Tuesday, September 26, 2017

New paper: “What’s Happening to U.S. Mortality Rates?”

The Center for Retirement Research at Boston College has released a new Issue in Brief:

“What’s Happening to U.S. Mortality Rates?”

by Anqi Chen, Alicia H. Munnell, and Geoffrey T. Sanzenbacher

The brief’s key findings are:

  • Mortality rates, which determine life expectancy, are a key factor in cost projections for the Social Security program.
  • Mortality rates consistently improve over time, but the pace of progress varies by year, by age, and by socioeconomic status.
  • Over the past 40 years, progress has been driven by medical advances, better access to health care, and a decline in smoking, partly offset by rising obesity.
  • Looking to the future, mortality improvements will continue to depend on the same drivers, but the net effects could play out differently.
  • The key debate is whether the future will mirror the past, with average rates of improvement of about 1 percent, or whether the pace of progress will slow.
This brief is available here.

3 comments:

WilliamLarsen said...

No real news here. The data since 1900 is easily obtained from the SSA website: period cohort life expectancy. When you look at all the data, the largest drop in mortality rates has been infants. However, as you progress in age, mortality rates have dropped.

What the authors do not seem to realize, is that increased life expectancy at age 65 has been slowing for more than five decades. This means that life expectancy is increasing, but at an ever slower rate than the year before.

An analogy would be a car traveling down the road. Every year it gets a bit further from its starting point, but it is velocity is not constant, it is decelerating.

For example looking at the growth in life expectancy at age 67 here are the values from 1900 in 10 year increments.

year @ 67
1900 14.82583685
1910 15.47524523
1920 15.94415651
1930 16.49207827
1940 17.09119252
1950 17.7173993
1960 18.31999344
1970 18.89195108
1980 19.43954295
1990 19.9648143
2000 20.46833536

Yes, it is increasing, but at what rate?

year @ 67 Rate of change
1900 14.82583685
1910 15.47524523 4.38%
1920 15.94415651 3.03%
1930 16.49207827 3.44%
1940 17.09119252 3.63%
1950 17.7173993 3.66%
1960 18.31999344 3.40%
1970 18.89195108 3.12%
1980 19.43954295 2.90%
1990 19.9648143 2.70%
2000 20.46833536 2.52%


Putting it into perspective is this a problem?

year - @ 67 - Rate of change - days per decade - day's per cohort - % of FRA
1900 - 14.826 - 0% - 0 - 0 - 0%
1910 - 15.475 - 0.044% - 237.03 - 23.7 - 0.004%
1920 - 15.944 - 0.03% - 171.15 - 17.12 - 0.003%
1930 - 16.492 - 0.034% - 199.99 - 20 - 0.003%
1940 - 17.091 - 0.036% - 218.68 - 21.87 - 0.004%
1950 - 17.717 - 0.037% - 228.57 - 22.86 - 0.004%
1960 - 18.32 - 0.034% - 219.95 - 21.99 - 0.003%
1970 - 18.892 - 0.031% - 208.76 - 20.88 - 0.003%
1980 - 19.44 - 0.029% - 199.87 - 19.99 - 0.003%
1990 - 19.965 - 0.027% - 191.72 - 19.17 - 0.003%
2000 - 20.468 - 0.025% - 183.79 - 18.38 - 0.002%

So the the increase for those born in 2000 is about 0.25% more to provide benefits. To amortize this over the working life, how much would it take to be set aside to pay 18.38 additional days in retirement? Using $30,000 in today's dollars, and wanting it to be replaced starting at age 67 by the change in wage growth of 3% yearly and then escalated by 2.1% thereafter would require $0.03694608 additional savings per day. based on saving a set % of wage per year (as wages go up, savings go up).

year Life replace this save this 1st year
2000 87.46833536 $30,000 $7,326.51
2001 87.51868747 $30,000 $7,339.99
$0.0369461 per day

Is increased a problem for Social Security? It is negligible. This a red herring.

Arne said...

$0.037 per day is $13.50 per year is 0.045% of $30K.
Repeat that increase every year for 75 years and you are at 3.4%

You have just mathematical confirmed that the SS shortfall is primarily due to increased lifespans.

WilliamLarsen said...

Arne

The 0.25% increase in cost is the total net increase cost over the beneficiary's years drawing SS. The added payment is only over the working life time of the worker which would be age 22 to age 67 or 45 years.

The amortize net cost of 0.25% (an additional 18.38 days) over the working person life time is tiny. The net cost is not 3.4%, but 0.25% over the persons benefit period.

If you were to raise 3.4% more revenue based on your words, you would pay for many more months which is unnecessary.