The Floor — Housing, Food, Energy, Healthcare as a Guaranteed Substrate

When we say “survival becomes guaranteed,” we are making a claim that has to cash out as something operational. It is not a slogan, and it is not the vague gesture toward “basic needs” that political rhetoric has been making for half a century. It is a concrete proposition about six categories of provision — housing, food, energy, water, healthcare, and basic connectivity — each of which has its own supply chain, its own current failure rate, and its own feasibility frontier. This essay tries to take that proposition apart, category by category, and ask what would actually be required to put a floor under the human condition.

What “the floor” actually contains

The phrase “basic needs” has done a lot of work in twentieth-century welfare economics, development theory, and political philosophy, and most of that work has been imprecise. The popular reference is Maslow’s hierarchy of needs, in which physiological survival sits at the base of a pyramid that ascends through safety, belonging, esteem, and self-actualization. It is a useful diagram for an undergraduate seminar and a poor empirical claim. The literature that has tried to test the hierarchy directly — most carefully Tay and Diener’s 2011 cross-national study in the Journal of Personality and Social Psychology — finds that while the underlying needs are recognizable across societies, they are not strictly hierarchical: people pursue belonging, esteem, and meaning in conditions of acute material deprivation, and the supposed prerequisite of physiological satisfaction is not, empirically, a prerequisite at all.¹ We mention Maslow only because we have to, and we set him aside.

A more defensible framework is the capabilities approach developed by Amartya Sen and elaborated by Martha Nussbaum. In Development as Freedom, Sen argued that the right unit of evaluation for human well-being is not income, not utility, and not a fixed list of commodities, but the set of substantive freedoms — the capabilities — that a person has reason to value.² Nussbaum’s Creating Capabilities extends this into a list of central human capabilities, each of which a just society has reason to secure for every person.³ Both authors are clear that capabilities are plural: “well-being” does not reduce to a single scalar, and a society that delivers high average income while leaving large numbers of people without adequate shelter, food, or health is failing in a way that aggregate statistics will not capture.

The capabilities approach is the framework we adopt here, with a deliberate narrowing. The full Nussbaum list runs to ten items, including some — affiliation, play, control over one’s environment — that are downstream of, or independent of, the material substrate we are discussing in this essay. What we are concerned with in this section of the abundance research is the material floor: the categories of provision that, if absent, foreclose the exercise of nearly every other capability. There are, as best we can tell, six of them.

The first is housing — durable, climate-appropriate shelter with security of tenure. The second is food — adequate caloric intake and the macro- and micronutrient diversity required for normal human functioning, available reliably and without indignity. The third is energy — electricity for lighting, refrigeration, communication, and increasingly for transport, plus clean cooking fuel. The fourth is water and sanitation — safely managed drinking water and the disposal of human waste in conditions that do not propagate disease. The fifth is healthcare — access to essential health services, including primary care, emergency care, and the prevention and treatment of major communicable and non-communicable diseases. The sixth, which is a relatively recent addition to the list and which would not have been recognized as foundational thirty years ago, is connectivity — meaningful internet access, because participation in nearly every other category now flows through the network.

These six are load-bearing in the technical sense. Their absence forecloses participation. A person without secure shelter cannot reliably hold a job; a person without adequate nutrition cannot reliably learn; a person without electricity cannot refrigerate insulin or charge a phone; a person without sanitation lives in a cloud of preventable disease; a person without healthcare faces routine medical events as catastrophes; a person without connectivity cannot file taxes, schedule appointments, or access most modern public services. Each of the six is necessary for the meaningful exercise of nearly every capability further up Nussbaum’s list, which is why we call this assemblage the floor.

We are not arguing that this list is the complete account of what a flourishing life requires. We are arguing that it is the part of the account that is amenable to coordination — that can be operationalized, costed, monitored, and delivered as a substrate, in the way that the Apik manifesto argues coordination is now technically tractable. The interaction with the cash-versus-services question, the political-economy question of how a floor of this kind is paid for, and the institutional question of who runs it are taken up below and in the companion essays at /abundance/economic-mechanism and /abundance/post-labor. What we do in the rest of this piece is examine each of the six categories on its own terms — current shortfalls, the structure of the supply chain, and the addressable feasibility frontier.

Housing

There are roughly 150 million people in the world today with no housing at all, by the most recent UN-Habitat figures, and on the order of 1.6 billion people whose housing is inadequate — lacking durable construction, sufficient living space, secure tenure, or access to basic services.⁴ These numbers are not the result of a global housing-stock shortage in the literal sense. The shortage is real but local. The deeper pathology is a set of compounding coordination failures: stocks where they are not needed, gaps where they are, and a regulatory and financial layer that prevents the two from converging.

The matching problem

It is worth dwelling on the matching problem before getting to the supply problem, because the matching problem is the one that is most often misdiagnosed. New York City’s vacancy rate, on the standard HUD measure, is around ten percent across the housing stock as a whole — meaning roughly one in ten units, by some accounting, is at any moment unoccupied.⁵ London’s residential vacancy rate is in the range of six percent.⁶ Tokyo, by official Japanese statistics, sits at around twelve percent — a figure that includes the country’s well-documented stock of akiya, abandoned homes in declining rural and peri-urban areas, but also includes substantial central-city vacancy.⁷ These cities also, simultaneously, have homelessness at scales that fluctuate between the merely shameful and the politically destabilizing. The simultaneous condition is not a contradiction; it is the diagnostic.

A city can have ten percent vacancy and ten thousand homeless residents because the housing market is a profoundly thin and friction-laden matching problem. Vacant units are vacant for reasons — they are mid-renovation, held empty by foreign investors, rented at prices the local labor market cannot bear, encumbered by inheritance disputes, kept off the market under rent-control regimes whose incentives interact poorly with each other — and the people who lack housing lack it for reasons that often have nothing to do with the existence of physical units somewhere in the same city. The market does not clear because the market is not, in any deep sense, a single market. It is a stratified, segmented, regulation-encrusted set of overlapping local sub-markets, each with its own clearing properties and its own coordination failures.

This is part of why we treat housing in the Apik stack as a coordination problem first and a construction problem second. It is not that more units would not help. It is that the relationship between marginal new construction and marginal homelessness reduction is mediated by so many layers of matching, finance, and regulatory friction that the unit-supply curve, even when it shifts outward, does not necessarily reach the people who need it.

The supply problem

That said, there is also a supply problem, and the empirical literature on it is unusually clear by the standards of contemporary economics. Glaeser and Gyourko’s 2018 article “The Economic Implications of Housing Supply” in the Journal of Economic Perspectives documented in painstaking detail how zoning, land-use regulation, and discretionary review processes in the most productive American cities have decoupled housing prices from construction costs.⁸ In San Francisco, Los Angeles, New York, and Boston, the price of a typical housing unit substantially exceeds what it would cost to build that unit on unencumbered land — a wedge that Glaeser and Gyourko interpret as a “regulatory tax” imposed on housing in supply-constrained metros.

Hsieh and Moretti’s 2019 paper in the American Economic Journal: Macroeconomics, “Housing Constraints and Spatial Misallocation,” made the macroeconomic stakes of this concrete.⁹ By estimating a spatial general-equilibrium model calibrated to U.S. metro-level data from 1964 to 2009, they conclude that lowering housing-supply constraints in the highest-productivity metros to the level of the median U.S. city would have raised aggregate U.S. GDP by roughly 36 percent over the period. The mechanism is that workers cannot move to where their marginal product is highest because they cannot afford to live there. The spatial misallocation of labor is a coordination failure with national-scale consequences.

Saiz’s 2010 paper in the Quarterly Journal of Economics on the geographic determinants of housing-supply elasticity is the natural complement.¹⁰ Using satellite-derived measures of developable land — accounting for slopes, water bodies, and topographic constraints — Saiz showed that a meaningful fraction of cross-metro variation in housing-price dynamics is explained by physical geography, with the rest substantially explained by regulation. Some cities cannot easily expand outward because they are hemmed in by ocean and mountain. Others have ample physical room and choose, through their land-use regimes, not to use it. The combination of inelastic supply and rising demand produces the price dynamics we observe.

What works

Two contemporary cases show that the housing problem is operationally tractable when the political economy permits a different design. Vienna’s social housing system, built and progressively expanded since the Rote Wien period of the 1920s, today houses roughly 62 percent of Vienna’s population in either municipally owned or limited-profit subsidized units.¹¹ Housing-cost burden across the income distribution in Vienna is among the lowest of any major European city. The system is not without its critics, and there are real questions about its fiscal sustainability under demographic pressure, but as a working demonstration of what a city looks like when the housing market is not the primary allocator of housing, Vienna is the data point.

Singapore’s Housing and Development Board operates at a different scale. Roughly 80 percent of Singaporeans live in HDB-built public housing, the majority of them as owners under a long-leasehold structure that combines public construction with private ownership.¹² HDB allocates housing through a deliberate matching process — racial-integration quotas, family-formation incentives, an explicit interaction with the country’s mandatory savings scheme — that would be politically infeasible in most other contexts but that has produced one of the most universally housed populations in the world.

Neither Vienna nor Singapore is a template that can be transplanted unchanged. Both rest on political and institutional foundations that took decades to build. What they show is that the housing component of the floor is not technically out of reach. The architecture that closes the gap is some combination of: a coordination layer applied to construction logistics and matching (the kind of thing a competent economic-orchestration substrate can directly address); a credible income floor that lets people transact in whatever residual market remains (see /abundance/economic-mechanism); and a deliberate Universal Basic Services-style housing component for the part of the population that the market cannot serve at the quality required. The exact weighting of these three is a political choice. The technical feasibility is not in question.

Food

The manifesto cited two facts about food that bear repeating in this context. First, the world produces roughly enough food, in aggregate, to feed everyone — current FAO production figures sit comfortably above the caloric and protein requirements of the global population. Second, of the food that is produced, on the order of one-third is lost or wasted between farm and mouth, with UNEP’s 2024 Food Waste Index Report placing the consumer- and retail-stage figure at approximately 1.05 billion tonnes per year and the FAO’s earlier work pegging the total figure, including production-side losses, closer to 1.3 billion tonnes.¹³ Food loss and waste is itself responsible for somewhere between 8 and 10 percent of global anthropogenic greenhouse-gas emissions — comparable to the emissions of all road transport.

Against that backdrop of physical sufficiency and routing failure, the FAO’s State of Food Security and Nutrition in the World 2023 estimated that approximately 735 million people were chronically undernourished in 2022, around 9 percent of humanity.¹⁴ The global trend is a familiar one: gradual decline through the 2000s and 2010s, reversal beginning in 2020 under the combined pressure of pandemic disruption, the war in Ukraine, and the climate-driven decline in agricultural productivity in already-fragile regions. The food problem, in the aggregate, is not a Malthusian one. It is a coordination one.

The cost of closing the gap

The relevant cost figures are smaller than the rhetoric of the problem suggests. The World Food Programme’s recurring estimate of the additional spending required to end acute hunger and put the world on track to eliminate chronic undernourishment sits in the neighborhood of $40 billion per year.¹⁵ For comparison, the IMF’s 2023 working paper on fossil-fuel subsidies estimated that explicit and implicit fossil-fuel subsidies — the latter calculated to include unpriced environmental and health externalities — totaled approximately $7 trillion in 2022, or about 7 percent of global GDP.¹⁶ Whatever one thinks of the specific accounting in either figure, the order of magnitude is the order of magnitude. The world spends, every year, roughly two orders of magnitude more on subsidizing the combustion of fossil fuels than would be required to end the most severe forms of hunger. This is a political-economy fact, not an economic one.

The question is then mechanical. What does it cost to deliver food where it is needed, given that the production exists somewhere? This is the supply-chain question that ARCANE addresses on the production side and that Economic Orchestration addresses on the distribution side. ARCANE’s brief, in our internal nomenclature, is the long-horizon agricultural and resource-flow planning layer — the system that knows, at planetary granularity and seasonal frequency, what should be planted where, harvested when, and routed under what contingencies. Economic Orchestration’s brief is the protocol layer that turns those plans into commitments — contracts with farmers, hedges against weather, allocations to logistics providers, settlements with buyers — without the friction that today causes a third of the harvest to never reach a mouth.

Cash transfers and the food problem

A particular subset of the empirical literature on poverty alleviation deserves attention here, because it has reshaped what economists believe is true about the relationship between cash and food security. Bastagli and colleagues’ 2016 review for the Overseas Development Institute synthesized the evidence from 165 cash-transfer studies across low- and middle-income countries.¹⁷ The findings on food are unusually consistent: across virtually all settings, cash transfers reliably increase caloric intake, increase dietary diversity, and reduce the share of household members reporting recent hunger. The effect sizes are not uniform — they depend on transfer size, duration, and the local food market — but the direction is.

The implication is that, in the food category specifically, the long-running debate between cash and in-kind benefits resolves in favor of cash for most contexts. Food markets, in most of the world, are thick: there are many sellers, prices respond to demand, and the marginal recipient of a cash transfer can convert that transfer into nutrition with relatively low friction. This is not true in all contexts — refugee camps, isolated rural communities with thin or monopolized local markets, and acute crisis settings can require direct food aid — but it is true in most contexts most of the time. We return to this argument in the section on cash-versus-services below.

The food floor, in summary, is the most operationally tractable of the six. The production exists. The distribution friction is large but not technically irreducible. The marginal cost of closing the chronic-hunger gap is small relative to the size of any major budget category in the world economy. What is missing is the coordination layer and the political will to deploy it.

Energy

Energy is the second tractable category, and it is becoming more tractable each year. The IEA’s World Energy Outlook 2023 estimated that approximately 675 million people remained without electricity access at the end of 2021, with the number declining slowly but having reversed direction temporarily during the pandemic.¹⁸ A larger figure — approximately 2.3 billion people — lacked access to clean cooking fuel, relying on biomass, coal, or kerosene for daily cooking with consequences that the WHO estimates contribute to roughly 3.2 million premature deaths annually from household air pollution.¹⁹ Energy poverty is not just an inconvenience. It is a leading cause of preventable death.

The empirical case

Lee, Miguel, and Wolfram’s 2020 paper in the American Economic Review, “Experimental Evidence on the Economics of Rural Electrification,” is the cleanest empirical study we have on what happens when an energy-poor population is connected to the grid.²⁰ Working in rural western Kenya, they used a randomized subsidy design to study the effects of household electrification on income, consumption, education, and quality of life. Their findings are honest and not entirely consonant with the most optimistic versions of the electrification narrative. Electrification raised consumption modestly, had limited effects on labor-market outcomes in the medium run, and the willingness-to-pay for connections was lower than the cost of provision under most realistic infrastructure assumptions. The lesson is not that electrification is unimportant — it is foundational — but that the benefits of electrification compound slowly and depend on complementary investments in productive uses, appliances, and downstream economic activity.

This is a useful corrective to a certain kind of techno-optimism that imagines electrification as a single switch. It is not. It is a long process of co-evolution between infrastructure and the economy that uses it. A reasonable energy floor needs to think not just about kilowatt-hours delivered but about what those kilowatt-hours are for and whether the rest of the household economy is positioned to use them productively.

The cost curves

The technological story is, however, the most encouraging in any of the six categories. IRENA’s 2023 Renewable Power Generation Costs report documented, in its most-cited figure, a roughly 89 percent decline in the global weighted-average levelized cost of electricity from utility-scale solar photovoltaic between 2010 and 2022.²¹ The corresponding decline for onshore wind was approximately 69 percent, and for battery storage at the pack level, the decline over a comparable period was roughly 85 percent, with the underlying lithium-ion cell cost decline tracking what observers have come to call Wright’s Law trajectories.²² In a substantial and growing share of the world’s grids, new utility-scale solar plus storage now beats new fossil-fuel generation on a pure marginal-cost basis, before any subsidy or carbon price.

What this means operationally is that the energy floor is no longer primarily a cost problem. It is a finance and coordination problem. The capital stack required to deploy renewable generation at the pace and geographic distribution required is large, but it is not unprecedented; it is comparable in scale to the capital deployed by the global oil and gas industry over the past several decades. The coordination problem — getting the right generation, transmission, distribution, and storage assets sited and interconnected in the right sequence, at the right places, in interaction with weather distributions and load patterns — is exactly the kind of high-dimensional, partial-observability problem that the Apik manifesto identifies as the place where computational coordination has the most to offer.

The longer horizon: ENERA

We treat the longer-horizon energy story under the ENERA program. ENERA’s working hypothesis is that within a multi-decade horizon, orbital and lunar energy infrastructure — space-based solar power, off-world mining of materials needed for the terrestrial energy transition, and ultimately closed-loop energy systems for off-world settlement — becomes both technically and economically tractable. We are deliberate about not promising any of this in the near term. The orbital story is a thirty- to fifty-year framing, not a ten-year one. But the energy floor is a moving target precisely because the technology underneath it keeps improving, and a serious account of the floor has to take that into consideration.

The argument, in summary, is that the energy floor is technologically the most tractable of the six. The unit cost of clean energy has been falling for two decades and shows no sign of stopping. The unit cost of storage is following. What remains is finance and grid coordination, both of which are addressable through a combination of patient capital, mechanism design, and the kind of computational planning layer that the research agenda on economic orchestration is working on.

Water and sanitation

Water and sanitation are the most chronically underdiscussed component of the floor and, in some ways, the most embarrassing. The technology to deliver safely managed drinking water to every person on the planet has existed, at industrial scale, for more than a century. The cost is well-understood. The public-health benefits are unambiguous and have been since John Snow’s 1854 cholera-map work in London. And yet.

The WHO/UNICEF Joint Monitoring Programme’s 2023 update estimated that approximately 2.2 billion people lacked access to safely managed drinking water services and that approximately 3.5 billion people lacked access to safely managed sanitation services.²³ “Safely managed” is the technical term for water that is on the premises, available when needed, and free from contamination, and for sanitation that includes a private toilet with safe disposal of excreta. The figures for “basic” water and sanitation — less stringent definitions — are smaller but still measured in billions. Roughly half of humanity lives without sanitation that meets the safely-managed standard.

The Lancet Commission on Pollution and Health, originally published in 2017 and updated by Fuller and colleagues in 2022, documented the disease burden directly attributable to water-, sanitation-, and hygiene-related causes.²⁴ The 2017 report estimated approximately 1.8 million deaths per year from unsafe water sources, with the broader category of water-, sanitation-, and hygiene-related disease contributing several million additional deaths through diarrheal disease and downstream conditions, particularly in children under five. The 2022 update revised some figures and noted that pollution-related mortality has continued to track demographic and economic patterns, with low- and middle-income countries bearing a disproportionate share of the burden.

We treat water and sanitation briefly here because the technical pathway is well-understood and the principal obstacles are coordination and finance, not invention. The capital cost of universal safely-managed water and sanitation has been estimated by the World Bank and others in the range of $114 billion per year through 2030 — a sum that is large in absolute terms but small relative to global infrastructure spending and trivial relative to the disease burden it would offset.²⁵ The infrastructure required is tractable. The implementation problem is one of governance, finance, and the matching of capital to capacity at the scale of cities and rural districts — exactly the kind of coordination problem the broader Apik thesis is concerned with.

There is one substantive technical caveat. Climate change is making water harder in ways that the static infrastructure models of the twentieth century did not anticipate. Drought-driven aquifer depletion, glacial retreat affecting downstream water supply for hundreds of millions of people in South and Central Asia, and the increased variability of precipitation patterns mean that “deliver water” is no longer a problem with a fixed solution. A robust water floor has to be dynamic, with demand-side management, storage, and inter-basin transfer all operating at a finer temporal resolution than the historical infrastructure model permits. This is itself a coordination problem.

Healthcare

Healthcare is the hardest category. Not because the empirical literature is thin — it is not — and not because the political economy is unique — it is, but most of the categories on this list have unique political economies. Healthcare is the hardest because of a structural feature that distinguishes it from every other category: it scales with technology. Every new treatment that medicine develops widens the floor that “guaranteed healthcare” has to underwrite. There is no analogous dynamic in food, water, or energy. The result is that the healthcare floor is a moving target in a way that the other floors are not, and any system that promises to deliver it is implicitly making a commitment about what it will and will not cover that the underlying technology will keep contesting.

The current shortfall

The WHO’s World Health Statistics 2023 and the WHO/World Bank universal health coverage monitoring report from the same year together provide the headline figures.²⁶ Approximately half of the global population lacks full coverage of essential health services as defined by WHO’s UHC service-coverage index. Approximately two billion people face catastrophic or impoverishing health expenditures — defined as out-of-pocket spending that exceeds either ten or twenty-five percent of household consumption, depending on the threshold used — in any given year. The figure represents an increase from previous reports, driven by the combination of population aging, rising treatment costs in middle-income countries, and the residual disruption from the COVID-19 pandemic.

A health system that protects against catastrophic financial risk is doing the basic job of a health system. A health system that does not is, in the technical sense, not functioning. By that standard, large portions of the global health architecture are not functioning, and the failures cluster in predictable patterns: lower-middle-income countries with growing private health sectors, post-Soviet transition economies with hollowed-out public systems, and the United States with its unique combination of high spending, high coverage gaps, and high catastrophic-expenditure rates among the uninsured and underinsured.

What we know about insurance

Two studies in particular have shaped what economists believe about the marginal value of health insurance, and they bear on the question of what a healthcare floor has to deliver. The RAND Health Insurance Experiment, conducted in the United States in the 1970s and analyzed by Manning and colleagues in their 1987 paper in the American Economic Review, was the first large-scale randomized study of health insurance generosity.²⁷ The headline finding is that more generous insurance led to more use of medical care but had relatively limited effects on most measures of health, with concentrated benefits for low-income participants and for specific high-value services. The RAND findings have been cited in support of cost-sharing as a way to control utilization without large health consequences. They have also, fairly, been criticized for the limitations of their measurement and the era they reflect.

The Oregon Health Insurance Experiment provides the more recent and arguably more relevant evidence. Beginning in 2008, Oregon used a lottery to allocate Medicaid coverage among a population of low-income adults whose demand exceeded available slots. Finkelstein and colleagues’ 2012 paper in the Quarterly Journal of Economics reported the year-one results, finding that Medicaid coverage substantially increased health-care utilization, reduced financial strain and depression, and improved self-reported health, with measurable effects across most categories.²⁸ Baicker and colleagues’ 2013 paper in the New England Journal of Medicine extended the analysis to year two and added measured biomarker data.²⁹ They found that Medicaid coverage produced large reductions in observed depression and large reductions in catastrophic financial outcomes, but produced no measurable improvements in three biomarkers — blood pressure, cholesterol, and glycated hemoglobin — over the two-year window. The biomarker null was widely discussed.

The two studies together produce a nuanced picture. Health insurance reliably reduces financial risk and improves access to care. Its effects on measured population health are smaller and more variable than its effects on financial security and mental health, and they materialize over longer time horizons than two-year experiments can capture. A healthcare floor cannot be evaluated on whether it improves a population’s average HbA1c. It has to be evaluated on whether it reliably prevents catastrophic financial outcomes, reliably treats acute conditions, and provides the range of preventive and primary services that compound over a lifetime. The Oregon results are, on those terms, strongly supportive of the value of coverage.

The aging-and-AI inflection

Two trends that interact awkwardly are reshaping the healthcare floor on a generational horizon. The first is global population aging. Most major economies, and an increasing share of middle-income economies, will be substantially older in 2050 than they are today, with the corresponding rise in the burden of chronic non-communicable disease — cardiovascular disease, cancer, dementia, musculoskeletal disorders — that requires expensive long-term management. McKinsey’s 2023 healthcare-economics work, IHME’s Global Burden of Disease projections, and the OECD’s health-spending projections all converge on the qualitative finding that healthcare spending as a share of GDP is going to keep rising in most advanced economies absent a major productivity intervention.³⁰

The second trend is the intersection of artificial intelligence with medical practice. Models capable of strong performance on diagnostic radiology, pathology, dermatology, and increasingly on aspects of clinical reasoning are now deployed or in late-stage trials at major health systems. Whether these systems substantially reduce the cost of care, improve quality, or simply add another layer of expense is genuinely an open question — the early productivity literature on health-care AI is mixed, and the implementation barriers are significant. But to the extent that AI medicine succeeds in raising the productivity of clinical labor, it eases the aging-cost crunch from one direction. To the extent that it expands what is treatable, it widens the floor that has to be underwritten from the other. The net effect on the size of the healthcare floor is genuinely uncertain.

Implicit rationing

One observation we treat as foundational. Every healthcare system rations. The question is how. The British NHS, through the National Institute for Health and Care Excellence, rations explicitly: NICE evaluates the cost-effectiveness of new treatments and produces recommendations that the NHS broadly follows, with an explicit incremental cost-effectiveness threshold around £20,000 to £30,000 per quality-adjusted life year.³¹ The American system rations implicitly, through price, insurance design, and the resulting cost-driven decisions of patients and providers. Whether one prefers explicit or implicit rationing is a value judgment with significant practical consequences. What is not a value judgment is the observation that no possible healthcare floor is unlimited — the technology will always exceed the budget — and the only question is whether the rationing is transparent and procedurally fair or hidden and procedurally arbitrary.

A guaranteed healthcare floor, in our view, has to be explicit about what it covers, how it decides what to add, and what the procedural mechanisms for change are. The Apik account of the floor does not promise unlimited care. It promises a credible floor that is large enough that no person in the system faces a routine medical event as a financial catastrophe. The exact contents of that floor are a political question. The mechanism by which the floor is sustained — costed, delivered, monitored, updated — is partly a coordination question, and is the part our research agenda is actually positioned to address. We say more about the interaction of healthcare and the broader economic mechanism in /abundance/economic-mechanism.

Connectivity

Connectivity is the newest item on the list and the one whose status as a floor primitive would have been disputed thirty years ago. It is no longer in dispute. The ITU’s Facts and Figures 2023 estimated that approximately 2.6 billion people remained offline as of late 2023.³² The figure is declining but not as fast as one might have expected given the pace of mobile-network expansion. The remaining offline population is concentrated in low-income countries, in rural areas of middle-income countries, and among older adults in high-income countries.

The argument for connectivity as a floor primitive is that participation in nearly every other category now flows through the network. Cash transfers in developing countries are increasingly delivered via mobile money. Healthcare appointments are increasingly scheduled and conducted via networked platforms. Tax filing, social-security enrollment, and voter registration in high-income countries now assume connectivity. Agricultural prices, in much of the world, propagate through SMS and mobile data. A person without connectivity is, increasingly, a person without access to the institutions that operate the rest of the floor.

Hjort and Poulsen’s 2019 paper in the American Economic Review, “The Arrival of Fast Internet and Employment in Africa,” is the cleanest causal study on the effects of connectivity expansion.³³ By exploiting the staggered rollout of submarine internet cables along the African coast and the resulting changes in inland connectivity, they identify the effects of fast-internet arrival on labor-market outcomes. The findings are large: significant increases in employment, particularly in skilled occupations, with the gains concentrated in cities that received connectivity earlier. The mechanism is the standard one — connectivity reduces information friction in labor markets, reduces the cost of running businesses, and connects local producers to broader markets — but the magnitude of the effect, in this carefully identified design, is substantial.

The connectivity floor is technically tractable. The combination of expanding low-Earth-orbit satellite networks, declining cost of mobile infrastructure, and the increasing fraction of useful services that work over modest bandwidth means that universal meaningful connectivity is, on a ten- to fifteen-year horizon, a feasible engineering target. The remaining barriers are cost, regulatory, and last-mile — particularly for the rural and low-income populations that are the residual offline population — and they are addressable through the same coordination-and-finance pattern as the other floors. We treat universal meaningful connectivity as a baseline assumption for what the broader post-labor economics framework presupposes.

What “guaranteed” has to mean operationally

If the floor is a list of six categories, “guaranteed” is a list of three properties. To say the floor is guaranteed is to say that it has each.

The first property is unconditional access. The floor is not contingent on employment, on means-tested verification, on documentary status, on having recently filed the right form, or on any of the other conditioning structures that contemporary welfare states use to limit the cost of provision. The conditioning structures are not arbitrary — they exist for reasons of cost, of fraud control, of political legitimacy — but they impose substantial friction on the population they nominally serve. The administrative burden of accessing means-tested benefits, well-documented in the work of Herd and Moynihan and others, is large enough that meaningful fractions of eligible populations do not access benefits they are entitled to.³⁴ An unconditional floor eliminates the friction at the cost of universal coverage. The mechanism for managing the cost trade-off is the subject of /abundance/economic-mechanism; the relevant point here is that conditional access is not, on the empirical evidence, a credible floor.

The second property is continuity. The floor is robust to local political failure and to supply-chain shocks. It does not collapse when a government is captured, when a natural disaster strikes, when a pandemic arrives, or when a regional war disrupts a key supply line. This is the property that points most directly at the coordination thesis. A floor that depends entirely on the competence of every local government in every locality, on every day, is not a floor; it is a collection of locally negotiated arrangements that fail at the worst possible times. A floor that is guaranteed in the operational sense has to have the property that the layer above the local government — the coordination substrate — is capable of stepping in when local provision fails. This is exactly the kind of resilience-through-redundancy that the broader Apik account of computational coordination is designed to deliver.

The third property is quality floors. The floor is not the lowest deliverable provision; it is a level that the median citizen of the political community would accept for their own family. A homeless shelter that is technically housing but that no member of the political community would actually sleep in if they had any choice is not housing in the operational sense. A clinic that exists on paper but is uncrewed, unstocked, or unsafe is not healthcare. A social tariff for electricity that flickers is not energy. The quality floor is what distinguishes a credible floor from a Potemkin one, and it is the property that political cycles tend to erode in the absence of a stabilizing layer below.

These three properties — unconditional access, continuity, quality floors — are the operational content of the word “guaranteed.” Together they imply that the floor cannot be delivered solely through the mechanisms of redistribution-via-cash that political economy has been working with for the last fifty years. The floor requires a coordination substrate underneath it, capable of responding to failures and stocking out the categories that cash alone does not adequately allocate. The argument that the manifesto makes about coordination as a layer below institutional politics is exactly the argument the floor needs.

The interaction with cash

It would be a mistake to read the foregoing as an argument that the floor must be delivered entirely in kind. It is not, and the empirical literature on cash transfers is too good to let that misreading stand.

The classic theoretical reference is Currie and Gahvari’s 2008 Journal of Economic Literature survey, “Why In-Kind Benefits?”³⁵ Their question was the obvious one: under what conditions does in-kind provision dominate cash, given that cash is preference-respecting and has lower administrative cost? Their answer, after surveying the theoretical and empirical literature, was that in-kind benefits dominate cash when one or more of the following conditions holds: when markets for the good in question are thin or imperfect; when there is significant information asymmetry between the recipient and the seller; when there are externalities in consumption that the recipient does not fully internalize; when paternalism toward a specific set of consumption choices is politically required for the program to exist; or when the cost of monitoring cash use is prohibitive in the specific institutional context.

Hanna and Olken’s 2018 Journal of Economic Perspectives paper on targeting in developing countries adds an empirical layer.³⁶ Using a series of careful field experiments, they showed that the choice between cash and in-kind benefits is not binary but is conditioned on local institutional capacity, the structure of the local market for the relevant good, and the distribution of preferences within the recipient population. Their bottom line is that cash dominates in most contexts most of the time, but the exceptions are real and predictable.

The mature design uses both. In domains with thick markets — most of the food category, most of the cash-side of housing in unconstrained markets, much of energy at the consumption side — cash dominates. The recipient is the best judge of how to allocate the marginal dollar across food, fuel, transport, and education for their children, and the administrative friction of in-kind provision is real. In domains with thin or imperfect markets — housing in supply-constrained cities, healthcare with rampant information asymmetry between patient and provider, water infrastructure that has the structure of a natural monopoly — in-kind provision dominates. The argument that cash transfers will solve homelessness in San Francisco fails on the supply elasticity of the local housing market. The argument that cash will produce optimal health outcomes fails on the patient’s inability to evaluate the quality of medical care in advance of receiving it.

A serious account of the floor uses both instruments. A meaningful cash floor — call it an unconditional basic income or, in the framing of /abundance/economic-mechanism, the income side of an economic mechanism — handles the categories where markets work. A deliberate Universal Basic Services-style provision handles the categories where markets do not. The two are complementary, not competitive, and the design problem is identifying which categories belong in which bucket and where the boundary moves under different local institutional conditions. We say more about the mechanism side in /abundance/economic-mechanism and about the political-economy questions in /abundance/transition.

Where to read further

The companion essays in this section take up adjacent threads. /abundance/post-labor addresses the economic structure of a society in which a credible floor is the foundation on which non-employment-based participation rests. /abundance/economic-mechanism addresses the mechanism design of how a floor of this kind is paid for, allocated, and updated. /abundance/precedents walks through the historical cases — Vienna, Singapore, Mongolia’s mineral-fund cash transfers, Alaska’s Permanent Fund — that have informed the architecture. /abundance/transition takes up the political-economy and sequencing questions of how a floor of this kind is built without breaking the systems it replaces. On the technical-substrate side, /civilizational-stack/arcane and /civilizational-stack/enera address the production layers — agricultural and resource flows, and the longer-horizon energy story respectively — that have to function for the floor to be deliverable.