Assessing the spatial dependence of welfare estimates obtained from discrete choice experiments

Abstract

We report results from a discrete choice experiment designed to assess the general public’s preferences regarding rural landscape improvements. Using a random parameters logit model to account for unobserved taste heterogeneity, we exploit the panel nature of the dataset to retrieve partworths, or willingness to pay (WTP) values, for every individual in the sample. We subsequently use the Moran’s I test statistic to gauge the spatial dependence of these estimates for a range of spatial extents. We find that WTP estimates are positively spatially autocorrelated, but the degree of homogeneity diminishes as the spatial extent becomes larger before it eventually becomes relatively stable.