Value is not a stock you defend. It is a move you make — and the best moves make more moves possible.

Begin with something physics already shows us. Every arrangement that holds — an enzyme, a written proof, a road, a shared standard, a trusted institution — does more than exist. It makes things reachable that were not reachable before it. The enzyme opens a reaction chemistry alone would not give; the notation opens a proof the same idea, unwritten, could not. Configuration generates configuration. Build enough of it, and the rate at which new things become possible begins to climb with what is already there.

Stated plainly, and rarely stated plainly: value is not stored, it is generated. It is a flow into the adjacent possible. A forest, a language, a craft — none costs much to keep once it exists, yet each is a standing engine for making the future easier to inhabit. If this is right, the question worth asking of an economy is not how much did it produce, nor even how much does it hold, but: is the engine still running? Are we making moves that keep more moves available — or moves that quietly close the board?

This is a change of question, and economics has traditionally asked a different one: how should scarce resources be allocated? For most of its history that question rested on assumptions that rarely needed stating — that human labour was scarce, that energy was cheap, that material throughput tracked value, and that growth through extraction was both possible and desirable.

Those assumptions no longer hold simultaneously. Automation has loosened the need for human labour in many domains. Energy and material systems are pressing against hard ecological limits. The cost of maintaining complexity rises while the benefit of additional throughput falls. And many of the activities most central to a life — care, learning, coordination, meaning-making — map poorly onto wage labour or conventional productivity.

This essay does not propose a political programme, a welfare policy, or a moral doctrine. It does not argue for what people should want, nor attempt to legislate values. Its purpose is narrower and more fundamental: to examine what kinds of value accounting stay coherent once physical reality is taken seriously — and to follow, without flinching, where that leads.

The throughline is simple. When the underlying constraints change, the proxies used to coordinate economic life must change with them. What follows explores what happens when value is understood not as throughput or labour, but as configuration that preserves and expands future option space under bounded energy flux — and treats the economy's task as keeping that generative game open.

Every economy operates inside a physical envelope that is not subject to negotiation.

Atoms are conserved.
Energy is conserved; what economic activity actually consumes is exergy — the fraction of energy available to do work.
Entropy increases globally. Local order persists for as long as living systems maintain it, which is to say: for billions of years, ongoing.

These facts do not prescribe how humans must live, but they strictly limit how any system of value exchange can persist over time. Economic theories that ignore this envelope may function briefly, but they accumulate hidden fragility that eventually dominates outcomes.

The binding constraint is not the raw thermodynamic ceiling. Earth receives solar exergy at a scale far above current civilisational use — roughly four orders of magnitude above. The constraints that actually bind operate elsewhere: in the rate at which atmospheric, oceanic, and biospheric sinks can absorb waste; in the rate at which specific materials can be extracted or regenerated; and in the integrity of the biosphere on which all stable configurations depend. The ceiling is far. The floor of what the biosphere can absorb is close.

For most of human history, this envelope was invisible at the scale of daily life. Energy flows were small, populations limited, and the distinction between renewable income and finite savings was not sharply felt. Today, that distinction defines the economic landscape.

An economy that lives on continuous exergy income can, in principle, persist indefinitely. An economy that depends on drawing down finite exergy savings cannot.

This is not ideology. It is accounting.

Modern industrial economies are built on a profound and largely unexamined accounting error: the treatment of exergy savings as income.

Fossil fuels are not a revenue stream. They are a one-time inheritance—dense stores of available work accumulated over geological time and released over a few human generations. Treating this inheritance as ordinary income produces a false signal of productivity, abundance, and growth.

The apparent success of industrial economics is inseparable from this drawdown. Productivity gains, GDP expansion, and rising material standards of living are amplified by spending exergy savings while externalising the future costs of depletion, pollution, and system maintenance.

Once this distinction is made explicit, many puzzles begin to clarify. Productivity appears to stagnate because easy exergy has already been spent. Increasing effort yields diminishing returns because complexity maintenance rises as the exergy inheritance shrinks. Declining energy-return-on-investment as the easiest fossil stocks deplete is a documented and load-bearing factor in the growth slowdown of advanced economies since the 1970s — though not the only factor; technological exhaustion, capital-allocation shifts, and demographic transitions are also in play. The claim is that EROI decline is one major factor among several, not the single cause.

The decoupling literature deserves direct engagement. Some advanced economies have grown GDP while territorial material footprint has declined. Two distinctions matter. Relative versus absolute decoupling: most cited cases are relative — GDP grows faster than throughput, but both grow. Territorial versus consumption-based accounting: under territorial accounting, where goods are produced, some advanced economies show apparent absolute decoupling. Under consumption-based accounting, where goods are consumed, the decoupling mostly disappears — it turns out to be production offshoring, not actual reduction. The Hickel–Kallis meta-analysis (2019) and related work conclude that absolute decoupling of GDP from material throughput and CO₂ at the scale and rate required for 2 °C targets has no empirical precedent. This is consistent with the EROI story, not in tension with it.

An economy that confuses savings with income will misprice risk, undervalue durability, and overproduce activities that narrow future option space.

In a world of apparent abundance, throughput became an acceptable proxy for value.

More exergy consumed implied more work done.
More labour hours implied more contribution.
More money exchanged implied more prosperity.

For a time, these correlations held well enough to coordinate large societies. They were never perfect, but the underlying exergy surplus masked their flaws.

That masking is now gone.

Throughput is a cost, not a measure of success. It represents exergy dissipated and materials transformed—necessary inputs, but not intrinsic goods. When exergy savings are plentiful, wasteful configurations can persist unnoticed. When savings decline, the same configurations become destabilising.

A reasonable objection arises here. Roughly 80% of advanced-economy GDP now sits in services—healthcare, education, finance, retail, software, consulting—sectors that look throughput-light at the point of measurement. If the economy has already shifted toward configuration rather than extraction, has the throughput problem not already begun to resolve itself?

Three considerations matter. First, services are not as throughput-light as their direct accounting suggests. Healthcare depends on pharmaceuticals, imaging equipment, hospital construction, transport. Education depends on buildings, heating, lighting, transport, technology. Finance depends on enormous data centres. Retail depends on logistics infrastructure. Under consumption-based accounting, where embodied throughput is attributed to the consuming sector, the services economy carries far more physical weight than its GDP share suggests. Second, even sectors with genuinely low direct throughput remain embedded in throughput-dependent supply chains—the software engineer's laptop, heated home, and food are produced with substantial fossil inputs. Third, and most important: within a services economy, configuration matters more, not less. A services sector is by definition a configuration of human time, skill, institutions, and infrastructure. A well-configured healthcare system delivers more option space per unit of physical input than a badly-configured one. The shift toward services increases the weight of configuration-centred analysis; it does not decrease it.

At the same time, the most valuable human activities increasingly involve configuration rather than throughput: learning rather than extraction, care rather than replacement, coordination rather than production, reuse rather than disposal, meaning rather than volume.

The failure is not moral. It is categorical. Throughput-based proxies are being asked to do work they were never designed to perform.

Atoms persist.
Energy flows.
Configuration accumulates.

What distinguishes a book from paper, a skill from calories, a culture from a crowd, is not the amount of matter involved but the patterning of matter across time. Configuration is the arrangement of elements such that future actions become easier, richer, or more varied.

Biological evolution is the most familiar example. Over billions of years, the same atoms have been reconfigured into increasingly complex organisms, not by increasing throughput indefinitely, but by refining structures that store, transmit, and reuse information. Culture, language, science, and technology are continuations of this process by other means. The underlying mechanism is the same in each case: each configured arrangement makes further arrangements reachable that were not reachable before, and accumulated arrangement lowers the cost of further arrangement.

Configuration differs from throughput in three crucial ways. It can compound without proportional energy increase — the rate at which new configurations become reachable grows with what is already there. It can be copied and transmitted at low marginal cost. And it can expand future option space rather than merely satisfy present demand.

An economy that fails to recognise configuration as a primary source of value will systematically undervalue learning, care, prevention, coordination, and meaning—while overvaluing extraction, churn, and short-lived production.

This is not a moral oversight. It is a measurement failure.

Configuration is often mistaken for information, symbols, or data. This mistake leads to a shallow optimism: the belief that digitalisation alone can resolve physical constraint.

Configuration is broader and more demanding. It includes bodies, skills, relationships, institutions, infrastructures, and ecological arrangements. It always has a maintenance cost, and it can be good or bad.

Some configurations are brittle, requiring continuous high throughput to sustain. Others entrench dependency, reduce autonomy, or concentrate risk. Configuration is valuable only insofar as it reduces future throughput requirements while expanding future option space.

There is no guarantee that configuration trends toward goodness. The distinction between productive and destructive configuration must be made explicit. Without it, societies risk replacing one form of waste with another—burning energy not on extraction, but on maintaining fragile arrangements.

For most of human history, work served two functions simultaneously: it allocated access to life-supporting configurations, and it produced those configurations in the first place.

As long as these functions coincided, tying survival to labour appeared natural and efficient. The deeper distinction — that allocation was the load-bearing function, and production was the form it took — could be left unmade without practical loss.

That coincidence is breaking.

Automation, computation, and accumulated configuration increasingly allow societies to meet material needs with less human labour. At the same time, many of the most valuable human activities—raising children, caring for others, learning, maintaining social coherence—do not map cleanly onto wage labour or market exchange.

When access to resources remains conditional on "having a job" under these conditions, societies are forced to invent work that satisfies accounting proxies rather than real needs. The result is not productivity but performative activity: effort expended to justify entitlement rather than to expand option space — an allocator continuing to operate after the thing it was allocating has changed shape.

The question, therefore, is no longer how to ensure everyone works, but how to ensure everyone remains a participant in the configuration process.

If participation is foundational, it must be bounded honestly.

A stable economy cannot promise unlimited consumption, infinite growth, or unconditional abundance. It can promise access to a fair share of what the system can sustainably provide.

Universal participation means that every person retains the right to exist, learn, care, and contribute within the physical envelope defined by energy income, material limits, and ecological stability. It does not guarantee comfort, equality of outcome, or freedom from trade-offs.

Scarcity remains. Choices remain. What disappears is exclusion as a control mechanism—the use of deprivation as a means of coordination.

Within this baseline, differentiated contribution remains possible and desirable. Scarce, difficult, or necessary work can still be rewarded. What changes is that rewards are no longer the price of admission to life itself.

Participation is primary. Labour is instrumental.

When growth is measured primarily as increased throughput, it eventually collides with physical reality. When it is measured as increased configuration quality, the picture changes.

A viable economic objective under constraint is a no-regret one: prefer moves that keep open the configurations humans and the wider community of life depend on, across the horizons each depends on; reject moves that asymptotically empty that set. It is a criterion over moves, not a scalar to maximise — closer to first do no harm than to maximise this number.

It takes this form for a structural reason. The destruction of option space is typically fast and local; the expansion of option space is typically slow and systemic. Under that asymmetry, a fast-local irreversible loss cannot be reliably offset by a slow-systemic uncertain gain, and expected-value reasoning systematically misprices what is at stake. No-regret preservation is not conservatism by temperament; it is the rule that prices the asymmetry correctly.

The reframe shifts success away from volume and toward qualities such as resilience rather than speed, learning rather than accumulation, care rather than replacement, time affluence rather than consumption, and option preservation rather than short-term yield.

Under this framing, many familiar trade-offs resolve themselves. Activities that burn energy savings for transient benefit are revealed as expensive. Activities that increase coherence, capability, and future possibility become central, even when they produce little measurable output.

The economy does not grow or shrink in any absolute sense. It reorients.

This framework replaces:

• GDP as a primary measure of success
• Labour as the gatekeeper of legitimacy
• Throughput growth as progress
• Energy-blind pricing and accounting

It leaves intact:

• markets as coordination tools
• trade and exchange
• ambition and differentiation
• experimentation and innovation
• disagreement about values within constraints

Configuration economics does not abolish existing mechanisms. It re-situates them within a reality they can no longer ignore.

Nothing in this framework depends on universal agreement, moral enlightenment, or political virtue.

It depends only on the interaction between bounded energy income, accumulated configuration capacity, and the physics of living systems.

As old proxies fail, societies will either adopt accounting systems that respect these realities or continue to burn option space to preserve familiar categories. Both paths are available. Only one is stable.

The transition to configuration economics is not guaranteed.
But the failure of throughput-blind economics is.

If value is configuration and cost is throughput, then any serious economics must make those quantities visible. The next task, therefore, is not persuasion, but accounting.