B1975
Title: Almost stochastic dominance hypothesis testing
Authors: Amparo Baillo - Universidad Autonoma de Madrid (Spain) [presenting]
Javier Carcamo - (Spain)
Carlos Mora-Corral - Universidad Autonoma de Madrid (Spain)
Abstract: Based on a bidimensional stochastic dominance (2DSD) index characterizing both strict and almost stochastic dominance, we obtain an estimator for the minimum violation ratio (MVR) of the almost stochastic ordering between two random variables. We will discuss the asymptotic properties of the empirical 2DSD index and MVR for the usual stochastic orders. Conditions under which the bootstrap estimators of these quantities are consistent will also be discussed. As a consequence, we are able to develop consistent bootstrap procedures for testing almost stochastic dominance. The performance of the test will be illustrated with the analysis of real data sets.
