Abstract
In the previous chapters we have looked for causes of differential death rates among various groups. These differences can have effects on public policy issues. In this chapter we consider the implications of differential death rates on the variation in expected private rates of return on individual and group “investments” in the Social Security System. These investments take the form of Social Security taxes paid (by employer and employee) while the returns take the form of old-age benefits received by the individual and sometimes a current or former spouse. We use the Retirement History Survey again (as in Chapters 2 and 5), which has been matched to Social Security records for each respondent. While the RHS is a random sample of household heads aged 58 to 63 in 1969, it may not be typical of other cohorts because of the rapid increase in the maximum taxable ceiling on covered wages, much of which occurred after 1970. However, the results are indicative of the impact of differential death rates on the rates of return to Social Security.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Until recently the rate of return on private annuities was small, e.g., a 2 percent nominal rate. More recently, Keogh’s, IRA’s and supplemental pension plans have been paying rates comparable to those offered on equally risky assets on a pre-tax basis while these annuities allow you to defer taxes. Much more money is invested now in these tax-deferred annuities.
This is called the “earnings tax.” The earnings are also subject to Social Security, income, and payroll taxes. Private annuities are not subject to the earnings tax though the interest component is subject to the income tax. In our sample period, the minimum age to waive the earnings tax was 72 and the earnings tax rate was 50 percent for amounts above about $5,000 (with annual variation).
Average lifetime earnings are calculated as the primary insurance amount (PIA). For example, 1990 the formula for PIA was based on a 20 year average monthly covered wage received from 1951–1990, but excluding from the average the years of lowest earnings.
Eligibility required 20 years of marriage beginning in 1956, and there was no eligibility before then. The widow’s benefit also began in 1956.
These percentages apply at age 65. Actuarially reduced benefits are given if the woman retires early, which can be as early as 60 for widows and 62 for other women.
We assume, for simplicity that retired people don’t work.
For all years after 1950, we have exact figures on taxes paid in each year through 1979. We have the total of taxes paid prior to 1951 and we interpolated these data to obtain annual figures. To estimate pre-1951 earnings, we selected heads of household born in 1904–13 from the CPS-IRS-SSA exact match study. For these individuals, we regressed 1937–50 earnings on 1951–76 earnings (separate regressions by race/gender). Using these coefficients, we get an estimate of pre-1951 earnings for each RHS head of household.
These results assume that the people in the RHS in 1969 are a random sample of the people in their birth cohort. The results in Chapter 5 on the Dorn sample suggest that there is selective mortality before age 58. However, since mortality rates are sufficiently low for this age group this should not greatly affect our results.
We use the 1977 rule for divorcees that the marriage had to last at least 10 years. For female heads of household who were widowed or divorced, we calculate that 30 and 37 percent respectively were receiving benefits based on their own PIA. In this calculation we assume the benefits are not based on one’s own history if they exceed by 5 percent or more what we calculate is due based on Social Security earnings’ records.
We can match by age and education the married women who do not draw the 50 percent of husband’s benefits to the widows in 1969 who draw their (82.5 percent) benefits. This gives us an estimate of the extra taxes paid (or investments made) by a couple. Problems with such a procedure include that some widowed women fall into the interval between the 50 and 82.5 percent differential, that some women who expected to be widows may have deliberately altered their labor force behavior, and those widowed young may have rejoined the labor force. The proportion of women heads of households receiving old-age benefits that were based on former husband’s PIA was 60 to 70 percent About 14 percent of widows and divorcees of both races would have had between 50 percent and 67 percent of actual benefits if they had used their own primary insurance amount.
Some members of the sample did not have twenty years of covered earnings after 1950. For these individuals we estimated the labor force experience missing from the twenty years necessary to use the standard PIA formula. We based earlier earnings on Social Security taxes paid by that individual before 1951.
While the inflation adjustment only comes into effect when the annual inflation rate exceeds 3 percent, the one time recently when this rate was less than 3 percent, Congress adjusted benefits anyway.
Again, defined as deaths before drawing first benefits.
We have not included in Cj the Social Security taxes paid by former spouses. We have included benefits based on former husbands’ PIA.
Recall that female heads of household do not include currently married women with non-institutionalized husbands.
Never married obviously rely on their own earnings.
Some idea of the magnitude of this bias from not including taxes paid by married women can be obtained from available data on widowed women who as head of household in 1969 were matched to Social Security earnings files, and who were drawing Social Security benefits based on their former husbands’ earnings. Including taxes paid as calculated from the widows’ records on earnings in covered employment, the rate of return using standard mortality is estimated at 8.2 and 9.2 percent for whites and non-whites. Thus, the either/or provision adds substantially to the r for married men even when “wasted” investments of spouse are accounted for.
Note that we use slightly different educational levels for whites and non-whites to have adequate sample sizes for non-whites.
Note that minimum PIA is sufficiently low that few observations are affected.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1998 Springer Science+Business Media New York
About this chapter
Cite this chapter
Behrman, J.R., Sickles, R.C., Taubman, P. (1998). Private Rates of Return on Social Security and Their Relation to Mortality for Groups Defined by Socioeconomic Characteristics. In: Causes, Correlates and Consequences of Death Among Older Adults. Springer, Dordrecht. https://doi.org/10.1007/978-94-011-4393-6_6
Download citation
DOI: https://doi.org/10.1007/978-94-011-4393-6_6
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-010-5887-2
Online ISBN: 978-94-011-4393-6
eBook Packages: Springer Book Archive