I recently finished the last session of Introduction to System Dynamics at MIT Sloan School of Management part of my MIT Executive MBA Program journey, led by the legendary Professor John Sterman. The session was a simulation of a UN plenary conference. The class was split into negotiating blocs, e.g., emerging markets, the big oil companies, developed nations, industry stakeholders, and climate activists (which is where I was sitting).

I was, let us say, volunteered as the group's spokesperson, cast in the Greta Thunberg role. I did it, after messing up my hair (nothing personal Miss Thunberg), and I brought a more “Scientific Greta’s version”. The casting invites a certain register. I went the other way and argued the engineering case: of all the levers available to us, the one that acts most powerfully on industrial emissions is the cost of carbon (of course, based on my prior knowledge of the matter, and extensive experience in modelling of the steel industry decarbonisation). Not because it is morally satisfying, but because it is the only intervention that reaches every sector, every fuel and every production decision simultaneously, upstream, without anyone having to pick which technology wins.

Then Professor Sterman opened En-ROADS (https://www.climateinteractive.org/en-roads/), the climate simulator built by Climate Interactive together with MIT Sloan, a full system-dynamics model of energy, economy and climate that lets you move a policy slider and watch the temperature curve to 2100 respond in real time. We went through the plausible favourites one by one. Renewables. Nuclear. Efficiency. Electrification. Afforestation. Methane. Each one moves the needle, and each one, alone, moves it disappointingly little. Then he pulled the carbon price slider, and the curve bent in a way nothing else had managed. The honest version of that lesson is not "carbon pricing solves it". En-ROADS is very good at killing silver bullets, and it takes a portfolio to get anywhere near 1.5 °C. The lesson is about leverage: among single interventions, the price of carbon is the highest-leverage point in the system, because it is the one that propagates through every feedback loop at once rather than optimising one node.

I came out of that room with a conviction and a caveat. The conviction is the one above. The caveat is the thing the model does not show you, because it sits outside the model's boundary: a carbon price only does any work if somebody actually pays it, and once somebody does pay for it, it matters enormously who collects it.

Which brings me to a recent paper:

A policy can look successful on paper: a climate law, an ETS, a pilot phase, allocation plans, benchmarks and a market stability mechanism. But if the system does not create a carbon cost actually paid by exporters, it cannot translate into CBAM relief. That is the issue identified in a recent Journal of International Trade Law and Policy article on Turkey. Article 9 of the CBAM Regulation allows EU importers to reduce the certificates they must surrender only by the carbon price effectively paid in the country of origin. Turkey's Climate Law establishes a national ETS, but its 2026–2027 pilot grants covered installations 100% free allocation. A price that is not paid cannot be deducted. The paper’s most persuasive argument is the customs-law parallel. “Price actually paid or payable” reflects the principle of economic reality in Article 70 of the Union Customs Code, not merely the formal existence of a price.

Some argue about Switzerland (where I live). But Switzerland is not the right comparison, as its ETS is linked to the EU ETS. Korea is more instructive, i.e., it began with near-total free allocation and gradually reduced it. The transition is possible. But 100% free allocation is not failure, it is a domestic political settlement with industry demonstrating that an ETS designed to exist before it is designed to bite.

The part of CBAM that deserves more credit than it gets:

Most of the commentary on CBAM argues about whether it is green protectionism or a legitimate anti-leakage device. I think that framing misses the more interesting thing the instrument does. CBAM does two jobs at once, and they pull in the same direction. The first is the obvious one. As the EU phases out free allocation in CBAM sectors (the CBAM factor opens at 2.5% in 2026 and reaches 100% in 2034) European producers begin paying the full carbon price. Without a border adjustment, that is a straightforward competitiveness transfer to importers. CBAM holds the line domestically while the phase-out happens, which is the only reason the phase-out is politically strong. The second job is the one I find genuinely a positive byproduct. Article 9 makes the EU's border tax a INDUCER for foreign carbon pricing, and it does so through the most reliable incentive there is, i.e., money. The carbon cost a third country declines to collect at home does not vanish. It is collected at the EU border and paid into the EU budget. Turkey’s own numbers make the point, roughly €5.6bn of CBAM-covered exports to the EU in 2025, and for steel alone an indicative annual exposure of some €300m that, under the current design, is almost entirely non-deductible. That is not a carbon price avoided. It is a carbon price exported, to a treasury in Brussels rather helping the descarbonisation at home.

The EU is not telling Turkey, Brazil, India or anyone else to price carbon. It is telling them that somebody is going to price the carbon in their exports, and they can choose who gets money. Framed that way, the domestic case for an ETS stops being an environmental argument and becomes a revenue argument, a far easier one to win in a ministerial discussion. Of course that this logic works best for economies with heavy EU exposure, and Turkey, sending roughly half its goods exports to the Union, is close to the ideal case. It works far less well for a diversified exporter with marginal EU trade, where the rational response is to redirect embedded emissions elsewhere rather than build a market. And Article 9 credits prices, not policies, as a country regulating emissions through standards, mandates or performance requirements gets no credit at all. The instrument rewards a particular institutional choice, and it does so whether or not that choice is the most efficient one locally.

The window is not ideal, but not totally lost either:

None of this is fatal for Turkey. The paper's prescription of auction 15-20% of allowances in the second pilot year, index a reserve price to the EU ETS, synchronise the free allocation phase-out with the CBAM factor, is modest, affordable. The uncomfortable point is the timing. 2026 and 2027 are not preparation years any more, rather they are the years the money starts moving. Every tonne shipped to the EU under a 100% free allocation pilot is a tonne on which Turkey chose, in effect, to let someone else collect the carbon revenue. And not only Turkey, it is for every other country exporting into EU. The question was never whether other country’s industry would pay a carbon price. The only open question is who cashes the cheque.

The end of the history is that while playing the “more scientific Greta Thunberg”, we managed to “convince the oil and industrial parties of our fictitious UN convention” to raise carbon price to $250/ton! And thanks to Prof. Sterman for the masterclass!

Source commented on: Ataç Aytuğ, "Turkish carbon pricing after Climate Law No. 7552: an assessment of the CBAM and the National Emissions Trading System", Journal of International Trade Law and Policy, 2026, DOI 10.1108/JITLP-07-2026-0122 (open access, CC BY 4.0).