There is a question buried inside most economic debates that rarely gets asked directly: what is the economy actually for?
Not in the trivial sense, not “to allocate resources efficiently” or “to maximize welfare.” The deeper version: of all the things humans spend energy, time, and attention on, how many of them are oriented toward physical survival and genuine functional need, and how many are oriented toward something else entirely?
This question matters because the answer isn’t obvious, and it’s substantially different from what introductory economics suggests. The modern economy is not primarily an engine for satisfying biological needs. It’s primarily a system for organizing status, maintaining shared belief, and managing the competing desires of a species that is, at its core, a symbolic animal.
Once you see that clearly, a lot of things that look irrational or parasitic start making a different kind of sense.
A Taxonomy Nobody Teaches You
Economics education tends to treat “value” as a monolithic concept. Utility is utility. Price reflects value. Demand reflects preference. The details of where preference comes from and what kind of value a given exchange is producing are usually left offstage.
But value, when you look carefully, is not one thing. It’s at least eight different things, operating through different mechanisms, with different relationships to physical reality.
Survival utility is the floor. Calories, clean water, thermal regulation, disease prevention. This is the category where physics has absolute veto power over economics. You cannot survive on a narrative.
Functional utility extends the floor into tool use: transport that moves you from one place to another, communication bandwidth that lets you transmit information reliably, materials that perform structural tasks with measurable efficiency.
Coordination utility is where things get interesting. Money, legal institutions, property rights, corporations, standards and protocols: these have no survival or functional utility in themselves, but they drastically reduce the cost of coordinating large numbers of people toward shared goals. A contract is a piece of paper with text on it. Its economic value is the behavior it enables, the trust it substitutes for, the conflicts it prevents.
Beyond coordination, we enter what might loosely be called the symbolic economy: positional value (derived from relative standing), narrative value (from stories about origins or futures), speculative value (from expectations about others’ expectations), and identity value (from alignment with self-concept and group membership).
Most real-world goods and transactions involve several of these simultaneously. A university degree has functional utility (skills), coordination utility (standardized credential), symbolic value (institutional prestige), positional value (relative rank), and identity value (alumni belonging). Trying to price any of these components independently is nearly impossible, which is why arguments about whether degrees are “worth it” are usually incoherent: they’re measuring multiple things simultaneously with a single number.
The Floor Is Smaller Than You Think
Here’s a number that tends to reorient how people think about this.
Global agriculture accounts for roughly 4% of global GDP. Food processing and handling adds energy consumption of approximately 1,933 TWh annually. These are the economic activities most directly connected to biological survival.
In high-income economies, agriculture often represents less than 2% of GDP. Services, which produce no physical goods and have no direct survival function, typically account for 70% or more of economic activity.
The global manufacturing sector, which includes all physical goods production from steel to pharmaceuticals to consumer electronics, generated roughly $16.83 trillion in 2024, about 15% of global GDP.
Everything else, the remaining 85% of economic activity in rich countries, is produced by coordination, symbolic, and informational systems that exist primarily in shared belief.
This isn’t a critique. It’s a description. The question is what it means that a species which requires calories, water, and shelter to survive devotes the overwhelming majority of its economic energy to things that exist primarily in social agreement.
Money: The Foundational Fiction
The clearest illustration of how symbolic value works is money, precisely because its fictional nature is now widely acknowledged, yet it continues to function at extraordinary scale.
Fiat currency has no intrinsic utility. A $100 bill is cotton and linen with ink. Its value comes entirely from the collective agreement to treat it as valuable, maintained by everyone who accepts it in exchange for real goods, labor, or time. This agreement is enforced through legal tender laws, tax obligations, and network effects (the more people accept a currency, the more useful it becomes to any individual), but ultimately it rests on belief.
When the belief fractures, as it did in Weimar Germany, Zimbabwe, and Lebanon in recent decades, the currency stops working regardless of what the central bank decrees. The paper doesn’t change. The belief does. And the belief is everything.
This seems like it should make fiat currency fragile. In some circumstances it is. But the fictional nature of money also makes it remarkably flexible in ways that matter for a modern economy. A gold-backed currency can only expand as fast as gold can be mined. A fiat currency can expand to meet the demands of an industrializing economy, a global financial system, or a pandemic-era fiscal response, because the constraint is institutional credibility rather than physical supply.
The fiction is load-bearing. The entire $92.6 trillion global economy runs on a coordination technology whose value is entirely a function of collective belief in it.
Understand this about money, and you start to see everything else differently.
Corporations don’t physically exist. When a court rules on a corporate liability case, it’s adjudicating the consequences of a shared fiction: a legal entity that has no physical form but has profound causal effects on the physical world. Nations don’t physically exist. The border between India and Pakistan is a set of agreements, maintained by mutual recognition and occasionally by military deterrence, but not by any physical property of the terrain.
Human rights don’t physically exist. Their force derives entirely from sufficient numbers of people agreeing that they exist and organizing institutions and norms around that agreement.
Harari’s observation that Homo sapiens is distinguished from other species by its capacity to believe in “imagined orders” is not poetic exaggeration. It’s a fairly precise description of what the coordination layer of civilization is made of.
The 50-to-1 Problem
The numbers become harder to dismiss when you look at the financial sector specifically.
The global derivatives market had a notional value of $846 trillion at mid-year 2025. The global manufacturing sector, all physical production, generated $16.83 trillion in value added. The ratio is roughly 50 to 1.
Derivatives serve real economic functions. Interest rate swaps protect borrowers from rate volatility. Currency forwards let businesses plan across exchange rate uncertainty. Agricultural futures let farmers lock in prices before harvest. The risk management function is genuine.
But a 50-to-1 ratio between the financial superstructure and the physical economy it ostensibly serves tells you that the superstructure has extended enormously beyond risk management into speculation, arbitrage, and what Keynes called the beauty contest: trying to guess what everyone else will think tomorrow, rather than assessing what things are worth today.
Keynes meant this critically, as evidence of a distortion in financial markets. But taken as pure description, his framing is remarkably precise about what financial markets primarily are. They are social prediction markets: mechanisms for betting on collective psychology about the future, operated at $846 trillion scale.
When markets work well, these social predictions pull future information into current prices, allowing better resource allocation. When markets work badly, they amplify the very psychology they’re trying to predict, creating feedback loops that disconnect prices from any relationship to underlying reality.
The fine art market, which sold about $57.5 billion in works in 2024, is a useful extreme case. A canvas with paint on it has no meaningful functional utility. Its valuation is entirely a function of historical narrative (who made this, when, in what context), social meaning (what does owning this signal about you), and positional scarcity (how many people can own the same thing). The market for contemporary art is a market for status signals that happen to be attached to physical objects.
This is not a critique of art. It’s an observation about what’s being priced. And the same pricing logic, the same mixture of narrative, positional, and identity value, operates throughout the luxury sector, the educational credential market, the real estate market in prestige locations, and increasingly in digital markets for attention and influence.
The Girard Layer
Most economic models of value treat desire as exogenous: preferences arrive from somewhere outside the model, and the economy’s job is to satisfy them efficiently. The question of where preferences come from is either ignored or attributed to some combination of genetics and individual psychology.
René Girard’s contribution was to argue that this is backward. Desire is not individual. It’s fundamentally social and imitative.
His model is triangular: a subject desires an object because a model (another person, a social group, a celebrity) desires or possesses it. The model confers desirability on the object. Without the model’s interest, the object might be valued very differently or not at all.
Once you apply this to markets, several things become clearer.
Brand premiums exist not because consumers rationally assess functional quality differentials, but because brands function as models: they tell you who desires this, and desiring it yourself places you in relation to those people. The $15,000 watch is not priced for timekeeping. It’s priced for the social meaning of being someone who has what admired people have.
Speculative bubbles follow mimetic cascades: the more people buy an asset, the more desirable it becomes, because others’ interest is itself evidence of value (or perceived evidence, which does the same work). The underlying asset barely matters during the mania phase. What matters is who else wants it.
Luxury markets deliberately manufacture model scarcity to intensify desire. By limiting supply, burning unsold inventory, and curating who is seen wearing or carrying their products, luxury firms maintain the mimetic mechanism that makes their goods worth wanting. The scarcity is often artificial, designed, and maintained as infrastructure for desire generation.
This is also why the Girardian analysis of speculative bubbles ends with scapegoating. When mimetic rivalry has driven prices to absurd heights and the narrative fractures, the subsequent loss and anger are collective, generated by the same imitative mechanism that drove the boom. But collective guilt is cognitively unsustainable. It resolves through the concentration of blame on specific actors: certain banks in 2008, specific exchanges or founders in the crypto corrections, particular regulatory agencies after every financial crisis.
The scapegoating mechanism doesn’t explain the crisis. It manages the social fallout from a crisis whose causes were distributed across everyone who participated.
What Thermodynamics Adds
All of the above involves following chains of reasoning through social science. Physics offers a blunter instrument, and it produces numbers that are harder to rationalize away.
Energy spent is energy spent. The laws of thermodynamics don’t distinguish between energy devoted to keeping people fed and energy devoted to maintaining the credibility of national borders. Both contribute equally to carbon emissions.
With that in mind:
The military-industrial complex consumes approximately 6,691 TWh annually. Food processing and handling uses about 1,933 TWh. We spend more than three times as much energy maintaining the fiction of national sovereignty as we spend ensuring the physical survival of the species.
The global banking system and Bitcoin network together consume roughly 425 TWh to maintain their respective ledgers of who owns what. Digital advertising infrastructure generates 3.5% of global greenhouse gas emissions, more than the aviation sector, in service of shaping consumer desire.
The ICT sector overall, which maintains the infrastructure for the digital economy including the symbolic and attention economy functions, consumes about 915 TWh annually (as of 2020, growing rapidly with AI workloads).
These numbers make something visible that economic accounting doesn’t: the physical cost of the symbolic economy is already enormous, and it’s growing faster than the physical economy itself.
The ecologically disturbing implication is that a significant portion of the anthropogenic disruption driving climate change isn’t the unavoidable thermodynamic cost of human survival. It’s the cost of maintaining status hierarchies, creating and directing desire, securing national fictions, and running the infrastructure of financial belief systems.
This doesn’t change the climate accounting. The atmosphere doesn’t care what the energy was for. But it changes what solutions are available. If the problem were purely about the thermodynamic cost of survival, the solutions would be purely technical. When the problem is substantially about the cost of maintaining symbolic systems, some of the solutions are institutional and cultural.
The Paradox That Makes This Harder to Dismiss
The straightforward critique of symbolic value is that it’s wasteful: resources spent on status signaling, narrative manipulation, and financial abstraction are resources not spent on genuine human welfare.
The historical record makes this critique uncomfortable.
The Space Race was a prestige competition. Kennedy’s mandate had no functional justification. It was a symbolic gesture toward ideological dominance, pure and simple.
The spillovers were: integrated circuit manufacturing, satellite communications, fuel cell technology, GPS, materials science advances, and the mathematical foundations of modern computing. The digital economy we inhabit now is, in part, a residue of a prestige race conducted between two ideological rivals who wanted to win an argument in the most expensive way imaginable.
The same pattern appears in the history of industrial research. Bell Labs was one of the most productive research institutions in human history, generating the transistor, information theory, cellular networks, and laser technology. It was funded by a regulated monopoly that had both the financial capacity and the reputational motivation to do science that had no obvious near-term commercial application. The symbolic goal, maintaining AT&T’s identity as a technology leader, generated material spillovers worth vastly more than the investment.
High-frequency trading is a more contemporary example. The latency race in financial markets is, in terms of competitive advantage, nearly self-canceling: when everyone has invested in the fastest infrastructure, the advantage disappears and everyone needs to invest in the next generation of speed. It looks like pure rent-seeking. But the infrastructure demands of the race have driven substantial improvements in global fiber-optic networks and computational hardware.
The uncomfortable inference is that some symbolic competition is genuinely generative: it drives investment beyond what rational calculation would justify, and the excess generates functional spillovers that benefit everyone.
The question is how to distinguish generative from extractive symbolic competition. And I don’t have a clean answer to that. The Space Race was generative. Much of the financial derivatives superstructure is probably not. Fast fashion’s prestige cycles are generating mostly emissions. The AI infrastructure race is probably somewhere in between: driving functional capability advances while also consuming growing shares of global energy on status and attention competition.
What Changes When Content Becomes Free
There’s an unresolved question in all of this that I find more interesting than most of the others.
One of the economists who studies this asks: in an AI-rich world where content and symbolic production become nearly costless, what new forms of scarcity emerge?
The answer, I think, is: scarcity of verified humanity, authentic experience, physical presence, and trusted provenance.
When AI can produce any text, image, or video at near-zero cost, the things that become genuinely scarce are the things that AI can’t replicate: the embodied experience of being in a place, the assurance that a particular human made or did a particular thing, the trust that a particular relationship is not simulated.
This would be an interesting inversion. The symbolic economy has for decades been running on manufactured scarcity, brands limiting supply, credentials controlled by institutions, attention mediated by platforms. When content production democratizes completely, the manufactured scarcity may be less powerful than the natural scarcity of verified human presence and authentic origin.
That’s speculative. But it follows from taking seriously the logic that symbolic value is real value, and that as material abundance increases, symbolic scarcity becomes the primary economic battleground.
The Honest Summary
Human civilization has never been primarily about survival. Even hunter-gatherer societies invested significantly in gift economies, ritual, and symbolic competition. The potlatch ceremony of Northwest Coast peoples involved the conspicuous destruction of wealth for social dominance. The Kula ring of the Trobriand Islands maintained elaborate ceremonial circuits of shells that had no functional use but served to cement alliances and confer prestige.
The modern economy has expanded this symbolic layer enormously, not because humans became less rational, but because as material abundance grew, the fraction of resources available for non-survival purposes grew with it. When you have enough food, you start competing for status. When you have enough status, you start competing for narrative.
This is what you get when a symbolic species achieves sufficient material productivity: a vast superstructure of belief, competition, and meaning-making, consuming an increasing share of the energy budget.
Whether that’s good depends on what the symbolic competition generates. When it produces scientific progress, cultural richness, better institutions, and coordination mechanisms that reduce the cost of global cooperation, it’s probably worth what it costs.
When it produces ecological overshoot, pure positional arms races that leave everyone worse off in absolute terms while sorting them in relative terms, and attention architectures optimized for addiction rather than meaning, it’s probably not.
The problem is that the symbolic economy doesn’t distinguish between these outcomes in advance. It generates prestige races indiscriminately. Some of them happen to be the Space Race. More of them are probably fast fashion.
What matters is whether the institutions and incentive structures we design push symbolic competition in the generative direction. That’s ultimately a political and cultural question, not an economic one. But it’s hard to even ask the question clearly if you’ve accepted the premise that the symbolic economy is just noise layered on top of the “real” one.
It isn’t noise. It’s most of what the economy is.