ECON 132 · Public Economics
Chapter 4 · Stiglitz / Rosengard

Market Failure

When does the market fail? The six classical sources of inefficient market outcomes — and why the government can (sometimes) help.

01 02 03 04 05

If the free market is Pareto-efficient under ideal conditions (Chapter 3) — why then do we need the government? Answer: because the ideal conditions are rarely met in reality. Stiglitz identifies six sources of market failure.

1 · Property Rights and Contract Enforcement

Before talking about market failure, you have to understand the basic preconditions of a functioning market: clear property rights and enforceable contracts. Without them, the market collapses.

Tragedy of the Commons Hardin's famous example: on a common pasture, all herders can graze their sheep. Each herder has an incentive to keep more sheep — the costs of overgrazing are borne by the community. Result: overuse. The absence of clearly defined property rights is a fundamental source of market failure.

There is also the reverse tragedy (tragedy of the anticommons): when too many actors have veto rights over a resource, it is underused (e.g., patent thickets in biotechnology).

2 · Failure of Competition

The perfect competition model assumes that every actor is a price taker. In reality there are:

  • Monopolies — a single supplier (Microsoft Windows in the 1990s).
  • Oligopolies — few suppliers (US mobile carriers: AT&T, Verizon, T-Mobile).
  • Monopolistic competition — many suppliers, but differentiated products (restaurants, branded goods).

In all cases, suppliers reduce output below the efficient level to keep prices high. Result: deadweight loss.

Monopoly deadweight loss
Figure 4.1 · Monopoly loss Under monopoly, quantity Qm is produced (where MR = MC) instead of the efficient quantity Qe (where P = MC). The price Pm is above MC. The shaded triangle is deadweight loss — welfare that no one captures.

Natural Monopolies

A special case: natural monopoly — industries in which average costs keep falling (massive economies of scale). Examples: electricity distribution, water supply, rail networks. Competition here would be wasteful (parallel infrastructure). Solutions:

  • Regulated private monopoly (US utilities).
  • Public production (water utilities, formerly EDF).

3 · Public Goods

Some goods are either not produced at all by the market, or are produced in too small quantities. Pure public goods have two properties:

Two Properties of Public Goods

Non-rival consumption: my consumption does not reduce anyone else's consumption. If I see a lighthouse, the next sailor can see it just as well.

Non-excludable: it is (nearly) impossible to exclude anyone from consumption.

Examples: national defense, lighthouses, freely accessible knowledge, clean air.

Free rider problem
Figure 4.2 · Free rider problem Because no one can be excluded, no one has an incentive to pay voluntarily. The market produces public goods in suboptimal quantities (or not at all) — the classic argument for government intervention.

4 · Externalities

An externality exists when one party's action has consequences for another that are not reflected in the market price.

  • Negative externality: air pollution, noise, traffic congestion, climate change.
  • Positive externality: education, research, vaccinations, attractive front yards.

With negative externalities, the market produces too much (private costs < social costs). With positive externalities, too little (private benefits < social benefits).

Externality and welfare loss
Figure 4.3 · Externalities With a negative externality, the market equilibrium lies above the socially optimal quantity. The welfare loss is the area between the social marginal cost (MSC) curve and market demand, between Qopt and Qmarket.

Coase Theorem: if property rights are clearly defined and transaction costs are zero, externalities can be resolved privately — regardless of who holds the right. The reality: transaction costs are rarely zero.

5 · Incomplete Markets

Some markets simply don't exist — even though there is a theoretical need. Classic examples:

Insurance Markets

There is no private insurance against:

  • Unemployment (too much moral hazard).
  • Inflation (too hard to hedge).
  • Obsolescence of job skills.
  • Lifelong long-term care costs (too little data, long time horizons).

These are precisely the gaps that public social insurance fills.

Capital Markets

The private market underserves:

  • Students without collateral (Stiglitz's own research area!).
  • Small entrepreneurs in developing countries (microfinance emerged as a response).
  • Research with uncertain returns.

Why incomplete?

Three main reasons:

  • (a) Innovation: new risks are not immediately insurable.
  • (b) Transaction costs: some markets would be too small to cover administrative costs.
  • (c) Asymmetric information: adverse selection and moral hazard destroy markets (see next section).

6 · Information Failures

The ideal competitive model assumes complete information. In reality, market participants often have asymmetric information:

Adverse Selection (Akerlof's "lemons")

When sellers know more about quality than buyers (used cars, health insurance), the good risks get pushed out of the market. Example: insurers cannot distinguish good from bad risks → uniform premium → good risks find it too expensive and drop out → only bad risks remain → premium has to rise → even more good risks drop out → adverse selection spiral.

Moral Hazard

Once someone is insured, their behavior changes — they become more careless (full-coverage effect) or consume more services (health insurance). Insurers cannot perfectly observe this.

Information is itself a public good

Information that would make markets efficient (product quality, the financial health of firms) is often itself non-rival — private firms produce too little of it. Hence mandatory disclosure (SEC), consumer testing (Consumer Reports), and government statistics.

7 · Unemployment, Inflation, and Disequilibrium

One of the most obvious examples of market failure: involuntary unemployment. If the market worked perfectly, wages should adjust so that all those willing to work find employment. In recessions they don't.

Similarly: persistent inflation, bank panics, financial crises — all of these point to structural market failure that justifies macroeconomic stabilization (monetary and fiscal policy).

8 · Redistribution and Merit Goods

Even if all markets work perfectly, the government can intervene for two further reasons:

Income Redistribution

The First Welfare Theorem guarantees only efficiency, not distribution. If the market distribution is perceived as unjust, redistribution is a separate government task in its own right.

Merit Goods

Merit goods are goods that the government believes individuals consume too little of — even if individuals themselves don't see it that way. Examples: education, cancer screening, seat belts. The opposite: demerit goods like tobacco and alcohol.

9 · Paternalism vs. Libertarianism

Here a fundamental conflict comes to light:

  • Paternalism: the government sometimes knows better than the citizen what is good for them (seat-belt laws, drug prohibitions, drinking age).
  • Libertarianism: the government should not interfere in individual decisions as long as no one else is harmed (Mill's "harm principle").

Stiglitz names two exceptions where even liberals accept paternalism:

  • Children: they cannot decide for themselves.
  • Drug addicts & similar behavioral addictions: when a person can no longer control their own will.
Chapter Take-away Markets fail for six reasons: lack of competition, public goods, externalities, incomplete markets, information asymmetries, and macroeconomic shocks. Each reason provides an economic argument for possible government action. But: market failure does not justify intervention — it opens the possibility. Whether the government actually solves the problem better has to be checked in each specific case.

Flashcards — Chapter 4

Click the card to flip it. Use the arrow keys to navigate.

Term
Click to flip
Definition
Click again to flip back

Quiz — Chapter 4

14 multiple-choice questions.

Problem Set — Chapter 4

Seven open-ended exam-style questions.

1 For each of the following programs: which market failure does it address? (a) Medicare (b) EPA emission limits (c) public broadcasting (d) Federal Deposit Insurance (e) government research funding (f) antitrust law

(a) Medicare: adverse selection in the health insurance market for seniors — a private market would lead to unaffordable premiums. Also positive externalities (preventing infection, quality of life for the family).

(b) EPA emission limits: negative externalities from air pollution. Without regulation, firms produce too much pollution (private costs < social costs).

(c) Public broadcasting: a public good (non-rival, non-excludable for free-to-air programming). Alongside that, an information market failure — the private market underproduces high-quality news content (ad revenue prefers entertainment).

(d) FDIC deposit insurance: prevents bank runs (a form of macroeconomic market failure). Information asymmetry: savers cannot tell whether their bank is solvent.

(e) Research funding: positive externalities from basic research (the private market underinvests because returns are hard to capture). Knowledge is also a public good.

(f) Antitrust: failure of competition — preventing monopolistic market power and its deadweight loss.

2 Explain adverse selection using the example of private health insurance without a mandate. Why is this an economic argument for an insurance mandate (as in Obamacare/ACA)?

Adverse selection spiral

The insurer cannot perfectly distinguish healthy from less-healthy enrollees. It sets an average premium. Healthy enrollees find the premium too high (they know they are healthy) and drop out. The pool is left with a disproportionate share of sick enrollees. The insurer has to raise the premium. Now the next-healthiest drop out. And so on — until only the sickest remain and the market collapses.

Why a mandate helps

If everyone has to participate, the insurer can price using a representative risk mix. The healthy subsidize the sick — but no one pays an astronomical premium. That is exactly the logic of the Obamacare individual mandate (2010), which was later (2017) weakened.

Alternative solutions

Instead of a mandate: a public insurance program like Medicare (in the US for ages 65+) — directly redistributive, financed by payroll taxes. Or risk equalization between insurers (the German model).

3 A steel plant emits sulfur dioxide and causes asthma in nearby residents. Describe the economic efficiency loss graphically (in words). Which three policy instruments could address the problem? Compare them.

Efficiency analysis

The steel plant only accounts for its private marginal costs (materials, labor, capital) — not the costs to residents. The social marginal cost (MSC) is higher than the private (MPC) by the value of the externality. Under a free market, the plant produces quantity Qprivate, where P = MPC. The efficient quantity Qopt would be where P = MSC, i.e. less. The difference × the marginal externality is the deadweight loss.

Three policy instruments

1. Pigou tax: a tax equal to the marginal externality on each ton of SO₂. Brings MPC + tax = MSC. The market then automatically reaches the efficient quantity. Pros: market-based, generates revenue. Cons: the size of the externality is hard to estimate.

2. Quantity regulation (emission limit): directly prescribe that the plant may emit no more than X tons. Pros: predictable. Cons: inefficient when abatement costs differ — plants with low abatement costs could reduce more than others.

3. Cap-and-trade (emissions trading): the total emission quantity is capped and permits are traded. Combines quantity certainty with market-based efficiency — the plants with the lowest abatement costs reduce the most. Examples: EU-ETS, the US SO₂ program of the 1990s.

Which instrument when?

Pigou tax and cap-and-trade are economically equivalent under certainty. Under uncertainty about costs/benefits, economists differ (Weitzman 1974). Quantity regulation is usually the second-best choice — widespread, but inflexible.

4 What are the two defining properties of a pure public good? Give three examples of pure public goods and three "impure" mixed cases.

Two properties

1. Non-rival: my consumption does not reduce anyone else's consumption.

2. Non-excludable: no one can be excluded from consumption at a reasonable cost.

Pure public goods

  • National defense — the protection I enjoy doesn't reduce my neighbor's; excluding anyone is impossible.
  • Lighthouse light — every ship can see it; excluding individual ships is technically absurd.
  • Knowledge / basic research — my reading a paper doesn't reduce anyone else's; without patents it is hard to exclude.

Mixed cases

  • Toll bridge (non-rival but excludable): as long as it is below capacity, one extra car costs nothing; but a toll booth can be set up.
  • City parks (excludable if fenced, otherwise not): can be a private club or open to the public.
  • Education (rival in class size, but with positive externalities): one more student costs resources, but society benefits from education beyond the student themselves.
5 What is the Coase Theorem? Under which conditions does it hold? Why does it not provide a solution to the real-world climate change problem?

Coase Theorem (1960)

If property rights are clearly defined and transaction costs are zero, then private negotiations lead to the efficient outcome — regardless of who is assigned the right. Example: factory vs. residents. Whether the factory has the right to emit or the residents have the right to clean air — they will negotiate their way to the efficient level of emissions either way.

Assumptions

  1. Clear and enforceable property rights.
  2. Negligible transaction costs.
  3. Complete information.
  4. No income effects (no large wealth transfers).

Why it fails for climate change

  • Massive transaction costs: negotiating between 8 billion people is absurdly impossible.
  • Unclear property rights: who "owns" the atmosphere?
  • Free-rider problem: every individual has an incentive not to pay.
  • Intertemporal dimension: future generations cannot negotiate.
  • Distribution: who bears the main burden (poor countries) vs. who caused it (rich countries)?

Therefore: global climate policy needs coordinated government/international mechanisms — cap-and-trade systems, international agreements, Pigou taxes.

6 Explain the difference between adverse selection and moral hazard. How do insurers try to mitigate both problems?

Adverse selection

Asymmetric information before contract signing. The insured knows more about their risk (e.g. pre-existing conditions) than the insurer. Result: bad risks get insurance; good risks are discouraged.

Moral hazard

Changed behavior after contract signing. The insured acts less carefully because they are insured. Examples: full coverage leads to more careless driving; insured patients consume more doctor visits.

Mitigation strategies

Against adverse selection:

  • Risk assessment at contract signing (health questions, driving data).
  • Waiting periods for pre-existing conditions.
  • Mandatory insurance (mandate) — all risks in the pool.
  • Group policies via employers (self-selection eliminated).

Against moral hazard:

  • Deductibles / co-payments.
  • Bonuses for claim-free years.
  • Monitoring (telematics boxes in cars).
  • Limits on covered services.

But: these measures never fully solve the problems — which is why some insurance markets do not exist at all (unemployment, long-term care).

7 "When a market fails, the government should step in." Discuss critically in 200 words.

Model answer:

The statement is a dangerous oversimplification. Market failure is a necessary but not sufficient condition for meaningful government action. Three objections.

First: governments fail too. Bureaucratic inertia, rent-seeking by lobbyists, asymmetric information between voters and politicians — all of these make government intervention itself a source of inefficiency. Stiglitz calls this "government failure". The right question is therefore not "is the market failing?" but "can the government do better?".

Second: the size of the market failure has to be taken into account. A small efficiency loss from a small externality may not be worth intervening over — the administrative costs of regulation could exceed the efficiency gain.

Third: there are usually several policy instruments available. An externality can be addressed via a tax (Pigou), cap-and-trade, regulation, or outright bans. Which instrument is optimal depends on information, abatement costs, and distributional effects.

Stiglitz's real point is therefore more subtle: market failure opens the possibility of meaningful government action. Whether and how the government should intervene must be checked in each specific case against the alternative — empirically, not ideologically.