Valuing Farmland Protection with Choice Experiments That Incorporate Preference Heterogeneity: Does Policy Guidance Depend On the Econometric Fine Print?

Although mixed logit models are common in stated preference applications, resulting welfare estimates can be sensitive to minor changes in specification. This can be of critical relevance for policy and welfare analysis, particularly if policymakers are unaware of practical implications. Drawing from an application to agricultural conservation in Georgia, this paper quantifies the sensitivity of welfare estimates to common variations in mixed logit specification and assesses practical implications for policy guidance. Results suggest that practitioners may wish to reevaluate modeling and reporting procedures to reflect the welfare and policy implications of common but often unnoticed variations in model specification.