Climate & Energy · Monday, 10 August 2026

01 · Briefing · what happened

Kenya fences in its carbon credits as the offset market's honesty problem resurfaces

Climate & Energy 3 min 13 sources

Kenya capped how many carbon credits it will sell abroad, one of Africa's most detailed rulebooks yet - just as fresh research shows most forest offsets never cut the carbon they claim.

10Mt

Kenya's export cap

carbon credits authorized for foreign sale through 2030

1.67Mt

allowed per year

annual limit on Kenya's international sales

~11 to 1

forest over-crediting

credits issued per genuine ton of cuts, one study found

$50bn

COP31 finance goal

annual money for developing countries via credits

At a glance

Kenya capped the carbon credits it will sell abroad at 10 million tons of CO2 through 2030.
It fears overselling cuts it will later need to meet its own Paris climate target.
A credit is a claim: one ton kept out of the sky, sold to someone offsetting a ton they emit.
New research finds most forest offsets never cut the carbon they were paid for.
A credit only counts if the cut would NOT have happened anyway - and no one can prove that.
Turkey, hosting COP31, is pushing to grow the market and unlock $50bn a year for poorer nations.

Forces in play

Market trust deficit High→

research keeps finding most forest offsets never cut the carbon claimed

Rulebook tightening Building↑

Kenya's cap and permanence rules push credits toward real cuts

Demand push Building↑

COP31 host Turkey wants to grow the market and its $50bn finance flow

Double-counting risk Easing↓

Kenya's cap directly stops selling a ton abroad and claiming it at home

In play Kenya — capped its credit exports at 10Mt to protect its own climate target Turkey (COP31 host) — pushing governments to grow carbon-credit demand Offset researchers — found fewer than a fifth of credits represent real cuts Foreign buyers — want cheap credits to offset their own emissions

How it unfolded

This decade carbon credit market booms, integrity scandals pile up
2026 studies confirm heavy over-crediting in forest offsets
Aug 4 Kenya publishes its rulebook and 10Mt export cap
Next COP31 pushes to grow demand while proving credits are real

Where this points

Watch whether COP31 ties its demand push to hard additionality tests - grow the market without them and the honesty problem scales with it.

Full briefing

Kenya has drawn a hard line around its carbon credits. On Monday it published the “Kenya Guide for Strategic Engagement in Carbon Markets 2026” [1]. The rulebook caps the credits it will authorize for foreign buyers at 10 million tons of carbon dioxide through 2030, with annual sales held near 1.67 million tons [2]. It is one of the most detailed rulebooks any African country has written for Article 6 of the Paris Agreement [1]. That clause lets one country pay for a cut somewhere else and count it as its own [3].

The reasoning is blunt. Kenya’s officials say the cap stops the country overselling credits it may later need to hit its own national climate target [1]. A carbon credit is a claim: one ton of carbon kept out of the sky, sold to someone who wants to offset a ton they emitted. If Kenya sells that ton abroad and also counts it at home, the same cut is counted twice - and the atmosphere sees no benefit [2]. Kenya’s new rules also steer credits toward engineering projects - things like clean cookstoves and waste plants - over harder-to-verify land schemes [3].

The move lands at a moment when the whole market is trying to prove it is worth trusting. Turkey, which hosts the COP31 UN climate summit, said this week that carbon markets should take their “rightful place” [4]. It is pushing governments to boost demand and unlock more than $50 billion a year in finance for developing countries [4]. Lawyers, meanwhile, are wrestling with “permanence”: a forest paid to stay standing can still burn down a decade later, reversing the cut a buyer already claimed [5].

Underneath all of it sits a harder problem the rulebooks cannot fully solve. A credit only cuts emissions if the clean outcome would not have happened anyway - what the trade calls additionality. Pay a landowner to protect a forest that faced no axe, and you have spent money and cut nothing. The evidence here is grim. A study in Science this year found tropical forest offset projects deliver only partial real gains amid “persistent over-crediting” [6]. A separate analysis of forest projects found over-crediting of roughly eleven to one [7]. For every genuine ton, about a dozen credits were issued with no real cut behind them. Wider reviews covering nearly a billion tons of credits have found fewer than a fifth represent real reductions.

Kenya’s cap does not fix that. It fixes double-counting, not the deeper question of whether a credit was ever additional. But a country writing clear, conservative rules is a signal the market is maturing past its wild-west decade [3][5].

Also moving this week

In the United States, the Trump administration is blocking billions of dollars in grants meant to modernize the aging power grid, holding up projects already approved [8]. Meanwhile, new solar tariffs raise the cost of panels for developers [9]. Both push against a clean build-out that is otherwise accelerating. The world has now installed a third terawatt of solar power [11], and in a first, Utah drew more electricity from solar than from any other source [12].

The heat is the backdrop to all of it. A brutal heatwave broke temperature records across central and eastern Europe, with Italy on red alert and Hungary and Romania dimming public lights to save power [10]. Scientists reported the world’s seas hit their hottest July temperature on record [13]. The transition is racing and stalling at once - and the carbon-credit fight is about who gets to claim credit for the parts that move.

02 · Lesson · why it matters

The world you paid to prevent, and why you can never see it

A carbon credit only cuts carbon if the cut needed your money - and the world where you didn't pay never happened.

How it works

A credit claims one ton of carbon was cut
But it only counts if the cut needed the money
The 'no-money' world never happened, so you can't observe it
So the whole claim rests on a guess about a counterfactual
Guess generously and you sell cuts that were never real

The twist

A carbon credit's value hinges entirely on a world that didn't occur - the one where you didn't pay - and you can never actually see it.

Where you've seen this

Job training grants

did the subsidy create the hire, or fund one the firm would have made anyway

Medical trials

the drug only works if patients would have stayed sick without it - hence a control group

Ad spending

half your ads are wasted on buyers who'd have bought regardless - you just can't tell which half

The catch

The counterfactual is unknowable, but not arbitrary: strict, conservative baselines and control groups shrink the guessing - they never erase it.

Full lesson

The claim inside every credit

When Kenya caps how many carbon credits it will sell, it is putting a fence around a strange kind of thing. A carbon credit is not a ton of carbon. It is a claim about a ton of carbon: that somewhere, because someone paid, a ton stayed out of the sky that otherwise would have gone up.

Read that again. “That otherwise would have gone up.” The whole value of the credit hangs on those five words - on a comparison with a world that did not occur. The forest that was saved is only worth paying for if it was actually going to be cut down. The methane that was captured only counts if it was actually going to leak.

This is the question the trade calls additionality, and it is the hardest question in climate policy: would this have happened anyway?

You can’t observe the road not taken

Here is what makes it so slippery. You can see the forest still standing. You can measure the methane not leaking. What you can never see is the other world - the one where you kept your money and did nothing.

That world is not hidden. It is simply gone. It never ran. So the emission cut you are buying is measured against a guess: an estimate of what would have happened without you. Guess that the forest faced the axe, and your credit is worth a lot. Guess that it was safe all along, and you paid for nothing.

And the two worlds look identical from the outside. A landowner paid to protect a forest that was never threatened, and a landowner paid to protect one that was, both hand you the same thing: a standing forest. The receipt is the same. Only the counterfactual differs, and the counterfactual is invisible.

Why the guess drifts generous

Now watch what the incentives do to that guess.

The person selling the credit wants the baseline - the picture of the dangerous world you avoided - to look as scary as possible. A forest “under grave threat” generates more credits than a forest that was probably fine. The buyer, quietly, wants the same thing: a company buying its way to “carbon neutral” wants cheap credits, and cheap credits come from generous baselines. Even the auditors are paid by the projects they check.

So almost everyone at the table has a reason to imagine the darkest possible version of the world that never happened. And the darker they imagine it, the more credits get printed for cuts that were never real.

This is exactly what the research keeps finding. A study this year found forest offset projects issued roughly eleven credits for every one that reflected a genuine cut. Broader reviews of nearly a billion tons of credits found fewer than a fifth stood for real reductions. The market did not lie, exactly. It just kept guessing generously about a world it could not see.

The pattern is everywhere money chases an outcome

Once you notice this, you find it under every policy that pays for a result.

A government offers a subsidy for solar panels. Some of that money goes to people who were going to install panels anyway - the subsidy changed nothing but the size of their wallet. A job-training grant is only worth its cost if the trainee would not have found work otherwise. A tax break for a factory only “created” jobs the company would not have created regardless.

In every case the spending is real, but the effect is measured against an invisible baseline: what would have happened without it. This is why serious drug trials use a control group. You cannot know a pill works by watching one patient recover; you have to compare against people who did not take it. Climate policy rarely gets a control group. It gets a guess.

Who is holding the guess

You are closer to this than it looks. If you have ever booked a “carbon-neutral” flight, bought a product that offsets its footprint, or paid a tax that funds a green subsidy, you have already done this. You handed money to a claim about a world you cannot inspect. You are trusting that someone, somewhere, guessed honestly about a road not taken.

Kenya’s rulebook does not answer the deep question - nothing can fully answer it. What a cap and a conservative baseline can do is shrink the room for generous guessing, and make the invisible comparison a little less flattering to the seller. That is real progress. But it is the progress of narrowing an unknowable, not removing it.

The humbling part is that this sits under far more of the world than carbon. Every time we pay to change an outcome, we are really betting on a world that didn’t happen. The ledger that would tell us if we were right is one none of us will ever get to read.

03 · Lab · your turn

The Counterfactual Test

Rehearse buying carbon offsets when you can't see whether each cut would have happened anyway - and feel how the invisible comparison bends the whole market.

04 · Hope · carry this

The market's honesty problem is finally being named out loud - by researchers, by lawyers, and by a Kenyan rulebook fencing in its own credits. Naming a flaw plainly is how the fixing starts.