Choice Modelling has been applied to a case of market failure in the provision of riparian buffers (strips of natural vegetation adjoining waterways) in the Lockyer catchment. With costs of water treatment increasing in the Brisbane area from increased turbidity and sedimentation in its sub-catchment areas, the implementation of riparian buffers has been identified as a possible solution to reducing these sediment loads and hence, water treatment costs. As the property rights of the land adjoining waterways in the Lockyer catchment are held by Lockyer landowners, the issue is one of willingness to accept (WTA) compensation for the loss in income-generating land. This is a unique application of Choice Modelling as the use of this technique has been predominantly restricted to willingness to pay applications and the trade-off here concerns less income and a higher quality of environmental attributes.
Conditional and mixed logit models are estimated using respondents' preferences in a choice modelling exercise. Using these models and results from the questionnaire, part-worths (or implicit prices) were estimated for the four identified attributes (compensation, water availability, topsoil erosion and environmental quality in the sub-catchment) and measures of compensating variation were calculated across 15 different alternatives to assess WTA. Measures from the compensating surplus were then used to infer values of the riparian attributes. In this way, the coefficients from the estimated logit models were combined with data concerning loss of income, as established from questionnaire responses, to determine how respondents change their WTA levels for the changed environmental conditions which would result from the riparian revegetation. Attribute values are in the expected direction (expect for a moderate el1vironmental improvement) with respondents valuing more water availability, less topsoil erosion and a big environmental improvement in the sub-catchment and respondents' utility reducing with more topsoil erosion and only a moderate environmental improvement result from riparian revegetation. This application has produced meaningful and consistent results. However, some results are sensitive to the levels of variables determined in the questionnaire rather than the Choice Modelling exercise.