Dead Economy Theory

The dead economy theory is Owen McGrann’s name for a failure mode in which AI-driven production keeps expanding while ordinary people lose the income, ownership, bargaining power, and political leverage needed to participate in the resulting economy. The economy is “dead” not because output stops, but because productive capacity no longer requires or belongs to most citizens.^{source: owen-mcgrann-dead-economy-theory-2026.md}

The three-turn mechanism

1. A firm replaces workers with AI, lowers costs, raises margins, and captures the benefit.

2. Displaced workers lose income and reduce consumption, weakening demand for many firms.

3. Automating firms eventually discover that their customers were workers elsewhere in the same system.

The coordination problem is that each firm receives nearly all of its own labor-cost savings but bears only a fraction of the aggregate demand it destroys. That creates a prisoner’s-dilemma dynamic: automation can be privately rational and collectively excessive.^{source: owen-mcgrann-dead-economy-theory-2026.md}

Why AI may differ from earlier automation

The essay argues that AI is unusually dangerous because it targets cognitive work across industries at once, its marginal cost is low, and the capital already committed to AI creates pressure for rapid adoption. Transition speed matters: even if new work eventually appears, the economic “short run” can last an affected worker’s lifetime. McGrann also highlights “so-so automation”—systems good enough to displace people but not productive enough to create broad abundance—as a plausible worse case.^{source: owen-mcgrann-dead-economy-theory-2026.md}

This belongs inside the wider uncertainty mapped in ai-labor-market. Historical job creation is evidence against simple permanent-unemployment forecasts, but it is not a guarantee that new tasks will appear quickly enough, in the same places, or for the same people.

Ownership and democratic leverage

The theory’s deeper claim concerns ownership. If AI systems and ai-compute-infrastructure are concentrated in a few firms, income moves from labor toward capital while the tax base, collective bargaining, and consumer demand weaken. Citizens lose leverage because elites need less of their labor and can route investment through systems with little democratic accountability.^{source: owen-mcgrann-dead-economy-theory-2026.md}

This connects the labor question to infrastructure governance: the same technical productivity gain has very different social effects depending on whether people own productive capital, can access AI cheaply, or receive only transfers after losing labor income.

Meaning is not only distribution

McGrann rejects the idea that universal basic income alone resolves displacement. Drawing on research about deaths of despair and precarity, he argues that work supplies status, structure, social belonging, and a credible future—not merely cash. This overlaps with ai-job-grief: economic displacement can become an identity and community crisis before or even without absolute poverty.^{source: owen-mcgrann-dead-economy-theory-2026.md}

Proposed interventions

The essay favors mechanisms that spread ownership and preserve public leverage rather than treating citizens only as transfer recipients:

public ownership stakes in AI infrastructure;
broader capital ownership;
taxes on automated labor and high capital income;
aggressive antitrust enforcement;
shorter workweeks and job guarantees;
democratic oversight of deployment incentives.

The AI 2040 Plan A scenario sketches one attempted answer: cap and auction permits for compute and robot production, then distribute part of the proceeds as a Citizen's Dividend. Its own discussion still recognizes that redistribution does not automatically preserve purpose, ownership, or political agency, so it addresses the income channel more clearly than the broader democratic-leverage problem.^{source: ai-2040-plan-a-2026.md}

What remains contested

The theory combines a strong coordination insight with uncertain empirical assumptions. Its outcome depends on the pace and breadth of displacement, whether AI creates complementary tasks and new markets, how prices respond, whether governments redistribute purchasing power, and whether productivity gains are real rather than merely promised. The essay is strongest as a warning about incentives and ownership, not as a settled forecast that aggregate demand must collapse.

Pascual Restrepo offers a direct macroeconomic rebuttal to demand-collapse scenarios. AI agents may not consume household goods, but they consume compute; revenue from that infrastructure accrues to human owners, including households indirectly through pension and mutual funds. If owners save more, he expects interest rates and monetary policy to adjust to sustain aggregate demand. He therefore treats distribution and transition insurance as serious problems but not permanent disappearance of demand.^{source: the-bell-pascual-restrepo-ai-labor-economy-2026.md}

The rebuttal narrows rather than closes the dispute. It assumes ownership income reaches people broadly enough, financial transmission works, and governments do not lock policy into a demand-deficient equilibrium. McGrann's strongest concern is precisely that ownership and political leverage may remain concentrated. The disagreement is therefore less about whether income exists somewhere than about who controls it, how quickly it circulates, and whether institutions redistribute it before social damage becomes entrenched.