EU Carbon Market History: From the EU ETS to CBAM and ETS2

Ask most people how a carbon market fails, and they picture prices spiking out of control. The EU ETS’s first real crisis went the opposite way: the price of a tonne of carbon dioxide fell from over €20 in April 2006 to a few cents by the end of Phase 1 in 2007. That is not a footnote in this market’s history — it is the event every later reform responds to.

Why the price fell to zero

Phase 1’s allowances were allocated through national allocation plans — each of the (then) 25 member states deciding, largely on its own, how many free permits its power plants and factories would receive. Those numbers were built on estimated emissions, submitted by industry and negotiated between governments and Brussels under real political pressure to protect domestic industry from cost increases. When the first year of independently verified emissions data came in during spring 2006, it showed emissions well below the allowances that had been issued. The market realised, almost overnight, that there were simply more permits in circulation than companies needed.

That alone might only have softened the price. What turned it into a collapse toward zero was a second design choice: Phase 1 allowances could not be “banked” — carried forward — into Phase 2. A permit that could not be used after 2007, and that nobody needed to buy in 2007 either, had no reason to hold any value as the deadline approached. Selling a worthless-at-expiry asset for whatever price the market would pay was the rational choice for every holder at once.

The failure was survivable, and arguably necessary, because it happened during a deliberately short three-year pilot phase rather than a long-term compliance period. Every reform since — Phase 2’s tighter, Kyoto-aligned cap; Phase 3’s move to a single EU-wide cap set centrally rather than negotiated nationally; and the 2015–19 Market Stability Reserve — targets some version of the same root problem: keeping the number of allowances in circulation close enough to actual demand that the price signal stays meaningful.

By the early 2010s, the EU ETS was carrying a structural surplus of more than two billion allowances — a legacy of the 2008 financial crisis’s demand shock layered on top of Phase 1’s original over-allocation. A market that oversupplied cannot fix itself simply by lowering the cap for future years; the old surplus keeps depressing the price regardless.

The Market Stability Reserve (MSR), established under Decision (EU) 2015/1814 and operating since 1 January 2019, is Europe’s answer: a rules-based mechanism that automatically adjusts how many allowances are auctioned each year, based on the Total Number of Allowances in Circulation (TNAC), published by the Commission annually by 1 June.

In plain terms: the EU created a mechanism that can remove allowances from circulation when the market contains too much supply — and, under the reverse condition, can release allowances back if circulation runs too low. If TNAC exceeds an upper threshold, a percentage of the surplus is withdrawn from the following year’s auctions and placed in the reserve, invisible to the market, rather than sold. If TNAC falls below a lower threshold, allowances can instead be released back from the reserve into auctions.

The scale of this mechanism in 2026 is large: the Commission’s 29 May 2026 notice confirms 190,494,202 allowances will be withdrawn into the MSR between 1 September 2026 and 31 August 2027, based on a 2025 TNAC of just over 1.02 billion allowances. That is not a rounded estimate — it is the exact legal output of the formula applied to that year’s published TNAC.

The MSR does not set the carbon price directly. What it does is remove the mechanism by which a one-off surplus, like the one created in 2007 or after 2008, can permanently depress the price for a decade or more. It is the structural fix for the exact failure mode this market experienced twice in its first ten years.

The Carbon Border Adjustment Mechanism is designed to solve a problem the EU ETS could not solve on its own: carbon leakage. If EU steelmakers pay for their carbon emissions and foreign competitors do not, production — and the emissions that come with it — can simply relocate outside the EU, or European buyers can switch to cheaper, carbon-intensive imports. Either way, global emissions do not fall; they just move somewhere the EU’s price signal cannot reach.

CBAM addresses this by requiring importers of covered goods to buy CBAM certificates matching the embedded carbon in what they bring into the EU, priced against the same weekly EU ETS auction average that EU producers already pay. It is best described as a border adjustment matching an existing internal carbon price, not a stand-alone tariff invented at the border — though the political debate over whether it functions partly like a trade-protection measure is real and unresolved (see below).

CBAM’s definitive regime covers six sectors: cement, iron and steel, aluminium, fertilisers, electricity and hydrogen — chosen because they are both carbon-intensive and exposed to international trade. An October 2025 “CBAM Omnibus” regulation simplified compliance (notably, a 50-tonne per importer per year de minimis exemption removing thousands of negligible importers from reporting obligations) but did not expand this sector list; a review of extending CBAM to organic chemicals and polymers remains under consideration, not yet adopted.

The European Commission published the first-ever official CBAM certificate price on 7 April 2026: €75.36 per tonne of CO2-equivalent for Q1 2026, calculated as the weighted average of EU ETS auction clearing prices for that quarter. Q2 2026’s price, published 6 July 2026, came in almost identical at €75.28. Because the certificate price tracks the EU ETS auction average directly, it moves with the same policy and market forces — the MSR’s withdrawals, industrial demand, energy prices — that move the EU ETS itself.

Free EU ETS allowances for CBAM-covered sectors are being phased out on a published schedule that mirrors CBAM’s own ramp-up: roughly 97.5% of previous free allocation retained in 2026, falling to around half by 2030, and toward zero by 2034. This timing matters — it prevents EU producers from simultaneously receiving free allowances and border protection against imports, which would have amounted to double protection.

ETS2, legislated under Directive (EU) 2023/959, is a separate emissions trading system covering fuel used in buildings, road transport and small industrial combustion not already covered by the original EU ETS. Crucially, it regulates fuel suppliers and distributors — roughly 11,400 entities across the EU — not individual households or drivers directly. In practice, a supplier’s ETS2 cost is expected to be passed through into the retail price of heating fuel and petrol, in the same way excise duties already are.

The date that changed: ETS2 was originally scheduled to become operational in January 2027. In November 2025, as part of agreeing the EU’s 2040 climate target, member states pushed that back one year — ETS2 is now scheduled to be fully operational, with binding compliance obligations, from January 2028. Limited voluntary early auctioning may still begin in 2027, but the regulatory start most sources cited through 2024 and early 2025 is superseded.

Because ETS2 puts a carbon cost into products — heating fuel, petrol, diesel — that households buy directly and cannot easily substitute away from in the short term, the EU built a companion mechanism: the Social Climate Fund (Regulation (EU) 2023/955, in force since June 2023). It is designed to help vulnerable households, micro-enterprises and transport users absorb ETS2-driven cost increases, funding measures like building renovation, clean heating and cleaner transport access.

The logic behind the Fund answers a fair question: why would Europe need a social fund if carbon pricing is supposed to change behaviour through cost signals? The answer is that a household’s ability to respond to a higher fuel price — by insulating a home, switching to an electric vehicle, or using public transport — depends heavily on income and existing infrastructure. Without support, the same price signal that nudges a wealthier household toward a heat pump can simply squeeze a lower-income household’s budget with no realistic way to respond. The Fund is Europe’s acknowledgment of that gap, not a concession that carbon pricing doesn’t work.

Carbon leakage is the risk that a carbon price applied in one jurisdiction simply pushes emissions somewhere else, rather than reducing them. The textbook version: a European steel producer faces real carbon costs; a competitor outside the EU does not; production, investment and jobs migrate to the lower-cost, higher-emissions location, and EU consumers switch to the now-cheaper imported steel. Global emissions do not fall — they may even rise, if the relocated production is less efficient.

This is exactly why the EU gave carbon-intensive, trade-exposed sectors free EU ETS allowances for two decades, rather than making them pay for every tonne from day one — a widely misunderstood policy choice. Free allocation was never meant to be permanent; it was a bridge, protecting competitiveness while the market matured, always intended to shrink as either global climate ambition converged or as a border measure like CBAM took over the leakage-prevention job instead. That handover is exactly what is happening now: free allocation for CBAM sectors declines on a published schedule as CBAM’s own coverage strengthens, precisely so industry is not protected twice, or left unprotected in the gap between the two mechanisms.

It is worth distinguishing carbon leakage’s different real-world forms: outright relocation of a factory is the most visible, but leakage can also happen through import substitution (EU buyers simply switch to a carbon-intensive imported product instead of a EU-made one) or through investment decisions (a company builds its next plant outside the EU rather than physically moving an existing one). CBAM is designed to address all three channels by pricing the imported good itself, not just the EU producer’s process.

Not a simple yes or no. On its own terms — creating a functioning, tradable price on a large share of Europe’s industrial and power-sector emissions, and sustaining it through a near-total collapse, a financial crisis and a pandemic — the EU ETS has clearly worked: it is the template every other major carbon market, from the UK to South Korea, has drawn on. Verified emissions from covered EU ETS installations have fallen substantially since 2005, though how much of that reduction is attributable to the ETS price signal specifically, versus renewables expansion, energy efficiency policy, fuel switching, and economic cycles including the 2008 crisis, COVID-19 and the 2022 energy crisis, is genuinely contested among researchers and should not be collapsed into a single “the ETS caused X% of the reduction” claim.

Where the record is more mixed: the 2007 collapse and the post-2008 surplus each cost the market close to a decade of weak price signals before repair mechanisms (Phase 3, the MSR) took full effect. Carbon leakage remained a real, unresolved risk for two decades before CBAM began to address it in 2026 — and CBAM itself remains contested, both legally and diplomatically, with major trading partners. The Social Climate Fund’s own headline figure is not even fully settled between official EU sources.

The next test is 2026–2028: whether CBAM survives its early diplomatic friction with India, China, the US and others without triggering a genuine trade dispute; whether ETS2 launches on its new 2028 timeline without the kind of political backlash that has already delayed it once; and whether a carbon price that now reaches EU factories, global steel and aluminium exporters, and, soon, European households, remains politically durable as its costs become more visible to more people at once.

Trace the arc of this timeline and a pattern is hard to miss: environmental policy became a financial market in 2005, that market became a lever of industrial policy through free allocation and the MSR, industrial policy became trade policy the moment CBAM reached the border in 2026, and trade policy is about to become social policy once ETS2 and the Social Climate Fund reach households in 2028. Few climate policies anywhere have been rebuilt this many times, this deliberately, in public view, with the failures left on the record rather than quietly buried. That is arguably the more interesting story than any single price chart: not that Europe priced carbon correctly on the first try, but that it kept adjusting the mechanism for two decades rather than abandoning the idea after a failure that would have ended a less durable policy outright.