Estimating Non-market Values in the Presence of Social Network Effects

by Bruno Wichmann, department of Agricultural Economics at the University of Alberta

Stated Preference (SP) valuation is often the only available approach to estimation of non-market values. Among SP elicitation formats, dichotomous choice has the potential to have important incentive compatibility properties and is therefore commonly used. This paper shows that yes/no responses are not independent when non-market values are influenced by social networks. However, the empirical literature has yet to attempt estimation of non-market values explicitly accommodating network effects. We investigate the statistical properties of estimates of mean willingness to pay obtained through standard econometric approaches that ignore social networks. Monte Carlo experiments using different types of simulated and real world social networks indicate that failure to account for network effects leads to underestimation of non-market values. We also show that spatial econometric models can be adapted to estimate non-market values when responses are socially connected.

A copy of this paper is available here.