Basis Reference Technique

Overview

The AXIS actuarial modeling system is being significantly enhanced to provide a new alternative approach to defining which actuarial assumptions are used for Cell calculations within an AXIS Dataset. This powerful new capability to define and manage actuarial assumptions was introduced in AXIS release 2021.04.00 (05 March 2021) under beta feature control.

The Basis Reference Technique (BRT) introduces a structural change in AXIS that improves the ability to define, manage, control, review and report larger and more granular sets of actuarial assumptions used in production models. The technique will be especially useful for efficiently managing large sets of actuarial assumptions used for the new valuation frameworks that are subject to frequent update (IFRS 17, US PBA, US GAAP TI, and Canadian GAAP).

The central feature of the BRT is that the definition of all actuarial assumptions for any given Assumption Set is extracted from the Cells and is consolidated in a new Basis Reference Technique database object in AXIS. The reference database of assumptions provides a generic rules-based approach to defining actuarial assumptions according to model "segment" (i.e. any portion of the model sharing attributes such as product line, plan, features, era, and risk class characteristics of the insured which in turn determine the actuarial assumptions). These rules can then be applied consistently across multiple AXIS applications regardless of model structure or Cell naming. This database of assumption rules also becomes an accessible, archivable, and authoritative specification of the actuarial assumption basis being used for each purpose at any given point in time.

Anticipated Benefits of Using Basis Reference Technique

The BRT is expected to have many benefits, including:

  • Improved model transparency in terms of actuarial assumptions being used
  • Assured consistency in assumptions used by AXIS for different applications
  • Convenient documentation of actuarial assumptions used currently and historically through accessible databases
  • Simpler and quicker processes for defining and implementing assumption updates within an AXIS model
  • Enables simplified model rebuilds without having to update all Assumption Sets in Cells