The Effect of Modeling Substitute Activities on Recreational Benefit Estimates : Is More Better?

The range of activities for which data are collected in many standard datasets for recreational fishing is limited by the mission of the funding agencies. We use a nested-logit model of recreational fishing to examine how varying the range of fishing alternatives included in the model affects the welfare results for various policies. The results are intuitive: a travel cost model that omits relevant substitute activities will understate gains and overstate losses for changes in the quality of activities. The magnitude of bias is directly related to the amount of substitution between the activity modeled and the activities omitted. We also use the results to highlight the importance of accounting for model range and portion of the user population represented in the sample, when comparing studies or when conducting benefit transfers