One word, “carbon”, covers two assets that behave nothing alike: a ~$950bn regulated compliance market worth owning, and a voluntary offset market that has twice been shown to be selling tonnes that were never real. Knowing which half you are buying is most of the game.
Part of the Alternative Fortune series on commodities and natural resources.
Key takeaways
- Two assets, one name. Compliance allowances (regulated cap-and-trade) and voluntary offsets (project-based) are not the same trade. The investable core is the compliance side.
- The scale gap is enormous. Compliance allowances traded ~$948bn in 2023; voluntary offsets ~$4bn, roughly 240x.
- It is a policy-beta asset. KRBN fell ~40% in days after Russia’s 2022 invasion of Ukraine on a regulatory signal. Treat it as policy exposure, not a bond.
- The forward driver is the cap. BloombergNEF forecasts the EU carbon price at €149/tonne by 2030.
- The voluntary market has a trust problem. Over 90% of Verra’s rainforest credits were found likely worthless in a 2023 investigation.
The 60-second version
There is one label here, and underneath it sit two assets that behave almost nothing alike. The first is a compliance allowance, a government-created permit to emit one tonne of carbon dioxide, issued under a cap-and-trade scheme like the EU Emissions Trading System. Governments make these scarce on purpose and tighten the cap over time, so the permit trades like a regulated commodity with a policy engine behind it. The second is a voluntary carbon credit, a project-based offset generated by planting or protecting a forest, distributing cleaner cookstoves, or pulling carbon out of the air, with no central cap and no government forcing anyone to buy it. Almost every headline you have read about carbon markets going wrong came from the second half. Almost all the money and liquidity sits in the first.
Get that split wrong and everything downstream goes wrong with it. The compliance market traded roughly $948 billion in 2023; the voluntary market traded about $4 billion in the same year, a ratio of roughly 240 to one by value. When somebody says “carbon credits are a scam” they are usually right about a corner of the voluntary market and wrong about the regulated core. When somebody says “carbon is the commodity of the century” they are usually talking about the regulated core and quietly ignoring the graveyard next door. Both statements describe only one half of the market and treat it as the whole.
The two halves stay apart from here on, because almost every question that matters, how big the market is, what drives it, what you can actually buy, and how badly it can go wrong, has a different answer depending on which one you mean.
I. What it actually is
A carbon credit is a claim over one tonne of carbon dioxide equivalent, one tonne either kept out of the atmosphere or removed from it. That is the whole idea. The complication is that two entirely separate systems both mint these claims, and they do it for different reasons, under different rules, with wildly different reliability.
Compliance allowances come first. A government caps the total emissions a set of industries is allowed to produce in a year, issues one permit (an “allowance”) per tonne under that cap, and requires covered companies to surrender enough allowances to match what they actually emit. A power station that emits a million tonnes needs a million allowances. If it cleans up and emits fewer, it can sell its surplus; if it emits more, it must buy. That is cap-and-trade: the government sets the ceiling (the cap) and lets a market discover the price (the trade). The scarcity is manufactured, deliberate, and tightened on a schedule. The best-known scheme is the EU Emissions Trading System (EU ETS), and the unit it trades is the EU Allowance (EUA). California runs its own version with the California Carbon Allowance (CCA); there are equivalents in the north-eastern United States (RGGI), the United Kingdom, and China.
Voluntary carbon credits come from the other direction entirely. Nobody is legally required to buy them. A company that wants to claim it is offsetting its footprint pays a project developer who has, in theory, either avoided emissions that would otherwise have happened (protecting a forest from being cleared) or actively removed carbon from the air (planting trees, or running a direct air capture plant, machinery that filters CO2 out of the atmosphere). A standards body such as Verra certifies the project and issues credits. There is no cap. The number of credits depends entirely on how many projects exist and how generously they are counted, and that generosity is where most of the trouble in this market has lived.
The investment case is strongest, cleanest, and most liquid on the compliance side, because the scarcity is legislated rather than estimated. The voluntary market is where most of the reputational damage has happened, because the underlying “tonne” was often not real. Keep that distinction in your head from here on, because it is the single most important fact about the asset.
II. Market history and growth
Carbon markets are not new, but their scale is. The global carbon market’s total traded value reached roughly $949 billion in 2023, a record, up around 2 per cent on 2022 according to LSEG (formerly Refinitiv). That is not a niche. That is a market the size of a major commodity complex, and it got there because governments kept adding pricing schemes and tightening the ones they already had.
By 2024 there were 75 carbon-pricing instruments in operation worldwide, a count that includes both emissions trading systems and straight carbon taxes, per the World Bank. Those instruments now cover roughly 24 per cent of global greenhouse-gas emissions, up from about 7 per cent when the World Bank published its first report on the subject. That direction of travel is most of the thesis, since more of the world’s emissions are being priced every year and the price is being pushed up on purpose.
The money flowing to governments tells the same story. Global carbon-pricing revenue hit a record $104 billion in 2023. That revenue is the flip side of the cost imposed on emitters, which is what sets the value of an allowance. When governments collect more, allowances are worth more, and the schemes have every fiscal incentive to keep the pressure on.
Now the split reasserts itself. Of that ~$949bn total, compliance allowances alone traded roughly $948 billion in 2023, while voluntary offsets traded about $4 billion, the roughly 240-to-one ratio again. The headline “billion-dollar carbon market” is almost entirely a compliance-market story.
The voluntary market’s recent history is a story of value holding while volume shrinks. Its value in 2024 was 1.9 times its 2018 level, even though 2024 posted the lowest transaction volume since 2018, according to Ecosystem Marketplace’s State of the Voluntary Carbon Markets 2025 report. Look closer and the mechanism is clear: transaction volume fell 25 per cent in 2024 while prices fell only 5.5 per cent. The market shrank in quantity but the average credit got dearer, because buyers stopped touching the cheap, low-quality stuff. What looks like a market in decline is closer to one clearing out its junk.
III. Demand drivers
Demand for carbon exposure comes from three distinct engines, and they pull on the two halves of the market differently.
The first and most powerful engine is legislated cap tightening on the compliance side. This is the structural driver, and it is not a forecast about behaviour. It is a schedule written into law. Governments reduce the number of allowances they issue year after year. As the cap falls, the same emissions chase fewer permits, and the price is designed to rise. BloombergNEF forecasts the EU ETS carbon price will reach €149 per tonne by 2030, which would be the highest carbon price in the world. Whether that exact number lands is beside the point; the direction is baked into the legislation, and that is the closest thing the asset class has to a tailwind.
The second engine is corporate voluntary demand, and here the signal is quality, not quantity. Credit retirements in the voluntary market (the moment a buyer permanently cancels a credit to claim the offset, as opposed to reselling it) reached 182 million tonnes of CO2 equivalent in 2024, roughly flat since 2021. So actual use held steady even as speculative trading dried up. But underneath that flat line, buyers are voting hard on quality. Removal credits, where carbon is physically taken out of the air, sold at a 381 per cent premium over reduction credits in 2024, up from 245 per cent in 2023. Credits issued in the last five years carried a 217 per cent premium over older vintages in 2024, against just 53 per cent in 2023. Recency and permanence are what buyers now pay for. The demand is real; it has simply become discriminating.
The third engine is financial demand: allocators buying carbon for return and diversification rather than compliance or offsetting. This is the smallest and most fickle engine, and it flows almost entirely into the compliance side through futures-based funds, because that is the only part of the market liquid enough to hold at size. It is also the flow that dries up fastest when policy sentiment turns.
Ricardo Bayon, a partner at Encourage Capital, put the voluntary side’s underlying demand plainly in the Ecosystem Marketplace 2025 report:
“Underlying demand indicators continue growing steadily. Companies retiring carbon remain undeterred; markets will boom when trust improves.”
That last clause is the whole voluntary-market case. The underlying demand exists; what is missing is the trust to convert it, and that has not yet arrived.
IV. The players
The carbon market has no single dominant firm the way some asset classes do. It has schemes, standards bodies, index providers, fund issuers, and a handful of very large corporate buyers whose purchases move the narrative.
On the compliance and policy side, the players are governments and the schemes they run. The EU ETS is the reference market for the entire world. The World Bank is the authoritative scorekeeper; its annual State and Trends of Carbon Pricing report is the source most of the market cites for coverage and revenue. Axel van Trotsenburg, the World Bank’s Senior Managing Director, framed the institutional view in the May 2024 edition:
“Carbon pricing can be one of the most powerful tools to help countries reduce emissions. That’s why it is good to see these instruments expand to new sectors, become more adaptable and complement other measures.”
On the standards side, the dominant name is Verra, the largest certifier of voluntary credits and, as the case studies will show, the body at the centre of the voluntary market’s worst crisis of confidence. Ecosystem Marketplace, run by Forest Trends, is the main independent tracker of voluntary market data.
On the investable side, the most visible player is KraneShares, whose family of carbon ETFs is the main way public-market investors get exposure. Their global fund tracks an S&P index of carbon futures. Barclays issues the older iPath carbon note. BloombergNEF is the forecaster whose price projections anchor the bull case.
On the corporate-buyer side, the players who matter are the very large companies making landmark purchases that set price signals for the removal frontier. Microsoft is the most active and most important; its direct air capture deal, covered in the case studies, is the largest carbon-removal purchase ever recorded. Occidental Petroleum, through its subsidiary 1PointFive, is the counterparty on that deal and one of the biggest builders of direct air capture capacity. On the cautionary side, buyers named in the voluntary-market scandals include Gucci, Shell, easyJet, Salesforce and BHP, a roll-call that tells you how mainstream voluntary offsetting had become before the credits were called into question.
V. Geography
Carbon pricing is regional by design. Each scheme is its own jurisdiction with its own cap, price, and rules, so geography is not decoration here. It is the structure of the asset.
Europe is the centre of gravity and it is not close. The EU ETS accounts for 87 per cent of the global carbon market by value as of 2023, the single dominant venue on Earth. Any exposure to “carbon” as an asset class is, in practice, overwhelmingly exposure to European climate policy. The United Kingdom runs its own separate ETS after leaving the EU scheme, with its own allowance (the UKA).
China runs the world’s second-largest carbon market by value with its national ETS, launched to cover its power sector. It is enormous in emissions covered but historically thinner in traded value and price than the EU system, and it is far harder for an outside investor to access directly.
North America has no federal scheme but two significant regional ones. California’s cap-and-trade programme trades the California Carbon Allowance and is the largest carbon market in the Americas; it is one of the few compliance markets outside Europe that public investors can get clean single-market exposure to. The Regional Greenhouse Gas Initiative (RGGI) covers power-sector emissions across a group of north-eastern US states and is smaller again.
The voluntary market has a different geography entirely. Its supply is concentrated in the developing world, in the forests of Latin America, sub-Saharan Africa, and South-East Asia, because that is where avoided-deforestation and nature-based projects live. The demand is concentrated in the corporate West. That mismatch of Western buyers, tropical projects, and thin verification in between is the fault line along which the voluntary market’s scandals opened up. The Zimbabwe project in the case studies below sits squarely on that fault line rather than being an exception to it.
VI. How to actually invest
For most public-market investors, direct spot allowances are not practical to hold. You are not going to open an account on the EU ETS registry. The exposure comes through exchange-traded products, and almost all of them share a structural detail that is easy to miss: they hold futures, not spot allowances. That changes both what the product costs you and how it is taxed.
The compliance markets you can reach this way span the main jurisdictions: the EU ETS, the UK ETS, and North America’s two regional schemes, California’s cap-and-trade and RGGI. The voluntary market has no clean listed vehicle for retail investors, for reasons the case studies make obvious. The real vehicles, with their fees and structures, are these.
| Vehicle | Ticker | Structure | Exposure | Expense ratio | Notes |
|---|---|---|---|---|---|
| KraneShares Global Carbon Strategy ETF | KRBN | US ETF (futures) | Global: EUA, CCA, RGGI, UK allowances | 0.79% | AUM ~$167M (Sep 2025); tracks the S&P Global Carbon Credit Index |
| KraneShares California Carbon Allowance ETF | KCCA | US ETF (futures) | Single-market: California CCA | 0.91% | Pure California cap-and-trade exposure |
| KraneShares European Carbon Allowance ETF | KEUA | US ETF (futures) | Single-market: EUA | See issuer | Pure EU ETS exposure; launched alongside KCCA |
| WisdomTree Carbon | CARB | LSE-listed ETC (futures), UCITS-eligible | Single-market: EUA | 0.35% | Fully collateralised; tracks the Solactive Carbon Emission Allowances Rolling Futures index. The main EUA route for non-US investors |
| iPath Series B Carbon ETN (Barclays) | GRN | US ETN (bank debt) | EU ETS + Kyoto CDM credits | See issuer | Unsecured Barclays debt: issuer-credit risk, not a fund holding assets; NYSE Arca |
Four things to understand before you touch any of these.
First, the futures-roll cost. Because these ETFs hold futures rather than spot allowances, they must continually sell expiring contracts and buy later-dated ones. Depending on the shape of the futures curve, that “roll” can quietly cost or add to your return over and above the stated expense ratio. The ~0.79% fee on KRBN is the visible cost; the roll is the invisible one.
Second, the ETN counterparty risk. GRN is an exchange-traded note, not a fund. It is unsecured debt of Barclays, the bank’s promise to pay you the index return. If Barclays fails, your note is a claim in the bankruptcy, regardless of what carbon did. That is a fundamentally different risk from an ETF that actually holds futures positions. You are stacking issuer-credit risk on top of carbon-price risk.
Third, where you can buy. The KraneShares funds and GRN are US-listed. Investors outside the US who cannot buy US-domiciled ETFs generally reach the same EU ETS exposure through the WisdomTree Carbon ETC (CARB) on the London Stock Exchange, which tracks EUA futures at a 0.35% fee. It is the cheapest single-market EUA wrapper on this list and the default route for a European or UK investor.
Fourth, the minimum. There is no fund minimum beyond the price of a single share and whatever your brokerage charges. Access is not the barrier here. Understanding what you are actually holding is.
Single-market funds (KCCA, KEUA, CARB) give you a cleaner, more concentrated bet on one jurisdiction’s policy; the global fund (KRBN) spreads across EUA, CCA, RGGI and UK allowances but is still, by weight, mostly a European bet because Europe is most of the market.
VII. Unit economics: a worked example
Real figures put through a compliance-allowance position show the unit economics cleanly, because there is a single allowance carrying a single price whose direction is set mostly by policy.
The unit is the EU Allowance (EUA), one permit to emit one tonne of CO2. In 2024, the EUA averaged roughly €65 per tonne, well down from the record highs above €100 per tonne seen in early 2023. By 15 December 2025, EUAs traded at about €83.79 per tonne, up roughly 30 per cent year over year.
Take the worked case. An investor who bought EUA exposure at the 2024 average of €65 and held to the €83.79 December 2025 level captured a gross move of +€18.79 per tonne, or about +29 per cent, in roughly 12 to 18 months. The driver was straightforward: cap tightening squeezing the same emissions into fewer permits.
Now net it down. That +29 per cent is gross. Against it you have the fund’s ~0.79 per cent expense ratio and the futures roll cost, which is not fixed and can run either way depending on the curve. Neither of those turns a +29 per cent gross move into a loss, but they are the frictions that separate the headline allowance price from what actually lands in your account. The gap between spot EUA and a futures-based ETF’s return is exactly the roll plus the fee, and over long holds it compounds.
Extend the same base out to the bull case and the arithmetic gets loud. If BloombergNEF’s €149 per tonne by 2030 forecast holds, that is a move from the €65 base to €149, roughly +129 per cent gross, before fees and roll, over several years. That is the shape of the compliance case, a legislated tightening of the cap that puts a rising floor under the allowance price. It is not guaranteed, and a forecast several years out can miss badly, but the mechanism behind it is written into law rather than merely hoped for.
VIII. Macro sensitivity
Carbon behaves oddly against the macro backdrop, and that oddity is both the reason to hold it and the reason to be careful with it.
On one hand, carbon allowances have shown low correlation with traditional asset classes (US equities, bonds, commodities, real estate, gold and oil) over the period August 2014 to June 2024, per KraneShares. That is the diversification case: a return stream driven by policy rather than by the business cycle, which theoretically zigs when your equity book zags.
On the other hand, the thing that drives it is policy, and policy moves on headlines, not fundamentals. The February 2022 episode shows how brutal that can be: KRBN fell roughly 40 per cent within days of Russia’s invasion of Ukraine, as the EU signalled it might release temporary allowance-supply measures to cushion energy prices. No change in emissions. No change in the long-run cap. Just a regulatory signal, and the asset nearly halved in days. This is a policy-beta asset: its dominant risk factor is the behaviour of the regulator, not the state of the economy.
Here is how it behaves across four broad regimes.
| Macro regime | Likely carbon behaviour | Why |
|---|---|---|
| Growth + tight policy | Strongest tailwind | High industrial activity means real allowance demand; a tightening cap adds a legislated squeeze on top |
| Growth + loose policy | Mixed / capped | Emissions demand is there, but any signal of allowance-supply relief (as in 2022) can slam prices regardless of activity |
| Recession + tight policy | Weak but supported | Industrial emissions fall, softening demand for allowances; the cap floor limits how far the price falls |
| Recession + policy retreat | Worst case | Falling emissions and a government minded to loosen supply for relief: the 2022 dynamic, but without the growth |
The pattern worth holding on to is that real-economy demand sets the floor while regulatory posture sets the ceiling and the tail risk. The bet is less about the level of emissions than about how committed governments stay to making those emissions expensive. When that commitment wobbles, even temporarily or rhetorically, the asset can move violently.
IX. Tax
This section is general and jurisdiction-neutral, and it is not tax advice. Carbon exposure is taxed by the wrapper you buy it through, and your treatment depends entirely on where you are resident and what you hold. Verify the specifics with a qualified adviser in your own jurisdiction before you allocate.
The one durable point worth carrying everywhere is this: the word “carbon” hides three different tax animals. The label says little about how a gain will be treated.
- A futures-based ETF like KRBN gains its exposure through futures contracts. In many jurisdictions, futures are taxed on a different basis from directly held assets, often marked to market or given blended treatment, so the tax character of your gain is set by the futures wrapper, not by “carbon” in the abstract.
- An ETN like GRN is structured as bank debt. A note can generate a different character of gain again, because you legally hold a debt instrument that pays the index return, not a position in the underlying.
- Spot allowances held directly, which most public investors will not touch, are different again, and their treatment varies widely by jurisdiction and by whether you are a compliance entity or a financial holder.
The practical takeaway: do not assume two “carbon” products are taxed alike just because they track similar prices. A fund, a note, and a spot allowance are three distinct instruments with three distinct tax profiles. Check the specific instrument’s wrapper in your jurisdiction before allocating, because the after-tax gap between them can be larger than the fee gap.
X. Case studies
Case 1: Microsoft and Occidental/1PointFive, the demand signal (9 July 2024)
On 9 July 2024, Microsoft signed what was reported as the largest single carbon-removal purchase ever: 500,000 tonnes of direct-air-capture credits over six years, sourced from the STRATOS plant in Texas run by Occidental’s subsidiary 1PointFive. The press valued the deal in the “hundreds of millions of dollars.” The lesson for an investor is not the specific plant. It is the signal. A hyperscaler is willing to write a nine-figure cheque for high-permanence removal credits with a verifiable, engineered tonne behind them. This is the credible end of the voluntary market: expensive, physically real, and in demand from exactly the buyers with the deepest pockets. It is also, for now, tiny relative to the compliance market.
Case 2: the Verra “phantom credits” scandal (18 January 2023, cautionary)
On 18 January 2023, a joint investigation by the Guardian, SourceMaterial and Die Zeit, backed by academic studies, found that more than 90 per cent of Verra’s rainforest credits (avoided-deforestation offsets) were likely “worthless”, with the deforestation threat these projects claimed to prevent overstated by roughly 400 per cent on average. The buyers exposed read like a corporate honour roll: Gucci, Shell, easyJet, Salesforce and BHP. Voluntary offset prices and, more importantly, trust cratered afterward. The lesson: when the underlying tonne is estimated rather than measured, and “how much deforestation would have happened without this project?” is a counterfactual rather than a fact, the whole asset rests on an assumption that can be shown to be inflated. And when it is, the credits do not just fall in price. They are exposed as never having represented what they claimed.
Case 3: South Pole / Kariba REDD+, Zimbabwe (2023, cautionary)
Kariba was one of the largest forest-carbon projects on Earth: roughly 36 million credits issued since 2011, with more than €100 million in credit sales. Then, in a Bloomberg investigation published 24 March 2023 and a later Verra review, it emerged that more than half of around 27 million credits did not correspond to real emissions reductions. Project developer South Pole terminated its contract on 27 October 2023 and asked Verra to cancel 2.5 million credits. The lesson is the sharpest one here: a project-based offset can be retroactively voided. The “asset” you bought and are holding can be reviewed, downgraded, and cancelled after the fact. There is no equivalent risk on a compliance allowance; a government does not retroactively decide your EUA never existed. The permanence of the asset is completely different between the two halves of the market, and Kariba is the proof.
XI. The core constraint
The constraint that governs carbon is blunt. The tonne has to be real, and only one half of the market can guarantee that.
A compliance allowance is real by construction. A government created it, capped its supply, and legally requires companies to surrender it. Its scarcity is legislated, its existence is registry-verified, and it cannot be retroactively found to have never existed. That is why the compliance side is the investable core, not because it is virtuous, but because the unit is genuinely scarce and genuinely enforceable.
A voluntary offset is real only if the counterfactual behind it is real, and counterfactuals are estimates. “This forest would have been cleared, so protecting it counts as a saved tonne” is a claim about a future that did not happen. When that claim is inflated by 400 per cent, or when half the credits from a flagship project turn out not to correspond to real reductions, the constraint bites. The tonne was not real, so the asset was not real.
This is the whole game. Compliance allowances pass the reality test by design. Voluntary offsets pass it only if verification is rigorous, and the market’s recent history is a list of times it was not. Almost everything else about the asset follows from that single constraint.
XII. Inside the asset
Look closely at what you actually own in each half, because they are different objects wearing the same word.
A compliance allowance is a regulated permit. It is fungible: one EUA is identical to another. It is registry-tracked. It has a deep, liquid futures market behind it, which is why funds can hold it at scale. Its value is a direct function of one variable you can watch: the gap between the cap and actual emissions, plus expectations about how the cap will move. It behaves like a commodity with a policy engine bolted on. When you buy KRBN or KCCA, you are, one layer down, holding futures on these permits, which is why the roll and the fee sit between you and the raw allowance price.
A voluntary credit is a certified claim about a specific project. It is not fungible in the way an allowance is. A credit from a well-run direct air capture plant and a credit from a questionable avoided-deforestation project are both “one tonne” on paper but worlds apart in reality, which is exactly why the market now pays a 381 per cent premium for removals and a 217 per cent premium for recent vintages. The value of a voluntary credit depends on the specific project behind it, the standard that certified it, and whether anyone has re-examined the counterfactual lately. It is closer to a private, heterogeneous, reputationally exposed instrument than to a commodity.
That is how one word ends up describing both an asset you can hold in a liquid ETF and an asset that can be cancelled after you buy it. The label is shared; the two objects underneath it are not.
XIII. The central dilemma
The dilemma sits precisely where the money and the meaning diverge.
The part of the market with the return, the liquidity, and the scale is the compliance allowance, and it is a bet on nothing more noble than governments staying committed to making pollution expensive. It does not fund a forest. It does not remove a tonne. It is a legislated-scarcity trade, and its entire value rests on political will that can, and did in 2022, wobble hard on a single headline.
The part of the market with the meaning is the voluntary credit that actually removes or protects carbon, and it is small, illiquid, heterogeneous, and has twice been shown to be riddled with tonnes that were never real. The credible frontier of it, engineered removals like Microsoft’s direct air capture deal, is real but priced at a 381 per cent premium and still a rounding error next to the ~$948bn compliance market.
So the dilemma is: the investable half is a policy bet with no direct climate impact, and the impactful half is barely investable and historically unreliable. You largely cannot buy both the scale and the meaning in one instrument. An honest allocator picks the scale, prices in the policy risk, and does not pretend it is saving the planet. Or the allocator picks the meaning, accepts the illiquidity and verification risk, and does not pretend it is a liquid asset class. Confusing the two is how people lose money and credibility at the same time.
XIV. The next frontier
The frontier is engineered removals, and the market is already pricing them as the future.
The single clearest signal is the 381 per cent premium removal credits commanded over reduction credits in 2024, up sharply from 245 per cent the year before. Buyers are voting with a widening margin for permanence, for a tonne that is physically gone rather than a tonne that was theoretically prevented. Direct air capture sits at the top of that hierarchy: an engineered, measurable, hard-to-fake tonne. Microsoft’s 500,000-tonne, six-year, hundreds-of-millions-of-dollars deal with Occidental’s 1PointFive is the flagship, and the fact that the biggest buyer wrote the biggest cheque for the most engineered tonne tells you where corporate demand is heading.
But keep the scale honest. This frontier is credible precisely because it solves the reality problem that sank the Verra and Kariba markets: the tonne is engineered and measured, not estimated. And it is expensive and small for exactly the same reason. There is, as yet, no liquid public vehicle that gives a retail investor clean exposure to the removals frontier the way KRBN gives exposure to allowances. For now, the frontier is a corporate-procurement story and a private-market opportunity, not something you buy in a brokerage account. The removal premium is the number to watch. If it keeps widening while volumes grow, the frontier is arriving in earnest; if it plateaus, the market has not yet made up its mind.
XV. Lessons from history
The record above teaches a few things that hold up outside carbon too.
Legislated scarcity is durable; estimated scarcity is not. The compliance market has grown to ~$949bn on the back of caps that governments keep tightening, with coverage climbing from 7 per cent to 24 per cent of global emissions. The voluntary market, built on estimated counterfactuals, produced Verra and Kariba. When scarcity is manufactured by law it holds; when it is inferred from a hypothetical, it can be dismantled by a single investigation.
Policy assets carry policy tail risk, and it is fast. The ~40 per cent KRBN drawdown in days in February 2022 was not a fundamentals event. It was a regulatory signal. Any asset whose value is set by a regulator can be repriced at the speed of a policy announcement, which is faster than any earnings cycle. The diversification benefit, that low correlation to traditional assets, and the tail risk are the same coin.
Quality wins the war of attrition. The voluntary market did not collapse after the scandals. It flushed. Volume fell 25 per cent while the surviving credits got dearer, with removals and recent vintages commanding 381 per cent and 217 per cent premiums. Markets built on trust punish the junk and reward the verifiable. That is a slow, healthy repricing, and it is the mechanism by which Bayon’s “markets will boom when trust improves” could actually come true.
XVI. The case for it
The bull case for carbon as an asset class is narrow, specific, and mostly about the compliance half.
The structural driver is legislated and directional. Coverage of global emissions by direct carbon pricing has climbed to 24 per cent, government revenue hit a record $104bn, and the number of pricing instruments has reached 75. Governments have every fiscal and climate incentive to keep tightening. BloombergNEF’s €149-by-2030 forecast is the numerical expression of that tightening, and the worked example shows what it would mean: roughly +129 per cent gross off the €65 base if it lands.
The diversification case is genuine. Carbon’s low correlation with equities, bonds, commodities, real estate, gold and oil over a decade means it is a return stream driven by a different engine than most of a portfolio: policy rather than the business cycle.
Access is easy and cheap by alternative-asset standards. You can buy global exposure through KRBN at 0.79 per cent, a single market through KCCA at 0.91 per cent, or EU allowances outside the US through WisdomTree’s CARB at 0.35 per cent, for the price of one share, in an ordinary brokerage account. And the demand under the voluntary side is not dead. 182 million tonnes of retirements in 2024 held roughly flat through the worst of the trust crisis. The floor under real corporate demand did not fall out.
XVII. The Risks
Most of the risks here come back to one thing: the asset is only ever as solid as the policy or the verification standing behind it.
Policy reversal is the headline risk. The ~40 per cent KRBN drop in days on a single EU supply signal is the whole risk in one data point. The bull case and the risk case are the same case, because both rest entirely on governments staying committed, and governments respond to energy prices, elections, and recessions. A policy retreat is the fastest way this asset falls.
Structural cost drag is permanent. The futures-based ETFs carry roll cost on top of the ~0.8 to 0.9 per cent fee. Over long holds, the gap between spot allowance returns and what your fund delivers compounds against you.
ETN counterparty risk is a hidden layer. GRN is unsecured Barclays debt. You are exposed to a bank’s solvency on top of carbon’s price, a risk that has nothing to do with carbon and everything to do with the wrapper.
Verification risk defines the voluntary half. Verra’s 90-per-cent-worthless finding and Kariba’s retroactive cancellations show that a voluntary credit can be exposed as never having been real, then voided after you own it. There is no such thing as a settled purchase in a market where the underlying tonne can be re-examined.
Concentration risk is baked in. With the EU ETS at 87 per cent of the market, any “global” carbon position is overwhelmingly a bet on European climate policy specifically. You are less diversified across jurisdictions than the word “global” suggests.
XVIII. The Alternative Fortune Verdict
Carbon credits are not one asset class but two, and pretending otherwise is how people get hurt.
The compliance side (EU ETS, California, the allowances behind KRBN, KCCA and KEUA) is a real, liquid, ~$950bn market with a legislated-scarcity engine driving toward a €149-by-2030 forecast. That is the investable core. It is also a policy-beta position that can lose 40 per cent in days on a regulatory headline. Treat it as policy exposure, not a bond.
The voluntary side is where money went to die twice, at Verra and Kariba, because the underlying tonne was never verifiably real. Its credible frontier, engineered removals like Microsoft’s DAC deal, is genuine but priced at a 381 per cent premium and still tiny.
Where the edge actually is: the edge is not in stock-picking a carbon company. It is in correctly separating the two markets, allocating (if at all) to the regulated allowance side through a cap-and-trade vehicle, pricing in the ~0.8 to 0.9 per cent fee plus futures roll, and sizing it as the volatile policy-beta position it is rather than a stable income asset. The mispricing most people carry is treating “carbon” as one thing; the edge is in knowing it is two.
Questions to ask, by vehicle:
- A global carbon ETF (e.g. KRBN): How much of this is really the EU ETS given it is 87 per cent of the market? What has the futures roll cost me on top of the 0.79 per cent fee over the last few years? Am I comfortable holding a position that can drop 40 per cent on a policy signal?
- A single-market ETF (e.g. KCCA, KEUA): Do I actually want concentrated exposure to one jurisdiction’s climate politics, and do I understand that jurisdiction’s cap trajectory well enough to hold through a policy scare?
- A carbon ETN (e.g. GRN): Am I being paid enough to take Barclays’ credit risk on top of carbon-price risk, versus an ETF that holds futures directly?
- Any voluntary-credit exposure: Is the tonne engineered and measured, or estimated from a counterfactual, and would this credit survive the kind of investigation that voided half of Kariba?
The compliance market is a legitimate, liquid way to take a directional view on climate policy, with a genuine diversification benefit and a violent policy tail. The voluntary market is mostly not investable for public-market participants and mostly should not be, until verification catches up with the rising quality premium. Knowing which of the two halves you are actually buying is most of what separates a sound position here from a costly one.
For more on how carbon fits alongside other real-resource plays, see the Alternative Fortune commodities and natural resources hub.