Conference Agenda

This paper empirically implements a structural framework to identify and estimate use and non-use values associated with water quality improvements. I combine revealed preference data from a random utility travel cost model of recreation demand with stated preference data from a contingent valuation referendum to jointly estimate welfare measures within a consistent utility-theoretic setting. The approach allows total willingness to pay for water quality improvements to be decomposed into use and non-use components, addressing an identification challenge in nonmarket valuation.

The empirical analysis uses web-based survey data from the general population of Michigan, which include detailed trip-level recreation behavior and referendum responses for statewide water quality changes. Recreation demand parameters are first identified from observed site choices and travel costs, while stated preference responses inform preferences over quality changes beyond observed use. Building on the structural model of Day et al. (2019), I extend the framework to allow for joint estimation of revealed and stated preference data, enabling separate identification of use and non-use values without imposing ad hoc assumptions.

The results indicate that structurally identified use and non-use values differ meaningfully from estimates obtained using simpler ad hoc approaches that rely on partial welfare measures or residual valuation. These differences have important implications for benefit–cost analysis of water quality policies, particularly when non-use values constitute a substantial share of total benefits. The paper contributes to the literature on nonmarket valuation by providing an empirically tractable and theoretically consistent method for welfare decomposition.