2Department of Statistics, Umeå university, 901 87 Umeå

ABSTRACT

This paper considers the interpretation of random components in random utility models underlying applied analyses of CVM data. Variation in responses is explained as taste variation interpreted as variation in valuation of choice alternative characteristics. The approach suggested has several interesting features and implications. Models obtained through the random utility difference approach is derived as a special case. Also, the approach gives theoretical justification for direct modeling of the distribution of individual willingness to pay. An illustration of results are given within an empirical example concerning angler’s willingness to pay for increased catch rates.