We’ve moved! Click here to view our most recent content!

Set Regional Preference
Required Field*
Set geographic preferences to highlight topics of greatest interest of you, written in your base currency.


Aon Retirement and Investment Blog

Life Expectancy Assumptions for Pension Plans – More Volatility Ahead?

Life expectancy is one of the key assumptions used when measuring pension liabilities. This assumption is used for calculating statutory requirements (including cash contributions and Pension Benefit Guaranty Corporation premiums), company balance sheet liabilities and income statement charges, and plan funded status measures for dynamic pension investment strategies. There have been significant discussions around this assumption for pension financing in recent years, as the Society of Actuaries (SOA) has published new research on how life expectancy in the U.S. has changed over time.

Updates From Society of Actuaries
In early July 2016, the SOA indicated that it expects to issue an updated set of life expectancy assumptions for U.S. pension plans in late October 2016. The SOA’s model for estimating future changes in life expectancy is based on broad U.S. population data, and can potentially be used for both private sector and public sector pension plans. While life expectancy assumptions were updated fairly infrequently in the past, the SOA plans to issue such updates on an annual basis going forward. This represents an evolution of U.S. practice toward that seen in some other countries such as the U.K.

Based on an Aon Hewitt analysis of new data released by the U.S. Social Security Administration (SSA), the SOA’s updated assumptions (Scale MP-2016) will likely show lower projected improvements in life expectancy than the previous assumptions (Scale MP-2015). In the near term, the updated assumptions may improve plans’ measures of funded status for purposes such as corporate accounting or monitoring de-risking glide paths. However, the underlying data illustrate significant volatility in year-to-year death rates for the general U.S. population, making it difficult to discern whether these lower longevity improvements are a temporary “blip” or a longer-term trend. To better manage volatility in plan financials, plan sponsors may want to take a longer-term view in assessing the new data and determining how to respond to it.

New Data Released by Social Security Administration
In late June 2016, the SSA released its 2016 Trustees’ Report, along with the assumptions it uses to project the solvency of the U.S. Social Security system. This information included new data on death rates for the general U.S. population for calendar years 2012 and 2013. The data show higher-than-expected death rates among retirement-age individuals, particularly those in the Baby Boom generation, continuing a pattern seen over the past few years. This is a slower improvement in longevity than anticipated, but in general, life expectancies are still increasing. 

The SOA’s model for projecting future life expectancy is designed to reflect recent experience in the near-term, and converge to a long-term improvement rate over time (a convergence period of 20 years was assumed for Scale MP-2015). Lower improvements in life expectancy observed during 2012 and 2013 would be assumed to persist over the near term, and then gradually converge with the long-term assumption. For a typical pension plan, updating the SOA’s longevity improvement model to reflect the recent SSA data is expected to reduce plan liabilities, although the size of the impact will vary based on plan demographics, plan design, and the underlying assumptions used in the model.

Volatility in Life Expectancy – The New “Normal?”
A reduction in plan liabilities based on the new SSA data may help offset increases in liabilities due to other factors, with discount rates testing historical lows and global capital markets exhibiting uncertainty in the wake of Brexit. However, the recent SSA data highlight the underlying volatility in year-to-year longevity improvement. This volatility has led to lower-than-expected improvements in recent years, but also could result in higher-than-expected improvements (and higher liabilities) in future years. This volatility can be clearly seen in the following graph of actual year-to-year changes in death rates for 65-year-old males, as compared to the SOA assumptions. (Note: an upward spike in the graph represents a more significant decrease in death rates, and a faster increase in life expectancy, while a downward spike represents a less significant decrease in death rates and a slower increase in life expectancy.) 

While the SSA has only released data through 2013, more current data from the Centers for Disease Control and Prevention (CDC) can provide insights on experience since then. Based on an Aon Hewitt analysis of final CDC data, longevity improvements for retirement age individuals were much higher in 2014 than in 2013, reversing the recent trend. Final data by age are not yet available for years after 2014, but preliminary data appear to show aggregate longevity improvements that were lower in 2015, and then higher again in the first quarter of 2016. This volatility makes it difficult to draw conclusions about longer-term trends from short-term experience.
In its July 2016 update, the SOA’s Retirement Plans Experience Committee indicated that it “continues to assess the committee-selected parameters, and will revise them, if appropriate, to reflect recent experience and future expectations and to enhance the year-to-year stability of its mortality projection model.” One way in which the SOA could potentially improve the stability of its model would be to use a somewhat shorter convergence period (e.g., 10 or 15 years rather than 20 years). However, more significant changes to the SOA model would likely require an exposure draft and comment process before being finalized.
Next Steps for Plan Sponsors
U.S. plan sponsors will need to work with their actuaries to understand the potential impact of Scale MP-2016 on their pension liabilities in planning for 2016 year-end. Companies with fiscal year-ends (or mid-year remeasurements) occurring before the SOA issues Scale MP-2016 should consider whether (and how) to reflect the new SSA data in their assumptions, since reflecting the data could be viewed as a better estimate. If a company reflects the new SSA data in its pension liabilities before the SOA issues Scale MP-2016, it also should consult with its auditors to determine whether this constitutes a precedent that may need to be applied consistently going forward.
In addition to understanding the impact of the new SSA data, plan sponsors should consider whether the use of a shorter convergence period or other modifications to the SOA model may be appropriate to reduce volatility in corporate accounting results and measurements of plan funded status for monitoring de-risking glide paths. It can be difficult to determine when short-term changes in longevity improvements are indicative of longer-term trends. As a result, some plan sponsors may not believe such changes are an appropriate driver for changes in asset allocation.
Eric Keener is a Partner and Chief Actuary of Aon Hewitt’s U.S. Retirement practice, and is based in Norwalk, CT. Grant Martin is an actuarial consultant in Aon Hewitt’s U.S. Retirement practice, and is based in San Francisco, CA.

The information contained above should be regarded as general information only. That is, your personal objectives, needs or financial situation were not taken into account when preparing this information. Accordingly, you should consider the appropriateness of acting on this information, particularly in the context of your own objectives, financial situation and needs. Nothing in this document should be treated as an authoritative statement of the law on any particular issue or specific case. Use of, or reliance upon any information in this post is at your sole discretion. It should not be construed as legal, tax or investment advice. Please consult with your independent professional for any such advice. The information contained within this blog is given as of the date indicated and does not intend to give information as of any other date. The delivery at any time shall not, under any circumstances, create any implication that there has been a change in the information since the date of publication, or any obligation to update or provide amendments after the original publication date. The blog content is intended for professional investors only.

Share:Add to Twitter Add to Facebook Add to LinkedIn   Print