Skip to content
Back to search results

Cost-Benefit Analysis (CBA)

Introduced by
Boardman, Greenberg, and colleagues (welfare economics)
Year
1970s
Type
Method
Family
economic evaluation framework
Font size

What does it do?

Cost-benefit analysis compares the total monetary value of benefits produced by a program against its total monetary costs, reporting net present value (NPV) or benefit-cost ratio (BCR).

Rooted in welfare economics and used extensively in public policy (transportation, environmental, education, health), CBA answers the question: 'Is this program worth doing from a societal perspective?' Unlike cost-effectiveness analysis, CBA monetizes both costs and benefits, enabling comparison across disparate program types.

The intuition behind the method

Imagine a government must choose between three projects: (A) Build a highway; (B) Fund a job training program; (C) Install water treatment. Each has different kinds of benefits and costs. CBA converts everything to dollars: highway benefits = reduced commute time (valued at wage rate per hour saved) + lower accident rates (valued at willingness-to-pay for safety); job training benefits = increased earnings (discounted lifetime gain) + reduced unemployment benefits (savings); water treatment benefits = improved health (reduced healthcare costs, lost productivity averted, willingness-to-pay for clean water). All are measured in dollars. Then: if highway NPV = $100 million, job training NPV = $50 million, water treatment NPV = $80 million, the highway is most efficient by this metric (though equity, distribution, and non-monetizable factors may alter the decision). CBA is powerful because it forces explicit valuation and enables cross-sector comparison; it is controversial because assigning dollar values to health, safety, and environment is ethically fraught.

History

CBA emerged from welfare economics in the 1950s–60s, developed by Kaldor, Hicks, and later formalized by Layard, Glaister, and others. U.S. adoption: water projects (Bureau of Reclamation) pioneered CBA in the 1960s–70s. By 1980s, CBA was standard for environmental impact assessment. Drawbacks became apparent in the 1990s: environmental valuation (e.g., assigning a price to biodiversity) and health valuation (value of life) sparked ethical criticism. Modern practice acknowledges CBA limitations and combines it with other frameworks (equity analysis, rights-based approaches, sustainability thresholds). Drummond et al. (2015) distinguished CEA (health sector standard) from CBA (cross-sector public policy standard).

When to use it

When evaluating large public investments or programs where benefits and costs span multiple sectors or years. CBA is standard for infrastructure (dams, highways, rail), environmental projects (pollution control, climate adaptation), education (scholarships, vocational training), and large health programs (vaccination campaigns, water/sanitation). Use when: costs and benefits are significant and distributed over time; comparability across programs is desired; decision-makers need a single summary metric (NPV); quantifiable data on costs and benefits are available. Not appropriate for: small, localized interventions with unclear long-term effects; settings where key benefits are non-monetizable (dignity, rights, culture); or when equity is paramount and CBA's utilitarian framework conflicts with equity values.

Limitations and considerations

  • Assigns monetary values to intrinsically non-market goods (health, environment, equity, human dignity), which many regard as ethically inappropriate or impossible.
  • Outcome is sensitive to discount rate choice: a 1% change in rate can flip sign of NPV for long-horizon projects (e.g., climate adaptation). Choice is normative, not empirical.
  • Valuation methods (contingent valuation, shadow pricing) are imperfect and context-dependent; two studies of same program may yield vastly different benefit estimates.
  • BCR and NPV can give conflicting rankings when project scales differ; choosing among projects is not straightforward if one is small/low-cost (high BCR, small NPV) vs large/high-cost (lower BCR, higher NPV).
  • Distributional issues invisible in single NPV metric: a program with large NPV but regressive distribution (benefits to rich, costs to poor) is not flagged as inequitable by CBA alone.

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Enables cross-sector comparison: a highway, vaccination program, and school can be compared using common metrics (NPV, BCR), facilitating system-level allocation.
  • Transparent and replicable: all assumptions (discount rate, benefit valuation, cost estimates) are explicit, enabling scrutiny and debate.
  • Forces discipline: requires decision-makers to articulate and quantify what benefits are expected, discouraging vague claims.
  • Incorporates time value of money: discounting ensures future costs and benefits are properly weighted, avoiding overestimation of long-distant impacts.

Applications

  • Infrastructure CBA: transportation authority conducts CBA of proposed metro line—capital cost $2 billion, annual operating costs $300M, benefits = time savings + congestion reduction + air quality improvement, monetized; NPV and BCR guide go/no-go decision.
  • Environmental CBA: air quality improvement rule: cost to industry = $500M; benefits = avoided respiratory illness + life expectancy gain (monetized using VSL), valued at $2 billion; net benefit = $1.5 billion, justifying regulation.
  • Health program CBA: cervical cancer screening program: cost = $200M; benefits = lives saved + DALYs averted (converted to dollars at $30,000/DALY), caregiving time saved; NPV used to compare against other health programs.
  • Education CBA: college tuition subsidy: cost = $100M; benefit = increased lifetime earnings of graduates (PV $500M) + positive externalities (public health, civic engagement, $100M); NPV = $500M > 0, recommend subsidy.
  • Regulatory impact analysis: new occupational safety standard; cost = $1 billion to industry; benefit = injury/illness reduction (lives saved, pain averted) = $3 billion; CBA supports regulation.

How it works

How it works

1

Step 1: Identify all program costs (direct: salaries, equipment, materials; indirect: capital depreciation, opportunity costs). Step 2: Identify all program benefits, including: direct benefits (user fees paid, output sold); indirect benefits (reduced crime, improved health, environmental gains, time savings); intangible benefits (equity, empowerment, peace of mind). Step 3: Monetize each benefit. Methods: (a) Market price: if a good is traded (e.g., income, fuel savings), use observed prices. (b) Contingent valuation: survey respondents 'What would you pay to reduce air pollution by X%?' Produces willingness-to-pay (WTP). (c) Revealed preference: infer valuation from actual behavior (e.g., people pay $1,000 more for houses in clean-air neighborhoods, implying air quality value). (d) Shadow pricing: use proxy markets; e.g., wage differential between hazardous and safe jobs infers value of occupational safety. (e) Health valuation: statistical value of life (VSL) ranges $5–$15 million in high-income countries (derived from occupational risk studies, environmental policy valuations); QALYs or DALYs converted to dollars using implicit cost-effectiveness threshold (e.g., £20,000/QALY × QALYs gained = monetary benefit). Step 4: Specify time horizon (e.g., 20-year infrastructure life, 30-year career span). Step 5: Choose discount rate (standard 3–5%; sensitivity test 0%, 7%).

Formula
Enumerate all program costs and benefits, both direct (medical, infrastructure) and indirect (productivity, societal). Assign monetary values to each: health benefits (e.g., via stated preference surveys, revealed preference from market data, or shadow pricing). Identify time horizon (years of program operation). Determine discount rate (3–5% standard).
2

Net Present Value (NPV) = Σ[Benefits_year_t − Costs_year_t] / (1 + r)^t for t = 0 to T, where r = discount rate. Example: Water treatment program costs $10 million upfront; produces $2 million/year in health benefits (reduced waterborne illness, avoided treatment costs) and $1 million/year in environmental benefits (wetland restoration value) for 20 years. Discount rate = 3%. NPV = −$10M + Σ($3M / 1.03^t) for t=1 to 20 = −$10M + $3M × 14.88 = −$10M + $44.6M = $34.6M. NPV > 0, so project is justified. Benefit-Cost Ratio (BCR) = Total PV of Benefits / Total PV of Costs = $44.6M / $10M = 4.46 (for every $1 spent, $4.46 in benefits returned). Internal Rate of Return (IRR) = discount rate at which NPV = 0; if IRR > society's opportunity cost of capital (e.g., 5%), project is attractive. Profitability Index (PI) = NPV / Initial Investment = $34.6M / $10M = 3.46 (return per dollar invested).

Formula
Calculate Net Present Value (NPV) = Σ(Benefits_t / (1+r)^t) − Σ(Costs_t / (1+r)^t), where r = discount rate, t = year. Alternatively, Benefit-Cost Ratio (BCR) = Total PV of Benefits / Total PV of Costs. Break-even analysis identifies discount rate at which NPV = 0 (Internal Rate of Return, IRR).
3

Decision rule: If NPV > 0, program generates net benefits; recommend implementation. If NPV < 0, program destroys value; do not implement. If comparing mutually exclusive programs (choose one), select highest NPV (if scales are similar) or highest BCR (if scales differ). IRR: if IRR exceeds society's discount rate or cost of capital, project is attractive. Break-even analysis: sensitivity testing—at what discount rate does NPV = 0? If break-even rate is 10% and true discount rate is 3%, project is robust. Distributional analysis: critical for equity. Example: highway project may have positive NPV but benefits accrue to wealthy commuters while poor residents bear pollution costs; CBA alone is silent on fairness (separate equity analysis required). For health programs, monetizing health outcomes is controversial. If valuing a QALY at £20,000 (implicit from NICE threshold), a program that gains 1,000 QALYs is assigned £20 million in health benefits—but this makes health trade-offs explicit and sometimes uncomfortable. Policy-makers often use CBA alongside other frameworks (equity, rights, sustainability) to inform decisions.

Formula
If NPV > 0 or BCR > 1, program is economically justified (benefits exceed costs). Compare NPV and BCR across alternative programs to rank by efficiency. Sensitivity analysis: vary discount rate, benefit valuation, cost estimates to test robustness. Report distributional analysis: who bears costs, who receives benefits? Equity assessment ensures benefits are not concentrated in wealthy groups if intent is to help disadvantaged.

Common pitfalls

  • Using arbitrary monetary values for health/environment: assigning WTP without empirical justification (e.g., stating 'life is worth $10 million' with no survey data). Always cite source and sensitivity-test valuation.
  • Mixing real and nominal values: costs/benefits in different years must be adjusted for inflation to real (constant-year) dollars before discounting; failing to do so overstates real value.
  • Ignoring externalities: valuing direct costs/benefits while overlooking spillovers. Example: job training program cuts unemployment; analysis counts new wages as benefit but misses childcare demand increase or crime reduction (externalities). Comprehensive CBA captures these.
  • Double-counting: if benefits include both avoided treatment costs AND health utility gained, both are being counted. Avoid double-counting of benefits or costs across categories.
  • Forgetting intangibles: some impacts (equity, sustainability, cultural preservation) resist monetization. CBA should explicitly note these and separate them from NPV calculation, allowing qualitative consideration.

Frequently asked questions

How is the statistical value of life (VSL) determined?

VSL is derived from labor economics: comparing wages of workers in hazardous vs safe jobs (wage premium for risk). If workers accept a job with 1 in 10,000 annual mortality risk for a $5,000 wage premium, implied VSL = $5,000 × 10,000 = $50 million. Reviews consolidate estimates: U.S. EPA and OSHA typically use $10–$15 million per life (2020 dollars). Criticism: VSL is context-dependent (poor workers may accept hazard at lower premium than rich); geographic variation; not everyone's life valued equally.

Why do economists sometimes get very different CBA results for the same program?

Differences in: (1) Discount rate choice (3% vs 7% → different NPV); (2) Time horizon (20 years vs 50 years → downstream benefits weighted differently); (3) Benefit valuation (WTP survey yields $X, revealed preference yields $2X); (4) Cost data (one study underestimated, another overestimated); (5) Included externalities (narrow scope vs broad scope). Standard practice: transparent reporting of all assumptions and sensitivity analysis over plausible ranges, showing which inputs drive results.

How are intangible benefits (equity, cultural value) incorporated into CBA?

They are not, typically. CBA monetizes what can be valued; intangibles are identified separately in qualitative sections ('This program also improves community social cohesion and cultural preservation, which are not monetized but valued by the community'). Some advanced frameworks assign weights or shadow prices to intangibles based on stakeholder surveys, but this is less standard.

What is the 'social discount rate' and how is it chosen?

Social discount rate reflects society's time preference and opportunity cost of public funds. U.S. standard: 3% real (Office of Management and Budget). UK: 3.5%. Developing countries: sometimes 5–10% (scarcer capital, higher opportunity cost). Debate: some argue lower rates (0–2%) for intergenerational projects (climate, infrastructure benefiting future generations); others defend 3% as market-based. Guidance: 3% is default; sensitivity-test 0%, 5%+ to show robustness.

If a CBA shows negative NPV, should the program definitely be rejected?

Not necessarily. NPV is one input to policy decisions. If a program has negative NPV but strong equity benefits (helps disadvantaged), redistribution goals, or non-monetizable benefits (dignity, rights), decision-makers may choose to proceed. Conversely, positive NPV does not guarantee implementation if equity or sustainability concerns exist. CBA is a tool to inform, not dictate, decisions.

Can CBA be used to evaluate health policies in low-income countries?

Yes, but with adaptations. CBA in low-income settings must (1) use local prices (not high-income country prices); (2) value health using country-specific VSL or willingness-to-pay (often much lower than U.S./Europe due to lower income); (3) use locally calibrated discount rate (may be higher, reflecting capital scarcity); (4) account for equity: program benefits to poorest groups may warrant higher weighting. WHO and World Bank have developed guidelines for CBA in developing countries.

Related methods

Sources

  1. Boardman, A. E., Greenberg, D. H., Vining, A. R., & Weimer, D. L. (2018). Cost-Benefit Analysis: Concepts and Practice (5th ed.). Cambridge: Cambridge University Press.

    Open publication reference
    Copied.
  2. Layard, R., & Glaister, S. (Eds.). (2003). Cost-Benefit Analysis (2nd ed.). Cambridge: Cambridge University Press.

    Open publication reference
    Copied.
  3. Drummond, M. F., Sculpher, M. J., Claxton, K., Stoddart, G. L., & Torrance, G. W. (2015). Methods for the Economic Evaluation of Health Care Programmes (4th ed.). Oxford: Oxford University Press.

    Open publication reference
    Copied.