Applying the System Asset Pricing Model to Nuclear Fission: Welfare Gains, Persistent Waste, and the Limits of Internalization

Nuclear fission operates in a hollow win equilibrium

Utilities, customers, and states transact in a bilateral payoff space that counts electricity revenues and near-term compliance costs but omits six system-welfare channels: accident losses, waste management, security and proliferation, decommissioning, liability subsidy, and intergenerational burden.

• Global annual revenue base: 150 billion (2,667 TWh at 56/MWh)
• Annual system welfare cost: $79.6 billion
• Average system beta β_W = 0.53: each dollar of revenue destroys ~53 cents of system welfare
• System-adjusted payoff: +$70 billion per year — fleet remains welfare-positive on average flow

Sellafield is the canonical hollow win: private surplus, public cost

The UK's Sellafield site, which has not generated electricity in decades, required £2.6 billion in annual public expenditure for cleanup and waste management, with total liabilities exceeding £130 billion. Electricity generation created private payoffs during reactor operating lives, while decommissioning and waste stewardship costs were shifted to future taxpayers.

• Bilateral transaction cleared on commercial terms
• System absorbed costs dwarfing original private surplus
• Illustrates pricing design failure, not isolated mismanagement

The literature is organized by discipline, not by a common welfare unit

Cost studies end at financing and dispatch; waste studies end at repository governance; risk studies end at accident frequency; public-acceptance studies end at legitimacy. No integrated welfare price exists for nuclear fission.

• Levelized-cost studies price capital, fuel, and operations
• Accident studies price catastrophic failures after the event
• Waste-governance studies examine siting, storage, and disposal
• Discounting debates ask how to value far-future harms

Six channels sum to $79.6 billion annual welfare cost

The welfare-cost estimate aggregates accident cost (21.6B), waste management (12.0B), security and proliferation (20.0B), decommissioning (15.0B), liability subsidy (3.0B), and intergenerational burden (8.0B). Each channel is anchored in established literature and converted into an annual civilian-fission welfare flow.

• Accident cost: annualized severe-event loss and legacy cleanup (Fukushima/Chernobyl literature)
• Waste management: interim storage, repository development, disposal obligations
• Security and proliferation: guard-force, safeguards, diversion risk allowance
• Decommissioning: gap between booked funds and revealed cleanup cost (Sellafield evidence)
• Liability subsidy: insurance wedge from statutory caps (Price-Anderson)
• Intergenerational burden: discounted stewardship transfer to future populations

Marginal beta for new build rises to 0.70, worse than fleet average of 0.53

The decision-margin reweighting applies multipliers to each channel: waste management (1.40), decommissioning (1.35), liability subsidy (1.40), intergenerational burden (2.00). The resulting marginal welfare cost is $104.1 billion, yielding β_W^m = 0.70.

• Incremental capacity adds fresh severe-event exposure, not sunk historical risk
• New build adds fresh spent-fuel obligations and repository-capacity demand
• Marginal project creates fresh long-duration stewardship claims with no historical amortization
• Policy implication: expansion is less attractive than legacy-fleet operation

91% of Monte Carlo draws keep average system beta below 1.0

A 100,000-draw simulation samples each channel from triangular distributions (low, central, high). Under baseline specification, 91% of draws yield β_W < 1. A fat-tail stress test on the accident channel (right-skewed draw) does not reverse the modal result.

• Low stack: W = 48B, β_W = 0.32, payoff +102B
• Central stack: W = 79.6B, β_W = 0.53, payoff +70.4B
• High stack: W = 124B, β_W = 0.83, payoff +26B
• Exact beta value is uncertain, but central case remains welfare-positive

High-level waste hazard horizon outlasts every demonstrated institutional duration

Fission creates radionuclides with hazard horizons of 10,000 to 100,000+ years. No human institution has operated on that timescale. Deep geological disposal can contain the burden but cannot erase the temporal mismatch between a commercial asset with a bounded revenue life and a waste stream requiring millennial stewardship.

• Corporate planning cycle: 5-30 years
• Regulatory and democratic cycle: 4-25 years
• Plant operating life and decommissioning tail: 40-100+ years
• Deep geological repository stewardship: 100-10,000 years
• High-level waste hazard horizon: 10,000-100,000+ years

No regulatory refinement can eliminate the persistence floor

The pricing problem is an intractability theorem: institutions can reprioritize and repricing can shift the equilibrium. The persistence floor is an impossibility theorem: no institutional refinement changes the physical fact that fission creates waste whose hazard horizon exceeds every demonstrated human institution. Finland's Onkalo repository shows the floor can be managed, not legislated away.

• Better governance can reduce burden, shift it into better containers, fund it more honestly
• It cannot turn long-lived spent fuel into an ordinary industrial residue
• The physical source of the burden is not a reporting failure or weak regulator

New build must compete against zero-carbon substitutes with no radioactive legacy

New nuclear cost estimates remain above utility-scale solar, onshore wind, and many firmed renewable portfolios. Once cheaper zero-carbon substitutes are available, the fact that nuclear still produces positive net value is not enough to justify expansion. The correct comparison is relative welfare productivity, not absolute survival.

• Nuclear's welfare-positive result would be more decisive if the alternative were still coal
• For new-build decisions in 2026, the operational comparison is against renewable portfolios and storage
• A technology that is expensive on private terms and carries a large persistent liability must compete against substitutes with lower private cost and no radioactive legacy

Rule change R transforms the flawed game G into G1 with tripartite payoff space

The rule change requires mandatory full-cost decision accounting at authorization, liability realism with funded back-end obligations, and a two-tier policy rule distinguishing legacy-fleet operation from new-build approval. Under G1, the operator cannot record a win unless the system also wins: liabilities are fully funded, disposal paths are real and financed, and the intergenerational burden is priced at the decision point.

• Component 1: Complete 17-field DA record with marginal-beta welfare quantification before decision
• Component 2: Replace liability caps with mandatory insurance pool; fully fund decommissioning at revealed-cost levels
• Component 3: Two-tier rule: existing plants can continue if they provide low-cost low-carbon power and fully fund liabilities

Two-tier rule: keep existing plants that fund liabilities; require stricter test for new build

Existing plants warrant continued operation where they supply low-cost low-carbon electricity, fully fund decommissioning and waste liabilities, and operate under liability regimes that price the public backstop. New build must compete with zero-carbon substitutes that do not create a radioactive legacy and must show a funded disposal path before construction.

• The rule follows from the SAPM theorem taxonomy
• Pricing problem is intractable but improvable; persistence floor is impossible to eliminate
• The paper identifies the flawed game, names the Hollow Win equilibrium, and specifies the rule change that moves the payoff space from bilateral to tripartite