Topic 15 of 64· Governmental economic appraisal
Revealed Preference Valuation
Revealed preference methods infer the value of a non-market good from observable behaviour in a related market, rather than asking people directly. Hedonic pricing and the travel-cost method are the two workhorse techniques: both start from a real transaction and back out an implicit price for the thing that was never directly sold.
Why it matters
Where stated preference methods ask a hypothetical question, revealed preference methods observe what people actually paid for, which the Green Book treats as generally more credible evidence, all else equal, because it is not subject to hypothetical bias — respondents in a hedonic house-price study genuinely paid the premium or discount being measured (https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government, Annex 2). Hedonic pricing decomposes a market price — typically house prices — into implicit prices for each attribute of the good, letting analysts isolate, for example, the price premium households actually pay to live somewhere quieter or with better air quality, controlling statistically for every other attribute that also affects house price (size, location, school catchment). The travel-cost method does the analogous thing for recreational sites with no entry fee: the time and money people spend travelling to a site reveals a lower bound on what the site is worth to them, because nobody incurs a cost exceeding what the visit is worth to them.
Both methods share a structural limitation: they can only value what is embedded in an existing market transaction. Noise near a runway shows up in house prices because people who care about noise sort into quieter housing; the existence value of a species nobody visits or lives near does not show up in any transaction at all, which is exactly the gap stated preference methods exist to fill.
The maths
Hedonic pricing:
House price = f(structural attributes, location attributes,
environmental attribute of interest, ...)
Estimate via regression; the coefficient on the environmental
attribute (holding all else constant) is its implicit price.
Implicit price of attribute X = ∂(House price) / ∂X
Travel-cost method:
Visit rate (visits per capita from zone i) = f(travel cost from zone i,
substitute sites, socioeconomic controls)
Estimate a demand curve for visits as a function of travel cost.
Consumer surplus = area under the estimated demand curve
= value of the site to visitors
Both methods require a statistically sound control set — omitting a confounding attribute (hedonic) or a nearby substitute site (travel-cost) biases the implicit price in a direction that is not always obvious in advance, which is why Green Book Annex 2 requires the regression specification and controls to be reported, not just the headline coefficient.
Worked example
National government: the Green Book's own shadow price of carbon methodology draws in part on hedonic evidence, but a simpler illustrative case is aircraft noise. A hedonic study regressing house sale prices in a flight-path area against distance-weighted noise exposure, controlling for size, age, and school catchment, finds each 1 decibel increase in average noise exposure is associated with a 0.5% reduction in house price. For a typical £280,000 house in the affected area:
Implicit price per decibel = £280,000 × 0.5% = £1,400 per household
Households affected by a 3dB increase from a new runway = 18,000
Aggregate implied cost of the noise increase = £1,400 × 3 × 18,000 = £75.6m
This is a one-off capitalized cost (embedded in house price), which the appraisal must be careful not to double-count against a separately estimated annual noise-annoyance cost stream.
Charity: an environmental charity uses the travel-cost method to value a free-to-enter nature reserve. Survey data on visitor postcodes gives average round-trip travel cost (time valued at the Green Book's recommended non-work time value, plus fuel) of £14 per visit, with 40,000 visits per year. The estimated demand curve — visit rates falling as travel cost from a zone rises — implies a consumer surplus per visit, above the £14 actually spent, of roughly £9.
Total annual value = 40,000 visits × (£14 spent + £9 consumer surplus)
= 40,000 × £23 ≈ £920,000/year
This dwarfs the reserve's zero entry-fee revenue and gives the charity's trustees a defensible figure for the site's recreational value when making the case to funders.
Software engineering connection
Revealed preference thinking shows up in public-sector product analytics more often than practitioners realise: usage data from a free government digital service is itself revealed-preference evidence of value (frequency, session length, and — most tellingly — repeat-versus-one-off usage patterns can be analysed the same way a travel-cost model treats visit frequency against distance). Where a service has genuine substitutes (a paper channel, a phone line), the "cost" citizens incur to use the digital channel instead (time, data, a device) can be estimated and compared to usage, echoing the travel-cost logic directly. See digital service standard and open data value, which faces exactly this valuation problem for a good with no direct market price.
Pitfalls
- Omitted variable bias in hedonic models. Leaving out a correlated attribute (school quality correlating with both house price and the environmental variable of interest) biases the implicit price estimate; specification needs to be reported and scrutinised, not just the result.
- Ignoring substitute sites in travel-cost studies. A visitor's revealed value for a site is understated if a closer substitute exists and is not controlled for — they may be visiting mainly because it's free, not because it's uniquely valuable.
- Applying revealed preference to a good with no market echo at all. Existence value, option value, and bequest value do not show up in any transaction and cannot be recovered by hedonic or travel-cost methods — that gap belongs to stated preference valuation.
- Confusing capitalized (one-off) value with an annual flow. Hedonic house-price effects are typically one-off capitalized values; treating them as an annual benefit stream inflates the appraisal.
Sources
- HM Treasury. "The Green Book," Annex 2: valuing non-market impacts. https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-government
- Department for Transport / Civil Aviation Authority. Aircraft noise valuation studies used in airport appraisal. https://www.gov.uk/guidance/aviation-noise
- Rosen S. "Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition." Journal of Political Economy, 1974.
- Clawson M, Knetsch JL. "Economics of Outdoor Recreation." Johns Hopkins University Press, 1966 (origin of the travel-cost method).