Abstract

Journal of Actuarial Practice

Volume 9, Numbers 1 and 2, 2001


Fitting Loss Distributions in the Presence of Rating Variables

Farrokh Guiahi

Abstract

This paper focuses on issues and methodologies for fitting alternative statistical models---parametric probability distributions---to samples of insurance loss data. The interactions of loss distributions, deductibles, policy limits, and rating variables in the context of fitting distributions to losses are discussed. Fitted loss distributions serve an important function in pricing insurance products. The methodology developed in this paper is applied to a sample of insurance loss data that has the lognormal as the underlying loss distribution.

Key words and phrases: quadratic generalized linear models, curve fitting, right-censored and left-truncated data, rating variables, maximum likelihood estimation, iteratively re-weighted least squares, parametric distribution.

Corresponding Author:

Farrokh Guiahi

Department of Business Computer Information Systems and Quantitative Methods,

Zarb School of Business,

Hofstra University,

Hempstead, NY 11549

USA

USA


© Copyright Absalom Press, Inc.