Inference for the positive stable laws based on a special quadratic distance

Inference methods for the positive stable laws, which have no closed form expression for the density functions are developed based on a special quadratic distance using negative moments. Asymptotic properties of the quadratic distance estimator (QDE) are established. The QDE is shown to have asymptotic relative efficiency close to 1 for almost all the values of the parameter space. Goodness-of-fit tests are also developed for testing the parametric families and practical numerical techniques are considered for implementing the methods. With simple and efficient methods to estimate the parameters, positive stable laws could find new applications in actuarial science for modelling insurance claims and lifetime data. © 2008 Elsevier B.V. All rights reserved.