Applying the System Asset Pricing Model to Gig Economy Platforms: Measuring the System Welfare Cost of Worker Misclassification

Gig platforms generate $0.76 in system welfare cost per dollar of annual industry revenue

The System Asset Pricing Model (SAPM) measures welfare cost per dollar of industry profit. For US ride-hail and delivery platforms, the system welfare beta β_W = 0.39 (90% CI: 0.6–0.9).

• Annual private payoff Π = $88B (platform take rate × GMV, 2024)
• Annual system welfare cost SW = $34.4B
• System-adjusted payoff ΠSA = Π − SW = +$10.6B — still positive, but narrow

The externality is the business model, not a byproduct

Pigouvian taxation fails because wage suppression and benefit denial are the revenue model, not side effects. Coasean bargaining fails because property rights are obscured, transaction costs are structurally infinite, and millions of workers cannot negotiate with three platforms.

• Uber raised passenger prices 50% (2017–2023) while cutting driver pay 11.9% per year — margin extraction, not efficiency
• A single driver cannot negotiate take rates with a $150B corporation; terms are set by algorithm and changed without notice
• Prop 22: 205.7M in political spending preserved 45B/year in private payoff — a 40:1 annual return

Wage suppression, foregone benefits, and shifted vehicle costs are the largest channels

Six channels trace how platform profit is organized through costs displaced outside the firm. The three largest account for 80% of total welfare cost.

• Channel 1 — Wage suppression: 10.8B. Median gig worker earns 11.75/hr vs. 17.62/hr for taxi drivers; bottom quartile earns 5.12/hr after expenses
• Channel 2 — Foregone benefits: $9.5B. Platforms avoid 7.65% FICA, UI, workers' comp, health insurance, paid leave
• Channel 3 — Shifted vehicle costs: 7.4B. IRS mileage rate 0.67/mile; full-time driver bears $16,750/year in vehicle costs; 43% fail to claim mileage deduction

Occupational injury, fiscal externalities, and regulatory capture add $14.1B more

The remaining three channels complete the welfare cost picture, with injury risk concentrated among delivery workers and regulatory capture preserving the classification arbitrage.

• Channel 4 — Injury and fatality risk: $2.8B. NYC delivery worker fatality rate 36/100K FTE — 5× construction; 21.9% report injury on the job
• Channel 5 — Fiscal externalities: $4.2B. Gig workers rely on SNAP, Medicaid; taxpayers subsidize platform profits
• Channel 6 — Regulatory capture: $2.7B. Prop 22 spending, lobbying, and litigation costs that preserve misclassification

The Drivers Cooperative proves the welfare gap is a choice, not a necessity

Functioning platform cooperatives operate at 15% take rates and return ~88% of revenue to drivers, showing that the current 40% take rate and $34.4B welfare cost are institutionally made.

• The Drivers Cooperative (NYC): 5.9M revenue in 2022, 5.2M returned to drivers, guarantees $30/hr
• Green Taxi Cooperative (Denver): 37% market share, returns ~90% of gross earnings to drivers
• Cooperative counterfactual payoff ΠC > 0 — observed, not assumed

β_W = 0.39 with 98% probability that β_W < 1.0

100,000 Monte Carlo draws propagate uncertainty across all six channels. The median β_W is 0.39; the 90% CI is [0.6, 0.9]. The probability that the industry is a net value destroyer (β_W > 1.0) is less than 2%.

• Marginal beta (at current take rate) = 1.29 — 1.7× the average beta, indicating accelerating damage at the margin
• System welfare ratio κ = SW / Π = 0.39
• Break-even Pigovian correction μ* = 1 − (1/β_W) = 0.786 per worker

Gig platforms rank 60 of 61 in the SAPM registry — borderline welfare-neutral

The SAPM registry compares welfare costs across otherwise incommensurable domains. Gig platforms sit near the bottom, with β_W lower than Bitcoin mining, PFAS, and nuclear power but higher than the least damaging sectors.

• Bitcoin mining: β_W ≈ 3.2; PFAS: β_W ≈ 2.1; nuclear power: β_W ≈ 0.5
• Gig platforms' β_W = 0.39 reflects a narrow welfare gap but a distributional burden falling disproportionately on workers
• The SOF (Missing System surface) is sharply concave: at 40% take rate, marginal cost exceeds average by 1.7×

2026–2028 is the structural break window

Three converging deadlines create a narrow window for institutional intervention before automation displaces human drivers.

• EU Platform Work Directive transposition deadline: December 2, 2026 — shifts burden of proof to platforms, bans automated deactivation without human review
• Waymo robotaxi scaling: 1M weekly rides by end of 2026 — automation begins to displace human drivers
• California AB 1340 (2025): collective bargaining for rideshare drivers — first US law allowing contractor organizing

Commission caps dominate minimum wages as a regulatory instrument

Fisher's (2024–2025) structural model shows that a commission cap of τ ≤ 0.20 would reduce β_W by approximately 1.8 points, holding all else constant. Minimum wages reduce utilization; commission caps transfer rents directly from platform to worker.

• Portable benefits mandates (delinking safety net from employment status) would close Channel 2
• Employment presumption (ABC test) would eliminate the classification-arbitrage floor
• The welfare gap cannot be closed through voluntary market mechanisms; it requires exogenous institutional restructuring

Misclassification is a policy choice, not a physical constraint

The paper recasts classification as an organizational mechanism that sorts obligations, allocates risk, and determines where welfare costs settle. The $34.4B annual cost is institutionally made and can be institutionally unmade.

• No impossibility theorem applies — this is institutional SOT, not a physical constraint
• The EU Platform Work Directive, UK Supreme Court's Uber v. Aslam, and functioning cooperatives show the game is being transformed
• The window for intervention narrows by 2030 as automation displaces human drivers