The CMI’s Mortality Projections Working Party has evaluated the P-Spline and Lee-Carter methodologies for projecting mortality

The CMI is not seeking approval for the methodologies, which we regard very much as ‘Work in Progress’. We certainly do not feel that we have the answer in this area, and it is only by exposing the work fully to the profession that the benefits and features of the methodologies will become apparent. Neither do we rule out the use of other stochastic or non-stochastic approaches. However, it is not the CMI’s role to set standards in this area.

From 2005 to 2008 the CMI made available software with the aims of illustrating these methodologies, assisting the profession to debate the issues and to help users validate the results of their own work. The CMI would like to express its thanks to James Kirkby and Dr Iain Currie of Heriot Watt University for developing this software.

The software uses the statistical package R, together with the R(D)COM interface. The software was withdrawn for new users early in 2009 after the CMI became aware that it is not compatible with newer versions of R and/or the interface. Registered users of the software were notified in January 2009 that they may need to retain existing versions of R and the interface for the software to remain functional.

The software is now available on request to anyone able to locate appropriate versions of R and the interface

To obtain the software please complete the Software Request Form and return it to the CMI.

The software requires the statistical package R to be installed together with the R(D)COM interface which allows Excel to call R. Please note that the software will not function on Excel 97 or earlier versions as these do not contain some of the VBA commands required by the software.

The software is of an experimental nature and is not commercial grade

It is made available on the understanding that you agree to the following conditions:

  • the software is illustrative and the CMI accepts no liability for its use
  • no support can be provided for the software
  • it is the user’s responsibility to source versions of R and the interface and to ensure that the software functions correctly with these
  • the software has been made available for your use only. Colleagues or others wishing to use the software should use the form to seek their own permission to use it
  • agreement to abide by the terms of the GPL licence.

In using the software we request that you communicate any errors or issues you discover to the Life Office Mortality Committee by emailing at the earliest opportunity. The CMI are keen that results produced by the software are discussed with the Committee, before any public comment is made. We hope you will feel able to abide by this request.

Instructions for installing the software

The software is provided in zipped format. You will need to copy all the contents into a single directory. Explanatory notes are contained in the User Guide and the GPL licence is contained in “GNU General Public License.doc”. The software itself creates an add-in to Excel. A spreadsheet is also supplied containing data for male assured lives from 1947 to 2005, for ages 20 to 90 and for all durations combined.

The latest version of the software was released in November 2007 and contains a number of changes to the previous version issued from April 2007, notably correction of an error relating to deviance residuals in the Lee-Carter functionality. Read the Exposure note for more information.

In addition the exposure data supplied with the software is central exposure and does not require adjustment before use (previously initial exposures were supplied).

All the changes from the previous version are documented within the User Guide. Registered users are automatically being supplied with the updated version.

Related documents

Contact Details

If you have any questions about the CMI please email