WP-2026-06: A New Economic Model in the Age of AI

The argument proceeds in three movements. First, that the discipline of economics rests on a category error, long identified but never resolved, in which historically constructed institutional arrangements are presented with the epistemic authority of natural law. Second, that artificial intelligence alters the practicality, though not the desirability, of alternatives that earlier critics proposed and could not implement. Third, that six propositions, each stated alongside its strongest counter-argument, define the design requirements of a genuinely reconstructed economics. The paper makes no claim to have built that economics. It claims only that the building has become possible, and states what it would require.

The discipline of economics presents a large part of its subject matter as natural. It speaks of laws, of equilibria, of a natural rate of unemployment, and of market outcomes that are described as discoveries about the world rather than as the products of particular institutional arrangements. This paper begins from the position that a significant portion of what economics presents as natural is in fact constructed: assembled by identifiable actors, at identifiable historical moments, in the service of identifiable interests, and capable in principle of being assembled differently.

This position is neither novel nor radical. It was stated, in its essentials, by Karl Polanyi in 1944, and elements of it were conceded by Simon Kuznets, the architect of national income accounting, as early as 1934. What is novel is the claim this paper adds to it: that artificial intelligence changes, for the first time, the practical feasibility of acting on the insight.

The scope of the paper must be stated precisely, because its title invites a larger expectation than it intends to meet. A new economic model, in the full sense, would require a theory of value not reducible to market price, a specification of how productive activity is coordinated without reliance on the price mechanism as the sole instrument, and an account of how investment and innovation are financed in the absence of capital seeking a return. This paper offers none of these in complete form. It offers instead an argument that the conditions for constructing them have changed, and a set of requirements against which any proposed construction should be tested. The distinction between a new political economy, meaning a revised set of rules and measurements operating within the existing market framework, and a new economic model, meaning a departure from that framework's foundational categories, is maintained throughout, and the paper is candid that the longer work it introduces achieves the former rather than the latter.

The claim that economic categories are constructed rather than natural requires careful statement, because in its loose form it is either trivially true or plainly false, and only in its precise form is it useful. It is trivially true that money is a human invention; no serious economist holds otherwise. It is plainly false that economic phenomena are arbitrary or that they can be reorganised at will without consequence; the constraints economies face are real, and the failures of attempts to ignore them are well documented. The useful form of the claim lies between these poles: that a specific and consequential set of arrangements, which determine who receives what and on what terms, are presented within the discipline and its public communication as though they possessed the necessity of natural facts, and that this presentation obscures both their origins and their alterability.

Simon Kuznets, who constructed the national income accounts that became gross domestic product, warned the United States Congress in 1934 that the welfare of a nation could scarcely be inferred from a measure of national income, and cautioned repeatedly across his subsequent career against the conflation of the two. The warning was not heeded, and gross domestic product became, through the second half of the twentieth century, the primary metric against which national economic performance was judged. This is a documented instance of a constructed measure, explicitly designed for a limited purpose by its own author, acquiring an authority its author disclaimed. Diane Coyle's historical account of GDP's design, political adoption, and subsequent life as the primary measure of economic performance provides the fullest documentation of this process, including the specific congressional and inter-agency negotiations through which a measure built for wartime planning was converted into a universal welfare indicator against its designer's stated intentions.

Karl Polanyi's argument in The Great Transformation, published in 1944, is the fuller statement of the constructed character of economic institutions. Polanyi held that land, labour, and money were transformed into commodities, priced and exchanged through market mechanisms, by a specific historical process rather than by the operation of any natural economic law. He described the enclosure of common land in England, extended over several centuries, as a principal instance: a legal and political process that converted a resource previously governed by custom and common right into private property governed by market exchange. Polanyi's claim was not that markets are unnatural in some moral sense, but that the self-regulating market society of the nineteenth century was a deliberate construction requiring continuous state action to establish and maintain, and that its presentation as a natural order was itself part of its ideological architecture.

Subsequent economic history has qualified aspects of Polanyi's account, and it would be inaccurate to present his thesis as uncontested; the timing and coherence of the enclosure process he describes remain matters of scholarly debate. The core observation, that market society was made rather than found, is nonetheless widely accepted across economic history, including by scholars who reject his broader conclusions.

The constructed character of economic arrangements extends beyond price and property to the temporal architecture within which economic activity is organised, and this dimension is less often examined. The standardised working week and the eight-hour day are frequently treated as though they reflected natural limits of human productivity. They do not. The eight-hour day was established, in Britain and the United States, through sustained organised labour campaigns extending across the nineteenth and early twentieth centuries, in opposition to substantially longer factory hours; the figure of eight was advanced through the slogan of eight hours for work, eight for rest, and eight for what we will, a demand for symmetry rather than a finding about optimal output. The historian Moishe Postone argued in Time, Labor, and Social Domination that industrial capitalism did not merely employ a pre-existing natural time but constructed a specific form of abstract, uniform, clock-measured time, and made that form the medium through which economic value was calculated, displacing the task-oriented and seasonally variable time that had preceded it.

This observation has an empirical corollary that constitutes evidence rather than assertion. If the Gregorian calendar and its associated production cycle were a natural necessity of commerce, one would not expect to find a large fraction of global economic activity operating successfully on a different temporal architecture. The Islamic Hijri calendar, being lunar, drifts against the Gregorian solar year by approximately eleven days annually, so that the fixed observances of Ramadan and Eid migrate through all seasons over a cycle of roughly thirty-three years. Commercial activity in economies organised substantially around this calendar cannot lock its principal demand events into fixed seasonal slots in the manner the Gregorian retail cycle permits, and it functions nonetheless. The persistence of two incompatible temporal architectures within the single global economy is evidence that neither is a natural necessity. The Gregorian production cycle achieved its global default status through the reach of European commerce and colonial administration, not through any property of commerce as such.

The longer work this paper introduces examines a range of contemporary economies, including Mauritius, Egypt, the Philippines, Zambia, Ghana, the states of the Sahel, and the United Kingdom, and identifies a structure that recurs across them despite substantial variation in their circumstances. The structure may be stated independently of the cases, and is summarised here because it motivates the propositions that follow.

The structure concerns the ordering of claims on economic output. In each case examined, a particular claim, that of an external creditor, of a currency peg's stability requirements, of a concentrated ownership interest, is satisfied first, as a precondition, and the residual remaining after that claim is satisfied is what becomes available to the population that produced the output. The claim satisfied first is characteristically presented not as a claim but as a technical necessity: a rating agency's methodology, a central bank's mandate, a condition attached to external finance. The presentation of a distributional choice as a technical requirement is the mechanism by which the ordering is maintained without appearing to be a choice at all. This paper terms the pattern the sequencing of the prior claim, and the longer work documents its operation across the cases named.

A second observation concerns the differential description of identical mechanisms. The same monetary operation is described as legitimate monetary policy when conducted by a state issuing a reserve currency, and as monetary indiscipline when conducted, at a fraction of the scale, by a state that does not. The same wage floor is described as a necessary protection of labour in one jurisdiction and as an impediment to competitiveness in another. The identical economic act carries a different name, and attracts a different judgement, according to the position within the currency and institutional hierarchy of the party performing it. This is not a claim of conspiracy; it requires no coordination to produce. It is a claim about how a shared framework of description can encode a hierarchy while presenting itself as neutral.

The critique of constructed economic categories is old, and the proposal of alternatives is nearly as old. E.F. Schumacher argued in Small Is Beautiful, published in 1973, for an economics organised at a human and ecological scale rather than at the scale demanded by the maximisation of output. Herman Daly developed, across the same decade and after, the framework of steady-state economics, holding that an economy operating within a finite biosphere must eventually cease to treat aggregate growth as an unquestioned objective. Kate Raworth proposed in Doughnut Economics, published in 2017, a framework bounded simultaneously by a social foundation below which no person should fall and an ecological ceiling beyond which collective activity should not press. Each of these is a serious contribution, advanced by a serious scholar, and none has achieved adoption as the organising framework of any major economy. The reason is instructive and is not that the arguments were refuted.

The reason, this paper contends, is that each of these proposals required administrative capacities, for verification, for coordination, and for measurement at scale, that the institutions of the existing model possessed and their proposed replacements did not. To measure the true ecological cost of an activity, to verify a supply chain's compliance with a labour standard, to coordinate the allocation of resources without relying wholly on the price signal: each of these was, at the time these proposals were advanced, either impossible or prohibitively costly. The existing model's institutions, however imperfectly, performed these functions, and no proposed alternative could match them. An argument may be correct and still fail to be adopted if it cannot be operationalised, and the alternatives to the growth model have historically failed at the point of operationalisation rather than at the point of argument.

Artificial intelligence bears on this specific failure. The opportunity is that the verification, coordination, and measurement functions that earlier alternatives could not perform may now be performable at a cost and speed that alter their feasibility. This claim is not hypothetical. The Anker Living Wage methodology, adopted by the Global Living Wage Coalition and now applied in more than forty-five countries, computes living wages against local cost-of-living baskets using standardised data collection frameworks. The ILO's Better Work programme deploys digital worker-voice systems that provide real-time monitoring of supply chain labour conditions across garment factories in seven countries, replacing the periodic and expensive physical inspection that made continuous verification economically unviable. These are operational instances, not proposals, of the verification capacity that Schumacher and Daly's alternatives required and could not access.

If the construction of a genuinely new economic model has become feasible, the question of what such a model must contain becomes pressing. This section states four requirements that distinguish a reconstructed model from a reformed version of the existing one. The requirements are stated as necessary conditions, not as a completed design, and the paper claims no more than that any adequate model would have to satisfy them.

The most substantial objection to any proposal that would reduce reliance on the price mechanism concerns coordination. The price system is not merely a method of distribution; it is a mechanism for transmitting information about relative scarcity and demand across an economy too large for any central authority to survey. The classic statement of this objection holds that no alternative can match the price system's capacity to coordinate the dispersed knowledge of millions of participants, and that attempts to replace it must therefore fail. This objection is serious, it is substantially correct as a description of the historical alternatives against which it was originally directed, and no proposal that ignores it deserves consideration.

It is therefore significant that coordination without a central price signal is not merely a theoretical possibility but an empirically observed phenomenon, occurring at scale in both natural and organisational systems. The point must be stated with precision, because it is easily overstated into a naturalistic fallacy. The claim here is narrower and strictly evidential: the proposition that large-scale resource coordination is impossible without a price mechanism is falsified by observed cases, and the falsification of an impossibility claim is logically sufficient to reopen the question, whatever else remains unresolved.

The biological cases complement the institutional evidence. Social insect colonies allocate labour and resources across many thousands of individuals through stigmergy, a mechanism in which individuals respond to signals deposited in a shared environment rather than to central direction; the term was introduced by the biologist Pierre-Paul Grasse in the 1950s and the mechanism has since been formalised in the field of swarm intelligence. In a study published in Science in 2010, researchers demonstrated that the slime mould Physarum polycephalum, an organism without a nervous system, produced a network connecting distributed food sources that closely resembled the topology of the Tokyo rail network, a system designed by human engineers over decades. These cases do not establish that human economies should imitate slime moulds or insect colonies, which would be the fallacious inference. They establish only that decentralised coordination of resource allocation, without central price direction, is a solved problem in the general case, which is sufficient to defeat the impossibility claim. Ostrom's institutional evidence is the stronger proof of the same point.

The more directly relevant evidence from human organisation at industrial scale comes from the Chinese appliance manufacturer Haier, which has since 2005 operated under a management model termed Rendanheyi, decomposing the corporation into several thousand largely self-governing micro-enterprises responding directly to signals from users rather than to central management direction. Two qualifications are essential. First, Rendanheyi decentralises coordination but does not dispense with price or profit; the micro-enterprises operate within a market. Second, the causal attribution of Haier's performance to Rendanheyi specifically, as opposed to its brand strength and domestic market position, has not been established in peer-reviewed work. The case establishes a bounded point: that decentralised coordination through distributed response to signals is viable at the scale of a major industrial enterprise. It does not establish more than this.

The following propositions do not constitute a model. They state the conditions a model would need to satisfy, and each is presented together with the most substantial objection to it, because a proposition insulated from its counter-argument is an assertion rather than a contribution. The objections are not disposed of; they are stated, because an honest account of what a reconstructed economics requires must include an honest account of what stands in its way.

This paper has argued three things and has been careful to claim no fourth. It has argued that a significant portion of what economics presents as natural is constructed, a claim established rather than asserted through the documented cases of Kuznets's disclaimed measure, Polanyi's account of the market's construction, and Coyle's historical documentation of GDP's adoption, and qualified by acknowledgement of where those accounts remain contested. It has argued that artificial intelligence alters the feasibility, though not the desirability, of alternatives that earlier critics proposed and could not implement, and has cited operational instances of the verification and measurement functions those alternatives required, in the Anker Living Wage methodology and the ILO Better Work programme's digital monitoring systems. And it has argued, through the Nobel Prize-level evidence of Ostrom's commons governance research and the biological evidence of stigmergy, that the impossibility claim on which the defence of the price mechanism as sole coordinator rests does not hold, whatever else remains to be established.

The paper has not argued that it has built a new economic model, and it has stated at each stage the requirements such a model would have to meet and has not met. A field that has too often presented the contestable as the settled is poorly served by a critique that repeats the error in the opposite direction, and a proposal that overstated its own completeness would forfeit the standing from which it criticises. The construction of a reconstructed economics, if it is undertaken, will be the work of many hands over a considerable period, and will require empirical development that a paper of this kind can identify but cannot supply. What this paper offers is an argument that the work has become possible, a specification of what it would require, and an honest account of the objections that stand against each requirement.