I n t e rd i s c i p l i n a r y S t a t i s t i c s STATISTICAL and PROBABILISTIC METHODS in ACTUARIAL SCIENCE C6951_FM.indd 1 1/24/07 1:45:48 PM CHAPMAN & HALL/CRC Interdisciplinary Statistics Series Series editors: N. Keiding, B. Morgan, T. Speed, P. van der Heijden AN INVARIANT APPROACH TO S. Lele and J. Richtsmeier STATISTICAL ANALYSIS OF SHAPES ASTROSTATISTICS G. Babu and E. Feigelson BIOEQUIVALENCE AND S. Patterson and STATISTICS IN CLINICAL B. Jones PHARMACOLOGY CLINICAL TRIALS IN ONCOLOGY J. Crowley, S. Green, SECOND EDITION and J. Benedetti DESIGN AND ANALYSIS OF D.L. Fairclough QUALITY OF LIFE STUDIES IN CLINICAL TRIALS DYNAMICAL SEARCH L. Pronzato, H. Wynn, and A. Zhigljavsky GENERALIZED LATENT VARIABLE A. Skrondal and MODELING: MULTILEVEL, S. Rabe-Hesketh LONGITUDINAL, AND STRUCTURAL EQUATION MODELS GRAPHICAL ANALYSIS OF K. Basford and J. Tukey MULTI-RESPONSE DATA INTRODUCTION TO M. Waterman COMPUTATIONAL BIOLOGY: MAPS, SEQUENCES, AND GENOMES MARKOV CHAIN MONTE CARLO W. Gilks, S. Richardson, IN PRACTICE and D. Spiegelhalter MEASUREMENT ERROR AND P. Gustafson MISCLASSIFICATION IN STATISTICS AND EPIDEMIOLOGY: IMPACTS AND BAYESIAN ADJUSTMENTS STATISTICAL ANALYSIS OF GENE T. Speed EXPRESSION MICROARRAY DATA STATISTICAL CONCEPTS J. Aitchison, J.W. Kay, AND APPLICATIONS IN and I.J. Lauder CLINICAL MEDICINE STATISTICAL AND PROBABILISTIC Philip J. Boland METHODS IN ACTUARIAL SCIENCE STATISTICS FOR ENVIRONMENTAL A. Bailer and W. Piegorsch BIOLOGY AND TOXICOLOGY STATISTICS FOR FISSION R.F. Galbraith TRACK ANALYSIS STATISTICS IN MUSICOLOGY J. Beran C6951_FM.indd 2 1/24/07 1:45:48 PM I n t e rd i s c i p l i n a r y S t a t i s t i c s STATISTICAL and PROBABILISTIC METHODS in ACTUARIAL SCIENCE Philip J. Boland University College Dublin Ireland Boca Raton London New York C6951_FM.indd 3 1/24/07 1:45:48 PM CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2007 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Version Date: 20110713 International Standard Book Number-13: 978-1-58488-696-9 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information stor- age or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copy- right.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that pro- vides licenses and registration for a variety of users. For organizations that have been granted a pho- tocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com Dedication To my wife Elizabeth, and my children Daniel and Katherine. v Preface This book covers many of the diverse methods in applied probability and statistics for students aspiring to careers in insurance, actuarial science and finance. Itshouldalsoserveasavaluabletextandreferencefortheinsurance analyst who commonly uses probabilistic and statistical techniques in prac- tice. The reader will build on an existing basic knowledge of probability and statistics and establish a solid and thorough understanding of these methods, but it should be pointed out that the emphasis here is on the wide variety of practical situations in insurance and actuarial science where these techniques may be used. In particular, applications to many areas of general insurance, including models for losses and collective risk, reserving and experience rat- ing, credibility estimation, and measures of security for risk are emphasized. The text also provides relevant and basic introductions to generalized linear models, decision-making and game theory. There are eight chapters on a variety of topics in the book. Although there are obvious links between many of the chapters, some of them may be studied quite independently of the others. Chapter 1 stands on its own, but at the same time provides a good introduction to claims reserving via the deterministic chain ladder technique and related methods. Chapters 2,3 and 4 are closely linked, studying loss distributions, risk models in a fixed period of time, and then a more stochastic approach studying surplus processes and the concept of ruin. Chapter 5 provides a comprehensive introduction to the concept of credibility, where collateral and sample information are brought togethertoprovidereasonablemethodsofestimation. TheBayesianapproach to statistics plays a key role in the establishment of these methods. The final three chapters are quite independent of the previous chapters, but provide solid introductions to methods that any insurance analyst or actuary should know. Experience rating via no claim discount schemes for motor insurance in Chapter 6 provides an interesting application of Markov chain methods. Chapter 7 introduces the powerful techniques of generalized linear models, while Chapter 8 includes a basic introduction to decision and game theory. There are many worked examples and problems in each of the chapters, with a particular emphasis being placed on those of a more numerical and practical nature. Solutions to selected problems are given in an appendix. Therearealsoappendicesonprobabilitydistributions,Bayesianstatisticsand basic tools in probability and statistics. Readers of the text are encouraged (in checking examples and doing problems) to make use of the very versatile and free statistical software package R. vii viii PREFACE The material for this book has emerged from lecture notes prepared for variouscoursesinactuarialstatisticsgivenatUniversityCollegeDublin(The National University of Ireland – Dublin) over the past 15 years, both at the upper undergraduate and first year postgraduate level. I am grateful to all mycolleaguesinStatisticsandActuarialScienceatUCDfortheirassistance, but particularly to Marie Doyle, Gareth Colgan, John Connolly and David Williams. The Department of Statistics at Trinity College Dublin kindly provided me with accommodation during a sabbatical year used to prepare this material. I also wish to acknowledge encouragement from the Society of Actuaries in Ireland, which has been supportive of both this venture and our program in Actuarial Science at UCD since its inception in 1991. Patrick Grealy inparticular providedvery useful advice and examples on the topic of run-off triangles and reserving. John Caslin, Paul Duffy and Shane Whelan were helpful with references and data. Ihavebeenfortunatetohavehadmanyexcellentstudentsinbothstatistics and actuarial science over the years, and I thank them for the assistance and inspiration they have given me both in general and in preparing this text. Particular thanks go to John Ferguson, Donal McMahon, Santos Faundez Sekirkin, Adrian O’Hagan and Barry Maher. Many others were helpful in reading drafts and revisions, including Una Scallon, Kevin McDaid and Rob Stapleton. Finally, I wish to thank my family and many friends who along the path to completing this book have been a constant source of support and encouragement. Introduction In spite of the stochastic nature of most of this book, the first chapter is rather deterministic in nature, and deals with Claims Reserving and Pricing with Run-off Triangles. In running-off a triangle of claims experience, one studieshowclaimsarisingfromdifferentyearshavedeveloped,andthenmakes use of ratios (development factors and/or grossing-up factors) to predict how future claims will evolve. Methods for dealing with past and future inflation in estimating reserves for future claims are considered. The average cost per claim method is a popular tool which takes account of the numbers of claims as well as the amounts. The Bornhuetter–Ferguson method uses additional informationsuchasexpectedlossratios(lossesrelativetopremiums)together withthechainladdertechniquetoestimatenecessaryreserves. Delaytriangles of claims experience can also be useful in pricing new business. Modeling the size of a claim or loss is of crucial importance for an insurer. In the chapter on Loss Distributions, we study many of the classic probabil- ity distributions used to model losses in insurance and finance, such as the exponential, gamma, Weibull, lognormal and Pareto. Particular attention is paid to studying the (right) tail of the distribution, since it is important to not underestimate the size (and frequency) of large losses. Given a data set of claims, there is often a natural desire to fit a probability distribution with reasonablytractablemathematicalpropertiestosuchadataset. Exploratory data analysis can be very useful in searching for a good fit, including basic descriptive statistics (such as the mean, median, mode, standard deviation, skewness,kurtosisandvariousquantiles)andplots. Themethodofmaximum likelihood is often used to estimate parameters of possible distributions, and various tests may be used to assess the fit of a proposed model (for exam- ple, the Kolmogorov–Smirnoff, and χ2 goodness-of-fit). Often one may find that a mixture of various distributions may be appropriate to model losses due to the varying characteristics of both the policies and policyholders. We also consider the impact of inflation, deductibles, excesses and reinsurance arrangements on the amount of a loss a company is liable for. Followingonfromastudyofprobabilitydistributionsforlossesandclaims, thechapteronRiskTheoryinvestigatesvariousmodelsfortheriskconsisting of the total or aggregate amount of claims S payable by a company over a relatively short and fixed period of time. Emphasis is placed on two types of models for the aggregate claims S. In the collective risk model for S, claims are aggregated as they are reported during the time period under consider- ation, while in the individual risk model there is a term for each individual ix
Description: