Fat Tails, Flat Tails, and Willingness to Pay

Abstract

This paper explores the limitations of two commonly used nonparametric approaches, the Turnbull estimator and the Kriström estimator. Fat tails and flat tails in the distribution of bid responses can result in a large divergence in willingness to pay estimates across different estimation methods. Since it is extremely rare to capture the tail of the distribution, we highlight the merit of an alternative approach using 120 data sets drawn from the literature. We calculate 4 non-parametric estimates of willingness to pay for each data set. Meta analysis is used to help determine when the Turnbull and Kriström estimators diverge.