WTP estimation for environmental goods

WTP estimates for environmental improvements are the primary output of environmental valuation DCEs - and they must be estimated and presented with appropriate care to be credible in policy processes.

This article explains how to estimate WTP for environmental goods from DCE data, how to aggregate household WTP to programme-level benefits, and how to present results for cost-benefit analysis.

Ben White, 07.07.2026

From model output to policy-relevant WTP

Estimating WTP from environmental DCE data is straightforward in principle: divide the attribute coefficient by the cost coefficient. But making those estimates policy-relevant requires additional steps: converting from per-choice-set WTP to annual WTP, aggregating across the affected population, accounting for households vs individuals, and expressing uncertainty around the estimates.

Getting these steps wrong - or not doing them at all - produces WTP estimates that cannot be used in cost-benefit analysis, even if the model estimation is technically correct.

Aggregation and benefit transfer

Policy cost-benefit analysis requires aggregate benefits - the total WTP of the affected population, not just the mean WTP of the study sample. Aggregation requires: a valid WTP estimate from the DCE model; information about the size of the affected population; and assumptions about how the study sample represents the broader population.

Benefit transfer extends WTP estimates from the study site to a policy site - using WTP estimates from one study to inform decisions about a different but similar environmental good. The validity of benefit transfer depends on the similarity between study and policy sites and the robustness of the original WTP estimates.

SurveyEngine's analytical output tools include benefit aggregation templates that apply the correct population scaling factors to DCE WTP estimates.

TLDR Quick links

Environment resources

Forum: modelling questions

Open a free account

Estimating and aggregating WTP in Apollo and R

Step 1: Estimate marginal WTP for each attribute. For each non-cost attribute, calculate WTP = -(beta_attribute / beta_cost). In a mixed logit model, also estimate the standard deviation of WTP across the population.

Step 2: Express WTP in policy-relevant units. Convert model WTP (which may be per choice task) to annual WTP per household. If your cost attribute is expressed as annual household cost, no conversion is needed. If it is expressed as a one-off payment, multiply by an appropriate annuitisation factor.

Step 3: Calculate total benefit for a proposed policy. Total benefit = annual WTP per household × number of affected households. Express as a present value using the appropriate social discount rate.

Step 4: Construct confidence intervals around the aggregate. Propagate uncertainty from the WTP estimate and the population size estimate through to the aggregate benefit estimate. Report a range as well as a point estimate.

Step 5: Test robustness with benefit transfer adjustments. If the study sample differs from the policy population in income, environmental values, or proximity to the good, apply income elasticity adjustments or geographic decay factors to the WTP estimates.

Worked example - aggregate WTP for coastal water quality

A coastal water quality DCE estimates mean WTP for Excellent vs Poor bathing water quality at £68/household/year (95% CI: £54–£82). The affected population is 480,000 households within 10km of the affected coastline.

Total annual benefit = £68 × 480,000 = £32.6 million/year (range £25.9M–£39.4M). Present value over a 30-year programme at 3.5% discount rate = £32.6M × 17.3 = £564 million (range £448M–£681M). The investment programme costs £280 million. Benefit-cost ratio = 2.01 (range 1.60–2.43) - the programme passes the cost-benefit test under all scenarios.