A utility-consistent, combined discrete choice and count data model Assessing recreational use losses due to natural resource damage

A two-stage budgeting approach can often be taken when analyzing consumer choice situations. In this paper we examine the particular situation where a consumer makes purchases of a discrete commodity of which there are a number of brands. In the first stage, the consumer decides how many purchases to make; in the second stage, the consumer decides how to allocate these purchases across brands. Our econometric approach to this type of situation uses a utility-consistent, combined discrete choice and count data model. The second stage, which is specified as a multinomial choice model, produces a price index for the commodity which is used to estimate the first stage, which is specified as a count data model. We apply the model to recreational demand behavior in Alaska in order to estimate the welfare losses suffered by recreational users due to the Exxon Valdez oil spill. These results may provide useful input to government agencies attempting to estimate the appropriate level of taxes, fines, or regulations for deterring damage to the environment.