B1499
Title: A new test of fit for one-sided stable laws based on the Laplace transform
Authors: Jaco Visagie - North-West University (South Africa) [presenting]
James Allison - North-West University (South Africa)
Simos Meintanis - University of Athens (Greece)
Leoni Snyman - North-West University (South Africa)
Abstract: The empirical Laplace transform is used to construct an L2-type test for the null hypothesis that a positive random variable follows a one-sided stable distribution with an unspecified tail index $(0,1)$. The large-sample properties of the test are investigated. The asymptotic null distribution of the proposed test statistic depends on the tail index. As a result, a parametric bootstrap procedure is used to arrive at critical values. Monte Carlo results compare the finite sample power performance of the new test to those of classical procedures and show that the newly proposed test compares favourably against several of the alternative distributions considered. The use of the new test is illustrated using observed financial data.
