Every mature market stands on a thin layer of shared infrastructure. Carbon doesn't — yet.
By Chris Canavan, CEO
Every mature, deep financial market rests on a thin layer of foundational infrastructure that performs the same basic functions: an authoritative record of title, settlement finality, common identifiers, and a common way to convey information between institutions. Equities have it. Bonds have it. Foreign exchange has it. The specific institutions differ, but the layer is always there, and it always does the same few things.
Carbon markets, for the most part, lack this layer.
That is worth pausing on. This layer’s absence is not the explanation usually offered for why carbon markets remain small, illiquid, and distrusted. The integrity of the underlying credits gets the attention, and it deserves the attention. But integrity alone does not explain why serious financial institutions — the ones that supply depth to every other market — remain on the sidelines. They stay out, in part, because the utilities they depend on in every other market cannot be found in carbon markets.
Why not?
It is not because we don't know how to build this infrastructure. The functions are well understood, the technology is mature, and financial markets have built this layer several times over in other markets. The layer is missing because the market’s most important players have not agreed on who should own it and how it should be governed. We face a problem of will, not engineering.
The infrastructure you cannot see
It helps to picture what the layer does when it exists. When someone buys shares on their phone, the transaction feels like a single event executed by a mobile app (think Robinhood). It isn’t. Underneath the app, systems that most of us have never heard of perform basic functions any financial transaction requires: establishing who owns what, moving shares and money so that neither side is cheated, and using a universal language to tell every institution involved what happened. All this happens in seconds. The mobile apps, brokerage houses, and banks can do what they do because they can rely on infrastructure for these functions. Most of us can name at least a couple of ride-hailing services, but very few of us could explain how GPS works. Yet without GPS, these ride-hailing services would have a hard time operating.
Little of this foundational infrastructure is a commercial product. It is shared and neutral. It is run to cover costs, not to maximize profits. It is generally owned cooperatively by the industry that relies on it, not sold to the industry as a commercial service. And it is invisible to most of us because it doesn't need a high profile.
In financial markets, each piece of this infrastructure exists because, at some point, the market was forced to answer the ownership and governance question that carbon has so far avoided.
What has forced the issue in other markets? Crisis or the threat of commercial capture
The financial system did not build foundational infrastructure in advance of scale. It was built under the pressure of scale. This pressure has historically come in two forms. Sometimes a crisis made the missing foundation impossible to ignore. And sometimes it was the looming threat of a commercial entity about to own the foundation. In both cases, it forced market players to have an epiphany: that collaborating rather than competing to build and govern the foundational infrastructure was the way forward.
Here are four episodes that illustrate my point:
The Depository Trust Company. In 1968, Wall Street still ran on paper share certificates couriered by hand around lower Manhattan. Volumes surged, back offices drowned, and the New York Stock Exchange (NYSE) was forced to close one day a week just to clear a backlog. This “Paperwork Crisis” was enough for brokerages and the NYSE to establish the Depository Trust Company, a cooperative owned by the financial institutions that use it. Its main function was and is to immobilise certificates and record changes of ownership electronically. Today, it processes almost $5 quadrillion of transactions.
CLS Bank. In 1974, German regulators closed Bankhaus Herstatt, a mid-sized Cologne bank, in the middle of the business day. Its counterparties in the foreign exchange market had already paid out their side of FX transactions with the bank, but Herstatt’s accounts were frozen before it had made its counterpayments. Herstatt’s defaults rippled dangerously through the global banking system, awakening banks and their regulators to foreign exchange settlement risk (often called, simply, “Herstatt” risk). It took almost three decades to coordinate the banking system to provide a solution. In 1998, sixty of the world's largest banks came together to build CLS Bank, where both legs of a currency trade settle simultaneously or not at all. It was launched in 2002, and is today owned by roughly 75 financial institutions and settles almost $7 trillion a day.
SWIFT. In the early 1970s, banks confirmed cross-border payments in free-form telex, with no standard codes and constant errors. This was a crisis waiting to happen, so one of the world's largest banks (Citibank’s predecessor) began building a proprietary messaging network of its own, pressing its correspondent banks to adopt it. The rest of the banking industry soon grasped that whoever owned the wires of interbank communication would gain invaluable insight into every competitor's business. So in 1973, 239 banks from 15 countries pre-empted that outcome by forming a cooperative and building their own network. SWIFT now connects more than 11,500 institutions and remains owned by its members. It exists in its current form because the alternative — a system controlled by one large commercial player — was unacceptable.
Euroclear. The Eurobond market's version of the paperwork crisis produced its own settlement system in 1968 built and owned by a single commercial bank, Morgan Guaranty, out of its Brussels office. The system worked for Morgan Guaranty (now JPMorgan Chase). But other banks refused to let a competitor hold records of their trades, and within two years a rival consortium launched a competing depository to escape Morgan's control. Morgan eventually sold Euroclear to a cooperative of its users in 1972, retaining only an operating role. The market had delivered its verdict: it would accept the infrastructure, but not the owner.
The DTC and CLS Bank were born of crises. SWIFT and Euroclear were born of the threat of commercial capture. All four alighted on the same institutional form: a layer of infrastructure owned and maintained cooperatively that performs the basic functions a market needs to operate at scale (e.g., a common ownership record, efficient title transfer, a common language). Commercial actors could compete fiercely for market share, but only after they had cooperated to put the foundation in place.
Why the answer is always the same
That convergence to cooperative ownership is not an accident of history. Three properties of foundational infrastructure are incompatible with running it as a profit-maximising competitive business.
It is more valuable when it is more widely adopted. Foundational infrastructure, even when it is spread over different jurisdictions, runs on a common set of principles and connections, so that, for example, a record of ownership in one place is legible and reliable everywhere else. This means the value of the network increases with every new participant.Everyone benefits from the existence and growth of the network, but it requires cooperative ownership. Commercial ownership would confer an unfair advantage to the owner, and the rest of the market would shun the network, reinforcing fragmentation. This is what we have in carbon markets today.
Its essential characteristic is neutrality. The second property follows from the first. A settlement system owned by one firm is a competitive threat to every other firm, and will be routed around. This is the Euroclear story, and the reason SWIFT's founders would not wait to find out how a rival's network would treat them. The system’s operator must remain neutral rather than compete with those who might rely on it.
Its outputs are public goods. The third property also follows from the first. Features like settlement finality, common identifiers, definitive title benefit every participant in the market at once, whether or not they paid for them. A profit-maximising owner of such outputs has every incentive to charge for them, to create frictions for their commercial advantage, and to tax growth rather than enable it. The right owner is an entity whose explicit mandate and only incentive is to perform these functions as a public service so the market above it can expand.
In short, the carbon markets cannot avoid the governance question any longer. Ownership of foundational infrastructure is a design decision on which everything else depends.
Where carbon markets stand
Carbon markets are not short of infrastructure in the ordinary sense. Exchanges, rating agencies, brokers, insurers, and data providers are all present, and all appropriately commercial.
What’s missing is the layer underneath them. There is no authoritative record of title. Trades settle bilaterally, with elaborate and bespoke arrangement to deal (imperfectly) with settlement risk. There are no common identifiers spanning programs, and no common language in which the market's institutions convey what happened. These deficiencies are made worse by the fragmentation of the market.
But the market has yet to experience the epiphany that delivered foundational infrastructure in other markets. The market isn’t big enough for scale to overwhelm the plumbing, as it did in the prelude to the DTC and CLS Bank. The market is also small enough to allow commercial firms to become the connective layer of the market, where registries link, data consolidates, and transactions flow. These platforms are useful and their commercial ambition is legitimate. But the lesson of Euroclear and SWIFT is that when the record of ownership and the language of the market sit inside one competitor's business, the rest of the market will eventually reject it.
Can we build the necessary infrastructure without a crisis or capture?
The layer that needs to be shared is thin: title, settlement finality, custody, common identifiers, a common way to convey information. Everything above it (e.g., ratings, analytics, structuring, insurance, execution, market-making) is commercial territory, and should remain so. Indeed, those businesses can flourish once the foundation is solid. DTC did not put brokers out of business; it made modern brokerage possible. CLS Bank did not displace FX trading desks; it let them trade with counterparties they would otherwise have refused. SWIFT's shared identifiers did not stop banks competing; they stopped banks misdirecting each other's payments.
This layer is missing in carbon markets because we have not learnt the lessons from other markets, which is that this layer must be owned and governed cooperatively, not by commercial actors.
At GCMU we are doing our part to address foundational deficiencies that the carbon market must overcome. GCMU’s register is built to provide the authoritative record of title the market needs. Moreover, it is a registry of title that meets bank-grade institutional standards, ensures registry assets are protected from insolvency, and is structurally free from the competing commercial interests embedded in every platform currently in the market. And, critically, it is governed and operates as a public utility putting climate impact and market functionality above profit.
We can wait for a crisis or capture to teach the carbon markets this lesson, just as they have in other financial markets. But why wait when we don't have time? We should learn from history and establish this foundational infrastructure while the market is still small, so that when it scales, as it must, the plumbing is ready for scale.
Now is the time for the market to rally around and build the entire layer of foundational infrastructure.
—--------
Sources: DTCC and DTC published volumes; CLS Group, ShapingFX series; SWIFT, 2025 Annual Review; Euroclear published history.
