Dependant assumptions are set for males and females on the dialogue box show below.
Basis Category and Member Status allow assumptions to be specified for all members and dependants, or just a subset.
Mortality assumptions apply to all dependants, whether in payment at the valuation date or generated during the valuation run in respect of members who die post valuation date. The mortality assumptions are selected from the normal list.
The assumption type determines the proportion of members married at each future date that are used to establish liabilities for dependant pensions. There are six assumption types reflecting the methods used in various valuation systems the authors have worked with:
Use data means to use the dependant data stored against the record only. This might be used for example for current pensioners if the data is present, with one of the other assumption methods used for deferred and active members.
Static at payment applies the married proportion from the married assumption box (eg 0.9 to represent 90%) at the first date a benefit is paid. That means if a member dies in deferment at age 45, they are considered to be 90% married at that time. If a member dies after retirement at age 80, they were considered to be 90% married when first retired at age 65 with the relevent reduction in the proportion married implied by the mortality rates experienced between 65 and 80. If a member is already in payment at age 100 at the valuation date, this will be 90% married at the valuation date.
Rate at payment works in a simular fashion to static at payment, applying the relevent proportion married at the same time, but applies a proportion married drawn from the table selected under married assumption instead of a flat rate. The use of an appropriate table allows the proportion married for older members to be reduced according to mortality.
Static at retirement applies the married proportion from the married assumption box (eg 0.9 to represent 90%) at the first date a benefit is paid, with a mortality adjustment for the period between retirement date and valuation date in the event the member was already retirement at the valuation date. That means if a member dies in deferment at age 45, they are considered to be 90% married at that time. If a member dies after retirement at age 80, they were considered to be 90% married when first retired at age 65 with the relevent reduction in the proportion married implied by the mortality rates experienced between 65 and 80. If a member is already in payment at age 100 at the valuation date, they will be assumed to have been 90% married at their own date of retirement with a mortality adjustment made for the period between the date of retirement and the valuation date.
Rate at retirement works in a simular fashion to static at retirement, applying the relevent proportion married at the same time, but applies a proportion married drawn from the table selected under married assumption (as at age at date of retirement) instead of a flat rate. This effectively means that for older pensioners at the valuation date, the parts of the proportion married table beyond age at date of retirement are ignored.
Static at death applies the married proportion only on the death of the member. So under the valuation simulation if death occurs at age 100, the proportion married will still be that entered in the married assumption box with no adjustment for mortality.
Rate at death works in a similar fashion to static at death, applying the proportion married at the death of the member, but applies a proportion married drawn from the table selected under married assumption instead of a flat rate.
Age diff determines the number of years (which can be a decimal number) the member is older than the generated dependant. It is ignored if the use data assumption type is made. To achieve the classic assumption of males three years older than females, +3 should be used for males and -3 should be used for females.