|Retirement Policy Outlook logo 130||
No. 3, October 2010
This version has been updated to reflect revised GDP figures.
In this Outlook, I use a detailed microsimulation model of the population to estimate the effects of increasing the Social Security Earliest Eligibility Age (EEA) from sixty-two to sixty-five on Social Security's finances, retirement income, and the economy. Increasing the EEA would extend the solvency of the Social Security trust fund by about five years, increase total annual retirement income as of age seventy for affected individuals by around 16 percent, and increase gross domestic product (GDP) by 3-4 percent. Raising the EEA may be one of the most effective options available for improving retirement-income security and would improve the federal budget in one year nearly as much as the recent health reform bill was projected to do over ten years.
Key points in this Outlook:
The earliest age at which Americans can claim Social Security benefits is sixty-two. Gradually raising the early retirement age to sixty-five would:
- Have only modest effects on Social Security’s finances, prolonging the life of the trust fund by around five years, from 2037 to 2042
- Significantly increase both Social Security benefits and private pensions once individuals did retire, by around 16 percent or $7,500 per year
- Boost GDP by 3-4 percent, adding billions to the economy and tax revenues. Increasing the early retirement age would improve the federal budget over ten years by around five times more than the recently passed health reform legislation
Other policies, such as reducing payroll taxes on older workers and making Medicare the primary payer of health care for individuals over sixty-five, could also facilitate longer work lives.
Increasing the EEA, while of small importance to Social Security's finances, could significantly increase retirement incomes while boosting the economy and federal tax revenues. Some individuals could not work longer due to poor health, but the health status of older Americans has improved significantly while the physical demands of work have declined. The evidence indicates that most Americans could work longer and would benefit from doing so.
Background on the Social Security Early Retirement Age
Three ages are important for claiming Social Security retirement benefits. Age sixty-two is the EEA, commonly known as the early retirement age. The Full Retirement Age (FRA) is currently sixty-six and will move gradually to sixty-seven by the early 2020s. Age seventy is the oldest age at which benefits are increased to account for delayed claiming; there is generally no reason to delay claiming after age seventy. Social Security benefits increase by around 7 percent for each year an individual delays claiming after age sixty-two, although the exact schedule differs by age.
When the Social Security program was established in 1935, the age of earliest eligibility and the normal retirement age were both sixty-five. In other words, there was no provision for early retirement. In practice, most individuals claimed benefits later than age sixty-five; in the 1950s, for instance, the average age of first benefit claiming was sixty-eight. At the time, Social Security had a strict earnings test, such that the receipt of even a small amount of earnings disqualified individuals from receiving benefits. In response, individuals generally waited until they were fully withdrawn from the labor force before claiming Social Security.
However, in response to concerns that some individuals could not work until age sixty-five, the Social Security Amendments of 1956 (for women) and 1961 (for men) allowed for early claiming at age sixty-two, with a reduction in benefits. In response to this change, the average age of initial benefit claiming declined by two years from 1960 to 1970, from 66. to 64.2 (see figure 1). The average claiming age as of 2007 was 63.6 for men and 63.5 for women.
RPO 2010-10 No. 3 Figure 1
Labor-force participation rates for older males also declined from the mid-1950s through the mid-1990s (see figure 2) in response to the availability of Social Security benefits at earlier ages, although other factors also contributed to this change. Rising incomes may reduce retirement ages, as individuals take part of the income gain in the form of leisure. Rising life expectancies, by contrast, may promote longer work lives as individuals offset the cost of supporting themselves in longer retirements. A number of other factors come into play, including physical and mental job requirements, the demand for older workers, and policy incentives regarding the retirement decision. For decades after World War II, the trend was toward falling labor-force participation at older ages and earlier claiming of Social Security benefits. While benefit-claiming ages remain low, over the past two decades labor-force participation among older individuals has risen, suggesting that forces promoting longer work lives may be having an effect. Some older individuals who might have left the workforce before claiming Social Security are now remaining at work, while other older individuals are continuing to work even after claiming retirement benefits. Labor-force participation rates for women rose over the same period, as part of a broader shift from working at home to paid employment, partially offsetting lower labor-force participation by men.
RPO 2010-10 No. 3 Figure 2
By far the most common age of initial retirement benefit claiming is sixty-two, when 42 percent of men and 48 percent of women claim benefits (see figure 3). The next most common age is sixty-five; this is driven both by the FRA long having been sixty-five and by the availability of Medicare benefits at that age. The FRA is currently sixty-six and gradually rising to sixty-seven; the norming effects of the FRA as well as the reduction in benefits associated with an increased retirement age may produce somewhat higher claiming ages.2
RPO 2010-10 No. 3 Figure 3
As figure 1 shows, the average age of benefit claiming has remained relatively steady since 1970. With the exception of 1997, when the average age spiked to 64.5, average claiming ages have tended to vary by only a few decimal points from year to year.
The increase in the FRA for Social Security implies a 13 percent reduction in monthly benefits if individuals do not respond by delaying benefit claiming. For this reason, some have proposed increasing the EEA along with the FRA. Raising the EEA from sixty-two to sixty-five may seem like a steep increase but would only bring the United States in line with other developed nations. In Canada, the Old Age Security pension is available at age sixty-five. The United Kingdom's Basic State Pension can be claimed at age sixty-five for men, with claiming as early as sixty for women. However, claiming ages for women are being increased to match those of men, and ages for both genders will gradually be increased to sixty-eight over the next four decades. Likewise, New Zealand's Superannuation benefits begin only at age sixty-five.
Simulating the Effects of Increasing the EEA
Using a suite of models from the Policy Simulation Group (PSG), I simulate the effects of increasing the EEA from sixty-two to sixty-five on the Social Security program, individual retirement income, and the economy. The PSG models are used by the Government Accountability Office, the Social Security Administration, and the Department of Labor to simulate the effects of policy changes on Social Security and private pensions. These models can provide a detailed view of how an increased EEA would affect Social Security's financing, how individuals would be affected by the changes, and how the economy might react to increases in the labor force.
The policy change I simulate is a gradual increase in the EEA from sixty-two to sixty-five over the period from 2015 to 2025. Current retirees would be unaffected, but the change would be phased in for individuals born between 1952 and 1960. Other variations are possible, and this exercise is intended merely to provide a feel for the results. No other changes are made to Social Security policy, though other changes will be necessary to make the program solvent over the long term.
Effects on Social Security's Finances
Social Security's finances would be largely unchanged by an increase in the EEA. The long-term deficit, which is 1.92 percent of taxable payroll under the 2010 Social Security Trustees Report, would rise to about 1.93 percent of payroll. The life of the Social Security trust fund would be extended by around five years, from 2037 to 2042.
While not inconsequential, these effects are modest. The reason, as noted above, is that Social Security increases benefits in response to delayed retirement. Under current law, if individuals delayed claiming benefits from age sixty-two to sixty-five, their monthly checks would rise by about 20 percent. Moreover, raising the EEA would increase benefits by an additional 3.5 percent because of how Social Security benefits are calculated. Thus, Social Security's financing profits from the extra taxes many individuals would pay if forced to delay retirement, but the gains to solvency are modest due to increased benefits paid out.
Effects on Retirement Incomes
An increase in the EEA may be perceived as merely a forced delay in retirement, and in many respects it is. Monthly Social Security benefits are increased by about 8 percent for each year that benefits are delayed, plus an inflation adjustment that averages around 3 percent. In addition, however, a little-known facet of the Social Security benefit formula would provide a small additional increase to benefits, such that total lifetime benefits would increase by about 6 percent.
But because higher lifetime Social Security benefits would be consumed over a shorter retirement, gains to individual retirement incomes from increasing the EEA would be significant. Social Security benefits paid at age seventy to individuals in the 1980 birth cohort would increase by an average of 17.1 percent. In addition, average income derived from defined contribution and defined benefit pensions would increase by 15.1 percent due to longer work lives in which savings would be amassed and shorter retirements in which savings would be drawn down. Combined, total Social Security and private-pension incomes would increase by 16.1 percent. Median annual Social Security benefits would increase by around $4,300, and median annual total Social Security plus pension income would increase by around $7,500, in constant 2010 dollars.
Another way of measuring retirement-income adequacy is through replacement rates, which represent retirement income as a percentage of preretirement earnings. Financial advisers generally recommend a replacement rate of 70 to 80 percent of preretirement earnings, although individual needs vary significantly. Under current law, the median couple in the 1980 birth cohort will receive a total Social Security benefit equal to 43 percent of preretirement earnings (see the table). Adding pension income raises total replacement rates to 67 percent of preretirement earnings. Mean values would be somewhat higher. On top of Social Security and pensions, households may rely on personal savings and other sources of income, so these should not be considered complete measures of retirement-income resources.
RPO 2010-10 No. 3 Table
Increasing the EEA to age sixty-five would raise the median Social Security replacement rate from 43 to 51 percent of preretirement earnings. The median combined Social Security and pension replacement rate would rise from 67 to 77 percent, making a significant dent in the percentage of individuals considered ill-prepared for retirement. Only around 6 percent of households in the bottom third of the earnings distribution would have replacement rates below 70 percent, with 4 percent having replacement rates below 60 percent. Given that most households, including low-income households, have limited resources beyond Social Security and employer-sponsored pensions, the share of low earners who would retire with inadequate replacements would be significantly reduced under this policy.
Moreover, when replacement rates are adjusted for household size and composition, both baseline rates and rates after increases in the EEA are significantly larger. Simple replacement-rate measures can misstate a household's preparation for retirement by ignoring how the costs of raising children and economies of scale in household size can affect retirement-income needs. Adjusting for these factors raises median baseline Social Security replacement rates by around 12 percentage points and would have similar effects on replacement rates after increasing the EEA. If retirement-income needs are properly measured, increasing the EEA would generate total incomes that make a large majority of households well-prepared for retirement. In fact, increasing the EEA may be the single most effective policy for raising total retirement income, as it would boost both Social Security benefits and pension savings.
Effects on the Economy and Tax Revenues
While increasing the EEA would not resolve Social Security's financial challenges, the economy would benefit due to a larger workforce. To the degree that individuals continue to work, their earnings would contribute to both GDP and federal tax revenues.
The PSG models estimate that raising the EEA would increase long-run GDP by 3-4 percent, meaning that in future years economic output would be 3-4 percent higher than it would be with a lower Social Security eligibility age. Increases in GDP would differ from year to year based on factors such as changes in the age structure of the population, the national saving rate, and the reaction of other workers' wages to changes in the labor force. Nevertheless, these results highlight the significant potential economic impact of greater labor-force participation among near-retirees. This increase in output, equivalent to around $550 billion annually in current terms, would improve Americans' standards of living and provide extra resources with which to address long-term budget shortfalls.
Based on these projections of changes in GDP, over the 2015-2024 period federal revenues would increase by about $1.1 trillion in nominal dollars. This estimate derives from the percentage of GDP collected under the Congressional Budget Office's "alternative budget scenario" in which most of the 2001 and 2003 tax cuts would be retained. Revenue increases would be larger over the longer term, as the projected increase in GDP over the first ten years of implementation averages only 2.3 percent versus a long-term increase of 3-4 percent.
By comparison, raising the EEA would improve the federal budget over ten years by about five times as much as is projected for the recent health reform bill. Over the period 2010-19, the Congressional Budget Office projects that health reform will improve the annual budget balance by an average of 0.08 percent of GDP; over the first ten years of increasing the EEA, from 2015 to 2024, the federal budget balance would improve by about 0.40 percent of GDP. While these new revenues are nowhere near sufficient to address the long-term budget gap, they are a very strong start.
Can Americans Work Longer?
It is reasonable to counter that Americans cannot work longer--and many Americans cannot. While the physical strains of work have been reduced as the United States has shifted toward a service and technology economy, in today's marketplace declining analytical and social skills may impose larger obstacles to work than physical strength. But these obstacles to longer work lives can be overcome. If an increase in the EEA is considered, a number of policies could reduce stresses on more vulnerable individuals.
If the EEA were increased, individuals who could not work would not be prohibited from claiming Disability Insurance benefits in the years from age sixty-two to sixty-five. These individuals would receive benefits payable as of the FRA of sixty-six or sixty-seven and thus would not suffer a financial penalty if forced to leave the workforce due to disability. Simulations using the PSG models show that almost twice the number of individuals would claim Disability Insurance benefits between age sixty-two and sixty-five, a figure that appears consistent with academic work on the subject. While rising disability costs are a significant part of Social Security's long-term funding shortfall, increased disability claims by older workers are an inevitable, and legitimate, offshoot of an increased EEA. The Disability Insurance program provides a safety net for individuals who truly cannot work longer.
To provide additional protections for vulnerable workers, the eligibility age for Supplemental Security Income (SSI) for the elderly could be lowered from sixty-five to sixty-two. SSI is a means-tested welfare program that provides benefits to the elderly, as well as the blind and the disabled. In 2010, the maximum monthly federal SSI payments are $674 for individuals and $1,011 for couples. In addition, most states provide additional payments on top of SSI to eligible individuals. Lowering the SSI eligibility age would protect low-income individuals who did not qualify for disability benefits.
But even without these extra protections, we should not overestimate the barriers to delayed retirement. One of the best pieces of evidence that delayed retirement is possible is that it has been done in the past, when work conditions were more strenuous and health conditions poorer than they are today. While policy should be designed to protect those who cannot delay retirement due to poor health or other factors, policy should also recognize that individuals who cannot work longer are the exception rather than the rule.
Individuals who might plausibly be regarded as unable to delay retirement are those who are both in poor health and wholly dependent on Social Security for income. As of the early 1990s, only around one in ten early retirees met both these conditions. Likewise, only around 15 percent of early retirees cited either losing their job or health problems as the reason for leaving their previous employment. By far the most common reasons are "other" or "unknown," which could include voluntary retirement, or "still working."
While poor health is obviously a precipitating factor for many retirements, early retirees in general are not markedly less healthy or less wealthy than those who delay retirement. As Gary Burtless and Joseph F. Quinn point out, "it should be plain the long-term trend toward earlier male retirement has had an important voluntary component," principally individuals' greater ability to afford to retire earlier. They continue:
Of all the explanations advanced for earlier retirement, two of the least persuasive are declining health and the changing physical requirements of work. While nearly all good retirement studies find that health plays an important role in the timing of retirement, there is no convincing evidence that the health of 60-year-olds or 65-year-olds was declining over the period in which older Americans' labor force participation rates were falling.
The number of Americans describing themselves as being in only fair or poor health has dropped significantly over the past several decades, according to data from the National Center for Health Statistics. Among individuals aged fifty-five to sixty-four, those with fair or poor health dropped from 25.1 percent in 1983 to 18.5 percent in 2007; among those aged sixty-five to seventy-four, the percentage dropped from 32.5 to 22.4; and for those aged seventy-five to eighty-four, it dropped from 34.6 to 25.5. Overall, 75 percent of individuals over age sixty-five reported themselves as being in good, very good, or excellent health in 2008 (see figure 4). Yet only slightly smaller numbers of individuals choose to claim Social Security benefits early, suffering a significant loss of income in later years.
Harvard University's David Cutler recently examined the capability of older men to continue in the workforce. Cutler concluded that
work capacity declines only slowly through the mid-70s, and then declines more rapidly thereafter. Relative to men in their late 50s,work capacity is about 90 percent as high at age 65, and 70 percent as high at age 75. The capacity for additional employment among near elderly and elderly men is thus very high.
Cutler finds that work capacity among men declines by only 8 percent between ages fifty-five and sixty-five and by only 20 percent up to age seventy-five.
Likewise, while some jobs are too physically demanding to allow for long work lives, fewer of them exist today than in the past. In 1950, over 20 percent of Americans worked in physically demanding jobs; today, only around one-third as many do.
The point is not that all Americans should be expected to work longer; there are millions of individuals who due to injury, ill health, or other causes cannot delay retirement past sixty-two. But there are millions more who can work longer and would benefit from doing so.
The Government Accountability Office points out that "the majority of workers aged sixty-two to sixty-seven do not appear to have health limitations that would prevent them from extending their careers, although some could face severe challenges in attempting to remain in the workforce." In addition to protecting those who cannot work longer, policy should remove impediments to longer work lives.
Increasing the number of older workers is the first step toward increasing retirement security, but there must also be demand for this large workforce. Some argue that no jobs exist for older workers. But this claim commits what economists call the "lump-of-labor fallacy," which holds that there are a limited number of jobs for which workers must compete. Economists almost universally reject this view, and the fact that employment continues to rise despite immigration and rising worker productivity (which presumably would reduce the need for extra workers) speaks against it. In the short term, particularly during the current recession, new jobs for older workers would be hard to find. However, any increase in the EEA would be phased in over a number of years, giving both individuals and markets time to adjust. Also, retiring baby boomers are leaving the labor market at the rate of ten thousand per day while birth rates and labor-force growth remain low. These forces by themselves may increase demand for labor and encourage employers to take various steps to retain older Americans in the workforce.
That said, a number of policies would smooth the transition to longer work lives. One such policy is to reduce or eliminate the Social Security payroll tax for older workers. As my previous work with Gayle Reznik and David Weaver of the Social Security Administration showed, most near-retirees receive little extra Social Security benefits in exchange for the extra payroll taxes they pay by delaying retirement. The median individual receives only around 2.5 cents of additional benefits in exchange for one dollar of additional taxes at the end of his work life. Lowering the Social Security payroll tax could sweeten the deal for older workers and send the signal to both workers and employers that longer work lives can pay off. Increased numbers of older workers would generate additional payroll and income taxes, partially offsetting the cost of a reduced tax rate.
In addition, Medicare could become the primary payer of health services for individuals who remain employed after age sixty-five. Under current law, employer-sponsored health insurance for Medicare-eligible individuals is the primary payer, while Medicare covers only secondary costs. This policy imposes an implicit tax of 15 to 20 percent of wages for a sixty-five-year-old, with rising rates thereafter. Making Medicare the primary payer of health care for eligible individuals would remove this implicit tax and make work significantly more profitable at older ages, and increased income-tax revenues from older workers could offset much of the additional cost of such a policy change.
Finally, while increasing the EEA would effectively eliminate the Retirement Earnings Test (RET) for individuals aged sixty-two through sixty-five since retirement benefits could not be claimed, the RET should also be eliminated for the years between sixty-five and the FRA of sixty-seven. The RET withholds $1 of Social Security benefits for every $2 in earnings above $14,160. A typical early retiree receiving a $1,000 monthly benefit would see his benefits eliminated if he earned more than $36,000. At the FRA, the RET ceases and benefits are increased to compensate for any reductions that occurred prior to the FRA. Nevertheless, most retirees are unaware of this benefit adjustment and treat the RET as a very high effective tax on earnings. Given this, retaining the RET for ages sixty-five through sixty-seven seems counterproductive.
Increasing the EEA for Social Security benefits could encourage Americans to remain in the workforce longer, significantly increasing their retirement income, boosting economic output, and increasing tax revenues. Extended work lives are possible for most Americans and would generate significant benefits to the economy, the federal budget, and, most importantly, individuals' own retirement security.
Andrew G. Biggs (firstname.lastname@example.org) is a resident scholar at AEI. A version of this Outlook appeared in the March 22, 2010, edition of National Review.
1. Geoffrey Kollmann, Social Security: Summary of Major Changes in the Cash Benefits Program: 1935-1996 (Washington, DC: Congressional Research Service, May 18, 2000), available at www.ssa.gov/history/pdf/crs9436.pdf (accessed October 12, 2010).
2. See Jae Song and Joyce Manchester, "Have People Delayed Claiming Retirement Benefits? Responses to Changes in Social Security Rules," Social Security Bulletin 67, no. 2 (2007): 1-23.
3. Service Canada, "Old Age Security Pension (OAS)," available at www.servicecanada.gc.ca/eng/isp/pub/oas/oas.shtml (accessed October 12, 2010).
4. Pensions Advisory Service, "State Pension Age Calculator," available at www.pensionsadvisoryservice.org.uk/state-pensions/state- pension-age-calculator (accessed October 12, 2010).
5. Ministry of Social Development, "New Zealand Super-annuation," available at www.workandincome.govt.nz/individuals/65-years-or-older/superannuation/index.html (accessed October 12, 2010).
6. For details on the PSG models, see Policy Simulation Group, "PSG Models," available at http://polsim.com (accessed October 12, 2010).
7. The PSG models are currently calibrated to the 2009 Trustees projections, so the increase in the EEA was simulated against the baseline in the 2009 Trustees Report.
8. Social Security's benefit formula is roughly neutral with regard to the age of claiming, which means that the increase in annual benefits from delaying claiming a year is sufficient to compensate an individual over his lifetime for the loss of a year's benefits. In other words, while early retirees receive lower benefits over a longer period, later claimants receive higher benefits over a shorter period. Over an average lifetime, total benefits are about the same. But raising the EEA would increase benefits because of the way the Social Security benefit formula is calculated. Social Security benefits are based upon an individual's average lifetime earnings, with low earners receiving a higher percentage replacement of average earnings than high earners. Average earnings are calculated by first "indexing" past earnings to wage growth through age sixty-two. For instance, if an individual earned the average economy-wide wage in a previous year, the indexing method would make these past wages equal to the average economy-wide wage as of the year he turned sixty. This method adjusts past earnings for inflation and compensates individuals for productivity-driven growth of real wages. After past earnings are indexed for wage growth, the highest thirty-five years of indexed earnings are averaged. The Social Security benefit is based upon these averaged indexed earnings. One little-known fact is that the age to which individual earnings are indexed is set at two years before the EEA. Thus, if the EEA were increased from sixty-two to sixty-five, the age to which earnings were indexed would rise from sixty to sixty-three. Since wages tend to grow around 1.1 percent faster than inflation, this would produce an average increase of around 3.4 percent in measured average earnings and an increase in benefits that would be nearly as large. This increase would come on top of normal adjustments to benefits due to delayed claiming. Thus, raising the EEA would increase total lifetime benefits for the typical retiree even if individuals did not respond by working longer. In addition, individuals who continued to work during the years from sixty-two to sixty-five might receive higher benefits as a result of these additional earnings, though the effects would generally be small due to low marginal returns paid by Social Security to near-retirees.
9. Benefit estimates are derived from the PSG models. For details on the interaction between raising the EEA and the Social Security benefit formula, see Pat Vinkenes, Alice H. Wade, Mark A. Sarney, and Tim Kelley, "Considerations for Potential Proposals to Change the Earliest Eligibility Age for Retirement" (Policy Brief No. 2007-01, Social Security Administration, Washington, DC, October 2007), available at www.ssa.gov/policy/docs/policybriefs/pb2007-01.pdf (accessed Octo-ber 12, 2010).
10. The increase in overall retirement incomes is the weighted average of the increase in Social Security and private-pension benefits.
11. Replacement rates are calculated only for individuals who are receiving benefits as of age seventy. Replacement rates are calculated by dividing Social Security benefits or Social Security benefits plus pension income as of age seventy by a steady preretirement earnings stream derived from the present value of lifetime earnings. This earnings stream is assumed to increase annually at the rate of wage growth.
12. See Glenn Springstead and Andrew G. Biggs, "Alternate Measures of Replacement Rates for Social Security Benefits and Retirement Income," Social Security Bulletin 68, no. 2 (November 2008): 1-19.
13. See John Karl Scholz and Ananth Seshadri, "Children and Household Wealth" (Working Paper No. 2007-158, Retirement Research Center, University of Michigan, Ann Arbor, MI, October 1, 2007).
14. For more information on measurement issues associated with replacement rates and retirement income, see Andrew G. Biggs, "Will You Have Enough to Retire On? The Retirement Security 'Crisis,'" AEI Retirement Policy Outlook (February 2009), available at www.aei.org/outlook/100001; and Glenn Springstead and Andrew G. Biggs, "Alternate Measures of Replacement Rates for Social Security Benefits and Retirement Income."
15. This does not imply that the annual growth rate would increase by 3-4 percent; rather, GDP growth would rise for the period until total GDP had increased by 3-4 percent relative to the baseline. Once the increase in the EEA was fully implemented, GDP growth would return to about the baseline rate.
16. These estimates should be treated as approximations, due to uncertainties regarding broader responses to an increase in the EEA. For instance, if national saving remained constant while the labor force increased, a given percentage increase in the labor force would produce a somewhat smaller increase in GDP because a larger labor force would be supplied with the same amount of capital. However, if saving increased along with longer work lives--which is not implausible, as the additional workers would in general continue to participate in their personal retirement saving plans--then a given percentage increase in the labor force would lead to an equal percentage increase in GDP. That is the implicit assumption underlying the results here.
17. See Richard V. Burkhauser, Kenneth A. Couch, and John W. Phillips, "Who Takes Early Social Security Benefits: The Economic and Health Characteristics of Early Beneficiaries," Gerontologist 36, no. 6 (December 1996): 789-99; and Congressional Budget Office, Raising the Earliest Eligibility Age for Social Security Benefits (Washington, DC, 1999), available at www.cbo.gov/ftpdocs/10xx/doc1058/ ssage.pdf (accessed October 12, 2010).
18. Congressional Budget Office, Raising the Earliest Eligibility Age for Social Security Benefits.
19. Gary Burtless and Joseph F. Quinn, "Retirement Trends and Policies to Encourage Work among Older Americans" (Working Paper in Economics 436, Boston College Department of Economics, Boston, MA, 2000), available at http://escholarship.bc.edu/cgi/viewcontent.cgi?article=1180&context=econ_papers (accessed October 13, 2010). See also Olivia S. Mitchell and John W. Phillips, "Retirement Responses to Early Social Security Benefit Reductions" (NBER Working Paper No. W7963, National Bureau of Economic Research, Cambridge, MA, October 2000).
20. David Cutler, "Estimating Work Capacity among Near Elderly and Elderly Men" (National Bureau of Economic Research, Retirement Research Center, Cambridge, MA, September 2009).
21. Kenneth G. Manton and Eric Stollard, "Medical Demography: Interaction of Disability Dynamics and Mortality," in Demography of Aging, ed. Linda G. Martin and Samuel H. Preston (Washington, DC: National Academies Press, 1994), 217-79; and Martin N. Baily, "Aging and the Ability to Work: Policy Issues and Recent Trends," in Work, Health, and Income among the Elderly, ed. Gary Burtless (Washington, DC: Brookings Institution Press, 1987), 59-96.
22. Richard W. Johnson, Christopher Spiro, and C. Eugene Steuerle, "Trends in the Physical Demands of Work: Implications for Social Security Reform" (working paper, Urban Institute, Washington, DC, 1999).
23. Senate Special Committee on Aging, Redefining Retirement: Options for Older Americans, 109th Cong., 1st sess., April 27, 2005, available at www.gao.gov/new.items/d05620t.pdf (accessed October 12, 2010).
24. Gayle Reznik, David A. Weaver, and Andrew G. Biggs, "Social Security and Marginal Returns to Work Near Retirement" (Issue Paper No. 2009-02, Social Security Administration, Washington, DC, April 15, 2009), available at www.ssa.gov/policy/docs/issuepapers/ ip2009-02.pdf (accessed October 12, 2010).
25. Gopi Shah Goda, John B. Shoven, and Sita N. Slavov, "A Tax on Work for the Elderly: Medicare as a Secondary Payer" (NBER Working Paper W13383, National Bureau of Economic Research, Cambridge, MA, September 2007).
26. See Andrew G. Biggs, "The Social Security Earnings Test: The Tax That Wasn't," AEI Tax Policy Outlook (July 2008), available at www.aei.org/outlook/28341.