SUGI 29 Systems Architecture - Paper 224-29 - Architectural Views of Building an Actuarial System using SAS/Warehouse Administrator - ABSTRACT

Sep 19, 2026

SAS/Warehouse Administrator supports development and documentation of a large actuarial liability valuation system through advanced code generation. The paper presents the system’s architecture and explains how to address practical problems, risks, and traps inherent in this kind of approach. It frames the context of Quebec’s no-fault automobile insurance plan and clarifies how data acquisition, liability valuation, and situation follow-up processes must coordinate across categories.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Introduction

Context of the actuarial liability valuation system

Global Architecture

Data acquisition process

Liability valuation process

Situation follow-up process

SAAQ Development Methodology

Methodology and business processes organization

Three infrastructure layers

Technological infrastructure layer

SUGI 29 Systems Architecture Paper 224-29 Architectural Views of Building an Actuarial System using SAS/Warehouse Administrator® Dominic Roy, DMR Conseil / Fujitsu Consulting, Sainte-Foy, Quebec, Canada Alain Baillargeon, DMR Conseil / Fujitsu Consulting, Sainte-Foy, Quebec, Canada ABSTRACT SAS/Warehouse Administrator is an advanced code generation tool used in building data warehouses. In our project, the tool was used in order to support and document the development of a large actuarial system. This paper discusses the architecture of the system, the ways and means to solve the problems and the traps and pitfalls of such an approach. INTRODUCTION Roy & Genois (2004) explains the business context of an actuarial liability valuation system built at the Société del’assurance automobile du Québec. The following presentation deals with architectural aspects of the project. However a short introduction to the context is necessary to understand the architecture. The “Société de l’assurance automobile du Québec“(SAAQ) is responsible for the Public Automobile Insurance Plan in the province of Quebec. This plan is a no-fault insurance regime that protects all victims of motor vehicle accidents. It is a mandatory plan for all drivers and all vehicles. It covers Quebec citizens circulating on provincial roads within and out of the province. It also protects non-Quebec citizens while using provincial roads. The plan is financed with a part of the driver’s license and auto registration fees. The plan’s actuaries are responsible for accurately predicting the amount of money that must be set aside (the liability) to cover the future needs of all new accident victims during the current year. They must also check whether the money set aside in previous years is sufficient to cover victims from those years. The liabilities are calculated and justified separately for each category of expenses. A new actuarial liability valuation system was necessary due to certain technological problems, but mainly because of the growing complexity of the valuation. At the same time, the upper management wanted a more secure system to sustain a corporate part of the mission. GLOBAL ARCHITECTURE An actuarial liability valuation system is composed of three major processes: data acquisition, liability valuation and situation follow-up. The data acquisition process is done by extracting the data, verifying the quality and making corrective actions on the data. The major proportion of the process can be taken over by SAS/Warehouse Administrator as a regular data warehouse activity. The liability valuation process is the most complex segment and the less conventional business intelligence activity. Each expense category has its own evolution behavior, so each category must have a specific predictive model based on a more general predictive models. The actual system deals with 3 income replacement rents evaluated on an individual basis and more than fifty expense categories evaluated on a global basis. Those valuations must take care of interdependence between categories, taxes, economic hypothesis and reimbursements for other organizations. | Actuarial Liability Valuation System
Data
acquisition | | | | --- | --- | --- | | | | | | | | | The situation follow-up process is a more conventional business intelligence part in which users watch the current economic and cash flow situation. Reports can be scheduled with SAS/Warehouse Administrator, as well as multidimensional data preparation. 1 SUGI 29 Systems Architecture SAAQ Development Methodology THE METHODOLOGY AND THE BUSINESS PROCESSES ORGANIZATION Because the proposed system’s predecessors were considered corporate level systems, it was essential that the new application would be a corporate level system as well. This meant that the highest standards of the SAAQ system development methodology had to be met, with a defined process, structured documentation and a

FAQs

What role does SAS/Warehouse Administrator play in the project?

It is used as an advanced code generation tool to support and document development, especially for regular data-warehouse style activities such as data acquisition.

What are the three major processes in the actuarial liability valuation system?

They are data acquisition, liability valuation, and situation follow-up.

Why was a new actuarial liability valuation system necessary?

It was required due to technological problems and mainly because valuation complexity had been growing, while management also wanted a more secure corporate system.

Sep 19, 2026

Uploaded by Chloe Bennett