The Efficient Markets Welfare Exclusion Theorem: Informational Efficiency, Allocative Welfare, and the Price of Everything Except the Cost

Informational efficiency does not guarantee allocative welfare

Fama (1970) defined efficient markets as those where prices fully reflect available information. The implicit welfare claim—that informational efficiency produces allocative welfare—is false. This paper derives the formal welfare decomposition W(p) = Π_I(p) − ΔW_U(p) − ΔW_S(p).

• Π_I(p) is the private surplus of informationally advantaged participants
• ΔW_U(p) is the adverse-selection cost borne by uninformed traders (Grossman-Stiglitz wedge)
• ΔW_S(p) is the system-level welfare cost from the gap between efficient prices and optimal prices

The EMH measures price discovery, not welfare

Event studies show stock prices respond to earnings announcements within hours. Mutual fund managers do not systematically beat passive benchmarks after fees. This evidence supports semi-strong informational efficiency but does not test whether the resulting capital allocation is welfare-optimal.

• EMH tests: speed and accuracy of price response to information
• Welfare question: does the allocation maximize total welfare including non-market participants?
• A stock price can be informationally efficient and allocatively catastrophic

EMH-efficient prices exclude social costs

Informationally efficient prices reflect private values. Social costs—climate damage, health costs, community disruption—enter prices only if they affect expected cash flows through regulation or litigation. When unpriced, EMH-efficient prices over-allocate capital to externality-generating activities.

• Global climate damages estimated at $23 trillion present value (Nordhaus 2017) are not in stock prices
• Fossil fuel β_W = 6.95: for every 1 of private value, 6.95 in social cost is excluded
• Tobacco β_W = 6.50, plastics β_W = 6.67, data brokerage β_W = 6.13

Missing markets prevent welfare-optimal risk-sharing

EMH's welfare claim requires complete markets—every risk must be tradeable. In reality, workers cannot insure against industry-specific unemployment, farmers cannot trade climate risk, and most welfare-relevant contingencies have no market. EMH-efficient prices are efficient only conditional on the existing incomplete market structure.

• Arrow (1964) and Debreu (1959): competitive equilibrium is Pareto efficient only with complete markets
• Magill and Quinzii (1996): incomplete markets generically produce constrained inefficient equilibria
• Labor income, health risk, and climate risk are largely uninsurable through capital markets

Informational efficiency requires a permanent welfare transfer

Grossman and Stiglitz (1980) showed that perfectly efficient markets are impossible: if prices fully reflect information, no one acquires it. The equilibrium requires a transfer from uninformed to informed traders. Kyle (1985) and Glosten and Milgrom (1985) formalize the adverse-selection mechanism through bid-ask spreads.

• Retail investors lose 0.5–2% of portfolio value annually to adverse selection (Barber and Odean 2000)
• A 500,000 portfolio over 30 years loses ~225,000–$300,000 in cumulative costs
• Three layers: trading-cost transfer, timing transfer, structural transfer (payment-for-order-flow)

EMH is blind to who bears the costs of efficient pricing

EMH evaluates efficiency without regard to distribution. Efficient allocation can concentrate returns among asset owners while dispersing costs across non-participants. The wealth distribution itself conditions what information is produced and reflected in prices.

• Carbon-intensive firms: shareholders capture gains, communities bear climate costs
• Piketty (2014): long-run wealth concentration driven by capital income
• Saez and Zucman (2016): top 0.1% wealth share at levels not seen since 1920s

Welfare decomposes into three terms: Π_I − ΔW_U − ΔW_S

Total welfare W(p) = Π_I(p) + Π_U(p) + Π_S(p) − L_U(p) − C_S(x(p)). Collecting terms: Π_I(p) is priced private surplus, ΔW_U(p) = L_U(p) − Π_U(p) is net adverse-selection cost, ΔW_S(p) = C_S(x(p)) − Π_S(p) is net system cost. EMH-efficient prices maximize Π_I(p) but ignore ΔW_U and ΔW_S.

• EMH pricing condition: p = E[V_private I]
• Social planner's condition: p_W = E[V_private − C_S + H_missing I^W]
• Welfare-completeness requires ΔW_U = 0 and ΔW_S = 0—conditions that fail in every real market

EMH-efficient prices are welfare-complete only under impossible conditions

Proposition 1: W(p) = Π_I(p) − ΔW_U(p) − ΔW_S(p). EMH-efficient prices maximize Π_I(p). They are welfare-complete iff ΔW_U = 0 (no adverse selection) and ΔW_S = 0 (no externalities and complete markets). Both conditions fail simultaneously in every real financial market.

• Corollary 1.1: When ∂ΔW_S/∂Π_I > 0, more efficient pricing of externality-generating assets increases welfare cost
• Corollary 1.2: Temporal asymmetry—Π_I is front-loaded, ΔW_S is back-loaded (climate damage, intergenerational effects)
• The EMH debate has focused on the wrong question: not Fama vs. Shiller, but whether efficient prices produce welfare-optimal allocation

Efficiency-welfare divergence grows with externalities, incompleteness, and information heterogeneity

Proposition 2: p_EMH − p_W increases in (i) unpriced externalities, (ii) market incompleteness, (iii) information heterogeneity. Under (i) alone, divergence equals the Pigouvian tax. Under (ii), the Arrow-Debreu gap. Under (iii), the Grossman-Stiglitz transfer.

• Carbon social cost $50–200/ton implies stock-price adjustment of 10–60% for carbon-intensive firms
• Bolton and Kacperczyk (2021): carbon risk premia of 1–2% annually—far below full social cost
• SAPM calibrations: coal β_W=6.95, tobacco β_W=6.50, payday lending β_W=7.08

No single instrument closes all three welfare channels

Proposition 3: Externality pricing (Pigou), market completion (Arrow), and adverse-selection mitigation (Kyle-Glosten-Milgrom) are complementary but individually insufficient. Carbon pricing closes the externality channel but leaves adverse selection open. Market completion leaves externalities open.

• EU Emissions Trading System: externality pricing, but adverse selection persists
• British Columbia's carbon tax: internalizes carbon cost, does not address market incompleteness
• Nordic financial governance model: adverse-selection mitigation through transparency and regulation

Welfare-complete capital allocation requires three complementary instruments

The theorem is an Intractability result, not an Impossibility. The welfare gap reflects institutional choices—disclosure-only regulation, absence of externality pricing, incomplete markets. Closing the gap requires externality pricing, market completion, and adverse-selection mitigation as complementary instruments.

• EU ETS and BC carbon tax show externality channel can be closed
• Denmark's pension system demonstrates market completion for retirement risk
• Nordic model shows adverse-selection mitigation through market design

The theorem shifts the object of the EMH debate from information to welfare

The Fama-Shiller debate asks whether prices are informationally efficient. The welfare theorem asks whether informationally efficient prices are welfare-optimal. Even if Fama is completely correct, the resulting capital allocation may be welfare-suboptimal because reflected information excludes social costs, incomplete markets prevent risk-sharing, and price discovery transfers wealth from uninformed to informed participants.

• Changes the test: from 'are prices efficient?' to 'is the allocation welfare-optimal?'
• Explains why EMH provided intellectual basis for leaving externalities unpriced and markets incomplete
• Connects to the Missing System Theorem: bilateral payoff spaces exclude system welfare by construction