Applying the System Asset Pricing Model
Decision Accounting
Applying the System Asset Pricing Model to Ultra-Processed Food: Measuring the System Welfare Cost of Engineered Dietary Harm
core
Core Claim
Every dollar of UPF processing-premium industry revenue is associated with $6.20 in system welfare
The System Asset Pricing Model (SAPM) translates CAPM logic to welfare economics: system-adjusted payoff = private payoff minus system beta times welfare at risk. For ultra-processed food, private payoff is 293B/yr from snack and SSB margins; system beta is 6.2 [5.2–7.6]; system welfare cost is 1.81T/yr.
- βW = 6.2 means 6.20 destroyed per 1 earned
- System-adjusted payoff = –$1.52T/yr
- No discount rate rescues this; no efficiency argument survives
bridge
SAPM Bridge
CAPM becomes SAPM: price risk → welfare risk
CAPM: E[Ri] = Rf + βi(E[Rm] – Rf). SAPM: ΠSA = Π – βW·W. System beta measures covariance of private payoff with system welfare destruction.
- CAPM asset return → SAPM private payoff Π = $293B/yr
- CAPM market portfolio → SAPM system welfare W0 = $4.2T/yr
- CAPM beta → SAPM system beta βW = 6.2
- CAPM risk-free rate → SAPM cooperative baseline ΠC = $68B/yr
extraction
Scale of Extraction
UPF supplies 60% of US calories; processing premium extracts $293B/yr
Farmers receive 1.3¢ of every soda dollar and 9.7¢ of every snack dollar. The remaining 90–99% is the processing premium: corporate margins disconnected from nutritional value. The Hall et al. RCT showed 508 excess kcal/day on a UPF diet.
- US snack market 186.4B; SSB market 293.5B
- Processing premium = $293B/yr in corporate margins
- Farm share: 1.3¢/soda dollar, 9.7¢/snack dollar
- Hall et al. (2019): 0.9 kg weight gain in 2 weeks on UPF diet
failure
Why Pigou and Coase Fail
Diffuse harm, captured regulator, infinite transaction costs
Pigouvian taxation requires a benevolent regulator; 95% of the Dietary Guidelines Advisory Committee had industry COIs. Coasean bargaining requires low transaction costs; organizing 330M consumers against Coca-Cola and the ABCD grain oligopoly is impossible.
- Marginal social cost of one Dorito is infinitesimal; aggregate is $1.81T
- Industry spends $178M/yr on federal lobbying
- Every FDA commissioner since Kessler joined an industry board
- Information asymmetry: 68% of US food supply is hyper-palatable, labeled as healthy
channels
Six Channels
Metabolic disease alone costs 1.10T/yr; five other channels add 0.71T
Channel 1 (metabolic disease) is anchored by Hall et al. RCT and Rockefeller Foundation $1.1T estimate. Channels 2–6: addiction/behavioral hijacking, agricultural distortion, environmental degradation, governance capture, intergenerational harm.
- Metabolic disease: βW,1 = 9.4, $1.10T/yr
- Addiction: 8–15% of adults meet YFAS food addiction criteria
- Governance capture: $178M/yr lobbying, 95% COI rate on DGAC
- Intergenerational: maternal UPF exposure harms child development
beta
Aggregate Beta
βW = 6.2 [5.2–7.6]; 100% of Monte Carlo draws show βW > 3.7
Channel-weighted system beta: SW = 1.81T / Π = 293B = 6.2. 100,000 Monte Carlo draws confirm βW > 5.0 in 94% of runs. Break-even mitigation rate μ* = 0.87: 87% of damage must be eliminated before net positive social value.
- SW = 1.81T/yr (metabolic 1.10T + other $0.71T)
- Π = 293B/yr; ΠC = 68B/yr (whole-food farm-share equivalent)
- System-adjusted payoff ΠSA = –$1.52T/yr
- κ = 0.43 (system capture ratio)
frontier
Pareto Frontier
Marginal expansion is catastrophically value-destroying
The Pareto System Frontier is strongly concave: marginal βW (10.0–12.9) exceeds average βW (6.2) by a factor of 1.6–2.1. Each additional dollar of processing-premium extraction destroys more welfare than the last.
- No safe threshold: Liang et al. (2025) linear dose-response, 10% higher all-cause mortality per 10% UPF increment
- Addiction channel accelerates as HPF market saturation rose from 49% (1988) to 68% (2018)
- Governance capture compounds with lobbying entrenchment
rank
Cross-Domain Rank
UPF ranks second-highest βW domain calibrated, below PFAS
UPF βW = 6.2 exceeds Bitcoin (5.0), frontier AI (7.4), auto emissions (6.8), and antimicrobial resistance (2.1). Only PFAS (35.2) is higher. The industrialized food system receives far less regulatory scrutiny than comparable domains.
- PFAS: βW = 35.2
- Bitcoin: βW = 5.0
- Frontier AI: βW = 7.4
- Nuclear energy: βW = 0.7
class
Classification
Institutional failure, no impossibility theorem, high system beta
Under the PST classification grid, UPF is an institutional failure: welfare cost arises through established market institutions and captured regulatory bodies. No formal impossibility blocks intervention; Chile, Mexico, UK, Finland, and Brazil have shown regulation works.
- Chile's FLAL: 23.8% reduction in calories from labeled products
- Mexico's SSB tax: 7.6% sustained decline in purchases
- UK SDIL: 6,600 kcal/person/yr removed, 80% via reformulation
- Political obstacle, not logical: $178M/yr lobbying protects status quo
change
What It Changes
SAPM reveals a $1.52T/yr negative system-adjusted payoff that standard metrics miss
Conventional metrics (consumer surplus, employment, GDP) show UPF as successful. SAPM corrects three errors: treating revealed preference as revealed welfare, netting processing premium against zero instead of cooperative baseline, and zeroing out governance capture costs. The binding constraint is political, not theoretical.
- Standard metrics: cheap calories (0.55/100 kcal vs. 1.45), 2.1M jobs, $900B+ output
- SAPM: –$1.52T/yr system-adjusted payoff
- μ* = 0.87: no existing policy approaches break-even
- Institutional redesign is feasible; the obstacle is political will