Volume 39 | Issue 121
Download this For Your Information as a printable PDF
The Society of Actuaries has released updated projection scales for use with the mortality tables it released in 2014 and 2015. The revisions reflect actual 2012 and 2013 mortality data and 2014 estimated mortality data instead of previously assumed projections for those years. The revised data will typically produce lower liability estimates. Plan auditors will require that these revisions be considered in setting assumptions for plan and corporate financial statements.
Background
In 2014, the Society of Actuaries (SOA) released mortality tables (RP-2014) based on a multi-year mortality study of participants in uninsured pension plans in the US. The release included MP-2014, a mortality improvement scale (quantifying how fast mortality rates are expected to change in the future). See our
October 27, 2014 FYI Alert for a discussion of the original release.
The SOA had planned to update the improvement scales on a triennial basis, but in 2015, Scale MP-2015, was created using two additional years of data applied to the 2014 model that had been used to create the MP-2014 scale. MP-2015 is intended to be used along with the RP-2014 base mortality table. The additional data became available from the Social Security Administration after Scale MP‑2014 was released. To decrease the lag time between the release of data and updates of the scale, and to produce more stable financial results, the SOA committed to releasing annual updates in the future. See our October 8, 2015 FYI Alert for a discussion of the first update.
SOA Releases MP-2016 Mortality Improvement Scale
Consistent with this commitment, the SOA has released mortality improvement scale MP-2016. As promised, the scale updates MP-2015 (and MP-2014 before it) to reflect data available from the most recent Social Security Trustees Report (Summer, 2016). In addition to 2012 and 2013 mortality rates provided with the Trustees Report, SOA was able to work with the Social Security Administration to calculate estimated 2014 rates on a consistent basis. This means the new study includes three additional years of US mortality data.
The MP-2016 scale also includes changes to the assumptions used to project future mortality improvement in an effort to increase year-over-year stability.
Reduction to Liabilities Expected
Mortality improvement during the three years added to historical data has continued to be lower than over the first ten years of the millennium. Based on present value factors shown in the study, it appears pension liabilities could decline about 1.5% to 2.0% compared to projecting with Scale MP-2015. The decline should be less for account balance designs (e.g., cash balance, pension equity, etc.) that pay mortality-independent lump sums, and the decline could be more for plans with large blocks of old, high liability retirees.
Plans and sponsors that have been using alternative assumption sets within the 2014 model, based on either MP-2014 or MP-2015 historical data, can now update for at least three years of additional information. In general, we anticipate small declines in liabilities, though that is highly dependent on the assumptions selected within the model. Pension plan sponsors and administrators will want to consider the information contained in the MP-2016 study relative to assumptions for employer and plan accounting going forward. It is likely, for example, that auditors of December 31, 2016, disclosures will bring up the mortality projection issue in light of the additional data.
We still do not know when or how the IRS will take into account data from the RP-2014, MP-2014, MP-2015 or MP-2016 studies. However, indications still point to a 2018 effective date for proposed changes as noted in our September 6, 2016 FYI Alert on IRS’ publication of the 2017 mortality table for valuations and minimum present values. We continue to monitor the situation closely and will provide updates as available.
More Than Mortality
Although the new Scale MP-2016 will temper the projected rate of change overall, using the latest SOA tables will still increase measured plan liabilities over those that reflected earlier mortality studies (RP-2000 and before). However, appropriate changes in other demographic assumptions may also be considered. For instance, in response to longer life expectancy statistics and the longer period of time retirement savings must stretch, many employees may plan to continue working beyond the plan’s “normal” retirement date instead of choosing retirement in their late 50s or early 60s. Aligning plan retirement assumptions with this new paradigm can potentially offset some of the increase due to the new mortality assumption. This is particularly true for retiree medical plans and pension plans with suspension-of-benefits provisions and generous early retirement subsidies.
In Closing
It is worth repeating that the ultimate cost of a defined benefit plan depends on the actual amount of benefits paid to plan participants. Reflecting longevity improvements in a plan’s mortality assumptions allows the actuary to provide a more reasonable estimate of future costs to avoid unnecessary surprises but doesn’t change the actual cost. Defined contribution plan sponsors will need to take notice of longevity improvements too, and then take steps to help employees plan for the additional years they will have to enjoy — and finance — after retirement.