The motivation for this book arose out of our many years of teaching actuarial students and analyzing insurance data. Generalized linear models are ideally suited to the analysis of non-normal data which insurance analysts typically encounter. However the acceptance, uptake and understanding of this methodology has been slow in insurance compared to other disciplines. Part of the reason may be the lack of a suitable textbook geared towards an actuarial audience. This book seeks to address that need.

We have tried to make the book as practical as possible. Analyses are based on real data. All but one of the data sets are available on the companion website to this book:

http://www.acst.mq.edu.au/GLMsforInsuranceData/.

Computer code and output for all examples is given in Appendix 1.

The SAS software is widely used in the insurance industry. Hence computations in this text are illustrated using SAS. The statistical language R is used where computations are not conveniently performed in SAS. In addition, R code and output for all the examples is provided on the companion website. Exercises are given at the end of chapters, and fully worked solutions are available on the website.

The body of the text is independent of software or software “runs.” In most cases, fitting results are displayed in tabular form. Remarks on computer implementation are confined to paragraphs headed “SAS notes” and “Implementation” and these notes can be skipped without loss of continuity.