Social Cost-Benefit and Cost-Effectiveness Analysis

Why this matters in public value

Every year, budget committees, boards, and legislatures choose between programmes that cannot be compared on price alone, because the thing each programme buys is different in kind: a flood defence scheme, a smoking-cessation service, and an early-years literacy programme all draw on the same finite pool of public money, yet none of their outputs shares a unit. Social cost-benefit analysis (CBA) and cost-effectiveness analysis (CEA) exist to answer the question a plain financial statement cannot: not just what does this cost, but is it worth it, compared with what else the same money could do.

The discipline matters more, not less, when a decision is contested. A minister proposing a new motorway, a council weighing a subsidised bus network against a rail link, or a health ministry choosing which vaccines to fund cannot settle the argument by political preference alone without eventually facing whether the choice created more public value than the alternative foregone. Formal CBA and CEA give that question a shared, auditable method — not a mechanical answer that replaces judgement, but a structured account of assumptions that can be checked, challenged, and revised, in a way that instinct cannot.

The method also carries real limits that a careless user mistakes for precision. Every number in a cost-benefit appraisal rests on assumptions — future prices, the worth of a life-year, how heavily to discount a benefit thirty years off — and each assumption embeds a judgement a spreadsheet can obscure but never eliminate. This topic draws on the outcomes vocabulary of Topic 2.1 (see Topic 2.1 — The Public Value Scorecard and Outcomes Frameworks) to give those judgements names, so a leader can defend or challenge them explicitly rather than let them hide inside a discount rate nobody in the room chose to interrogate.

Core concepts

Cost-benefit analysis converts every material cost and benefit of a proposed action into a common monetary unit, discounts each to a common point in time, and compares the resulting present values to judge whether a programme creates more value than it consumes. Its theoretical foundation is the Kaldor–Hicks efficiency criterion: an action is worth taking if those who gain could, in principle, fully compensate those who lose and still be better off, even if that compensation never happens. This hypothetical-compensation test is what lets CBA add up gains and losses across different people as though a pound or a dollar means the same thing to everyone, an assumption this topic returns to under distributional weighting.

Net present value (NPV) is the single figure a cost-benefit appraisal usually produces: the sum of all future benefits minus all future costs, each converted to today's money using a discount rate. A positive net present value means the programme is worth doing on efficiency grounds alone; comparing the net present values of competing programmes ranks them by how much value each creates per unit of public money committed.

Shadow pricing assigns a monetary value to a good, cost, or benefit with no observable market price — clean air, a life-year, travel time saved, a wetland preserved — because a price either does not exist or is distorted by a tax, subsidy, or monopoly. Analysts derive shadow prices two ways: revealed preference methods infer value from what people actually do, such as hedonic pricing, which extracts the value of a characteristic (quiet, clean air, a good school nearby) from how it changes a marketable good's price, typically a house; stated preference methods ask people directly, most commonly through contingent valuation, a survey technique that asks respondents their willingness to pay for a hypothetical change in a non-market good, such as protecting a stretch of coastline.

Value of a statistical life (VSL) is the shadow price most appraisals reach for when a policy changes the probability, not the certainty, of someone dying — a road-safety measure that reduces annual fatalities by a fraction of a person, for instance. It is derived from what people are observed to accept, or say they would accept, for a small change in mortality risk, then scaled up; it does not price an identified individual's life, a distinction guidance stresses because the two are routinely and wrongly conflated in public debate.

Cost-effectiveness analysis compares programmes using a single non-monetised outcome measure instead of converting every benefit into money — a cost per case prevented, per tonne of emissions avoided, or, most commonly in health economics, a cost per quality-adjusted life year (QALY), a measure combining survival and health-related quality of life into one unit, where a year in full health scores 1. CEA is the right tool, rather than CBA, whenever the outcome resists credible monetisation, or where monetising it would itself be ethically or politically contested enough to swamp the analysis; it ranks options per unit of the chosen outcome rather than delivering a single verdict on whether a programme is worth doing in absolute terms.

Distributional weighting adjusts the basic CBA arithmetic to reflect that a pound or a dollar is worth more, in welfare terms, to a poor household than a rich one — the standard economic assumption of diminishing marginal utility of income. Unweighted CBA treats a £100 benefit to a low-income household and the same benefit to a high-income household as identical, exactly what Kaldor–Hicks efficiency permits and distributional weighting exists to correct; applying it systematically can shift the ranking of programmes that otherwise look equivalent on efficiency alone (see Topic 3.5 — Equity, Fairness, and Distributive Justice).

Social discount rate converts future costs and benefits into present values, and is one of the most consequential, contested numbers in the whole appraisal. A higher rate shrinks the present value of anything decades away, systematically disadvantaging programmes — flood defence, biodiversity protection, climate mitigation — whose costs fall now and benefits fall on future generations, raising a direct question of intergenerational equity this topic returns to under discussion questions and failure modes.

Best practices

1. Fix the counterfactual before fixing the numbers. Every cost and benefit in a CBA or CEA is a comparison against what would have happened anyway, so agree the do-nothing baseline before any figure is estimated, or the appraisal will silently compare the proposal against the wrong alternative.
2. State the discount rate, and its source, on the cover page of every appraisal. A rate chosen without explanation invites the suspicion that it produced a preferred answer; citing the source lets a reader check it rather than take it on trust.
3. Prefer CEA over CBA whenever monetising the outcome would be more contested than the decision itself. Forcing a price onto a good with no defensible shadow price does not make the analysis more rigorous; a well-built cost-effectiveness ratio is often the more honest tool.
4. Publish the shadow prices used, not just the totals they produce. A reader who can see a life-year's specific, sourced value can judge whether it is reasonable for their context; shown only a final net present value, they cannot audit anything.
5. Run a distributional weighting variant alongside the unweighted appraisal, not instead of it. Showing both figures together keeps the efficiency case and the equity case visible side by side, rather than letting one silently substitute for the other (see Topic 3.5 — Equity, Fairness, and Distributive Justice).
6. Test the ranking's sensitivity to the discount rate before committing to a decision. If switching between two defensible rates reverses which programme looks better, that instability is itself a finding the decision-maker needs, not a footnote to bury in an annex.
7. Separate optimism-bias correction from genuine uncertainty. Appraisers systematically underestimate costs and overestimate benefits on novel projects; apply a documented adjustment for that known bias, distinct from modelling genuine uncertainty in the underlying data.
8. Use willingness-to-pay estimates from a context close to your own, and say so. A shadow price estimated in one country's income level does not automatically transfer to another; note the source study's context and flag the gap rather than importing the number silently.
9. Report a cost-effectiveness ratio's plausible range, not a single number. A point estimate of cost per outcome invites false precision; showing the range lets a board see how much confidence the figure deserves.
10. Route genuinely non-monetisable outcomes to a complementary method rather than forcing a shadow price onto them. Where an outcome resists both credible monetisation and a single-unit effectiveness measure, bring in social return on investment or theory-of-change methods instead of abandoning rigour altogether (see Topic 2.3 — Social Return on Investment and Impact Measurement).

Questions to discuss with your team

Which discount rate are we using, and could we defend it in public? The discount rate is not a neutral technical parameter; it encodes a view about how much future generations' welfare counts relative to the present, and a rate borrowed uncritically from another jurisdiction's guidance may not reflect your own fiscal position or values. Ask who chose the rate in use, whether it has been reviewed since adoption, and whether the team could explain to a citizen why that specific number was used rather than a higher or lower one. If nobody can answer beyond 'it is what the template uses,' treat that as a gap to close before the next major appraisal.

Where has our CBA or CEA quietly assumed everyone's pound is worth the same? Standard cost-benefit analysis treats a benefit to a wealthy household and an equal benefit to a poor household as identical, an assumption easy to forget is an assumption once buried inside a spreadsheet. Walk through your most recent major appraisal and check whether its winners and losers are evenly spread across income groups or concentrated at one end. If skewed, discuss whether distributional weighting would change the recommendation, and if it would, ask why that variant was not presented to the decision-maker.

Is this decision better suited to cost-benefit or cost-effectiveness analysis, and did we choose deliberately? Some benefits genuinely resist credible monetisation, and forcing a shadow price onto them can produce a spuriously precise number that erodes trust once challenged. Discuss the specific outcome at stake and whether a defensible monetary value exists for it, or whether a cost-per-outcome measure would better withstand scrutiny. If the method was chosen only because it is the team's default template, say so and consider switching.

What would change our recommendation if we varied our key assumptions within a plausible range? A headline net present value or cost-effectiveness ratio that looks robust can depend entirely on one contestable assumption — a shadow price, a discount rate, a take-up forecast — nobody has stress-tested. Walk through the assumptions most likely to be challenged and discuss what happens to the ranking if each moves to the edge of a defensible range. If the recommendation flips under a plausible variation, the decision-maker needs to see that instability directly, not a single confident-looking figure.

Whose welfare have we left out of this appraisal, and why? A cost-benefit analysis is only as complete as the stakeholders it accounts for, and costs or benefits falling on a group with weak voice — future generations, non-citizens, an ecosystem with no market price — are commonly left out through the appraisal's default boundary rather than through malice. Discuss who is inside that boundary and who is outside it, and whether the boundary matches who the decision actually affects. Where a group has been excluded for practical reasons, say so in the write-up rather than letting the omission pass silently.

How will we know, after the fact, whether this appraisal's assumptions were right? Too many appraisals are produced once, at the business-case stage, and never revisited once the money is spent, so a systematically wrong shadow price or discount rate can go uncorrected for years. Discuss what would need to be measured after implementation to check the appraisal's key assumptions, and who is accountable for running that check. If no ex-post evaluation is planned, treat the appraisal as provisional, not settled (see Topic 2.4 — Public Sector Econometrics and Programme Evaluation).

In practice: a public value example

Meridian County, a fictional mid-sized regional authority used here purely to illustrate the topic's concepts, faced a choice between two flood-resilience proposals competing for the same £38 million capital allocation. The first, a hard-engineering scheme of raised river walls and a pumping station, protected roughly 4,200 households from a one-in-100-year flood event within four years. The second, a programme of upstream wetland restoration, protected fewer households directly and more slowly but reduced flood peaks across a much wider catchment, sequestered carbon, and created habitat value that would grow over decades.

The appraisal team began with cost-benefit analysis for elements with defensible shadow prices: avoided property damage from insurance-claim data, avoided disruption costs from published travel-time and business-interruption values, and avoided risk to life from a nationally published value of a statistical life. On these categories alone, the hard-engineering scheme showed a higher net present value, because its benefits arrived sooner and the discount rate shrank the wetland scheme's slower-building benefits more heavily.

Two further steps changed the picture. The team ran a distributional weighting variant: the wetland scheme disproportionately protected the county's lowest-income postcodes, while the hard-engineering scheme mainly protected a higher-income riverside district, and weighting benefits by household income narrowed the gap substantially. The team also judged that the wetland scheme's carbon and habitat value resisted a single defensible shadow price, so rather than force one, they reported it as a cost-effectiveness supplement — cost per hectare restored, cost per tonne of carbon sequestered — and flagged the remainder for a social-return-on-investment exercise already under way (see Topic 2.3 — Social Return on Investment and Impact Measurement). Presented with the efficiency, distributional, and non-monetised evidence together, the council chose a phased approach: the wetland scheme first, with a smaller hard-engineering element retained for the highest-risk streets it would not protect within the funding period.

Four sector lenses

Local government

Local authorities run appraisals under tighter capital constraints and thinner in-house economics capacity than national government, so discount rates and shadow prices are more often inherited wholesale from national guidance than derived locally. This makes distributional weighting especially important at local scale, because a single scheme's winners and losers usually sit within one authority's boundary and are more visible to councillors than a national programme's diffuse effects, making an unweighted appraisal harder to defend publicly.

National government

National governments typically hold the authoritative discount-rate and shadow-price guidance other tiers and sectors borrow, published through a central finance ministry's appraisal handbook, and carry the greatest responsibility for getting it right, since it is applied across the widest range of programme types and the largest capital sums. National appraisal units also carry the heaviest burden of ex-post evaluation, since a biased shadow price embedded in national guidance propagates into every subordinate appraisal that cites it.

Social sector and nonprofit

Nonprofits and non-governmental organisations rarely have the capacity to run a full monetised cost-benefit analysis, and the outcomes they pursue — dignity, empowerment, community cohesion — more often resist credible shadow pricing altogether, making cost-effectiveness analysis and complementary impact-measurement methods the natural default (see Topic 2.3 — Social Return on Investment and Impact Measurement). Where a funder requires a formal CBA, the sector's distinctive risk is a shadow price borrowed from an unrelated context and applied uncritically because no alternative is readily available.

Multilateral and international

Multilateral development institutions apply cost-benefit and cost-effectiveness analysis across vastly different income levels within a single portfolio, which makes shadow pricing — particularly the value of a statistical life and the discount rate — far more consequential and contested than in a single-country appraisal, since a shadow price calibrated to a high-income country's willingness to pay can systematically undervalue benefits to lower-income populations if transferred uncritically. Cross-country appraisal also puts intergenerational equity at the centre of the discount-rate debate, since decisions made by an international body affect generations across jurisdictions that had no seat in setting the rate.

Common failure modes

The borrowed discount rate. A team facing a major appraisal adopts whatever discount rate a template or a previous project used, without checking whether it is still current or fits the programme's time horizon. The cost is a ranking that looks authoritative but rests on a number nobody chose deliberately, and that collapses under the first serious challenge.

Monetising what should have been counted, not priced. Under pressure to produce a single net present value, an analyst forces a shadow price onto an outcome — dignity, cultural value, an ecosystem's existence value — with no defensible market analogue, producing a number precise enough to look rigorous and weak enough to collapse the whole appraisal's credibility once challenged.

Unweighted efficiency mistaken for the whole answer. A programme with a strongly positive net present value is approved without anyone asking who gains and who loses, because Kaldor–Hicks efficiency does not require that question to be asked. The cost falls on whichever group's losses were real but never counted, and surfaces later as a legitimacy problem rather than an appraisal problem.

The appraisal that is never checked against outturn. A business case's shadow prices and demand forecasts are treated as settled once funding is approved, and nobody returns to compare them against what actually happened. A biased assumption embedded in one appraisal is then reused, uncorrected, in the next.

Cost-effectiveness ratios compared across incompatible outcome measures. A cost-per-outcome figure from one programme is set directly against another whose outcome measure is defined differently, producing a ranking that looks like a fair comparison but is not, and a resourcing decision is then made on a comparison that was never valid.

Maturity model

Checklist

• Document the do-nothing counterfactual before estimating any cost or benefit.
• State the discount rate used, its source, and the date last reviewed.
• Confirm whether CBA or CEA fits this decision, and record why.
• Publish the specific shadow prices used, not only the resulting totals.
• Produce a distributional weighting variant alongside the unweighted result.
• Run a sensitivity analysis on the discount rate and largest shadow prices, and report the range.
• Apply a documented optimism-bias correction to cost and benefit forecasts.
• Check that any borrowed willingness-to-pay or value-of-statistical-life figure fits this context.
• Report a cost-effectiveness ratio with a plausible range, not a single point estimate.
• Identify who is inside and outside the appraisal's boundary, and disclose any material exclusion.
• Schedule an ex-post evaluation for any programme whose case relied on a contested assumption.

Key sources

The cost-benefit and cost-effectiveness framework here follows the welfare-economics tradition running from Kaldor and Hicks's compensation criteria to modern applied practice, most comprehensively set out in Boardman, Greenberg, Vining, and Weimer's Cost-Benefit Analysis: Concepts and Practice. National and multilateral appraisal guidance grounds the discount-rate and shadow-pricing discussion: HM Treasury's Green Book, the US Office of Management and Budget's Circular A-4, and the World Bank's guidance on discounting. The distributional-weighting and intergenerational-equity discussion draws on the OECD's applied environmental cost-benefit work and the Stern Review's use of a low discount rate for climate policy, one of the most publicly contested applications of this topic's method.

References

1. Cost–benefit analysis — Wikipedia — https://en.wikipedia.org/wiki/Cost%E2%80%93benefit_analysis
2. Cost-effectiveness analysis — Wikipedia — https://en.wikipedia.org/wiki/Cost-effectiveness_analysis
3. Shadow price — Wikipedia — https://en.wikipedia.org/wiki/Shadow_price
4. Social discount rate — Wikipedia — https://en.wikipedia.org/wiki/Social_discount_rate
5. Willingness to pay — Wikipedia — https://en.wikipedia.org/wiki/Willingness_to_pay
6. Value of a statistical life — Wikipedia — https://en.wikipedia.org/wiki/Value_of_a_statistical_life
7. Contingent valuation — Wikipedia — https://en.wikipedia.org/wiki/Contingent_valuation
8. Welfare economics — Wikipedia — https://en.wikipedia.org/wiki/Welfare_economics
9. Kaldor–Hicks efficiency — Wikipedia — https://en.wikipedia.org/wiki/Kaldor%E2%80%93Hicks_efficiency
10. Quality-adjusted life year — Wikipedia — https://en.wikipedia.org/wiki/Quality-adjusted_life_year
11. Intergenerational equity — Wikipedia — https://en.wikipedia.org/wiki/Intergenerational_equity
12. Stern Review on the Economics of Climate Change — Wikipedia — https://en.wikipedia.org/wiki/Stern_Review
13. The Green Book: Central Government Guidance on Appraisal and Evaluation — HM Treasury — https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government
14. Circular A-4, Regulatory Analysis (revised 2023) — US Office of Management and Budget — https://bidenwhitehouse.archives.gov/wp-content/uploads/2023/11/CircularA-4.pdf
15. Discounting Costs and Benefits in Economic Analysis of World Bank Projects — World Bank — https://documents.worldbank.org/en/publication/documents-reports/documentdetail/099610503022315638
16. Cost-Benefit Analysis and the Environment: Further Developments and Policy Use — OECD — https://www.oecd.org/en/publications/2018/06/cost-benefit-analysis-and-the-environment_g1g8b70e.html
17. Cost-Benefit Analysis: Concepts and Practice (5th edition) — Anthony E. Boardman, David H. Greenberg, Aidan R. Vining, David L. Weimer, Cambridge University Press — https://www.cambridge.org/highereducation/books/cost-benefit-analysis/484720E57798B7E7A29C7156407CD4A1