Finance
Cost-Benefit Analysis
An appraisal method that identifies every material cost and benefit of a proposal, expresses them in money, discounts them to present value and compares the totals, so that options are judged on evidenced net benefit rather than on the persuasiveness of their sponsors.
Also known as Benefit-cost analysis, CBA, Social cost-benefit analysis. First set out by Jules Dupuit (theoretical foundations); developed by many hands in 1844; the primary source is cited in full below.
Where this is contested
A layered lineage rather than a single invention: Dupuit supplied consumer surplus and the theory in 1844, the US Flood Control Act of 1936 made benefits-exceed-costs a statutory test, and the Kaldor-Hicks criterion (1939-1940) plus government appraisal manuals such as HM Treasury's Green Book formalised modern practice.
- Format
- Scoring model
- Level
- Corporate · Business unit
- Best for
- Evaluate options · Allocate resources
- Decision stage
- Explore options · Decide
- Difficulty
- Advanced
- Time to apply
- A day or two for a disciplined commercial appraisal with data to hand; public-sector business cases run to weeks.
Plate · The model
The components
Identify and scope the costs
A complete inventory of what the proposal consumes over its whole life: capital, operating and opportunity costs, and costs imposed on third parties, all measured against the counterfactual.
Signals of strength
Whole-life costs, and never just the build · Opportunity costs of tied-up assets included · Costs to others counted, not just costs to the promoter · Scoped against an honest do-nothing base case
Identify and scope the benefits
A complete inventory of the value the proposal creates, each stream with a causal story and an evidence base, and with transfers and displacement stripped out of the count.
Signals of strength
Each benefit traceable to a mechanism, and to a payer · Displacement and transfers excluded from the total · Non-cash benefits identified and flagged for valuation · Benefit estimates referenced to comparable real cases
Monetise and discount to present value
Conversion of all streams into money and then into present values using a stated discount rate, making costs and benefits that arrive in different years genuinely comparable.
Signals of strength
Valuation methods stated for every non-market item · A justified discount rate, applied consistently · Real and nominal values never mixed · Optimism bias adjustments applied and disclosed
Compare and decide
The reading of the result: net present value, benefit-cost ratio, sensitivity of both to the fragile assumptions, and the distribution of gains and losses, feeding a decision rather than dictating one.
Signals of strength
NPV and BCR reported with their sensitivities · The switching values that reverse the decision identified · Winners and losers named, not netted into silence · The decision recorded with its reasoning
When it earns its keep
- A large, mostly irreversible commitment of capital is proposed and the organisation needs a disciplined answer to whether it is worth it against doing nothing.
- Several options compete for the same funds and a common monetary yardstick is needed to compare unlike things fairly.
- A proposal's benefits are spread over many years, so timing matters and discounting is the only honest way to compare money now with money later.
- A decision must withstand external scrutiny, from lenders, boards or, in the public sector, appraisal regimes such as the Green Book, and the reasoning needs a recognised, auditable form.
And when it doesn't
- The decisive factors resist monetisation. Where reputation, safety culture or strategic option value dominates, forcing them into pounds produces confident nonsense; use a decision matrix that lets them stay in their own units.
- The decision is small or reversible. A full CBA on a modest, recoverable spend costs more than being wrong would.
- Deep uncertainty swamps the estimates. When the plausible range of outcomes spans success and ruin, scenario planning and staged commitments serve better than a single net present value.
- The analysis is commissioned to defend a decision already taken. CBA is the tool most frequently abused this way, and a motivated analyst can make almost any scheme clear the bar.
How to run it
Before starting, gather the inputs the analysis depends on:
- A clearly defined proposal and a counterfactual: what happens, and what it costs, if you do nothing.
- A full inventory of costs over the appraisal period: capital, operating, opportunity costs and any negative externalities.
- A full inventory of benefits, with the evidence base for each, including which are cash and which must be valued indirectly.
- A discount rate with a stated justification, whether a commercial cost of capital or a social time preference rate.
- An appraisal period, and reference data from comparable past projects against which to test the estimates for optimism.
- 1
Define the proposal, the counterfactual and the appraisal period
State precisely what is being appraised, against what alternative, over how many years. The counterfactual does the quiet work: benefits are always relative to doing nothing, and doing nothing is rarely free. An appraisal without an honest base case flatters the scheme by default.
- 2
Identify and scope the costs
List capital costs, operating costs over the whole period, opportunity costs of assets and attention tied up, and any costs imposed on others. The commonest scoping error is truncation: counting the build and forgetting the running, maintaining, insuring and eventually replacing.
- 3
Identify and scope the benefits
List the benefit streams with the causal story for each: where the money or value actually comes from, and from whom. Distinguish genuine new value from displacement, revenue captured from your own other operations or from neighbours is a transfer, and counting transfers as benefits is the oldest trick in scheme promotion.
- 4
Monetise and discount to present value
Express each stream in money, using market prices where they exist and stated valuation methods where they do not, then discount future values to the present. The discount rate embodies a judgement about time and risk: HM Treasury's Green Book prescribes a 3.5 per cent social time preference rate for public appraisal, while a commercial appraisal uses the firm's cost of capital. State the rate and defend it, because the conclusion often turns on it.
- 5
Adjust for risk and optimism bias
Test the result before trusting it. Flex the fragile assumptions, and correct for the documented tendency of appraisals to understate costs and overstate demand: the Green Book mandates explicit optimism bias uplifts, and Flyvbjerg's reference-class evidence explains why. An appraisal that has never been compared with the outcomes of similar past projects is a hypothesis, however detailed its spreadsheet.
- 6
Compare and decide
Read the net present value and the benefit-cost ratio across the central case and the sensitivities, note who gains and who bears the costs, and decide. A marginal NPV that survives pessimistic assumptions is worth more than a handsome one that dies under the first stress test. The analysis informs the judgement; it was never meant to replace it.
Reading the result
A net present value and benefit-cost ratio for each option against the counterfactual, built from itemised, discounted cost and benefit streams, with sensitivity analysis, optimism bias adjustments and a statement of who gains and who pays.
- The central NPV is the least interesting number in the report. Read the sensitivities first: the assumptions whose plausible variation flips the sign are where the decision actually lives.
- A benefit-cost ratio near 1 means the case rests on the precision of the estimates, and appraisal estimates are systematically imprecise in the optimistic direction.
- Check what was left unmonetised and where it points. An honest CBA lists its exclusions; a dishonest one buries them, and the exclusions are usually the interesting part.
A worked example
A holiday park group weighs an indoor waterpark extension
A family-owned group of five holiday parks on the east coast of England is considering a 6.5 million pound indoor waterpark at its flagship park, intended to lift shoulder-season occupancy and justify higher nightly rates. The board commissions a cost-benefit analysis over a ten-year appraisal period at the group's 9 per cent cost of capital, against a counterfactual of continued seasonal trading with minor refurbishment.
- Identify and scope the costs
- Capital: 6.5 million pounds for build and fit-out. Operating: about 480,000 pounds a year in lifeguards, energy, water treatment and maintenance, with energy the most volatile line. Also scoped: 150,000 pounds of pitch income lost during the nine-month build, higher insurance and business rates, and a 400,000 pound plant refurbishment pencilled for year eight. The do-nothing case carries its own cost, a slow decline in shoulder-season bookings as rival parks invest.
- Identify and scope the benefits
- Central case: shoulder-season occupancy up nine percentage points, a modest uplift in peak nightly rates, gated day-visitor admissions, and secondary spend in the cafe and shop. Bookings displaced from the group's two nearby sister parks are stripped out as transfers, which cuts the gross benefit estimate by roughly a fifth and annoys the marketing director. A claimed brand halo across the group is left unmonetised and recorded as a qualitative factor.
- Monetise and discount to present value
- Incremental net cash flow builds to about 1.05 million pounds a year from year two. Discounted at 9 per cent over ten years with the year-eight refurbishment included, the central case gives a net present value of roughly plus 0.9 million pounds and a benefit-cost ratio of about 1.1. Comfortable at first glance, and thin the moment anyone leans on it.
- Compare and decide
- Sensitivity testing finds the switching values: NPV turns negative if the occupancy uplift is five points rather than nine, if energy costs run 30 per cent above plan, or if capital overruns by 15 per cent. Reference-class comparison with similar UK leisure builds suggests capital overruns of that size are nearer the norm than the exception, which is the optimism bias warning in one line.
The read. The honest reading is that the central case is positive and fragile, with the decision hanging on capital discipline and demand. The board declines to approve the scheme as presented, and instead proceeds conditionally: a fixed-price design-and-build contract to cap the overrun risk, and a pre-sale of season passes to test the occupancy assumption with real money before ground is broken. If either condition fails, the analysis says walk away, and the board minutes now say so too. The CBA earned its fee by locating exactly where the bet was, rather than by blessing it.
Pitfalls
- Counting transfers as benefits. Spend displaced from your own sister sites, or from the town next door, is movement rather than creation, and inflating benefits with it is the classic promoter's error.
- Truncating the cost base. Appraisals that count construction and forget operations, maintenance, refurbishment and decommissioning flatter every capital scheme.
- Tuning the discount rate to the desired answer. Long-dated benefits are exquisitely sensitive to the rate; choosing it after seeing the cash flows reverses the burden of proof.
- Presenting the central case as the forecast. The point estimate will be wrong; the sensitivities and switching values are the analysis.
- Double counting the same value twice, once as revenue and again as an uplift in asset value, which are two views of one benefit.
- Treating unmonetised factors as zero. Leaving something out of the arithmetic and then out of the discussion is how the arithmetic quietly becomes the decision.
What the critics say
Monetising non-market goods is the method's deepest wound. Valuing life, health, landscape or community through willingness-to-pay imports market logic where many argue it has no jurisdiction, and the resulting numbers carry a precision their foundations cannot support. Ackerman and Heinzerling's critique is that pricing the priceless does not neutrally inform the decision; it changes what the decision is about.
Ackerman, F. and Heinzerling, L. (2004) Priceless: On Knowing the Price of Everything and the Value of Nothing. New York: The New Press.
The Kaldor-Hicks foundation makes CBA distributionally blind: a project passes if winners could hypothetically compensate losers, whether or not they ever do. Net benefit can therefore rise while the losers, often the poorest affected, are simply worse off. Modern treatments propose distributional weights as a repair, and standard practice mostly ignores them.
Adler, M. D. and Posner, E. A. (2006) New Foundations of Cost-Benefit Analysis. Cambridge, MA: Harvard University Press.
The empirical record indicts the inputs. Flyvbjerg's work across thousands of projects shows costs systematically underestimated and benefits overestimated, through optimism bias and strategic misrepresentation, to the point where he argues conventional ex ante CBA is broken without reference-class forecasting: the survival of the unfittest, in his phrase, where the most misrepresented schemes win funding.
Flyvbjerg, B. and Bester, D. W. (2021) 'The Cost-Benefit Fallacy: Why Cost-Benefit Analysis Is Broken and How to Fix It', Journal of Benefit-Cost Analysis, 12(3), pp. 395-419. https://www.cambridge.org/core/journals/journal-of-benefit-cost-analysis/article/abs/costbenefit-fallacy-why-costbenefit-analysis-is-broken-and-how-to-fix-it/608C8A0D37D38653846B9CF9DBC1DB49
Sen's foundational assessment accepts the discipline's core, explicit valuation and comparison, but rejects the mainstream variant's additional commitments: insisting that all values be routed through market-price analogies admits only a narrow class of values, ignores distribution, and demands that people be indifferent to distinctions, such as how an outcome comes about, that they demonstrably care about. The general discipline survives his scrutiny; the standard operationalisation does not.
Sen, A. (2000) 'The Discipline of Cost-Benefit Analysis', Journal of Legal Studies, 29(S2), pp. 931-952.
Kelman's ethical critique targets the method's utilitarian spine: in areas such as safety, health and environment, deciding by summed willingness to pay is not a neutral procedure but a moral theory, and a contested one. Some things are wronged by being priced at all, and the act of valuation changes them; his argument drew a full round of published replies and remains the classic statement of the position.
Kelman, S. (1981) 'Cost-Benefit Analysis: An Ethical Critique', Regulation, 5(1), pp. 33-40.
Nussbaum's contribution to the same symposium as Sen distinguishes the obvious question, what shall we do, from the tragic question, whether every available option involves serious moral wrongdoing. CBA answers the first and cannot even pose the second: by producing a best option it suggests the matter is settled, obscuring situations where the right response includes acknowledging, compensating and preventing the recurrence of a wrong the chosen option still commits.
Nussbaum, M. C. (2000) 'The Costs of Tragedy: Some Moral Limits of Cost-Benefit Analysis', Journal of Legal Studies, 29(S2), pp. 1005-1036.
Work it through
Enter the cost and benefit for each year against the do-nothing counterfactual, year zero being now (undiscounted). Set a discount rate and the present values, net present value and benefit-cost ratio compute below. The central NPV is the least interesting number; watch how few years of shortfall it takes to turn the sign.
| Year | Cost (£) | Benefit (£) | Present value |
|---|---|---|---|
0 | £0 | ||
1 | £0 | ||
2 | £0 |
- Net present value
- £0
- Benefit-cost ratio
- –
A benefit-cost ratio near 1 means the case rests on the precision of the estimates. Flex the fragile years and see how few points of shortfall turn the net present value negative.
Sources and further reading
- Dupuit, J. (1844) 'On the Measurement of the Utility of Public Works', translated in International Economic Papers, No. 2 (1952). Original in Annales des Ponts et Chaussees, 2nd series, vol. 8.
- HM Treasury (2022) The Green Book: Central Government Guidance on Appraisal and Evaluation. London: HM Treasury. ↗
- Mishan, E. J. and Quah, E. (2007) Cost-Benefit Analysis, 5th edition. London: Routledge.
- Flyvbjerg, B. and Bester, D. W. (2021) 'The Cost-Benefit Fallacy: Why Cost-Benefit Analysis Is Broken and How to Fix It', Journal of Benefit-Cost Analysis, 12(3), pp. 395-419. ↗