Data Used to Estimate U.S. Companies' small impact because early retirees have only a few years of eligibility for this benefit. Once the retirees reach age 65, we assumed, companies pay the same for them as they do for those who retire at 65. First-year contributions could range from a low of $26 billion to a high Changing only the length of the amortization period in the unit credit method caused our estimate of first-year contributions to vary by $7 billion (in 1988 dollars) around the baseline value of $32.4 billion. The shorter period-15 years—would raise annual contributions by 15.7 percent (to $37.5 billion), while the longer period-35 years—would reduce annual contributions by 5.6 percent (to $30.6 billion). Table IV.1: Sensitivity of Estimated Total Liabilities of $402 Billion to Different Values for Selected Variables (1988) Appendix V GAO's Method for Estimating 1988 Retiree Health Liabilities and Annual Contributions to Begin Advance Funding Our estimates of employers' 1988 total and accrued liabilities for retiree health benefits were based on a model of companies' projected benefit payments. The data sources and assumptions underlying the model are set forth in appendix IV. We calculated retiree health liabilities by 1. projecting annual benefit payouts for active workers and retirees, using assumptions about mortality and future increases in medical care costs; 2. converting these benefit payouts to present values at an assumed discount rate; and 3. selecting a pension-funding method to estimate the accrued portion of total liabilities and the contribution that would be needed as of 1988 to begin advance-funding the benefits. Projecting Numbers of Using data on pensioners and active workers with employer-sponsored Retirees and Costs for Table V.1: Distribution of the Initial group health coverage, we constructed populations of retirees and work- We then projected each cohort into the future, using age-specific probabilities of death or mortality rates. For example, of the 2 million retirees age 65-69 receiving health benefits in 1988, we estimated that about 45,000 would die within a year. This left about 1.95 million alive in 1989. The number of expected deaths was based on mortality rates from a table commonly used by actuaries in pension valuations (see app. IV for a discussion of this table). We continued the process into future years until there were no more survivors from the cohort. GAO's Method for Estimating 1988 Retiree Table V.2: Projected Employer Health As an example, table V.2 shows every fifth year of the projection for the 65-69 cohort. Benefit Payouts for Present Value of Calculated Employers' average annual cost to provide health benefits for a retiree Employers' total liabilities for retiree health benefits are presented in For example, table V.3 shows the results of this conversion for every fifth year of the 65-to-69 cohort projection. We totaled the payments for each year (the last column of the table) and all intermediate years to arrive at the present value in 1988 of $28 billion for the health benefits expected to be paid to this cohort. Table V.3: 1988 Present Value of Future Benefit Payments for Retirees Ages 6569 in 1988 (1988-2023)
The average employer annual health cost for cohorts of retirees younger To reach the present value of $99.5 billion for future health benefits for the 6.6 million persons already retired in 1988, we estimated the initial per capita early and normal retiree health costs, and the initial numbers of retirees. Also, we used certain mortality, medical inflation, and interest rate assumptions. Were alternate values of health costs, numbers of retirees, or assumptions used, a different present value would result. Age of Retirement All of the $99.5 billion would be accrued as of 1988, under the projected unit credit funding method that we used. Accrued liabilities represent amounts attributed by this method to the past service of workers. Because retirees' service is all in the past, the present value of their future health benefits is 100-percent accrued, by definition. Calculating employers' liabilities for future retiree health benefits of active workers required us to make assumptions about the ages at which workers would retire. In pension plans, actuaries often assume one or two ages of retirement, even though they know that workers retire at several ages. These ages—often the plan's early and normal retirement ages-produce present values reasonably close to those that would result from several retirement ages.
Benefits Projected for Table V.4: Projected Benefits of the Because retiree health costs differ before and after age 65, we selected two retirement ages for use in our model: 60 and 65. Using CPS data on the age at retirement of private pensioners, we assumed for our baseline estimates that 60 percent of workers would retire early at age 60 and 40 percent at the normal age of 65. For each 5-year cohort of active workers in the U.S. in 1988, we estimated that 60.5 percent of those retiring at age 60 and 58.7 percent of those retiring at age 65 would be covered. According to our estimates, the cohort ages 45-49 of 3.3 million workers was divided into 1.3 million workers expected to retire at age 65 and 2.0 million expected to retire early. As with the retiree cohorts, we applied mortality rates to estimate the future survivors from these subgroups. We then multiplied the number of future retirees by our projected average health costs to determine the benefit payouts for each of these groups. Finally, we calculated the present value of the projected benefit payouts. As an example, table V.4 shows every fifth year of this projection for the 45-49 age cohort. Accrued Liabilities To determine what portion of the total 1988 liabilities for active workers age 40 and over was accrued, we prorated the liabilities for each cohort by the ratio of average service since age 40 to expected service at retirement. (This proration is commonly employed in the actuarial funding method we used, the projected unit credit method.) |