a= CENTER for
RETIREMENT
RESEARCH

at BOSTON COLLEGE

ARE OLDER WORKERS CAPABLE OF WORKING LONGER?
Laura D. Quinby and Gal Wettstein

CRR WP 2021-8
June 2021

Center for Retirement Research at Boston College
Hovey House
140 Commonwealth Avenue
Chestnut Hill, MA 02467
Tel: 617-552-1762 Fax: 617-552-0191
https://crr.bc.edu

Laura D. Quinby and Gal Wettstein are senior research economists at the Center for Retirement
Research at Boston College. The research reported herein was pursuant to a grant from the U.S.
Social Security Administration (SSA) funded as part of the Retirement and Disability Research
Consortium. The findings and conclusions expressed are solely those of the authors and do not
represent the views of SSA, any agency of the federal government, or Boston College. Neither
the United States Government nor any agency thereof, nor any of their employees, makes any
watranty, express or implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of the contents of this report. Reference herein to any specific
commercial product, process or service by trade name, trademark, manufacturer, or otherwise
does not necessarily constitute or imply endorsement, recommendation or favoring by the United
States Government or any agency thereof.

© 2021, Laura D. Quinby and Gal Wettstein. All rights reserved. Short sections of text, not to
exceed two paragraphs, may be quoted without explicit permission provided that full credit,
including © notice, is given to the source.
About the Center for Retirement Research

The Center for Retirement Research at Boston College, part of a consortium that includes
parallel centers at the National Bureau of Economic Research, the University of Michigan, and
the University of Wisconsin-Madison, was established in 1998 through a grant from the U.S.
Social Security Administration. The Center's mission is to produce first-class research and forge
a strong link between the academic community and decision-makers in the public and private
sectors around an issue of critical importance to the nation's future. To achieve this mission, the
Center conducts a wide variety of research projects, transmits new findings to a broad audience,
trains new scholars, and broadens access to valuable data sources.

Center for Retirement Research at Boston College
Hovey House
140 Commonwealth Ave
Chestnut Hill, MA 02467
Tel: 617-552-1762 Fax: 617-552-0191
https://crr.bc.edu/

Affiliated Institutions:

The Brookings Institution
Mathematica - Center for Studying Disability Policy
Syracuse University
Urban Institute
Abstract

Disability-free life expectancy had been rising continuously in the United States until
2010, suggesting working longer as a solution for those financially unprepared for retirement.
However, recent developments suggest improvements in working life expectancy have stalled,
especially for minorities and those with less education. This paper uses data from the National
Vital Statistics System, the American Community Survey, and the National Health Interview
Survey to assess how recent trends in institutionalization, physical impediments to work, and
mortality have affected working life expectancy for men and women age 50, by race and

education.

The paper found that:

e The capacity to work to older ages is still increasing for high-education individuals and

low-education Black women.

e However, no progress has been observed for low-education whites of all genders and

Black men.

e As aresult, large shares of those still working at age 62 will be incapable of working

even two more years.

The policy implications of the findings are:
e Raising Social Security eligibility ages may reduce the financial security of large
segments of the population.

e These impacts will be particularly pronounced for Black men and low-education white

individuals of all genders.
Introduction

Working longer is a key to securing a comfortable retirement (Munnell and Sass 2008;
Bronschtein et al. 2019). However, health shocks are known to push older workers out of the
labor force before their intended retirement date (Coile and Levine 2007; Munnell,
Sanzenbacher, and Rutledge 2018). Until 2010, the trend of rising disability-free life expectancy
in the United States suggested increasing scope for longer working lives (Munnell, Soto, and
Golub-Sass 2008; Cutler 2009; Cutler, Ghosh, and Landrum 2014; Crimmins, Zhang, and Saito
2016; and Chernew et al. 2017), but recent developments may have stalled this progress.

This paper examines the period 2006-2018 to explore how long individuals can expect to
be able to keep working, and how these expectations vary across racial and socioeconomic status
(SES) groups. The analysis mirrors that of Munnell, Soto, and Golub-Sass (2008), combining
mortality data from the National Vital Statistics System (NVSS) with data on the total and
institutionalized population from the American Community Survey (ACS), and data on work-
limiting impairments from the National Health Interview Survey (NHIS).

These datasets are used to estimate cumulative probabilities of mortality,
institutionalization, and disability at each age after 50 for different demographic and education
groups.! The resulting profiles of working life by age are then used to calculate working life
expectancy at 50 for each group. These expected additional years of working ability are
estimated for each year between 2006 and 2018.

The resulting trends suggest cause for concern. While overall life expectancy at age 50
displays moderate improvement for the general population and for every demographic group
since 2006, the same is not true for working life expectancy. Working life expectancy at 50 has
increased slightly for high-education groups - Black and white, men and women. However low-
education groups have, with the exception of Black women, experienced stagnation in their
working life expectancy. This pattern suggests that calls for older workers to delay retirement,
which have proven successful over the past couple of decades, may be less fruitful going

forward.

 

1 Following Dowd and Hamoudi (2014), absolute levels of education are recognized as capturing increasingly more
selection by unobserved characteristics. Thus, similar to Bound et al. (2015), Sanzenbacher et al. (2019) and
Wettstein et al. (2021), education in this analysis is defined in relative, rather than absolute, terms.
The rest of the paper proceeds as follows. Section 2 presents the current state of the
literature. Section 3 discusses the data and methodology for the analysis. Section 4 estimates the
components of working life expectancy - mortality, institutionalization, and work-limiting
disability - and working life expectancy at age 50 for the full population segmented by gender.
Section 5 shows the same estimates broken out by the various racial and SES groups. The final
section concludes that working life expectancy has improved since 2006 primarily among the
more highly educated, while lower-educated individuals have experienced stagnation in working

life expectancy.

Background

The main question in this paper is how long people will be able to work and how does
that vary by education, race, and gender. Prior work on disability-free life expectancy - how
long individuals can expect to live without a disability - does not answer this question. Most
studies of health trends define poor health as an inability to perform activities or instrumental
activities of daily living (ADLs and IADLs). But ADLs and IADLs are an extreme form of
functional limitation; ailments that are both more prevalent and less severe are also likely to
prevent people from working (Burkhauser et al. 2002).

Coile and Duggan (2019) conduct one of the few analyses of work-limiting disability.
They find that, among men ages 45-54 without a college degree, 1.7 percent reported an ADL in
2015, compared to 16.0 percent reporting a work-limiting health condition; the comparable
shares for those with a college degree were 0.9 and 7.5 percent. But since the authors limit their
sample to prime-age men, the prevalence of work-limiting health conditions among near-retirees
of both genders remains unexamined. Analysis by race and education is similarly unexplored.

The period covered by the current analysis includes the troubling acceleration of deaths
of despair among middle-age workers (Case and Deaton 2015), growing inequality, the Great
Recession, and the subsequent recovery. Following these upheavals, the analysis explores both
whether past trends in working life expectancy have changed and whether these trends vary by
SES and race. Since less-educated whites and non-whites tend to approach retirement with
fewer resources, declines in work capacity could prevent delayed retirement among the very
individuals who would most benefit from it (Hou and Sanzenbacher 2020; Biggs, Chen, and
Munnell 2021; Munnell, Wettstein, and Hou 2021 forthcoming).
Recent trends suggest cause for concern. Health status - measured by self-reported
health, activities and instrumental activities of daily living, and obesity - has worsened over the
past two decades (Martin et al. 2010; Lezzoni, Kurtz, and Rao 2014). This decline has been
particularly acute for workers without a college degree (Cutler, Ghosh, and Landrum 2014; Coile
and Duggan 2019). At the same time, the separate trend of rising educational attainment, which
helped spur past improvements in disability-free life expectancy, has largely played out
(Munnell, Soto, and Golub-Sass 2008). And a long-term decline in nursing home use has mostly
reduced institutionalization over age 80, when work is largely irrelevant (Kaiser Family
Foundation 2021).

Adding to worries about the work ability of low-education individuals, particularly Black
individuals, are rising incarceration rates (U.S. Bureau of Justice Statistics 2021). Despite
incarceration rates falling for younger men over the period 2000-2016, they nevertheless
continued to increase among middle-age men, reflecting the increase in incarceration of younger
men in years past (Coile and Duggan 2019). Much like institutionalization in long-term care
facilities, incarceration is a nearly insurmountable impediment to work.

Compounding trends in health, trends in mortality among the working-age population are
also not encouraging. Although life expectancy has risen across the population over the past
several decades, the gains have mostly occurred at older ages when individuals are well past
retirement age (Chernew et al. 2017)." More troubling, recent studies have even found an
increase in prime-age mortality rates among less-educated whites (Case and Deaton 2020).

The analysis brings together these disparate trends in mortality, institutionalization, and
work-limiting disability. It explores the trends by gender, race, and SES, and combines the

different risks to work capacity into a summary measure of working life expectancy.

Data
"Working life expectancy" - the additional years of work ability an individual can expect
at a given age - depends on three components: 1) mortality risk; 2) the risk of

institutionalization; and 3) the risk of work-limiting disabilities in the non-institutionalized

 

? Increases in life expectancy have an ambiguous effect on retirement preparedness. On the one hand, reductions in
mortality in mid-life allow workers to work longer and accumulate more resources for retirement (both their own
and their spouses'). On the other hand, reductions in late-life mortality rates present households with longer
retirements, increasing the need for resources.
population.? Each of these probabilities is estimated based on some combination of the NVSS,
the ACS, and the NHIS, for the years 2006-2018. The NVSS gives the number of deaths in each
demographic group defined by gender, race, and education. The ACS gives the total population
in each such group as well as the institutionalized population. The NHIS is used to estimate the
share of each demographic group in the community that suffers from a work-limiting disability.

These components are estimated for each five-year age bracket after age 50 and for each SES

group.*

Defining the Demographic Groups

To estimate working life expectancy for different SES groups, it is first necessary to
define the groups. Race and gender are relatively straightforward and the definitions follow
those of the Census Bureau; the analysis focuses on non-Hispanic white and Black men and
women. However, defining relative education groups involves some discretion.

The assignment of individuals to their appropriate educational group is as follows. First,
the ACS is used to determine the median level of education for each gender-race-cohort group.°
Next, individuals in each of the datasets are assigned to be either above or below the median for
their demographic group. To allocate marginal absolute levels of education to above/below the

median, individuals in the marginal groups are randomized in the appropriate proportions.°

Mortality Estimates
Age-specific mortality rates, g, for each demographic group defined by gender, race, and
education are calculated year by year with the following formula:

d .
qxi,j = (1)

 

3 See Crimmins, Saito, and Ingegneri (1989 and 1997) and Munnell, Soto and Golub-Sass (2008).

4 Mortality is estimated at each year of age. The other two measures are estimated by 5-year age bins to increase
sample size.

5 Note that education quantile is, therefore, race-cohort-gender specific. This approach recognizes the fact that
attaining an equivalent absolute level of education (e.g., a college degree) implies greater selection for Black
individuals than for whites, given the overall lower educational attainment among Black individuals. For further
discussion see Leive and Ruhm (2021), who take the complementary approach of assigning quantiles across race. In
their setting, therefore, Black individuals are overrepresented in lower-education groups, and whites are
overrepresented in higher-education groups. Our approach ensures that each race has equal shares in the education
quantiles.

6 This approach is similar to that of Meara, Richards, and Cutler (2008), Bound et al. (2015), and Leive and Ruhm
(2021).
where x is age, i represents each demographic group, j represents year, l, ; is the number of
individuals alive in group i at age x using ACS data, and d,41; is the number of individuals in
group i who die between ages x and x + 1 using the NVSS data. To correct for small cell size,
the analysis adjusts the age-specific mortality rates using the Gompertz-Makeham formula (see
Brown et al. 2002 and Sanzenbacher and Ramos-Mercado 2016), estimated with non-linear least

squares.

Institutionalization Estimates

The share of each gender-race-education group living in institutions in each year is
estimated from the ACS. Institutions include both long-term services and supports (LTSS)
facilities (such as nursing homes) and carceral institutions such as prisons. Incarceration has
played a large and growing role in low labor force participation over the past few decades,

particularly among Black men (Coile and Duggan 2019).

Work-limiting Disability Estimates

The estimate of the risk of work-limiting disability relies on three questions in the NHIS:
1) Does a physical, mental, or emotional problem keep you from working? 2) Are you limited in
the kind/amount of work you can do because of your health? and 3) Are you limited in any way
because of physical, mental, or emotional problems? Following Crimmins, Saito, and Ingegneri
(1997), individuals are considered disabled if they respond yes to any of these questions.'

The final stage of the analysis estimates the probability of being able to work at a given
age by multiplying the probabilities of being alive, non-institutionalized, and without a work-
limiting disability. Working life expectancy, conditional on current age, is the sum of this

product over all future years.

 

7 The results look qualitatively similar when disability is defined more restrictively, as responding "yes" only to one
or both of the first two questions. See Appendix for results using this restrictive definition.

8 This calculation assumes that mortality, disability, and institutionalization are independent events. This approach
is common in the literature as the best approximation of the intersection of these events (see, for example,
Crimmins, Saito, and Ingegneri 1997).
Estimates of Key Metrics for the Full Population

The methods outlined above produce three sets of probabilities - dying, entering an
institution, and developing a work-limiting disability - for each demographic group, and for the
full population. These estimates are calculated for every year between 2006 and 2018, producing

time trends in the three risks.

Life Expectancy Trends

Tables 1a and 1b show the evolution of life expectancy at age 50 between 2006 and 2018,
for men and women. At this age, life expectancy has steadily increased for both genders. From
2006 to 2018, period life expectancy increased by just over one year for men and by almost 1.2
years for women. Cohort life expectancies, which are generally higher since they reflect the
predicted improvement of mortality in future years, also show steady increases. Even though
period life expectancy overstates likely mortality, estimates of future improvement in mortality
by demographic group are not available. Hence, the rest of the analysis will proceed using

period mortality estimates.

Institutionalization Trends

Tables 2a and 2b show trends in the share of the population that is institutionalized in
each five-year age bin from age 50 and up, for men and women respectively. The data display a
number of patterns pertinent to working life expectancy.

First, institutionalization rates generally increase with age. This pattern is not surprising
as institutionalization in mid- to late-life typically reflects residence in LTSS facilities, and LTSS
needs rise sharply with age." More unusual is the departure from this pattern for men, whose
institutionalization rates decline in their 50s only to begin rising in their 60s. A substantial
portion of men (but not women) in their 50s are incarcerated, with this share declining with age.
Because so few individuals at these ages have LTSS needs, the age-related decline in
incarceration swamps the small increases in LTSS facility use. However, by age 60, a growing
need for LTSS results in the expected increasing slope of institutionalization with age going

forward.

 

9 U.S. Department of Health and Human Services (2018).
The other noteworthy patterns in the data concern time trends. Institutionalization at each
age declines over time, particularly at older ages, reflecting a long-term reduction in nursing
home use.!° Again, men in their 50s and 60s stand out. While institutionalization is stable or
declining for women at these ages (and sharply declining at older ages for both genders), the
percentage of men in their 50s and 60s who are institutionalized has increased since 2006. This
pattern likely reflects the increasing prevalence of long prison sentences in the last few decades,
which are imposed primarily on younger men who then reach their 50s and 60s in correctional

institutions. !!

Work-Limiting Disability Trends

Tables 3a and 3b show the share of the non-institutionalized population with a work-
limiting disability, for men and women respectively. Unsurprisingly, disability rates increase
with age. More importantly, disability rates within a given age group do not seem to have
systematically changed over time. In other words, holding age constant, work-limiting disability
rates have held steady; thus the growth in total life expectancy implies more expected years of
disability now than fifteen years ago, as additional years are added at older ages where disability

rates are high. Working life expectancy will, indeed, be shown to reflect this fact.

Working Life Expectancy at Age 50

Tables 4a and 4b show how many additional years a 50-year old man (woman) can
expect to live, in total and in each of the states relevant to working life expectancy: not
institutionalized and free of disability (1.e., capable of work), not institutionalized with a
disability, and institutionalized. Furthermore, the tables also show how expected years in each of
these states have changed since 2000 and since 2006, decomposing the change in total life
expectancy into changes in the three relevant states of life.

In 2018, a 50-year-old man could expect to live an additional 29.8 years, and in 21.8 of

those years he would be expected to be capable of work. For a woman, the corresponding

 

10 This trend predates COVID-19 but has accelerated during the pandemic.

"! Prison sentences are not only an impediment to work while they last, but also lead to a permanent loss of earning
capacity even for those who do reenter the labor market (see. for example, Agan and Starr 2018). While the latter is
beyond the scope of this analysis, the estimated impact of institutionalization on the ability of individuals,
particularly men, to provide for themselves and their families must therefore be seen as a lower bound of the total
effect.
numbers are 33.6 and 23.9. The remaining years are expected to be life in the community with
some work-limiting disability, with only about half a year in an institution. '?

How have these expectations changed since 2006? Men (women) have gained 1.2 (1.0)
years of total life expectancy over the time period. These years can be decomposed into 0.5 (0.6)
years of work-capable life, 0.7 (0.7) years of life in the community with a disability, and slight
declines in years of life in an institution (primarily among women, who have seen a decline of
0.23 years of expected life in this state). Thus, every year of added life is approximately evenly
divided between work-capable and work-incapable. '°

This pattern is very different from trends in the preceding period, from 2000 to 2006. In
the earlier period, total life expectancy rose more in six years than it did in the 12 years between
2006 and 2018. The difference in the evolution of disability-free life expectancy is even more
stark: between 2000 and 2006, expected years with a disability rose only by about one quarter of
a year compared with 1.5 years of total life, unlike the latter period where over half of additional
years of expected life were years with a disability.

Nevertheless, on their face, the absolute numbers in Tables 4a and 4b may seem
encouraging. While only about half of the additional year of life gained since 2006 is time that
can be used for work, even this slow progress still means that the average person can work until
their early 70s. However, the average does not tell the full story: certain groups have made more
progress than others, and even within groups a substantial share of individuals may not be able to
work as long as the average group member. The next section explores heterogeneity in working

life expectancy to get at these disparities.

Heterogeneity in Working Life Expectancy
Tables 5a and 5b show the percentage of the non-institutionalized population with a

work-limiting disability for each demographic group at ages 50-64. Disability rates have

 

12 This estimate is slightly lower than the estimate in Hurd, Michaud, and Rohwedder (2017) of an average of 272
nights, or 0.75 years, in a nursing home for individuals ages 57-61. The difference is likely accounted for by the
older ages of individuals in that study versus the current one and the earlier time period of that previous study, which
included years 1992-2010. Indeed, looking at women in 2006 in Table 4b yields an estimate very close to the
previous study's.

13 As expected, using an alternative definition of work-limiting disability that omits the question "are you limited in
any way because of physical, mental, or emotional problems?" results in more of the additional years of life being
healthy and fewer being years with a disability, although the differences are small. The qualitative patterns are
robust to using this alternative definition. See Appendix Tables Ala and A1b.
increased for every gender, race, and education group over the sample period.'* Disability rates
will be joined with mortality estimates to form working life expectancy.

The analysis now turns to estimating total life expectancy for each demographic group.
The top panels of Tables 7a and 7b show, for men and women at age 50, the number of years an
individual of each race-education group can expect to live.!> All of the groups experienced gains
in life expectancy between 2006 and 2018. These gains ranged from small (low-education white
women and men gained only about half a year) to large (Black men with high education and
Black women with low education had gains of over two years). !°

The modest gains by low-education whites are emblematic of the opioid epidemic, and
"deaths of despair" more generally, which have been particularly devastating among low-SES
whites. In contrast, high-education whites experienced more robust gains in life expectancy over
the analysis period. The lack of systematic patterns by SES among Black individuals echoes
similarly disparate results in other recent work (Leive and Ruhm 2021 and Wettstein et al. 2021).

In contrast to total life expectancy, where all groups saw at least nominal improvement
over the past two decades, working life expectancy displays qualitative differences across
groups. The bottom panels of Tables 7a and 7b show the additional years of working life that
individuals of each demographic group can expect at each year between 2006 and 2018.
Examining the differences between 2006 and 2018 gives a sense of these disparities, with high-
education white women gaining 1.4 years of expected work capacity, while Black men with low
education had 1.4 fewer years of expected work capacity at the end of the period than at the

beginning.

 

'4 How can this finding be reconciled with the relatively stable rates of disability in Tables 3a and 3b? The
explanation rests on the fact that the 50-64-year-old group, as a whole, has grown older over time. Table 6 shows
the average age of this group, by gender. The population of those who might conceivably work longer has gotten
0.7 and 0.6 years older, for men and women respectively. This increase in average age, while generally small, has
likely led to increases in work-limiting disability in the older working-age population for all the demographic
groups.

15 The patterns of working life expectancy are generally similar when using the more restrictive definition of work-
limiting disability that does not consider those who answer the question "are you limited in any way because of
physical, mental, or emotional problems?" as having a work-limiting disability. As expected, working life
expectancy is higher under this definition, but trends over time and across demographic groups are qualitatively
similar. See Appendix Tables A2a and A2b.

16 These groups also had the lowest and highest average annual gains in life expectancy over the period. Low-
education white men and women gained 0.05 (p=0.021) and 0.02 (not significantly different from 0) years of
expected life per year. In contrast, Black men with above-median education gained 0.18 (p<0.001) years of
expected life per year, on average, between 2006 and 2018. The corresponding number for below-median education
Black women was 0.16 (p<0.001).
Just looking at the change between 2006 and 2018 is noisy, especially for Black groups
due to small sample sizes. Hence, Figure 1 shows the average annual change in working life
expectancy over the analysis period for each group.!' A stark divide is apparent: Both high-
education Black and whites individuals experienced an increase of about 0.1 years of working
life expectancy per year on average. A similar improvement was seen for Black women with
low education. However, the other low-education groups saw no annual improvement, on
average, in working life expectancy.

The lack of growth in working life expectancy for most low-education groups means that,
even as their total life expectancy has improved, their ability to work longer has not. In
particular, low-education Black men begin and end the period with roughly the same low
working life expectancy at age 50; the average member of this group will not be able to work
past age 63. This lack of progress is, in fact, a step back in terms of retirement security, since the

inability to work to a later age is now accompanied by a need to finance a longer retirement.

Simulating How Long Different Individuals Can Work

To more fully explore how long different types of people can work, the analysis uses the
estimated life, institutionalization, and work-limiting disability tables to track how ability to
work declines with age using the most recent data, from 2018. Figures 2a and 2b show the
estimated share of individuals in each group who will be unable to work to each age. The figure
is generated by simulating the experience of 100,000 individuals of each demographic group
who are capable of work at age 50. As these simulated individuals age, more and more of them

die, enter institutions, or develop work-limiting disabilities. !°

 

' This average is calculated using an OLS regression with a linear term in year, with each year-demographic group
considered as a single observation. This calculation does not account for the fact that each such working life
expectancy estimate is itself an estimate. With this caveat, the simple average improvement is highly statistically
significant for all the high-education groups, besides Black men for whom the change is marginally significant
(p<0.08). The positive slope for low-education Black women is also highly significant (p<0.001).

18 The simulation accounts for the fact that while death is an absorbing state, institutionalization and work-limiting
disability are not. The ability to leave institutions is particularly important, considering the non-monotone
relationship between age and disability among men, due to declining incarceration rates with age in their 50s and
60s. Thus, each individual in the simulation receives a single random draw between 0 and 1, which is compared to
his cumulative survival probability to determine age of death; and a random draw for each age which is
independently compared to the probability of being resident in the community without a work-limiting disability.
However, this calculation assumes independence within individual across years in this latter probability. This
independence assumption is surely incorrect but is conservative in the following sense: it will overestimate the share
of individuals who are capable of work, since each individual gets a new chance to leave the work-incapable state at
each age.

10
The patterns of decline in work capability are starkly different across demographic
groups, consistent with the estimates in Tables 7a and 7b. For example, of those capable of work
at age 50, fully 81 percent of high-education white men will still be capable of work at age 70,
the latest age for claiming Social Security. In contrast, only 19 percent of low-education Black
men will have that capability.

The analysis above begs the question: if individuals are expected to work longer, how
many of them will be unequal to the task? For example, the raising of the Social Security Full
Retirement Age from 65 to 67 implied no loss of monthly benefits - for those who could
postpone claiming by two years. What share of people in each demographic group would be
able to postpone retirement by two years from age 62, the Early Eligibility Age?

To answer this question, the analysis builds on the estimated probabilities of mortality,
institutionalization, and work-limiting disability at each age after age 50 in 2018. Using the
same simulation approach described to generate Figures 2a and 2b, the analysis takes an
individual at age 62 and calculates the probability that they will still be capable of work by age
64. Figure 3a shows the result of this exercise. Unsurprisingly, in the general population, only 7
percent of men and 4 percent of women would be forced to drop out of the labor force by death,
institutionalization, or disability by age 64. However, the picture is much less rosy when
considering those with low education and Black individuals. Among those with low education,
over 10 percent of each group would be unable to work even to age 64 (except for low-education
white women for whom the share is 7 percent). Similarly, among Black individuals, the different
gender and education groups have a greater than 10 percent chance of being unable to work until
64, except for high-education Black women (for whom the share is 6 percent).

The intersection of the two most disadvantaged groups is, unsurprisingly, the least likely
to be capable of working to age 64. Sixteen percent of low-education Black women capable of
work at age 62 will no longer be able to work by age 64. For low-education Black men, almost a
quarter of those capable of work at 62 will no longer be capable two years later.!? Thus
substantial shares of the Black population, particularly men, cannot be expected to work much

later than age 62.

 

19 Recall that even the average individual in the latter group cannot work to age 64; their working life expectancy is
63.4 at age 50. The 24 percent in Figure 3a is conditional on still being alive and capable of work at age 62.

11
Looking beyond age 64, outcomes continue to look grim for Black and low-education
groups - particularly low-education Black individuals. Figure 3b shows that around a quarter of
low-education white men who can work at 62 will not be capable of working to the Full
Retirement Age (FRA) of 67. A similar share of high-education Black men will be in the same
predicament. Strikingly, more than half of low-education Black men capable of work at age 62
will prove incapable of working to the FRA. When it comes to working until the maximal
claiming age of 70, Figure 3c shows that only high-education whites will not experience rates of
inability to work in excess of 20 percent. Among all other groups of both genders, of those who
can work at age 62, more than a quarter will not be able to work until age 70. For low-education

Black men, this share exceeds three quarters.

Conclusion

Both life expectancy and expected years of disability-free life had been trending up in the
United States for decades until 2010. The resulting need to fund a longer retirement was met by
calls to work longer, and the expanding capacity to work longer justified those calls. However,
in the last fifteen years, slowing declines in mortality have coincided with negative health trends,
raising the possibility of even slower growth in working life expectancy. A crucial question,
then, is whether working longer is even possible for many people?

To answer that question, policymakers need to know whether individuals are physically
capable of working: are they alive, in the community, and not encumbered by work-limiting
disabilities? The analysis presented here shows that improvement in life expectancy has
moderated since 2006, while improvement in working life expectancy has slowed even more,
such that every year of life expectancy gained is associated with only about half a year of work
capacity.

When looking across demographic groups, the picture is more concerning. The
population-level gain, however modest, is driven almost entirely by high-education groups
(although low-education Black women have seen similar growth). As a result, a large share of
those with less than median education will not be able to work even two years beyond the early
eligibility age for Social Security, even if they managed to work to 62. This problem is

particularly acute among low-education Black men, who had very low working life expectancies

12
in 2006 and experienced no improvement in the past fifteen years. A majority of this group will
be incapable of work to the FRA.

In thinking of solutions for inadequate retirement savings, working longer may be a fine
response for those with more education, but Black and low-education individuals, who are the
least likely to have sufficient savings, are also the least well-positioned to work longer."° They
would also be the groups most vulnerable to further increases in Social Security's eligibility age
thresholds. New solutions for these groups need to account for their high probability of not
being physically capable of extending their working lives.

 

20 Munnell, Hou, and Sanzenbacher (2018).

13
Tables and Figures

Table la. Period and Cohort Life Expectancy at Age 50 for Males, 2006-2018

 

 

Year Period LE Cohort LE
2006 28.59 30.81
2010 29.33 31.11
2014 29.66 31.40
2018 29.77 31.73

 

Table 1b. Period and Cohort Life Expectancy at Age 50 for Females, 2006-2018

 

 

Year Period LE Cohort LE
2006 32.52 34.55
2010 33.03 34.85
2014 33.38 35.13
2018 33.56 35.41

 

Source: Authors' calculations using American Community Survey (ACS) (2006-2018) and National Vital Statistics
System (NVSS) (2006-2018).

Table 2a. Percentage of Male Population Institutionalized

 

 

Age group 2006 2010 2014 2018
50-54 1.26% 1.55% 1.75% 1.76%
55-59 1.04 1.17 1.39 1.54
60-64 1.00 1.02 1.07 1.27
65-69 1.23 1.06 1.17 1.26
70-74 1.60 1.37 1.39 1.37
75-79 2.65 2.05 1.77 1.95
80-84 4.52 3.50 3.16 2.78

 

14
Table 2b. Percentage of Female Population Institutionalized

 

 

Age group 2006 2010 2014 2018
50-54 0.30% 0.28% 0.29% 0.33%
55-59 0.39 0.32 0.36 0.33
60-64 0.61 0.50 0.51 0.49
65-69 0.93 0.72 0.79 0.73
70-74 1.70 1.43 1.31 1.22
75-79 3.49 2.53 2.51 1.87
80-84 6.93 5.23 4.48 4.15

 

Source: Authors' calculations using ACS (2006-2018).

Table 3a. Percentage of Non-Institutionalized Male Population with Limitation of Activity

 

 

Age group 2006 2010 2014 2018
50-54 14.90% 13.50% 14.19% 14.49%
55-59 16.96 17.92 17.64 19.41
60-64 22.94 23.52 21.73 21.85
65-69 21.49 23.25 24.94 23.77
70-74 27.09 25.54 25.64 28.32
75-79 31.93 30.65 28.64 30.39
80-84 40.70 39.38 38.54 43.05

 

Table 3b. Percentage of Non-Institutionalized Female Population with Limitation of Activity

 

 

Age group 2006 2010 2014 2018
50-54 15.80% 16.00% 15.05% 15.24%
55-59 20.23 21.77 20.59 20.21
60-64 22.74 23.49 23.08 23.06
65-69 23.47 24.74 24.77 23.97
70-74 30.93 28.91 27.19 28.79
75-79 32.86 37.91 33.35 36.66
80-84 41.80 41.97 47.85 43.98

 

Source: Authors' calculations using National Health Interview Survey (NHIS) (2006-2018).

15
Table 4a. Expectations at Age 50 of Years Spent in Various States for Men

 

 

 

Change
Expectation of life 2000 2006 2018 2006-2018 2000-2006 2000-2018
Total 27.00 28.59 29.77 1.18 1.59 2.77
Free of disability 19.99 21.30 21.80 0.50 1.30 1.80
With disability 6.53 6.77 7.46 0.69 0.24 0.93
Institutionalized 0.47 0.52 0.51 -0.01 0.05 0.04

 

Table 4b. Expectations at Age 50 of Years Spent in Various States for Women

 

 

 

Change
Expectation of life 2000 2006 2018 2006-2018 2000-2006 2000-2018
Total 30.98 32.52 33.56 1.04 1.54 2.58
Free of disability 22.00 23.26 23.85 0.60 1.26 1.85
With disability 8.29 8.55 9.22 0.67 0.26 0.93
Institutionalized 0.69 0.72 0.49 -0.23 0.03 -0.20

 

Sources: Authors' calculations using NHIS (2000-2018); ACS (2000-2018); and NVSS (2000-2018).

Table 5a. Percentage of the Non-Institutionalized Male Population Ages 50-64 with Limitation of
Activity, by Race and Education Group

 

 

 

 

White Black
Year Below median Above median Below median Above median
2006 23.4% 12.6% 30.7% 19.3%
2010 23.5 13.0 31.1 22.0
2018 24.6 13.5 36.9 22.0

 

Table 5b. Percentage of the Non-Institutionalized Female Population Ages 50-64 with Limitation
of Activity, by Race and Education Group

 

 

 

White Black
Year Below median Above median Below median Above median
2006 24.6% 14.5% 30.1% 17.2%
2010 24.7 16.0 32.3 22.1
2018 24.8 15.7 33.9 19.4

 

Source: Authors' calculations using NHIS (2000-2018).

16
Table 6. Average Age and Race of Respondents Age 50-64 in Various NHIS Rounds

 

Average age

 

 

Year Men Women
2006 56.2 56.3
2010 56.4 56.5
2018 56.9 56.9

 

Source: Authors' calculations using NHIS (2006-2018).

17
Table 7a. Total Life Expectancy and Working Life Expectancy for Males at Age 50, by Education
and Race

 

 

 

White Black
Year Below median Above median Below median Above median
Total life expectancy
2006 26.60 30.91 23.26 26.32
2007 26.73 31.17 23.46 26.77
2008 26.57 31.33 23.38 27.10
2009 26.94 31.47 23.49 27.58
2010 27.31 31.50 24.14 27.65
2011 27.33 31.55 24.42 27.93
2012 27.36 31.82 24.36 28.27
2013 27.35 31.77 24.36 28.34
2014 27.38 31.89 24.76 28.29
2015 27.09 32.15 24.29 28.84
2016 27.24 32.07 24.31 28.46
2017 27.13 32.13 24.40 28.39
2018 27.15 32.34 24.06 28.62
2006-2018 change 0.55 1.43 0.80 2.30
Working life expectancy
2006 18.95 24.25 14.77 18.83
2007 18.44 24.98 13.37 18.46
2008 18.46 24.70 14.64 19.40
2009 18.37 24.87 13.22 20.04
2010 19.14 25.17 14.57 19.10
2011 18.38 24.99 13.08 19.95
2012 18.93 25.16 14.24 20.69
2013 19.15 25.48 15.35 20.56
2014 19.04 25.50 14.46 20.14
2015 18.73 25.78 14.69 21.32
2016 18.79 25.39 14.53 19.59
2017 18.80 25.11 14.48 19.78
2018 18.51 25.44 13.39 19.67
2006-2018 change -0.44 1.19 -1.38 0.84

 

18
Table 7b. Total Life Expectancy and Working Life Expectancy for Females at Age 50, by
Education and Race

 

 

 

White Black
Year Below median Above median Belowmedian Above median
Total life expectancy
2006 31.20 33.78 28.74 30.47
2007 31.44 34.03 29.34 30.57
2008 31.14 34.15 29.07 31.08
2009 31.57 34.37 29.15 31.55
2010 31.63 34.12 29.89 31.28
2011 31.66 34.25 30.35 31.17
2012 31.67 34.36 30.41 31.93
2013 31.60 34.43 30.35 31.57
2014 31.70 34.60 30.51 31.81
2015 31.31 34.71 30.61 31.73
2016 31.59 34.72 30.43 31.76
2017 31.47 34.70 30.54 31.93
2018 31.57 34.91 30.82 32.08
2006-2008 change 0.37 1.13 2.08 1.61
Working life expectancy
2006 21.28 25.58 16.34 21.56
2007 21.46 25.32 15.58 19.90
2008 20.14 25.68 15.63 20.72
2009 21.19 25.70 16.37 20.71
2010 21.34 25.72 16.69 20.63
2011 21.10 25.20 16.46 20.85
2012 21.86 26.08 17.08 21.38
2013 20.96 25.71 17.49 22.20
2014 21.40 26.01 16.50 20.96
2015 20.82 26.33 17.18 21.95
2016 21.25 26.26 17.66 22.22
2017 20.61 26.44 17.18 21.64
2018 21.03 26.94 16.66 22.24
2006-2008 change -0.25 1.36 0.32 0.68

 

Sources: Authors' calculations using NHIS (2000-2018); ACS (2000-2018); and NVSS (2000-2018).

19
Figure 1. Average Annual Change in Working Life Expectancy between 2006 and 2018 by
Demographic Group

 

 

 

 

 

 

 

 

      

 

 

 

 

0.15
0.13
& Men #& Women 0.12
0.10 0.10

0.10

0.05 -

0.00 Ly T T T

-0.01

-0.05

 

 

White -low White - high Black - low Black - high

Note: Solid bars are significantly different from 0 at the 10-percent level.
Sources: Authors' calculations using NHIS (2000-2018); ACS (2000-2018); and NVSS (2000-2018).

20
Figure 2a. Fraction of Men Capable of Work at Age 50 Who Can Still Work at Future Ages, by
Race and SES

 

 

 

 

 

 

 

 

 

   

 

 

  

1.00 -
Black - low
0.75 --- Black - high
' -- White - low
<= White - high
0.50
0.25
0.00 T T T T T T
50 60 710 80 90 100 110
Age

Figure 2b. Fraction of Women Capable of Work at Age 50 Who Can Still Work at Future Ages,
by Race and SES

 

 
 
 
 
  

 

 

 

 

 

 

 

 

 

 

1.00 ~
Black - low
0.75 oon Black - high
-- White - low
--= White - high
0.50
0.25
0.00
50 60 70 80 90 100 110

Note: Based on 100,000 simulations for each group.
Sources: Authors' calculations using NHIS (2000-2018); ACS (2000-2018); and NVSS (2000-2018).

21
Figure 3a. Percentage of Individuals Capable of Work at Age 62 Who Will Not be Capable by
Age 64, by Demographic Group

 

 

 

 

 

 

 

 

 

30%
@ Men ;

gw Women 24.2%

I

I

20%
I

I

I

I

I

I

I

I

10%
'

I

I

I

I

I

I

I

0% -

 

 

White - low White - high Black - low Black - high All

Figure 3b. Percentage of Individuals Capable of Work at Age 62 Who Will Not Be Capable by
Age 67, by Demographic Group

 

 

 

 

 

 

 

100%
G@ Men
g@ Women
75%
50.9%
50%

 

 

25%

 

 

0%

White - low White-high Black-low Black - high All

22
Figure 3c. Percentage of Individuals Capable of Work at Age 62 Who Will Not Be Capable by
Age 70, by Demographic Group

 

100%

 

@ Men

 

 

@ Women 76.3%

 

75%

 

50%

 

 

25% -

 

 

0% -

White - low White-high Black-low Black - high All

Note: Based on 100,000 simulations for each group.
Source: Authors' calculations.

23
References

Agan, Amanda and Sonja Starr. 2018. "Ban the Box, Criminal Records, and Racial
Discrimination: A Field Experiment." Quarterly Journal of Economics 133(1): 191-235.

Biggs, Andrew G., Angi Chen, and Alicia H. Munnell. 2021. "The Consequences of Current
Benefit Adjustments for Early and Delayed Claiming." Working Paper 2021-3. Chestnut
Hill, MA: Center for Retirement Research at Boston College.

Bound, John, Arline Geronimus, Javier Rodriguez, and Timothy Waidmann. 2015. "Measuring
Recent Apparent Declines in Longevity: The Role of Increasing Educational
Attainment." Health Affairs 34(12): 2167-2173.

Bronschtein, Gila, Jason Scott, John B. Shoven, and Sita Nataraj Slavov. 2019. "The Power of
Working Longer." Journal of Pension Economics and Finance 18(4): 623-644.

Brown, Jeffrey R. 2002. "Differential Mortality and the Value of Individual Account Retirement
Annuities." In The Distributional Aspects of Social Security and Social Security Reform,
edited by Martin Feldstein and Jeffrey B. Liebman, 401-446. Chicago, IL: University of
Chicago Press.

Burkhauser, Richard V., Mary C. Daly, Andrew J. Houtenville, and Nigar Nargis. 2002. "Self-
Reported Work Limitation Data: What They Can and Cannot Tell Us." Demography
39(3): 541-555.

Case, Anne and Angus Deaton. 2020. "Deaths of Despair and the Future of Capitalism."
Princeton, NJ: Princeton University Press.

Chernew, Michael, David M. Cutler, Kaushik Ghosh, and Mary Beth Landrum. 2017.
"Understanding the Improvement in Disability Free Life Expectancy in the U.S. Elderly
Population." In Insights in the Economics of Aging, edited by David Wise, 161-201.
Chicago, IL: University of Chicago Press.

Coile, Courtney C. and Mark G. Duggan. 2019. "When Labor's Lost: Health, Family Life,
Incarceration, and Education in a Time of Declining Economic Opportunity for Low-
Skilled Men." Journal of Economic Perspectives 33(2): 191-210.

Coile, Courtney C. and Phillip B. Levine. 2007. "Labor Market Shocks and Retirement: Do
Government Programs Matter?" Journal of Public Economics 91(10): 1902-1919.

Crimmins, Eileen M., Yasuhiko Saito, and Dominique Ingegneri. 1989. "Changes in Life

Expectancy and Disability-Free Life Expectancy in the United States." Population and
Development Review 15(2): 235-267.

24
Crimmins, Eileen M., Yasuhiko Saito, and Dominique Ingegneri. 1997. "Trends in Disability-
Free Life Expectancy in the United States, 1970-90." Population and Development
Review 23(3): 555-572.

Cutler, David M. 2009. "Estimating Work Capacity among Near Elderly and Elderly Men."
Working Paper RRC NB09-18. Cambridge, MA: National Bureau of Economic
Research.

Cutler, David M., Kaushik Ghosh, and Mary Beth Landrum. 2014. "Evidence for Significant
Compression of Morbidity in the Elderly U.S. Population." In Discoveries in the
Economics of Aging, edited by David A. Wise, 21-51. Chicago, IL: University of Chicago
Press.

Dowd, Jennifer B. and Amar Hamoudi. 2014. "Is Life Expectancy Really Falling among Groups
of Low Socio-Economic Status? Lagged Selection Bias and Artefactual Trends in
Mortality." International Journal of Epidemiology 43(4): 983-988.

Hou, Wenliang and Geoffrey T. Sanzenbacher. 2020. "Measuring Racial/Ethnic Retirement
Wealth Inequality." Working Paper 2020-2. Chestnut Hill, MA: Center for Retirement
Research at Boston College.

Hurd, Michael D., Pierre-Carl Michaud, and Susan Rohwedder. 2017. "Distribution of Lifetime
Nursing Home Use and Out-of-Pocket Spending." Proceedings of the National Academy
of Sciences 114(37): 9838-9842.

Kaiser Family Foundation. 2021. "Certified Nursing Facility Occupancy Rate." San Francisco,
CA. Available at: https://www.kff.org/other/state-indicator/nursing-facility-occupancy-
rates/

Leive, Adam A. and Christopher J. Ruhm. 2021. "Education Gradients in Mortality Trends by
Education and Race." Working Paper 28419. Cambridge, MA: National Bureau of
Economic Research.

Lezzoni, Lisa L., Stephen G. Kurtz, and Sowmya R. Rao. 2014. "Trends in U.S. Adult Chronic
Disability Rates Over Time." Disability Health Journal 7(4): 402-412.

Martin, Linda G., Vicki A. Freedman, Robert F. Schoeni, and Patricia M. Andreski. 2010.
"Trends in Disability and Related Chronic Conditions Among People Ages Fifty to Sixty-
Four." Health Affairs 29(4): 725-731.

Meara, Ellen R., Seth Richards, and David M. Cutler. 2008. "The Gap Gets Bigger: Changes in

Mortality and Life Expectancy, by Education, 1981-2000." Health Affairs 27(2): 350-
360.

25
Munnell, Alicia, H., Wenliang Hou, and Geoffrey T. Sanzenbacher. 2018. "Trends in Retirement
Security by Race/Ethnicity." Issue in Brief 18-21. Chestnut Hill, MA: Center for
Retirement Research at Boston College.

Munnell, Alicia, H., Mauricio Soto, and Alex Golub-Sass. 2008. "Will People Be Healthy
Enough to Work Longer?" Working Paper 2008-11. Chestnut Hill, MA: Center for
Retirement Research at Boston College.

Munnell, Alicia H. and Steven A. Sass. 2008. Working Longer: The Solution to the Retirement
Income Challenge. Washington, DC: Brookings Institution Press.

Munnell, Alicia H., Geoffrey T. Sanzenbacher, and Matthew S. Rutledge. 2018. "What Causes
Workers to Retire Before They Plan?" Journal of Retirement 6(2): 35-52.

Munnell, Alicia H., Gal Wettstein, and Wenliang Hou. 2021 (forthcoming). "How Best to
Annuitize Defined Contribution Assets?" Journal of Risk and Insurance.

Sanzenbacher, Geoffrey T. and Jorge Ramos-Mercado. 2016. "Calculating Expected Social
Security Benefits by Race, Education, and Claiming Age." Working Paper 2016-14.
Chestnut Hill, MA: Center for Retirement Research at Boston College.

Sanzenbacher, Geoffrey T., Anthony Webb, Candace M. Cosgrove, and Natalia Orlova. 2019.
"Rising Inequality by Life Expectancy by Socioeconomic Status." North American
Actuarial Journal 25(1): $566-S581.

U.S. Bureau of Justice Statistics. 2021. Key Statistic: Total Correctional Population.
Washington, DC. Available at: https://www.bjs.gov/index.cfm?ty=kfdetail&iid=487

U.S. Department of Health and Human Services. 2018. "An Overview of Long-Term Services
and Supports and Medicaid: Final Report." Washington, DC: U.S. Department of Health
and Human Services, Assistant Secretary for Planning and Evaluation, Office of
Disability, Aging and Long-Term Care Policy.

Wettstein, Gal, Alicia H. Munnell, Wenliang Hou, and Nilufer Gok. 2021. "The Value of

Annuities." Working Paper 2021-5. Chestnut Hill, MA: Center for Retirement Research
at Boston College.

26
Appendix

Table Ala. Expectations at Age 50 of Years Spent in Various States of Health for Males

 

 

 

Change
Expectation of life 2000 2006 2018 2006-2018 2000-2006 2000-2018
Total 27.00 28.59 29.77 1.18 1.59 2.77
Free of disability 21.55 22.93 23.67 0.73 1.39 2.12
With disability 4.98 5.14 5.59 0.45 0.15 0.61
Institutionalized 0.47 0.52 0.51 -0.01 0.05 0.04

 

Table Alb. Expectations at Age 50 of Years Spent in Various States of Health for Females

 

 

 

Change
Expectation of life 2000 2006 2018 2006-2018 2000-2006 2000-2018
Total 30.98 32.52 33.56 1.04 1.54 2.58
Free of disability 24.23 25.55 26.29 0.74 1.32 2.06
With disability 6.06 6.25 6.78 0.53 0.19 0.72
Institutionalized 0.69 0.72 0.49 -0.23 0.03 -0.20

 

Note: Disability is defined using the more restrictive definition, of answering affirmatively one of the following: 1)
Does a physical, mental, or emotional problem keep you from working? 2) Are you limited in the kind/amount of

work you can do because of your health?

Sources: Authors' calculations using NHIS (2000-2018); ACS (2000-2018); and NVSS (2000-2018).

27
Table A2a. Total Life Expectancy and Working Life Expectancy for Males at Age 50, by
Education and Race

 

 

 

White Black

Year Below median Above median Belowmedian Above median
Healthy life expectancy

2006 20.40 26.13 16.03 20.25
2007 19.92 26.67 14.67 20.68
2008 20.09 26.59 15.67 21.13
2009 19.60 26.65 14.36 22.37
2010 20.57 26.89 16.00 21.42
2011 20.10 26.76 14.70 21.53
2012 20.41 27.04 15.89 21.92
2013 20.57 27.08 16.89 21.87
2014 20.60 27.18 16.68 21.85
2015 20.11 27.53 16.23 23.03
2016 20.26 27.10 16.73 21.01
2017 20.34 27.19 16.13 22.15
2018 20.22 27.59 14.76 20.76

 

28
Table A2b. Total Life Expectancy and Working Life Expectancy for Females at Age 50, by
Education and Race

 

 

 

White Black

Year Below median Above median Below median Above median
Healthy life expectancy

2006 23.44 27.89 18.64 23.75
2007 23.66 27.87 17.38 22.72
2008 22.55 28.47 18.68 23.02
2009 23.52 28.14 18.32 22.99
2010 23.65 27.90 19.74 23.38
2011 23.66 27.71 18.99 23.48
2012 23.81 28.44 19.76 24.46
2013 23.44 28.27 19.92 24.54
2014 23.77 28.77 19.62 23.37
2015 23.04 28.82 19.88 24.97
2016 23.67 28.75 20.39 24.32
2017 23.06 29.02 20.01 23.87
2018 23.29 29.21 19.90 24.47

 

Note: Disability is defined using the more restrictive definition, of answering affirmatively one of the following: 1)
Does a physical, mental, or emotional problem keep you from working? 2) Are you limited in the kind/amount of
work you can do because of your health?

Sources: Authors' calculations using NHIS (2000-2018); ACS (2000-2018); and NVSS (2000-2018).

29
RECENT WORKING PAPERS FROM THE
CENTER FOR RETIREMENT RESEARCH AT BOSTON COLLEGE

 

 

Do Stronger Employment Discrimination Protections Decrease Reliance on Social Security
Disability Insurance? Evidence from The U.S. Social Security Reforms
Patrick Button, Mashfiqur R. Khan, and Mary Penn, June 2021

Trends in Opioid Use among Social Security Disability Insurance Applicants
April Yanyuan Wu, Denise Hoffman, and Paul O'Leary, March 2021

The Value of Annuities
Gal Wettstein, Alicia H. Munnell, Wenliang Hou, and Nilufer Gok, March 2021

Will Women Catch Up to Their Fertility Expectations?
Angi Chen and Nilufer Gok, February 2021

The Consequences of Current Benefit Adjustments for Early and Delayed Claiming
Andrew G. Biggs, Angi Chen, and Alicia H. Munnell, January 2021

Intended Bequests and Housing Equity in Older Age
Gary V. Engelhardt and Michael D. Eriksen, January 2021

The Effect of Early Claiming Benefit Reduction on Retirement Rates
Damir Cosic and C. Eugene Steuerle, January 2021

Financial Security at Older Ages
Barbara A. Butrica and Stipica Mudrazija, December 2020

Do People Work Longer When They Live Longer?
Damir Cosic, Aaron R. Williams, and C. Eugene Steuerle, December 2020

Do State and Local Government Employees Save Outside of Their Defined Benefit Plans
When They Need To?
Laura D. Quinby and Geoffrey T. Sanzenbacher, November 2020

How Much Taxes Will Retirees Owe on Their Retirement Income?
Angi Chen and Alicia H. Munnell, November 2020

A Behavioral Economics Assessment of SSDI Earnings Reporting Documents
Denise Hoffman, Jonah Deutsch, and Britta Seifert, November 2020

How Accurate Are Retirees' Assessments of Their Retirement Risk?
Wenliang Hou, July 2020

All working papers are available on the Center for Retirement Research website
(hitps://crr.bc.edu) and can be requested by e-mail (crr@bc.edu) or phone (617-552-1762).

30