A Nested Logit Model of Recreational Fishing Demand in Alaska

This paper (and the large study behind it) was in some ways a labor of love. We set out to build a discrete choice model to predict recreational fishing behavior that would extend the literature along several dimensions. our interest in doing so stemmed from earlier work that we were individually or collectively involved in. hanemann’s dissertation (1978)1 suggested the power of the discrete choice random utility model (RUM) to look at a large set of recreational alternatives, in this case beaches in the Boston area, and identify the role that attributes like water quality played in consumer choice behavior.2 Carson worked as a research assistant on a large recreational demand project for Jeff Vaughan and Cliff Russell at Resources for the Future (Vaughan and Russell 1982a,b). This project attempted to expand the travel cost framework to dealing with water quality issues and different types of fishing on a large spatial scale using the 1975 Survey of Hunting, Fishing, and Wildlife-Associated Recreation and a much smaller survey undertaken by the research team aimed at gathering specific information for placing a monetary value on different types of fishing days. The project used a discrete choice framework and had bumped up against both computational limits and limits of what could be estimated using