Benefit Transfer Valuation

Unit-value transfer takes a point estimate of value — say willingness to pay per household — from one or more study sites and applies it, perhaps with an income or price adjustment, to the policy site. Function transfer instead borrows an estimated value function that relates value to site and population characteristics, and computes the policy-site value by inserting that site's own characteristics. Johnston and colleagues note that function and meta-analytic transfers generally outperform naive unit transfers because they explicitly adjust for measured differences, whereas unit transfer assumes the borrowed value applies unchanged.

Convergent-validity studies, which value the same site both primarily and by transfer, find that transfer errors vary widely and can be substantial, often tens of percent and sometimes more, growing as the correspondence between study and policy sites weakens. The 2015 guide treats managing this error as central: choose closely matching, high-quality studies, prefer function or meta-analytic transfer where possible, conduct sensitivity analysis over transfer choices, and report the expected error rather than a single deceptively precise number so decision-makers understand the uncertainty.

Benefit transfer should be avoided or treated as merely provisional when the decision is high-stakes or precedent-setting and demands a defensible site-specific value, when the resource is novel or unique with no comparable studies in the literature, or when the policy context differs sharply from available study sites in population, baseline scarcity, culture, or market scope. In such cases the transfer error is likely too large or too uncertain to justify reliance, and a primary valuation study — for example a choice experiment — is the appropriate course.