Post Scarcity Economics

This brief opens Category IX, and the category is new to this map. The eight categories before it ask what can be built — propulsion, habitats, biology, minds, energy, climate, infrastructure, institutions. Category IX asks the question those leave untouched: how the output gets produced, claimed, distributed and counted. It is the part of the frontier where the constraint is not a material or a law of physics but an accounting rule, a transfer schedule, a tax base or a price. The subjects behave differently from the rest of the map — the good studies are randomised rather than instrumented, the failure record is legislative rather than experimental, and several of the most confident claims in circulation have no measurement behind them at all.

Post-scarcity economics is the category's theoretical slot: it asks whether scarcity itself can stop binding. Two neighbouring briefs take the practical questions — Universal Basic Abundance examines one specific instrument for a world where scarcity still binds, and Wealth Distribution Systems examines the mechanisms that move claims on output and what they measurably do. This slot sits underneath both. If its strong version were true, neither of the others would have a subject.

The claim decomposes into three propositions that are almost never separated, and separating them is most of the work. (a) Some goods have had their real cost collapse by orders of magnitude. (b) That collapse generalises, until scarcity ceases to bind for the goods people actually care about. (c) Once it does, the allocation problem has no content and the institutions built around it — prices, property, wage labour — go vestigial. Almost the entire persuasive force of the argument runs through the step from (a) to (b), and in essentially every popular statement of it that step is made by assertion. The thinness of the evidence is not a gap in this brief; it is the brief's finding.

Established Lighting is the cleanest case in the whole economic record, and it is genuinely spectacular. On Fouquet's long price series, the inflation-adjusted price of a million lumen-hours fell from about £34,000 in the early fourteenth century to £2.15 in 2023 in year-2000 prices — a 16,000-fold decline, with more than 99.9% of it since 1700. For the specific good “artificial light at night”, the budget constraint that governed human behaviour for all of prior history has effectively vanished in rich countries. Established Note precisely what the case does not show. Light became cheap, not free, and it became cheap through a two-century sequence of innovations in appliances, fuels, infrastructures and institutions — not through the arrival of a general condition of abundance that lighting happened to share in.

Established Software is the second clean case, and it is better than lighting because the counterfactual can be priced. Hoffmann, Nagle and Zhou estimate the supply-side value of widely-used open source software — the one-time cost of recreating it once — at $4.15 billion at global average wages, bracketed by $1.22bn at Indian wages and $6.22bn at US wages. They estimate the demand-side value — what firms would have to spend if each had to recreate the open source it actually uses — at $8.80 trillion at global average wages, with a range of $2.59tn to $13.18tn across wage scenarios. The ratio is roughly 2,000 to 1. This is as close to a measured demonstration of the non-rival-goods abundance argument as exists anywhere.

Established The same paper carries the finding that complicates it, and it is the cutting one: 96% of the demand-side value is created by 5% of open source developers. The authors' own distributional analysis shows this arises because top contributors participate in far more projects, not because a handful of repositories are hyper-productive. The good is non-rival. The input is not. An abundance that rests on a few thousand identifiable people is an abundance with a scarce factor sitting directly underneath it, and the factor is unpriced rather than absent. Frontier That structure — free output, concentrated and uncompensated input — recurs everywhere the zero-marginal-cost argument is made, and no version of the argument has a theory of what happens when the input is exhausted.

Established Zero marginal cost holds for the reproduction step and for nothing around it. Copying a post costs nothing; deciding whether the post may exist costs a workforce the size of a mid-sized city. Blackwell's 2025 study of trust-and-safety practitioners — 33 professionals with 230 collective years of experience across Meta, Twitter, Google, Discord and TikTok — records Meta alone reporting 40,000 “safety and security” personnel, and the Trust & Safety Professional Association estimating over 100,000 professionals globally. Her participants report that automated tools “flatten behaviors” and cannot make contextual judgements, so human review persists for the hard cases. That cost scales with volume — which is to say, with the very abundance that made the copying free. Frontier Attention is the same point in its purest form: total human attention is fixed at roughly sixteen waking hours per person per day and does not respond to technology. An economy whose only remaining scarce input is time has not dissolved the economic problem. It has re-specified it.

Established Energy is where the abundance case is quantitatively strongest. Way, Ives, Mealy and Farmer report that solar photovoltaic costs fell about 15% per year over 2010–2020, having fallen “roughly exponentially at a rate near 10% per year” over several decades and by more than three orders of magnitude since 1958; lithium-ion battery costs fell 12% per year from 1995 to 2018 while production grew 30% per year; wind deployment grew 23% per year over thirty years. Their method is a stochastic version of Wright's law, backtested on more than 50 technologies and roughly 6,000 forecasts. Established Their most damaging result is about forecasters rather than technologies: over 2010–2020 integrated assessment models projected solar investment costs would fall at a mean of 2.6% per year, every projection under 6%, against a realised 15%. Frontier Their headline — roughly $26 trillion in median net present savings from a fast transition against continued fossil dependence through 2070, at a 1.4% discount rate — is a projection contingent on the discount rate and on the experience curves continuing, and INET Oxford advocates the fast transition, so the interest runs with the finding.

Established Now the counter-fact, and it is the single most important number in this brief. US real retail electricity prices in 2024 were the same as in 2019 and about 8% below 2010. Lawrence Berkeley National Laboratory and The Brattle Group report nominal prices up 23% over 2019–2024, which is to say retail prices “largely tracked inflation”. Established Generation costs fell in real terms across that window; distribution and transmission capital spending rose in real terms in every region and ate the saving. The gap between “solar modules fell 15% a year” and “your bill fell 8% in fourteen years” is the entire abundance dispute in miniature. Production cost is not delivered cost, and the non-cheapening components — poles, wires, hardening, permitting, land, labour — are precisely the ones that do not sit on an experience curve.

Established The Baumol effect is the general statement of that fact, and it is the strongest structural argument against the framing. The objection is not that technology will slow down. It is that technology speeding up in one sector mechanically raises the relative price of the sectors it does not touch — health, education, care, courts. Tabarrok's account carries the record: a Beethoven quartet required 2.66 labour-hours in 1826 and still does, while economy-wide real wages rose roughly 23-fold by 2010 (average hourly production-worker wages $1.14 in 1826 to $26.44 in 2010), so the opportunity cost of staging the identical performance rose from $3.02 to $70.33. Between 1950 and 2016 quality-adjusted car prices fell while repair costs rose sharply, and clothing prices fell substantially while tailoring rose to roughly six times clothing costs from near parity in 1950. Established Nordhaus's econometric work found industries with low productivity growth show percentage-point-for-percentage-point higher growth in relative prices, with productivity differences explaining about 85% of relative price variance. Tabarrok writes from the Mercatus Center, which has a stated free-market position; the chapter nonetheless argues against regulation-based explanations of rising prices, which is against its own side's usual line.

Established And the same author supplies the best pro-abundance evidence in the literature, which is why the Baumol material cannot simply be filed as the pessimistic case. The mechanism predicts rising prices and rising quantities simultaneously, because the productivity gains elsewhere raise incomes. Baumol and Bowen predicted the arts would decline. Instead, between 1960 and 2016, actor employment grew 4-fold against 2.3-fold labour-force growth, musicians 7-fold, entertainers 17-fold, and US symphony orchestras went from 65 in 1965 to over 1,000 by 2020. Relative price up; absolute consumption also up. That is a real, if partial, abundance result, and it is a much stronger empirical version of the post-scarcity claim than the zero-marginal-cost version — which is a claim about a price, where this is a claim about a quantity people actually got.

Established The forecasts with dates on them have a consistent scorecard: direction survives, magnitude and date do not. Keynes in 1930 predicted output per head four to eight times higher within a century and a fifteen-hour week. Crafts's assessment: on output he was right — UK real GDP per capita was about 5.15 times its 1931 level by 2011, inside the band. On hours he was badly wrong — annual hours fell from about 2,500 in 1931 to 1,840 in 2011 against his implied 780. Frontier But Crafts's explanations are not the usual ones and they partly rescue the prediction: expected retirement lengthened from 4.66 years in 1931 to 16.37 in 2011, requiring sustained pre-retirement work; longevity gains were underestimated; married women's employment rose from 10% to 49%. On a lifetime-equivalent basis holding 1930s demography fixed, the fifteen-hour week implies a 53% increase in lifetime leisure — less than the 58.4% that actually accrued. The gains were taken as leisure. They were taken at the end of life rather than the end of the week.

Established The automation forecasts age the same way, and the record is now long enough to score. Frey and Osborne in 2013 put 47% of US employment at high risk of computerisation. Arntz, Gregory and Zierahn at the OECD asked the same question with a task-based rather than occupation-based method and got 9% across 21 OECD countries, ranging from 6% in Korea to 12% in Austria and Germany; their illustrative case is that retail salespeople score 92% automatable occupation-wide while only 4% of actual retail workers do jobs with neither group work nor customer interaction. Established Then the outturn: Georgieff and Milanez find employment growth over 2012–2019 of 6% in high-risk jobs against 18% in low-risk jobs. High-risk employment grew. It grew a third as fast, which is a real and adverse relative result, but it is not displacement. Frontier The authors add that employment rates of low-educated workers did not fall relative to other groups because the supply of low-educated workers contracted to match — a compositional explanation that would not have been visible from the 2013 method at all.

Established What the failure record contains, stated exactly. There is no case of a society that reached and sustained a state in which allocation stopped being a live problem. What exists instead is: goods whose real price fell by three to five orders of magnitude while remaining priced — lighting, computation, long-distance communication, calories in rich countries; one large class of goods produced and distributed outside the price system at trillion-dollar scale with production concentrated in a few thousand people; and repeated forecasts of imminent post-work or post-scarcity conditions at roughly generational intervals since 1930, none of which arrived on schedule. Speculative The absence of a success case is not proof that propositions (b) and (c) are false. It is the reason they are tagged speculative and handwave here rather than frontier: there is nothing to contest, because there is nothing to measure.

Frontier Does cost collapse in non-rival goods raise or lower the relative price of the goods people most value? The Baumol mechanism says it raises them, and the mechanism is not in dispute. What is in dispute is the welfare implication: an economy of near-free information plus expensive care is one in which scarcity has concentrated rather than stopped binding, and the concentrated part is the part welfare depends on. Speculative Nobody has run the counterfactual — there is no study that takes a good whose relative price rose in a stagnant sector and asks what its price would have been had the dynamic sector not sped up. The identification problem is hard but not obviously harder than others this literature solves.

Frontier The same mechanism read optimistically is the live question, and it has a clean test that has not been run. Tabarrok's arts evidence shows relative prices rising while quantities rose four-to-seventeen-fold. Whether the same holds for health and education access — where the relative-price rise is the political grievance — is contested and testable with existing data. Frontier If quantities of care consumed have risen in proportion to their relative price, the cost-disease complaint is a complaint about an accounting artefact. If they have not, Baumol pessimism wins on the goods that matter.

Speculative Has an abundance of any good ever reduced conflict over its distribution? The post-scarcity claim implicitly assumes distributional politics is a function of scarcity. No study tests that on a good that actually became abundant. This is a strange hole: lighting, calories in rich countries and bandwidth are all available as cases, and the political-economy literature has not used any of them for this question. Handwave The step from “there is enough” to “the fight stops” is the least examined link in the entire argument and the one the argument most needs.

Frontier Whether satiation exists at all. Robbins defined economics as the allocation of scarce means among competing ends; the post-scarcity claim requires that the ends are finite. A standing methodological position — rarely written down because economists treat it as obvious — holds that human ends are unbounded, so no level of means dissolves the problem and post-scarcity is definitionally incoherent rather than empirically false. Speculative As stated it is not testable. A version that would be: measure whether consumption of a good stops rising in a population whose means for that good are effectively unconstrained. The measurement runs straight into the wellbeing-measurement problem, which a separate brief in this category owns.

Frontier Whether experience curves survive at scale. Wright's law is an empirical regularity across roughly fifty technologies, not a law; Way and colleagues model it stochastically precisely for that reason. Every projection of continuing energy abundance inherits that assumption. Speculative Nobody has identified the conditions under which a curve breaks — materials constraints, land, interconnection queues, or simply the exhaustion of the learning available in a given design — and the deployment record is not yet long enough at current scales to tell.

Frontier The most interesting live hypothesis is that scarcity is socially produced rather than technical — that intellectual property, zoning, occupational licensure and enclosure manufacture scarcity in goods that are technically abundant. The open source and proprietary software contrast is the natural experiment sitting in plain sight, and it is confounded in every direction that matters. Frontier What would settle it is narrow and doable: a jurisdiction removing one artificial scarcity, with output measured before and after. Housing supply reform is the case with the most attempts and the least clean measurement.

Speculative And the persistent fringe position: energy is the master resource, so cheap enough energy dissolves everything else — desalination, synthesis, vertical farming, materials, all become energy problems and energy becomes free. The cost curves in the first half of the argument are real. The propagation step is untested, and the retail-price record is direct counter-evidence for the delivery link in the chain. Handwave A version that would be evidence: a region with genuinely cheap delivered energy showing across-the-board real price declines in energy-intensive goods. Iceland and Quebec are the obvious candidates and neither has been analysed this way.

Established The binding constraint on delivered abundance is the set of components that do not sit on an experience curve, and in electricity they have been measured. Distribution and transmission capital spending rose in real terms across every US region over 2010–2024, offsetting a real fall in generation cost so completely that the retail price change was indistinguishable from inflation. Frontier Wires, poles, land, hardening, permitting and skilled field labour have no learning rate anyone has demonstrated; some of them have a negative one, because the easy corridors are used first. Any abundance argument that stops at the module price has stopped at the wrong number.

Established The second bottleneck is the scarce input hiding under the non-rival good. Open source is the best-evidenced abundance in the record and 96% of its demand-side value comes from 5% of its developers, who are concentrated by participating across many projects rather than by writing unusually productive repositories. Frontier That is a bus-factor problem at the scale of the world economy: the maintenance of $8.8tn of demand-side value is a labour input with no funding channel proportional to it, and the two largest attempts to build one — foundation stipends and corporate employment of maintainers — have never been evaluated against the concentration figure.

Established The third is moderation, which is the cost of deciding what may exist rather than of making it. Meta alone reports 40,000 safety and security personnel and the professional association estimates over 100,000 globally, in a workforce whose practitioners report that automated tools cannot make contextual judgements. Frontier This cost rises with volume, so it rises with abundance, and it is the clearest case in the record of a constraint created by the removal of another.

Frontier The fourth is measurement, and it cuts against the sceptics as much as the enthusiasts. National accounts price goods at transaction value, so a good that becomes free leaves the accounts entirely: the $8.8tn demand-side value of open source appears in no country's GDP. Speculative Estimates of the consumer surplus from free digital goods exist and vary by more than an order of magnitude depending on method, and this brief does not adopt any of them. The honest statement is that the abundance the framing describes is partly invisible to the instrument used to deny it — which is a real problem for both sides and is not resolved by picking the number you prefer.

Established Nothing on this map produces a result this brief waits on, and no typed depends-on edge is claimed. Post-scarcity is not blocked by a discovery. What it waits on is measurement and institutional machinery: a delivered-price decomposition that separates experience-curve components from the rest, an accounting treatment for goods consumed at zero price, and a funding channel proportional to the concentrated labour input underneath the best abundance case. All three are things the world could choose to supply and has not.

Frontier What it waits on from research is narrow and specific. A counterfactual on relative prices in stagnant sectors; a quantity series for health and education access to test the optimistic reading of Baumol against the pessimistic one; a satiation measurement that does not reduce to a self-reported wellbeing question; and a break condition for experience curves. Speculative None of these is a frontier physics problem. Each has been available to run for a decade or more, and the reason none has been run is that the question is treated as rhetorical rather than empirical — which is itself the most reliable finding about this literature.

Established The highest-value experiment is also the cheapest: decompose delivered prices into experience-curve and non-experience-curve components across jurisdictions, using data that already exists. The Berkeley and Brattle work does this for US electricity for one period. Doing it for several countries, several decades and several goods would convert the central dispute from an argument about anecdotes into a measured ratio. Frontier The prediction the abundance case would have to make is specific: the non-curve share should be falling somewhere. In the one place it has been measured it is rising.

Frontier Second: a natural experiment on manufactured scarcity. Take one artificial constraint — a licensure rule, a zoning limit, a patent term — removed in one jurisdiction and not in a comparable one, and measure output, price and consumption. This is a standard design that the post-scarcity literature has never used, and it is the only route to testing the “scarcity is socially produced” hypothesis that does not rest on the confounded software comparison.

Frontier Third: pre-register a displacement forecast. The 2013 forecasting exercise and its successors share an evidentiary structure — expert judgement over occupational task descriptions, extrapolated without a stated date or a falsification condition. The current generation of AI-displacement claims has the same structure and will age the same way unless someone commits to a horizon, a measure and a threshold in advance. Speculative That is not a research programme; it is a discipline, and its absence is why the 47% figure survived a decade of contradicting data.

Speculative Fourth, and hardest: test whether abundance in a good reduces conflict over its distribution. Lighting, calories and bandwidth all crossed from scarce to effectively abundant within measurable political records. Whether distributional conflict over each declined, persisted or migrated is answerable from existing data and has not been asked.

Established Making a good free at the point of use is an infrastructure problem, not a production problem, and the infrastructure is where the money is. The electricity case makes this concrete: generation cost fell in real terms, delivery capital spending rose in real terms in every region, and the consumer saw nothing. Frontier Any engineering programme aimed at abundance in a physical good has to attack the delivery layer — interconnection throughput, right-of-way, undergrounding cost, workforce — and those are the components with the weakest learning rates and the strongest institutional friction on this map.

Frontier Storage is the one place where an experience curve is attacking a delivery-side cost. Lithium-ion fell 12% a year from 1995 to 2018 while production grew 30% a year, and storage substitutes for some transmission and some peaking capacity. Whether that substitution is large enough to bend the delivered price is an open empirical question with a decade of data now accumulating. Speculative It is the single best candidate for a delivered-cost collapse that the retail record has not yet seen.

Established For non-rival goods the engineering requirement is maintenance, not production. The replication cost of the world's widely-used open source is $4.15bn at global average wages — recreating it once is affordable to a mid-sized company. Keeping it alive is a recurring labour cost carried by a small concentrated group with no proportional funding channel, and that is the engineering problem the abundance framing consistently misses because it is priced at zero.

The boundaries inside Category IX, stated because these three briefs are being read together for the first time. This slot is the theoretical claim: whether scarcity can cease to bind. Universal Basic Abundance is a specific instrument for a world where it does bind — and a working basic income is evidence against post-scarcity in the strong sense, because it is a rationing mechanism. Wealth Distribution Systems covers mechanisms and their measured effects on stocks of claims; the wealth-share measurement dispute belongs there and is deliberately not imported here, exactly as the Baumol material on relative prices belongs here and not there.

Two further Category IX slots are being authored in parallel and this brief does not reach into either. Human Flourishing owns wellbeing measurement — including the question of whether “more goods did not make people happier” is a measurable claim at all, which is superficially a post-scarcity argument and is not this brief's to make. Human Development Metrics owns composite indices and what publishing them does to policy. The GDP-critique literature sits on that boundary: “GDP misses the consumer surplus from free goods” is a real argument and is noted in section 4, but index construction and Goodhart effects are not adjudicated here.

Elsewhere on this map: Commercial Fusion and Advanced Fission, whose cost cases are the energy-abundance argument in its strongest form; Energy Storage Revolutions, which carries the one experience curve pointed at a delivery cost; Space-Based Solar Power and Planetary-Scale Energy Systems, where the master-resource hypothesis is stated in engineering form; Automated Construction Systems and Autonomous Supply Chains, the production-side automation claims; Circular Infrastructure Systems and Industrial Ecology, where material rather than monetary scarcity is the subject; Artificial General Intelligence, the current carrier of the displacement forecast; and Future Education Systems, one of the two stagnant sectors the Baumol argument turns on.

Established The interested parties here run in both directions and it is load-bearing to say which. INET Oxford, whose forecasting paper supplies the cost curves, advocates a fast energy transition — interest runs with the finding. Tabarrok writes from a think tank with a stated free-market position, and the chapter argues against regulation-based explanations of rising prices, which cuts against his own side's usual line. The trust-and-safety study's author is a former platform researcher with an interest in the field's expansion. The open source valuation is a working paper by authors who publish on open source value. Frontier The Berkeley and Brattle retail-price material is the least interested source in the brief and carries the finding that cuts hardest, which is part of why it is given the weight it is.

Established The institutional machinery that would settle the argument does not exist and is nobody's job to build. No statistical agency produces a delivered-price decomposition separating experience-curve components from the rest. No national accounting standard values goods consumed at zero price, so the best-measured abundance in the world is invisible to the measure used to assess whether abundance is happening. Frontier Both are institutional choices, not technical limits, and both are recorded as typed requirements below.

Frontier And the manufactured-scarcity question is institutional all the way down. Intellectual property terms, zoning envelopes, occupational licensure and spectrum allocation are the instruments that determine whether a technically non-rival or technically cheap good is available cheaply. The literature that studies each of them separately is large; the literature that adds them up into a statement about how much of observed scarcity is chosen does not exist. Speculative That total is the number the socially-produced-scarcity hypothesis needs and nobody has attempted it.

Established The distributional question does not wait for post-scarcity; it is sharpest in the abundance cases themselves. Open source delivers $8.8tn of demand-side value, and the 5% of developers producing 96% of it are largely uncompensated by the firms capturing that value. An abundance whose producers are unpaid is a transfer, not a miracle, and the framing's language of dissolved scarcity obscures who is bearing the cost. Frontier The same structure appears in moderation labour, which is concentrated, outsourced, psychologically hazardous and priced as an overhead on a free good.

Frontier The claim that abundance dissolves politics is the ethically consequential one, and it is unsupported. If distributional conflict is not a function of scarcity — and no study has tested whether it is — then a politics that defers questions of distribution until abundance arrives is deferring them indefinitely. Handwave Every strong statement of post-scarcity contains this deferral, usually implicitly, and it is the step at which a technological argument becomes a political one without announcing it.

Established There is a countervailing ethical fact and it should not be lost in the scepticism. Lighting really did fall 16,000-fold, and the fall was overwhelmingly a gain for people who previously could not read after dark. Software really is produced and distributed outside the price system at trillion-dollar scale. Arts employment really did rise four-to-seventeen-fold while its relative price rose. Frontier These are large welfare gains delivered by exactly the mechanism the framing describes, and a brief that only recorded the counter-evidence would be describing a different record than the one that exists.

Established The terminal position is a declared tie, and the tie is the finding rather than a failure to reach one. Some scarcity genuinely dissolved: lighting by four orders of magnitude, software into an $8.8tn non-market flow, arts consumption by a factor of four to seventeen while its price rose. And the aggregate economic problem did not dissolve: real retail electricity prices are flat through the largest generation-cost collapse in history, the goods that resist cheapening are the ones welfare depends on, and the abundance cases each rest on a scarce input that the framing does not price. Both are supported by the record and they are not in contradiction. Picking one requires ignoring half the measurements.

Frontier What follows is a change of question, not a compromise between the answers. The useful question is not whether scarcity ends but where it relocates, because the record shows it relocating every time: from lumens to wires, from copying to moderating, from making to deciding, from goods to attention. Speculative A civilisation that got very good at producing things and never got better at delivering, adjudicating or attending to them would look exactly like the one in the data — spectacular production statistics and flat delivered prices.

Speculative The long-run version of the optimistic case is the one worth taking seriously, and it is Tabarrok's rather than Rifkin's. If abundance arrives as more rather than as cheaper — more performances, more care, more education consumed at higher relative prices out of higher incomes — then the post-scarcity condition would arrive unrecognised, because every price index would be rising while consumption rose faster. Frontier That is a testable civilisational claim and it is the one the arts data supports. Whether it extends to health and education is the open question on which the whole framing turns.

Handwave And the strong version — that prices, property and wage labour go vestigial — has no evidence at any scale, in any period, in any polity. Not a weak case: none. That is recorded here rather than omitted, because a map of the frontier should mark the places where a widely held position is held entirely on assertion, and this is the largest of them in Category IX.

These horizons track the ageing of dated claims, the deployment of experience curves and the measurement work that is not being done — not a technology readiness level:

• 10 yr: Established One date arrives inside this window and is already effectively scored: Keynes's century ends in 2030. Output per head landed inside his 4–8 band at about 5.15 times; the fifteen-hour week did not arrive and will not, with annual hours at roughly 1,840 against an implied 780. Frontier Expect the delivered-versus-production-cost gap in electricity to persist or widen, because distribution and transmission capital programmes are already committed and their unit costs are not falling. Frontier Expect storage to be the one component with a real chance of bending a delivered price, and expect the current generation of AI-displacement forecasts to reach their implied horizons with the same shape of outturn as 2013's: adverse relative employment growth in exposed categories, no absolute displacement at the forecast magnitude.
• 25 yr: Speculative If the optimistic reading of Baumol holds, this is the window in which quantities of health, education and care consumed would visibly rise against rising relative prices, which is the only way the abundance claim can be true in the sectors that matter. Speculative If it does not hold, the stagnant-sector share of income keeps rising and the political salience of cost disease keeps rising with it. Handwave Claims that a specific technology — fusion, general-purpose robotics, general artificial intelligence — collapses the stagnant sectors on a stated date belong here and none of them carries a falsification condition. Dated claims of this shape have a fifty-year record of arriving late or not at all.
• 50 yr: Speculative At this horizon the only defensible statement is conditional: if experience curves in energy and storage continue and if the delivery layer is rebuilt at falling real unit cost, delivered energy abundance is possible and the master-resource hypothesis becomes testable for the first time. Handwave Both conditions are assumptions, the second has no historical precedent at national scale, and every published version of this claim asserts rather than argues the propagation from cheap energy to cheap everything.
• 100 / 250+ yr: Handwave Beyond useful forecasting. The lighting series is the only data point in the corpus at this horizon — four orders of magnitude over seven centuries, achieved through a sequence of unrelated innovations no contemporary could have forecast. Handwave It supports the claim that enormous cost collapses happen and supports nothing whatever about when the next one arrives or what it will be. Anyone extrapolating a post-scarcity date from it is reading a single time series as a schedule.

• Depends on Nothing on this map. No brief in this corpus produces a result this one waits on: the abundance cases are already measured, the counter-evidence is already measured, and the disputed steps are untested rather than unresolved. No typed depends-on edge is claimed. The dependencies that bind are measurement machinery, delivery capacity and a price, and they are recorded as typed requirements below.
• Requires (not on this map) Four constraints, none of them a research result. Delivery capacity is the physical one. Distribution and transmission capital spending rose in real terms across every US region while generation costs fell, which is why real retail electricity prices in 2024 equalled 2019 and sat only about 8% below 2010 through a decade in which solar module costs fell roughly 15% a year and battery costs 12%. Until poles, wires, interconnection and hardening can be built at falling real unit cost, an experience curve in generation cannot reach a consumer. The delivered price is the market one, and it is the test the abundance claim keeps failing: production cost is not delivered cost, and no version of the post-scarcity argument in the literature explains why the non-experience-curve complements do not become the whole cost. The accounting treatment is institutional and it cuts both ways: national accounts value goods at transaction price, so the $8.8 trillion demand-side value of widely-used open source — against a $4.15 billion one-time replication cost — appears in no country's output statistics, and the best-measured abundance on record is invisible to the instrument used to argue about whether abundance is occurring. And the funding channel is the market constraint under the best case: 96% of that demand-side value is produced by 5% of open source developers, concentrated because they participate across many projects, with no compensation channel proportional to the value they carry. A non-rival good resting on a scarce, unpriced and concentrated input is not an abundance that has been made durable; it is one that has not yet been tested by the loss of its input.
• Enables If the strong version held, this brief would make Universal Basic Abundance and Wealth Distribution Systems unnecessary rather than enabling them — a rationing instrument and a distribution mechanism both presuppose that scarcity binds. No typed enabling edge is claimed, and that is the honest shape of the relationship: the three slots are alternatives at the level of premise, not stages in a sequence.
• Adjacent Growth and productivity economics, which supplies the experience curves; public finance, which supplies the cost-disease literature; the economics of information goods, which supplies the open source valuation; energy systems analysis, which supplies the delivered-price record; and within this map Energy Storage Revolutions, Commercial Fusion, Autonomous Supply Chains and Artificial General Intelligence.

Established “Zero marginal cost means zero cost.” The best case for the abundance argument shows the opposite. Open source has a reproduction cost indistinguishable from zero and a demand-side value of $8.8 trillion resting on the labour of a few thousand people, 5% of whom produce 96% of it. Around the free copy sit moderation — over 100,000 professionals globally, 40,000 at one firm — and attention, which is fixed at roughly sixteen waking hours a day and does not respond to technology. Frontier The marginal cost of the copy is zero; the marginal cost of the system that makes the copy usable is not, and it rises with volume.

Established “Solar got cheap, so electricity is getting cheap.” Over the decade in which solar module costs fell about 15% a year and lithium-ion battery costs 12% a year, US real retail electricity prices in 2024 came in equal to 2019 and about 8% below 2010, with nominal prices up 23% over 2019–2024 — tracking inflation. Generation cost fell in real terms; distribution and transmission capital spending rose in real terms in every region and absorbed the difference. Frontier This is not a regulatory scandal or a conspiracy. It is what happens when the components on the experience curve become a small share of the delivered cost.

Established “Forty-seven per cent of jobs are about to be automated.” That 2013 figure was an occupation-level expert assessment. The task-level re-estimate across 21 OECD countries put it at 9%, ranging 6% to 12%; the illustrative gap is retail salespeople, scored 92% automatable occupation-wide while only 4% of actual retail workers do jobs with neither group work nor customer interaction. Established And the outturn is now measurable: employment in high-risk jobs grew 6% over 2012–2019 against 18% in low-risk jobs. Frontier The direction survived and the magnitude did not, and the current AI-displacement forecasts share the earlier method's evidentiary structure closely enough that the same discount applies.

Frontier “Keynes was wrong about the fifteen-hour week.” On weekly hours, yes, and badly — about 1,840 annual hours in 2011 against an implied 780. But on output he was right, at 5.15 times the 1931 level inside his 4–8 band, and on a lifetime-equivalent basis holding 1930s demography fixed his prediction implied a 53% increase in lifetime leisure against the 58.4% that actually accrued. Established The productivity gains were taken as leisure; they were taken at the end of life rather than the end of the week, with expected retirement rising from 4.66 to 16.37 years. A verdict of “wrong” on this prediction is as incomplete as a verdict of “right”.

Speculative “We have already achieved post-scarcity in food and information.” Both goods became dramatically cheaper and neither became free, and both remain rationed by price, by delivery and by attention. The claim usually rests on a rich-country vantage point and on aggregate production statistics rather than on delivered consumption. Handwave It also inverts the interesting question: if information abundance is achieved, the constraint has moved to what to read and who decides what may be published, which is where the moderation workforce came from.

Speculative “Cheap enough energy dissolves every other scarcity.” The energy master-resource claim is the most durable technical version of post-scarcity and its first half is well evidenced: the cost curves are real and steep. The propagation step — from cheap generation to cheap water, food, materials and housing — has never been demonstrated, and the retail-price record breaks the first link in the chain. Handwave Where a claim of this shape carries a date, the date is doing the work. Dated abundance forecasts have arrived at roughly generational intervals since 1930 and none has landed on schedule; several — the mid-century leisure society, the 1990s frictionless economy, the 2010s zero-marginal-cost transition — have passed their stated horizons with the world visibly still allocating.

Handwave “Abundance ends distributional politics.” This is the step at which the argument does its work by assertion, and it is stated here in full rather than glossed because it is load-bearing for the entire literature. There is no study anywhere testing whether an abundance of any good has ever reduced conflict over its distribution — not a weak study, none. Frontier The available indirect evidence points the other way: distributional conflict in rich countries has not obviously fallen as goods became cheaper, and the abundance cases in this brief each generated new distributional questions about their own scarce inputs.

Frontier And the reflexive dismissal is a misconception too, which is why this brief declares a tie rather than a verdict. Lighting fell 16,000-fold. Software of $8.8tn demand-side value is produced and given away. Arts employment rose four-to-seventeen-fold while its relative price rose, which is abundance arriving as quantity rather than as price and is the strongest form of the claim available. Established Both hypotheses — that some scarcity genuinely dissolves, and that the aggregate economic problem does not — are supported by the record, and a page that picked one would be misreporting it.