Home > Valuations > Valuation Assumptions > Reference Tables > Mortality Improvement Scales

Mortality Improvement Scales

Mortality improvement scales are referenced within mortality rate tables and are used to specify the changes in mortality for years other than the specified base year(s). Note that improvement is applied retroactively in the case of years prior to the base year (e.g., for entry age normal and Teilwert cost methods). This article focuses on the interface and input methods for these scales rather than the mortality improvement calculation methodology; for a detailed discussion of the calculations, see the Technical Reference article entitled “Mortality improvement (generational mortality) ”.

A mortality improvement scale has a table interface that permits improvement rates (or rates of decline) to vary by the age corresponding to the mortality rate or by both age and year of improvement (i.e., the calendar year associated with the improvement rate), in which case the table is “two dimensional” and has a column for each year of improvement. Furthermore, improvement scales can be unisex or may vary by sex.  For a two dimensional table, the improvement rate appears in the cell for the end of the year of improvement, for example, improvement from year 2014 to 2015 is associated with calendar year 2015, not 2014, and thus should be entered in the 2015 column (this follows the convention noted in the Society of Actuaries RPEC Response to Comments on Mortality Improvement Scale BB Exposure Draft, Section 7.8).

ProVal looks up table values based on age as defined by the Date of birth (or attained age) parameter of the Census Specifications. Age for table lookups is always computed as whole (not fractional) numbers. For details, see Active Data, Inactive Data - pension modes and Inactive Data - OPEB mode.

If the Table Type (selected behind the Options button) is “Age by Year of Improvement”, the Improvement Scale rates are parameter appears, to tell ProVal whether the improvement rates entered in the table are Annual reductions to the mortality rate or Cumulative factors. These options correspond, respectively, to whether an improvement rate represents a percentage per year (e.g., 0.02), which should be accumulated in a “rolling product,” or instead is a single multiplicative factor (e.g., 0.95) that encompasses mortality improvement from the base year to the year of improvement. Thus an assumption that mortality is expected to deteriorate rather than improve would correspond to annual reductions that are negative but cumulative factors that are greater than one (that is, the mortality rate in the year of improvement exceeds the mortality rate in the base year).

Annual reductions have a valid input range of - 0.2 to + 0.2 and cumulative factors have a valid input range of 0 to 5. Therefore, because the option to adjust the “Min / Max Allowed Values” (selected behind the Options button) is not accessible for mortality improvement scales, if invalid values exist when you switch between the Annual reductions and Cumulative factors options, ProVal will ask you to indicate whether invalid values should be replaced with valid ones or changed to blanks (which you can then replace with the desired values).

Under either option for the nature of improvement scale rates, the values entered for the last calendar year are used for all years thereafter and the values entered for the earliest calendar year are used for all preceding years. Note: when the rates are Cumulative factors, ProVal will use the factor entered for 2030, for example, for 2031, 2032, etc., effectively stopping mortality improvement at 2030. If, instead, you wish to reflect mortality improvement in 2031 and beyond, all potential decrement years and payment years must be included in the table (and their associated cumulative factor values entered), which may require extending the table’s years (columns). This can be done by changing the entry for Latest year under “Row & Column Limits” (selected behind the Options button) to a year far enough beyond 2030 to “cover” the last payment year of the Valuation or Core Projection. Similarly, the Earliest year of “Row & Column Limits” could be changed if needed.  

If the Table Type is "Age by Year of Improvement", the Interpolate rates from January to valuation month option will also appear.  If this option is checked, ProVal will assume that the input rates are on a calendar year basis (as of January of each year) and will be interpolated to the valuation month. 

The following mortality improvement scales are automatically included in an existing ProVal client, when the client files are updated, and in every new ProVal client. The name of each begins with the abbreviation of its issuing body (e.g., "SOA"). They may be hidden from the current Project; click the Unhide button to gain access. You cannot erase, or save changes to, mortality improvement scales that are automatically included in a ProVal client. The following tables are currently available:

  1. SOA Scale AA as published in the Transactions of Society of Actuaries 1995 Vol. 47, which presented the 1994 Group Annuity Mortality Table and the 1994 Group Annuity Reserving Table;

  2. SOA Scale BB as published in Mortality Improvement Scale BB Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in September 2012 – this scale has been suggested as an interim replacement for Scale AA for valuation systems that do not support two-dimensional rates;

  3. SOA Scale BB2D, the two-dimensional rates underlying Scale BB, as published in Questions and Answers Regarding Mortality Improvement Scale BB by the RPEC of the Society of Actuaries in September 2012 – see Q&A A3;

  4. SOA Scale D as published in the Transactions of Society of Actuaries 1971 Vol. 23 Pt. 1 No. 67, which presented the 1971 Group Annuity Mortality Table;

  5. SOA Scale H as published in the Transactions of Society of Actuaries 1983 Vol. 35, which presented the 1983 Group Annuity Mortality Table;

  6. SOA Scale MP-2014 as published in Mortality Improvement Scale MP-2014 Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in October 2014;

  7. SOA Scale MP-2015 as published in Mortality Improvement Scale MP-2015 Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in October 2015;

  8. CIA Scale CPM-B as published in the Canadian Pensioners' Mortality Final Report by the Canadian Institute of Actuaries Pension Experience Subcommittee in February 2014; and

  9. CIA Scale MI-2017 as published in the Task Force Report on Mortality Improvement (Final) ​by the Canadian Institute of Actuaries Task Force on Mortality Improvement in September 2017.

  10. SOA Scale MP-2016 as published in Mortality Improvement Scale MP-2016 Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in October 2016;

  11. SOA Scale MP-2017 as published in Mortality Improvement Scale MP-2017 Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in December 2017;

  12. SOA Scale MP-2018 as published in Mortality Improvement Scale MP-2018 Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in October 2018;

  13. SOA Scale MP-2019 as published in Mortality Improvement Scale MP-2019 Report by the Retirement Plans Experience Committee (RPEC) of the Society of Actuaries in October 2019.

  ​