Reforming Medicare
Appendix
Remaining lifetime spending for the 1912 birth cohort, last observed at age 85 in 1997, is imputed using the experience of the 1897 birth cohort that reached age 100 in 1997. Data are available for the 1897 birth cohort for the ages 77 to 100 and are used as follows. First, imputations are limited to members of the 1897 cohort who survived to age 85. Their real Medicare reimbursements are then discounted to age 65 using a 4 percent rate. Reimbursements between ages 80 and 85 are summed for each enrollee, as are reimbursements between ages 86 and 100. The ratio of post-85 reimbursements to the reimbursements between ages 80 and 85 is then calculated for each beneficiary who survives to age 85. Spending ratios cannot be calculated for beneficiaries who had no spending between ages 80 and 85. Also, the top and bottom 1 percent of ratios are set to the 99th and 1st percentile, respectively. These ratios are used to impute remaining lifetime reimbursements for the 1912 birth cohort by applying them to this cohort's members' summed reimbursements between ages 80 and 85.
The older cohort's ratio of spending after age 85 to spending between ages 80 and 85 is assigned to the younger cohort by first stratifying the two samples by spending quintiles and sex, then randomly matching within these cells. Spending between ages 80 and 85 for members of the 1912 cohort who survive to age 85 are ranked and divided into quintiles. Within quintiles they are divided between men and women and randomly assigned a spending ratio from the 1897 birth cohort similarly categorized by quintile and sex. The spending ratio is then used to impute remaining lifetime spending after the age of 85 for survivors to age 85.25