direct air capture needs several hundred.
Actual prices span three orders of magnitude. Some schemes charge under a dollar a tonne and exist mainly to establish the administrative machinery. The EU sits above 80, Sweden above 100, and the majority of covered emissions worldwide are priced below 20 dollars - which is below almost every abatement threshold that matters.
The honest summary is that carbon pricing works where it is high, broad and credible over time, and that most existing schemes fail at least one of those tests. The credibility criterion is the one most often overlooked: a firm deciding whether to build a plant that will run for 40 years needs to believe the price will still be there in 2050, and no government can bind its successors. That uncertainty is why carbon pricing alone has rarely been enough, and why it usually appears alongside standards and subsidies rather than instead of them.
Frequently asked questions
What is the difference between a carbon tax and emissions trading?
A tax fixes the price per tonne and lets emissions settle wherever they settle. A trading system fixes the total quantity of allowances and lets the price move. One gives certainty about cost, the other certainty about the environmental outcome. Most modern systems are hybrids with a price floor and ceiling.
What is carbon leakage?
Production moving to a jurisdiction without a carbon price, so the emissions relocate rather than disappear. The exporting region loses the industry and the atmosphere gains nothing. Fear of leakage led to free allocation of allowances to exposed industries, and later to border adjustment mechanisms.
What does the EU border mechanism do?
It charges importers of steel, cement, aluminium, fertiliser, hydrogen and electricity for the emissions embedded in those goods, at the EU allowance price, minus any carbon price already paid at origin. This equalises the carbon cost between domestic and imported production so free allocation can be withdrawn.
How high does a carbon price need to be?
It depends what you want it to trigger. Coal-to-gas switching happens around 30 to 50 euros per tonne, industrial carbon capture needs perhaps 100 to 150, and direct air capture several hundred. Most covered emissions worldwide are priced below 20 dollars, which is beneath almost every meaningful threshold.
Isn't carbon pricing regressive?
By default yes, because energy is a larger share of a poor household's spending. Every durable scheme has addressed this by returning revenue - as dividends, tax cuts or targeted support. Schemes that raised prices without visible compensation have generally been repealed, which is a political finding rather than an economic one.
