NORMA B. COE, University of Washington - Department of Health Services, Boston College - Center for Retirement Research
APRIL YANYUAN WU, Boston College - Center for Retirement Research
Low take-up by elderly Americans in most means-tested federal programs is a persistent and puzzling phenomenon. This paper seeks to measure the causal effect of the benefit levels on elderly enrollment in two public assistance programs – the Supplemental Nutrition Assistance Program (SNAP) and the Supplemental Security Income (SSI) program – by using the variation in SNAP and SSI eligibility and benefit levels introduced by Social Security retirement benefits. Our findings are three-fold. First, the low take-up among the elderly is not driven by changes in the composition of the eligible pool: individuals who become eligible as they age exhibit average take-up patterns that are similar to those who were eligible before reaching Social Security benefit claiming ages. Second, Social Security has a significant impact on the use of public assistance programs among the elderly, because the increase in income decreases the potential benefits available from public programs. Third, we estimate different behavioral responses to SNAP and SSI programs: a $100 increase in SSI benefits leads to a 4-6-percentage-point increase in the probability of taking up SSI, but we are unable to estimate consistent results on how benefits impact the take up for SNAP. Together with the fact that eligible individuals who begin receiving Social Security benefits continue to participate in SSI more often than they maintain SNAP enrollment, we posit that the different estimated behavioral responses could be due to individual preferences for cash over in-kind transfers.
This analysis uses data from the Health and Retirement Study (HRS) to examine the sources of variation in mortality for individuals of varying socioeconomic status. The use of the HRS allows a distinction between education and a measure of career earnings as primary determinants of socioeconomic status for men and women separately. We use those predictions of mortality to estimate the distribution of annual and lifetime Old Age, Survivors, and Disability Insurance benefits for different birth cohorts spanning the birth years from 1900 to 1950. We find differential rates of mortality have had substantial effects in altering the distribution of lifetime benefits in favor of higher income individuals.