Don’t Subsidize ‘Middle-Age Retirement’

The author argues that his fellow liberals shouldn’t oppose an increase in the minimum age for claiming Social Security.

The idea of changing the retirement age, or the age of eligibility for Social Security benefits, has long been part of a narrow debate about how to ensure the long-term fiscal sustainability of that program. But we should really confront the retirement age as part of the much broader debate about societal progressivity: Who in our society has the ability to give more, who needs more support, and when do they need it?

The income tax, for example, is progressive in that households with higher incomes are expected to contribute a higher share of it, while lower-income households pay little or no tax. In the case of the retirement age, we should be asking whether we really intended to create what Social Security has become—increasingly a retirement system for people in middle age, most of whom report good or excellent health.

For over half of beneficiaries retiring in their early 60s today, the result is that one member of a couple with average life expectancy spends well more than a quarter century retired. We need to ask whether such a system leaves Americans adequately prepared for old age and, more broadly, allows society to take advantage of older citizens’ tremendous potential. That is, does the typical 62-year-old have more ability to give and less need for support than the typical 90-year-old?

It is true that many people in late middle age need help, for reasons of health or the economy. But let’s be honest: Most of the benefits from expanding years of retirement support have gone to relatively healthy middle-income and affluent people, who still have much to contribute. It’s not just that we and everyone else get these extra years of benefits but that those of us with above-average income get higher levels of benefits, often twice as much or more than those with lower lifetime incomes.

Think of it this way: Social Security has expanded years of support universally, resulting in both lower revenues and a lost opportunity to target resources progressively to the disabled, those with low- and moderate-lifetime incomes, and older individuals with greater health needs and fewer capabilities to care for themselves. Such a program, whether chosen or arising by accident, is akin to an across-the-board cut in tax rates that seems fair on the surface but happens to give thousands of dollars in cuts to higher-income taxpayers as a way of giving much less to middle- and low-income taxpayers.

Long-term unemployment and disability are real issues, but they don’t really depend on whether one is 62 or 65. Many workers in their 50s and younger also find themselves unable to handle the physical demands of their jobs or adjust to the fact that the job in which they’ve invested no longer exists anywhere, and no other position currently matches their skills. Nor, despite claims to the contrary, do more years in retirement disproportionately benefit low-income minorities. Yes, some have lower life expectancies, but many of those people die before age 62 or receive disability payments that don’t change with the retirement age. If we really wanted to help disadvantaged minorities and other workers with lower lifetime earnings, we could use the money more progressively by targeting benefits directly to them, through devices like minimum benefits and removal of the discrimination against parents — minority women particularly — who were not married and therefore get nothing out of the current spousal and survivor benefits, although they pay for them.

Unfortunately, in negotiations with the several commissions and task forces on the budget deficit in 2010, many liberals focused so intently on protecting the current benefit system that they overlooked real opportunities to improve the equity and progressivity of our retirement system. The consequence is to endanger a once-in-a-generation chance to improve benefits and security for those in old age, including reductions in poverty and near poverty among the elderly.

The question we should be asking is a simple one: For how many years should people who are not disabled receive benefits to support their retirement?

A single person retiring at the earliest retirement age (65) in 1940 was likely to receive close to 14 years of benefits. By 1975, this person could retire at 62 and would likely receive benefits for 19 years, and today that number has risen to almost 22 years. When it comes to couples retiring at the earliest retirement age, by 2030, at least one partner is expected to receive benefits for nearly 30 years.

Is the decision to provide benefits for an ever-increasing number of years one that we, as a society, have made thoughtfully? Is it one that we would have made independent of other societal needs, such as children’s programs?

Put another way, if you define old age by, say, the last 15, 18, or even 20 years of life, then Social Security has become, almost by definition, a middle-age retirement system.

Arguments about the retirement age often center on a chronological age like 65, the normal (and early) retirement age when Social Security was first created in 1935. However, the number of years spent in this undifferentiated period known as retirement is all the more dramatic when we look to both actual retirement ages and life expectancy. Americans on average retired at age 68 in both 1940 and 1950. If they were to retire for the same number of years on average today as they did in 1940, they would be retiring at age 75. In 2070, they’d quit working at about age 80. Instead, they retire on average at about age 64.

At the same time that people are retiring earlier and living longer, Americans have been having fewer children. Unlike increases in life expectancy, which doesn’t really increase the proportion of the population in the last years of life, fewer births do create fewer younger people (say, in the first three-quarters of their lives) per older person (say, in the last quarter of her life).

These demographic effects — longer lives, more years of support, and lower birth rates — can be combined. Rather than four workers per Social Security beneficiary in 1965, or the current ratio of 3-to-1, we are projected to move to 2.1 workers for every Social Security recipient by 2035. Social Security costs, as a share of gross domestic product, in 2030 will be about twice as high as they were in 1970, mainly because of these demographic effects.

These demographic pressures stretch well beyond Social Security. But so would the benefits of encouraging people to spend fewer years in retirement and more time in the labor force. To the extent we can improve employment rates, we can also increase national income and personal income and in turn, raise income-tax revenues and therefore the level of benefits that government can provide to the truly needy.

Within the Social Security system, reducing the incentives for early retirement would boost payroll-tax revenues. This could allow the system either to provide a higher level of benefits for all or to concentrate benefits progressively on those who are older and who, in our current middle-age retirement system, receive increasingly smaller shares of total benefits over time. Less than 35 percent of benefits to men go to those expected to live 10 or fewer years; in 1968 these older recipients received more than half of all benefits.

Here’s one way this shift toward a more progressive system could work: Suppose reform gradually increases the earliest retirement age, which today is 62. For an average couple retiring at that age, at least one of the two (more often the woman) is likely to make it to about age 90, and that life expectancy is rising.

Under current practices, Social Security balances would change little because that loss of any one year of benefits for some workers is offset by higher benefits they receive later. If people would work one year longer, however, personal incomes would rise by about 8 percent for every year of life thereafter. And as people earn more and delay drawing benefits, they receive higher annual Social Security benefits, draw down their private retirement balances for fewer years, and save longer. The individual or family is likely to be far better off and so, too, is the government — but mostly through higher income-tax revenues. In fact, at any tax rate, the additional revenues to Social Security would support higher average lifetime Social Security benefits as well.

The phenomenon of continually expanded years of retirement, however, is not just a question for Social Security or government policy more generally. It is a question about how we decide to live our lives and best use our nation’s pool of human capital in the 21st century. In the next few decades, older workers will be seen as women were in the last half of the 20th century: the largest pool of underutilized talent in the economy. If we modernize the signals and symbols around the idea of when old age begins — not just in Social Security but in the private sector as well — labor demand is likely to shift toward this pool of talent.

This argument is supported by evidence from history. Over the last half of the 20th century, the entry of women and baby boomers into the workforce made it easy to meet labor demand and allow other workers to retire for ever more years. The employment rate of adults, measured as a percentage of all adults, has increased in almost every non-recession year. That suggests that it’s a mistake to project the future work of those aged 62 to 68 by their recent employment rates, just as it was wrong to project the labor supply of women in 2000 by what they were doing in 1950. And increasing employment for older Americans need not come at the expense of younger workers, because the income older workers earn translates to more demand and more jobs for all. Put another way, by inefficiently pushing people into retirement, we’re reducing their output and personal income in ways that reverberate throughout the economy in the same way as an increase in unemployment does.

We are at a time when we need to look at all the elements of our nation’s social contract — those we created deliberately and those that arose by accident — and question whether they are achieving the goals of social justice and progressivity as well as making the best use of our most valuable resource — our people. The unintended transformation of Social Security into a middle-age retirement system has proved to be a poor way of promoting progressivity and dealing with larger macroeconomic questions, and a gradual shift in the other direction would have countless benefits for older Americans, those with greater needs among old and young alike, and the economy as a whole.

Eugene Steuerle holds the Richard B. Fisher Chair at the Urban Institute in Washington, D.C. and writes the on-line column “The Government We Deserve.”