Chapters 2 and 3 showed that competitive markets produce an equilibrium that maximizes total surplus. The price system, as we argued in Chapter 1, coordinates decentralized decisions with remarkable efficiency. But this result depends on conditions that sometimes fail to hold. When they do, markets allocate resources inefficiently — producing too much of some things and too little of others.
The conditions for market efficiency include: (1) no costs or benefits fall on third parties outside the transaction, (2) goods are rival and excludable, (3) buyers and sellers have adequate information, and (4) there are many buyers and sellers (no market power — addressed separately in Chapter 7). When any of these conditions breaks down, we have a market failure — a situation where the market equilibrium is not Pareto efficient.
This chapter identifies four categories of market failure: externalities, public goods, common resources, and information asymmetry. These are not exceptions to be memorized; they are systematic patterns with a common structure. For each, we ask the same questions: Why does the market get it wrong? How far off is it? What, if anything, can be done — and at what cost?
Externalities are everywhere. When a factory pollutes a river, it imposes costs on downstream fishers that don't appear in the factory's cost calculations. When a homeowner maintains a beautiful garden, it raises the property values of neighbors — a benefit the gardener doesn't capture. When a driver enters a congested highway, she slows down every other driver — a cost she doesn't pay. In each case, the private decision-maker considers only their own costs and benefits, not the effects on others.
A negative externality exists when a transaction imposes costs on third parties. The producer or consumer makes a decision based on private costs, ignoring the costs imposed on others. The result: too much of the activity.
The market equilibrium occurs where demand (marginal benefit) equals supply (MPC). But the socially optimal quantity is where demand equals MSC — which accounts for all costs, including those borne by third parties. Since $MSC > MPC$, the socially optimal quantity is lower than the market quantity. The market overproduces the externality-generating good.
The deadweight loss from the externality equals the area between MSC and demand, from $Q^*$ (social optimum) to $Q_M$ (market quantity). This triangle represents the net cost to society of the excess production — units where the full social cost exceeds the benefit to consumers.
Figure 4.1. Negative externality. Drag the MEC slider to see how the marginal external cost drives a wedge between private and social cost. The MSC curve separates from MPC, the socially optimal quantity falls, and the DWL triangle grows. The optimal Pigouvian tax equals the MEC. Hover for values.
Real-world examples of negative externalities:
A positive externality exists when a transaction confers benefits on third parties. The market produces too little of these goods because the private benefit understates the social benefit.
where MSB is the marginal social benefit, MPB is the marginal private benefit (reflected in the demand curve), and MEB is the marginal external benefit.
Real-world examples of positive externalities:
How can we fix externalities? One approach: change the prices to reflect true social costs.
After the tax, the producer's effective cost becomes $MPC + t^* = MSC$, and the market equilibrium coincides with the social optimum. The deadweight loss from the externality is eliminated.
For positive externalities, the Pigouvian subsidy is equal to MEB at the socially optimal quantity. The subsidy lowers the effective price to consumers, encouraging them to buy more — pushing quantity up to the social optimum.
Demand for steel: $P = 100 - Q$. MPC (supply): $P = 20 + Q$. Constant $MEC = 10$ per unit.
Market equilibrium: \$100 - Q = 20 + Q \Rightarrow Q_M = 40$, $P_M = 60$.
Social optimum: $MSC = 30 + Q$. Set \$100 - Q = 30 + Q \Rightarrow Q^* = 35$, $P^* = 65$.
DWL: $\frac{1}{2}(10)(5) = 25$.
Optimal Pigouvian tax: $t^* = MEC = \\$10$ per unit. With the tax, producers face \$10 + Q = MSC$. New equilibrium: $Q = 35$, $P_B = 65$, $P_S = 55$. DWL eliminated.
Tax revenue: \$10 \times 35 = \\$150$. Pigouvian taxes generate a "double dividend" — they correct the externality and raise revenue.
Figure 4.2. Pigouvian tax correction. Toggle between the unregulated market and the optimal tax. With the tax, the effective supply curve shifts up to MSC and the DWL is eliminated. Hover for values.
Pigouvian taxes work beautifully in theory but face practical challenges:
An alternative to government intervention: let the affected parties bargain with each other.
A factory's pollution damages a neighboring farmer by \$10 per unit. The factory earns \$10 profit per unit. Efficient outcome: no production (cost \$10 > benefit \$10).
Case 1 — Farmer has rights: Factory needs permission to pollute. Must pay farmer ≥ \$10, but only earns \$10. Cannot afford it. Result: no pollution. Efficient.
Case 2 — Factory has rights: Farmer pays factory between \$10 and \$10 to stop. Both gain. Result: no pollution. Efficient.
Same outcome either way. Only the distribution of wealth differs.
Figure 4.3. Coase bargaining. Toggle property rights and slide transaction costs. When TC = 0, the efficient outcome (no production) emerges regardless of rights allocation. As TC rise, the bargaining surplus shrinks and eventually bargaining fails. Hover for details.
The Coase theorem requires three conditions that often fail in practice:
1. Well-defined property rights. Who owns the right to clean air? To a stable climate? In many externality situations — especially environmental ones — property rights are ambiguous, contested, or unenforceable.
2. Low transaction costs. Bargaining must be cheap. The Coase theorem works well for two neighbors negotiating over a barking dog. It fails spectacularly for air pollution, where millions of affected parties would need to negotiate with thousands of polluting firms.
3. No strategic behavior or information asymmetry. Parties must bargain honestly. In practice, each side has an incentive to misrepresent their costs or benefits. The holdout problem can prevent agreement even when a mutually beneficial deal exists.
The Coase theorem is most useful not as a practical solution but as a diagnostic tool. It identifies the reason markets fail at handling externalities: transaction costs.
These two properties — non-rivalry and non-excludability — create distinct problems. Non-rivalry means the efficient price is zero (the marginal cost of an additional user is zero). Non-excludability means private firms cannot charge any price. Together, they imply that private markets cannot provide public goods efficiently.
| Excludable | Non-excludable | |
|---|---|---|
| Rival | Private good: food, clothing | Common resource: ocean fish, clean air |
| Non-rival | Club good: cable TV, toll road | Public good: national defense, lighthouse |
What is the efficient level of a public good? For a private good, efficiency requires $MB_i = MC$ for each consumer. For a public good, all consumers consume the same quantity simultaneously. Efficiency requires the sum of marginal benefits to equal marginal cost:
This is the Samuelson condition (Samuelson, 1954). Graphically, we vertically sum the individual MB curves (because everyone consumes the same quantity) and find where the aggregate MB equals MC.
3 households: $MB_1 = 10 - Q$, $MB_2 = 8 - Q$, $MB_3 = 6 - Q$. Marginal cost: $MC = 6$.
$\sum MB = 24 - 3Q$. Samuelson condition: \$14 - 3Q = 6 \Rightarrow Q^* = 6$ hours.
Private provision: Household 1 provides where $MB_1 = MC$: \$10 - Q = 6 \Rightarrow Q = 4$ hours. Others free-ride. Underprovision: 4 instead of 6.
Figure 4.4. Public goods: vertical summation. Adjust each household's willingness to pay. The bold green curve is the vertical sum of all three MB curves. The Samuelson optimal quantity is where ΣMB = MC. Private provision (where the highest individual MB = MC) always falls short. Hover for values.
Examples abound: ocean fish stocks, groundwater aquifers, the atmosphere as a carbon sink, common grazing land, public roads during rush hour, and wild game. In each case, the resource is depletable (rival) but open to all (non-excludable).
The logic is identical to a negative externality. Each fisher who takes an additional fish receives the full market value of that fish but imposes a cost on all other fishers by reducing the remaining stock. The private marginal cost is below the social marginal cost, so the resource is overexploited.
Figure 4.5. Tragedy of the commons. Drag the slider to add users. Each user takes more than their socially optimal share because they ignore the depletion externality they impose on others. With a single owner, extraction is efficient; with many users, the resource is severely overexploited. Hover for values.
1. Property rights (privatization). Assign ownership to an individual or firm. The owner internalizes the full depletion cost. Iceland's individual transferable quota (ITQ) system for fishing is a successful example.
2. Regulation. Government-imposed limits on extraction: fishing quotas, hunting seasons, water use permits, emission standards.
3. Pigouvian taxes. Tax each unit of extraction at a rate equal to the marginal external cost. Congestion pricing on roads is an example.
4. Community governance (Ostrom). Elinor Ostrom (Nobel 2009) studied communities that successfully manage commons without privatization or government regulation. Success requires: clearly defined boundaries, rules adapted to local conditions, participation of users in rule-making, effective monitoring, graduated sanctions, and accessible conflict resolution.
Markets assume that buyers and sellers have sufficient information to make good decisions. When one side knows materially more than the other — asymmetric information — markets can malfunction in predictable ways.
Sellers know whether their car is reliable ("peach," worth \$10,000) or defective ("lemon," worth \$1,000). Buyers cannot tell. With 50/50 odds, buyers offer \$1,500. But peach owners refuse — their car is worth \$10,000. Only lemons sell. Buyers learn this and offer only \$1,000.
Result: The market for good used cars disappears. High-quality sellers exit, leaving only low-quality sellers.
Real-world solutions to adverse selection:
With fire insurance, a homeowner may become less careful about fire prevention. With health insurance, patients may visit the doctor more often. Moral hazard is fundamentally a problem of hidden action. Solutions include:
Both adverse selection and moral hazard are introduced here intuitively. Chapter 11 formalizes adverse selection through the revelation principle and mechanism design. Chapter 10 provides the formal framework for thinking about information and incentives.
Maya's lemonade stand generates a positive externality. Neighbors report that foot traffic from Maya's customers has increased visits to nearby shops. The estimated marginal external benefit is \$1.30 per cup.
Should the city subsidize Maya?
$MSB = MB + MEB = (5 - Q/20) + 0.30 = 5.30 - Q/20$. Setting $MSB = MPC$:
\$1.30 - Q/20 = 0.50 + Q/20 \Rightarrow Q^{**} = 48$ cups (vs. market $Q = 45$).
A Pigouvian subsidy of \$1.30/cup would achieve this. But the city taxed Maya \$1.50/cup (Chapter 3), pushing output to 40 — the wrong direction. The tax was motivated by revenue needs, not efficiency. Understanding the externality framework clarifies what we're trading off.
| Label | Equation | Description |
|---|---|---|
| Eq. 4.1 | $MSC = MPC + MEC$ | Marginal social cost with negative externality |
| Eq. 4.2 | $MSB = MPB + MEB$ | Marginal social benefit with positive externality |
| Eq. 4.3 | $t^* = MEC$ at $Q^*$ | Optimal Pigouvian tax |
| Eq. 4.4 | $\sum_{i=1}^{N} MB_i = MC$ | Samuelson condition for public goods |