Is climate action possible under capitalism at all?

On the map on this site, Germany’s emissions appear as dots. Zoom in and you see at once what statistics conceal: they are not spread evenly across the country, they cluster in a few places. A single power station can emit more than a mid-sized city with all its households, all its cars and all its heating put together.

Meanwhile the public debate about climate action is largely about shower times, plastic bags and holiday flights. Not that these are nothing. But there is an odd mismatch: the largest objects on the map barely feature in it.

This text is about that mismatch, and about why it persists so stubbornly.

Terms

Which growth?

Before one can argue about whether capitalism needs growth, it has to be clear what is being talked about. In everyday use, three very different quantities collapse into one.

Gross domestic product measures value added in money. It rises when more steel is cast, but equally when more software is licensed, more care is given, more advice is sold. A repair raises it in much the same way as a new purchase, although the two move very different amounts of material.

Throughput measures precisely that: how much energy and material flows through the economy. Tonnes of steel, cubic metres of gas, kilowatt-hours of electricity. This is the quantity the ecological problem hangs on, because the atmosphere responds to tonnes of greenhouse gas, not to euros. Whether a tonne of CO₂ cost three euros or three hundred when it was produced changes nothing about its effect: it stays up there just as long and warms just as much.

Return on capital is the expectation that money invested will earn interest. Anyone financing a factory, extending a grid or building housing generally does so with borrowed money and with the prospect of getting back more than was put in.

The stability of the system hangs on this third quantity. If the expectation goes permanently unmet, it is not that nothing happens: lending becomes more cautious, investments are postponed, positions go unfilled. And the social security systems come under pressure along with them, because they are calculated on rising wages and contributions.

To this comes a fourth term, and it decides the whole debate: decoupling. It means that economic output rises without throughput rising with it. The term comes from environmental economics and has carried, for decades, the promise that prosperity and resource use can be separated. It is therefore the bridge between the first and the second quantity, and everything that follows in this text depends on whether that bridge holds.

Two degrees of it need to be distinguished. Relative decoupling means throughput is still growing, but more slowly than the economy. That is not enough, because what accumulates in the atmosphere is absolute quantities, not ratios. Absolute decoupling means the economy grows and throughput falls. Only the second helps, and even then only if it happens fast enough.

Distinguishing these quantities is not academic, it decides the debate. Someone saying “growth destroys the climate” means throughput but says GDP. Someone answering “but we decoupled long ago” means GDP and says nothing about throughput. The two talk past each other, and each is right about their own quantity.

Germany is currently providing a case study. After two years of recession the economy grew by 0.2 percent in real terms in 20259, which is effectively standstill. Emissions fell by 0.1 percent in the same year12. Three years without meaningful growth, in other words, brought no meaningful relief. Anyone claiming a simple coupling between GDP and emissions is refuted by this.

I regard this as the most important finding in this text, and it cuts against part of the critique of growth as well: shrinking GDP is not a climate strategy. Falling throughput would be one, but the first does not compel the second. A stagnating economy does not automatically reduce throughput, because the building stock, the vehicle fleet and the infrastructure keep running. It does immediately produce distributional conflict, and the political response to that in recent years has been less climate policy, not more.

So what remains of the premise? Not the strong claim that capitalism cannot function without growth at all. Japan has shown for thirty years that a market economy can persist close to zero. And the economists Tim Jackson and Peter Victor used their LowGrow SFC model to run the Canadian economy through half a century to 2067; in their third scenario ecological and social indicators improve even when the growth rate approaches zero17. That is a simulation for one country, not a proof, but it shows the claim that it simply cannot work is unfounded. In the research literature, too, the concept of a growth imperative is explicitly contested: whether it exists, for whom it holds and which mechanism produces it are answered differently across disciplines; the older monetary argument is considered weak, while arguments based on resource access and capital accumulation have proved more robust16.

What remains is the weaker and more defensible statement: growth is not a law of nature for this system, but it is its condition of stability. And because that is so, every measure that would reduce throughput fast enough is measured against precisely that condition and regularly declared unreasonable. It is not physics that prevents the necessary path, but the question of who carries the burden of adjustment.

This allows the question of this text to be put more precisely than in its headline. Not: is climate action possible under capitalism? But: does the decoupling this would require come about fast enough under the current distribution of power? The next two sections examine the first half of that question, the pace. After that comes the second half, the distribution of power.

Evidence

Is it fast enough?

Decoupling is happening, and in the stronger of the two forms distinguished above. Since 1990 German greenhouse gas emissions have fallen by 48 percent12 while economic output has grown. That is absolute decoupling, not an accounting trick. Anyone claiming that climate action and the economy are fundamentally incompatible is arguing against the data.

The question is therefore not whether, but how fast. And there the picture changes.

First: the window. The remaining global CO₂ budget for 1.5 degrees stands at roughly 170 billion tonnes according to the Global Carbon Budget 2025. At today’s output it would be used up before 203010. Global fossil emissions did not fall in 2025, they rose to a new record10. Anthropogenic warming currently stands at around 1.36 degrees11. “Too slow” is not a judgement here, it is a subtraction.

Second: the pace has slackened, not picked up. Germany emitted around 649 million tonnes of CO₂ equivalent in 2025, 0.1 percent less than the year before12. That is not a reduction, that is standstill.

For the coming years the Umweltbundesamt calculates how far the instruments already adopted will carry. Three pairs of figures from those projections, achievable against required13:

• 2030: 62.6 percent reduction achievable, 65 percent is the target.
• 2040: around 80 percent achievable, 88 percent is required by law.
• 2045, the target year for greenhouse gas neutrality: 212.5 million tonnes remain on the basis of instruments adopted today.

The most telling figure here is not the gap itself but its movement. A year earlier the buffer to the 2030 target still stood at around 81 million tonnes. It is now 3.8 million tonnes13.

Third: the reductions are coming from the wrong sources. Looking at the sectors is the real finding. In 2025 industrial emissions fell by 5.7 million tonnes, not because industry had transformed itself but because it was running below capacity12. In the same year emissions rose by 3.4 million tonnes in buildings and 2.1 million tonnes in transport12. The independent Council of Experts on Climate Change reviewed the figures and confirmed them: what industry and the energy sector saved was almost entirely used up again by buildings and transport14. Where a real transformation is under way, in electricity, it works. Where millions of individual decisions about heating systems and vehicles add up, it does not.

Fourth: part of the balance has emigrated. The German climate balance is a territorial one. It counts what is emitted within the country’s borders. What arises in supply chains abroad appears in the balance of the producing country. The Umweltbundesamt itself points out that a consumption-based balance, including imports and international travel, is not directly comparable with the national greenhouse gas inventory15. For an industrial nation with a high share of imports this means part of the German decline is not a disappearance but a shift on the map. How large that part is cannot be cleanly quantified with the publicly available figures, which is remarkable in itself.

Taken together: decoupling is real, but it is slower than the budget, concentrated in one sector, and partly outsourced. That is precisely what I mean when I say growth has nowhere yet come apart from resource use fast enough.

Scale

Beyond our means

That we live beyond our means is not a feeling but, for some years now, a measurement programme. The concept of planetary boundaries describes nine systems whose stability carries the conditions for life on Earth, and determines a safe operating space for each.

According to the Planetary Health Check 2025 of the Potsdam Institute for Climate Impact Research, seven of these nine boundaries have been crossed: climate change, the integrity of the biosphere, land-system change, the freshwater cycle, biogeochemical flows, the introduction of novel entities and, new in this report, ocean acidification21. Only ozone depletion and aerosol loading remain within the safe range21.

Those two exceptions are exactly why I am not writing this section as a diagnosis of doom. The ozone layer has largely recovered and global aerosol loading is declining21. In both cases because there was coordinated international action. There is evidence, then, that changing course is possible. What is notable is what the two success stories have in common: they concerned clearly identifiable substances and a manageable number of producers, not the foundation of the entire energy supply.

I deliberately use the planetary boundaries rather than the better-known Earth Overshoot Day. The Overshoot Day has been criticised in some detail for converting very different environmental pressures into a single figure of land area. The planetary boundaries keep the nine systems separate and are the more robust basis for it.

Power

Whoever decides is deciding about themselves

The resource question explains why things are getting tight. It does not explain why so little is happening. For that one has to look at who controls the levers.

The state knows what would have to be done and simultaneously finances the opposite. The Umweltbundesamt puts environmentally harmful subsidies in Germany at more than 65 billion euros; the figure is from 2018, and no more recent one exists1. Around 90 percent of it is harmful to the climate, close to half of it falls on transport and a further 39 percent on the supply and use of energy1. What is notable is less the size than the direction: between 2012 and 2018 transport subsidies alone rose from 28.6 to 30.8 billion euros, in parallel with the expansion of climate funding1. And the trend continues. A calculation by the Forum Ökologisch-Soziale Marktwirtschaft, commissioned by the environmental organisation Greenpeace, arrives at eleven billion euros of additional incentives for oil and gas consumption introduced for 20262.

This is not an oversight. It is the result of a balance of forces that can be quantified, now that a lobby register exists. According to an analysis by the organisation LobbyControl, among the hundred largest lobbying actors in Berlin, 84 representing business interests face seven organisations working in the public interest, and the business side spends more than seven times as much3. The analysis particularly highlights the high spending from the energy sector and the sharply increased outlay of the car industry3.

Whether this shows up in actual decisions is not a matter of speculation but of research. The political scientist Lea Elsässer, together with Svenja Hense and Armin Schäfer, demonstrated a systematic bias in the Bundestag in favour of upper occupational and income groups, on the basis of 746 policy questions on concrete reform proposals from 1980 to 2013, across all governing coalitions4. The mechanism they describe is a downward spiral: those whose concerns find no resonance disengage, whereupon politics orients itself still more closely to the better off5.

This data ends in 2013, and no comparable study exists for the years since. More recent work takes a different approach but points in the same direction: for German tax policy, a systematic misperception among members of parliament shows up. Around 80 percent of the population want higher incomes taxed more heavily; politicians estimate the figure at about 63 percent6. In fairness it belongs to the picture that the study taking up these figures reaches a milder verdict in its own findings: the perception of the preferences of lower income groups turned out to be less distorted there than previously assumed6.

Honesty requires saying that the strength of the effect is disputed in the research. A comparative analysis across several democracies concludes that the wealthy have only a slight edge7, and a counter-current disputes the extent of the distortion altogether8. The direction is not in dispute; the size is.

Anyone demanding decarbonisation today is demanding of specific actors that specific assets be written down. That they use their means to delay this is not a moral failing of individuals but rational behaviour under the prevailing incentives. That is exactly where the problem lies: it continues even though everyone involved knows the facts.

At this point an objection regularly arises, and it is a fair one: Bitterfeld was not a capitalist project. Nor was lignite-fired power in the GDR, nor the Aral Sea. Anyone attributing environmental destruction to capitalism has to explain why centrally planned alternatives sometimes had worse records.

I consider the objection correct, and a refinement of my argument rather than a refutation of it. What is destructive is not the price system but concentrated power without effective accountability. Under state socialism it lay with the party apparatus; under present-day capitalism it lies with private wealth that has political access. The mechanics are the same: those who decide how resources are used do not bear the consequences, and those who bear the consequences have no say.

What follows for me is not a grand question about the system but a smaller and more uncomfortable one: which decisions about energy, land and infrastructure are actually open to democratic negotiation today, and which count as pre-political because they have already been settled in property rights?

Discourse

Thinking with the doors closed

One could object: if the majority wants something else, it should simply prevail. This is where the most interesting finding lies: the majority exists, but it does not know about itself.

A study published in Nature Climate Change surveyed around 130,000 people in 125 countries. 69 percent would be willing to give one percent of their income for climate action, 86 percent support corresponding social norms, 89 percent demand more political action18. The real finding, though, is the gap behind these numbers: asked how many of their fellow citizens would be willing, respondents estimate 43 percent on average, 26 percentage points below the actual value. In all 125 countries the assumed level of support lies below the real one18. The authors call this pluralistic ignorance: a condition in which almost everyone wants something and almost everyone believes they are alone in it. The same effect has been demonstrated separately for the United States19.

Set this alongside the finding from the previous section and a double blind spot emerges: the population underestimates itself, and elected representatives underestimate the population6. Both sides take their decisions on the basis of a picture of the majority that is demonstrably wrong.

A caveat belongs here, and I would rather state it myself: agreement in a survey is not the same as behaviour at the ballot box or in the hardware store. One percent of income is a low threshold, and the gap between what people say and what they do is well documented in the research. What the figures show is not that the majority would act. It is that it believes itself to be smaller than it is.

That is the ground on which the debate about alternatives gets stuck. Anyone criticising the existing arrangement in Germany finds two ready-made drawers waiting: naivety or socialism. No sooner is the criticism voiced than it is assigned to a historical model nobody wants back, and is thereby settled without ever having been examined. The effect is not that alternatives are rejected. The effect is that they are never formulated in the first place. Someone who believes they are in a minority, and who knows which drawer their position will be filed in, does not think it through to the end.

It is easier to imagine the end of the world than the end of capitalism, runs a sentence that Mark Fisher, in Capitalist Realism, attributes to Fredric Jameson and Slavoj Žižek20. It is usually read as a diagnosis of hopelessness. I read it as a statement about imagination, and the figures above suggest the problem lies less in a lack of agreement than in a lack of knowledge about how widespread that agreement is.

Test

What would change my mind

A position that cannot be refuted is not a position. So here are four observations that would make me revise my assessment, all checkable against data published every year.

First: transport and buildings fall for three consecutive years without a recession being the reason. These are the two sectors where nothing has moved so far12. If they shift while the economy runs normally, then transformation is happening and not merely a business cycle.

Second: the projection gap is closed by instruments, not by adjusting the target. The Umweltbundesamt reports around 80 instead of the required 88 percent for 2040, and 212.5 million tonnes remaining for 204513. If that gap narrows in future projections because adopted measures are working, I was too pessimistic. If it narrows because the target was lowered, I was not.

Third: consumption-based emissions fall at least as fast as territorial ones. That would show that less is actually being consumed here rather than merely produced elsewhere15.

Fourth, and this is the real test: a decision of consequence goes against concentrated wealth interests and in favour of a majority. Cutting environmentally harmful subsidies in transport would be such a case1. At its core my argument is not a claim about growth but about power. It stands or falls on whether it turns out that the distribution of power does not determine the outcome.

What would not change my mind, I will say as well: a good year, a programme of announcements, or a recession that pushes emissions down. None of that is evidence of transformation.

Conclusion

What follows from this

None of this amounts, for me, to a call for a change of system. I have no worked-out alternative to offer, and anyone claiming to have one should give you pause.

What does follow is more modest and more checkable: we should stop talking about the wrong levers.

Germany’s emissions are distributed very unevenly. The map on this site records 67,253 emitters22, and a single power station can emit more than a mid-sized city. That is precisely what the atlas shows: not as an accusation, but as a map. Here are the things at issue, and here they are not.

Three sober consequences follow from that.

The debate about individual behaviour is not wrong, but it is wrongly scaled. Anyone who has once seen the comparison of scale between a person and an industrial plant starts asking different questions.

The levers are where decisions about infrastructure are made: power stations, heating systems, vehicle fleets, grids. These are investment decisions with lifetimes measured in decades, not consumption decisions.

And who takes those decisions is not a fact of nature but a question of jurisdiction, and therefore negotiable. That is the point at which a data map becomes a political question.

Negotiable, however, does not mean that anything moves by itself. It means there are places where the negotiating happens: supervisory boards of municipal utilities, land-use plans, wage rounds, procurement procedures, elections at every level. These places are unspectacular and mostly poorly attended, and that is exactly why they are within reach.

The difference between individual behaviour and political action then lies not in the effort but in the scale. The same hour that goes into one’s own consumption acts, in an association, a citizens’ initiative or local politics, on decisions that apply to thousands. How that sorts by effect is not in this opinion piece but next door in the atlas: three levels of action side by side, individual, collective and systemic, each with what they actually move. That is not a finished strategy. A direction, though, it is.

You do not have to share my assessment to make use of the atlas. The data are the same whether you consider me too pessimistic or too cautious. That matters more to me than agreement with this text.

What the atlas shows stands in any case: where the large quantities really are, and how far that is from what we usually argue about. If this text achieves nothing more than getting the next discussion about climate action to start with the orders of magnitude instead of with shower times, it has achieved enough.