Abstract
In this paper, we investigate and compare several truncated count data models of travel cost recreation demand. Using on-site data from Jizerske hory Landscape Protection Area in the Czech Republic, we derive the single site recreation demand model with a Poisson and a Negative Binominal specification.
The models account for the truncated and overdispersed nature of the data, we also treat the problem of endogenous stratification due to the oversampling of more frequent recreation users. We also use the more flexible approach that allows the overdispersion parameter to vary according to the characteristics of the visitors and we compare it with the more restrictive approach.
The results show that the negative binomial model that accounts for truncation and endogenous stratification provides the better goodness-of-fit. If don't correct overdispersion by relying on Poisson estimates, the consumer surplus will be underestimated.
If we account for endogenous stratification, the consumer surplus will be adjusted appropriately upwards.