public-policy

Public Policy

Public policy economics asks when government intervention improves on market outcomes and when it makes things worse. The field sits at the intersection of positive analysis (what will happen if a policy is enacted) and normative judgment (should it be enacted). This skill focuses on the economic toolkit for policy analysis: taxation, regulation, externality correction, and public goods provision, with attention to both market failure and government failure.

Agent affinity: keynes (fiscal policy, stabilization), hayek (limits of central planning, knowledge problem), ostrom (commons governance, polycentric solutions), varian (policy design, mechanism design)

Concept IDs: econ-fiscal-policy, econ-market-failures, econ-trade-offs, econ-marginal-thinking, econ-opportunity-cost

The Public Policy Toolbox at a Glance

| # | Topic | Core question | Key framework | |---|---|---|---| | 1 | Market failure | When do markets produce bad outcomes? | Externalities, public goods, asymmetric info, market power | | 2 | Taxation | How do taxes affect behavior and welfare? | Incidence analysis, deadweight loss, optimal tax theory | | 3 | Regulation | When and how should government regulate? | Cost-benefit analysis, regulatory capture | | 4 | Externalities | How do we correct unpriced costs and benefits? | Pigouvian taxes, cap-and-trade, Coase theorem | | 5 | Public goods | Why do markets underprovide some goods? | Non-rivalry, non-excludability, mechanism design | | 6 | Government failure | When does intervention make things worse? | Rent-seeking, regulatory capture, knowledge problem | | 7 | Policy evaluation | How do we measure whether a policy worked? | Causal inference, natural experiments, RCTs |

Topic 1 -- Market Failure

The benchmark. The First Welfare Theorem states that competitive markets produce Pareto-efficient outcomes under standard assumptions: complete markets, perfect competition, no externalities, no public goods, full information. Market failure is the departure from these assumptions.

Four categories. Externalities (unpriced spillovers -- pollution, congestion, network effects). Public goods (non-rival, non-excludable -- national defense, basic research). Asymmetric information (one party knows more -- insurance markets, used cars). Market power (firms with pricing power restrict output below the efficient level).

The policy question. Market failure is necessary but not sufficient for intervention. The cure must be better than the disease. Every intervention has costs: administration, enforcement, unintended behavioral responses, political distortion. The comparison is always between imperfect markets and imperfect government.

Worked market failure example. Consider traffic congestion. Each driver's decision to enter the road imposes a cost on every other driver (slower speed, longer commute), but the driver does not bear this cost -- it is an externality. The result: more driving than is socially optimal. A congestion tax (London, Singapore, Stockholm) prices the externality, reducing traffic to a level where the marginal social cost of an additional trip equals the marginal social benefit. Stockholm's congestion tax reduced traffic by 20% and improved air quality measurably. But the tax has distributional effects: low-income commuters who cannot avoid driving bear a disproportionate burden. The policy response (using revenue for public transit improvements) addresses this, but only if the revenue is actually spent that way -- which requires political institutions that are themselves subject to capture and distortion.

The Demsetz critique. Demsetz (1969) warned against the "nirvana fallacy": comparing imperfect market outcomes to ideal government outcomes. The correct comparison is between imperfect markets and imperfect government. A market failure that produces 90% of the efficient outcome may be preferable to a government intervention that produces 85% of the efficient outcome after accounting for administrative costs, unintended consequences, and political distortions.

Topic 2 -- Taxation

Tax incidence. Who bears a tax depends not on who writes the check but on relative elasticities. A payroll tax nominally split 50-50 between employers and employees will fall primarily on whichever side is more inelastic. If labor supply is relatively inelastic (workers can't easily stop working), workers bear most of the burden regardless of the legal assignment.

Deadweight loss. Taxes create a wedge between the price buyers pay and the price sellers receive. The resulting reduction in quantity traded produces deadweight loss -- surplus that vanishes rather than being transferred. Deadweight loss is proportional to the square of the tax rate, which is why a single high rate is more distortionary than multiple low rates raising the same revenue.

The Ramsey rule. Optimal taxation minimizes deadweight loss for a given revenue target. The Ramsey rule says to tax goods in inverse proportion to their elasticity of demand -- tax inelastic goods more. This minimizes behavioral distortion but raises equity concerns (necessities like food and medicine tend to be inelastic).

Optimal income taxation. Mirrlees (Nobel 1996) showed that the optimal income tax balances the equity gain from redistribution against the efficiency cost of discouraging work. The optimal marginal rate schedule depends on the distribution of abilities, the labor supply elasticity, and the social welfare function. There is no universal answer, but the framework clarifies the trade-offs.

Laffer curve. Beyond some tax rate, further increases reduce revenue because the behavioral response (reduced work, increased evasion) dominates the higher rate. The revenue-maximizing rate is an empirical question that depends on the elasticity of taxable income. Estimates for the US top rate range from 50% to 80%.

Worked tax incidence example. A $1 per gallon gasoline tax is imposed. Gasoline demand is highly inelastic in the short run (E_d approximately -0.2) because people cannot quickly change commuting patterns or replace cars. Gasoline supply is moderately elastic (E_s approximately 0.5). The burden falls primarily on consumers: they bear roughly 0.5/(0.5 + 0.2) = 71% of the tax, paying about $0.71 more per gallon, while producers absorb about $0.29. In the long run, demand becomes more elastic (people buy fuel-efficient cars, move closer to work, use transit), so the consumer burden shrinks. This illustrates why short-run and long-run incidence can differ substantially and why time horizon matters for policy design.

Tax expenditures. Deductions, credits, and exemptions in the tax code are economically equivalent to government spending on the favored activity. The mortgage interest deduction is equivalent to a housing subsidy. The charitable contribution deduction is equivalent to a subsidy for philanthropy. Calling these "tax cuts" rather than "spending" obscures their cost and shields them from the budget scrutiny that direct spending faces. The US tax expenditure budget exceeds $1.5 trillion annually -- larger than Medicare and defense combined.

Progressive vs. regressive. A progressive tax takes a larger share of income from high earners (income tax). A regressive tax takes a larger share from low earners (sales tax, payroll tax cap). The overall progressivity of the tax system depends on the mix. Many countries have progressive income taxes but regressive consumption and payroll taxes, and the net effect is less progressive than the income tax alone suggests.

Topic 3 -- Regulation

Cost-benefit analysis. The standard tool for evaluating regulations. Enumerate all costs and benefits, monetize them, discount to present value, and compare. The regulation passes if benefits exceed costs. In practice, CBA is difficult: some costs and benefits are hard to monetize (the value of a statistical life, ec