Keywords

1 Introduction

Older Americans on the cusp of retirement today face a very different retirement income landscape than prior generations faced, with a higher degree of exposure to market forces and a higher risk of financial insecurity later in life (Quinn & Cahill 2016, in press).Footnote 1 This exposure and risk are due to a variety of interactions among demographic changes, the evolution of the traditional three pillars of retirement income (Social Security, private pensions, savings), and market volatility. Public sector budgets, already strained, will be strained even further with the continued retirements of the baby boomers and the individual challenges they will face in maintaining their standard of living in retirement. At the same time, older Americans have exhibited a remarkable degree of flexibility when it comes to continued work later in life, both by remaining in (and returning to) the labor force and by transitioning into new jobs, often in new lines of work. This flexibility of older American workers is a reason for some optimism as we confront the stresses and strains of an aging society.

In many ways, the changing retirement income landscape presents a perfect storm that threatens financial security later in life. First, the long-term demographic changes behind societal aging—increases in longevity and declines in fertility—mean that the ongoing shift in the age distribution will be permanent (Ortman, Velkoff, & Hogan, 2014). The combination of low fertility rates during the Great Depression and high fertility rates following WWII resulted in a pause in societal aging from 1990 to 2010, followed now by a rapid increase that will continue through 2030 (Werner, 2011). Second, the retirement income landscape has evolved over the past three decades and continues to do so, altering the retirement incentives and opportunities that older Americans face (Cahill, Giandrea, & Quinn, 2015a). The well-documented transition from defined-benefit (DB) to defined-contribution (DC) pension plans in the private sector is an important example (Copeland, 2009; Munnell, 2014). Finally, the changing retirement landscape has increased the exposure of older Americans to market forces at a time when the macroeconomic environment is volatile.

This chapter summarizes the changes to the retirement income landscape, highlighting some key causes for concern that are associated with societal aging. We then describe how older Americans have responded to these changes in recent years. Their remarkable flexibility and ability to adapt are noteworthy in many respects, and may just be the antidote to our societal aging challenges.

2 Causes for Concern

2.1 Demographic Changes

The story of societal aging in America has long-term and short-term components. The long-term components are increases in life expectancy and decreases in fertility. Life expectancy at birth for men and women were, respectively: 46 and 48 years in 1900; 66 and 71 years in 1950; 70 and 77 years in 1980; and 76 and 81 years in 2014 (National Center for Health Statistics, 2016). The same is true for life expectancy at age 65. Between 1950 (when these data were first available) and 1980, life expectancy at age 65 increased 1.3 years for men (to 14.1 years) and 3.3 years for women (to 18.3 years), and between 1980 and 2014 it increased another 3.9 years for men (to 18.0 years) and 2.2 years for women (to 20.5 years; National Center for Health Statistics, 2016). These steady gains in life expectancy contribute to the societal aging expected over the next several decades.

The second long-term trend is a pronounced decline in fertility rates dating as far back as the early 1800s, when the average number of births per American female was approximately seven (Munnell, Cahill, Eschtruth, & Sass, 2004). By the beginning of the twentieth century, this figure had dropped nearly in half, and the decline persisted through the Great Depression. Fertility jumped temporarily to about 3.5 births per female following World War II (a reversal that has important implications today, some 65 years later), but then the long-term decline resumed and fertility rates eventually stabilized in the 1990s at approximately 2.0. The decline in fertility means that societal aging is not just a baby boomer phenomenon. If the boomers had had children at the same rate as prior generations did, the shift toward a more rectangular population distribution would not have occurred.

The important short-term component of societal aging has to do with two brief breaks from the long-term fertility trend, one during the Great Depression and the other following World War II. As noted above, fertility rates declined during the economic hardships of the 1930s, and these lower rates—decades later—yielded a brief respite from the tide of societal aging. Between 1990 and 2010 the percentage of the population aged 65 and over changed little, from 12.6 to 13.0% (Werner, 2011; see Fig. 7.1). In contrast, the temporary spike in fertility following World War II is now generating a rapid aging of our population. The percentage of the population aged 65 and older—13.0% in 2010 and 14.9% in 2015—will increase to nearly 20% by 2030, an increase of over 50% in just 20 years, and is expected to remain between 20 and 22% between 2030 and 2050 (Ortman et al., 2014; U.S. Census Bureau, 2016a, 2016b). So, while demographers and researchers have raised awareness about societal aging, in practice such changes are only now just starting to take place. Individuals, employers, and society as a whole are about to witness an unprecedented demographic change over the next 10 years.

Fig. 7.1
figure 1

Percent of the US population aged 65 or older, 1880–2080

Full size image

2.2 The Evolution of Retirement Income Sources

Important changes to Social Security, private pensions, and savings have occurred over the past three decades, with some directly related to societal aging.

2.2.1 Social Security

One well-known change is the gradual increase in Social Security’s Full Retirement Age (FRA) from 65 to 67 per the 1983 Social Security reforms (U.S. Social Security Administration, 2018a, 2018b). The increase in the FRA began with those individuals who turned 62 in 2000 (i.e., born in 1938; U.S. Social Security Administration, 2018a, 2018b). Their FRA was increased by 2 months, from age 65 to age 65 plus 2 months. The FRA was then increased an additional 2 months per year until it reached age 66 for those who turned 62 in 2005 (i.e., born in 1943). These increases were followed by a 12-year hiatus, after which the increases resumed with those who turned 62 in 2017 (born in 1955). The FRA is now scheduled to increase 2 months per year until it reaches age 67 for those who turn 62 in 2022 (born in 1960 or later). No further increases in the FRA are currently legislated although they would likely be considered in any further Social Security reform.

Other important changes to Social Security have taken place as well. Individuals who receive Social Security benefits prior to their FRA and earn income above a threshold ($17,040 in 2018) have a portion of their benefits withheld according to Social Security’s retirement earnings test (U.S. Social Security Administration, 2018c). Prior to 2000, the earnings test also applied to individuals older than their FRA (Burtless & Quinn, 2002; Gruber & Orszag, 2003). Importantly, although not generally known, benefits withheld due to the earnings test are returned to the beneficiary when they reach their FRA in the form of a subsequent actuarially fair increase in monthly benefits. Still, the earnings test is oftentimes viewed as a work disincentive and, in turn, its removal is commonly viewed as a work incentive. Given the need to promote continued work later in life (discussed below), the elimination of the earnings test above the FRA is a positive step.

Another important pro-work change in the Social Security program was the adjustment to the Delayed Retirement Credit (DRC), the percentage by which monthly benefits are increased for each year benefit receipt is delayed beyond the FRA. The gradual increase in the DRC (from 3 to 8% for each year of delay beyond the FRA) removed a significant work disincentive (an implicit pay cut for those who worked beyond their FRA) by making expected lifetime benefits approximately the same for the average worker regardless of when they are first claimed between the FRA and age 70. With this incentive removed, Munnell (2013) argues that the “true” Social Security retirement age is really age 70, as the FRA has little meaning in terms of one’s expected lifetime Social Security benefits before then.

Social Security is by far the most important of the three traditional sources of retirement income, providing over 80% of retirement income for those age 65 and older in the bottom two income quintiles and about two thirds and 40% of retirement income for the next two quintiles, respectively (U.S. Social Security Administration, 2016a, 2016b). The program faces a long-term financial shortfall. Since 2010, outlays have exceeded revenues, absent interest on the Trust Fund (Fig. 7.2; Congressional Budget Office, 2016). The gap between outlays and tax revenues is expected to widen considerably over the next 20 years as the baby boomers continue their transition from paid employment to retirement. By 2034 the Trust Fund is expected to be exhausted, at which point payroll taxes would be sufficient to cover only about 80% of scheduled benefits (Board of Trustees of OASDI, 2017). Numerous options to avoid such an outcome were proposed by a commission of experts though little action has taken place since the report was released (National Commission on Fiscal Responsibility and Reform, 2010).

Fig. 7.2
figure 2

Social Security tax revenues and outlays with scheduled and payable benefits, 1985–2085

Full size image

2.2.2 Private Pensions

Private pensions have also been impacted. Prior to the 1980s, private sector pensions were dominated by defined-benefit (DB) type plans, in which beneficiaries are paid an annuity typically based on years of service and some measure of final average salary. With increases in longevity and in administrative costs from government regulations, however, employers began shifting away from defined-benefit plans to defined-contribution (DC) plans, which operate like a tax-deferred individual savings account (Butrica, Iams, Smith, & Toder, 2009). In addition to the tax incentives (workers do not have to pay taxes on dollars contributed to their accounts until they are withdrawn), some employers also offer matching contributions, providing an additional incentive to participate.

The transition from DB to DC plans was remarkably quick, with little change in the prevalence of overall pension coverage. About one half of private sector workers participate in an employment-based retirement plan. The actual level of participation has been the subject of some debate, as Munnell (2014) and Morrissey (2016) found that pension participation has declined slightly in recent years while Dushi, Iams, and Lichtenstein (2015) argue that participation has increased. Little disagreement exists, however, with respect to the dramatic shift in the type of pensions offered in the private sector. The percentage of private sector workers with a defined-benefit plan only declined from more than 60% in 1980 to less than 10% by 2006, and to less than 5% today (Fig. 7.3; Employee Benefit Research Institute, 2018). In contrast, fewer than 10% of private sector workers had a defined-contribution plan in the early 1980s, compared with more than 70% today. The percentage of workers with both a DB and DC plan is currently the same as it was in 1980 (23%) and have remained between 20 and 35% over the past three decades.

Fig. 7.3
figure 3

Private-sector participants in an employment-based retirement plan, by plan type, 1979–2014 (among all workers)

Full size image

DC plans do have certain advantages. They are portable from employer to employer, and the current value of the assets (which can rise or fall) is clear. On the other hand, DC plans often require workers to decide whether or not to enroll (about 1 in 5 eligible employees do not participate), how much to contribute, how to invest the funds, whether to withdraw assets before retirement, and when and how (via lump sum and/or annuity) to withdraw the assets remaining at retirement (Munnell, 2014; Munnell & Sundén, 2004). DC plans expose workers to both market and longevity risk (living longer than anticipated), two risks assumed by employers in traditional DB plans.

2.2.3 Savings

The traditional third pillar of retirement income, individual savings, is of modest importance for the majority of older Americans. Excluding home equity and the value of defined-benefit pensions, the typical American has less than $25,000 in financial assets, and more than one quarter of American workers have less than $1000 in financial assets (Helman, Copeland, & VanDerhei, 2016). These percentages, combined with the fact that roughly one half of private sector workers do not participate in an employer pension, are consistent with the income statistics mentioned above, namely, that Social Security benefits constitute 80% or more of retirement income for the bottom 40% of older Americans (U.S. Social Security Administration, 2016a). For a sizable minority of older Americans, Social Security provides the only lifeline between a decent standard of living and poverty. For example, the 2017 Social Security retirement benefit for an individual with career average earnings equal to the national Average Wage Index (AWI) was $20,190, or less than $1700 per month, if benefits are claimed at the FRA. For an individual with career average earnings of 45% of the AWI, the benefit is $12,256, or about $1000 per month (Board of Trustees of OASDI, 2017).

A review of poverty rates from 1959 to 2016 illustrates this point (Fig. 7.4). In 1959, roughly one out of three older Americans was in poverty. This poverty rate was higher than that of children or other adults. The poverty rate for older Americans fell precipitously during the 1970s when real Social Security benefits increased, and then fell gradually thereafter to the point where today, the poverty rate of older Americans (9.3%) is below the rate of other adults (11.6%) and far below the rate of children (18.0%; Semega, Fontenot, & Kollar, 2017). In many ways, the decline in poverty among older Americans is a dramatic success story though it is important to keep in mind that poverty rates among older Americans continue to vary significantly by race, marital status, and gender (Quinn & Cahill, 2016). For example, in 2014, when the overall poverty rate for those aged 65 and older was 10%, Black men and women aged 65+ (with poverty rates of 17 and 21%), and Hispanic men and women aged 65+ (16 and 20%) were much more likely to be poor than White men and women aged 65+ (6 and 11%). Women aged 65+ were more than 60% more likely to be poor than older men (12% vs. 7%), and nonmarried men aged 65+ (13%) and women aged 65+ (18%) were much more likely to be poor than married couples aged 65+ (5%; U.S. Social Security Administration, 2016a).

Fig. 7.4
figure 4

Poverty rates by age, 1959–2016

Full size image

A key question is whether these relatively low levels of poverty will be maintained in the future. Importantly, the increase in Social Security’s FRA implies a reduction in benefits for those who claim benefits at age 62, the first year of Early Eligibility Age (EEA). When the FRA was 65, the reduction in benefits for claiming at age 62 was 20%. With an FRA of 67, the reduction is 30% (Board of Trustees of OASDI, 2017). This 30% reduction might be of only modest consequence if other income sources including earnings were available in one’s 60s, but the reduction would be much more consequential if Social Security became the sole source of income when retirees reached their 80s or 90s. Further, among those fortunate enough to have a private pension plan, the switch from DB to DC plans leaves these individuals more exposed to market forces. While conceptually it is possible to rebalance one’s portfolio as one ages, most do not, leaving individuals vulnerable to market fluctuations (Coombes, 2015).

2.3 Macroeconomic Volatility

Macroeconomic changes present additional challenges for older Americans. For one, as more older Americans are responsible for the investment risk of their assets, the stock market has fluctuated considerably. For example, in the past two decades, the Dow Jones Industrial Average has dropped more than 25% on three different occasions (Wall Street Journal, 2016). While the markets eventually rebounded, each run-up exposed investors to the risk of another sharp decline. Indeed, as we write this chapter, the market has experienced a prolonged upward trajectory. What happens next is uncertain, but the chance of another sharp decline is ever present, and older Americans who are significantly invested in equities run the risk of losing a sizable portion of their assets, with little time to recover from their losses.

Another macroeconomic challenge is the ability of the federal government to mitigate the effects of a downturn similar in magnitude to the recent Great Recession and the ensuing historically sluggish recovery (Desilver, 2014). One response to the Great Recession, for example, was a sizable increase in deficit spending. Figure 7.5 shows the cumulative impact of these large annual deficits over the past decade. As a percentage of GDP, federal debt held by the public is currently higher than it has ever been, with the one important exception of the short-lived, high-debt period following World War II (Congressional Budget Office, 2017). Moreover, the CBO projects the federal debt to increase sharply over the next 20 years. With the federal debt projected to far exceed 100% of GDP, the ability of future governments to rely on deficit spending to address periods of economic weakness might be limited, leaving older Americans and Americans in general economically vulnerable.

Fig. 7.5
figure 5

Federal debt held by the public as a percentage of GDP, 1790–2050

Full size image

Finally, it is worth noting that the interaction between these risk factors could play an important role in the economic vulnerability of older Americans. The volatility in equity markets, for example, comes at a time when older Americans’ retirement income sources are more exposed to investment risk than they were in the past. These interactions present additional risks over and above those posed individually by the evolution of Social Security, private pensions, and savings.

3 Reasons for Optimism

Individuals have little or no influence over future Social Security, Medicare, or Medicaid reform, or over employers’ decisions regarding pension coverage, pension type, or post-retirement health insurance. Additional savings later in life can help a little, but nowhere near as much as saving done much earlier, with decades of asset accumulation ahead. If one cannot rewrite one’s savings history and cannot personally influence important Social Security parameters, trends in employer pensions, or the details of medical insurance coverage, what can one do?

3.1 Older Americans Are Extending Their Work Lives

One option which many have is to continue working later in life, and many older Americans are doing just that. Additional years of work can have a profound impact on economic security in retirement, as each year of work increases important retirement income flows (additional earnings and savings, and larger Social Security and pension benefits) that can be allocated for retirement and reduces the number of years in retirement that need to be financed. For example, the Congressional Budget Office (2003) examined the level of assets needed in retirement for a married couple earning $62,000 per year before taxes. This couple would need more than $330,000 in assets to finance their retirement years if they retired at age 62. In contrast, if the couple worked until age 67, they would need less than $160,000 in assets—a reduction of more than 50% with only 5 years of additional work. Munnell and Sass (2008) illustrate the point more simply using the example of an individual who works from age 20 to 60 (40 years of work) and lives to age 80 (20 years in retirement). The ratio of working years to retirement years is 2:1. With just 5 years of additional work, the ratio of work years (45) to retirement years (15) is increased by 50%, from 2:1 to 3:1.

Fortunately, the road to continued work later in life among older Americans has been well paved as a century-old trend toward earlier and earlier retirement came to a halt in the 1980s (Quinn, Cahill, & Giandrea, 2011). Since 1910, the average age of retirement for men, defined here as the age at which the labor force participation rate drops to 50%, declined steadily from age 73 to 65 in the mid-1970s and to age 63 in 1985 (Burtless & Quinn, 2002). Dora Costa (1998) documents similar dramatic declines in the employment rates of men aged 55–64 and 65+ going back to 1880. But since the mid-1980s, these trends have all reversed. The labor force participation rates of older American men are higher than their pre-1985 trends would have predicted (Quinn & Cahill, forthcoming). For example, as described in Chap. 1, among men aged 65, labor force participation rates increased by about 50% between 1985 and 2016; at age 68, they are up about 60%; and even at age 70, participation rates have increased by about 70%.

The labor force participation rates of older American women have changed since the mid-1980s as well. During the 20-year period between the mid-1960s and the mid-1980s, labor force participation rates for older women were more or less flat, as the early retirement trend was counteracted by the large influx of women into the labor force. Once the early retirement trend stopped, the labor force participation rates of older American women increased dramatically, and have approximately doubled since 1985 for women aged 65–70. Interestingly, the increases for women have flattened and even declined slightly in the past 5–10 years, at rates that are roughly 10% points below those of older men.

The reversal of the early retirement trend is consistent with important changes in retirement incentives noted above. The Social Security earnings test has been eliminated for those older than their FRA, and the delayed retirement credit (DRC) has increased to become actuarially fair for the average worker (Cahill, Giandrea, & Quinn, 2015b; Gruber & Orszag, 2003). The fact that Americans appear to be responding to these financial incentives, combined with the potentially large financial benefits that come with continued work later in life, is one reason to feel optimistic about our society’s ability to address the challenges of an aging society.

3.2 Older Americans Show Remarkable Flexibility with Respect to Their Work Decisions

The zero-one construct of labor force participation—one is either in or out of the labor force—masks a very important attribute of retirement transitions in America. For most older Americans, retirement is a process, not a one time, permanent event. The majority of older Americans retire gradually, in stages, as they transition from career jobs to complete labor force withdrawal. The term “gradual retirement ” encompasses three types of job transitions later in life: phased retirement, a reduction in hours with one’s current employer; bridge employment, a transition from career employment to a new employer prior to labor force exit; and reentry, a return to the labor force following an initial retirement (Cahill, Giandrea, & Quinn, 2006; Cahill et al., 2015b; Kantarci & Van Soest, 2008).

Data from the Health and Retirement Study (HRS), an ongoing longitudinal survey of older Americans that began in 1992, reveal that the most prevalent form of gradual retirement among those with career jobs is bridge employment. Between one half and two thirds of older Americans with career jobs transition to a bridge job prior to labor force exit. The next most common form of gradual retirement is reentry. Approximately 15% of older career workers who exit the labor force for at least two biennial HRS survey waves return to paid employment. Finally, approximately 10% of career workers experience phased retirement, reducing their hours with their career employer by 20% or more (Cahill, Giandrea, & Quinn, 2011, 2012; Cahill et al., 2015b; Quinn, 1999, 2010; Ruhm, 1990).

The relatively low prevalence of phased retirement seems counterintuitive at first, but barriers on the demand side may explain it. First, only a small minority of career workers have access to flexible hours arrangements. Approximately one quarter of career workers report being able to reduce the number of paid hours, and many fewer (10–12% of men and 14–17% of women) report being able to reduce their hours by one half or more (Cahill, Giandrea, & Quinn, 2014). The availability of job-sharing arrangements in which one or more employees split the responsibilities of a full-time position is even lower than that of reduced-hours arrangements (Hardy, 2008; Matos, Galinsky, & Bond, 2017). Employees could be reluctant to take up phased retirement if their subsequent pensions depend on their last few years of earnings. Employers, on the other hand, might restrict hours worked because of regulatory barriers that might prevent employees from claiming pension benefits and remaining with their employer (Hoffman & Andrew, 2010; Sheaks, Pitt-Catsouphes, & Smyer, 2010). Options for phased retirement policies are also limited by antidiscrimination rules with respect to age and income (Johnson, 2011; Penner, Perun, & Steuerle, 2002).

Still another complication with phased retirement pertains to scheduling. As an older worker reduces his or her hours from full-time to part-time, the employer might then need to find a worker to complete the now-unstaffed projects left by the worker who is taking phased retirement (Sloan Center on Aging & Work at Boston College, 2013a, 2013b). Further, closing the gap could involve more than just filling hours, as coordinating job tasks may involve changes for managers who need to oversee multiple workers for a series of tasks that were previously completed by one. One way to alleviate these strains is for employees to “compensate” their employers for being flexible, potentially in the form of reduced hourly wages. Aaronson and French (2004) find that reductions in the number of hours worked are associated with declines in hours wages; however, they also find that reductions in hourly wages are most pronounced among those who changed employers. The key takeaway is that, despite these barriers to phased retirement, older Americans still retire gradually by transitioning to bridge jobs and reentering the labor force.

Older Americans are also remarkably flexible with respect to the types of job changes they make. For example, a sizable minority of older Americans transition between wage-and-salary work and self-employment, and vice versa. An analysis of HRS respondents finds that between 12 and 17% of career wage-and-salary workers transition into self-employed bridge jobs and between 35 and 37% of career self-employed workers transition into wage-and-salary bridge jobs (Giandrea, Cahill, & Quinn, 2013). Although the percentage of career self-employed workers transitioning into wage-and-salary bridge jobs is higher than the reverse, the absolute number of older Americans who transition into self-employed bridge jobs is higher than the number who transition into wage-and-salary bridge jobs because there are many more wage-and-salary career workers than self-employed career workers.

Further, older self-employed workers stay in the labor force longer on average than older wage-and-salary workers, causing the prevalence of self-employment among those working to increase steadily with age. For example, the prevalence of self-employment among full-time career males in the HRS at the time of their first interview was 20%. Over the next 20 years, the prevalence of self-employment among those who remained working doubled to more than 40%. The prevalence of self-employment among older career women who remained working also doubled over the same time frame, from 10% at the time of the first interview to 20% 20 years later (Cahill & Quinn, 2014).

Among those older Americans who change jobs later in life, a sizable fraction also change occupations or “re-career.” Studies of occupational changes have found that between 30 and 40% of older career workers who change jobs also change occupations (Cahill, Giandrea, & Quinn, forthcoming; Johnson, Kawachi, & Lewis, 2009). The prevalence of re-careering is similar among career men and women. Further, these changes take place between white collar and blue collar workers, and a high prevalence of switching remains when two-digit occupational codes are used instead of the more refined three-digit codes. Moreover, when combining occupational changes with changes in the number of hours worked in bridge employment, nearly 80% of older Americans who transition from career to bridge employment either change occupations, reduce hours from full-time to part-time work, or both. Clearly, older Americans exhibit a great deal of flexibility when it comes to continued work later in life.

The degree to which this flexibility is critical to the retirement income security of older Americans has been the subject of some debate in recent years. Munnell and colleagues at the Center for Retirement Research at Boston College have constructed a National Retirement Risk Index (NRRI) using data from the Federal Reserve’s Survey of Consumer Finances (SCF) (Munnell, Hou, & Sanzenbacher, 2018). The SCF is conducted every 3 years and was conducted most recently in 2016. Using the SCF data, the NRRI compares working-age households’ pre-retirement income and expected retirement income to assess the degree to which Americans will be able to maintain their pre-retirement standard of living in retirement (Munnell et al., 2018). Specifically, the measure quantifies the percentage of households whose projected retirement income falls more than 10% below the estimated amount needed in retirement. The NRRI has increased from 31 to 50% over the past 30 years, with a high of 53% in 2010 (Fig. 7.6). A sizable percentage of Americans—roughly half—will not have enough income in retirement to maintain their pre-retirement standard of living.

Fig. 7.6
figure 6

National Retirement Risk Index , 1983–2016

Full size image

Other researchers come to a different conclusion. Biggs, for example, argues that there is no retirement “crisis” (Biggs, 2017). He points to the fact that, in the past, just as today, most private sector workers did not have a pension. The potential negative implications of the shift away from defined-benefit plans, therefore, apply to about one half of private sector workers only. He further argues that the advent of 401(k)s has actually increased the level of savings among workers generally. Biggs also points out that the American Community Survey (ACS) and the Current Population Survey (CPS)—two datasets commonly used by researchers to quantify retirement income—understate retirement income. One reason is that 401(k) balances and IRAs distributions are not paid out regularly and, therefore, are not counted as retirement income.

Schieber (2015) discusses three sets of estimates of retirement unpreparedness by birth cohort, by researchers using different datasets, different income definitions and annuitization assumptions, and different criteria for being at risk. Munnell et al. (2018) are the most pessimistic because, according to Schieber (2015), they believe “that current workers will have defined-contribution balances at retirement that are no larger than that of current retirees” (p. 15). One researcher presents two sets of estimates, with and without expected long-term care and home health care expenditures, with much higher risks when these are considered. Schieber (2015) concludes that “it is clear that many low earners face retirement with inadequate resources to provide an income that will allow them to maintain either a socially acceptable standard of living or one that matches that achieved while they were working” (p. 16).

The disagreement among researchers with respect to the adequacy of retirement income is but one outstanding issue that is unresolved in the retirement literature. Below we discuss other key outstanding questions.

4 Key Outstanding Questions

4.1 To What Extent Does Prolonged Labor Force Participation Translate into More Work Hours over the Retirement Transition Period?

A priori, the impact of extending working lives on total hours worked over the retirement transition period is ambiguous. Policies that promote continued work later in life might simply alter how older individuals allocate their work hours rather than increase the total number of hours worked. Employment decisions, retirement timing, and total hours worked are all jointly determined. This is a potentially fruitful area of research and some preliminary work on this topic suggests that voluntary job changes later in life increase the likelihood of remaining in the labor force to age 65 (Sanzenbacher, Sass, & Gillis, 2017). This analysis could further examine the impact of extending working lives on particular subgroups, especially vulnerable populations. Earlier in this chapter, for example, we commented on how poverty rates among older Americans as a whole have declined dramatically over the past 50 years, yet poverty rates remain elevated among older ethnic minorities and for unmarried men and women. Similarly, it would be worthwhile to examine how the overall impacts of extending working lives differ by population subgroups.

4.2 To What Extent Does Bridge Employment Per Se Increase Labor Supply over the Retirement Transition Period?

For many, bridge employment is a way to extend labor force participation later in life; for example, for those who move from a physically demanding career job they can no longer handle to a less strenuous one. But others may leave a career job that they could have continued. Since about one half of bridge jobs are part-time, it may be the case that the total number of hours worked over the retirement transition period would have been higher with continued career employment than with bridge employment. Studies of the impact of bridge employment on total hours worked would be worthwhile.

4.3 How Do Bridge Jobs Compare with the Career Jobs That Older Workers Leave Behind?

Bridge employment has the potential to offer older workers a variety of advantages, including flexible work arrangements and new career opportunities (rewirement!). Indeed, job transitions later in life are mostly voluntary (Cahill et al., 2015b; Maestas, 2010), suggesting that older workers expect to be better off after changing jobs. A deeper understanding of the trade-offs that older workers face when changing jobs later in life could improve policymaking around the impacts of pro-work incentives. For example, how does bridge employment compare with career employment in terms of wages, fringe benefits, job security, and other considerations? Further, how do these compare for those leaving career employment voluntarily versus involuntarily?

4.4 How Can Society Address Some of the Challenges Associated with Hiring and Retaining Older Workers (e.g., Age Discrimination and the Cost of Fringe Benefits)?

To this point, we have focused on the labor supply decisions of older workers under the implicit assumption that jobs will be available to those who want them. This is a reasonable assumption when economic conditions are favorable, such as now when the unemployment rate among workers aged 55 or older is about 3% (US Bureau of Labor Statistics, 2018). The sudden onset of the Great Recession, spanning December 2007 to June 2009, however, with large increases in the extent and duration of unemployment, reminds us that the labor demand side can be critical. Rather than wait until the next downturn, we should address the persistent challenges in hiring older workers, including age discrimination and extra costs that might be associated with older employees. Doing so might help avoid or reduce a spike in long-term unemployment similar to the one recently observed (Rix, 2014).

4.5 To What Extent Are Low Interest Rates Masking the Financial Instability of Older Americans with Substantial Debt and Suppressing the Urgency of Continued Work Later in Life?

The baby boomers are entering their retirement years with substantial amounts of debt—more than one quarter have credit card debt and nearly one third have housing debt—and a sizable minority (11%) of the middle boomers are entering retirement with negative net assets (Munnell, 2015; The Center for Retirement Research at Boston College, 2017). The historically low-interest rates for much of the past decade might be masking the extent to which these individuals are vulnerable, as low-interest rates reduce the cost of financing debt. When interest rates return to more normal levels, the cost of borrowing will increase, and this increased cost will exacerbate the impacts of being in debt. Research now on the prevalence of this hidden vulnerability could be very helpful in order to better understand just how problematic this issue could be in the years ahead.

4.6 For Savers, Is the Low-Return Environment Altering Appetites for Risk in Order to Achieve Higher Returns?

In contrast to those in debt, many with savings experience low returns on their assets in the current low-interest rate environment. One outstanding question is how these savers have reacted. On the one hand, one might expect little change, if they see low levels of inflation in tandem with the low returns and therefore see little reason to change behavior. In contrast, some might seek higher rates of returns by taking on more risk. To the extent that risk-taking has occurred, savers too might be vulnerable in a market downturn, just as those with equity-heavy portfolios were in 2008.

5 Conclusion

Older Americans are living longer and also working longer. While societal aging is a given, its implications will depend on how we respond to the challenges and the opportunities presented by these trends. The new world of retirement income security in America suggests many causes for concern. Social Security, the bedrock of financial stability for most older Americans, is now experiencing annual shortfalls, and the Social Security Trust Fund is projected to be depleted by the mid-2030s. The financial outlooks for Medicare and Medicaid are precarious. The shift to defined-contribution pension plans in the private sector for most of the approximately one half of workers covered by an employer-provided pension along with low personal savings suggest that for many older Americans these two legs of the traditional retirement income stool are unlikely to fill any void created by negative changes to Social Security, Medicare, and Medicaid.

On the other hand, many older Americans are responding to these important changes by remaining in the labor force later in life. A century-long trend toward earlier retirement ended in the mid-1980s and has since reversed. The change is even more notable because the labor force participation rates of older Americans have increased even as those among younger workers have declined. Older Americans have also shown a remarkable ability to adjust the ways in which they work by changing employers, re-careering, switching between wage-and-salary jobs and self-employment, and altering the number of hours they work. The flexibility of older workers is one of the true (and few!) bright spots in the financial outlook for older Americans. Policymakers would be wise to consider additional ways to support continued work later in life to help mitigate the many challenges that our aging society will face in the decades ahead.