Modeling non-compensatory preferences in environmental valuation

Abstract

While the compensatory model of choice dominates the environmental valuation literature, non-compensatory models, where individuals do not tradeoff one attribute for another, are sometimes found to be better representations of choice behavior. Most non-compensatory models employ “cutoffs”, the point at which utility abruptly changes. But cutoffs are usually elicited directly from respondents using a stated preference question. Such elicitation could be inaccurate and might introduce bias to the decision process. In this article, we develop a model that estimates cutoff levels endogenously. Our model has two error components, one for the utility function and another for the cutoff function. This facilitates the estimation of the cutoff as a function of individual specific variables. We estimate the model by maximizing the log-likelihood function that involves the weighted sum of the two error components. We test the model using synthetic data and find that estimated parameters are close to the true parameters. When applied to actual empirical data, our model appears to be a better fit than the compensatory preference model; however it is somewhat different than the self-reported cutoff model, highlighting the need for an approach that does not rely on stated cutoff information.