Report | Trade and Globalization

Raising the Retirement Age: The wrong direction for Social Security

Download PDF

September 2000 Briefing Paper

Raising the Retirement Age

The wrong direction for Social Security

by Christian E. Weller The debate over Social Security has emerged as a key election-year issue, as the presidential candidates offer proposals for reforming the system or enhancing retirement security. The Republican presidential nominee, Gov. George W. Bush, has proposed fundamentally changing the Social Security program by “carving out” private accounts from the current system. Under the Bush proposal, workers would be able to divert part of their Social Security payroll tax, reportedly two percentage points, into private accounts. In order to pay for these private accounts, Gov. Bush has left open the possibility that the resulting gap in Social Security funds could be made up through benefit cuts for younger workers, including an increase in the retirement age. The required benefit cuts under Bush’s proposal could be substantial. A recent Century Foundation study (Aaron et al. 2000) estimates that, to compensate for the fund’s loss of revenue, Bush’s proposal will require benefit cuts of 46% for those 35 years old in 2002 and 25% for those age 55. Though there are a variety of ways to address the Social Security shortfall, Gov. Bush has said specifically that he would not rule out raising the retirement age as an option (Des Moines Register 1999), thus reducing the number of years workers would receive full benefits. In order to cover the Social Security shortfall, the retirement age would have to be increased by an extra 6.6 years, to age 73.6, for those age 35 in 2002 and by an additional 3.0 years, to age 69, for 55 year olds. Even if only half of the shortfall was to be financed by pushing the normal retirement age into the future, the retirement age would still have to go up by 1.4 years, to 67.4 for 55 year olds, and by 3.2 years, to age 70.2 for 35 year olds. Assuming that workers remain in good enough health to work the extra years, raising the normal retirement age results in additional Social Security revenues for three reasons. First, since people will have to work longer before receiving full benefits, they will also have to pay taxes longer. Second, workers will receive full benefits later in life and, hence, receive fewer total benefits. And finally, as the retirement age is pushed back, fewer people will reach it, reducing the number of people who will ever see benefits at all. Another factor to consider is that most people retire early and do not wait until they reach normal retirement age. Currently, the earliest age at which workers can receive any benefits is 62, at which point workers receive about 20% less than they would receive at the normal retirement age of 65. Recent legislation has already mandated that the normal retirement age will increase to 67 in 2027, resulting in a 30% reduction in benefits for those retiring early at 62. The cuts required under Bush’s proposal would increase these benefit reductions, perhaps to as much as 55% for early retirement at age 62. In other words, workers will have to either work for up to six years longer than they would under the current system, or they could take a substantial benefit cut and work fewer years, leaving them with significantly less retirement income for the remainder of their lives. Raising the retirement age also means less time spent in retirement for those who leave the workforce at the new normal age. Although, on average, more people are living long enough to make it to their retirement years, they aren’t necessarily living to be much older than in the past. While the life expectancy at birth – 76.5 years in 1997 – grew by 0.2 years annually between 1960 and 1997, the life expectancy at age 65 – 82.7 in 1997 – increased by only 0.09 years annually. Furthermore, Social Security’s trustees expect the life expectancy at age 65 to increase only at about 0.05 years annually for the next 75 years. To cut benefits and make up the shortfall, the normal retirement age has to go up faster than that. Consequently, a higher retirement age means significantly fewer retirement years, or, if a worker retires early, a substantial benefit cut. If the retirement age goes up to 73.6 years for 35 year olds in 2002, and if workers continue to work for the extra 6.6 years, they would have, on average, about 14.3 years in retirement, as compared to nearly 18 years under the current arrangement (Table 1). If the retirement age is instead increased by just 3.2 years to age 70.2, a 35-year-old worker would still have only 16.5 years in retirement instead of the current average of 18 years. Given the slow growth rate of life expectancy at or above age 65, it would take decades to bring the average number of retirement years back to where it is now. To reduce benefits across the board by 41%, the retirement age has to be raised to age 72.8 for those 35 years old and to 71.2 years for 55 year olds (option I in Table 1). Alternatively, benefit cuts could be phased in at a rate of 25% for workers age 55 in 2002 to 46% for workers age 35 that year. Then, the retirement age for the 55 year olds in 2002 would increase to age 69 and for 35 year olds to age 73.6 (option II in Table 1). Another option would be a scheme that increases the retirement age to cover only part, say half, of the benefit cut that would be made necessary by Bush’s privatization proposal. The other 50% could come from cutting benefits directly. Under such a plan, the retirement age would increase to 67.4 for 55 year olds and to 70.2 for 35 year olds if benefit cuts were phased in (option III in Table 1). It is worth noting that these increases in the retirement age would be more dramatic than any option discussed so far in the debate over fixing Social Security’s shortfall. Under current law, the retirement age is scheduled to increase to 67 by 2027. The 1994-96 Advisory Council on Social Security (SSA 1997) considered the savings that would result from even greater increases in the retirement age. But even the most aggressive increases examined by the advisory council – raising full retirement to age 70 – paled in comparison to almost all of the options possible under Bush’s proposal.

Table 1

The demographic impact 

Discussions of benefit cuts are usually modeled on an average worker, but it is important to remember that there are great disparities among different demographic groups in the U.S. workforce. African Americans, the poor, blue-collar workers, and workers with less education will bear the brunt of the proposed benefit cut, mainly because of their lower income levels and poorer health. Different workers have different life expectancies depending on their demographic and economic characteristics (Table 2). Women live longer than men; whites live longer than African Americans; average-income workers live longer than low-income workers. Raising the normal retirement age by a fixed number of years obviously will hit some workers harder than others.

Table 2

Looking at life expectancy once a worker reaches normal retirement age helps illustrate
these inequities. Table 3 shows the number of years that workers, given their personal characteristics, can expect to receive full retirement benefits if they choose to work to the new retirement age. As mentioned, to fund Bush’s privatization proposal, an average benefit cut of 41% would be necessary. If benefits were cut 41% across the board (option I in Table 3), the average 55-year-old woman in 2002 could expect about 16 years of full benefits upon retirement. Low-income women, however, would receive fewer than 13 years and low-income men about eight.

Table 3

Compounding these disparities between different groups is the fact that certain workers are less likely to survive to the new normal retirement age. For those who live to age 65, women have the highest chance of reaching age 71 (89%) or age 73 (84%). Low-income men, however, have the lowest chance of reaching either age 71 (74%) or age 73 (65%). Thus, more than a third of the low-income men who reach age 65 typically die before 73, and those who survive still have a significantly shorter retirement to look forward to than everybody else. Aside from the high probability of dying between age 65 and the new retirement age, there are other quality-of-life issues involved in raising the retirement age. Some workers will be forced to retire early and take a significant benefit cut in order to preserve their health. This seems to be particularly true for blue-collar and low-income workers (Garrett 1995; Iams and McCoy 1991; SSA 1986). And such health concerns are only compounded by age, with certain groups experiencing faster health deterioration than others (Bovbjerg 1998; Crimmins et al. 1996). The physical deterioration of those over 65 is more pronounced for women, African Americans, the less educated, blue-collar workers, and the poor. How the plan adds up

So what are the ramifications of the cuts needed under Bush’s plan? Life expectancies at age 65 vary by economic and demographic factors. In 1997, the average life expectancy at age 65 was another 17.7 years (Table 2). However, men can expect only another 15.9 years, African Americans an additional 16.1 years, and African American men just 14.2 years. Finally, low-income men – those in the bottom income quintile – can look forward to only another 11.2 years of life expectancy. Clearly the benefit cut will hurt the workforce’s already most-disadvantaged workers. An average benefit cut of 41%, for example, might mean an actual cut of 40%, on average, for women but a 52% average cut for poor men. But one possible way of funding Bush’s proposal would be to phase in the benefit cuts. In such a scenario, the disparity in benefit cuts across different groups actually grows larger and results in even higher average cuts. Based on an average cut of 25% for people turning 55 in 2002, the benefit cuts range from 24% for women to 32% for poor men. But for those turning 35 in 2002, the spread increases from 8 percentage points to 13 – women will see a 46% cut, but poor men will see benefits cut 59%. African American men would also move further away from the average over time, with the disparity growing from a 3 percentage-point difference for 55 year olds to a 6 percentage-point difference for 35 year olds. Conclusion and recommendations

A higher normal retirement age means that 35-year-old workers in 2002 would have to work an additional six or seven years before they could receive full retirement benefits or retire early with substantially reduced annual benefits. Some workers may want to continue working until they reach the normal retirement age, but others may not have that option. Sometimes a worker’s health has already deteriorated to a point where working additional years is not possible. This is especially true for low-income and blue-collar workers, women, and African Americans. Furthermore, fewer workers will qualify for disability insurance, even if their health is deteriorating. Thus, some workers may find themselves between a rock and a hard place after the normal retirement age is raised. They may need to work longer to avoid retirement in poverty, but their health may prohibit them from doing so. It must be remembered that Social Security was designed as social insurance for exactly these workers. Ultimately, raising the retirement age would be most burdensome for those who depend on Social Security the most. Some will argue that Bush’s proposed private accounts will make up for the cuts in benefits, but one recent study (Aaron et al. 2000) estimates that, even under optimistic assumptions, the private accounts in Bush’s plan will replace only 50% of the Social Security benefit cut. Another study (Baker 2000) estimates that, under more realistic assumptions, private accounts can cover only 22-37% of the cuts to Social Security benefits. In the end, benefit cuts or tax increases may be necessary to cover the anticipated shortfall in Social Security funding after 2037, but raising the retirement age is clearly the wrong path to take. Instead, progressive changes to the system could be introduced, such as eliminating the cap on annual taxable earnings – currently at $76,200 – and raising benefits accordingly, a change that would pay for more than 75% of the expected funding shortfall. For policy makers who want demographics to play a part in the solution, taxes could be increased at the same rate as life expectancy at retirement age. In this way, workers would pay more during their working years to finance longer retirements.

  September 2000 Methodology

A higher normal retirement age translates into benefit cuts for a number of reasons. First, raising the normal retirement age (NRA) changes the amount of benefits Social Security has to pay. Second, it also increases the tax payments that Social Security receives since workers have to work longer in order to receive full benefits. Third, more savings to Social Security result from the fact that each year a certain share of workers dies and hence will not receive any retirement benefits. There are also a few factors that offset the savings from a higher normal retirement age. In particular, as each year passes, the likelihood of living one year longer does not decline by a full year, but rather by 0.65 years. In other words, if life expectancy at age 65 is 18 years, life expectancy at age 66 is 17.35. As the normal retirement age is moved up, we correct for this declining reduction each year in the calculations. Most workers decide to retire early, instead of waiting to reach the normal retirement age, which is the point when they receive full benefits. Should workers decide to take early retirement, it is assumed that their benefits will be reduced in an actuarially fair manner. In other words, the net present value of all benefits paid out is the same whether a worker takes early retirement or retires at the normal age. The benefit cut calculations in this paper relate the benefits under a new normal retirement age to the benefits a worker would have received under current law. The benefit reductions are hence a sum of fewer years of benefit payments, of more tax payments, and of the declining number of possible recipients each year. Also, the number of years for which benefits have to be paid declines by 0.65 years with each year added to the normal retirement age. It is further assumed that the additional savings that arise from more non-payments of benefits are invested in the
trust fund at the long-term interest rate. The parameters for the calculations are based on average benefits, average earnings, and the assumptions of the 2000 Trustees Report. The tax rate on additional working years is 10.4% instead of 12.4%, since it is assumed that all workers will invest the maximum possible amount into private accounts. Moreover, the additional working years are assumed to be benefit neutral, which most likely understates reality, and therefore overestimates the impact of each additional year. Furthermore, the rate of change in life expectancy at different ages after age 65 – 0.35 years for each year – is taken from the 1996 Interpolated Abridged Life Tables from the National Center for Health Statistics (NCHS). Similarly, the life expectancy at age 65 for whites, African Americans, and African American men are taken from the 1997 Interpolated Abridged Life Table from the NCHS and are assumed to grow at the rates projected for men and women in the 2000 Trustees Report. Finally, the life expectancy for low-income workers is derived by adjusting the NCHS life tables for men and women after age 65 assuming, a mortality ratio of 1.70 for low-income women to average-income women and a mortality ratio of 1.50 for low-income men to average-income men (Iams and McCoy 1991). The results of this analysis are consistent with the results reported elsewhere, and they are slightly more conservative than other estimates for necessary increases in the normal retirement age. In general, each year that is added to the normal retirement age reduces benefits by 7-8% (see Table 1). The American Academy of Actuaries reported that each year added to the normal retirement age reduces benefits by about 7% (AAA 19997; Gebhardtsbauer 1998). Thus, we are slightly overstating the impact of each year and consequently understating the necessary increases in the normal retirement age to generate the required benefit cuts. Put another way, our estimates are conservative in the sense that the years by which the normal retirement age would have to increase to pay for the required benefit may be underestimated. References

Aaron, H., Blinder, A. Munnell, and P. Orszag. 2000. Governor Bush’s Individual Account Proposal: Implications for Retirement Benefits. Issue Brief No. 11. New York, N.Y.: The Century Fund.

Advisory Council on Social Security. 1997. Report of the 1994-96 Advisory Council on Social Security, Volume I: Findings and Recommendations. Washington, D.C.: Department of Health and Human Services.

American Academy of Actuaries (AAA). 1997. Raising the Retirement Age for Social Security. Issue Brief. Washington, D.C.: American Academy of Actuaries.

Baker, Dean. 2000. Governor Bush’s Individual Account Proposal: A Reassessment Using Realistic Stock Return Projections. Briefing Paper. Washington, D.C.: Center for Economic and Policy Research.

Bovbjerg, Barbara D. 1998. “Raising Retirement Ages Improves Program Solvency but May Cause Hardship for Some.” Testimony before the Special Committee on Aging, U.S. Senate, July 15, 1998.

Crimmins, E. M., M.D. Hayward, and Saito Yasuhiko. 1996. “Differentials in Active Life Expectancy in the Older Population of the United States.” Journal of Gerontology. Vol. 51, No. 1, pp. 111-20.

Des Moines Register. 1999. “Saving Social Security.” December 25, 1999.

Garrett, D. 1995. “The Effects of Differential Mortality Rates on the Progressivity of Social Security.” Economic Inquiry. Vol. 33, July, pp. 457-75. Gebhardtsbauer, R. 1998. “The Future of Social Security for This Generation and the Next: Raising the Retirement Age.” Testimony presented to the Social Security Subcommittee, Committee on Ways and Means of the U.S. House of Representatives, February 26.

Iams, H., and J. McCoy. 1991. “Predictors of Mortality Among Newly Retired Workers.” Social Security Bulletin. Vol. 54, No. 3, pp. 2-10.

National Center for Health Statistics. Various years. Abridged Life Tables. Hyattsville, Md.: NCHS.

Social Security Administration. 2000. The 2000 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds. Washington, D.C.: Social Security Administration.

Social Security Administration. 1986. “Increasing the Social Security Retirement Age: Older Workers in Physically Demanding Occupations and Health.” Social Security Bulletin. Vol. 49, No. 10, pp. 5-23.

Wolff, E., and H. Chernick. 1996. The Distributional Effects of Raising the Social Security Retirement Age and Partially Indexing Benefits. Working Paper No. 115. Washington, D.C.: Economic Policy Institute.

See related work on Trade and Globalization

See more work by Christian E. Weller