Assumption Sets in AXIS (October, 2016)
GGY has been working hard to prepare AXIS for the future demands of new financial reporting and capital frameworks. One important element in the AXIS design strategy is a fundamental change in the way sets of actuarial assumptions are arranged and stored within the Cell object. This change enables more sets of assumptions to be defined and permits more flexible use of each assumption set. The change also impacts the user experience with the Cell interface and any process that creates or updates Cell contents as part of AXIS model maintenance. Assumption Sets have been introduced in AXIS 2017.05.01. This article elaborates on the changes and on the steps that AXIS users should consider for implementing the Assumption Set feature in AXIS. Further information has been provided through a Webinar and training will be available to help clients prepare for the change.
Why have the Liability Module Cells in AXIS changed?
The fundamental purpose of AXIS Liability Module Cells is to define the product features for all contracts intended to be modeled within each Cell as well as the actuarial and other related assumptions that drive the projection of cash flows and financial results into the future. Currently the Cells support 7 sets of these “actuarial” assumptions, with each set designated for either the projection of the initial inforce cohort through time (Pricing/Projection assumptions) or for one of the six Reserve Calculation slots on the Assumption screen (Statutory, Tax, Second through Fifth Reserves).
Any application using the Cells can only refer to assumptions in one of these seven places which are each linked to a single calculation purpose, and there is only limited ability to temporarily change the purpose of any given set (i.e. using the Keycode field).
Furthermore the underlying storage requirements of the Cell are driven by the seven purpose screen design regardless of the number of sets actually needed or used.
Under pending financial reporting frameworks such as IFRS for Insurance Contracts, it will be important to store and refer to three or more sets of assumptions to capture the more complex presentation and calculation requirements of the new frameworks. In fact certain US GAAP reserve requirements already demand more than one set of actuarial assumptions to handle all the components of the reserve.
In addition, the Cells in AXIS are being used in new ways such as the stochastic processing functionality controlled by the Block and Embedded Block objects that may require additional sets of assumptions to be more conveniently specified.
This new approach to the Cell design addresses all these challenges.
How have the Cells Changed with Assumption Sets?
The fundamental change with Assumption Sets is the relocation of the seven existing sets of actuarial and related assumptions as shown in the screen shot above to a new flexible sized storage location within the Cell. This results in the Cell having a fixed size main screen for all of the Cell assumptions other than the actuarial assumption sets, plus a new expandable area for all of the Assumption Sets. This is just a relocation and does not by itself change any of the Cell fields in appearance or function.
However, now the user can choose to have up to 99 separate sets of actuarial assumptions in all Cells of the Dataset, if they are needed. To manage the Assumption Sets, the user must predefine at the Dataset level how many Sets are required, how each Assumption Set is named, and the order in which they are stored and displayed. The Assumption Set names and sequence numbers are how the system links each Assumption Set to the calculation purpose such as for the inforce projection or for a specific reserve calculation.
Accordingly the Cell Screen interface has been changed to reflect the use of named Assumption Sets that can be independent of purpose, and the definition of which Assumption Sets are used for each of the seven possible calculation purposes inside each Cell. We also took the opportunity to rearrange the order of all the actuarial and related assumption fields that may differ by calculation purpose to be more natural and intuitive (e.g. grouping all mortality related assumptions together).
What has Not Changed in the Cell
Please note that the introduction of these new named Assumption Sets does not by itself result in any change in the way assumptions are specified or applied. Reserve and projection calculations should produce exactly the same answers in the absence of any other new enhancement or bug fix that may impact your models using the new release with the Assumption Sets included.
Also while there is the flexibility to define more than 7 sets of independent assumptions in the Cells, there are still only 6 Reserve Sections, the same number as before, which means that any Batch Cell recalculation will only be able to produce 6 independent policy based reserves during that run, and that if more than 6 reserves are required, you will need to set up two or more batch runs and apply Cell Overrides to switch either or both of the Reserve Method and the Assumption Set for the subsequent recalculations. This is however a simpler process than it was.
How is DataLink Impacted by Assumption Sets?
DataLink can be used for the following three stages of a production process:
- To load inforce business information, map individual policy records to specific Cells, and export that seriatim data to the corresponding Cells.
- Optionally, to define within the seriatim data specific assumptions that are policy specific and thus override or adjust information defined at the Cell level.
- Optionally, to create full definitions of Cells and their assumptions, for the purposes of updating or recreating the Cells in the Dataset.
The first function of attaching basic seriatim data to the Cells is unaffected by Assumption Sets, but the second and third functions above may be impacted a) if seriatim optional fields are used within the seriatim data to specify actuarial assumptions for one of the seven calculation purposes record by record, or b) if the Cell assumptions defined in the Cells Table in DataLink includes any actuarial assumption fields that are relocated by the introduction of Assumption Sets.
For practical reasons, it will be necessary to set up and maintain a separate list of Assumption Sets within DataLink, which is independent of the list in Dataset Parameters that controls how Cells link to their assumptions. This list will be created on conversion but must be maintained separately.
More information on how Assumption Sets will impact various DataLink processes can be found in Knowledge Base article 2001 "Actuarial Assumptions - Special Considerations For DataLink" which is referenced below.
Will Datasets Convert Automatically to the New Release with the Assumption Sets?
Yes. The process of opening a Dataset with a new release of AXIS that includes the new Assumption Set design will result in a conversion of the Cell definitions and of the Cells Table and Policy Info table structures in DataLink. All references to actuarial assumptions that are impacted by the new Assumption Sets will be automatically converted to refer to the new assumption field locations. This should allow production batch processes that load seriatim data and generate reports by running these Cells to produce the same results as prior to the conversion. This will include Optional Fields and Tables in seriatim data and Cells table fields in DataLink that are used to update Cell assumptions if these fields are already populated at the time of conversion.
Users will however need to carefully check any model maintenance procedures that rely on information outside of AXIS objects as these cannot be converted by AXIS. To the extent that users want to reorder and expand the numbers of Assumptions Sets in the Cells, once the Dataset has been updated to the new release, this may also require further maintenance work by the user.
We also strongly recommend that each company adopt a standard approach within the organization for the numbering and naming of Assumption Sets, and ensure that all Datasets that may need to be consolidated with or transferred to other Datasets have a consistent definition of Assumption Sets, which is necessary before AXIS will combine Cell data. It will also be necessary to separately maintain the list of valid Assumption Sets.
GGY has developed more comprehensive documentation and guidance on the implementation of Assumption Sets in AXIS including a set of demonstration videos and training sessions on how to prepare for the change. You will find a page dedicated to Assumption Sets Resources on our website.