November 22, 2010

November 22, 2010 | Backgrounder on Retirement Security

Time to Raise Social Security’s Retirement Age

Key Points

  1. Since Social Security has promised to pay trillions of dollars more in benefits than it will receive in taxes, a comprehensive overhaul of the program is necessary to increase retirement security and improve the program’s finances.
  2. Today’s workers receive Social Security payments for a much higher proportion of their lives because the program’s retirement ages have not kept pace with longevity improvements.
  3. Part of the solution is to gradually increase Social Security’s full retirement age from 67 to 68, and to increase the early eligibility age from 62 to 65. Those physically unable to work longer will receive disability benefits until they reach retirement age.
  4. This alone will reduce the program’s deficit by about 35 percent.
  5. As an incentive to work longer, seniors who work past the full retirement age would be exempted from paying social security taxes.

Abstract: Americans are living longer, which means they are spending a higher proportion of their lives in retirement, receiving Social Security payments. Yet the government program is a mere five years away from being unable to pay out all of the claims it has promised. Because today’s retirees enjoy longer lives and better health, both Social Security retirement ages (“normal” and “early eligibility”) must be increased. It is common sense, and it is fair. While the retirement ages must be increased, this alone is not change enough to provide a secure future for today’s younger workers. A higher retirement age should be part of a whole package of Social Security reforms, including ending the payroll tax for workers willing to work past their normal retirement age. Heritage Foundation retirement and economic policy expert David C. John explains.

As Social Security reaches its 75th birthday, the program is beginning to show signs of age. While it remains a major source of retirement income for millions of Americans, Social Security is running a significant cash flow deficit for the first time in several decades.[1] To make matters worse, Social Security is greatly underfunded, having promised $7.7 trillion[2] more in benefits over the next 75 years than it can afford to pay from its payroll taxes. Deficits will begin in 2016, and continue indefinitely.

Worse still, studies show that younger workers will receive a much lower return for their tax dollars than that of their parents and grandparents.[3] In many cases, they will receive less in benefits after retirement than they paid into Social Security from payroll taxes during their working lives. Further, it is possible for a career low-wage worker to pay Social Security taxes for 35 years, and still retire with a benefit that is well below poverty.

Fixing Social Security will not be easy. There is no magic bullet that will painlessly eliminate the program’s deficits, improve a younger worker’s rate of return, and allow Social Security to keep paying the same level of benefits that it has in the past. Social Security can and must be fixed, but that will require a combination of several reforms, some difficult.

A major overhaul to the system is necessary in order to improve the program’s long-term financing and to strengthen retirement security for those who need it the most.[4] Reforms should promote personal savings, strengthen the safety net and move to a more needs-based system.[5] On balance, these changes and others to make the program more affordable will leave workers in a better financial position than under the status quo.

One of those elements crucial to improving the program’s financing is to increase the Social Security retirement age, accompanied by other incentives, such as eliminating the Social Security payroll tax for employees who are willing to work beyond their normal retirement ages. These reforms, along with other necessary changes, will improve the long-term financing of the program, prevent radical benefit reductions in the future, and avoid pressure for tax hikes that would affect younger generations as they enter the workforce or plan for their own retirements. Social Security was never intended to be the sole source of retirement income that it has become for many Americans, so increasing personal savings is vital to the retirement security of seniors. Improvements in retirement savings should be accomplished through voluntary personal accounts within Social Security, and by improving retirement savings in general through such mechanisms as the Automatic IRA[6] for small businesses. Another necessary element is to strengthen the safety net for low-income retirees.

Social Security Not Keeping Pace with Longevity

It is time to increase the age at which American workers can receive Social Security benefits—both the full benefits age and the early eligibility age. Longevity trends show that not only are workers living longer and staying healthier longer than in the past, but that this improvement is likely to continue.

Unfortunately, Social Security has not kept pace with these changes. Congress has not changed the age at which workers can receive full benefits since 1983 (when it was increased from 65 to eventually 67 in 2022), or the early eligibility age of 62 since 1961,[7] and both the program and the workforce have changed a great deal since then. While increasing Social Security’s retirement ages is not a step that should be taken lightly, increases in longevity combined with the fact that the program’s finances are unsustainable make this change both fair and necessary. However, unlike 1983, when only the normal retirement age (NRA) was increased, this time the early eligibility age (EEA) needs to be increased as well.

Social Security’s eligibility ages should be increased simply to reflect the longevity increases that have already taken place. Additional increases may well be necessary in future years to reflect additional advances in average longevity, but even if the growth in future longevity increases slows or even stops,[8] advances since the last change in the program’s retirement ages justify raising the NRA to 68 by 2023, and the EEA to 65 by 2032. After those eligibility ages are reached, both the NRA and the EEA should be indexed to automatically rise along with longevity. In addition, those who are willing to work beyond their normal retirement age should be exempt from paying any further payroll taxes, as should their employer. The combination will provide additional employment opportunities for older Americans. Increasing longevity is not a situation that exists only in the United States. Across the world, countries are recognizing both that their workers can and should delay retirement and that doing so will reduce the cost pressures their public pension systems face. Over the past few years, Germany, Ireland, France, and Italy have already increased their public pension age, while the United Kingdom, the Netherlands, Spain, and Greece are in the process of increasing theirs.[9]

Retirement: Getting Longer and Longer

The increase in life expectancy since 1950 has been substantial. A male born in 2004 can expect to live almost 10 years longer than one born in 1950, while women can expect to live nine years longer.[10] When the Social Security program was created in 1935, an adult man who reached age 65 could expect to spend about 13 years in retirement, which was 16 percent of his life; a woman averaged 15 years, or 18 percent of her life, in retirement. However, at that time only 54 percent of men (and slightly more women) aged 21 were expected to live to age 65, and there were approximately 8 million Americans ages 65 or older.[11]

A male retiree, born in 1940, will spend anywhere from 19 percent to 25 percent of his life collecting Social Security benefits (depending on whether he retired at the normal retirement age of 65 or chose early retirement), and a female born in the same year will spend 21 percent to 27 percent of her life collecting benefits.

Life Expectancy and Retirement

In addition to supporting individual retirees longer, Social Security also has more people to support: Presently, there are 36 million people over age 65, which is more than four times the number that had to be supported when the system was created. Those who retire early in 2035, two years before the program will reach insolvency, will spend as much as 28 percent of their lives in retirement, and trends will continue to increase from there.

The age for full and early retirement should be gradually increased, and both retirement ages should be indexed for future increases in longevity. While the normal retirement age is currently scheduled to reach age 67 by 2022, even with this change Social Security cannot afford to pay all of the benefits it has promised without massive tax increases on working Americans. To keep the proportion of one’s life spent in retirement roughly the same as what it was in the 1940s, the NRA needs to increase to 70 years and five months no later than 2035.

The Relationship Between Social Security and Retirement Ages

While there is little direct proof that the introduction of the EEA caused a change in Americans’ timing of retirement, there are a number of empirical studies that strongly suggest a correlation between the two. For instance, economists John Rust and Christopher Phelan state that their “results suggest that Social Security creates significant disincentives to labor force participation, and is largely responsible for the peaks in retirements at ages 62 and 65.…”[12] This is especially true for the EEA of 62. Boston College scholars Alicia Munell and Steven Sass emphasize the point: “The fact that Social Security offers benefits at age 62 is a major obstacle to older workers’ offering their services in the future.”[13]

However, these studies also point out that the decision to retire at age 62 is complex, and includes several factors other than just the availability of Social Security, such as the presence of disability insurance and private pensions.[14] In fact, there is evidence that the proportion of men claiming benefits at age 62 stabilized at 52 percent between 1985 and 2006, while the proportion of women dropped from 64 percent in 1985 to 56 percent in 2006.[15]

While welcome, these trends still show more than half of workers leaving the labor force at age 62, and there is a valid reason to raise the EEA along with the NRA. While Social Security benefits at age 62 are actuarially adjusted so that raising the EEA has little or no effect on system finances,[16] such a move would benefit the overall economy by keeping many workers in the workforce longer. A 2008 report by McKinsey & Company found that increasing the median retirement age from 62.1 to 64.1 by 2015 would add $13 trillion to the economy over the next 30 years.[17]

A Short History of Social Security Retirement Ages
Although “mid-sixties” is typically the age range defined as the beginning of retirement, history shows that until fairly recently, it was common for men to be employed after they reached 65. In 1880, 76 percent of men were employed at age 65, a proportion that declined to 43 percent in 1940, and 18 percent in 1990.1 Although the current recession has caused more workers to postpone retirement, a 2009 survey of retirees found that 84 percent had entered retirement at age 65 or earlier.2
When Social Security was established in 1935, state pension systems were split equally between those that determined 65 as the retirement age and those that determined 70 as the retirement age. The Commission on Economic Security, which designed the system under FDR, was swayed to adopt age 65, partly because the federal Railroad Retirement System, which was established in 1934, used 65, and partly because analyses at the time showed that 65 was actuarially feasible at low levels of taxation.3
Since the program’s inception, the retirement age has undergone only two modifications. The first was the addition of the early eligibility age (EEA) (under which recipients receive partial benefits), which was created for women in 1956 and for men in 1961. The modification for women was motivated by politeness rather than policy considerations: Wives were generally three years younger than their husbands, and a gracious Congress wanted to allow couples to retire at the same time. Men were later offered early retirement as a mechanism to cope with high unemployment by encouraging workers to leave the labor force.4
The second modification was the increase in the normal retirement age (NRA) from 65 to 67, which was implemented in the early 1980s because Social Security was nearing a point where it did not have the funds to print benefits checks. In 1982, Chairman of the House Ways and Means Subcommittee on Social Security Congressman J. J. Pickle (D–TX) first proposed phasing in an increase to age 68 over a period of 10 years beginning in 1990. Pickle’s proposal was panned at first by Speaker Tip O’Neill in what Time magazine then dubbed “one of the more egregious examples of partisanship.”5 The Reagan Administration had landed itself in hot water over reform proposals it introduced in 1981, and O’Neill was not inclined to let Democrats absorb any of the heat with a proposal that would reduce benefits. In 1983, when a reform package was signed into law, it was clear that some increase in the retirement age would have to be part of the compromise. The Senate initially passed an increase to age 66, but Pickle was able to push through an increase to 67 in conference committee. Even then, the increase was not scheduled to take effect until 2000, when an increase to 66 was phased in year by year in two-month increments. A further increase to 67 using the same phase-in will begin in 2017.


1 Munnell and Sass, Working Longer, p. 36.

2Ruth Helman, Mathew Greenwald & Associates, Craig Copeland, and Jack VanDerhei, “The 2009 Retirement Confidence Survey: Economy Drives Confidence to Record Lows; Many Looking to Work Longer,” Employee Benefit Research Institute Issue Brief No. 328,April 2009, p. 15, at (November 10, 2010).

3It is sometimes claimed that 65 became the prevailing retirement age because Germany’s chancellor from 1862 to 1890, Otto von Bismarck, who created the world’s first social insurance program, was 65 at the time the German system was designed. In fact, Bismarck was 74 when the German program began, and the retirement age was set at 70. See Social Security Administration, “Frequently Asked Questions: Age 65 Retirement—The German Precedent,” at (September 11, 2010).

4Sylvester J. Schieber and John B. Shoven, The Real Deal: The History and Future of Social Security (New Haven, Conn.: Yale University Press, 1999), p. 133.

5George J. Church, “Social Security: A Debt-Threatened Dream,” Time, May 24, 1982, p. 11, at time/printout/0,8816,953496,00.html (November 10, 2010).

Increases in Longevity and Lifespan Differ Between Racial and Economic Groups

Increases in longevity have not been the same for all income and racial groups. Since 1950, overall life expectancy at birth for non-white males has risen at about the same rate as that for white males, while that for non-white women has risen at an even faster pace than for white women.[18] This still leaves a gap of six years between African-American and white men, and four years between African-American and white women.

However, much of this difference reflects those who die before age 65. There is a much smaller difference between racial and ethnic groups when life expectancy is measured at 65. Counting both genders together, an African-American who reaches age 65 has a life expectancy of 1.2 years less than a non-Hispanic Caucasian. However, a Hispanic worker at age 65 has a life expectancy that is 4.2 years longer than that of an African-American and 3 years longer than that of a non-Hispanic Caucasian.[19]

Similarly, the longevity gap between the highest and lowest income groups has grown since 1980, but a significant amount of that difference reflects those who die before reaching the age of 65. When life expectancy is measured at birth, the gap between the longest-expectancy and shortest-expectancy groups has increased by 1.7 years— from 2.8 years in 1980 to 4.5 years in 2000; the gap in life expectancy when measured at age 65 has risen by 1.3 years—from 0.3 years in 1980 to 1.6 years in 2000.[20] If trends continue, that gap can be expected to continue to widen. While this growing gap is very disturbing, it is not sufficient reason to justify leaving the current retirement ages in place. However, the situation should be closely monitored, and if the gap continues to widen, it could be a reason for a future revision to retirement standards specifically for lower-income workers.[21]

How to Increase Social Security Retirement Ages

Increasing the Normal Retirement Age. The 1983 reforms set up a process to gradually increase the NRA for workers who reached age 62. Starting in 2000, for workers born in 1938, the reforms increased the NRA by two months per year until it reached age 66 in 2005 for workers born in 1943. At that point, the process paused until 2017, when it will again begin to increase the NRA by two months annually for workers born in 1955 until it reaches 67 in 2022 for workers born in or after 1960.[22] A similar method should be used to further increase the retirement age to 68. Once the NRA has reached 68, it should then be indexed to reflect further changes in longevity.

There is little justification for retaining the delay in re-starting the process until 2017, since all this serves to do is delay improvement of the program’s finances. Instead, the increases to age 67 should begin in 2012 for people born in 1950. Once the NRA reaches 67 in 2017 for workers born in 1955, it should continue without interruption until it reaches 68 in 2023 for workers born in or after 1961.

Starting in 2025, the NRA should be indexed for further increases in longevity. This would create an automatic mechanism that, depending on the method chosen, would either further increase the NRA or reduce benefits unless the worker delayed retirement beyond 68 without requiring additional legislation. One method to index the NRA was found in a 2008 Social Security reform plan proposed by Representative Paul Ryan (R–WI),[23] which would adjust the NRA so that on average, a retiree who reached age 68 would receive benefits for 20 years.[24] According to the Social Security actuaries, this would result in a one-month increase in the NRA every two years.[25] Because of the delay in collecting accurate mortality and longevity data, the Social Security Administration (SSA) estimates[26] that it would be necessary to use data from three years before the date that it makes the calculations.

Another method, used in Sweden, calculates benefits payable to each retirement cohort by using a complex formula intended to ensure that each cohort receives roughly the same level of lifetime benefits. The formula includes changes in life expectancy as one of several factors that would increase the obligations of the system. The revised liability figure is then divided into the assets available to pay those benefits. If the result shows that the system is in imbalance, benefits are gradually reduced for each new retirement cohort until it comes back into balance.[27] While the Swedish mechanism is probably too complex for the U.S., and goes well beyond just adjusting to changes in longevity, it shows that such a move is possible. American legislators should include such an automatic mechanism in Social Security to prevent the political pain caused by having to legislate further increases in the NRA.[28]

Some may complain that starting the process of increasing the NRA in 2012 places workers born near 1950 and 1956 at a disadvantage, since they may have planned to retire at 66 and would now receive reduced benefits unless they delay retirement. But the change in these workers’ benefits would be fairly small, and workers could reclaim the full amount by delaying retirement by only a few months. It is far more important to both reduce Social Security’s fiscal problems and to increase economic growth. One could slightly increase those workers’ retirement benefits, but experience with “notch babies”[29] shows that even minor benefit differentials between retirement cohorts can create resentment and claims of unfair treatment that is easily manipulated by political interest groups. Thus, such adjustments should be avoided.

Increasing the Early Eligibility Age. A similar mechanism could be used to gradually increase the EEA from the current 62 to 65. Also in this case, the benefit age would increase by two months per year until it reaches its final scheduled level. After that, the EEA would be indexed in the same manner as the NRA. However, because increasing the earliest age at which a worker can receive Social Security benefits could have a real impact on retirement plans, the start of the process should be in 2015 for workers born in 1953, rather than using the 2012 start date for increasing the NRA.

Using this schedule, the EEA would reach 63 in 2020 for workers born in 1957, 64 in 2026 for workers born in 1962, and 65 in 2032 for workers born in or after 1967. After that, it would be indexed for further increases in longevity beginning in 2040.

Although this mechanism is essentially the same as that used to raise the NRA, its effect on a worker’s benefits would be very different. Increasing the NRA reduces the worker’s lifetime Social Security benefits, while increasing the EEA either has no effect on most workers’ lifetime benefits or slightly increases them. While this increased EEA will not cause problems for most workers, there will be some who are physically unable to work that long. (Discussed in more detail below.)

Additional Incentives to Work Longer

Increasing the retirement age also requires increased employment opportunities for older workers. While older workers offer a high level of skill and motivation, certain employers are reluctant to hire them. Similarly, some older workers with needed skills and knowledge seem determined to retire even though their employers still have a great need for them. As an additional incentive, a comprehensive Social Security reform plan should also include measures that would exempt both those workers who have reached their NRA, as well as their employers, from paying additional payroll taxes. This tax exemption would only apply to workers who delay taking benefits after their NRA, and would not apply to workers who delayed beyond their EEA.

In return for remaining in the workforce, and thus delaying taking their Social Security benefits until after their NRA, older employees would not have to pay 6.2 percent of their pay in payroll taxes, and could use that exemption to increase their take-home pay in advance of retirement. The Social Security benefits of workers who delay their retirement would be calculated using the work history they had prior to reaching their NRA, and just as they are currently, their benefits could be actuarially adjusted to reflect their delay in taking those benefits.

Thus, those retirees would receive higher take-home pay while they continued to work, as well as slightly higher benefits once they do file for benefits. Since employers would not have to pay the 6.2 percent match to the payroll taxes withheld from the pay of these workers, such a change would also lower their cost of hiring or retaining older workers. This second piece of the proposal should provide older Americans who need or want to work longer with additional employment opportunities along with the benefits of continuing to be a valued employee. However, it is important to stress that both the employer and the employee payroll tax exemptions are intended to be part of an overall reform of Social Security, and not as stand-alone legislation.

Higher Retirement Ages and Social Security Finances

Increasing the NRA Reduces Social Security Underfunding. Increasing the NRA will alleviate Social Security’s fiscal problems by reducing the $7.7 trillion gap (in net present value terms) between the level of benefits the program has promised and what it can afford to pay from its revenues. While the exact effect of increasing the NRA will depend on the details of the plan and the exact year in which it goes into effect, past scoring memos by SSA actuaries give some indication of the level of savings.

The SSA has issued a series of scoring tables showing the effects of a number of options for changing the program, including increasing the NRA to age 68 beginning in 2009. Based on the 2008 trustees report, the memo[30] found an improvement in Social Security’s long-range actuarial balance, which is the size of the surplus or deficit over the next 75 years as a percentage of taxable payroll,[31] of 0.58 percent, reducing the deficit from –1.70 percent to –1.12 percent.[32] While this is an improvement of about 35 percent of the total long-term 75-year deficit, increasing retirement ages alone does not eliminate the long-term problem: A higher retirement age adds about nine years to the life of the trust fund—in 2082, the program still faces a balance of –3.47 percent of taxable payroll, about 18 percent better than the 2008 trustees report estimate of a –4.20 percent.

Indexing the NRA after 2025 would result in additional improvements to the program’s finances. While not enough to close a substantial part of the remaining deficit, indexation is still an important feature of sound policy that will also help eliminate any future politicization associated with changing the eligibility age. Clearly, other changes to the program will also be necessary to make Social Security fiscally healthy over the long term.

Increasing the EEA Will Not Save Social Security, But Has Other Benefits. Some may argue that increasing the EEA will unfairly affect those workers who are physically unable to work longer, but the reality appears to be that only about 18 percent of workers must stop working before age 65, while over half claim benefits at age 62 alone. Rather than basing the EEA on the relatively small share of workers who face physical challenges, it would be far better for them to receive benefits from the disability program, while the much larger number of workers who remain able to work after age 62 is required to wait until a later age. This would both improve Social Security’s finances and maintain the safety net.

According to the actuaries at the Social Security Administration, increasing the EEA has “only a very small net effect” on the program’s finances.[33] This is due to the fact that Social Security benefits are calculated so that a worker living an average lifespan will receive the same amount of lifetime benefits regardless of his age when he applies for benefits. If the worker applies before reaching the NRA, the monthly amount he receives will be lower, but that reduction will be made up by the additional number of years during which he receives benefits. For SSA scoring purposes, those factors offset each other.

In practice, raising the EEA may increase at least some workers’ benefits. Since the SSA’s benefits formula indexes a worker’s past wages for the growth in wages between the time when they were earned and two years before a worker applies for benefits, raising the EEA would increase the number of years of past indexed wages. While the wages of a worker who retires at the current EEA of 62 would be indexed up to those earned when the worker turns 59,[34] the wages of those retiring at the higher EEA of 65 would be indexed for past wages earned through age 62. The additional three years would increase the real value of their Social Security benefits by a few percentage points. In addition, raising the EEA increases a married couple’s benefits.[35] This is due to the fact that survivors and spousal benefits (if any) are based on the primary earner’s retirement benefits. If the primary earner is forced to delay taking benefits, and his or her benefits are higher as a result, then both spousal and survivors’ benefits are also higher.[36]

To the extent that raising the EEA increases labor force participation, it is also likely to have beneficial effects for the overall economy. Most economic projections make the explicit point that in the intermediate future, the drop in the size of the labor force will depress economic growth.[37] Thus, any action that helps to arrest that decline by encouraging workers to remain in the labor force longer is almost certain to improve economic growth. Since the baby boom generation reaching retirement is much larger than cohorts that will succeed it, there should be little worry such a move would deny future workers job opportunities.

What About Those Who Physically Cannot Work Longer?

Increasing the retirement age for Social Security to age 68 by 2023 while increasing the early retirement age from 62 to 65 is not a step that The Heritage Foundation recommends lightly. Although many workers may wish to work longer, they may have health issues that make it impossible for them to delay retirement. Workers with physically demanding jobs are more likely to face these issues, though they are not limited to a particular job classification. Another factor is that some racial, ethnic, or income groups may have a higher incidence of disability than the overall population. These workers should not be penalized by increasing Social Security’s retirement ages, but those workers who can work longer should certainly be encouraged to do so.

Rather than leave Social Security’s current retirement ages as they are, workers who are physically unable to work longer should receive benefits through Social Security’s Disability Insurance (SSDI) program until they reach the EEA. However, SSDI faces continued managerial problems with long backlogs of unprocessed claims. While SSDI is a federal program, state agencies that are under contract to Social Security make the determination about physical and vocational aspects of a claim. The way that the SSDI program is currently administered, approval for disability benefits for a specific condition depends on the state in which a worker files the claim, and certain states take much longer to process a claim than others Several Social Security commissioners have focused a large part of their attention on this situation, but serious problems remain.[38] Even without the increase in the NRA and EEA, SSDI must be reformed so that it makes quick and uniform decisions about paying benefits to workers who are no longer able to work.

At the same time that Congress increases the NRA and EEA, it should also set some explicit performance standards for the disability program including uniform national deadlines for disability decisions that include both a sharp reduction in the time that it takes to get an initial determination and a greatly improved appeals process for those who are turned down when they first apply. Improving SSDI during the years while the increase in retirement ages is phased in is both achievable and essential.

However, the number of workers who need disability benefits should be limited, since, in theory, increasing the EEA would have little effect on retirees. As recently as 1960, the average age at which men claimed benefits was 66,[39] and the major health care advances since then should make it much easier for tomorrow’s workers to work until at least that age. Further, over the last several decades, the proportion of jobs that are physically demanding has shrunk from about 20 percent in 1950 to about 8 percent in 1996.[40]

It is true that a relatively small proportion of workers affected by an increase in the EEA would not be able to work longer. One study of workers who retired before age 65 found that 18 percent did so for health reasons.[41] The proportion of these workers was highest in the lowest income segments, declining as workers moved up the income scale.[42] Researchers have found a major improvement in the health of older Americans,[43] which should reduce the proportion of workers who are unable to work beyond today’s EEA, but there is some question whether that will continue into the future.

One study found that younger workers have a rising incidence of disability,[44] and suggested that the increasing number of overweight and obese adults may be at least partly responsible. The study found increases for both whites and non-whites in all education levels both within and without the labor force. If this remains the case, then at least some of the workers affected by an increase in the EEA will move directly from the labor force to SSDI.

To some extent, this shift into SSDI is inevitable. One study of the effects of the 1983 reforms showed that for every $5,000 decrease in the average present value of retirement benefits, SSDI enrollment increased by 0.6 percentage point, with the rate for men increasing by 0.8 percent and for women by 0.4 percent.[45] This study further estimates that because of the 1983 law, the SSDI enrollment rate for men was 0.58 percent higher for workers aged 45 to 64 in December 2005, and 0.89 percent higher for women.[46] If both the EEA and NRA are raised, it is very likely that enrollment in SSDI will increase even more.

Even assuming that the entire 18 percent of workers who retired before 65 due to health reasons will receive SSDI, the overall cost to Social Security is likely to be much lower than if the current retirement ages are retained, because there will be substantial cost savings from the 82 percent of workers who remain in the workforce longer than they otherwise would. Estimates from the SSA’s Office of the Chief Actuary show that while the increase in the number of those on disability will increase the long-range actuarial deficit by 0.01 percent of taxable payroll,[47] that amount is more than offset by the 0.58 percent of taxable payroll reduction which results from raising the retirement age to 68.[48]


Social Security is and should remain a part of Americans’ retirement security system. However, Americans of all ages are increasingly recognizing that hard decisions will be necessary if Social Security is to remain financially stable.[49] Future retirees will live much longer on average than their grandparents, and Social Security must change to reflect that increased longevity by increasing retirement ages for both full benefits and early retirement benefits. Otherwise, recipients will spend an ever higher proportion of their lives living at the expense of their children and grandchildren.

Merely increasing Social Security’s retirement ages will not solve the system’s financial pressures, but this step should be part of a major package of changes that includes increasing the rate of return that the program provides younger workers. Otherwise, the younger workers face a series of tax increases that will both reduce their standard of living during their careers and make it ever harder for them to save enough to supplement their Social Security income.

Social Security will begin to run continuous cash flow deficits as soon as 2017, and Congress does not have the luxury of delaying hard decisions—or of implementing them. Instead, it must have the courage to let some of today’s workers know that they will likely have to work longer before they can receive Social Security benefits. This may not be a pleasant message, but it is a fair one. Further delay will only force Congress to make an even more unpalatable decision as Social Security’s finances continue to worsen.

David C. John is Senior Research Fellow in Retirement Security and Financial Institutions in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation. The author thanks former Roe Institute Deputy Director Nicola Moore for all of her help in researching this paper.

About the Author

David C. John Senior Research Fellow in Retirement Security and Financial Institutions
Thomas A. Roe Institute for Economic Policy Studies

Show references in this report

[1]Mary Williams Walsh, “Social Security to See Payout Exceed Pay-In This Year,” The New York Times, March 24, 2010, at 10, 2010).

[2]Net present value measures the amount of money that would have to be invested today in order to have enough money on hand to pay deficits in the future. In other words, Congress would have to invest $7.7 trillion today in order to pay all of Social Security’s promised benefits between 2016 and 2083. This money would be in addition to what Social Security receives during those years from its payroll taxes.

[3]William W. Beach and Gareth Davis, “Social Security’s Rate of Return,” Heritage Foundation Center for Data Analysis Report No. 98-01, January 15, 1998, at

[4]David C. John, “How to Fix Social Security,” Heritage Foundation Backgrounder No. 1811, November 17, 2004, at

[5]Stuart Butler and Maya MacGuineas, “Rethinking Social Insurance,” Heritage Foundation and New America Foundation White Paper, February 19, 2008, at

[6]For details on the Automatic IRA, see David C. John, “Automatic IRA Builds Retirement Security,” Heritage Foundation WebMemo No. 2789, February 5, 2010, at

[7]This paper uses the term “normal retirement age” (NRA) to indicate the age at which a worker can receive full benefits, and “early eligibility age” (EEA) to indicate the age at which a worker can first receive early retirement benefits.

[8]There are some forecasts suggesting that the pace of future longevity increases will slow in the future. See “Pension Groups Warned Over Longevity Model,” Financial Times, July 13, 2010.

[9]“Factbox: European Retirement Ages on the Rise,” Reuters, June 24, 2010, at (November 10, 2010).

[10]National Center for Health Statistics, “Health: United States, 2007,” Table 27.

[11]Social Security Administration, “History: Life Expectancy for Social Security,” at (November 10, 2010).

[12]John Rust and Christopher Phelan, “How Social Security and Medicare Affect Retirement Behavior in a World of Incomplete Markets,” Econometrica, Vol. 65, No. 4 (July 1997), p. 825, at (September 11, 2009). Rust and Phelan also point out that when they take Social Security out of their model, the peaks in retirements at ages 62 and 65 disappear, just as they did in similar studies.

[13]Alicia H. Munnell and Steven A. Sass, Working Longer: The Solution to the Retirement Income Challenge (Washington, D.C.: The Brookings Institution, 2008), p. 57.

[14]Rust and Phelan, “How Social Security and Medicare Affect Retirement in a World of Incomplete Markets.”

[15]Dan Muldoon and Richard W. Kopcke, “Are People Claiming Social Security Benefits Later?” Center for Retirement Research at Boston College Issue Brief No. 8–7, June 2008, p. 2.

[16]Memorandum from Stephen C. Goss to Representative Paul Ryan, May 21, 2008, Social Security Administration, p. 4, at (September 11, 2010). However, such a move could have an effect on individual benefits as discussed later in this paper. In addition, it would increase the number of people claiming disability, which is also discussed later in this paper.

[17]Eric D. Beinhocker, Diana Farrell, and Ezra Greenberg, “Why Baby Boomers Will Need to Work Longer,” The McKinsey Quarterly, November 2008, at (September 11, 2010).

[18]Joyce Manchester and Julie Topoleski, “Growing Disparities in Life Expectancy,” Congressional Budget Office Economic and Budget Issue Brief, April 17, 2008, p. 1.

[19]U.S. Census Bureau, “2004 Life Tables—Projections of the United States by Age, Sex, Race, Hispanic Origin, and Nativity: 1999–2100,” reproduced in “Retirement Security for Latinos: Bolstering Coverage, Savings, and Adequacy,” The Retirement Security Project and National Council of La Raza, July 2005, p. 6.

[20]Manchester and Topleski, “Growing Disparities in Life Expectancy,” p. 2.

[21]If this longevity gap between different population cohorts grows, there are a number of potential solutions that could be examined, including allowing workers to stop paying Social Security taxes after a certain number of years. See Munnell and Sass, Working Longer, p. 130, for a discussion of this option.

[22]Social Security Administration, “Retirement Planner: Retirement Benefits by Year of Birth,” at (September 11, 2010).

[23]Portions of the language scored by the SSA were incorporated into H.R. 6110, the Roadmap to America’s Future Act of 2008.

[24]The actual language in H.R. 6110 is “maintai[n] the ratio between life expectancy at NRA and the difference between the NRA and 20.”

[25]Goss Memorandum to Ryan, Social Security Administration, p. 4.

[26] Ibid.

[27]Ole Settergren, “The Automatic Balance Mechanism of the Swedish Pension System,” The National Social Insurance Board Working Papers in Social Insurance, August 20, 2001, pp. 10–11.

[28]Regardless of the method used, adding longevity indexing to Social Security would also protect future retirees if life expectancy ever started to decline. While a decrease in longevity is very unlikely, barring some massive economic disruption, at least one study suggests that a sedentary lifestyle could cause some younger workers to be more likely to have significant health problems than those who are a decade or so older who had more active lifestyles. If this and similar studies prove to be correct, projected growth in lifespan could be lower than forecast. The probability of a reduction in longevity actually happening is far too low to undermine the case for increasing Social Security’s eligibility ages, and it makes the case for automatic longevity indexation even stronger.

[29]In 1972, Congress implemented an automatic cost-of-living adjustment for Social Security benefits, but used a faulty formula that counted inflation twice, and thus increased benefits to unsustainable levels. To correct the problem, Congress further revamped the benefits formula in 1977 to lower benefit levels but exempted those already in retirement. To minimize the loss that the switch to the new calculation generated for those close to retirement, a “transitional computation method” was devised for those born between 1917 and 1921, also known as the “notch.” Individuals born during that time period, or “notch babies,” received benefits that were nominally lower than comparable “pre-notch” babies, resulting in substantial outcry and an investigation by the Commission on the Social Security ‘Notch’ Issue in 1994. Most of the outrage was hype, as the commission found that there was no attempt to single out the “notch babies” and that those affected will still receive a greater return on their Social Security taxes than subsequent generations. For further information, see Social Security Administration, “Final Report on the Social Security ‘Notch’ Issue,” December 31, 1994, at (September 11, 2010).

[30]For estimates for increasing the full retirement age, see Social Security Administration, “Actuarial Publications: Summary Measures and Graphs,” at (September 13, 2010). These estimates are based on the 2008 trustees report.

[31]“Percent of Taxable Payroll” represents the proportion of wages and self-employment income that is subject to Social Security payroll taxes. Since payroll taxes are only assessed on the first $106,000 (in 2009) of salary income, this number is lower than the total wage and salary bill. It also does not include other types of income.

[32]The SSA estimates of Social Security’s long-run actuarial deficits do not include the cost of repaying the trust fund. Thus, the actual cash flow cost of financial Social Security benefits for the next 75 years is far higher than this estimate indicates.

[33]Memorandum from Stephen C. Goss and Alice H. Wade to Jeffrey Liebman, Maya MacGuineas, and Andrew Samwick, Social Security Administration, November 17, 2005, p. 3, at Liebman_20051117.pdf (September 11, 2010).

[34]This calculation assumes that a worker applies for benefits that begin payment on the date of his 62nd birthday, and thus last earns wages when he is 61.

[35]Steven A. Sass, Wei Sun, and Anthony Webb, “Why Do Married Men Claim Social Security Benefits So Early? Ignorance or Caddishness?” Center for Retirement Research at Boston College Working Paper 2007-17, October 2007, p. 25, at (September 14, 2010).

[36]In practice, this would only apply in a system with a higher EEA and NRA if benefits are higher at the new EEA than they are today.

[37]Kevin L. Kliesen, “As Boomers Slow Down, So Might the Economy,” The Regional Economist, July 2007, pp. 12–13, at (September 14, 2010).

[38] As one example of many reports on the problems of SSDI, see Government Accountability Office, Social Security Disability: Better Planning, Management, and Evaluation Could Help Address Backlogs, GAO-08-40, December 7, 2007, at (November 8, 2010).

[39] Munnell and Sass, Working Longer, p. 144.

[40] Ibid., p. 94, and Eugene Steuerle, Christopher Spiro, and Richard W. Johnson, “Adjusting for Life Expectancy in Measures of Labor Force Participation,” Straight Talk on Social Security and Retirement Policy 10 (Washington, D.C.: Urban Institute, 1999).

[41]Alicia H. Munnell, “Policies to Promote Labor Force Participation of Older People,” in Workforce Policies for a Changing America,Harry Holzer, ed. (Washington, D.C.: Urban Institute Press, 2007).

[42]Munnell and Sass, Working Longer, p. 31.

[43]Darius N. Lakdawalla, Jayanta Bhattacharya, and Dana P. Goldman, “Are The Young Becoming More Disabled?” Health Affairs, Vol. 23, No. 1 (2004), p. 168.

[44] Ibid., p. 172.

[45]Mark Duggan, Perry Singleton, and Jae Song, “Aching to Retire? The Rise in the Full Retirement Age and Its Impact on the Disability Rolls,” National Bureau of Economic Research Working Paper No. 1181, December 2005, p. 4.

[46] Ibid.

[47]Traditionally, Social Security’s finances have been expressed in terms of the percent of taxable payroll. Taxable payroll is considered to be the total amount of wages and salary that are subject to the payroll tax.

[48]For estimates for increasing the early retirement age, see Social Security Administration, “Actuarial Publications: Summary Measure and Graphs,” at (September 14, 2010). For estimates on increasing the full retirement age, see Social Security Administration, “Actuarial Publications: Summary Measures and Graphs,” at (September 14, 2010). These estimates are based on the 2008 trustees report. However, because Social Security’s finances are very complex, and various parts influence others, one should not assume that one can simply add or subtract various proposals. In fact, those interactions may change the overall result to some degree. However, it is unlikely that changing the two retirement ages (EEA and NRA) proposed here would interact in a way that would come close to cancelling each other out. Thus, we can safely assume that the overall result would still be a substantial savings.

[49]Susan Page, “Poll: Faith in Social Security System Tanking,” USA Today, July 20, 2010, at (November 10, 2010).