Choice Set Definition Issues in a Kuhn-Tucker Model of Recreation Demand

This line of research considers econometric methods for dealing with agent-level consumer choice data in the presence of binding non-negativity constraints. Specifically, the Kuhn-Tucker (KT) model and its dual counterpart (Wales and Woodland (1983), Lee and Pitt (1987), Phaneuf (1999), Phaneuf, et. al. (2000)) are applied to the problem of recreation site choice. The conceptual attractiveness of these methods is generally acknowledged, but the difficulty of estimation has prevented wide application. Extensions of these works to allow wider application are emphasized here. These include the use of more general functional forms and error structures, expanding the dimensions of models that can be estimated, improving the tractability and usability of these models, and providing additional applications. Good progress has been made in these areas. The paper included for consideration provides results for a 16 good complete demand system KT model applied to a fresh data set on the demand for visits to wetland recreation sites in Iowa. Recreation demand provides a good platform to study binding non-negativity constraints, since few individuals visit all recreation sites available to them and hence the data contain many zeros. The KT method recovers estimates of the utility function, from which utility consistent welfare measures resulting from changes in wetland characteristics and availability are calculated via Monte Carlo integration. A second paper is included for reference, providing more detail on the calculation and interpretation of welfare effects.