Thursday, December 14, 2017

Changing lifespans make funding 

assumptions hard to predict

Mortality rates, that morbid annual measure of deaths, continue to decline in the United States. However, progress hasn’t been steady, according to a recently released research brief by the Center for Retirement Research at Boston College.

The mortality rate is the percentage of people who die each year. It is one of the most critical assumptions in the analysis of the sustainability and the assessment of appropriate contribution rates of public pension funds as well as for the U.S. Social Security system. Many in the general population don’t like to think about when they will die. But measuring the number of deaths in populations and age groups is common talk for trustees and administrators of public employee pension plans and their actuaries who analyze the financial costs of risk and uncertainty. A slowing decline has many wondering what future lifespans could look like and how that will impact the future of their pension funds.

In the U.S., roughly 823 people die per 100,000 population.  Americans can expect to live for a little more than 78 years, according to the U.S. Centers for Disease Control and Prevention. Some studies put the age at 79. Mortality rates of improvement – the increase of life expectancy – fluctuated on average around 1 percent per year between 1900 and 2016, according to U.S. Social Security Administration data. The keyword is “fluctuate.”

“Sometimes mortality rates decline very rapidly, and sometimes they decline slowly,” according to the brief titled, “What’s Happening to U.S. Mortality Rates?”

Released in September and written by Anqi Chen, Alicia H. Munnell, and Geoffrey T. Sanzenbacher, the brief’s authors examined swings in the mortality rate since the year 1900. The authors also took a closer look at the years since 1969, when more detailed data on cause of death became available.

What the report’s authors and pension fund actuaries want to know is will the current trend of improved mortality rates fluctuating around 1 percent per year continue in the future or will substantial improvements occur less rapidly.

History of Mortality

TEXPERS spoke to one of the study’s authors, Sanzenbacher, to get a better understanding of the mortality rate’s past and future.

“There are two big things to understand,” he says. “Over the long run, mortality tended to improve at a constant rate. Right now, we do see a slowdown.”

Sanzenbacher points to the development of life-extending medical drugs as well as technology; access to health care for the elderly, poor and disabled; and improvements in education and overall standard of living as boosting life expectancy. Here’s a brief history of mortality in the U.S.:
  •       Starting in the year 1900, mortality rates have dropped year after year, with the exception of the flu pandemic in 1918. Sanzenbacher and his colleagues attribute the improvements to better water delivery and waste removal infrastructure.
  •       Then, the discovery of antibiotics helped doctors treat infectious diseases. An infection that would normally kill a person was reduced to a few days of bed rest. These progresses also reduced infant mortality rates.
  •       During the second half of the 20th century, Medicare and Medicaid increased access to medical care for the elderly, poor and disabled. 
  •        And as access to education became more evenly spread across age groups, standards of living improved resulting in a continued decline in mortality. 
“It’s very obvious that life expectancy tends to go up,” Sanzenbacher says. “However, when you look at mortality rates, demographics must be considered.”

Age groups, gender and even where a person lives all play a part in judging life expectancy. Mortality rates, for example, have always been higher for men than for women. Although the gap is narrowing, Sanzenbacher says. Also, the rate of change varies significantly over time for both men and women.

“The United States is now in a downward part of the cycle,” Sanzenbacher says. “The rate of improvement has been falling for men and women in recent years.”

Rate Variations

There is also an age variation. Mortality rates improve at younger ages, which are those younger than 30 years old. Mortality improvement among that age group benefits the actuarial status of America’s Social Security program as well as pension funds. However, growth at older ages worsens the actuarial status.

The variances and demographics of mortality rates can be confusing. Another addition to the mix: rates can change state to state, even county to county. That could be reflective of places with less access to higher education and medical care.

“Mortality rates may be lower in areas where obesity is higher than other parts of the country,” Sanzenbacher says.

In fact, according to another report published on the Institute for Health Metrics and Evaluation’s website (, "babies born today in 13 U.S. counties have shorter expected lifespans than their parents did when they were born decades ago." In that study, published in May, researchers calculated life expectancy by county from 1980 to 2014. The report’s authors also considered the risk of dying among five age groups plus the extent to which risk factors, socioeconomics, and health care contribute to inequality.

“Looking at life expectancy on a nation level masks the massive differences that exist at the local level, especially in a country as diverse as the United States,” says Laura Dwyer-Lindgren, lead study author and a researcher at the IHME. “Risk factors like obesity, lack of exercise, high blood pressure, and smoking explain a large portion of the variation in lifespan, but so do socioeconomic factors like race, education and income.”

According to the report, 11.5 percent of U.S. counties had an increased risk of death in adults between 25 and 45. Inequality in the probability of dying also rose for people between 45 and 85 since 1980. Obesity, lack of exercise, smoking, hypertension and diabetes explained 74 percent of the variation in longevity. Socioeconomic factors – a combination of poverty, income, education, unemployment, and race – were independently related to 60 percent of the inequality. Access to and quality of health care explained 27 percent, according to the IHME study.

By the Numbers

Nationwide, the county with the most significant increases in life expectancy from 1980 to 2014 is Aleutian East Borough, Aleutians West Census Area, in Alaska. Life expectancy increased 18.3 percent. Residents in that county live an average of 83.73 years. The average U.S. life expectancy is 79.08 years. The county with the largest decrease is Owsley County in Kentucky. That county experienced a 3 percent decline in life expectancy between 1980 and 2014, according to the study. Residents there are expected to live an average of 70.21 years.

Texas fell roughly in the middle of all U.S. states in term of life expectancy in 2014. Texans are, on average, expected to live 78.5 years. The Texas county with the highest life expectancy in 2014 was Presidio County. People there are expected to live an average of 83.72 years, an 11 percent increase since 1980. Polk County, on the other hand, had the lowest life expectancy in 2014. People there are expected to live 72.75 years, only slightly up from an average age of 72.17 in 1980.

How it's Used

Mortality improvement must be addressed as a part of the mortality assumption either by projection to a future date or by use of generational mortality projection, says Mark Fenlaw, a consulting actuary with the firm Rudd and Wisdom Inc. It is the responsibility of a pension fund’s actuary to select an appropriate mortality assumption.

“The large statewide public employee retirement systems in Texas have enough experience to develop their own mortality tables or to develop adjustments to published mortality tables,” he says. “Most other public plans in Texas do not have enough experience to rely on completely. They generally use published tables and projection scales.”

Tables consist of a series of annual rates of dying at each age from an early age, sometimes birth, to a very advanced age, such as 120 years old. An educational and professional membership organization for actuaries, the Society of Actuaries, published a new set of tables for pension plans in 2014. However, Fenlaw says many public plan actuaries asked questions about the development of the SOA mortality tables and appropriateness for public employee pension plans. As a result, the SOA agreed to develop a new set of mortality tables by studying the experience of only public employees – yet another demographic subset.

“The study will have three sets of tables: teachers, public safety employees and all other public employees,” Fenlaw says. “These are expected in late 2018 or early 2019. All of these will have separate tables for males and for females. Of course, mortality rates vary by age.”

So, what will happen to mortality rates in the future? Improvements will continue to depend on many of the same drivers. However, experts say the effects could play out differently. That’s where actuaries like Fenlaw often come into play.

“It is not uncommon for mortality assumptions to be changed only every four or five years,” he says. “There is not a single correct assumption; so the actuary is responsible for selecting a reasonable assumption for the particular pension plan. There is an actuarial standard of practice (ASOP 35) that actuaries are to follow in selecting the mortality assumption. The ASOP is principles-based, not prescriptive.

Allen Jones
About the Author:
Allen Jones is the communications manager for the Texas Association of Public Employee Retirement Systems. Email him at or call 713-622-8018.

No comments:

Post a Comment