The role that outdoor recreation plays in consumers’ daily lives has gradually increased over the last centuries. There is a long-standing tradition of recreation demand studies capturing preference heterogeneity in outdoor recreation, and this tradition sketches a useful but diffuse picture. Many applications succeed in pointing out the effect that distance and cost have on the number of trips to some particular site, or uncover substitution patterns between sites. However, the choice set is usually confined to alternatives exhibiting some degree of equivalency, for example a set of different forests. Many of the activities typified as outdoor recreation can indeed be exercised at different locations, but also, and rather important, at dissimilar types of locations.

We evaluate time allocation to a set of dissimilar landscape types for home-based outdoor recreation trips, by developing and estimating a Kuhn-Tucker demand system which is in line with Bhat (2008). This should allow policy makers to evaluate the effects of, for example, the addition of a municipal park to the set of existing outdoor recreation sites, or the degradation in quality of countryside locations due to lesser maintenance.

To accommodate differences in accessibility to landscape types, we introduce a separate stage for mode and destination choice. This stage tests the applicability of two types of logit structures, namely multinomial logit (MNL) and nested logit (NL). Across the range of landscapes, each consumer has a choice between three modal types and a set of Dutch municipalities, for which the size depends on the availability of a landscape type. The estimates of the mode-destination choice model are integrated back in the Kuhn-Tucker time allocation model, and the joint parameter estimates offer a behaviorally intuitive view on how consumers value accessibility to landscape types.