IRS Finalizes Mortality Tables for DB Plans

The IRS has finalized the regulations proposed in 2007 regarding the mortality assumptions used to determine present values for minimum funding purposes for defined benefit (DB) pension plans under the Pension Protection Act of 2006. There are no surprises in the final rules for the standard mortality tables (see Watson Wyatt Insider, July 2007). Despite comments urging the IRS to make some changes, the rules for plan-specific mortality tables remain mostly the same, too, except that the IRS expanded the study period for making a credible case for using substitute tables from four to five years.

Sponsors must use the applicable tables for 2008 and later plan years.

Standard Mortality Tables
Separate tables are required for annuitants and nonannuitants (i.e., before and after distributions begin). Sponsors may use either a static projection or a fully generational projection to reflect mortality improvements, and, unlike with other assumption options, sponsors may change the projection for each valuation without getting IRS approval first. A combined table option is available for plans with 500 or fewer participants on the valuation date.

The static tables are based on the RP-2000 mortality tables projected forward with assumed mortality improvement to the date of the actuarial valuation and further projected (by seven years for annuitants and 15 years for nonannuitants) to approximate the fully generational table. The fully generational table projects mortality improvement for each participant for each year he or she is assumed to survive. Scale AA assumptions from the RP-2000 mortality tables are used to project future mortality improvement.

It is anticipated that the RP-2000 table with mortality improvement in accordance with Scale AA will be used to update the static tables at least until 2013. After that, a different table might be adopted.

Substitute Mortality Tables
Sponsors may use their plan’s own mortality experience to devise substitute tables as long as they get approval from the IRS first. The substitute tables must be generational and gender-specific.

Before using substitute tables, the plan sponsor must demonstrate “credible experience,” which the IRS defines as at least 1,000 deaths within a two- to five-consecutive-year study period. The last day of the study period cannot be less than three years before the first day of the first plan year at issue. In response to comments, the final regulations extend the maximum study period from four to five years.

Sponsors also must demonstrate that other groups for which substitute tables are not being requested lack credible experience.

Generally, for a plan to experience 1,000 male deaths over five years, it must have roughly 10,000 or more male annuitants. The plan would need even more female annuitants, thereby implying a total plan participant population of at least several tens of thousands. The requirement significantly restricts the number of plans whose sponsors could even try to make a case for using a substitute mortality table.

Many commentators suggested allowing a smaller study population, such as 1,000 annuitants or more, but requiring the sponsor to demonstrate significant actuarial gains or losses (as measured against the standard IRS table). However, the IRS indicated that such a demonstration would not be sufficiently credible and rejected the proposal.

If a sponsor’s plan population has higher-than-average mortality but the sponsor cannot use its own experience, the plan’s year-to-year actuarial gains will (eventually) reduce funding requirements. The funding liability will always be overstated, however, and at best the plan can only partially offset this excess funding. Moreover, liabilities will be permanently overstated in determining the Pension Benefit Guaranty Corporation’s variable-rate premiums, which will result in premiums higher than warranted by the plan’s experience. No mechanism exists for recouping overstated variable-rate premiums.