Solutions
Industries
Regulations
Resources
Company
Key Takeaways
In the previous blog, I ended with what seemed like the obvious question: if carbon is the problem, why not just make carbon expensive and let the market figure out everything else?
I find this idea rather attractive. Partly because it is simple, but mostly because it requires the regulator to know remarkably little.
Imagine two factories making exactly the same product. Factory A can reduce one tonne of CO₂ for €20. Factory B would need to spend €200 to achieve the same reduction. If the regulator simply tells both factories to reduce emissions by the same amount, we could end up spending €200 somewhere when the same tonne could have been eliminated for €20 elsewhere.
Put a sufficiently meaningful price on carbon instead and the calculation changes. Factory A looks at its €20 abatement option and has every reason to take it rather than continue paying for carbon. Factory B may reasonably decide that, for now, paying for its emissions is cheaper.
The regulator doesn’t have to decide which factory should install which technology. It doesn’t need to know whether electrification, efficiency, hydrogen, carbon capture or something nobody has thought of yet is the best solution. It creates the constraint and allows millions of individual decisions to determine where emissions can be reduced most cheaply.
As climate regulation goes, this is surprisingly respectful of markets.
It is also the basic idea behind the EU Emissions Trading System.
The EU ETS works on a cap-and-trade principle. Europe decides how much greenhouse gas the sectors covered by the system can collectively emit and expresses that cap through allowances. One allowance, or EUA, gives its holder the right to emit one tonne of CO₂ equivalent. Companies need to surrender enough allowances to cover their emissions, while the total number available under the cap falls over time. Allowances can be auctioned and traded, so the carbon price itself emerges from the market rather than being directly fixed by the regulator.
That distinction from a carbon tax is worth lingering on. With a carbon tax, the government essentially decides the price of emitting and the market determines how much emissions fall in response. With an emissions trading system, the government decides the quantity of emissions it is prepared to allow and the market determines what the right to emit that quantity is worth.
There is something almost elegant about this arrangement. The government creates scarcity around something that previously had very little scarcity attached to it, the atmosphere’s ability to absorb carbon, and then allows a market to form around that scarcity.
The commodity itself exists because the regulator created it. The regulator decides the cap, who participates, how allowances enter the market, what needs to be reported, what happens if a company doesn’t surrender enough allowances and how quickly the supply of allowances falls.
The market discovers the price, but the regulator designs the playing field.
This distinction becomes important because the history of the ETS is a rather useful lesson in what happens when that playing field is designed badly.
The EU ETS began in 2005 and its first phase was explicitly treated as a three-year learning period. Europe had a fairly basic problem: it did not yet have sufficiently reliable emissions data from the installations it was trying to regulate. Caps were therefore set largely on estimates.
The estimates were wrong.
More allowances were issued than the market needed. Once participants realised that supply exceeded emissions, there wasn’t much reason for the allowances to retain significant value. By 2007, the price of the first-phase allowances had fallen to zero.
I find this episode fascinating because it strips an emissions market down to its economic bones. The allowance has no intrinsic value. You cannot eat it, burn it or manufacture something from it. Its value exists because regulation makes it scarce.
Remove the scarcity and you remove the value.
Europe learned from that experience, collected better emissions data and tightened the cap for the second phase. Then the 2008 financial crisis arrived. Industrial activity and emissions fell much more than anticipated, demand for allowances weakened and a large surplus accumulated, again depressing the carbon price.
The lesson this time was slightly different. Even if the regulator has better information when designing the market, the economy doesn’t politely behave according to the assumptions in the model.
Europe subsequently changed the system considerably. A single EU-wide cap replaced the earlier national caps, auctioning increasingly replaced free allocation as the default, allocation rules became more harmonised and eventually the Market Stability Reserve was introduced to help deal with structural imbalances in allowance supply. The system that exists today is therefore not the elegant little market mechanism someone designed once and then left alone. It is the result of two decades of learning, adjustment and increasingly sophisticated market design.
Which makes the phrase “let the market decide” a little more complicated than it first appears.
The ETS is probably one of the more market-friendly instruments in Europe’s climate-policy arsenal. Once the emissions cap exists, firms retain considerable freedom over how they respond. A company can reduce emissions, change its energy source, improve efficiency, invest in a new process, buy allowances or combine several of these approaches.
But the market can only optimise within the boundaries it has been given.
If Europe creates too many allowances, carbon becomes cheap. If it creates fewer, carbon becomes expensive. If the cap falls faster, scarcity increases faster. If sectors receive allowances for free, their exposure changes. If new sectors enter the ETS, demand changes. If allowances are moved into or released from the Market Stability Reserve, supply changes.
In other words, there is still an enormous policy decision sitting before the market gets to do any price discovery: how scarce should the right to emit carbon actually be?
The current system makes that scarcity progressively tighter. Following the 2023 revision, the ETS cap is intended to bring emissions in the covered sectors down by 62% by 2030 compared with 2005. The system has also expanded, including maritime transport from 2024, while free allocation is being reduced in parts of the system.
So when we say the ETS “puts a price on carbon,” that price is not some naturally discovered fact about the economic cost of climate change. It is the market price of a politically determined scarcity.
That doesn’t make it artificial in the sense of being meaningless. Plenty of economically important rights exist because societies define and enforce them. But it does mean that anyone trying to understand the carbon market has to understand both economics and regulation. You cannot really separate the two.
This is where the theory runs into industrial reality.
If the entire purpose of the ETS is to make emitting expensive, giving companies free allowances initially sounds like a spectacular way of defeating the point.
But imagine a European steel producer competing with a producer outside Europe. The European producer faces a carbon constraint that increases its cost of production while its competitor may not. If the European plant loses market share, shuts down and the same steel is simply produced elsewhere with equal or higher emissions, Europe has reduced its territorial emissions without necessarily reducing global emissions.
This is carbon leakage.
Europe’s response has historically included free allocation of allowances to sectors considered particularly exposed to this risk. Free allocation is benchmarked rather than simply reimbursing whatever a company emits; benchmarks are linked to the emissions intensity of the more efficient installations in a sector and tighten over time.
There is an important economic point here. Receiving an allowance for free doesn’t make carbon economically irrelevant. If I have an allowance that I can use or sell, using it to cover my own emissions still has an opportunity cost. Reducing emissions can leave me with an allowance of value rather than consuming it.
But free allocation does soften the immediate financial exposure of industries Europe is worried about losing.
And this is where a climate policy that began with the beautifully clean idea of pricing an externality runs head-first into trade policy, industrial competitiveness and politics.
If Europe wants the carbon price to become stronger while continuing to protect industry from competitors that do not face an equivalent carbon constraint, eventually it needs another answer.
That answer is CBAM.
But I want to leave CBAM aside for now because there is an even bigger problem with the idea that carbon pricing should be enough.
Suppose producing a tonne of conventional material costs €1,000. The low-carbon alternative costs €1,400. A carbon price adds €100 to the conventional route.
The cleaner option is still €300 more expensive.
There is nothing obviously broken about the market in this example. The market has seen the carbon price and responded perfectly rationally. The low-carbon technology is simply still expensive.
We then have to ask why it is expensive.
Perhaps it needs infrastructure that nobody wants to build before demand exists. Perhaps it is an early technology whose costs could fall dramatically with scale and learning. Perhaps the first few projects are extraordinarily difficult to finance because nobody knows whether there will be buyers ten years from now. Those are all plausible reasons for additional intervention.
But there is another possibility that I think climate discussions sometimes glide past rather quickly: perhaps the technology is simply expensive.
Markets can squeeze inefficiency out of technologies. They can create scale, competition and learning. They cannot repeal physics.
If renewable hydrogen, e-fuels or another pathway remains structurally more expensive even after infrastructure, scale and learning have done their work, then subsidising or mandating it is no longer merely helping an infant technology get started. We are choosing to pay a premium for that pathway because we value the emissions reduction or some other strategic benefit enough to justify the cost.
That may still be the right decision. But economically, it is a different decision.
And this is where the Green Deal begins moving away from simply correcting the price signal.
Once I started looking at EU climate regulation this way, I found it useful to imagine a spectrum.
At one end is carbon pricing. The regulator identifies the thing the market is failing to price, creates a price for it and gives businesses considerable freedom to respond.
Move a little further and the regulator sets a performance outcome. A fuel, building or vehicle has to become cleaner, but companies retain some choice over how they achieve it.
Move further still and the regulator starts requiring particular categories of solutions: a renewable-energy share, a Sustainable Aviation Fuel mandate, a renewable-hydrogen target.
And then there are subsidies, infrastructure programmes, domestic manufacturing targets and other forms of industrial policy designed to make particular parts of the transition happen.
The further we move along that spectrum, the more information the regulator needs to have about what the future should look like.
Sometimes there is a very good reason for that. Markets can have coordination problems, infrastructure bottlenecks, innovation spillovers and very short investment horizons relative to a climate target several decades away.
But every additional intervention deserves a question that I think we ask too rarely:
What is the market failure that this regulation is fixing, and is the intervention narrower and cheaper than the problem it is trying to solve?
That question becomes especially interesting with renewable-energy mandates.
Because once Europe moves from “carbon must become more expensive” to “renewable energy must occupy this much of the energy system,” it has made a fundamentally different economic choice. It is no longer only telling the market what problem to solve. It has started telling the market something about what the answer should look like.
And that is where I want to go next.
