The climate crisis, approached through a Marxist analysis, becomes a question about production rather than about weather. Physical findings belong to climate science, and this article credits them to the institutions and assessment literatures responsible for them, restating none of them and quantifying none of them. What a class analysis examines sits upstream of every physical finding: the social machinery that keeps fossil energy flowing through factories, freight networks, heating systems, and supply chains, decade after decade, even as the knowledge of consequences accumulates. Emissions, on this reading, are an output of a system organised around competitive accumulation, and understanding the crisis means understanding why that system generates the output so reliably, so impersonally, and with such resistance to correction from within.

Consider one concrete illustration before any theory. In the early nineteenth century, British cotton manufacturers faced a choice between water power and steam power. Water was abundant along the streams of the northern valleys, and in many districts it was cheaper than coal-fired steam, whose engines consumed fuel, demanded maintenance, and broke down with exasperating regularity. Manufacturers nevertheless migrated toward steam, and the migration founded the fossil economy. A study of that transition, examined in detail further on, argues that steam won because it could be installed in towns where labour was plentiful, recruitable, and disciplineable, while water chained the mill to remote valleys where workers had to be attracted, housed, and retained on terms the owner could not fully control. Fossil energy entered large-scale production as an answer to a labour problem. That single documented finding reframes the entire question of why fossil fuels prove so difficult to surrender.

Everything that follows builds outward from two anchors of that kind. One anchor is historical: technologies are adopted for reasons embedded in social relations, and the reasons persist inside the machinery long after the moment of adoption has passed. Another anchor is structural: a firm competing against rivals cannot unilaterally restrain its own material throughput without surrendering position, so the pressure toward expansion operates regardless of what any individual owner intends or believes. Around those anchors the analysis assembles four mechanisms, a counter-case stated at full strength, and a set of concessions that any honest version of the argument must carry. Readers who want the framework beneath the mechanisms will find it in metabolic rift explained, which this article treats as background rather than repeating.

The climate crisis as a problem of production

Begin with a distinction that popular discussion routinely blurs. One question asks what greenhouse gases do in the atmosphere; another asks why an economic system produces them in ever-renewed volumes. The first question belongs to physics, chemistry, and the assessment literature that synthesises those sciences. The second belongs to political economy. Confusing the two produces a characteristic muddle in which a social argument is expected to do the work of a physical one, or in which physical findings are treated as if they settled social questions by themselves. A Marxist analysis confines itself to the second question, and its claims stand or fall as claims about accumulation, technology choice, asset values, and state capacity, all of which are checkable against historical and economic evidence without adjudicating any scientific dispute.

Production, in this vocabulary, means far more than factory output. Goods must be transported, buildings heated and cooled, food grown and distributed, devices manufactured and replaced, and each stage draws on energy. Fossil fuels came to dominate that energy supply not through a single decision but through layered commitments: engine designs, transport networks, urban forms, agricultural methods, and financial expectations all grew around coal, then oil, then gas. Describing the outcome as a failure of awareness insults the evidence, since awareness has been widespread among governments and publics for a long span of years. Describing it as a failure of pricing, as though one missing signal explained the whole pattern, captures a fragment and mistakes the fragment for the mechanism. The Marxist claim is that the pattern is produced, continuously and impersonally, by relations that reward expansion and penalise restraint.

Class enters at exactly this point. Ownership of the productive apparatus is concentrated, while the people who operate that apparatus sell their capacity to work and exercise no direct authority over what gets built, burned, or shipped. Decisions about energy systems are therefore investment decisions, taken where capital is allocated, and judged by returns. Workers inside a carbon-intensive plant may understand its emissions perfectly and still depend on the plant for wages, housing credit, and pensions. Communities hosting such plants defend them for the same reason. An analysis pitched at individual consumption choices floats above these relations and explains almost nothing about why the underlying structure reproduces itself, which is why the Marxist reading relocates attention from shopping baskets to boardrooms, balance sheets, and the labour process itself.

What does the analysis explain that a price failure account leaves unexplained?

A price failure account explains why emissions continue when their costs fall on parties outside the transaction, and it prescribes a corrective charge. The class analysis explains a further fact: why corrective measures arrive weakened, delayed, or designed around threatened asset values, and why the pattern recurs across political systems and electoral cycles.

Notice, too, what the production framing does to the idea of responsibility. Aggregate humanity did not design the steam economy, does not own the freight networks, and does not sit in the rooms where generation capacity gets financed. Emissions are radically uneven in their origins when traced through ownership and income, a point developed below without recourse to any figure. Treating the crisis as the product of an undifferentiated species hides the levers, because the levers are specific: investment criteria, infrastructure lifetimes, competitive pressure, and the political defence of capitalised value. Naming those levers precisely is the service the analysis performs, whether or not a reader accepts its wider theoretical commitments.

A further consequence follows for how change is imagined. If emissions were a by-product of ignorance, education would suffice. If they were a by-product of mispricing, a tax schedule would suffice. If, instead, they are a by-product of how production is owned and compelled to expand, then the relevant questions concern ownership, planning capacity, and the terms on which existing assets are retired. Readers need not endorse any particular answer to see that the diagnosis redirects the inquiry. The diagnosis also generates predictions, which separates it from a mood or a posture: it predicts that restraint will be hardest exactly where capitalised fossil value is largest, that efficiency gains will be partially consumed by expanded output, and that unilateral restraint by individual firms will be punished by markets. Each prediction can be examined, and each has its own section below.

The question the analysis asks, and the questions it refuses

Every framework earns its keep partly by the questions it declines to answer. This one declines to rank personal virtues. Whether a reader flies, drives, or heats a large house belongs to moral biography, not to an account of why the energy system has the shape it has. The analysis also declines to treat population as its explanatory variable; aggregate numbers of people say nothing about who owns the smokestacks or why the smokestacks multiply. Equally, it declines to explain political outcomes by the personal corruption of officials, since the structural story it tells would predict similar outcomes under honest administrators facing the same asset values and the same competitive constraints. Declining those questions is not evasion. It is the discipline that keeps the explanation at the level where its mechanisms operate.

Positively, the analysis asks four questions. Why does output keep expanding even when the material consequences of expansion are documented by the relevant sciences and acknowledged across the political spectrum? Why was fossil energy adopted in its founding industrial case, and what does that adoption reveal about the forces selecting technologies? Why does the existing stock of fossil infrastructure defend itself so effectively in political arenas, out of proportion to the numbers of people it employs? And why do improvements in resource efficiency so often coincide with growth in total resource use? The four mechanisms table, presented in its own section below, assigns one mechanism to each question: competitive accumulation, the labour-control origin of fossil energy, capitalised sunk assets, and rebound in resource use.

Holding the four questions apart matters because public argument habitually merges them into a single fog of blame. A commentator proves that consumers enjoy cheap flights and concludes that ownership is irrelevant. Another proves that a firm lobbied against regulation and concludes that competition is irrelevant. Another cites engineering progress and concludes that asset defence is irrelevant. Each proof, even when accurate, addresses one mechanism while the others continue operating. The table format is chosen precisely to resist that merger: a reader can accept one row, contest another, and reject a third, and the analysis is structured to survive such partial verdicts. Scholarship in this tradition is at its best when disaggregated in exactly that way, since the mechanisms rest on different kinds of evidence and deserve separate trials.

There is also a question the analysis asks about itself, which honest exposition should surface early. What would count against it? Several findings would wound it gravely: demonstration that fossil adoption in the founding case tracked price signals closely and ignored labour supply; evidence that owners of fossil assets accept devaluation calmly when policy requires it; sustained absolute reductions in material throughput across wealthy economies on consumption-based measures, achieved inside ordinary competitive conditions. The counter-case section below presents the strongest available movement toward those findings, attributed to the research that reports it, and the answer offered there distinguishes what that movement settles from what it leaves open. A framework that specifies its own defeat conditions can be argued with; one that absorbs every fact as confirmation deserves the suspicion it attracts.

Finally, the analysis asks a question about time. Infrastructure built for fossil energy operates across spans far longer than electoral cycles or business quarters. A power station, a port, a motorway network, a suburban housing stock: each embodies a commitment made by one generation of investors that constrains several successor generations. Political argument, dominated by short horizons, systematically undervalues such commitments, treating the built environment as a backdrop rather than as frozen decisions. The sunk-capital mechanism is the analysis at its most temporal, and it explains a great deal of the peculiar slowness that frustrates observers who expect acknowledgement of a problem to produce proportionate motion.

Competitive accumulation: why no firm can restrain itself

The first mechanism requires no villain and no conspiracy, which is precisely its force. Firms produce under competitive pressure. A firm that voluntarily restricts its material throughput, raises its energy costs, or retires productive equipment ahead of schedule hands advantage to rivals who decline to follow. Investment flows toward the unrestrained. Managers who internalise restraint are replaced, acquired, or outcompeted; the selection operates impersonally, through margins and share prices rather than through anyone’s intentions. Marx described capital as self-expanding value, a formulation that sounds metaphysical until it is translated into the mundane observation that a business which stops growing inside a growing market is a business being liquidated in slow motion. The mechanism explains why exhortation fails. Appeals to conscience address the occupants of roles, while the roles themselves are defined by the competitive relation.

Extend the mechanism from firms to states and its grip tightens. Governments compete for investment, employment, and tax base. A jurisdiction that raises energy costs sharply while neighbours do not invites relocation of the most mobile production, a prospect that disciplines policy even where relocation threats are exaggerated. Hence the recurrent pattern in which climate measures arrive hedged with exemptions for the very sectors whose emissions prompted them. Observers often read those exemptions as evidence of lobbying power, and lobbying is real, yet the structural account predicts exemptions even under weak lobbying: the competitive position of national capital is itself a state interest in a system of rival states. The treatment of the state in this series, marxist state theory today, supplies the fuller account of that double character, manager of accumulation and arena of class forces at once.

Labour is caught inside the same mechanism, and the analysis loses credibility whenever it pretends otherwise. Workers in carbon-intensive industries face a credible threat: restrain the industry and their jobs, pensions, and towns contract before any alternative provision materialises. Their defence of the industry is therefore rational within the options actually presented to them, and a politics that sneers at it understands nothing. The mechanism generates a genuine alignment, temporary and tragic, between the owners of fossil capital and the workers whose livelihoods that capital organises. Breaking the alignment requires changing the options, which is a matter of provision and planning rather than of persuasion. Any version of the analysis that skips this point has confused a structural explanation with a morality play.

Why can a single firm not restrain its own throughput for long?

A single firm operates inside markets that reward lower costs and punish voluntary expense. Restraint raises its unit costs while rivals keep theirs unchanged, so orders migrate, margins compress, and capital withdraws toward less scrupulous competitors. The selection needs no malice; it runs through purchasing decisions and credit judgements until the restrained firm shrinks, sells, or abandons the restraint.

Growth, under this mechanism, is not a preference that could be talked out of the system; it is a survival condition imposed on each unit by the existence of the others. That is why the analysis treats aggregate expansion as the default outcome and asks what institutional arrangements could override the default, rather than asking why individuals fail to choose differently. It also explains the curious sincerity of much corporate environmental language. Executives may genuinely accept the relevant science and genuinely intend reductions, while the portfolio decisions their firms announce keep expanding total throughput, because each individual project clears the return threshold that competition imposes. Hypocrisy is the least interesting explanation available. The structure produces the outcome with honest occupants as readily as with cynical ones, and an analysis that needs villains to function would collapse exactly where this one stands firm.

One more refinement deserves emphasis. Competitive accumulation does not merely expand output; it expands output of whatever clears the market, with no internal register of material composition. A unit of revenue earned hauling coal and a unit earned insulating homes count identically in the accounts. The system is, in that precise sense, indifferent to the physical content of its own growth, and the indifference is not a correctable oversight but a consequence of value accounting itself. Critics of the analysis sometimes reply that prices could encode the missing register. The reply concedes the mechanism while proposing an external patch upon it, and the patch then faces the political defence of sunk assets described further below. Mechanisms interlock; that interlocking, more than any single row of the table, is the analysis in its mature form.

The four mechanisms table

The four mechanisms table gathers the argument into a single findable artifact. One row belongs to each mechanism: the compulsion to accumulate competitively, fossil energy’s labour-control origin, the sunk stock’s capitalised value, and rebound in resource use. Columns record what each mechanism claims, the evidence offered for it, what it predicts, and the strongest objection it must survive. Readers in a hurry can grade the whole analysis from this table alone; readers with time will find each row expanded in the sections surrounding it.

Mechanism	What it claims	The evidence	What it predicts	The strongest objection
Competitive accumulation	No firm or state can restrain throughput unilaterally without losing position	The recurrent pattern of exemptions, relocation threats, and reversed measures across rival jurisdictions	Restraint will be weakest where competition is most exposed, whatever the governing rhetoric	Coordination among states could in principle override the selection pressure
The labour-control origin	Fossil energy was adopted in its founding case because it freed production from site-bound labour	The British cotton industry’s migration from water sites to steam-powered towns despite water’s cost advantage	Attachment to fossil energy will outlast purely technical or price-based reasons for it	Price and reliability explanations may account for more of the transition than the labour evidence allows
Capitalised sunk assets	Existing fossil infrastructure embodies value whose destruction owners and states will resist	The political defence of long-lived assets and the design of measures around their book values	Resistance will concentrate where capitalised value is largest, not where employment is largest	Asset owners have accepted write-downs in other industries without blocking transition
Rebound in resource use	Efficiency gains lower the effective price of a service and thereby expand its consumption	The nineteenth-century coal record and a long subsequent literature with contested magnitudes	Efficiency policy alone will underdeliver against its engineering projections	Magnitudes vary by sector, and in some settings rebound is small enough to ignore

Read the table as a set of wagers rather than as a creed. The accumulation row wagers that competitive selection, not intention, drives expansion. The origin row wagers that a specific historical episode reveals the social logic of fossil adoption. The sunk assets row wagers that balance sheets explain political geography. The rebound row wagers that engineering projections systematically overstate savings because they ignore behavioural and structural responses. Each wager invites a distinct literature, and the strongest version of the analysis is the one that has read those literatures rather than gestured at them. Where the literatures push back, the sections below say so, and the counter-case gathers the pushback into its most formidable combined form.

A table also disciplines the writer. Claims that sound majestic in prose shrink to testable size when confined to a cell. The origin row, for instance, cannot hide behind atmosphere: it asserts something about one industry, one country, and several decades of site decisions, and the archives either sustain it or they do not. The rebound row cannot claim that efficiency is futile, because its own cell records that magnitudes are contested and sometimes small. Students meeting the analysis for the first time should be directed to the table before the rhetoric, since the table is where the argument consents to be measured.

The labour-control origin: fossil capital and the move to steam

Andreas Malm’s Fossil Capital, a study of the British cotton industry, supplies the analysis with its most concrete historical argument, and the argument deserves slow treatment because so much depends on its details. Early industrial cotton spinning and weaving drew power from water. Rivers and streams turned wheels; wheels drove shafts; shafts moved machinery. The arrangement worked, and in many valleys it worked cheaply, since the fuel was rainfall and the infrastructure, once built, asked little beyond maintenance. Yet the industry’s centre of gravity shifted toward steam, an energy source requiring mined coal, expensive engines, boilers, engineers, and constant attendance. Explaining that shift by cost or efficiency requires the evidence to cooperate, and Malm’s archival work argues that the evidence refuses: water remained abundant, frequently cheaper, and in several applications technically superior well into the period when manufacturers were committing to steam.

Why did industry switch from water to steam?

not because steam was cheaper. In the founding British case water power was abundant and often cheaper, but tied production to remote sites. Steam allowed mills to be located in towns with available labour that could be recruited and disciplined, which was its decisive advantage.

Unpack the labour problem that water imposed. A mill seated on a remote stream needed workers where workers were scarce. The owner had to attract families to a valley, build cottages, tolerate a workforce that knew its own scarcity value, and accept that the river set the schedule: drought idled the machinery, frost stopped it, and flood damaged it. Water power also resisted the employer’s desire to extend and regularise the working day, because the flow arrived on nature’s timetable rather than the market’s. Steam dissolved every element of that predicament at once. An engine could be erected in a populous town, beside a canal or a coal wharf, amid a labour market crowded with people who needed wages and could be replaced when they resisted. The engine ran when the owner chose, for as many hours as the owner chose, at a pace the owner set. Control over location, over the supply of hands, and over the clock constituted steam’s decisive product, more valuable than any saving on the power bill.

What evidence supports the labour-control reading of steam adoption?

The reading rests on manufacturers’ own testimony and business records from the transition decades: complaints about scarce and insubordinate rural labour, about housing and retaining workers at water sites, and about interrupted flow, set beside calculations in which water frequently appears the cheaper power. Firms then paid steam’s higher costs willingly once urban labour markets lay within reach.

Social historians had long documented the discipline dimension of factory production: the bells, fines, and supervised floors through which a workforce accustomed to task-oriented rhythms was remade into a workforce governed by the clock. Malm’s contribution was to connect that familiar story to the energy transition itself. The factory system and fossil energy were not two separate developments that happened to coincide; the second was selected, in part, because it served the first. Labour historians who studied time discipline and energy historians who studied prime movers had each held half of the explanation. Joining the halves produces the labour-control origin in its full form: the fossil economy was born as a solution to the problem of commanding other people’s time.

The claim’s precision should be noted, since critics and defenders alike blur it. Nothing in the argument requires that every manufacturer consciously calculated labour advantages, or that water sites vanished, or that steam lacked genuine technical virtues. Superior regularity, scalability, and independence from season were real virtues, and in some applications decisive ones. The argument is comparative and aggregate: across the industry’s decisive decades, the pattern of adoption, including expensive adoption against apparent cost interest, is better explained when labour supply and labour discipline enter the account than when price and efficiency stand alone. Historical explanation rarely offers cleaner terms than that, and readers should be suspicious of any participant, on any side, who demands them.

How did water power constrain the working day in ways owners resented?

Water arrived on a natural schedule. Flow strengthened and weakened with season and weather, halting machinery during drought and frost and surging beyond use during flood. An employer bound to a stream could not lengthen production into the evening, equalise output across the year, or synchronise several processes under one roof at a steady pace, however urgently orders pressed.

Coal, by contrast, waited in the yard until summoned. Stored energy is obedient energy, and the obedience extended from the boiler to the workforce: a mill that can run at any hour can also demand attendance at any hour, in shifts arranged around the machine rather than around daylight or river level. The celebrated extension of the working day in early factories, against which the first factory legislation pushed, presupposed a power source indifferent to season and sunset. Steam supplied that presupposition. When later generations inherited an economy in which production runs continuously and energy waits on demand, they inherited a social achievement disguised as a technical fact: the conquest of the working timetable, built into the fuel itself.

Geography carries the same lesson. Water sites were fixed points in a landscape, often picturesque, always specific. Capital seated at such a point was a hostage to that valley’s labour supply, that river’s moods, and that landlord’s terms. Steam unbound capital from place, letting it migrate toward labour, toward markets, toward coal measures, and away from any locality whose workers grew expensive or organised. Mobility became a permanent feature of the employer’s bargaining position: the credible threat to move elsewhere, which disciplines workforces in every subsequent century, was amplified enormously once production no longer had to sit beside falling water. Fossil energy and footloose capital grew up together, each enlarging the other.

Objections to the fossil capital argument, fairly weighed

Serious objections exist and an honest exposition states them in their strong forms. Economic historians have argued that water’s apparent cheapness dissolves under fuller accounting: the capital cost of dams, leats, and wheels; the losses from interrupted operation; the expense of transmitting power across a large mill; the difficulty of expanding a water site whose stream was already fully committed. On such accounting, steam’s total cost per unit of reliable power may have been lower than the simple comparison of fuel against flow suggests, and the transition becomes a story about reliability and scalability rather than about labour. A related objection holds that the industry’s geography had multiple drivers, including proximity to ports, to coal, to machine makers, and to finance, so singling out labour discipline mistakes one strand of a braid for the whole rope.

What do critics of the fossil capital argument reply at their strongest?

Critics reply that complete cost accounting rehabilitates the price explanation: water’s hidden expenses in construction, interruption, transmission, and exhausted sites narrow or erase its advantage, while steam offered reliability and room to grow that a rational manufacturer would purchase without any thought of discipline. Where the archives show labour complaints, critics answer, they show employers complaining about everything, always.

Malm’s rejoinder, and the rejoinder of the literature following him, leans on the pattern of choices rather than on testimony alone. Manufacturers repeatedly selected steam in settings where water remained available and, by the surviving calculations, economical; they expressed the labour advantages explicitly and often; and the industry’s urban concentration proceeded faster than the technical exhaustion of water sites can explain. The debate is genuinely open, and this article takes no position on climate science while taking the historical debate seriously on both sides: readers should register that the labour-control origin is a well-evidenced interpretation contested by competent specialists, not a settled datum. Even its partial truth carries weight, since a technology regime adopted substantially for control will not be dislodged by arguments addressed only to cost, a point whose contemporary implications the next sections develop.

A second objection cuts from the opposite direction. Suppose the labour reading is wholly correct about cotton in one country across several decades. What licences generalising from that case to the entire fossil economy, including sectors and centuries far removed from it? The licence claimed by the analysis is structural rather than statistical: the founding case reveals a selection logic, namely that technologies are chosen inside social relations and acquire constituencies from the advantages they give their owners, which then recurs wherever energy systems are built. Oil in transport, gas in heating, coal in electricity generation each have their own histories, and no serious writer collapses them into the cotton story. The case functions the way a fossil functions in geology: one preserved specimen, unusually complete, from which the anatomy of a much larger organism is inferred and then checked against other specimens.

There is also an objection about motives worth separating from the evidence. To say employers valued steam for the discipline it enabled is not to say they conspired in a room, nor that they were unusually cruel men by the standards of their class and period. Hiring, housing, and retaining a rural workforce were real managerial burdens; the urban labour market genuinely relieved them. Structural explanation assigns outcomes to positions and pressures, and it is fully compatible with the manufacturers having been, individually, ordinary commercial men responding intelligibly to their circumstances. The critique lands on the relations that made labour discipline a decisive commercial advantage, not on the characters of the men who noticed that it was.

Capitalised sunk assets and the anatomy of lock-in

The third mechanism shifts attention from adoption to persistence. Fossil infrastructure represents an enormous stock of capitalised value: extraction sites, pipelines, refineries, power stations, ports, vehicle fleets, road networks, and the buildings and urban layouts arranged around them. Each asset was financed against expected returns spread across decades of operation. Retiring such assets early converts expected returns into losses, and the losses land on identifiable balance sheets: firms, pension funds, municipal budgets, and states whose revenues lean on the sector. Political resistance to rapid transition is therefore legible as the defence of capitalised value, a structural explanation that neither requires nor denies the existence of lobbying. Lobbying is the visible instrument; the asset stock is the reason the instrument finds such receptive ground.

Carbon lock-in names the broader condition that the asset stock creates. Technologies, institutions, skills, and expectations interlock around the incumbent energy system. Engineers are trained for it, regulations presuppose it, supply chains serve it, consumers organise daily life around its conveniences, and each element raises the cost of departing from the others. Lock-in is not a conspiracy and not a metaphor; it is the ordinary path dependence of large technical systems, intensified by the sheer scale of the capital committed. The concept entered energy studies through the work of Gregory Unruh, who described how industrial economies become persistently organised around fossil energy through mutually reinforcing technical and institutional commitments. Marxist writers adopt the concept and give it a class inflection: among the institutions doing the locking, they emphasise ownership and the state’s interest in the value of the stock.

How does sunk capital turn into political resistance without any conspiracy?

An asset financed over decades must operate across those decades to repay its backers. Announce early closure and the loss is immediate, specific, and legally legible, while the gains are diffuse and deferred. Owners therefore litigate, lobby, and bargain; workers and regions tied to the asset join them; finance reprices the sector. Resistance follows the balance sheet predictably.

Consider the temporal asymmetry inside the mechanism. Building fossil infrastructure was a series of discrete, financeable decisions, each defensible on the information and incentives of its moment. Unbuilding it requires coordinated decisions of a different order, because the value of each asset depends on the survival of the network around it: a refinery without pipelines, a fleet without fuel distribution, a generating station without a grid designed for it. Partial retirement strands assets whose owners then demand compensation, and compensation at full book value would make transition programmes staggeringly expensive, while compensation below book value teaches investors that energy assets carry political risk, raising the cost of capital for the replacement system as well. Governments discover that every path runs through somebody’s balance sheet. The analysis predicts that governments will therefore prefer measures that leave existing assets to expire on schedule while steering new investment elsewhere, a preference visible in the recurring design of real policy, which favours subsidising the new over retiring the old.

Which commitments count as locked in, beyond power stations and pipelines?

Lock-in extends through vehicle fleets and the settlement patterns built for them, through heating systems embedded in housing stock, through farming methods organised around synthetic inputs, through professional training and regulatory templates, and through household routines and expectations. Together they mean an energy transition is simultaneously a renovation of ordinary life, which is why its friction surprises commentators.

The labour dimension of sunk capital deserves its own emphasis. Skills are assets too, held by workers rather than firms, and a transition that devalues a region’s accumulated skills destroys working-class capital as surely as it destroys corporate capital. The welder, the driller, the plant operator possess expertise purchased with years of their lives; telling them that the expertise must be written off, without equivalent provision, asks them to absorb a loss that shareholders are simultaneously fighting to avoid. Marxist analysis, which began with the labour process, is well placed to insist that the sunk-capital mechanism operates on both sides of the class divide, and that any credible account of transition politics must reckon with working-class sunk investment as carefully as with the corporate kind. Failure to reckon with it manufactures the very coalition, owners plus workers of the incumbent sector, that blocks change most effectively.

Financialisation deepens the mechanism further. Fossil assets sit inside pension funds, index portfolios, insurance reserves, and sovereign wealth vehicles, so their defence is mounted not only by energy firms but by institutions speaking for retirees and public budgets. A finance minister contemplating accelerated retirement confronts the exposure of the national pension system in the same briefing book as the emissions inventory. None of this implies that the assets cannot be written down; financial history is littered with written-down certainties, from canal stocks to telegraph networks. It implies that the write-down will be politically organised against, with broad institutional sponsorship, and that analyses treating resistance as the eccentricity of one industry’s public relations will be perpetually surprised by its breadth.

Rebound: Jevons and the efficiency paradox

William Stanley Jevons, studying British coal use in a work published in 1865, arrived at the observation that gives the fourth mechanism its name. Improvements in the efficiency of steam engines, he argued, did not reduce the nation’s coal consumption; they increased it, by lowering the cost of steam power and thereby multiplying the uses to which steam was profitably put. The Jevons paradox generalises the observation: raising the efficiency with which a resource delivers a service cheapens the service, and the cheapened service is consumed in greater quantity, sometimes to the point where total resource use rises rather than falls. Efficiency, intended as a conservation measure, behaves as an expansion measure whenever the rebound is complete.

The honest treatment of rebound requires care in both directions, because the contemporary magnitude of the effect is contested and varies by context. At the level of a single device or household, direct rebound is often partial: a better-insulated home is heated more warmly, yet usually not so much more warmly that the entire saving evaporates. Economy-wide effects are harder to bound, since money saved on energy is spent on other goods and services that themselves require energy, and since cheaper energy services can restructure whole industries around greater throughput. Energy economists dispute the size of these indirect effects with genuine technical arguments on multiple sides. The Marxist use of Jevons is narrower than the rhetorical use often made of it: rebound demonstrates that efficiency operates inside a system that converts savings into expansion, so efficiency policy alone cannot carry the weight that political rhetoric loads onto it. That is a claim about the system’s metabolism, not a claim that insulation is pointless.

Why does efficiency so often raise total resource use instead of reducing it?

Efficiency lowers the price of the service the resource provides: the mile travelled, the room warmed, the unit computed. A lower price invites more consumption directly, and frees income then spent on other energy-using goods. Where those responses exceed the engineering saving, total resource use climbs even as each act of consumption grows leaner.

How large is the rebound effect, on the honest reading of the literature?

The literature reports a spectrum. Direct rebound for household energy services in wealthy economies is frequently estimated as partial, sometimes modest. Economy-wide estimates spread wider, with some researchers arguing that indirect and structural effects can consume most or all of the engineering saving in some settings. Writers quoting one magnitude as settled have selected a conclusion, not reported a field.

Marx’s own notebooks engage the machinery question from the side of labour: a cheaper machine displaces workers in one place while cheapening the product and expanding the market that employs workers elsewhere, and the net effect on employment and output is an empirical tangle he refused to simplify. The rebound mechanism belongs to that same family of reasoning, in which a local improvement propagates through a competitive system and emerges, at system level, wearing an unexpected sign. Competitive accumulation is the transmission belt: savings achieved by one firm become lower prices, lower prices become expanded markets, expanded markets become new investment, and the resource saving is converted into growth with mechanical reliability. Rebound, in other words, is the first mechanism viewed from the standpoint of a fuel gauge. The four rows of the table are not four separate curiosities; they are one system photographed from four angles, and the efficiency case is where their interconnection is hardest to deny.

A policy corollary follows, and stating it is description rather than recommendation: programmes that rely on efficiency gains to deliver absolute reductions must either cap the expansion channel directly or accept that their projections will be eroded by the responses the system generates. Historical experience supplies the pattern repeatedly, in lighting, in vehicle fuel economy, in aviation, in computing, where spectacular gains in efficiency per unit of service have accompanied enormous growth in the total service consumed. The pattern does not refute efficiency as a goal; efficient provision is plainly better than wasteful provision at any given scale. It refutes the inference from efficiency to aggregate decline, which is the inference the mechanism exists to block.

The treadmill of production and the sociology of expansion

Allan Schnaiberg’s treadmill of production, developed within environmental sociology, offers a companion mechanism that Marxist writers frequently borrow. Producers, compelled to expand in order to maintain profits and employment, draw ever more material from ecosystems and return ever more waste to them; the state, dependent on the resulting revenue and employment, facilitates the expansion; workers, dependent on the jobs, acquiesce. Each participant behaves intelligibly, and the treadmill accelerates. The framework’s value for the climate argument is its insistence that expansion is a coalition outcome, not a corporate imposition: the growth machine enlists the state and organised labour as junior partners, which explains the political difficulty of braking it far better than any story about public misunderstanding.

What does the treadmill of production add that accumulation alone does not supply?

The treadmill specifies the coalition. Accumulation explains why firms must expand; the treadmill explains why states subsidise the expansion and why workers defend it, by tracing revenue dependence and employment dependence respectively. It thereby converts an economic compulsion into a political sociology, showing how each major institution acquires a material interest in the continuation of throughput growth.

Borrowing has limits, and the differences instruct. Schnaiberg’s framework treats the drive to expand as a feature of industrial production organised through markets and states generally, including states that described themselves as socialist, a generality Marxists handle in different ways. Some accept the generality and locate the problem in industrialism itself, which raises hard questions for any tradition promising emancipation through developed productive forces. Others insist that the specifically capitalist form, production for profit under competitive compulsion, sharpens the treadmill beyond what other arrangements produce, and point to the counter-case’s own evidence about planned economies as a separate matter, treated in socialist states environmental record. The dispute between those answers is live inside the tradition and is flagged here rather than resolved, because an application article that concealed it would misrepresent the state of the scholarship.

The treadmill also clarifies why information campaigns disappoint their sponsors. If participation in expansion were a mistake, better information would correct it. If participation is a livelihood strategy, information changes little until livelihoods are secured by other means. Publics that appear indifferent to findings they demonstrably accept are usually protecting incomes, mortgages, and municipal budgets, and the sociological tradition behind the treadmill documented that protection dynamic decades before climate politics made it famous. Analysis that begins from livelihoods rather than from attitudes predicts both the indifference and the specific flashpoints where resistance ignites: plant closures, fuel prices, and land use, the points where the treadmill touches household budgets directly.

Class, distribution, and the refusal of aggregate blame

Emissions attach to people very unevenly, and the unevenness is a matter of structure before it is a matter of anyone’s virtue. Output is organised by owners of capital; consumption is stratified by income; and the carbon content of a wealthy household’s life, its housing, travel, diet, and portfolio, reflects the services and assets that wealth commands. Within the working population, emissions are largely embedded in the necessities of participation: commuting to employment, heating rented housing whose insulation the tenant does not control, buying food from supply chains whose logistics the shopper did not design. A class analysis therefore distributes explanatory weight in a particular order: first to the organisation of production and investment, then to the consumption possibilities that income strata open or close, and only last, and weakly, to individual choice.

Why does the analysis refuse the language of individual blame?

Blame directed at consumers mislocates the decisions that determine emissions. The shopper did not choose the grid mix, the freight system, the building stock, or the product lines among which a constrained choice is made, and moral pressure leaves every one of those determinants intact. Worse, blame divides the population whose livelihoods a transition must secure.

How are emissions distributed, on the class reading offered here?

The reading distinguishes three locations. Emissions arise in production, where investment decisions fix the carbon content of goods before any consumer appears. They arise again in the consumption of higher-income strata, whose larger dwellings, vehicles, and travel multiply commanded throughput. Finally they arise as the unavoidable residue of ordinary working life inside systems nobody individually designed.

That distribution carries an analytical consequence that advocates and critics both sometimes miss. Because emissions concentrate where capital is allocated and where income is highest, measures addressed to the median consumer’s conscience touch the smallest and least elastic part of the problem while generating the largest political friction. The friction is not incidental: framing the crisis as a failure of ordinary people’s habits recruits ordinary people into defending the habits, and thereby into defending the system that assigned the habits to them. Class analysis predicts that framing, incidentally, because a system in difficulty benefits when scrutiny travels downward toward consumers and away from investment. Whether any particular commentator intends that benefit is beside the point; the structural effect operates regardless.

Responsibility, in this framework, is therefore assigned the way an engineer assigns load: to the members that actually bear it. Owners of fossil assets bear it as decision makers over the stock. Financial institutions bear it as allocators of the capital that renews the stock. States bear it as the only actors capable of coordinating retirement at scale. High-income consumers bear a real but derivative share, since their demand patterns reward the system’s most carbon-intensive luxuries. Workers in carbon-intensive employment bear the least share of any group routinely blamed, because their emissions arrive as a condition of employment they did not design and cannot unilaterally refuse. An ethic that inverts that ordering is not radical; it is merely confused about where causation lives.

The counter-case, first part: technology costs and the speed of learning

Against the four mechanisms stands a body of evidence that any honest exposition must place at full strength, attributed to the research that produced it. Begin with technology costs. Over recent decades, the costs of several low-carbon technologies have fallen substantially, in some cases dramatically, and the declines repeatedly outran the forecasts of official agencies and planning exercises alike. Research by Gregory Nemet, examining how solar energy became inexpensive, attributes the decline to a combination of forces operating inside market economies: publicly funded research and development, deployment policies that created protected early demand, manufacturing scale, and relentless incremental learning by competing firms. No central plan directed the process. Competition, which the first mechanism casts as the villain of restraint, appears in this literature as an engine of cost destruction, grinding down the price of the substitute technologies faster than the defenders of the incumbent system expected and faster than most transition models assumed.

The implication cuts deep, and it should be allowed to. If competitive markets, supplemented by public research and industrial policy, can collapse the cost of alternatives within a few decades, then the claim that the system is constitutionally incapable of addressing the problem loses its easiest evidence. Substitution becomes an ordinary capitalist process: a new sector grows by outcompeting an old one, creative destruction proceeds, and the sunk assets of incumbents are written down by market forces rather than defended forever. Defenders of the Marxist account reply that cost is not the binding constraint the analysis identifies, a reply examined below, yet the reply must first concede the fact itself: on cost, the system’s performance surprised its critics. Forecasts issued from within the Marxist tradition that treated renewable energy as permanently marginal read poorly against the record Nemet documents.

What has driven low-carbon technology costs down, on the strongest available research?

Nemet’s research assigns the decline to interacting causes: public research funding that solved early technical problems, deployment support that guaranteed initial markets, manufacturing competition that punished inefficiency, and learning by doing across an expanding industry. The combination operated inside market economies with active states, which is why the finding challenges simple stories told by both market celebrants and critics.

Learning curves deserve a paragraph of their own, because they complicate every participant’s favourite narrative. Each doubling of cumulative production for a young technology tends to reduce its cost by a roughly characteristic proportion, as engineers refine designs and factories specialise. Incumbent fossil technologies sit at the flat end of their curves, mature and optimised, while challengers descend steep early slopes, so time itself works for the challenger once deployment begins. States that funded early deployment therefore purchased not only installations but position on the curve, an industrial policy outcome achieved through market instruments. Marxist readers will recognise a familiar pattern in modern dress: the state socialising the risk of a new productive force while private firms appropriate the returns, exactly the division of labour the tradition has always described in railway building and armaments. The pattern confirms one Marxist claim about how capitalism innovates even while embarrassing another about how slowly it can move.

Sectoral unevenness, however, already announces the answer that the analysis will give. Cost collapse has been spectacular in electricity generation and in some categories of equipment, while other domains, aviation, shipping, heavy industry, agriculture, building heat in cold climates, have moved far more slowly, because the physics, the infrastructure, or the asset lifetimes resist. A system that transforms its easiest sector rapidly and its hardest sectors barely at all presents a genuinely mixed picture, and both camps in the debate survive by quoting their preferred half of it. The analytical task is to hold the whole picture in view at once: direction of travel in the transformed sectors is real, and so is the inertia elsewhere, and any verdict pronounced on the system as a whole has to be an aggregate judgement about rates across unlike domains.

The counter-case, second part: decoupling and its discontents

The second body of counter-evidence concerns decoupling, the separation of economic output from emissions. Research associated with Corinne Le Quéré and collaborators, tracking emissions across wealthy economies over recent decades, reports cases in which emissions declined while output grew, including on consumption-based measures that attribute the emissions embedded in imported goods to the importing economy rather than to the producer. That inclusion matters enormously, because the standard criticism of wealthy-economy decoupling is that it merely exports emissions: factories close at home, goods arrive from abroad, and territorial accounts improve while the global total does not. The consumption-based findings answer that criticism in part. For the economies and periods studied, the decline survives the correction, which means at least some wealthy economies have absolutely decoupled, on some measures, over some periods, while continuing to grow.

State the finding’s full force before qualifying it. The Marxist account, in its strong form, holds that competitive accumulation binds output and throughput together so tightly that absolute decoupling at the level of a whole economy should not occur under ordinary capitalist conditions. Documented cases of exactly that occurrence are therefore not a debating point but a falsification pressure on the strong form, and the tradition’s more careful writers have absorbed the pressure by retreating to claims about rate and sufficiency: decoupling, they argue, proceeds too slowly, covers too few economies, and rests partly on one-off transitions such as fuel switching in power generation that cannot be repeated indefinitely. That retreat may be justified, yet it is a retreat, from impossibility to insufficiency, and intellectual honesty requires naming it as one. The weaker claim, that the system moves too slowly relative to what the relevant science describes as necessary, depends on scientific findings this article attributes rather than adjudicates, and it is a claim about pace that evidence could continue to erode.

Why is the measurement of decoupling disputed among specialists?

Measurement disputes begin with boundaries. Territorial accounts credit an economy for emissions inside its borders, while consumption accounts reassign emissions embedded in trade to final consumers, and the two tell different stories for economies importing manufactures heavily. Further disputes concern which gases enter the ledger, how international transport is allocated, and whether favourable periods reflect durable structure or passing conditions.

The dispute matters beyond methodology because both camps in the political argument select the ledger that favours them. Celebrants of green growth quote territorial declines; critics quote consumption-based totals or global aggregates, where no decoupling of the required kind has been demonstrated. A scrupulous reader should hold three facts simultaneously: decoupling in some wealthy economies, on consumption-based measures, over recent decades, is documented by named research; the global aggregate continues to be dominated by growth in emissions elsewhere, driven substantially by industrialisation in poorer economies whose claim to development the Marxist tradition itself defends; and the rate question, whether observed decoupling is fast enough and broad enough, cannot be settled by the decoupling literature alone, since it imports its benchmark from climate science. The analysis that emerges from holding all three is less satisfying than either camp’s pamphlet and considerably more accurate.

Attention should also fall on what decoupling, where it occurred, was made of. Studies decomposing the declines find contributions from energy efficiency, from fuel switching within electricity, from structural change toward services, and from policy instruments including carbon pricing in some jurisdictions and regulation in others. Almost none of the decomposition attributes the decline to markets operating without state action; the states involved taxed, regulated, subsidised, and procured. The counter-case therefore refutes a caricature, the claim that capitalist states can do nothing, while leaving open the Marxist question of whether states can act against the interests of capital at the speed and scale the situation is said to require. Evidence about what states did under favourable conditions, in power sectors with cheap substitutes available, does not by itself answer what states will do in sectors where substitutes remain expensive and asset values remain enormous.

The strongest objection, assembled, and the answer the analysis can honestly give

Assemble the counter-case into the single objection the brief for this article demands be stated at full strength, attributed to the named research surveyed above. Competitive markets combined with public research and industrial policy have driven cost reductions in low-carbon technologies faster than any planning exercise anticipated, as Nemet’s research documents. Several wealthy economies have reduced emissions while growing, including on consumption-based measures, as the research associated with Le Quéré and collaborators records. And the historical record of planned economies on environmental performance is poor, a record examined in its own right in socialist states environmental record, where the disasters are stated plainly before any explanation is attempted. If all three claims hold, then the assertion that the system cannot address the problem appears to be refuted by what the system has actually done, and the Marxist analysis survives, if at all, as a theory of obstacles rather than of impossibility.

What is the strongest objection to the Marxist account of the crisis?

The strongest objection is empirical and cumulative. Markets plus active states collapsed the cost of substitutes faster than planners predicted; several wealthy economies decoupled emissions from growth even after imports are counted; and the planned economies whose example might have vindicated an alternative compiled a wretched environmental record. Together, the objection concludes, capability is demonstrated and the impossibility thesis collapses.

How does the analysis answer the decoupling and cost evidence?

The answer distinguishes rate from direction. Direction, in the sectors and economies the objection cites, is conceded: substitution occurred, costs fell, emissions declined under growing output. Rate raises the questions the mechanisms were built to ask: how quickly change spreads to the hard sectors, how much repeats once easy fuel switching ends, and whether asset defence slows each later stage.

Concede, first and without varnish, that the strong impossibility claim is not supported. A tradition that prizes materialism should be the first to revise a thesis when the material record turns, and the record shows capitalist economies reducing emissions in absolute terms in a number of cases while innovating at a pace their critics dismissed. Writers who continue to assert that nothing can change within the system are defending a slogan against evidence, and this article does not join them. The defensible claim is narrower, conditional, and still substantial: the mechanisms retard, distort, and unevenly distribute the transition, so that its pace and shape are set by the defence of capitalised value and the compulsions of competition rather than by the requirements the relevant science describes. Whether the resulting pace suffices is a judgement that depends on scientific findings attributed here and not adjudicated, which is exactly where a social analysis should stop.

Second, identify which sectors have not moved and ask why, since the mechanisms earn their keep precisely there. Electricity generation, where substitutes became cheap and assets are comparatively modular, moved first and fastest. Domains organised around long-lived, network-bound, or energy-dense requirements, heavy industry, long-distance freight, aviation, shipping, and the heating of an ageing building stock, have moved slowly or not at all, and in each case the obstacle profile matches the table: sunk capital is largest, infrastructure interlock most complete, competitive exposure most punishing for first movers, and rebound most available, since cheaper transport historically expands the volume of transport demanded. The analysis does not merely survive the sectoral pattern; it predicts the ordering. An objection that quotes the power sector and falls silent about freight has answered the easy half of the examination.

Which parts of an economy have proved hardest to shift away from fossil energy?

Heavy industry stands high, because its processes demand intense heat and its plants embody decades of committed capital. Long-distance freight and shipping follow, bound to energy-dense fuels and infrastructure. Aviation shares the density problem and adds rebound, since efficiency gains are spent on cheaper flight. Building heat lags because the asset is the housing stock itself, renewed at glacial pace.

Third, the planned-economy point requires the analysis to be precise about what it claims. The poor environmental record of states that industrialised under Marxist governments is a fact of the historical record, documented in the article devoted to it, and it forbids any lazy inference from public ownership to ecological care. Those states pursued output growth with coal-heavy technology, suppressed the independent scrutiny that might have checked the damage, and treated nature as a storehouse to be conquered, failings the tradition’s own ecological writers have catalogued bitterly. What the record does not show is that the mechanisms in the table were absent there; output targets imposed their own accumulation compulsion, a treadmill without private owners, which suggests the deeper variable may be growth compulsion under any ownership form that subordinates ecology to expansion. That suggestion is uncomfortable for several Marxisms at once, and it is left here as the open question it is, with the fuller comparison deferred to ecosocialism vs green capitalism.

Where the analysis criticises itself

Internal criticism keeps a tradition alive, and ecological Marxism has accumulated pointed criticisms of its own climate writing from within. One charge is determinism: popular versions of the argument move from the existence of mechanisms to the inevitability of outcomes, forgetting that mechanisms operate through politics, where organisation, ideas, and accidents alter what structures dispose. A second charge is functionalism, the habit of explaining every outcome as what capital needed, including outcomes, such as the cost collapse of substitutes, that plainly injured incumbent capitals; a system whose needs explain both an event and its opposite explains nothing. A third charge concerns agency: writing that casts capital as the sole protagonist and everyone else as audience offers workers and publics no role except spectatorship, and then wonders at its own political sterility.

Where do Marxist writers criticise the climate analysis from inside the tradition?

Internal critics target three habits: determinism, which converts pressures into certainties; functionalism, which retrofits every outcome as a requirement of capital, including outcomes that damaged incumbent firms; and the disappearance of agency, which leaves no actor capable of interrupting the mechanisms. Each criticism has produced corrections in the better scholarship, where mechanisms are stated as pressures whose outcomes politics decides.

A fourth internal criticism concerns the treatment of science, and it connects to the policy governing this article. Some climate writing in the tradition borrows the authority of climate science rhetorically, presenting contested social claims as though they carried the warrant of physical findings. The borrowing is both dishonest and unnecessary: the social claims are strong enough to be argued on social evidence, and wrapping them in borrowed certainty teaches readers to distrust the genuine article. Hence the discipline observed throughout these pages, under which physical findings are attributed to the scientific bodies and literatures that produced them, never restated as this article’s own discoveries, and never conscripted as applause for a political economy argument. Where an economic claim depends on a scientific finding, such as the claim that transition pace is insufficient, the dependence is flagged at the point of use.

Fifth, the tradition argues with itself about growth in the poorer world. A climate analysis addressed to wealthy economies generalises badly when exported: societies lacking sanitation, electricity, and steel cannot treat throughput as a surplus category, and Marxist writers from those societies have said so forcefully, insisting that the critique of accumulation targets the compulsion to expand for profit, not the construction of water systems. The richer version of the analysis holds both thoughts: expansion compelled by competition is the pathology, while expansion answering unmet need is the substance of development, and confusing the two produces a politics that lectures the poor on restraint while the wealthy defend assets. Readers pursuing the throughput question itself should continue to degrowth and marxism, where the dispute between the tradition’s own camps receives full treatment.

The state inside the structural account

States appear in this analysis in three capacities at once, and collapsing them into one produces most of the confusion in public argument. A state is a manager of the conditions for accumulation, responsible for the infrastructure, legal order, and energy security within which firms operate. A state is also an owner and financier of fossil assets directly, through nationalised industries, sovereign funds, and public pension exposure, and therefore a party to the sunk-capital defence rather than a referee above it. Finally, a state is an arena in which classes and coalitions contend, so its policy at any moment reflects the balance of organised forces pressing upon it, including the treadmill coalition of producers, dependent workers, and revenue-hungry treasuries. Measures emerge from the intersection of the three capacities, which is why the same state can subsidise substitutes with one hand and license new extraction with the other without any individual official being insincere.

What role does the state play in the structural account given here?

The state manages accumulation’s conditions, owns and finances part of the fossil stock itself, and hosts the contest among classes and coalitions over what to do. Because the three capacities coexist, policy characteristically advances on substitutes while protecting incumbent assets, and the structural account reads that doubleness as the normal output of the arrangement rather than as betrayal or capture.

Understanding the state this way reframes the recurring disappointment of climate summits and legislative seasons without naming any of them. Observers ask why governments that accept the relevant science, and employ the scientists who produce it, nonetheless deliver measures sized to the tolerance of the asset stock. The structural answer points to the treasury bench rather than the press release: fossil sectors supply revenue, employment in specific regions, energy security in the short term, and collateral inside the financial system, and a government moving faster than those dependencies allow must either replace them deliberately or be punished through investment strikes, regional revolts, and electoral losses. None of that makes faster movement impossible; states in emergencies have overridden comparable dependencies. It makes faster movement a deliberate act of statecraft against the grain of the state’s own first capacity, which is why it happens rarely and usually under duress.

Historical perspective widens the point. The fossil economy was itself a state-assisted construction: coal measures surveyed and enclosed, canals and railways chartered, navies converted, grids nationalised or franchised, suburbs zoned around the motor car. A system built with that much public sponsorship will not be unbuilt by public abstention, a conclusion that follows from the sunk-capital row of the table and that every serious participant in the debate, whatever their politics, in practice accepts when designing policy. The argument among traditions concerns the terms of the unbuilding: who is compensated, who is stranded, whose skills are honoured with equivalent provision, and whose assets are treated as the public’s problem. Marxist analysis insists those distributional questions are the real content of transition politics, hidden inside technical vocabularies of pathways and portfolios.

For the wider setting of these arguments, including how the tradition reads the present form of the system it criticises, see marxism and modern capitalism.

Origin explains persistence: the series pattern in this case

One pattern recurs across this series: establishing why an institution or technology was adopted explains why it resists displacement better than cataloguing the interests that now defend it. Present interests are real, yet they are also replaceable; firms diversify, funds rebalance, lobbies dissolve and re-form. The deeper resistance comes from fit: the technology was selected because it served the organisation of production, and it continues to serve that organisation, so substituting it means reorganising production and not merely swapping a fuel. Steam was not adopted as an energy source that happened to have social effects; it was adopted as a social arrangement that happened to burn coal. An economy built on that adoption inherits the arrangement along with the fuel, which is why technical substitutes, however cheap, meet friction that price comparisons cannot see.

Why does the origin of a technology matter for understanding its future?

Origins reveal the selection criteria that chose a technology from among its rivals. Where the winning criteria included control over labour, location, and time, the system embodies those criteria in its layouts, skills, and expectations. A substitute matching the incumbent on price alone still fails the unlisted criteria, and adoption stalls for reasons appearing irrational until the selection is recovered.

What separates a structural explanation of persistence from a list of interests?

Lists of interests name the parties who benefit from the status quo and would stop defending it if bought off or defeated. A structural explanation identifies the relations that regenerate the defence: competition punishing restraint, capitalised value demanding returns, livelihoods bound to the stock. Buy off every interest and the relations recruit defenders, which is why structural accounts predict resilience.

What does the analysis predict about purely technical substitutes?

The prediction is asymmetrical uptake. Substitutes that serve the incumbent organisation of production, dropping into existing grids, fleets, and business models with minimal reorganisation, spread rapidly once cheap. Substitutes demanding new settlement patterns, new labour processes, or premature scrapping of valuable stock spread slowly regardless of technical elegance, until the organisational question is settled by policy, crisis, or generational turnover.

Stress-test the pattern against the counter-case evidence and it holds up better than the impossibility thesis did. Solar generation, once inexpensive, fitted the existing organisation of electricity supply: centralised or distributed, it feeds the same grids, serves the same appliances, and asks nothing of households except roof space or a tariff change. Its spread was correspondingly swift. Building heat, by contrast, asks millions of property owners to renovate the fabric of their homes and replace the appliance at the system’s heart, inside a housing stock whose renewal is measured in generations; cheap heat pumps alone do not dissolve that organisational problem, and uptake lags in precisely the way the pattern predicts. The origin lens converts a miscellaneous set of policy frustrations into a single intelligible gradient, which is the kind of service a theory is supposed to perform.

The pattern also disciplines optimism about invention. A breakthrough fuel or reactor design, announced tomorrow in the thought experiment’s permissive sense, would still enter a world of financed assets, trained workforces, zoning codes, and grid architectures built for its predecessors. History’s verdict on superior techniques that arrived before their organisational moment is unsentimental: they wait, sometimes for generations, until the relations that would reward them exist. None of this argues against invention; it argues against the theory of change in which invention is sufficient, a theory the fossil record of the nineteenth century, with water’s unexploited virtues sitting in plain sight, refutes in advance.

A note on method: how this article handles climate science

Method deserves explicit statement in an article whose subject invites conflation. Physical climate findings, throughout these pages, are credited to the bodies and literatures responsible for them; they are not restated in this article’s own voice, not quantified, and not adjudicated. No emissions figure, temperature figure, cost figure, or decoupling measurement appears anywhere above, by design rather than by oversight: the analysis defended here concerns the social mechanisms that produce emissions and resist transition, and those mechanisms can be stated, evidenced, and contested without borrowing a single number from the physical sciences. Where an economic or social claim depends on a scientific finding, the dependence has been flagged at the point of dependence, most visibly in the discussion of transition pace, where the judgement that observed decoupling is insufficient imports its benchmark from climate science and is labelled accordingly.

How does this article treat findings that originate in climate science?

Findings are credited to the bodies and literatures that produced them, and the article asserts none of its own. Where a social claim requires a scientific premise, the borrowing is flagged at the point of use. Scientific consensus is never used here as a rhetorical instrument, and disagreement about policy is never redescribed as disagreement about physics.

Two further rules follow from the method and bind the whole cluster to which this article belongs. Scientific disputes internal to a field are reported as disputes, with the contest named, rather than settled by editorial preference; the rebound literature received that treatment above, and so did the decoupling measurements. And the moral valences are kept separate: establishing that a mechanism operates says nothing, by itself, about what ought to be done, which is why this article recommends no policy and evaluates none. Readers who finish it knowing exactly what the analysis claims, what evidence would wound it, and where its own tradition argues against it have received everything an application article in this series promises.

Misreadings of the analysis, corrected

Five misreadings circulate so reliably that correcting them amounts to a summary of the argument. First, the analysis does not blame individual consumption; it explains why consumption is structured by production and income, and assigns explanatory priority accordingly. Second, it does not treat emissions as evenly distributed across humanity; the entire class apparatus exists to deny that premise. Third, it does not claim fossil fuels triumphed because they were cheapest; its founding case argues nearly the opposite. Fourth, it does not claim that efficiency is pointless; rebound qualifies the aggregate effect of efficiency, while efficient provision remains preferable at any scale. Fifth, it does not claim that falling technology costs settle the argument in either direction; costs govern one constraint among several, and sectors where costs collapsed behave differently from sectors ruled by sunk assets and organisational fit.

Does the analysis claim that emissions cuts are impossible under capitalism?

No, and the claim should be retired wherever it appears. Documented absolute reductions in several wealthy economies, alongside collapsing substitute costs, refute impossibility, and this article has presented that evidence at full strength. The defensible claim is conditional: competition, capitalised value, and rebound retard and distort reduction, so pace and distribution are set by those mechanisms rather than by need.

A sixth misreading deserves addition because it governs how the other five survive. Sympathetic summaries often present the analysis as a complete theory of climate politics, when its actual ambition is narrower: an account of obstacles, located in production relations, that any adequate politics must reckon with. It does not supply a theory of social movements, of party formation, of international negotiation, or of the cultural meanings of nature, and writers who stretch it across those territories produce the bloated version that critics then enjoy refuting. The four mechanisms table is deliberately small. Its modesty is a methodological commitment, and the closing sections of this article respect it by ending with an inventory of what remains unexplained rather than with a peroration.

What remains unexplained, then, should be listed in prose. The analysis does not explain why some states move faster than others with similar asset stocks, a variation that sends the reader toward comparative politics and the study of institutions. It does not explain the timing of cultural shifts, the moments when publics suddenly reward or punish environmental positions, which belong to the sociology of opinion. It does not explain innovation’s direction in any detail, beyond noting that competition steers research toward appropriable returns. And it does not, by itself, rank the strategies available to those who accept its diagnosis, since strategy depends on judgements about organisation and opportunity that a mechanisms table cannot supply. Naming those limits is not a confession of failure; it is the boundary that keeps the claims inside the evidence.

Water’s long persistence and the unevenness of the transition

A careless reading of the fossil capital argument imagines water power collapsing the moment steam appeared, swept away by a superior rival. The historical record shows something slower and more instructive. Water remained the cheaper and often the preferred power source in numerous districts for decades after steam engines became commercially available, and new water installations continued to be built, improved, and enlarged well into the period of steam’s ascendancy. Wheels grew more efficient under the pressure of competition, and the great turbine eventually gave water a second industrial life. The transition was a contested overlap lasting generations, during which the two power sources coexisted, competed, and were chosen mill by mill according to local circumstances. That protracted character is itself evidence for the labour-control reading: a transition governed purely by cost should have proceeded at the speed of the price difference, whereas a transition governed by the fit between a power source and the organisation of labour proceeds at the speed of urbanisation, labour market formation, and the exhaustion of particular rural arrangements.

Site-level decision making rewards close attention, because it is there that the competing explanations can be weighed against each other most fairly. A manufacturer inheriting a water site with established housing, a trained workforce, and a reliable stream faced one calculation; a manufacturer founding a new enterprise faced another, freer to weigh the town against the valley. The pattern Malm documents is consistent with that distinction: steam’s advance was fastest where new capacity was being created and where urban labour markets were deepest, while water held its ground longest where sunk investments in dams, housing, and community bound employer and workforce together in an arrangement both sides had learned to operate. Sunk capital, the third mechanism, thus appears inside the first transition as well as in the fossil economy that transition created. The owners of water sites defended their capitalised value against steam exactly as the owners of fossil assets defend theirs against substitutes, and the parallel is more than rhetorical: it suggests that resistance to energy transition is a generic feature of capitalised infrastructure, amplified in the fossil case by the scale of the stock and by the labour relations the stock was built to serve.

Regional variation sharpens the lesson further. Districts with powerful streams and sparse populations retained water long after districts with weak streams and crowded towns had converted, and the difference maps onto labour supply at least as neatly as onto hydrology. Where workers were plentiful and cheap at the valley’s edge, the valley remained attractive; where the valley demanded that an employer conjure a workforce out of moorland and then keep it, the town won. A purely technical history can describe these outcomes district by district without ever noticing the common variable, rather as a tide table describes each harbour without mentioning the moon. The labour-control reading names the moon. Critics are entitled to answer that coal measures, transport links, and machine-making centres also cluster in the towns, and the clustering is real; the scholarly question is which variable does the explanatory work when the accounts are compared case by case, and that question remains genuinely open in the literature, as this article has already conceded.

Why did water power survive for decades beside the steam engine?

Water survived because its advantages were real and locally decisive: negligible fuel cost, mechanical simplicity, and, in well-sited mills, adequate power across most of the year. Owners with established valleys, housed workforces, and reliable streams faced no reason to scrap assets for engines. Steam conquered new ground and labour-hungry districts first, which is why the two systems coexisted across generations.

Coal, geography, and the remaking of industrial space

Energy sources are also geographies, and the shift from water to steam redrew the map of production in ways that still structure economic life. Water power dispersed industry along valleys, scattering small industrial communities through landscapes that had been agricultural. Steam concentrated industry. Coal measures, canal junctions, and ports became the poles around which factory towns assembled, drawing population from the countryside into dense urban districts built at extraordinary speed and with minimal provision for health or comfort. The concentration was not an accidental by-product of the fuel; it was among the fuel’s attractions, since density meant labour markets, and labour markets meant the replaceability that disciplined wages and effort. Urbanisation and fossil energy are usually narrated as parallel stories. On the labour-control reading they are the same story, told once in the vocabulary of demography and once in the vocabulary of power.

Concentration generated the classic pathologies that nineteenth-century reformers documented and that the socialist tradition made central to its case: overcrowded housing, contaminated water, epidemic disease, and the peculiar misery of populations gathered for production and abandoned by every other institution. Those pathologies belong to the history of public health rather than to climate analysis, yet they matter here for one reason. They demonstrate that the social costs of an energy system have always been distributed by class, falling on the populations the system assembled while its returns accrued elsewhere. The climate crisis globalises that distribution, spreading the costs of the fossil economy across populations who never shared its returns, including generations who did not participate in its construction. A class analysis recognises the pattern because it has seen the pattern before, at the scale of a mill town rather than of a planet.

Transport completed the remaking of space. Coal had to move from pit to boiler, and the networks built to move it, canals first and then railways, became the skeleton on which the rest of the industrial economy hung. Railways then consumed coal in their own right while teaching capital a new lesson about scale: networks reward whoever builds them largest and earliest, and they strand whoever guesses wrong. The lesson generalises directly to the fossil economy’s later layers, pipelines, refineries, tanker routes, and motorway systems, each a network whose value depends on the survival of the whole. Lock-in, encountered earlier as a balance-sheet phenomenon, is equally a property of networks: a single pipeline is a stranded tube, while a pipeline system is a civilisation’s circulatory system, and the difference explains why retiring the parts feels, to those who govern, like retiring the whole.

The motor vehicle, the road, and the reorganisation of settlement

No technology illustrates lock-in at civilisational scale better than the motor vehicle and the settlement patterns built around it. The private car became the organising unit of urban form across a vast swathe of the industrial world: suburbs spread along roads, shops migrated to the periphery, public transport withered in districts designed for driving, and the daily commute lengthened until the vehicle was no longer a luxury but a condition of employment itself. Workers in such districts do not choose their transport emissions in any meaningful sense; the built environment chooses for them, having been designed, financed, and zoned under assumptions about cheap fuel and individual mobility that no household can revise alone. Here the sunk-capital mechanism takes a material form visible from any window: a world built for the car resists departures from the car even when the vehicles themselves become cleaner, because the problem was never only the engine.

The labour dimension of road transport extends in two directions. On the supply side, the road system sustains vast employment in construction, maintenance, vehicle manufacture, freight, and fuel distribution, reproducing the treadmill coalition in the most literal sense: the people who build and serve the system depend on it and defend it. On the demand side, the vehicle disciplines the worker’s time, transforming commuting hours into unpaid labour time spent serving the geography of production, a transfer of cost from employer to employee that the wage bargain rarely acknowledges. The motor vehicle also reorganised the household division of labour, since the school run, the shopping trip, and the care journey all became motorised errands, and the burden of those errands fell unevenly within households. A class analysis that ignored these everyday disciplines would be incomplete; the fossil economy governs the working day, the working week, and the geography of care, not merely the generating station.

Substituting a different vehicle into the same road system therefore addresses the smallest part of the lock-in, which is why the analysis predicts the slow, contested character of transport transition even where the substitute technology performs well. Electrifying the fleet leaves the settlement pattern, the road network, the commute distances, the freight logistics, and the employment dependencies untouched, while every element continues to demand the energy and materials that sustain it. Whether that partial substitution counts as success or as postponement is precisely the kind of judgement the mechanisms leave to the reader’s wider politics; the analysis itself contributes only the structural map, showing which elements of the system move when a single variable changes and which remain anchored by their own interlocks.

Housing, heat, and the slowest asset of all

If transport is the most visible fossil lock-in, housing is the slowest. The dwellings in which populations live were built across many decades, overwhelmingly heated by fossil combustion or by electricity generated from fossil fuels, and embedded in settlements whose energy demands were set by the construction standards of their era. A building stock renews itself at a pace measured in generations, and each dwelling embodies a set of irreversible or nearly irreversible decisions: orientation, insulation, pipework, appliance type, connection to a grid or a gas main. The owner of an older dwelling faces a calculation that is individually rational and collectively paralysing: a deep retrofit costs more than the fuel saving returns within any horizon that matters to a household budget, while the alternative, doing nothing, costs nothing now and contributes its share to a problem the household did not create alone.

Ownership complicates the mechanism in ways that repay class analysis. Landlords, who decide what is installed, do not pay the fuel bills; tenants, who pay the fuel bills, cannot decide what is installed. The split incentive is well known to energy economists, and the Marxist gloss is straightforward: property relations sever the person who bears the cost from the person who controls the asset, a separation that markets do not spontaneously repair. Owner-occupiers face the mirror problem: the capital cost of retrofit falls on one balance sheet while the climate benefit accrues to everyone, which is the textbook definition of an externality, and the political economy question is why the textbooks’ remedies arrive so slowly for housing while arriving promptly for sectors with stronger lobbies. The answer, on the sunk-capital reading, is that housing’s capital is fragmented among millions of owners who cannot lobby as a class, while its fuel supply is concentrated among firms that can. Fragmentation is itself a political fact: the interests of the many are dispersed, and the interests of the few are organised, so policy moves where organisation concentrates.

Heat also exposes the limit of generation-centred transition thinking. An economy can replace its power stations while its dwellings still demand combustion for warmth, and the seasonal peak of that demand, cold evenings when everyone heats at once, is exactly the moment when supply is hardest to furnish from intermittent sources. Engineers know this; what the class analysis adds is the social translation: the seasonal peak is a property of buildings, settlements, and working hours, not of the weather alone, and addressing it means addressing the built environment rather than the fuel mix. No measurement is needed to state that dependence; the structural claim is that the demand pattern is socially produced, and the social production is visible in the construction history of the housing stock itself.

Finance, insurance, and the pricing of a stranded world

Follow the money through the fossil economy and the sunk-capital mechanism acquires a balance-sheet geography. Extraction, refining, and generation assets are financed by loans, bonds, and equity whose holders are rarely energy firms themselves: they are banks, pension funds, insurance companies, and millions of households saving for retirement through products that hold a slice of everything. That diffusion creates the political breadth of the asset defence, because an accelerated retirement of fossil assets sends losses ricocheting into retirement incomes and public finances that have nothing visible to do with energy. Finance ministers reading the problem see a distributional minefield, and their caution, whatever else explains it, is also the sunk-capital mechanism operating through the most abstract and therefore the most hidden of channels. The losses are socially diffused in advance, which means the resistance is socially diffused in advance as well.

Insurance supplies a stranger and more instructive angle. The insurance industry is the one corner of capital that must price long-run physical change as a condition of its own survival, since its liabilities are claims on a future world. Actuarial assessments therefore absorb the findings of the relevant physical sciences without waiting for political permission, and premiums in exposed regions adjust to risk as the models dictate. Yet the same industry invests its reserves across the whole economy, fossil assets included, and the tension between its two activities is a small, precise model of the larger structural contradiction: knowledge of the physical problem, incorporated rigorously where knowledge protects capital, coexisting with continued financing of the activities the knowledge implicates. The Marxist reading of that tension is not that insurers are hypocrites but that each unit of capital incorporates information selectively, exactly where its own balance sheet compels attention, and the system as a whole has no corresponding compulsion. No board of directors exists for the climate; there are only boards for firms.

Stranded assets complete the financial picture. The phrase entered economic discussion to describe fossil reserves and infrastructure that cannot be exploited or operated for their full expected life if climate policy is taken seriously, and its significance lies in the fight over who bears the stranding. A rapid transition strands value held by identifiable owners; a slow transition strands coasts, harvests, and water supplies held by unidentifiable everyone. The conflict over pace is therefore also a conflict over which assets a society treats as real, and the analysis predicts, correctly so far, that the assets with owners, lawyers, and lobbyists are defended as real with an energy that the assets of the unorganised cannot match. That prediction does not require a view about science; it requires only a view about how property and politics interact, which is the native territory of the tradition under examination.

Manufactured demand and the engineering of needs

Production for profit must sell what it produces, and at the scale of an advanced economy the aggregate of individual wants falls short of the aggregate of output unless wants are cultivated. Advertising, product design for obsolescence, and the fashion cycle are the visible instruments of that cultivation, and an older tradition of Marxist writing treated them as central evidence that the system’s expansion is not a response to human need but a machinery for generating need compatible with profit. Applied to energy, the argument gains a concrete edge: the throughput of the system depends substantially on goods and services whose consumption is socially manufactured, from the shortened replacement cycles of electronic devices to the marketing of ever-larger vehicles, and the energy embodied in that manufactured sector is a product of the accumulation mechanism rather than of any freely formed popular appetite.

The point deserves defence against its own caricature, because the claim is not that consumers are dupes or that pleasures are fake. A larger vehicle genuinely carries more; a new device genuinely performs; a holiday flight genuinely delivers the traveller. The claim concerns the institutional production of the demand curve: which satisfactions are developed, marketed, and built into infrastructure, and which, public space, leisure time, repairable goods, quiet streets, are left undeveloped because no accumulation strategy monetises them. A society’s energy profile is therefore partly a fossil record of which needs found investors. That formulation also clarifies why appeals to consumers to choose differently underperform: the choice set itself is produced, priced, and marketed within the same system that profits from the present selection, so exhortation asks shoppers to outvote, at the till, one of the largest industries on earth.

There is a classical root for this reasoning in the distinction between production for use and production for exchange, though the ecological application gives it a more specific form. When production is organised for exchange, success is measured by sale, and sale expands by creating new wants as readily as by satisfying old ones; the material content of the want is indifferent to the seller. Energy throughput inherits that indifference. A unit of revenue from a disposable product and a unit from a durable one count identically in the accounts, while their physical consequences diverge completely. The indifference is structural, not moral, and it is why the system can be full of ecologically conscientious managers while its aggregate metabolism expands: the accounting registers value, and value has no column for rivers.

The uneven world: industrialisation, development, and the export of emissions

Any climate analysis written from the wealthy world carries a provincialism risk, and the Marxist tradition, with its internationalist inheritance, is obliged to confront it directly. The fossil economy was built first in a small number of industrialising societies, which drew the materials of their industrialisation from a global periphery and emitted freely into an atmosphere treated as an infinite sink. Societies industrialising later inherit a world in which the sink is no longer treated as infinite and in which the demand that they industrialise differently arrives, conveniently, from the very economies that grew rich by the older method. The politics of that inheritance are unavoidable in any honest account: the global distribution of the remaining atmospheric space is a distributional conflict between nations and classes simultaneously, and analysis that treats emissions as a single planetary aggregate dissolves the conflict into a fog of shared guilt that suits the earliest industrialisers best.

Production geography compounds the problem. Wealthy economies that report decoupling have frequently relocated their most material industries abroad, importing the embodied energy and emissions of their consumption while exporting the smokestacks, a pattern the consumption-based accounting discussed earlier is designed to catch. Where the catch succeeds, some wealthy-economy progress survives; where it fails, the celebrated decoupling is revealed as relocation wearing statistical disguise. The methodological point and the political point coincide: the unit of analysis cannot be the national economy when the production system is global, and the Marxist insistence on analysing the world market as a whole, unfashionable in some decades, returns as a simple requirement of honest accounting. The relevant comparison for a wealthy economy is not between its own emissions at two dates but between its consumption footprint at two dates, and the relevant social unit for the crisis is the global organisation of production, in which classes exist within and across borders at once.

Labour enters the global picture through the supply chain. The goods consumed in wealthy economies are produced by workers whose wages, conditions, and energy systems are set in the poorer economies where production has migrated, and the carbon intensity of a product is often the obverse of its labour cheapness: production moved to where labour discipline was harshest and environmental regulation weakest, which is the labour-control mechanism of the nineteenth-century mill restated at planetary scale. The analysis therefore refuses to separate the climate question from the organisation of global labour. Cheap goods arrive carbon-heavy and labour-cheap for connected reasons, and a transition that ignores the connection will reproduce it, shifting extraction and manufacture to new peripheries under new flags. The tradition’s insistence that emancipation is international or nothing, a slogan in its party programmes, functions in the climate context as an analytical warning: national success stories that depend on displacement are not successes at the level of the system the crisis actually occupies.

How does the global organisation of production change the national picture?

It dissolves the national frame. Emissions counted at the border of a wealthy economy improve through relocation, while the consumption that economy enjoys continues to command production abroad. Classes cross borders: the owners of the relocated production, its workforce, and its consumers live under different flags. Analysis confined to one flag mistakes displacement for progress and virtue for systemic change.

Knowledge, expertise, and the professions of the fossil economy

Every energy system trains the experts who will defend it, and the fossil economy’s professional layer deserves examination alongside its asset layer. Engineers, geologists, chemists, economists, and planners built careers inside the incumbent system, and their expertise, genuine and hard-won, is calibrated to its continuation: the petroleum geologist reads strata the solar installer never studies, the grid engineer optimises for the generation profile that existed when the grid was designed, and the economist’s models embed energy assumptions invisible to the person running them. Resistance to transition inside expert communities is therefore not mainly corruption or ignorance; it is the defence of professional capital, the intellectual analogue of sunk capital, and it operates through the most respectable channels imaginable: standards committees, design codes, peer review, and the curriculum.

The point generalises uncomfortably, and the tradition that makes it must apply it to itself: Marxist expertise is also professional capital, and the mechanisms of this article have their own careers invested in them. The methodological rule that follows is symmetrical scepticism, treating every expert consensus, including this article’s preferred scholarship, as evidence to be weighed rather than as a verdict to be quoted. That rule is one reason the article states the counter-case at full strength and attributes its claims to named research; a structural analysis that exempted its own producers from structural explanation would refute itself in the act of writing.

Professions also shape what counts as a realistic proposal. The feasibility judgements that dominate public discussion are issued by communities trained under the incumbent system, whose sense of the possible was formed by its parameters: dispatchable power, centralised generation, liquid fuels, just-in-time logistics. Judgements of impossibility pronounced from within that formation are data about the formation as much as about the proposal. Historians of technology can supply the parallels, since every mature system declares the next one impractical in the accents of engineering common sense. The labour-control origin offers the standing antidote: the present system’s features were choices made under identifiable pressures, not deliverances of physics, and the feasibility judgements of its heirs deserve the same historical treatment as the choices themselves.

Petrochemicals, plastics, and the material afterlife of oil

The fossil economy does not end at the smokestack or the exhaust pipe. A substantial stream of oil and gas flows not into combustion but into materials: plastics, synthetic fibres, fertilisers, solvents, pharmaceuticals, and the chemical intermediates inside almost every manufactured object. That material afterlife matters for the analysis because it breaks the intuitive picture in which fossil fuels are simply burned and could therefore be swapped for electricity at the point of use. A plastic object is fossil carbon in solid form, produced through an industrial chain organised around refineries and crackers whose scale and integration represent sunk capital in its most concentrated form. The petrochemical complex is also where the industry’s own expectations have concentrated as transport demand is questioned: the material stream is widely treated inside the sector as its durable future, which means the asset defence adapts, migrating from fuels to materials as the debate shifts.

Waste completes the circuit. Products designed for brevity, packaging engineered to be discarded at first touch, and devices sealed against repair convert the material stream into a waste stream at unprecedented speed, and the management of that waste, its collection, sorting, export, and burial, is itself an energy-using industry staffed by some of the worst-paid labour in the world. The political economy of waste reproduces the class geography of production at the point of disposal: wealthy districts export their discards, poorer districts import and process them, and the workers who sort the world’s refuse by hand absorb its hazards. A climate analysis focused only on emissions accounts misses this material dimension entirely, yet the throughput concept, the total flow of matter and energy from extraction to disposal, was designed to catch it. What the table calls rebound has a material twin: as goods grow cheaper, their lifespans shrink, and the volume of matter processed per unit of satisfaction climbs even where satisfaction itself stands still.

The repair question shows the mechanisms in miniature. A repairable product threatens the throughput model twice, once by deferring a new sale and once by sustaining an independent repair trade outside the manufacturer’s control. The movement of design toward sealed units, proprietary parts, and software locks is therefore intelligible in accumulation terms without any assumption of ill will: durability is a cost to a seller who profits from replacement. Campaigns for repair rights exist and belong to other accounts; the analytical point is that the disposable object is not a consumer preference discovered by manufacturers but a production strategy selected by the same pressures that selected steam, namely the fit between a product and the organisation of profit. Origin explains persistence in the drawer of discarded devices as faithfully as in the mill.

Electricity systems, grids, and the architecture of supply

Electricity occupies a peculiar place in the fossil economy, at once its most fossil-dependent large system and the sector where substitution has advanced furthest. The grid is a machine for balancing supply and demand instant by instant across a territory, and its architecture embodies assumptions inherited from the age of large thermal stations: centralised generation, one-way flows, demand treated as destiny and supply as the obedient variable. Renewable generation inverts several of those assumptions at once, arriving variably, distributing itself across rooftops and fields, and asking the grid to treat demand as flexible and supply as given. The technical literature on integration is enormous and genuinely contested in its details; the social point sits beneath the technical one. The grid’s architecture is a frozen set of decisions about who controls generation and on what terms, and its redesign is an institutional contest disguised as an engineering problem.

Control is the thread that ties the grid back to the labour-control origin. Centralised generation matched the organisational preferences of the utilities and states that built it: few decision points, professional gatekeeping, capital intensity, and a public cast as passive consumers at the end of the wire. Distributed generation threatens that settlement not primarily on cost, where it now competes strongly, but on governance: millions of small producers complicate metering, planning, and the recovery of network costs, and they redistribute a measure of productive property to households and communities. Opposition framed as concern for reliability or for the poor customer’s cross-subsidy often carries, underneath, a defence of the centralised settlement, and the analysis trains its readers to ask of every technical objection the historical question: which organisation of production does this objection protect? The question does not refute the objection, since grids really must balance and poor customers really must not be exploited. It restores the suppressed premise that the existing architecture is itself a choice with beneficiaries.

Storage and demand management belong in the same frame. A system that stores energy and shifts demand is a system negotiating with time, and the terms of that negotiation, who waits, who is interrupted, whose consumption is firm and whose is flexible, are social terms. Industrial users have always sold flexibility; households are now being recruited into flexibility markets whose design will determine whether the burdens of variability fall on those best able to bear them or on those least able to refuse. Energy transitions are often narrated as substitutions of hardware. The grid case shows them to be renegotiations of command over time, which is to say, continuations of the oldest argument in this article by other means.

Energy poverty and the household end of the system

At the household end of the fossil economy stands a population for whom energy is not an abstraction but a weekly arithmetic of dread: the prepayment meter, the cold room kept unheated, the choice between cooking and warmth that recurs in the testimony of poverty research across wealthy societies. Energy poverty disciplines its subjects as effectively as any factory bell, organising the day around tariffs, top-ups, and the management of scarcity inside the home. The class analysis insists on beginning from this end of the system when questions of burden arise, because the households least responsible for the fossil economy’s structure are the most exposed to its prices, its housing stock, and its pollution, living disproportionately beside the refineries, motorways, and waste facilities that wealthier districts refuse.

That exposure pattern has a history as old as the mill towns. Industrial capital assembled workers beside its works and externalised its smoke, noise, and effluent onto their streets, and the geography of environmental harm has followed income ever since, through the siting of incinerators, the routing of freight, and the distribution of green space. Environmental justice research, a literature largely independent of Marxism, documented the pattern empirically across decades, and the two traditions converge on the structural conclusion: the burdens of the metabolism are allocated by power, not by accident or by the preferences of the burdened. A climate politics that imposed transition costs through the same allocation, raising household energy prices while leaving asset holders whole, would reproduce the founding injustice in green colours, and the analysis predicts the ferocious public resistance such an allocation reliably provokes. The resistance is not ignorance of science. It is a rational defence by populations who have learned, from long experience, exactly how costs are distributed when distribution is left to the market.

Household arithmetic also reframes the consumer question one final time. The aggregate emissions attributed to households in national accounts are dominated by the necessities of housed, heated, transported, fed existence inside systems designed upstream. Moralising that aggregate confuses an accounting category with a decision maker. The decisions that bind households, the stock they inhabit, the distances they must travel, the appliances on sale, were taken in boardrooms, planning departments, and legislatures. Energy poverty is the extreme case of a general condition: the consumer’s sovereignty extends over the last few percent of a chain whose every earlier link was forged elsewhere, and an analysis honest about causation distributes its attention accordingly.

Recession, crisis, and the emissions that follow the business cycle

One grim regularity in the emissions record functions as a natural experiment for the accumulation mechanism: when capitalist economies contract, emissions fall, and when they recover, emissions resume their climb. No policy achieves that reduction; the treadmill slows because production slows, freight idles, construction halts, and households retrench. The correlation states the structural thesis in its purest available form, since it shows emissions tracking the accumulation process itself rather than tracking awareness, virtue, or technology, all of which continue improving through the downturn while emissions obediently follow output down and then back up. Defenders of the analysis cite the pattern as its most economical proof. An economy whose emissions fall only when its people are being impoverished is an economy whose emissions are a function of its metabolism, and every crisis that demonstrates the function also demonstrates the human cost of achieving reduction by contraction, which is precisely the outcome the analysis exists to distinguish from a planned and provisioned transition.

Critics answer, correctly, that correlation across a cycle does not settle the long-run question, since the decoupling evidence concerns trends across decades rather than oscillations across quarters. The distinction deserves its full weight: cyclical sensitivity proves that emissions remain coupled to output in the short run, while the trend evidence assembled by Le Quéré and collaborators suggests the coupling can loosen across longer spans in favourable conditions. Both can be true, and the interesting analytical work lies in specifying the conditions: which sectors decouple, under which policies, at what speed, and with what dependence on one-off fuel switching that cannot be repeated. That specification is where the debate inside energy economics lives, and a reader who can state the cyclical fact and the trend qualification in a single breath understands more than most participants in the shouting version of the argument.

The crisis pattern also illuminates the treadmill’s human wiring. Downturns reduce emissions by destroying livelihoods, and the populations harmed first are those with the least buffer, so the system’s only demonstrated brake is one nobody can defend pulling. Any serious climate politics is therefore a search for a second brake: a way of reducing throughput that does not pass through unemployment, eviction, and hunger. The Marxist tradition’s contribution to that search is the insistence that provision be detached from accumulation, that incomes, housing, and care be secured by arrangements that do not require the treadmill to accelerate in order for people to eat. Whether such arrangements are built is a political question outside this article’s remit. That the search is necessary follows from the cyclical evidence alone, and the evidence is visible in every downturn’s emissions inventory.

Ecological modernisation: the rival theory stated fairly

The strongest theoretical rival to the Marxist account is not denial or delay but ecological modernisation, the research programme associated with scholars such as Maarten Hajer and Arthur Mol, which holds that environmental reform can proceed through the further modernisation of existing institutions: cleaner technology, smarter regulation, market instruments, and reflexive governance that learns from its own monitoring. The programme points to genuine achievements in wealthy democracies, where certain pollutants were drastically reduced, rivers recovered, and industrial processes redesigned under regulatory pressure without any transformation of property relations. Sulphur, lead, and urban smog furnish its case studies, and the case studies are not fabricated; they demonstrate that capitalist states can force industries to internalise specific, bounded harms when the technical fixes are cheap, the culprits are identifiable, and the asset stock at stake is modest.

The Marxist reply begins by granting the case studies and then examining their common structure. Each success concerned a pollutant peripheral to the energy metabolism itself: an additive, an emission at the margin of a process, a waste stream that could be captured without redesigning the productive core. Carbon is different in kind rather than in degree, on this argument, because carbon is the metabolism: the overwhelming mass of fossil throughput passes through the centre of production, transport, heat, and food, and the asset stock implicated is not a scrubber on a smokestack but the smokestack, the mine, the fleet, and the suburb together. Ecological modernisation’s reply is equally clear: peripheries expand, the core of one era becomes the margin of the next as technology advances, and the cost evidence assembled above shows the centre itself becoming tractable. The exchange is the live frontier of the debate, and this article presents it as such rather than awarding a verdict its own framework has not earned.

How does ecological modernisation differ from the account given here?

Ecological modernisation expects reform from modernisation itself, as cleaner technology, learning institutions, and market signals discipline production from within. The Marxist account expects retardation by competition, capitalised value, and rebound, so reform arrives late and unevenly, sized to what the asset stock tolerates. Both predict cleaner production; they differ on pace, distribution, and where the binding constraint sits.

A reader grading the two programmes should ask what each treats as the residual problem when its preferred instruments have done their work. For ecological modernisation, the residual is governance quality: better monitoring, better design, better politics. For the Marxist account, the residual is property: the stock’s owners must still be bought, beaten, or waited out, and the waiting is measured in the lifetimes of assets rather than of parliaments. History will grade both, and is already grading them sector by sector, which is why the sectoral pattern, rapid movement in electricity, slow movement in freight and heat, functions in this article as the shared evidence both programmes must digest. Neither programme is refuted by a mixed pattern; each is embarrassed by a different half of it, and honest exposition says which half embarrasses whom.

Pricing instruments, described rather than recommended

Few topics in climate policy generate more heat per unit of light than carbon pricing, and an article bound to recommend no policy can still perform the analytical service of describing what the instrument is and how the mechanisms in the table predict its career. A carbon price attaches a cost to emissions at the point of combustion or production, allowing each emitter to decide whether to pay, abate, or pass the cost along. Its theoretical attraction is informational: one signal propagates through millions of decisions without a planner adjudicating any of them. Its practical history, across the jurisdictions that have adopted versions of it, shows the signal arriving weakened by exemptions, free allocations, and price ceilings, and being revoked or frozen when its incidence on household fuel costs becomes politically legible. The four mechanisms predict exactly that career. Competition buys the exemptions; sunk assets bargain for the free allocations; the treadmill coalition of threatened workers and regions supplies the crowds; and the household burden, allocated by price without regard to income, detonates the politics.

None of that makes the instrument useless, and the analysis, properly stated, does not claim uselessness. Prices demonstrably shift behaviour at the margin, and the decomposition studies behind the decoupling evidence credit pricing among the instruments that worked. The structural critique is about sufficiency and incidence, not about existence: a single price signal asked to retire a defended asset stock, reorganise settlement patterns, and protect the poorest households simultaneously is being asked to do the work of an entire state while armed with a tax schedule. Writers in the Marxist tradition who dismiss pricing entirely overstate their case exactly as writers who treat pricing as sufficient overstate theirs. What the tradition contributes is the insistence that the instrument’s career be analysed politically, as a negotiation among the interests the table names, rather than technically, as a calibration problem awaiting the correct number.

There is also a distributional asymmetry inside the instrument that deserves the tradition’s vocabulary. A price on carbon is a price on the necessities of life for households with no slack, and a rounding error for households with much. Revenue recycling schemes answer that asymmetry on paper, returning the proceeds as dividends or targeted support, and their design determines whether the instrument functions as a transfer or as a extraction from the bottom of the distribution. Which design emerges from a legislature is again a question of organised interests rather than of economic logic, and it is the question on which real pricing regimes have repeatedly foundered. The analytical lesson generalises: in climate policy, incidence is destiny, and instruments are adopted or destroyed according to who is made to pay, which is the oldest theorem in political economy wearing new clothes.

The nineteenth-century witnesses: coal smoke and the condition of the towns

The fossil economy’s social character was visible to its first generation of critics, and returning to their testimony disciplines the temptation to treat the climate analysis as an invention of environmentalism. Friedrich Engels, sent to manage a family firm in the greatest cotton city of the age, published his account of the working class in 1845, and its descriptions of the industrial districts remain the founding documents of urban environmental description: courts without drainage, cellars inhabited below the level of flooded streets, air thickened by the smoke of hundreds of chimneys, rivers converted into open sewers by the dye works and tanneries along their banks. The book’s argument was not about pollution as an unfortunate side effect. Engels insisted that the squalor was the predictable product of a system that assembled labour for production and accepted no responsibility for the conditions of its reproduction, a formulation that stands behind the whole later tradition of social reproduction and environmental class analysis alike.

Notice the structure of that early critique, because it prefigures the argument of this article almost exactly. The harm is not attributed to the personal callousness of individual mill owners, several of whom Engels treats as comparatively decent men trapped in the same competitive machinery. The harm is attributed to the relations: to competition that forbids any owner to spend on conditions his rivals skip, to property forms that sever the factory’s balance sheet from the street outside its gate, and to a state whose local organs were staffed by the very interests requiring regulation. Replace smoke with carbon and the Victorian street with the planetary atmosphere, and the analytical skeleton is unchanged, which does not prove the modern argument but does establish its pedigree. The tradition did not borrow an environmental concern late in its life; it began, in one of its founding texts, as a description of what industrial energy does to the people housed beside it.

Marx’s own treatment of machinery belongs in the same archive. In the chapters of Capital devoted to machinery and modern industry, the analysis of the factory is simultaneously an analysis of power, in both senses: the motive force that drives the line shaft and the social force that subordinates the worker to the machine’s rhythm. Marx observed that machinery, in the form capital gives it, confronts the worker as an alien power, and that the technical composition of production, the ratios of machinery to labour, is chosen under the pressure of the wage relation rather than by engineering logic alone. Malm’s fossil capital argument is, in that light, less an innovation than a recovery: it asks of the prime mover the question Marx asked of the machine. Who chose this configuration, under what compulsions, and whose command does the configuration embody? The continuity explains the speed with which ecological Marxists recognised the steam research as their own, and the equal speed with which critics recognised it as a target worth answering.

What did the earliest critics of the factory system blame for its squalor?

Engels’s 1845 account blamed relations rather than characters: competition that punished any owner who spent on conditions his rivals neglected, property forms that severed the factory’s accounts from the streets beyond its gate, and government staffed by the interests needing regulation. The squalor of the districts appears in his pages as an outcome produced by those relations, not as accident or failure of manners.

The population question, refused in full

No misreading of the climate analysis is more persistent than the suspicion that it is a population argument in disguise, and the refusal deserves a section of its own. Aggregate human numbers explain almost nothing about emissions, because emissions per person vary across societies by factors so large that the aggregate conceals the entire phenomenon: a population with universal electricity, heated housing, motorised transport, and industrial diets emits on a scale that a far larger population without those systems does not approach. The variable doing the work is the production system a population is inserted into, which is why the analysis consistently substitutes organisation for enumeration. Where population arguments direct attention toward the fertility of the poor, the class analysis directs it toward the investment decisions of the rich, and the political difference between those directions is the whole ballgame.

The tradition’s historical reasons for refusing the population frame are equally specific. Malthusian reasoning entered political economy as an argument against poor relief, attributing poverty to the reproductive habits of the poor rather than to the distribution of property, and Marx’s demolition of Malthus, repeated across his work, treated the doctrine as an apology dressed as arithmetic. Ecological writing that revives the frame, however green its intentions, inherits that function: it relocates causation from the owners of the productive system to the bodies of the least powerful, and it reliably licenses coercion against those least able to resist. The refusal is therefore not squeamishness. It is a conclusion drawn from watching the argument operate for two centuries, and it is one of the points on which the tradition’s judgement has aged best.

A subtler version of the same question concerns consumption growth in the developing world, and here the analysis must be as clear with its friends as with its critics. Rising consumption among populations that lack sanitation, refrigeration, and mobility is not a climate pathology to be preached against by the already comfortable; it is development, the substance of every promise the modern economy makes to the poor. The analysis locates the crisis in the form that production takes and in the compulsion that drives its expansion where needs are already met many times over. Conflating the two kinds of growth, the growth of provision for the unprovided and the growth of throughput for the saturated, is the characteristic error of aggregate thinking, and correcting it is among the services a class framework performs for a debate perpetually tempted by undifferentiated totals.

Grading the mechanisms sector by sector: a worked exercise

The four mechanisms earn their keep when they are forced to grade unlike sectors and survive the comparison, so this section performs the exercise in the open. Take electricity generation first. Accumulation pressure is high, since generators compete on cost, yet substitution has advanced furthest here, because the substitute technologies became cheap, the product is homogeneous (a unit of electricity is a unit of electricity), and the organisational form of the sector, centralised plants feeding a grid, survived the substitution intact. Sunk capital resisted, as predicted, strongest where the fleet was youngest, but each retired plant weakened the coalition defending the next. Rebound operates weakly at the point of generation and more strongly through the uses cheap electricity enables. The sector therefore reads as the objection’s best evidence and the analysis’s limiting case simultaneously: given a cheap substitute that fits the existing organisation, the system’s own competitive mechanism becomes the transition’s engine.

Freight inverts nearly every term. The product is not homogeneous in the relevant respects, since shippers buy speed, reliability, and network reach as much as carriage; the fuel’s energy density underwrites the entire logistics architecture of just-in-time production; the asset stock, of vehicles, vessels, aircraft, ports, and warehouses, is enormous and interlocked; and rebound is notorious, since every fall in the cost of moving goods has historically expanded the distance goods are moved, dissolving production into global chains whose length is itself an energy decision. Accumulation pressure, far from driving substitution, drives the expansion of the network and the deepening of its fossil dependence. The sector reads as the analysis’s best evidence, and the honest expositor should say plainly that the two sectors together produce the mixed verdict the whole article defends: the system transforms what fits and defends what does not, and the boundary between the two is drawn by capitalised value and organisational fit rather than by the state of public concern.

Buildings sit between the poles and repay the closest study, because their inertia is neither technological nor exactly financial. Substitute technologies for heat exist and perform; costs are falling; yet the stock turns over at geological pace, the decision makers are fragmented among millions of owners, the split incentives between landlords and tenants block the obvious investments, and the disruption of retrofit falls on occupants who did nothing to design the problem. The mechanisms distribute the explanation among themselves: sunk capital in its most fragmented form, accumulation pressure nearly absent (a dwelling does not compete), rebound present in the temptation to heat more rooms once heating is cheap and clean, and the labour-control origin visible in negative, since the difficulty is precisely that nobody commands the stock. A sector where the binding constraint is the absence of a commanding owner is a useful embarrassment for any crude version of the theory and a useful confirmation of the sophisticated one, which predicts that transitions proceed where an organised interest can capture the returns and stall where returns are as diffuse as the stock.

Agriculture, examined earlier through its inputs, grades differently again: substitution is partial, the biological calendar resists rescheduling, and the treadmill coalition includes the most politically protected producers in most democracies. The exercise’s cumulative lesson is methodological. Grand verdicts on capitalism and climate, in either direction, dissolve under sectoral grading into a set of local judgements with local evidence, and the four mechanisms function best as grading instruments rather than as a creed. That is a deflationary conclusion for an article in a series about a grand tradition, and it is offered deliberately, because the tradition’s own best work, from the working-day chapter onward, proceeded by exactly such patient disaggregation.

Reading the primary literature: a short guide to the shelf

Readers who want to test this article against its sources can do so with a manageable shelf. Begin with Malm’s Fossil Capital for the founding case, reading it against the cost-based replies of the economic historians so the contest is visible rather than curated. Jevons’s study of the coal question, a nineteenth-century work whose author worried about exhaustion rather than emissions, supplies the rebound mechanism in its original and surprisingly modern form. Schnaiberg’s treadmill writings give the sociological coalition its classic statement, and Unruh’s work on lock-in gives the institutional mechanism its technical grounding. For the counter-case, Nemet’s study of solar costs is the indispensable text, both for its evidence and for its method of attributing a cost decline among interacting causes; the decoupling literature associated with Le Quéré and collaborators should be read in its consumption-based form, where its claims are strongest and its methods most exposed to scrutiny. Ecological modernisation is best approached through its own theorists, Hajer and Mol among them, rather than through hostile summaries.

Reading strategy matters as much as the list. Structural arguments are most honestly tested at their joints: the places where a mechanism is asked to explain a case it was not designed for. Does the labour-control reading illuminate oil, whose adoption story differs from coal’s? Does lock-in explain the failure of transitions nobody resisted? Does rebound survive contact with sectors where consumption is saturated? A reader who carries those questions through the shelf will finish with a graded, personal version of the analysis, which is a more durable possession than agreement with any summary. The summaries, this one included, are ladders; the scholarship is the wall.

Seminar leaders might also note the generational layering of the shelf. Jevons wrote as a Victorian political economist addressing scarcity; the treadmill and lock-in literatures wrote as late twentieth-century social scientists addressing institutions; Malm and the decoupling researchers write as contemporaries addressing a crisis whose scientific framing they inherit rather than construct. Watching the same system described across those formations, each generation discovering the mechanisms its predecessor missed, is itself an education in how social knowledge accumulates: not by refutation alone, but by the slow addition of angles, until the object stands in something like the round.

Organised labour between the wage and the atmosphere

The labour movement’s position inside the climate argument is structurally tragic, and the analysis loses all credibility when it treats that tragedy as a failure of consciousness. Unions organise workers as sellers of labour power inside existing industries; their first duty runs to the wages, safety, and continuity of the members in front of them. When a transition threatens a carbon-intensive plant, the union defending the plant is performing its function exactly as designed, and the workers voting for that defence are protecting mortgages, pensions, and towns, not expressing an opinion about atmospheric chemistry. The treadmill framework predicted this alignment: employment dependence enlists labour as the growth coalition’s most sympathetic partner, because no other partner’s stake is as humanly legible. Any politics that treats those workers as obstacles has misunderstood the structure it claims to analyse and will be defeated by it, deservedly.

History offers the tradition its own counter-examples, and they matter because they show the alignment is structural rather than natural. Building workers in one celebrated construction boycott movement refused environmentally destructive projects at the cost of their own employment, once their organisation had been persuaded that the union’s power could defend a wider definition of members’ interests, including the city they lived in. Dock and transport workers have at times refused cargoes on political grounds. Such episodes share a precondition: the workers possessed organisation strong enough to act beyond the immediate wage bargain and leadership capable of framing the longer interest. Where those conditions are absent, the treadmill alignment holds by default. The lesson for the analysis is that labour’s position is a variable, determined by organisation and by the credibility of alternative provision, not a constant to be assumed by either camp.

The phrase just transition, whatever its career in policy documents, names in this structural vocabulary a precise bargain: the workers whose sunk skills and regional economies a transition devalues must be made whole by deliberate provision, or they will rationally defend the stock that employs them, and their defence will lend the asset owners a human shield. Asset owners understand the shield’s value and have historically been willing to fund its rhetoric. The class analysis therefore evaluates every transition proposal by a single prior question: who is made whole, in what order, and with whose money? A proposal that compensates capital while lecturing labour has the order backwards, and the tradition’s entire history entitles it to say so. Whether any given proposal passes the test is a judgement this article leaves open, since it recommends no policy; the test itself, however, follows from the mechanisms as directly as any prediction in the table.

Status, position, and the consumption that competition manufactures

One sector of emissions resists every efficiency improvement aimed at it, because its consumption is positional: the good consumed is not the service but the rank the service displays. Larger vehicles, larger houses, more distant holidays, and faster replacement of visible goods all carry a positional premium, meaning their value to the purchaser depends on comparison with the purchases of others. Positional competition has no satiation point, since the reference group upgrades alongside the individual; it is, structurally, an arms race conducted through the shops. Climate analysis that ignores position will perpetually misread the demand side, expecting incomes to plateau into contentment when the social mechanism guarantees escalation instead. The Marxist tradition has a native vocabulary for this, since the critique of commodity fetishism was always partly an account of how social relations disguise themselves as properties of things, yet the positional mechanism deserves statement in plain terms: a share of the fossil economy’s throughput is the exhaust of a status competition that no participant can exit alone, because exit is itself a positional loss.

Positional demand also clarifies a distributional asymmetry that aggregate figures hide. The households competing at the visible frontier of consumption are drawn from the upper reaches of the income distribution, and their escalation sets the standards that cascade downward through marketing and credit, so that even households far from the frontier consume inside a built environment, a vehicle fleet, and a holiday geography designed at the frontier’s specifications. Blame, once again, distributes badly when aimed at the individual shopper: the shopper did not set the frontier, does not control its movement, and faces real social penalties for falling visibly behind it. The mechanism is nobody’s fault and everybody’s trap, which is the signature of a structural explanation and the reason the tradition keeps returning to structures when moral vocabularies run out.

Advertising’s role in position deserves one further distinction. The industry does not merely inform consumers of goods; it manufactures the comparisons themselves, attaching products to ranks, ages, and identities with a sophistication that constitutes one of the great cultural industries of the age. Its own account of its function, persuasion at scale, is not in dispute. What the analysis adds is the connection between that industry and the energy system: the throughput of a consumer economy is partly the product of a persuasion economy whose output is dissatisfaction, calibrated and renewable. A transition politics addressed only to supply, to fuels and power stations, leaves the dissatisfaction factory running at full capacity, and should not be surprised when demand reconstitutes itself around whatever the factory promotes next.

Why the forecasters kept missing the transition they were measuring

An under-examined episode in the recent history of energy deserves a section, because it functions as a controlled test of the knowledge mechanism. Across the decades in which low-carbon technologies collapsed in cost, the official forecasting agencies of the energy world, staffed by competent specialists with unrivalled data, underestimated the speed of the collapse in forecast after forecast, revising each year’s projection upward while keeping the curve’s shape conservative. Nemet’s research and the wider innovation literature treat that pattern as a puzzle about models: linear assumptions applied to exponential learning, and institutional caution inside agencies whose constituencies included the incumbents being displaced. The structural reading adds the professions argument developed earlier. Forecasters formed inside the incumbent system absorbed its sense of the possible, and their errors were not random; they leaned, with metronomic consistency, in the direction that flattered the existing order.

The episode cuts in two directions, which is why it belongs in an honest article. Against complacent Marxism, it shows the system innovating faster than its own priesthood believed possible, a reminder that accumulation’s dynamism is real and that any analysis freezing the system into a static caricature will be refuted by the system’s own engineers. Against complacent modernisation theory, it shows the official knowledge of the system systematically misdescribing the system’s trajectory in the incumbents’ favour, a reminder that expertise is produced inside relations and carries their watermark. Both reminders are worth the price of the section. A reader who absorbs them will treat every confident projection in this debate, from whatever camp, as a social artefact requiring the same question the labour-control origin teaches: whose organisation of production does this account of the future protect?

Innovation systems research supplies the constructive counterpart. Studies of how new technologies actually mature describe protected niches, public procurement, patient finance, and learning networks, a machinery of support that the successful transitions all used and that pure market accounts omit. The machinery is neither capitalist nor socialist in itself; it is the standard equipment by which societies have always midwifed new productive forces, from canals to semiconductors. The analytical payoff for this article is modest and firm: the observed speed of the low-carbon cost decline is evidence about what organised support can do inside market economies, not evidence that the unaided price mechanism performed a miracle. Both the objection and its answer must therefore be restated in the vocabulary of organised support, and the debate becomes, properly, a debate about which interests the organisation serves and which assets it is willing to strand.

The home as a workplace: energy and unpaid labour

One final territory repays the analysis, because it joins the climate question to the social reproduction themes treated elsewhere in this series. The household runs on energy twice over: once as purchased fuel and electricity, and once as the unpaid labour, overwhelmingly performed by women, that converts fuel and ingredients into meals, warmth, cleanliness, and cared-for people. Fossil energy entered the home as a labour-saving promise, and the appliances that burned or consumed it genuinely abolished whole categories of drudgery: hauling water, splitting fuel, laundering by hand, cooking over managed fires. Any account of the fossil economy that cannot acknowledge that emancipation, partial and uneven as it was, has no business lecturing the populations who received it. The drudgery was real, its relief was real, and the relief helps explain the depth of popular attachment to the energy system that delivered it.

Yet the domestic settlement contains the rebound mechanism in its most intimate form. Labour saved by appliances was recaptured by rising standards: laundry done more often, rooms heated throughout, cleanliness norms escalated until the hours returned in new configurations, a pattern feminist historians of technology documented long before energy analysts noticed it. The household thereby reproduces the system’s signature in miniature: productivity gains converted into expanded output rather than into free time. Standards, like status, are socially set and competitively enforced, among neighbours as among firms. The parallel is not a metaphor but a genuine structural rhyme, and it warns against imagining that the expansion compulsion lives only in boardrooms. It lives wherever output standards are set by comparison and enforced by respectability, which is to say it lives in the texture of ordinary life.

The policy-free conclusion the analysis draws is about attention. Energy debates centred on power stations and vehicles miss the household as a site of production, and therefore miss both a share of demand and the entire politics of care that will decide whether transition measures feel like liberation or like a new round of imposed scarcity in the home. Populations asked to accept colder rooms and longer journeys while asset holders are compensated will draw their own conclusions about the order of sacrifice, and the history of energy politics suggests those conclusions arrive at the ballot box with interest.

Subsistence and luxury: the moral anatomy of a tonne

The philosopher Henry Shue introduced a distinction that the class analysis absorbs naturally: between subsistence emissions, generated in securing the necessities of life, and luxury emissions, generated in securing comforts and status beyond necessity. The distinction is moral in form but structural in consequence, because it assigns the burden of any reduction in an order exactly inverse to the order in which aggregate accounting assigns it. A tonne emitted heating a single room in a cold climate, by a household with no alternative system available, and a tonne emitted heating a little-used second property are identical in the atmosphere and opposite in every other respect that matters for judgement. Analysis that sums them into national totals and then prescribes uniform restraint has performed an arithmetic operation and mistaken it for an ethics.

Shue’s distinction also protects the development claim discussed earlier from its crudest critics. The emissions of a population acquiring electricity, sanitation, and transport for the first time belong overwhelmingly to the subsistence side of the ledger in any honest application of the terms, while the emissions of saturated economies are weighted toward the luxury side, including the vast category of positional consumption examined above. Climate argument that ignores the distinction will therefore always sound, to most of the world, like a demand that the poor freeze in the dark so that the rich may continue. The Marxist tradition’s contribution is to give the distinction a materialist grounding: subsistence and luxury are not properties of individual souls but of positions in a production system, and the same household can be driven to luxury-pattern consumption by the built environment and the status competition it did not design.

There is a harder edge beneath the ethics, concerning bargaining. In any negotiation over a shrinking global budget of emissions, the subsistence claim is the strongest card held by the poorer parties, and its strength is precisely moral: no settlement that begins by taxing the heating of the poor will be accepted as legitimate by the populations asked to sign it, whatever its aggregate arithmetic. The analysis therefore predicts that distributional ordering, who reduces first and who is compensated, will dominate the politics of the crisis at every scale from the household budget to the international conference table, and that technical optimal paths which ignore the ordering will fail on contact with legislatures. That prediction requires no position on climate science; it requires only the observation that legitimacy is a material force in politics, a theorem the tradition has maintained since its founders watched revolutions fail for want of it.

From coal to oil: the second adoption and its lessons

The fossil economy renewed itself once already, when oil displaced coal across transport and much of heating, and the second adoption tests the labour-control pattern against a new case. Oil won its domains on properties that map closely onto the old criteria. It was energy-dense and liquid, hence transportable, storable, and dispensable by the unit, freeing vehicles from fixed infrastructure and fixed fuelling points as steam had freed mills from fixed streams. It suited decentralised, mobile, individualised consumption, above all the private vehicle, and it weakened specific labour chokepoints: the stokers and coal heavers whose collective position had given them leverage over the steam economy found no equivalent place in a system of pipelines, tankers, and pumps. Historians of energy have noted how the transition altered the geography of labour militancy, dissolving some of the most strike-prone occupations in the industrial world. Whether that effect was a motive or a windfall is debated in the same terms as the cotton case, and the debate’s recurrence is itself the point: in each adoption, a social advantage accompanies the technical one, and the pattern of advantage shapes the politics of the following century.

Oil also globalised the asset problem in a new way. Coal economies were substantially national, dug and burned within the same broad territories, while oil’s geography separated extraction from consumption across oceans, creating chokepoint politics, strategic reserves, and the permanent entanglement of energy with foreign policy. The sunk-capital mechanism acquired a military-strategic dimension: states defend oil infrastructure not only as value but as security, and the security framing immunises the asset defence against ordinary economic argument. A minister who might shrug at a balance sheet cannot shrug at a briefing about supply vulnerability. The analysis therefore expects oil’s political defence to outlast its economic rationale by the widest margin of any fossil fuel, and expects the defence to be conducted in the vocabulary of security rather than of profit, a prediction the public rhetoric of energy debates confirms so routinely that its confirmation attracts no notice.

The second adoption carries one consoling implication that the analysis should not suppress. Energy systems have been replaced before, completely and within living memory of the societies that replaced them, when the successor offered decisive advantages on the criteria the adopters actually used. The obstacle revealed by the two adoptions is not that replacement is impossible but that replacement follows the adopters’ criteria, which are social before they are thermodynamic. A successor that wins on cost alone while losing on control, convenience, and organisational fit will wait in the laboratory and the pilot plant, as water’s virtues waited in the valleys while the towns chose steam. A successor that wins on the full set of criteria spreads with a speed that surprises every forecaster, as the cost literature’s subjects did. The analysis, read carefully, is a manual for telling the difference.

Method in practice: how a historical materialist reads an energy system

Step back from the case material and the method itself can be stated as a sequence of questions, portable to any energy system a reader encounters. Ask first what the system produces and for whom, since the pattern of outputs reveals the pattern of command. Ask second how its technologies were selected, and by whom, and against which alternatives, since the adoption moment preserves the selection criteria in recoverable form. Ask third where the capitalised value sits, and whose incomes, pensions, and public revenues depend on it, since that map predicts the geography of defence. Ask fourth where the compulsion to expand operates, and what happens to any participant who refuses it, since the answer distinguishes a system that could rest from a system that cannot. Ask finally what the system does to time: whose schedule it serves, whose hours it lengthens, whose waiting it abolishes and whose it imposes. A reader who puts those five questions to a national electricity system, a freight network, or a food chain will produce a recognisably Marxist analysis without quoting a line of Marx, which is as it should be, since the method was always meant to be a practice rather than a liturgy.

The method’s discipline is as important as its questions. Each question directs attention to evidence that exists independently of the analyst’s wishes: archives of adoption decisions, balance sheets, labour market records, the testimony of participants, the physical layout of the built environment. Where the evidence contradicts the expected pattern, the method instructs the analyst to revise, as the decoupling evidence has instructed this tradition to revise its impossibility claims. Historical materialism degenerate in the hands of its custodians becomes a vocabulary for redescribing conclusions reached in advance; in its functioning form it is a research programme with unusual explanatory reach and ordinary obligations to fact. This article has tried to exhibit the functioning form, including its concessions, because the concessions are part of the method and not decorations upon it.

One more methodological commitment deserves naming, since it separates the analysis from the two genres that surround it. Advocacy writing selects evidence for its effect on a conclusion already chosen; engineering writing brackets the social relations as externalities to its models; the analysis practised here treats the social relations as the object and the physical system as their expression, while holding its own claims to the evidential standards of the first genre’s opponents and the second genre’s precision. The hybrid is uncomfortable to write and to read, and the discomfort is diagnostic: the climate crisis sits exactly at the junction where the genres fail separately, which is why a literature that works the junction, however imperfectly, continues to attract readers whom neither pamphlets nor models can satisfy.

The demand side of the state: public consumption and strategic appetite

States are not only regulators and owners in the fossil economy; they are among its largest direct consumers, and their appetite has a structure worth isolating. Armed forces burn fuel on a scale no other single organisation matches, and their procurement sets technical standards, underwrites extraction and refining capacity, and defines the security interest that protects the whole complex from ordinary scrutiny. Public construction pours the concrete and lays the asphalt of the road system; public fleets, buildings, and logistics consume steadily in every year regardless of the business cycle. The state’s demand is therefore countercyclical ballast for the fossil economy, a guaranteed market that operates when private demand falters, and the analysis counts it among the mechanisms of persistence even though it appears in no row of the table: the customer of last resort is also a producer of lock-in.

Secrecy compounds the effect. Military consumption is the least transparent category in national energy accounts, exempted from many reporting conventions and shielded by classification, so the public picture of national emissions systematically omits the fuel burned by the institution most completely organised by the state itself. Analysts who reconstruct the omission describe a consistent pattern rather than a conspiracy: the state’s coercive arm consumes without accounting because accounting is a civilian discipline. For the class analysis, the episode is a parable about visibility. The parts of the fossil economy most tightly fused with state power are the parts least available to democratic inspection, and any politics of transition that begins from the published accounts begins with a flattering fiction about who consumes what.

Procurement, finally, is a lever whose existence the analysis notes without prescribing its use. A buyer of the state’s scale shapes markets by specification: vehicle standards, building codes for public estates, fuel contracts, and infrastructure designs all teach industries what to produce. Whether that lever is pulled toward transition or toward incumbency is a political outcome, and the historical record shows it pulled in both directions by different administrations facing different coalitions. The structural point stands independent of the record: the state is inside the market it regulates, as customer, owner, and guarantor, and accounts of climate politics that cast it as an external referee misdescribe the game from the first whistle.

Coal regions, identity, and the politics of a threatened world

The fossil economy is not evenly spread across the map; it concentrates in regions whose entire social fabric was woven around extraction. Coal valleys, oil towns, and refining coasts developed cultures as distinctive as any in the industrial world: occupational pride transmitted across generations, union traditions, sporting and musical institutions, and an ethic that dignified dangerous labour as the foundation of national prosperity. When transition arrives in such a region, it arrives not as an adjustment of fuel mix but as a threat to a civilisation in miniature, and the resistance it meets is cultural and existential before it is economic. Compensation formulas that price the lost wages and ignore the lost world systematically underestimate what is being defended, and the political explosions that follow are then misread as ignorance or manipulation by commentators who never priced the world either.

The tradition should be the last to make that misreading, since its own classics are saturated with the dignity of mining communities and the solidarity their danger produced. A Marxism that celebrates the coal miner as the archetype of the industrial worker in one chapter and treats coal regions as an emissions problem in the next has divided its own inheritance against itself. The coherent position, and the one the mechanisms support, is that the defence of a coal region is the sunk-capital mechanism operating at the level of a whole society: the capital at risk includes the cultural and organisational capital of the community, its chapels and clubs and accumulated mutual aid, none of which appears in the asset registers and all of which its inhabitants will fight to preserve. Transitions that honour that capital, carrying institutions across rather than dissolving them, face a different politics from transitions that arrive as liquidation notices written in a distant capital.

Regional concentration also explains the electoral geometry that so often frustrates national climate majorities. The costs of transition fall on identifiable districts with names, accents, and representatives, while the benefits diffuse across the whole electorate and the whole century. Political systems weight the concentrated and the present over the diffuse and the future as a matter of structure, not of anyone’s bad faith, so fossil regions punch above their demographic weight in every legislature. The analysis predicts that the politics of transition will be fought region by region and plant by plant rather than in a single national reckoning, and that prediction, like the sectoral one, has the dreary merit of matching the observed record.

The skies and the seas: freight’s far domains

Long-distance freight by sea and air concentrates every difficulty the table contains, and a section devoted to it tests the mechanisms at their extreme. Shipping moves the overwhelming bulk of world trade on vessels whose working lives span decades, powered by the cheapest and dirtiest fraction of the oil barrel, through jurisdictions designed to fragment responsibility among flag states, owners, charterers, and cargo interests across every maritime nation at once. The industry’s structure is a masterpiece of the accumulation mechanism’s logic: competition so intense and margins so thin that no operator can absorb a cost its rivals escape, regulation so international that no single state can impose one, and an asset stock so long-lived that the fleet embodies the fuel assumptions of a previous generation. If a structural account of emissions has a showcase, it is a bulk carrier.

Aviation presents the mirror image: a young industry by capital-stock standards, yet locked in by physics and by the positional economy. Flight’s energy density requirements exclude most substitutes on any near horizon the engineers describe, while its demand is driven by the status competition and business practices of the most mobile class in history. Efficiency in aviation is the Jevons paradox’s most faithful servant: each generation of aircraft burns less per seat and the system responds by selling more seats to more destinations at lower prices, so that the sector’s total consumption climbs through every improvement. The honest treatment of aviation in climate writing requires holding two facts that partisans separate: the sector is a small share of the present total, and it is among the fastest-growing and least substitutable, disproportionately consumed by the wealthiest fraction of humanity. Class analysis exists for such facts, since the sector is nearly useless as transport for the world’s majority and nearly indispensable as a marker of membership in its managerial stratum.

Both domains expose the limits of national accounting one final time. International bunkers, the fuels of ships and aircraft between jurisdictions, belong to no national inventory, a diplomatic arrangement that leaves the fastest-moving emissions in a statistical no man’s land. The omission is not a technicality; it is the world market declining to be measured by the units of its member states, and it vindicates the article’s insistence that the system, not the nation, is the unit the crisis occupies. A tradition formed by analysing the world market as a single, uneven whole finds in the bunker fuels a small, perfect confirmation of its oldest methodological habit.

Overaccumulation and the search for outlets

Beneath the four mechanisms runs a deeper current in the tradition’s economics that deserves surfacing, because it connects the climate argument to the theory of crisis. Capital, on the classical Marxist account, periodically accumulates beyond the profitable outlets available to it, and the surplus seeks discharge in new spheres: new territories, new industries, new forms of consumption, and the built environment itself. Energy infrastructure has historically been among the grandest of those outlets, absorbing enormous masses of capital into mines, railways, grids, dams, and motorways whose construction employed armies of labour and whose returns unfolded across generations. The fossil build-out was therefore not merely a response to energy demand; it was one of the principal forms in which an expansionary system housed its own surplus, and the stock it left behind is correspondingly vast.

That genealogy sharpens the sunk-capital mechanism with an uncomfortable corollary. Retiring the fossil stock early does not only destroy the value of the assets themselves; it closes one of the system’s historic outlets at the very moment when the surplus seeking outlets has not diminished. Capital displaced from fossil investment does not evaporate; it migrates, into property speculation, into financial instruments, into the next material frontier, which in the present period prominently includes the minerals and land demands of the replacement technologies themselves. The analysis therefore predicts that transition will generate its own extractive frontiers and its own displaced populations, in mining regions and sacrifice zones attached to the new economy, unless the distributional questions are confronted directly. Green extractivism, the phrase its critics use, names that frontier, and the tradition’s framework accommodates the phenomenon without strain: the compulsion is indifferent to the colour of the commodity.

Crisis theory also explains the peculiar temporality of energy investment, its booms of construction followed by gluts and write-offs, from railways to shale. Each boom is narrated by its participants as a response to scarcity or opportunity; the pattern across booms is of capital herding into the outlet of the moment, overbuilding it, and then defending the overbuilt stock through the long trough that follows. The fossil economy’s defenders are therefore not defending a rationally proportioned system; they are defending the sediment of several herding cycles, frozen at various stages of amortisation, and the incoherence of the stock, its redundancies and stranded corners, is evidence of the process that built it. Understanding the stock as sediment rather than as plan is clarifying for everyone: it lowers the expectation that its owners possess a unified strategy, and raises the expectation that its defence will be fierce, fragmented, and prolonged.

Framing the crisis: narrative, imagery, and the manufacture of common sense

Every public argument arrives inside a frame, and the frames through which the climate crisis is commonly seen deserve the same structural scrutiny as the economy itself. The dominant imagery of the crisis is meteorological and zoological: storms, floods, ice, and emblematic animals, a visual vocabulary that presents the phenomenon as an event happening to humanity from outside. The production system that generates the phenomenon appears in that vocabulary hardly at all; smokestacks figure as scenery rather than as decisions, and the boardrooms, balance sheets, and labour processes of the preceding sections are nowhere in the frame. The effect, whatever the intention of any individual image-maker, is to naturalise the crisis, relocating it from the category of things societies do to the category of things that befall them, and a naturalised crisis invites endurance rather than alteration.

Alternative frames exist and compete. The consumption frame centres the shopper and the footprint calculator, individualising responsibility in the pattern this article has criticised throughout. The technology frame centres the inventor and the breakthrough, promising rescue without reorganisation. The security frame centres the state and the threat, recruiting the crisis into the oldest vocabulary of emergency power. Each frame selects a protagonist and therefore a politics, and the struggle among frames is a real part of the phenomenon, not a distraction from it: publics act on the crisis they can picture. The Marxist intervention in that struggle is to insist on the production frame, in which the protagonists are owners, workers, states, and assets, and in which the crisis appears as the output of arrangements that have authors and alternatives. Whether the production frame persuades is a question of politics; that it identifies the level at which the levers sit is a question of analysis, and the two should not be confused by either side.

Journalism occupies a peculiar position inside the framing struggle, since it must render a slow, structural, statistical phenomenon as news, a genre built from events, persons, and novelty. The mismatch deforms coverage in predictable directions: toward the storm and away from the stock, toward the announcement and away from the amortisation schedule, toward the villain and away from the mechanism. Reporters who understand the deformation can compensate for it, and the best climate journalism of the period does exactly that, treating infrastructure as a beat and balance sheets as sources. The persona section addressed to journalists below builds on that observation; here it is enough to register that the crisis’s public form is itself produced, by institutions with their own accumulative pressures, and that a materialist analysis is incomplete until it includes the production of the picture among the things it explains.

The digital economy and the myth of dematerialisation

Every generation of the fossil economy produces a story in which the newest sector escapes materiality, and the digital economy is the present teller of that story. Software, streaming, finance, and platform services present themselves as weightless, an economy of information floating above the grubby world of fuel and ore. The material substrate tells a different story: data centres consuming the output of power stations, networks of fibre and masts, devices manufactured from mined metals and discarded on short cycles, and a logistics empire of warehouses and delivery vehicles standing behind every dematerialised purchase. The digital sector did not replace the material economy; it reorganised it, accelerated its cycles, and concealed its throughput behind interfaces of exemplary cleanliness. A shopper who has dematerialised a purchase has usually materialised a delivery.

The rebound mechanism operates in the digital realm with unusual purity. Computation’s efficiency has improved at a pace unmatched in any other technology, each unit of calculation cheaper in energy than the last by orders of magnitude across the industry’s history, and the system’s response has been to multiply calculations beyond every projection: more data retained, more models trained, more video streamed at higher resolutions to more screens for more hours. No sector demonstrates more cleanly that efficiency inside an expansionary system is a recipe for growth in the service, not for conservation of the resource. The analysts who celebrate digital dematerialisation are measuring weight per transaction while the transactions multiply without bound, an error the Jevons framework predicts and names.

Class enters through the labour the interface hides. Behind the frictionless platform stand the warehouse pickers, delivery riders, content moderators, miners of specialty metals, and assemblers of devices, a global workforce whose conditions are screened from the consumer by the same design that screens the smokestack. The digital economy’s self-image as post-industrial is therefore an artefact of perspective: from the consuming end, industry has vanished; from the producing end, it has merely moved, geographically and socially, out of the consumer’s sight. The climate analysis, whose founding gesture is to follow production rather than appearances, treats the digital sector as continuous with the older fossil economy it supplies, powers, and accelerates, and its conclusion generalises the whole article: there is no dematerialised capitalism, only capitalism whose materiality has been better hidden.

Sufficiency: the question the system cannot ask itself

Running beneath the mechanisms is a question so simple that the system’s accounting cannot formulate it: how much is enough? Needs, in any defensible account, are finite and satiable; a household can be warm enough, fed enough, housed enough, and moved enough, and beyond satiation additional throughput purchases position rather than welfare. An economy organised by accumulation cannot register satiation, because its units survive by expanding regardless of the welfare content of the expansion, and its measures of success, output and return, ascend identically whether the increment feeds the hungry or gilds the gilded. The concept of sufficiency, developed in the ecological literature adjacent to this tradition, names the missing register: a standard of enough against which production could in principle be judged complete.

Sufficiency deserves careful separation from austerity, with which hostile commentary deliberately confuses it. Austerity is the reduction of provision for the many in order to protect the claims of creditors; sufficiency is the capping of throughput at the top of the distribution in order to universalise provision at the level of need. The two policies move in opposite directions along the class axis, and the distinction is precisely the subsistence and luxury distinction restated as a programme-shaped idea. This article recommends no programme, yet it can record that the sufficiency concept performs analytical work inside the debate: it identifies the point at which the growth question stops being technical and becomes distributional, since an economy that can satiate needs and chooses throughput beyond them is making a choice about whose consumption the surplus serves.

The tradition’s own history with satiation is double, and honesty requires registering both sides. Marx imagined a society in which the realm of necessity shrinks and free development begins, a vision organised around time rather than goods; the states that later claimed his name organised themselves instead around output targets and heavy industry, treating sufficiency as a bourgeois deviation. Ecological writers inside the tradition have used that contrast as an internal reproach, arguing that the vision was betrayed by the productivism of its executors. The reproach connects this article to the record treated in the socialist states article, and it leaves the modern reader with a genuinely open question: whether the sufficiency standard is a recovery of the tradition’s founding hope or an amendment forced on it by a crisis its founders did not foresee. Either answer is available in the literature, and the debate between them is among the most productive the contemporary tradition sustains.

Neighbours and boundaries: ecological economics and the wider shelf

The Marxist analysis does not occupy its territory alone, and mapping the neighbours clarifies what is distinctive in its own claims. Ecological economics, the heterodox discipline associated with figures such as Herman Daly and Joan Martínez-Alier, shares the throughput concept, the scepticism about decoupling at scale, and the insistence that the economy is a subsystem of a finite ecosphere. Its centre of gravity, however, is biophysical rather than relational: it asks how large the subsystem can grow before it overwhelms its container, while the Marxist account asks who owns the subsystem, whose labour powers it, and which mechanisms compel its expansion. The two questions cooperate more often than they collide, and the contemporary literature is full of fruitful borrowings in both directions, yet the difference matters at the point of remedy, where ecological economics reaches for scale limits and the Marxist account reaches for property relations, each suspecting the other of treating the symptom.

Environmental history is a second neighbour, and the debt runs the other way: the labour-control origin is a work of history before it is a work of theory, built from archives, trade journals, and the material record of mills and streams. The discipline’s methods, its suspicion of retrospective inevitability and its attention to contingency, supply the analysis with its best internal check, since a mechanisms story that cannot survive the archive is merely a philosophy with footnotes. Political ecology, a third neighbour, contributes the ethnography of particular places, the village beside the mine and the neighbourhood beside the refinery, where the abstractions of this article acquire faces and names. An exposition that confined itself to mechanisms would miss what those neighbours know: that throughput is lived, place by place, by populations whose resistance and adaptation are data no table can hold.

Boundary disputes among the neighbours are mostly productive, and one deserves a sentence because it illuminates the whole map. Ecological economists sometimes accuse the Marxist account of anthropocentrism, of valuing nature chiefly as the condition of human production and reproduction. Marxist writers reply that the accusation mistakes the level of analysis: a theory of society is anthropocentric in the way a theory of tides is lunar, by designation of its object, and the charge properly lands only when the theory claims to be an ethics of nature as well. The exchange has generated a useful clarification on both sides, and readers who follow it will understand why this article, faithful to its designation, makes claims about society’s metabolism with nature while attributing nature’s own behaviour to the sciences that study it.

The concessions ledger: what this article has given away

Scholarly honesty is easier to audit when the concessions are gathered in one place, so let them be gathered. This article has conceded that several wealthy economies absolutely decoupled emissions from growth over recent decades on consumption-based measures, contrary to the strong form of its own tradition’s claim. It has conceded that low-carbon technology costs collapsed faster than planning or the tradition anticipated, through mechanisms internal to market economies. It has conceded that rebound magnitudes are contested and often partial. It has conceded that the environmental record of states industrialised under Marxist governments is poor, and that the labour-control reading of steam, while well evidenced, is disputed by competent economic historians on cost-accounting grounds. It has conceded that the analysis explains obstacles rather than outcomes, and that whole territories, movement politics, cultural change, innovation systems, belong to other accounts.

A ledger of that length might look like a defeat, and in the pamphlet version of the tradition it would be edited out. In the scholarly version, the concessions are the credential. What survives them is a compact, graded set of claims: competition punishes unilateral restraint; technologies are selected inside social relations and the selection persists; capitalised fossil value organises its own defence through owners, workers, regions, and treasuries; efficiency is converted by the system into expansion at rates that vary and must be measured rather than assumed; and the pace of transition, whatever its direction, is set by conflict among those mechanisms rather than by the state of public knowledge. A reader who rejects every one of those claims has a coherent position available, the ecological modernisation programme, stated fairly above. A reader who accepts them has not thereby accepted any party’s programme, because the claims underdetermine strategy entirely, which is why the tradition’s political arguments continue among people who share its analysis completely.

The ledger also records a methodological debt the tradition owes its critics. Every concession in the list was forced by evidence assembled outside the tradition’s own institutions: by innovation researchers, atmospheric accountants, economic historians, and the dissident scholars of the planned economies themselves. A research programme that can be corrected only from outside is in trouble; one that institutionalises the corrections, as this cluster of articles attempts, converts its critics into its quality control. The test of the next decade’s ecological Marxism will be whether its concessions ledger keeps growing at the same rate as its claims, since a tradition whose ledger has stopped growing has stopped reading.

Cement, steel, and the materials that build the world

Two industries deserve individual treatment because they embody the hard core that sectoral grading keeps encountering. Cement and steel together underpin the entire built environment, and both are fossil-bound in ways that exceed fuel choice: cement production releases carbon from the chemical transformation of its raw material itself, so that even a perfectly clean energy supply leaves a process emission untouched, while steel’s traditional chemistry uses carbon as a reducing agent, a reagent rather than a fuel. These are not sectors awaiting a cheaper substitute on a learning curve; they are sectors whose incumbent processes are woven into the material basis of construction, and whose alternatives remain, on the engineers’ own accounts, works in progress rather than products on shelves. The analysis did not invent that difficulty, and citing it is not a Marxist manoeuvre; what the analysis adds is the observation that both industries also carry the classic political profile, concentrated ownership, regional employment dependence, strategic status in every state’s imagination, and asset stocks amortised across decades.

The built environment those materials create then becomes the next century’s inheritance, in the pattern this article has traced repeatedly: decisions taken for reasons of cost, speed, and the organisation of the construction trades harden into a stock whose energy demands and demolition costs constrain every successor generation. A tower built cheaply around a steel frame and a concrete core is a commitment to a century of heating, cooling, and eventual disposal, entered into by developers whose horizon ended at the sale. The temporal structure deserves emphasis, because it inverts the usual story about markets and time: the market actors are rational at exactly the horizon their incentives specify, and the irrationality appears only at the level of the stock, where nobody’s incentives were ever stationed. A structural explanation is the name for an account that can locate irrationality at the level of the whole without attributing stupidity to any of the parts.

Substitution research in these sectors, pursued seriously by engineers and materials scientists, examines alternative chemistries, recycled feedstocks, and design strategies that use less material to achieve the same structure. The design strategy is the sleeper in that list: a large share of structural material in modern construction is specified by convention, liability practice, and professional habit rather than by physical necessity, so that the cheapest tonne of cement is the one the drawings never required. Professional conservatism, defended earlier in this article as a form of intellectual sunk capital, here appears in its most literal embodiment, poured and cured. Changing it requires no breakthrough, only the slow revision of codes, insurance expectations, and curricula, the unglamorous strata where lock-in actually lives.

Leisure, tourism, and the geography of escape

The fossil economy also built the modern experience of escape from itself, and the paradox is worth more than an ironic sentence. Mass tourism depends on cheap aviation, on heated and cooled resorts, on the global logistics of food and entertainment, and it sells, as its core product, relief from the industrial environments its own transport helps to sustain. The destinations marketed as untouched are reached by the most carbon-intensive consumer activity most households ever undertake, and the economies of those destinations reorganise around the arrivals, replacing local production with imported provision for visitors and remaking coastlines into the built environment of leisure. None of that makes the holidaymaker a villain; the desire for sun, rest, and elsewhere is among the most human of desires, and the industry’s achievement was to industrialise its satisfaction along the standard lines: standardised, marketed, and expanded without internal limit.

Positional analysis applies here with special clarity, because the geography of fashionable escape is a positional market in near-perfect form. Each cohort of travellers devalues the destination it publicises, and the frontier of the authentic retreats before the arrivals, generating an endless expansion of range that maps directly onto aviation’s growth. The mechanism has no villain and no satiation point, which is why it belongs in a structural account: a positional economy converts the universal and innocent desire for a holiday into an escalator of distance, and the escalator’s energy bill is nobody’s decision and everybody’s outcome. Writers who locate the problem in the traveller’s character have mistaken the escalator for a staircase.

Working time completes the circuit. The annual holiday is the scheduled negation of the working year, sold back to the workforce as the compensation for its disciplines, and the fossil economy supplies both the discipline and the compensation from the same stock of fuels. An analysis of leisure therefore belongs inside an analysis of production, not as a garnish upon it: the society that organises work around the employer’s timetable, the theme of the steam sections, organises rest around the airline timetable, and the two timetables are printed by the same civilisation. Readers inclined to treat tourism as a trivial coda to the serious economy might reflect that few sectors reveal the system’s unity, its fusion of production, consumption, status, and time, in so compact and so cheerful a form.

Risk, discounting, and the economics of a distant harm

One technical corner of the debate deserves translation into plain language, because it decides more than most manifestos. Economic models of climate policy must compare costs incurred now against harms avoided later, and the comparison runs through a discount rate, the device by which future values are shrunk into present equivalents. At the rates conventional in finance, harms arriving a century hence weigh almost nothing in the present ledger, however catastrophic they are described as being; at the low rates defended by the most famous dissenting review in the literature, the same harms dominate the calculation and justify immediate, expensive action. The entire policy conclusion pivots on a parameter that looks like a technicality and behaves like a philosophy, since choosing it means choosing how much the unborn count. The debate between the high-rate and low-rate camps was conducted in the journals as a dispute about the pure rate of time preference, which is the ethics of intergenerational justice wearing an economist’s vocabulary, and no amount of technical refinement can conceal that the parameter is chosen, not measured.

The Marxist observation about discounting is characteristic and worth stating precisely. A system that allocates capital by expected return has the discount rate built into its bones: every investment decision in the economy already shrinks the future at the rate competition imposes, because a firm that valued the distant future at par would be outcompeted by firms that do not and would cease to exist as a decision maker. The models’ parameter is therefore not an analyst’s whim but a transcription of the system’s own temporal constitution, which is why asking the system to act on a low discount rate amounts to asking it to behave as if its allocation mechanism were something else. States can and do override the transcription in particular domains, through public investment criteria and regulation, and the history of public works is partly the history of such overrides. The structural point is only that the override must be deliberate and defended politically each time, since the default, the rate at which the market values the future, reasserts itself wherever attention lapses.

Generational language enters here, and the analysis should handle it with the same suspicion it applies to aggregate humanity. The future generations invoked in climate rhetoric are not a class and do not share interests uniformly: the descendants of asset holders inherit portfolios as well as coastlines, while the descendants of the energy poor inherit the housing stock and the flood plain. Intergenerational justice and class justice are therefore entangled rather than alternative framings, and the discounting debate, read through the tradition’s lens, is a debate about whose grandchildren the present is willing to tax itself for. That reading does not settle the parameter. It explains why the parameter will never be settled technically, and why its career in politics resembles the career of every other distributional question the system processes.

What the reader now holds: an inventory before the apparatus

The argument can now be inventoried in the order a critical reader should test it. First, a definition: the analysis locates the crisis’s causes in competitive accumulation, the historically specific adoption of fossil energy, the capitalised defence of the existing stock, and the conversion of efficiency into expansion. Second, a founding case: the water-to-steam transition, in which the winning technology was chosen substantially for the command it gave employers over labour, location, and time, a case whose evidence is archival, whose critics are competent, and whose partial truth alone would establish that technology selection is a social process with permanent consequences. Third, a political mechanism: sunk capital, which predicts that resistance maps onto balance sheets and livelihoods rather than onto opinion, and which explains the shape of real policy, generous to the new, protective of the old. Fourth, a dynamic: rebound, honestly bounded, which blocks the inference from efficiency to aggregate decline. Fifth, a counter-case at full strength: cost collapse under market conditions, documented absolute decoupling in several wealthy economies, and the poor environmental record of the planned economies, which together retire the impossibility thesis and demote the analysis from prophecy to a theory of retardation. Sixth, a method: attribute the physics to the physicists, quantify nothing on borrowed authority, name the dependence wherever a social claim borrows a scientific premise, and keep the concessions ledger open.

Held together, those six elements are neither a programme nor a mood. They are an instrument for grading the debate a reader will meet everywhere else: in the news segment that blames shoppers, in the editorial that announces a breakthrough, in the ministerial speech that protects an asset while praising a target, in the pamphlet that promises the system’s innocence or its doom. The instrument’s readings are fallible and revisable, as this article has been at pains to show, yet a reader equipped with it will ask, of every climate claim, the questions that sort the serious from the decorative. Who owns the stock in question? What selection chose the technology? Which balance sheet does the proposal threaten, and whose livelihood stands behind it? What happens to the saving once the system has finished converting it? Those four questions, learned once, apply for life, and they are the closest thing to a conclusion that an article forbidden to recommend policy is permitted to leave behind.

Biomass, muscle, and the world before fossil power

The fossil economy looks inevitable only from inside it, and a glance at the energy regime it replaced restores a sense of its contingency. Pre-industrial production ran on muscle, wood, wind, and water: on the bodies of people and animals, on forests managed or plundered for fuel and charcoal, on sails and millstreams. That regime had its own scarcities and its own cruelties, and no part of this article romanticises it; the labour of the pre-fossil world was grinding, its energy poverty absolute by modern measure, and its hierarchies enforced scarcity as a technology of rule. What the comparison establishes is narrower and more useful. For most of human history, energy was a flow constrained by land and season, harvested at rates the landscape set, and economic life organised itself inside that constraint as inside a law of nature. Fossil fuels broke the constraint by substituting a stock for a flow: millions of years of stored sunlight, drawn down at whatever rate machinery allowed. The break, not any particular machine, is the founding event of the modern economy, and everything distinctive in its subsequent behaviour, its growth rates, its urban forms, its liberation from season and locality, follows from the substitution.

The stock character of fossil energy also explains a feature of the system that purely economic accounts leave mysterious: its insensitivity to depletion signals in the short run. A flow regime disciplines its users continuously, since overuse in one season becomes shortage in the next within the same valley. A stock regime defers the discipline across generations and continents, allowing each era of users to experience the fuel as effectively limitless while the drawdown proceeds. Political argument formed inside that experience inherits its intuitions: abundance feels like a fact about the world rather than a phase in the life of a stock. The analysis does not need to adjudicate reserve estimates, which belong to geology, to make the sociological point that institutions built during the phase of apparent limitlessness will misjudge the phase that follows, and that the misjudgement is a datum about institutions rather than about anyone’s intelligence.

Land tells the same story from the other side. The pre-fossil economy’s energy was land, worn as woodland, pasture, and arable, so its growth collided constantly with the food supply and the forest. Fossil energy released production from that collision, which is one reason contemporaries experienced coal as emancipation and why its critics, from the start, were answered with the genuine rejoinder that the alternative was the wood famine and the horse. A transition away from fossil energy reintroduces land into the energy question, through biofuels, through the footprints of generation, through the mining of new materials, and the reintroduction surprises a civilisation that forgot the connection. The historical memory matters practically: societies that remember energy as a land question ask different questions of a transition plan, about acres and ownership and food, than societies that remember it only as a bill.

Reading a policy announcement with the table in hand

The mechanisms repay one final exercise, of the kind a reader can perform on any morning’s news without specialist equipment. Take a government announcement of a new climate measure, of whatever government and whatever design, and grade it against the four rows. Ask what the measure does to the competitive position of the firms it touches, and whether their rivals abroad or in adjacent sectors escape it, since the first row predicts that the escape routes will be used. Ask whether the measure retires capitalised value or merely adds new capacity beside the old, since the third row predicts that announcements will systematically prefer addition to retirement, celebration of the new plant to mourning for the closed one. Ask what the measure assumes about efficiency savings, and whether any cap prevents the savings from being spent on expansion, since the fourth row predicts the conversion. Ask, finally, what the announcement reveals about the selection criteria of the technologies it favours: which of them were chosen because they fit the existing organisation of production, with its grids, its settlement patterns, its timetables, and which demand that organisation change, since the second row predicts that the fitting technologies will be built and the demanding ones will be studied.

The exercise is deliberately non-partisan, and its results will embarrass governments of every complexion in rotation, which is the mark of a structural instrument rather than a factional one. It also explains the characteristic emotional rhythm of climate politics, the alternation of announcement euphoria and implementation despair, better than theories about sincerity do. Announcements are made in the register of intention, where governments are sovereign; implementation happens in the register of the mechanisms, where they are not. A public trained to grade announcements against the table will discount the euphoria appropriately and will locate its scrutiny where the outcome is decided: in the exemptions, the timelines, the compensation terms, and the treatment of the stock. That training, multiplied across a readership, is the modest political effect an analytical article can honestly claim.

Business announcements submit to the same grading with even less resistance. A corporate pledge can be read for its boundary (which subsidiaries, which scopes of emission, which dates), for its reliance on offsets purchased elsewhere rather than on the retirement of its own assets, and for the relation between the pledged trajectory and the firm’s capital expenditure, which is the document in which corporations tell the truth. The expenditure either bends toward the pledge or it does not, and no quantity of narrative alters the arithmetic. Journalists equipped with the table have turned that comparison into a genre of reporting, and its conclusions, applied firm by firm, amount to an empirical literature on the sunk-capital mechanism written in real time by the financial press. The tradition could ask for no better confirmation of its method than to have its central mechanism independently operated, under other names, by the business pages.

The corporation, the charter, and the legal form of accumulation

Behind the first mechanism stands a legal artefact so familiar that its strangeness has worn off: the joint-stock corporation with limited liability, governed by officers charged with returning value to shareholders. The form matters to the climate argument because it hardwires the accumulation compulsion into law and fiduciary duty. A manager who proposed to shrink the firm’s throughput for the common good, at the cost of the return shareholders could obtain elsewhere, would face not merely market discipline but legal challenge, and the officers who replaced that manager would be selected for their refusal to repeat the experiment. The compulsion, in other words, is not a mood among businesspeople; it is inscribed in the constitution of the firm itself, and reform proposals that ignore the inscription will be processed by it, as a river processes a stone.

The historical depth of the form deserves a moment’s attention, because it clarifies what is and is not being claimed. Chartered companies raised capital for voyages and colonies centuries before the factory system, and the modern corporation generalised their device: permanent capital, transferable shares, liability capped at the stake. Permanence is the climate-relevant feature. A partnership dies with its partners and discounts the future accordingly; a corporation is chartered to outlive everyone in the room, accumulates across generations, and defends its asset base with an institutional memory no individual can match. The sunk stock of the fossil economy is guarded by entities designed, with great legal ingenuity, never to retire and never to forget an interest. That design is a human artefact, revisable in principle by the legislatures that grant the charters, and the analysis simply notes that the defence of fossil value is mounted by the most durable organisations our law knows how to build.

Shareholder composition adds the final twist encountered earlier in the finance section. The owners of the fossil corporations are, through pension funds and index products, substantially the workforce itself, saving for old age in instruments that hold the market as a whole. The class line, so clean in the mill town where the owner lived above the works, now runs through every retirement account, and the politics of asset defence recruits the savings of the very populations a transition is supposed to serve. Marxist analysis that ignores that interpenetration will misjudge the coalition it faces; analysis that notices it understands why the defence of the stock presents itself, sincerely, as the defence of everyone’s grandmother. Disentangling retirement security from fossil returns is a technical and political problem of the first order, and it is invisible to any framework that pictures the classes as two camps in separate fields.

War, peace, and the energy of the state reconsidered

A second look at the state’s appetite is warranted because the first pass treated only its scale. The deeper point concerns priority: in every recorded emergency, states have demonstrated that the constraints celebrated as iron in peacetime, budgets, market signals, property expectations, dissolve overnight when the survival of the state itself is engaged. Production is commandeered, consumption is rationed, capital is directed, and populations accept impositions they would riot against in ordinary times, because the emergency frame reorganises what sacrifice means. Climate politics returns obsessively to that precedent, usually in the form of a demand that the crisis be treated with equivalent seriousness. The analysis contributes a colder observation: emergencies are declared by states, for purposes states define, and the fossil economy’s entanglement with the state’s own security apparatus means the emergency frame, if it comes, will arrive shaped by the interests this article has catalogued. Nothing in the precedent guarantees the direction of the mobilisation, only its possibility, and a tradition that studies the state will be the last to confuse the two.

Peacetime has its own quieter militarisation of energy. The geography of extraction is patrolled; shipping lanes are policed; infrastructure is hardened and surveilled; and the costs of that protection are socialised across taxpayers while the returns remain with the asset holders, an arrangement the tradition recognises from every earlier chapter of accumulation’s history, in which the state supplies the coercion and the commons supplies the budget. Energy security, in short, functions as the sunk-capital mechanism’s diplomatic passport, valid in every capital, exempting the stock’s defence from the scrutiny applied to ordinary commercial interests. Analysts grading a transition debate should listen for the passport being stamped: the moment an argument shifts from value to security, the balance-sheet discussion has ended and a different, harder politics has begun.

Children, inheritance, and the temporal politics of the stock

The fossil economy’s longest commitments are made on behalf of people who cannot contest them, and the fact gives the crisis a distinctive moral geometry. A generating station financed now will still be defending its amortisation when children now at school hold mortgages; a suburb laid out now will dictate the transport emissions of families not yet formed; the atmosphere itself, on the findings attributed to the relevant sciences, keeps the longest ledger of all. Political systems, by contrast, are built from short cycles: elections, budgets, quarterly reports, news days. The mismatch between the stock’s horizon and the polity’s horizon is not a failure of attention that better communication could repair. It is a structural property of the institutions involved, each faithfully pursuing the incentives of its own clock, and it explains why the crisis is perennially described as urgent and perennially processed as routine. Urgency belongs to the stock’s timescale; routine belongs to the polity’s; the commentary that marvels at the gap has usually mistaken a property for a paradox.

Inheritance complicates the generational story in the class direction this article keeps insisting on. What the young inherit is not a uniform damaged world but a stratified one: portfolios and properties for some, flood-prone rentals and stranded skills for others, public debts incurred to compensate asset holders distributed across all. The fossil stock’s defence, if it succeeds in socialising its losses, will bequeath the bill preferentially to those who never held the assets, a transfer conducted in the vocabulary of fiscal necessity. Analysts of the transition should therefore read compensation schemes as inheritance law written early, and ask of each one the question inheritance law always asks: who is being made whole, at whose expense, and in what order of priority? The tradition’s founders wrote about the inheritance of capital as the quiet mechanism by which class reproduces itself across generations; the climate version of the mechanism is being drafted now, in the fine print of transition finance, largely unread.

Youth movements enter the analysis at exactly this point, and they should be read structurally rather than sentimentally. A generation that inherits the stock without having consented to its construction, and that faces the longest exposure to its consequences, has an objective interest in the question that no other constituency shares with the same purity, and its periodic irruptions into climate politics express that interest in the only currencies available to the assetless: attention, disruption, and the vote it will soon hold. The tradition has a category for such irruptions, the moment when a structural position becomes a political force, and it also has a warning earned across its history: structural position alone organises nothing, and the distance between an interest and a movement is measured in organisation, programme, and time. Whether the irruptions cross that distance is among the open questions the analysis bequeaths, honestly, to events.

A short genealogy of the word crisis

The title of this article contains a word whose career is worth a paragraph, because naming is part of the phenomenon. Crisis entered political economy from medicine, where it named the turning point of an illness, the moment at which the patient recovers or dies, and Marx’s usage preserved that charge: a crisis was the violent moment in which the system’s contradictions, accumulated quietly, forced themselves into the open and demanded resolution. The climate usage inherits the charge and alters the tempo. An atmospheric crisis does not turn in a season; it accumulates in a stock, expresses itself in probabilities and distributions, and offers no morning on which the patient is declared saved or lost. The mismatch between the word and the phenomenon deforms public reasoning in both directions at once: the absence of a turning point is misread as the absence of the condition, while the rhetoric of turning points promises resolutions that the phenomenon’s structure cannot deliver.

The tradition’s own crisis vocabulary needs the caution as much as anyone’s. Writers formed by the economic cycles reach instinctively for the grammar of the crash, the moment when the old order’s impossibility is revealed and history accelerates. Applied to the atmosphere, that grammar misleads, because the relevant processes are cumulative and lagged, and because a politics organised around awaiting the revealing moment will wait through the entire period in which the stock grows. The better inheritance from the tradition is not its crisis rhetoric but its crisis method: the insistence that a system’s normal functioning, not its interruptions, is what requires explanation, and that the contradictions are visible in the ordinary operations of a Tuesday. The fossil economy’s crisis, in that sense, is not coming. It is running, at nameplate capacity, in every power station and freight corridor of the ordinary world, and the analysis that grasps the ordinariness has understood the tradition’s deepest lesson about how systems fail: not with a crash that announces itself, but with the smooth continuation of everything that was already happening.

The seminar room test: explaining the argument to a sceptical engineer

Every structural argument should survive an encounter with a competent practitioner of the system it describes, and the climate analysis is best imagined in conversation with a sceptical engineer: a person who designs real plants, respects thermodynamics, and suspects social theory of being fog. The engineer asks what the mechanisms predict that a competent spreadsheet does not already know. The answer can be given without fog. The spreadsheet knows the cost of the substitute; the mechanisms predict the fate of the incumbent, the behaviour of the owners whose assets the substitute strands, the response of the workers whose skills it devalues, and the conversion of the engineering saving once prices fall. Those are not quantities a spreadsheet refuses; they are quantities its categories omit, because the model begins where the social relations have already been assumed. The analysis is the study of the assumptions, and an engineer who has watched a technically superior design die in committee, killed by procurement rules written for its predecessor, has already operated the labour-control and lock-in mechanisms under their office names.

The engineer then asks the harder question: whether the mechanisms add up to a counsel of despair, in which nothing can move until everything changes. The honest answer returns to the sectoral evidence. Electricity moved, within the system’s ordinary operations, once the substitute fitted the organisation and undercut the incumbent; freight has not moved, because nothing yet fits its organisation at a price competition will transmit. Despair is the wrong conclusion from a theory that predicts differential motion and then watches it occur. The better conclusion is a research agenda disguised as a warning: study the fit, the stock, and the coalition in any sector before predicting its trajectory, and distrust equally the booster who reads one sector’s success as the system’s absolution and the catastrophist who reads another sector’s inertia as its damnation. Both are quoting half the evidence, and the mechanisms exist to make the halves visible as halves.

A last exchange completes the test. The engineer observes that engineers, given a clear specification and a funded mandate, have solved every problem of this shape put to them, from sanitation to electrification itself. The observation is just, and the analysis should absorb it whole: the history of the fossil economy is not a history of technical incapacity but of specifications written by the relations this article describes. Electrification was specified, funded, and built by states and utilities pursuing their own accumulation and security; sanitation was specified when the diseases of the industrial city began killing the classes who wrote the specifications. The question the climate crisis puts to the present is whose interests will write its specification, and with what urgency, and the analysis’s whole contribution is to show that the specification is a social document before it is a technical one. On that point, the engineer and the materialist can shake hands: both know that nothing gets built until somebody with power decides what building means.

A note on the evidence base and its silences

An article that has refused all figures owes its readers an account of what its evidence consists of, since the refusal might otherwise masquerade as a shortage. The case assembled here rests on five kinds of material, each with its own standards. Archival and business history grounds the founding case: manufacturers’ correspondence, trade testimony, and the site decisions themselves. The history of technology grounds the claims about selection and fit, a literature built from artefacts, patents, and the records of the firms that chose among designs. Institutional economics and energy studies ground lock-in, where the evidence is the observable career of infrastructures and the rules written around them. Innovation research grounds the counter-case on costs, where the evidence is the documented trajectory of prices and the reconstruction of the policies and learning processes behind them. Finally, the emissions accounting literature grounds the decoupling discussion, where the evidence is national inventories rebuilt on alternative boundaries, and where the disputes are methodological in the open, honourable way that lets a reader watch the construction.

The silences deserve equal billing. Absent from the article, by the policy stated at the outset, is any adjudication of the physical science, and a reader seeking the state of that knowledge must go to the assessment literatures themselves, which is where such knowledge legitimately lives. Absent also is any treatment of the ecological damage that does not pass through the climate system, except where the mill-town archive forced it into view; the framework beneath that larger subject sits in the metabolic rift article linked at the start. Absent, finally, is the strategic literature, the arguments about parties, movements, and programmes, which the analysis underdetermines and which other articles in this series take up on their own terms. A synthesis that confessed no silences would be a sales document. This one has tried to be a map, and maps are judged by whether their blank spaces are labelled.

The labelled blanks also serve the reader who disagrees. A critic who believes the founding case fails on cost accounting knows exactly which archive to attack. A critic who believes decoupling is broader and faster than the article allows knows exactly which accounting literature to marshal. A critic who believes the whole framework mistakes the level of analysis knows that the objection must be aimed at the mechanisms’ explanatory claims, not at their author’s politics, which the article has declined to supply. Scholarship advances through such well-aimed disagreement, and the tradition examined here, whatever its pamphleteers do, has in its best quarters always preferred a precise enemy to a vague friend. The article is offered in that spirit: as a set of claims with addresses, each of which can be visited, tested, and, where the evidence demands it, condemned.

The argument in one paragraph, for the reader who arrived late

Everything claimed here can be compressed without violence to its qualifications. An economy organised by competitive accumulation must expand its throughput and punishes any participant who restrains it unilaterally. The fossil energy at that economy’s heart was not selected for cheapness in its founding case but for the command it gave employers over labour, place, and time, and the selection persists inside the built world it created. That built world is a stock of capitalised value whose owners, workers, regions, and treasuries will defend it in proportion to their dependence, so transition is resisted structurally rather than merely lobbied for. Efficiency, the remedy most often proposed as painless, is converted by the same system into expanded consumption at rates that are real, variable, and argued over in good faith by the specialists. Against the whole construction stands a documented record of collapsing substitute costs and of wealthy economies whose emissions fell while their output grew, which retires the claim that the system can do nothing and replaces it with a harder, more useful claim: that the system does what it does at a pace and in a shape set by the defence of value, and that anyone who wants a different pace must argue with balance sheets, livelihoods, and selection pressures rather than with ignorance. That is the analysis. Its evidence has addresses, its concessions are on the record, and its four mechanisms sit in a table near the front, waiting to be graded.

A reader who retains only the table has retained the working core, since the surrounding thirty-odd sections are commentary upon four wagers. The first wager is that selection, not intention, drives expansion. The second is that a single well-preserved historical case can reveal the anatomy of a whole energy order. The third is that political resistance follows the money with enough fidelity to be mapped. The fourth is that savings inside an expansionary system are a form of fuel. Each wager can lose on its own evidence without dragging the others down, and the willingness to let them be graded separately is what separates the analysis offered here from the rhetoric it is so often confused with, on both sides of the argument. Grade them, then, and keep the grades; the next article in the cluster takes up the comparison between this tradition and its green capitalist rival, where several of the wagers placed here are called.

One discipline remains to be named before the apparatus takes over, because it governed the writing and should govern the reading. Nothing in the preceding pages has asked the reader to feel anything in particular about the fossil economy: not guilt, which the analysis regards as a misdirection of causal attention; not nostalgia, which its archive forbids; not despair, which its own counter-case refutes; and not the warm sense of being on the correct side, which it treats as the most reliable sign that an argument has stopped. The analysis asks instead for a habit of attention, directed at selection criteria, balance sheets, timetables, and the conversion of savings, and it is indifferent to the reader’s prior allegiances because the mechanisms operate identically on the sympathetic and the hostile observer. A conservative reader who accepts the sunk-capital mechanism has acquired a tool for understanding why governments of the right protect incumbents; a radical reader who accepts the decoupling evidence has acquired a check against slogans. The article succeeds if both readers leave better equipped than they arrived, and it is content to be judged by that standard rather than by the warmth of its welcome in any camp.

For students and exam candidates

Build the answer around the four mechanisms table, since it converts a sprawling topic into a gradeable structure. Open with the canonical definition: the analysis locates the causes of the crisis in competitive accumulation, in the historical adoption of fossil energy for labour control, and in the capitalised value of the existing stock, rather than in aggregate behaviour or unpriced costs alone. Then give the founding case its paragraph, because the water-to-steam transition is the concrete evidence examiners reward and the claim most candidates omit. Close by stating the strongest objection fairly, markets and decoupling research included, before offering the rate-versus-direction answer; scripts that present the objection honestly and answer it precisely outscore scripts that pretend it away. Practise the distinctions until they are reflexes: territorial against consumption-based accounting, cost against control, persistence against conspiracy. The companion study notes at VaultBook Marxist study notes carry a card on the four mechanisms designed for revision, and the cluster’s dedicated guide, ecological Marxism exam answers, works through full specimen answers. Keep every figure out of the answer; the analysis, as this article models it, is argued from mechanisms and cases, and a script stuffed with unattributed measurements advertises the opposite of the method being examined.

The working day, the clock, and stored energy

Energy transitions are also revolutions in time, and the dimension of time is where the labour-control origin shows most clearly to a modern reader. Agricultural and artisanal production had always been governed by task and season: work expanded and contracted with daylight, weather, harvest, and the completion of the job in hand. The water-powered mill began to break that governance, since machinery that turned only when the river allowed still imposed attendance when it turned, yet the river kept a share of sovereignty over the schedule. Steam transferred the whole of that sovereignty to the employer. The engine’s indifference to season and sunset meant that the limiting factor on production became the endurance of the workforce and the state of the order book, and the working day lengthened accordingly, into hours that reformers spent decades documenting and legislatures spent decades contesting.

Marx’s chapter on the working day in Capital reads, among other things, as an energy history written from the receiving end. The struggle over hours was a struggle over how far the new sovereignty of stored energy would extend into human life, and every factory act that limited the day was a partial socialisation of the timetable that steam had privatised. Remembering that struggle clarifies the present analysis in two ways. It shows that resistance to the fossil economy’s disciplines is as old as the fossil economy, mounted by the workers subjected to them long before any environmental framing existed. And it underlines the analysis’s central methodological habit: follow the energy source into the labour process, and the social relations become visible, because energy is never merely burned; it is burned under somebody’s command, on somebody’s schedule, for somebody’s return.

Shift systems, night work, and the continuous-process industries all descend from that transfer of sovereignty. A modern economy that produces around the clock has normalised the idea that energy waits upon demand at every hour, and the normality conceals the achievement: continuous production requires continuous labour discipline, continuous maintenance, and continuous fuel supply, coordinated across whole societies. When analysts ask why fossil energy is difficult to displace, part of the answer hides in that normalisation. Substitutes are not only evaluated as fuels; they are evaluated against a civilisation organised around the expectation that power is available instantly, everywhere, in any quantity, at any hour. An energy source with different rhythms asks that civilisation to renegotiate expectations embedded in everything from hospital rotas to food logistics. The renegotiation may be entirely feasible, and its feasibility is not this article’s question. Its scale, however, is the sunk-capital and labour-control mechanisms stated in the vocabulary of everyday life.

Food, fibre, and the fossil inputs of agriculture

Discussion of the fossil economy gravitates toward electricity and transport, yet agriculture binds fossil energy into the most basic of all provisions. Synthetic fertiliser manufactured from fossil feedstocks underwrites a large share of the world’s food production; diesel powers the machinery that plants and harvests; fossil-heated glasshouses extend growing seasons; and fossil-fuelled transport moves food across distances that separate production from consumption almost completely. Each of those dependencies was adopted through the same selection process examined in the cotton case: the winning techniques were those that served the organisation of production, in this instance the substitution of purchased inputs and machinery for agricultural labour, and the integration of farming into industrial supply chains owned elsewhere.

The class anatomy of food energy repays attention. The farmer operating inside the fossil system is less an independent producer than a manager of throughputs, buying fertility in bags and power in tanks, selling into markets whose prices are set far away. Dependence on purchased inputs disciplines the farm as surely as the factory bell disciplined the mill, through debt, through contracts, and through the impossibility of pausing a biological process while disputing a price. Meanwhile the nutritional emissions of a population are largely fixed by the supply chains offered to it: the shopper choosing between two products chooses between two arrangements of fertiliser, freight, refrigeration, and packaging that were settled upstream. Food therefore illustrates, at the most intimate scale, the article’s recurring correction of consumer-blame framings. The carbon content of dinner was decided in fertiliser plants, shipping lanes, and supermarket logistics long before it was decided at the table.

Agriculture also tests the rebound mechanism in a distinctive register. Efficiency in food production, higher yields per hectare and per animal, has historically been accompanied not by the retirement of land and inputs but by expanded total production, cheaper calories, and diets richer in the most resource-intensive products. The pattern matches the Jevons template: gains at the level of the unit, expansion at the level of the system, mediated by markets that convert every saving into an opportunity for growth. Researchers dispute the proportions in particular cases, and land-sparing optimists can point to regions where yields rose while farmland contracted, usually under the influence of policy and protection rather than of the market unaided. The dispute belongs to the agricultural literature; what the climate analysis takes from it is only the structural observation that efficiency inside an expansionary system behaves differently from efficiency inside a capped one.

For teachers and seminar leaders

The water-to-steam case is the most portable teaching asset in this cluster, because it works in history, geography, and politics classrooms alike and because its evidence can be placed directly in students’ hands. Stage the seminar as a decision problem: give the class the constraints a manufacturer faced, water sites with their labour scarcity and interrupted flow, steam with its fuel costs and its urban freedom, and let groups argue the investment before revealing the labour-control reading and its critics. The exercise teaches the difference between price explanations and relational explanations more efficiently than any lecture, since students discover in their own deliberations how readily discipline and location crowd out the power bill. Follow with the four mechanisms table as a sorting instrument: present four short accounts of transition resistance drawn from different sectors and ask which rows explain which account, and where each row fails. Close by assigning the objection in its strong form as an essay, requiring named research on both sides. Students who can state Nemet’s findings and the consumption-based decoupling results accurately, and can say exactly what those findings do and do not settle, have learned the transferable skill this material exists to teach: grading a structural argument against evidence without either worshipping or dismissing it.

For researchers, historians, and journalists

The sunk-capital mechanism is the exportable product here, because it converts a familiar newsroom observation, that transition politics moves at the speed of the slowest balance sheet, into a researchable claim. Treat asset stocks as the independent variable: map the age, book value, and network dependence of the fossil infrastructure in any sector under examination, and the geography of political resistance becomes a hypothesis to test rather than a mystery to lament. The labour-control origin offers a second programme, this time archival: Malm’s cotton evidence invites parallel studies of other energy transitions, asking in each case what social advantages, beyond cost, the winning source delivered to its adopters. Historians of technology will find the origin-explains-persistence pattern a useful organising wager, provided it is stated as a wager and checked against price-based rivals sector by sector. For journalists, the discipline this article models is worth stealing outright: credit physical findings to the bodies that produced them, refuse the rhetorical use of consensus, and never let a source convert a disagreement about policy into an insinuation about science. The framework beneath the mechanisms, for those who need it, sits in the cluster’s opening survey, ecological Marxism complete guide.

Frequently Asked Questions

Q: What does a Marxist analysis of climate change say?

It says the crisis is produced by the organisation of production rather than by aggregate human behaviour. Competitive accumulation compels firms and states to expand throughput; fossil energy was adopted historically for advantages of labour control and location; the existing fossil stock defends its capitalised value politically; and efficiency gains are partly consumed by expanded output. Causes are located in relations of ownership and compulsion, and remedies, on this account, must reckon with those relations rather than with consumer virtue.

Q: Why can capitalism not solve the climate crisis internally?

The seeded question states the strong version of the claim, and the honest answer qualifies it. Internal solutions face four structural obstacles: competition punishes unilateral restraint, sunk fossil assets mobilise their owners and dependents against early retirement, rebound converts efficiency into expansion, and states depend on the growth the system generates. Absolute reductions have nevertheless occurred in several wealthy economies, so impossibility is not established. The defensible claim is that internal mechanisms systematically retard and distort a solution, setting its pace by asset defence rather than by need.

Q: What is the fossil capital argument?

The argument, developed most fully in Andreas Malm’s study of the British cotton industry, holds that fossil energy was adopted in its founding industrial case for social rather than purely technical reasons. Manufacturers chose steam over cheaper and abundant water power because steam could be installed where labour was plentiful and disciplineable, ran on the owner’s schedule, and freed production from remote sites. Fossil capital names the resulting formation: an energy system whose social advantages to owners were built in at birth and which therefore resists displacement by price arguments alone.

Q: Why did industry switch from water to steam?

Because steam solved the manufacturer’s labour problem. Water power tied mills to remote streams where a workforce had to be gathered, housed, and kept, where scarcity strengthened the workers’ hand, and where drought, frost, and flood set the working calendar. Steam could be erected in populous towns amid deep labour markets, operated for as many hours as the owner chose, and moved if a locality grew expensive or organised. Water was frequently cheaper as power; steam was decisively superior as command over labour, location, and time, and the industry paid the difference willingly.

Q: What is carbon lock-in?

Carbon lock-in is the condition in which technologies, institutions, skills, and expectations interlock around the incumbent fossil energy system until departure becomes expensive from every direction at once. Power stations presuppose grids, grids presuppose generation profiles, vehicles presuppose roads and fuel distribution, buildings presuppose heating systems, professions presuppose training, and finance presupposes the book values of all of it. The concept, associated with Gregory Unruh’s work on industrial economies, describes path dependence at civilisational scale. Marxist writers add the class inflection: ownership of the locked stock, and state dependence on its revenues, convert inertia into organised defence.

Q: What is the Jevons paradox?

The paradox, drawn from William Stanley Jevons’s 1865 study of British coal use, observes that improving the efficiency with which a resource is used can increase rather than decrease total consumption of it. Efficient engines cheapened steam power, and cheapened power found multitudes of new applications, so coal consumption climbed. The general form: efficiency lowers the effective price of a service, the lower price expands demand for the service, and freed income is spent on further energy-using goods. Contemporary magnitudes are contested and vary by sector, so the paradox warns against relying on efficiency alone without claiming efficiency is useless.

Q: Is climate change a class issue?

On the analysis presented here, yes, in a precise structural sense. Emissions are generated inside production organised by owners of capital, distributed unevenly across income strata, and defended politically where capitalised value and dependent livelihoods coincide. The class claim is not that every wealthy person emits identically or that workers emit nothing; it is that the decisions fixing the carbon content of economic life, investment decisions about plant, networks, and products, are taken within class relations, and that the burdens and the capacity to act are distributed along class lines as well.

Q: What is the treadmill of production?

The treadmill of production, Allan Schnaiberg’s framework within environmental sociology, describes a coalition of expansion. Firms must grow to sustain profits, states facilitate growth to sustain revenue and employment, and workers acquiesce because their livelihoods ride on the same expansion. Each participant behaves intelligibly, and material throughput rises as the joint result. The framework’s contribution to the climate argument is political sociology: it explains why braking the treadmill is so difficult even when its direction is widely regretted, since every major institution draws income from its motion.

Q: Does the falling cost of renewables change the argument?

It changes the argument’s centre of gravity without dissolving it. Research such as Gregory Nemet’s shows that competition, public research, and industrial policy collapsed substitute costs faster than planning exercises anticipated, which refutes any claim that the system cannot innovate its way toward substitutes. The mechanisms in this article were never mainly about invention, however. They concern the retirement of a defended asset stock, the sectors where substitution remains organisationally hard, and the conversion of efficiency into expansion. Cheap generation answers the easiest part of the problem; lock-in governs the rest.

Q: What is the strongest objection to the Marxist account?

The strongest objection is empirical. Competitive markets combined with public research and industrial policy drove low-carbon technology costs down faster than any planning exercise anticipated. Several wealthy economies reduced emissions while growing, including on consumption-based measures that count imported goods. And the environmental record of planned economies was poor, which removes the easy contrast the account might wish to draw. Together these findings refute the strong impossibility thesis. The account’s surviving form is a theory of retardation and distortion, and its defenders should say so plainly.

Q: How does the labour-control origin grade an energy transition?

It grades a transition by organisational fit rather than by price alone. A substitute that slots into the existing organisation of production, serving the same grids, appliances, shift patterns, and business models, is predicted to spread quickly once inexpensive. A substitute that demands new settlement patterns, new labour processes, or the premature scrapping of valuable stock is predicted to stall however elegant its engineering, because it fails criteria that were decisive when the incumbent system was selected and that remain embedded in how production is arranged.

Q: What should a student separate between a technology’s cost and the control it gives its owner?

A student should separate the power bill from the command structure. Cost asks what a unit of energy or a machine does to the accounts. Control asks what the technology does to the owner’s authority over location, over the supply and discipline of labour, and over the working timetable. The fossil capital argument is precisely the claim that the second ledger decided the founding transition. Keeping the ledgers distinct prevents the commonest examination error on this material, which is refuting a control argument with cost evidence that was never in dispute.

Q: Why did employers prefer steam power when water was cheaper?

Employers preferred steam because cheap power at an unusable place and on an unusable schedule was not cheap in the only sense that mattered commercially. A water site imposed housing costs, retention problems, a scarce and therefore confident workforce, and production hostage to season and weather. Steam cost more at the boiler and repaid the difference at the gate: abundant replaceable labour, hours set by the owner, and freedom to locate beside markets and canals or to leave a district altogether. The preference reveals what employers were actually buying, which was command over the conditions of production.

Q: How does the four mechanisms table order the analysis?

The table assigns one mechanism to each of the four questions the analysis asks. Competitive accumulation answers why output keeps expanding; the labour-control origin answers why fossil energy was adopted; capitalised sunk assets answer why the existing stock defends itself politically; rebound answers why efficiency underdelivers against engineering projections. Ordering matters because the mechanisms rest on different evidence and can be graded separately: a reader may accept the sunk-assets row, contest the origin row, and remain undecided on rebound, and the table is built to survive exactly that partial verdict.

Q: What separates a structural explanation of the crisis from a policy failure explanation?

A policy failure explanation treats the crisis as the residue of correctable mistakes: a tax set too low, a regulation delayed, an election lost. A structural explanation treats the recurring pattern of those mistakes as the phenomenon to be explained, and locates its cause in relations, competition, capitalised value, and state dependence, that reliably reproduce the failures across different governments and decades. The structural claim is falsifiable in its own way: show the pattern dissolving under ordinary conditions, with restraint unpunished and assets calmly written down, and the explanation loses its object.