Topic 10 of 64· Governmental economic appraisal
Social Cost-Benefit Analysis (SCBA)
Social cost-benefit analysis converts every cost and benefit of a policy or programme — market and non-market — into a common monetary unit, discounts future flows to present value, and nets them off to produce a single number: does this proposal make society better off, and by how much?
Why it matters
SCBA is the default quantitative method in the economic case of Green Book appraisal: HM Treasury's guidance requires proposals to demonstrate a positive net present social value (NPSV) wherever benefits can be credibly monetized, using willingness-to-pay as the basic valuation principle for non-market goods (https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government, Chapter 5). The discipline it enforces is that "social" cost-benefit analysis is not the same exercise as a private-sector investment appraisal: it must include costs and benefits that fall on third parties who are not party to the transaction (externalities), it must use the social discount rate rather than a commercial cost of capital, and it should apply distributional weighting where a pound matters more to a poorer household than a richer one.
Where SCBA breaks down is exactly where its critics expect: goods with no market analogue — clean air, social cohesion, the value of a life saved — have to be monetized using stated preference or revealed preference methods, or a shadow price has to be constructed. When monetization is contested rather than merely difficult, the Green Book itself recommends falling back to cost-effectiveness analysis or multi-criteria decision analysis rather than forcing a number that nobody believes.
The maths
NPSV = Σ [t=0 to T] (Benefit_t − Cost_t) / (1 + r)^t
where:
Benefit_t = all monetized benefits in year t, including non-market
goods valued via stated/revealed preference or shadow price
Cost_t = all monetized costs in year t, including opportunity cost
of resources (see ../opportunity-cost-in-public-spending/)
r = social discount rate (HM Treasury sets 3.5% declining to
lower rates beyond year 30, per Green Book Annex A)
T = appraisal period
Benefit-cost ratio (BCR) = Σ PV(Benefits) / Σ PV(Costs)
A BCR above 1 (or NPSV above zero) indicates net social value. The Green Book's value-for-money categories (as used in transport and infrastructure appraisal) label BCR ranges: below 1.0 is poor value for money, 1.0–1.5 is low, 1.5–2.0 is medium, 2.0–4.0 is high, and above 4.0 is very high. Sensitivity analysis — re-running the NPSV under pessimistic and optimistic assumptions — is mandatory, not optional, because monetized non-market benefits carry wide uncertainty bands.
Worked example
Local authority: a council appraises a £3m investment in a new cycling and walking network over a 20-year appraisal period at a 3.5% discount rate.
Costs: £3m capital in year 0, £50,000/year maintenance (years 1-20)
PV(maintenance) ≈ £50,000 × 14.2 (20-year annuity factor at 3.5%) ≈ £710,000
Total PV(costs) ≈ £3.71m
Benefits (all monetized via published DfT/WHO valuation tools):
Health benefit from increased physical activity: £180,000/year
Absenteeism reduction: £40,000/year
Decongestion (fewer car trips): £60,000/year
Total benefit stream: £280,000/year
PV(benefits) ≈ £280,000 × 14.2 ≈ £3.98m
NPSV = £3.98m − £3.71m = +£0.27m
BCR = 3.98 / 3.71 = 1.07 → "low" value for money
The scheme clears the bar but only just; a sensitivity run at a 20% lower health-benefit estimate (reflecting genuine uncertainty in physical-activity valuation) flips the BCR below 1.0, which is exactly why the Green Book requires the sensitivity table to be published alongside the headline number, not just the central estimate.
Charity: an infant-mortality prevention programme costing £500,000/year is evaluated using the value of a statistical life (VSL) — a shadow price, not an observed market price — of roughly £2.1m (HM Treasury's 2023-updated figure, itself derived from stated-preference studies). Averting one infant death per year against a £500,000 cost gives a BCR of 4.2, comfortably "very high" value for money — but the entire result rests on the VSL figure, which is why any SCBA using VSL must disclose it as an assumption, not a fact.
Software engineering connection
SCBA is the natural frame for platform and infrastructure investment decisions in government software — comparing a shared identity platform against departmental point solutions, for instance, requires monetizing benefits like reduced duplicate onboarding cost, reduced fraud, and faster time-to-service that have no market price on their own. Engineers building the underlying service should expect programme leads to ask for inputs to this analysis: unit costs of transactions (see cost per transaction), expected volumes, and degradation/downtime costs. The discipline that matters most to import: discount future benefits, name the counterfactual baseline explicitly (see counterfactual analysis), and never present a single point estimate without its sensitivity range.
Pitfalls
- Double-counting benefits. Counting both "time saved" and "productivity gained from that time" as separate benefit lines overstates the case; time saved is the benefit, its downstream use is not an additional one unless independently evidenced.
- Omitting displaced costs. A scheme that moves congestion from one road to another, or moves fraud from one channel to another, has not created the net benefit its headline NPSV implies — see displacement and attribution.
- Using a private discount rate. Applying a commercial cost of capital (say 8–10%) instead of the social discount rate systematically undervalues long-horizon public benefits like health and environmental gains — see social discount rate.
- Monetizing the uncontested and hand-waving the contested. If two-thirds of a proposal's benefit is a confidently monetized efficiency saving and one-third is a shakily monetized wellbeing gain, the headline NPSV silently blends a hard number with a soft one; report them separately.
Sources
- HM Treasury. "The Green Book: appraisal and evaluation in central government." 2022, Chapter 5. https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government
- HM Treasury. "Green Book supplementary guidance: value of a statistical life." 2023. https://www.gov.uk/government/publications/green-book-supplementary-guidance-value-of-a-statistical-life
- Department for Transport. "TAG unit A1.1: cost-benefit analysis." Transport Analysis Guidance. https://www.gov.uk/guidance/transport-analysis-guidance-tag