Committed Emissions: The Debt Fixed at Financing
An oath is cheap to swear and dear to keep, for the swearing is a moment and the keeping is all the years after. A thing built to last is a promise made to time — that it will go on doing what it was built to do, long after the builder has forgotten he made the promise, and long after he would wish to break it.
The Emissions Are Committed at Financing
The stranded-asset debate is almost always told in the future tense: will a carbon price, a mandate, a technology shift one day render this fossil asset uneconomic before its time? That framing hides the decisive fact. The emissions — and therefore the stranding exposure — are not created by some future policy. They are committed at the moment the asset is financed.
“Committed emissions” is the future CO₂ that an asset’s mere existence implies: build a coal plant, a blast furnace, a gas boiler, and you have signed up for decades of emissions that will follow as surely as the asset operates. The quantity is fixed at the final investment decision — capacity, lifetime, expected utilisation and fuel are all known then. Everything after is just the slow delivery of a bill already incurred. Which means the interesting question is not the usual one. It is: at the point of financing, is this commitment even compatible with the carbon budget it will be judged against — and if not, which way does it resolve?
Stranding is not a surprise that happens to a fossil asset. It is the resolution of a commitment that was made, and was visible, the day the asset was financed.
Read that way, the analysis moves from forecasting a future shock to reading a present fact. The commitment is on the balance sheet from day one; the market simply chooses whether to price it. The mispricing is the gap between the asset’s modelled economic life and the emissions budget that life cannot fit inside.
The Problem Is Largely Already Built
Start with the arithmetic that makes this urgent rather than academic. If the world’s existing fossil-fuel infrastructure is simply operated as it historically has been, it will emit on the order of 660 gigatonnes of CO₂ over its remaining life — and the coal fleet alone accounts for roughly 330 Gt, more than every coal plant that has ever operated has emitted in all of history combined. Set that against the remaining carbon budget: the allowance left to hold warming to 1.5°C is now about 170 Gt — roughly four years of current emissions.
The implication is the one the reform debate resists: the climate problem is, to a first approximation, already built. Existing infrastructure alone commits several times the remaining 1.5°C budget; add the plants proposed, planned and under construction and the total exceeds it further and eats deep into the 2°C budget too. No new fossil asset is required to blow past the target — the standing stock does it. This is why the serious question shifts from “stop building” to “what happens to what is already there,” and that is a question about money as much as molecules.
Every Long-Lived Asset Faces a Fork
Put the commitment and the budget together and every long-lived fossil asset confronts the same unavoidable fork. Either it runs for its full life — delivering its committed emissions and, in aggregate, blowing the budget — or it is retired before its time, which is the definition of a stranded asset. There is no third state in which the emissions are both delivered and the budget is held. The arithmetic forces the choice; it only leaves open which side of the fork each asset lands on, and who absorbs the cost.
| Asset | Typical operating life | What its existence commits | How the fork resolves |
|---|---|---|---|
| Coal power plant | ~40 years | Decades of high-utilisation emissions | Run to term (budget) or retire early (strand unrecovered capital) |
| Gas power plant | ~30–40 years | Lower per-MWh, still decades of commitment | The same fork, with softer edges and more optionality |
| Blast furnace / cement kiln | ~40 years (relined mid-life) | Industrial lock-in — “2050 is one investment cycle away” | Retrofit (CCUS, H₂, DRI), curtail, or strand |
| ICE vehicle fleet | ~15–20 years | Tailpipe emissions across the turnover cycle | Slow turnover locks in; fast turnover strands residual value |
| Gas boiler / building heat | ~15–25 years | A fresh ~20-year commitment with every install | Each new unit deepens the lock-in one household at a time |
Two escape routes soften the fork without eliminating it. An asset can be run less — lower utilisation delivers fewer of the committed emissions, a gradual partial unwinding rather than a clean break — or it can be abated in place through capture or fuel-switching, paying capital to keep the asset while shedding the emissions. Both change the shape of the commitment. Neither makes it disappear, and both cost money that the original financing did not price.
Stranding Is a Commitment Coming Due
This is where the mispricing lives. A fossil asset is typically valued and financed on its full economic life — thirty or forty years of modelled cash flow — as though the carbon budget and the commitment did not exist. But the budget arithmetic says a large share of that fleet cannot run to term. Studies put it starkly: only around 42–49% of existing and pipeline power generators can be operated to the end of their economic life under a 2°C path. The rest must be curtailed or retired early. The valuation assumes a full life the budget has already ruled out for half the fleet.
Estimates of global stranded coal-power assets run from roughly $150 billion to $1.4 trillion depending on policy stringency and timing, and for individual listed owners the exposure can reach up to ~78% of share price or more than 80% of equity. The crucial point is not the size of the number but its origin: this loss was not created by the policy that eventually triggers it. It was written into the asset at the final investment decision, the moment its committed emissions were set against a budget they could never fit. The policy is only the alarm clock. Anyone underwriting a long-lived fossil asset on full-life cash flows is buying a commitment the budget has already discounted — and calling the discount a surprise when it arrives.
The Exposure Sits Where the Fleet Is Young
If the loss is fixed at financing and delivered at retirement, then the exposure is largest wherever the most committed life still lies ahead — and that is a question of fleet age. Here the world splits cleanly in two. The coal fleets of the United States and Europe average more than 40 years old: their committed emissions are mostly behind them, and retiring a nearly-depreciated plant strands little. The fleets of developing Asia average about 13 years, with over half built in the past two decades — decades of committed emissions, and unrecovered capital, still ahead.
That age gap is the whole exposure. Retiring Indonesia’s fleet is estimated at over $114bn, Vietnam’s at over $57bn, and the combined China-and-India coal transition at $3–5 trillion — because you are writing off young plants with decades of contracted life left, often shielded by long-term power-purchase agreements that guarantee their revenue. The financial exposure follows the same logic as the debt tenor: wherever the capital’s horizon outlasts the asset’s viable life, the financier holds the commitment when it comes due. This is precisely where the Western stranding debate goes wrong for the emerging world — it assumes the emissions are already sunk, which is true of a 40-year-old fleet and false of a 13-year-old one.
Positioning: Price the Commitment at Signing
The discipline follows directly: treat committed emissions as a liability fixed at the final investment decision, and price it there — not when the policy arrives. That single move re-sorts the opportunity set.
Underwrite the carbon-liability-adjusted life, not the full economic life — and finance the unwinding of commitments already made.
Three places to stand. First, the unwinding market: transition and early-retirement finance, securitisation and replacement structures that pay to resolve a commitment ahead of its policy-forced date — a $3–5 trillion need concentrated in young-fleet Asia. Second, abate-in-place: capture, co-firing and fuel-switching that keep the asset while shedding the committed emissions, the only path that avoids both blowing the budget and stranding the capital. Third, avoid the tenor mismatch: refuse the long-dated fossil exposure whose viable life is shorter than the capital committed to it, and discount every full-life fossil valuation by the fraction of that life the budget has already ruled out. Price the commitment at signing, and stranding stops being a surprise.
Reading It Through the Frameworks
Where does policy become the cash flow? Committed emissions are the cleanest case in the whole Strain thread, because the policy does not create the loss — it merely triggers a loss that was fixed at financing. The carbon price, the retirement mandate, the phase-out date are alarm clocks on a debt already owed. So the analytical task is not to forecast the policy but to read the commitment on the balance sheet today and ask whether the asset’s modelled life can survive contact with the budget. Where it cannot, the cash-flow impairment is already present; it is only unrecognised.
Structural moat or temporary bottleneck? Neither — it is a one-way ratchet. Every new long-lived fossil FID deepens the lock-in and enlarges the eventual stranding; every early retirement or in-place abatement resolves a piece of it. The discipline is to separate the asset whose commitment can be gracefully unwound (short remaining life, abatement-ready, no PPA lock) from the one whose commitment is rigid and long (young, contracted, un-abatable), and to treat the near-term FIDs being signed now — not the distant 2050 target — as the decisions that actually set the mid-century outcome.
Committed emissions collapse the future tense the stranding debate hides behind. A long-lived fossil asset fixes its emissions — and its stranding exposure — the day it is financed, not the day a policy arrives, and the world’s existing infrastructure already commits several times the remaining 1.5°C budget. That leaves every such asset on a single fork: run for its full life and blow the budget, or retire before its time and strand. The valuation that assumes a full economic life is assuming away a budget that has already ruled out roughly half the fleet.
Stranding is a commitment coming due, not a surprise. Price it at the investment decision, not at the policy alarm: underwrite the carbon-liability-adjusted life, finance the unwinding of commitments already made, back abate-in-place where the asset can shed its emissions, and refuse the long fossil tenor whose horizon outlasts its viable life. And read fleet age as the map of exposure — the West’s emissions are largely behind it; Asia’s young fleet carries decades of them ahead, at a cost measured in trillions. The reckoning was fixed the day the foundation was laid; the years between are only its slow arrival.
They thought the reckoning lay far ahead, in some year not yet come; but the reckoning was fixed the day the foundation was laid, and all the years between were only the slow arrival of a debt already owed.
Leave a Reply