Lower Social Security Benefits Reduced Mortality

04/01/2003
Summary of working paper 9197
Featured in print Digest

Retirees with smaller Social Security benefits had a lower age-specific mortality rate than retirees with more generous benefits.

A considerable body of research suggests that those with lower incomes have poorer health and higher mortality rates than higher income individuals. In 1996 when an advisory commission found that the Consumer Price Index overstated the growth in prices by about 1.1 percentage points per year, and therefore recommended that federal programs -- such as Social Security -- take into account this over-indexation for inflation by reducing benefits, critics cited that body of research. Witnesses at a Congressional hearing on the matter suggested that this would raise mortality rates among the elderly by making them poorer and thus less healthy. They painted a grim picture of the elderly on fixed incomes, forced to choose between purchasing food or prescription drugs.

The problem with this research, though, is that income and health are jointly determined. For example, the healthy may find it easier to earn more money. Those with lower incomes may not be able to afford proper care and thus may have other financial troubles that weigh on their health. So, researchers have found it difficult to figure out which way the causation goes in this correlation between income and life expectancy.

In The Impact of Income on Mortality: Evidence From the Social Security Notch (NBER Working Paper No. 9197), Stephen Snyder and William Evans explore a way to get around this puzzle. They compare the mortality rates of two groups of elderly males affected by a major change in the Social Security laws which arbitrarily trimmed the pensions of later retirees compared to those before them. To the surprise of the authors, they find that those later retirees with smaller Social Security benefits had a lower age-specific mortality rate than retirees with more generous benefits.

Concerned with rapidly rising costs, the federal government changed the way that benefits were calculated for new beneficiaries in 1977. This substantially decreased the size of payments for recipients born after January 1, 1917. As a result of these changes, two people with identical earnings histories but different birth dates would receive substantially different retirement incomes. Those born after what is called the "Notch" had little time to adjust since the changes happened late in their work lives. Most did not even realize the impact of the law's changes on payments until after they retired.

Snyder and Evans compare the five-year mortality rates after age 65 for those born in the fourth quarter of 1916, just before the Notch, with those born in the first quarter of 1917. To the surprise of the authors, they find that those younger retirees with smaller Social Security benefits had a lower mortality rate than retirees with more generous benefits. Since there is little difference between the cohorts except their Social Security income, the authors attribute this difference to the lower incomes generated by the "Notch." The authors test this counterintuitive result by examining the mortality rates for women from the same cohorts. Most women from these birth cohorts receive Social Security benefits as a result of their husbands' contributions to the system, and there is little difference in Social Security earnings across these groups. Therefore, there should be no difference in mortality across these groups, which is exactly what the authors find.

So, why did the poorer retirees live longer? Snyder and Evans find that smoking patterns do not explain the higher mortality rate for the higher-income retirees. The younger cohort, those born after the Notch, responded to lower incomes by increasing the amount of their post-retirement work by 5 percentage points more than those born earlier; there was a large increase in work after age 67. Some probably returned as part-time workers, often in different industries, sometimes at reduced wages from their primary career employment. "This work could have positive health benefits if the work keeps the seniors connected to the community and reduces social isolation," the authors speculate.

-- David R. Francis