Government Failure and Policy Evaluation
Comparing imperfect institutions
Market failure creates a possible case for collective action, not a proof that a particular policy improves welfare. Governments also act with limited information, conflicting objectives, administrative constraints, and political incentives. Serious analysis compares realistic alternatives rather than an imperfect market with an idealized regulator.
Government failure occurs when public intervention produces avoidable social costs or fails to achieve its stated objective because of institutional incentives or constraints. The relevant benchmark includes a feasible status quo, private ordering, local governance, and alternative public designs. Doing nothing also has costs and distributional consequences.
Sources of failure
Information is dispersed: agencies may not know marginal damages, compliance costs, or local conditions. Implementation is costly: eligibility checks, inspections, appeals, procurement, and data systems consume resources. Rules can be captured by organized interests or protect incumbents from competition.
Officials and contractors respond to measured targets. A school judged only by pass rates may exclude difficult students or teach narrowly to the test. Budget cycles can favor visible short-run projects over maintenance and prevention. Overlapping agencies may create gaps, duplication, or unclear accountability.
These risks are not arguments for automatic inaction. They are costs and incentive problems to include in policy design.
A structured evaluation
Begin with a measurable problem and a causal theory of change. State the affected population, baseline, market or institutional failure, proposed mechanism, and expected responses. List alternatives, including a smaller pilot and the status quo.
Cost-benefit analysis compares incremental social benefits and costs over time:
Transfers such as tax payments are not automatically resource costs, though collection, avoidance, and distribution matter. Report ranges and sensitivity to uncertain values rather than hiding uncertainty in one number.
Policy evaluation workbench
Unlock the complete interactive lab, deeper explanations, and curated reference answers.
Worked appraisal
A congestion charge is forecast to create 6 million, administration 9 million, and annualized equipment costs 35 million is mainly a transfer before collection cost. Forecast net social benefit is therefore 14 million per year.
That positive estimate is not the end of evaluation. Analysts should test lower time values, higher diversion, behavioral adaptation, and alternative transit investment. They must also show who pays, who benefits, and how rebates or revenue use change access.
Evidence and counterfactuals
The counterfactual is what would have happened without the policy. A before-after comparison can confuse policy effects with economic trends, weather, construction, or simultaneous reforms. Randomized pilots, phased rollout, matched comparisons, and interrupted time series can strengthen causal inference when used appropriately.
Evidence quality depends on relevance as well as method. A successful pilot may not scale if staffing, participant mix, or compliance changes. Qualitative evidence can reveal mechanisms and burdens that an average outcome misses.
Metrics, gaming, and distribution
Use a compact set of complementary indicators: final outcomes, full costs, service quality, compliance, equity, and unintended effects. Define denominators and data ownership before launch. Audit intake and exclusions so providers cannot improve scores by selecting easy cases.
Efficiency and equity should be reported separately before making an explicit social judgment. An aggregate gain can coexist with concentrated losses, liquidity barriers, or violations of rights. Compensation may improve acceptability, but promised transfers must be administratively and politically credible.
Adaptive policy and accountability
Pilot programs, staged expansion, sunset clauses, and scheduled reviews preserve learning and reversibility. A useful sunset rule specifies a date, evidence owner, decision authority, and thresholds for continuing, changing, or ending policy. Without precommitted criteria, review can become a ritual that entrenches the program.
Independent audits, transparent procurement, appeal rights, conflict disclosures, and public data reduce some implementation risks. They also cost money and can delay action, so oversight should be proportional to stakes and reversibility.
KNOWLEDGE QUIZ
Appraise a congestion policy
A city proposes a downtown congestion charge after observing slow traffic. Build an appraisal that compares doing nothing, pricing, transit investment, and a time-limited pilot.
- Easy: Calculate annual net benefit from stated time savings, operating cost, and traveler payments, treating charge revenue as a transfer before collection cost.
- Medium: Identify a credible counterfactual and define outcome, cost, equity, leakage, and compliance metrics before launch.
- Hard: Compare alternatives under uncertainty, assign implementation owners and risks, and write precommitted thresholds for expansion, modification, or sunset.
Forecast, millions per year
Traveler time and reliability benefit: 48
Extra transit crowding cost: 6
System administration and enforcement: 11
Charge revenue: 35
Estimated diversion harm in nearby districts: 9
Capital cost annualized: 8The chapter began with the competitive efficiency benchmark and ended with comparative institutional judgment. Good public policy diagnoses a specific failure, chooses a mechanism that can work under real constraints, measures results, and remains willing to revise.