Should Actuaries be Random

Abstract

Section 2 of the paper outlines the general methodology of simulation and serves as a guide to those appendices of the paper where some technical details are given. Section 3 of the paper considers various actuarial problems and discusses how simulation methods might be applied to them. We have not worked on all the problems described in this section ourselves. The intention is partly to suggest problems that could be tackled by simulation and to consider whether this approach is worthwhile. Throughout the paper we talk in terms of computers but many simple models can be simulated using a programmable pocket calculator. A simulation experiment is not necessarily a long-term project; useful answers may be obtained in a matter of hours. We have used simulation for probability problems where the theoretical solution of the problem would have taken much too long to achieve although such a theoretical solution has often been welcome as belated confirmation of advice given on the basis of a simulation experiment. The most recent occasions on which the actuarial profession in this country discussed the use of simulation were first in 1966 when S.