1 / 21

Projecting Mortality, Different Models for Different Purposes and Views About the Future

Projecting Mortality, Different Models for Different Purposes and Views About the Future. Steve Goss, Chief Actuary U.S. Social Security Administration Longevity 12 Conference, Plenary Session IV September 30, 2016. Projections Based on Data. Extrapolating past trends of:

fsnyder
Download Presentation

Projecting Mortality, Different Models for Different Purposes and Views About the Future

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Projecting Mortality, Different Models for Different Purposes and Views About the Future Steve Goss, Chief Actuary U.S. Social Security Administration Longevity 12 Conference, Plenary Session IV September 30, 2016

  2. Projections Based on Data • Extrapolating past trends of: • Life expectancy at birth or other ages? • Improvement by cohort? • Mortality rate by age and sex? • Mortality rate by age, sex, and cause? • Use a statistical extrapolation? • Consider factors affecting mortality change?

  3. Can Life Expectancy Rise Linearly? • Requires accelerating rate of decline in mortality rates if retain age gradient • LE most affected by lowest ages—only so much gain possible • Is there an omega? Squaring the curve

  4. Extrapolation by Cohort • U.K. and RPEC • “Phantoms never die” data issues • What does change up to age x say above age x? • Is cohort healthier at x if lower mortality up to x? • Or is cohort compromised by impaired survivors? • What does one cohort imply for the next cohort? • Period effects from known conditions appear to be stronger—at least in the U.S.

  5. Mortality Decline Varies Over TimeAntibiotics/economy 1936-54; health spending 1968-82 5

  6. Extrapolation by Age and Sex • Lee and Carter • Future conditions replicate past • Age gradient never changes • No deceleration in mortality decline

  7. Projection by Age, Sex, and Cause • SSA/OCACT/Trustees Reports • Requires selecting ultimate rates of decline • Allows change in age gradient • Results in deceleration in mortality decline

  8. No Effect on Social Security 75-yr Actuarial Status Lee maintaining full age-gradient offsets lack of deceleration

  9. Canada Assumes Little Age Gradient; UK None 9

  10. Developing Assumptions by Cause • Controversial in some quarters • Trustees (SSA OCACT) develop in consultation with NIH and others • Johns Hopkins recent survey of medical researchers and clinicians came to very similar medium term expectations—independently • Trustees’ medium-term rates by cause had not been published

  11. Cardiovascular: JHU Less Optimistic than Trustees over Age 50 for Next 30 Years 11

  12. Respiratory: JHU More Optimistic under Age 50, Less Optimistic over Age 85 12

  13. Cancer: JHU Very Similar to Trustees’ Expectations 13

  14. What Future Conditions Might Change • Smoking decline for women • Started and stopped later than men • Obesity—sedentary lifestyle • Health spending—must decelerate • Advances help only if apply to all • Human limits? • Has omega changed? Will it?

  15. Trends in Obesity: US 1971-2006 Sam Preston 2010—must consider cumulative effects Increasing duration of obesity for aged in future

  16. Does Health Spending Affect Mortality?Note stability 1990-2000 and since 2009 16

  17. Interesting Correlation with Slow Mortality Improvement 1990-2000 and Since 2009 Trustees Reports have overestimated lately 17

  18. Health Spending Cannot Continue to Rise at Historical RatesNote Trustees’ deceleration 18

  19. Is There an Omega?It appears we are squaring the survival curve?

  20. Death Rates Will Continue to Decline: But How Fast and for Whom? • Must understand past and future conditions • Persistent historical “age gradient” • Death rates drop fastest at young ages • Avoid simple extrapolation of past periods • Cannot ignore changing conditions, or “limits” on longevity • Latter half of 20th century was extraordinary • So deceleration seems likely

  21. For More Information…http://www.ssa.gov/oact/ All OASDI Trustees Reports Detailed single-year tables for recent reports Our estimates for comprehensive proposals Our estimates for the individual provisions Actuarial notes Actuarial studies Extensive databases Congressional testimonies

More Related