Since the market meltdown of 2008, pension plans have a renewed focus on managing investment risk. However, few plans can afford to lose sight of another significant risk: longevity risk, the risk of plan members living longer than anticipated under the plan’s funding assumptions.
Despite being overshadowed by investment risk, the importance of managing longevity risk is growing, both in Canada and around the world. In fact, a U.K.-based survey by Aberdeen Asset Management found that longevity risk ranks second (after investment risk) among the most important risks that pension plans face. To ensure the long-term financial health of the plan, sponsors must understand the impact of longevity risk and find ways to mitigate it.
Why it’s an issue
In short, people are living longer. Life expectancy is increasing with improvements in living conditions and better treatments for chronic life-threatening illnesses. More importantly for pension plans, most of the increase in life expectancy at birth in Canada over the last 30 years has come from improvements in mortality at ages 65 and older. As indicated in Figure 1 (see pdf), life expectancy at age 65 has increased from 14 to 18 years for males and from 18 to 21 years for females, and it’s projected to continue increasing—meaning that it will continue to cost plan sponsors more to fund lifetime pension benefits.
Recognizing mortality improvements in pension plan valuations can result in significant added costs. For example, when the mortality basis used for solvency valuations was strengthened in 2005, many plans in Canada experienced a one-time increase in solvency liability values of 5% to 10%.
For plan sponsors, the challenge is to account for the increases in average life expectancy in the mortality tables used to value the liabilities of their plans. Mortality tables are often based on population statistics rather than on specific plan member experience. To the extent that plan member mortality differs from the general population, mortality may be overstated or understated. Certain groups may have worse mortality than the general population—for example, construction trades—while others, such as university groups, may have better mortality.
It’s important for plan sponsors to get an accurate picture of the mortality characteristics of their plans in order to control costs. If the mortality improvements over time are significant, accounting for these improvements in advance is critical. At the same time, setting aside too much for assumed mortality by being too conservative in the mortality assumptions, or by using a mortality table that doesn’t accurately reflect a plan’s mortality characteristics, may create unintended surpluses.
Quantifying longevity risk
There are two types of longevity risk that apply to all pension plans:
• systematic risk, in which the mortality assumption used is incorrect as a whole; and
• specific risk, the risk that certain individuals will live longer than expected. This is the volatility of the mortality assumption.
Specific longevity risk is a bigger risk for small plans because they don’t have the economies of scale that allow them to spread that risk over larger groups. For example, if a plan has only two retirees and both of those retirees live to age 90, the cost (loss) to the plan will be higher than expected. However, if a plan has 2,000 retirees, it’s unlikely that all 2,000 retirees will live to age 90. The realized lifetime will be more evenly spread across the life expectancy of the group, with some retirees living longer than expected and others dying sooner than expected.
Plan sponsors already get a comparison of how well the mortality assumption is aligned with the plan membership through valuation reports. As part of a valuation, actuaries provide a gain and loss analysis that reconciles the actual mortality experience of the plan with the expected experience from the last actuarial valuation. If the plan consistently experiences losses because pensioner mortality is lower than expected, it could be subject to significant longevity risk. To assess whether or not the mortality tables properly account for the plan-specific mortality, large plans often undertake a mortality study. The results of the study can then be used to adjust the mortality assumptions, if necessary.
More recently, mortality measurement services, such as Club Vita in the U.K., have been introduced. These services more finely tune the mortality characteristics by comparing plan-specific data across a number of different pension plans. Because of the large amount of data collected, they can look at where members live, their income levels and other factors to help plan sponsors better understand the mortality characteristics of their membership.
The key variable is longer life expectancy. To account for increasing longevity, the mortality tables used by plan actuaries typically include an allowance for mortality improvement. This allowance is usually based on historically observed mortality improvements and can be prescribed in some cases.
While plan sponsors may be comfortable with these mortality assumptions over the short to medium term, some are less comfortable with assumed mortality improvements over the long term. But tools are available to help eliminate or at least mitigate this risk.