The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
The Demographics
of Wealth
How Age, Education and Race Separate
Thrivers from Strugglers in Today’s Economy
Essay No. 3: Age, Birth Year and Wealth | July 2015
Authors
Ray Boshara is senior adviser and director of the Center for Household
Financial Stability at the Federal Reserve Bank of St. Louis. Before joining
the Fed, Boshara was vice president of the New America Foundation, a
think tank in Washington, D.C., where he started and directed programs
promoting financial well-being, college savings and a new social
contract. He has testified several times before the U.S. Senate and House
of Representatives. He has also worked for CFED, the United Nations in
Rome and the U.S. Congress. Boshara is the co-author of the book
The Next Progressive Era, published in 2009. Boshara has a bachelor’s
degree from Ohio State University and master’s degrees from Yale Divinity
School and the John F. Kennedy School of Government at Harvard.
William R. Emmons is senior economic adviser at the Center for Household Financial Stability. He is an assistant vice president and economist
at the Federal Reserve Bank of St. Louis, where his areas of focus include
household balance sheets and their relationship to the broader economy. He also speaks and writes frequently on banking, financial markets,
financial regulation, housing, the economy and other topics. Emmons
received a Ph.D. in finance from the J.L. Kellogg Graduate School of
Management at Northwestern University. He received his bachelor’s and
master’s degrees from the University of Illinois at Urbana-Champaign.
Bryan J. Noeth is a lead policy analyst for the Center for Household
Financial Stability. Noeth conducts primary and secondary research and
policy analysis on household balance sheet issues and helps to organize
conferences, roundtables and other efforts. Noeth received bachelor’s
and master’s degrees in economics from the University of Missouri and
a master’s degree in finance from Washington University in St. Louis.
An Introduction to the Series
The Demographics of Wealth
How Age, Education and Race Separate
Thrivers from Strugglers in Today’s Economy
By Ray Boshara, William R. Emmons and Bryan J. Noeth
A
describe eight of them as thriving financially.
the population who, generally, are accumulating
These groups include families headed by some-
wealth, and the strugglers, the other three-quarters
one who is typically middle-aged or older, white
who, generally, are not. As we show, race, educa-
or Asian, and with a college degree alone or with
tion and age increasingly determine whether
a graduate or professional degree. These families
someone is a thriver or a struggler.
generally earn above-average incomes, make
new economic reality is emerging in the U.S.
It’s between the thrivers, the one-quarter of
After considering each of the 48 groups, we
This is the third in a series of essays that the
or respond to good financial choices, and have
Center for Household Financial Stability at the
accumulated substantial wealth. These families
Federal Reserve Bank of St. Louis is publishing
constituted 24 percent of all U.S. families in 2013;
on how a family’s race or ethnicity, educational
they owned 67 of the economy’s wealth.
attainment, and age are related to its financial
The groups we describe as struggling finan-
choices and the financial outcomes it experi-
cially—the remaining 76 percent of all families—
ences. Our primary data source is the Federal
are typically younger, less educated, or black or
Reserve’s triennial Survey of Consumer Finances,
Hispanic. They earn average or below-average
which provides the most comprehensive picture
incomes, make or respond to less-conservative
of American families’ balance sheets and financial
financial choices, and have accumulated little or
behavior over time. We use information from over
no wealth; they own 33 percent of the nation’s
40,000 families, each of which was surveyed in
total wealth. Many, although not all, of these
one of nine waves between 1989 and 2013.
families are financially unstable.
By partitioning the sample in each wave into
Demography may not be destiny, but it is pow-
48 nonoverlapping groups based on four racial
erful in predicting family wealth. By documenting
or ethnic groups, four levels of educational
the links between race and ethnicity, educational
attainment, and three age ranges, we document
attainment, and age on the one hand, and financial
profound and persistent differences in financial
behaviors and financial outcomes on the other, we
decision-making, balance-sheet choices and
hope to inform policymakers, community prac-
wealth outcomes across groups. We show that
titioners, financial institutions and others in their
each demographic dimension is important in
efforts to improve the financial health of American
its own right.
families and the nation as a whole.
The Demographics of Wealth 3
E X E C U T I V E S U M M A RY O F E S S AY N O. 3
A
lthough there may be downsides to old age, those 62 and older can take heart in knowing that the
odds are in favor of their being wealthier than younger people. And the gap has widened considerably
over the past quarter-century—in favor of old people. That said, being old isn’t what it used to be. Baby
boomers, who are now retiring in droves, are likely to be less well-off than their “old” counterparts in the
two previous generations. And it looks as if members of the next two generations—Generation X and
Generation Y (the millennials)—might also end up less wealthy than the generation before them.
These are just some of the connections between age and wealth that were found in researching this
essay—the third—in our “Demographics of Wealth” series. (The first looked at the link between race/
ethnicity and wealth. The second examined the connection between education and wealth. Both can be
read at www.stlouisfed.org/hfs; accompanying videos can also be viewed there.) All of the essays are the
result of our analysis of data collected between 1989 and 2013 by the Federal Reserve for its Survey of
Consumer Finances. More than 40,000 families were interviewed by the Fed over those years.
For this essay, we looked at age in two ways: where a person stands in the life cycle (young, middle-aged or old) and how birth-year cohorts stack up against one another. This latter approach allows us
to make some comparisons of generations, from “the greatest generation” of WW II fame to the millennials of today.
Among our findings:
• The median wealth of old families (headed by someone at least 62) rose 40 percent between 1989
and 2013, from just under $150,000 to about $210,000. The median wealth of a middle-aged family
(40-61) in 2013 was 31 percent lower than in 1989, declining from $154,000 to about $106,000. The
median wealth of a young family dropped more than 28 percent, from $20,000 to just over $14,000.
(All figures are adjusted for inflation.)
• The explanations for this growing gap are difficult to pin down. The lack of education does not
appear to be to blame, given that each succeeding generation is better-educated than the previous
one. Younger families could be losing ground, in part, because they are more racially and ethnically
diverse than ever before—and we know that race- and ethnicity-based disadvantages continue to
loom large in our society.
• Baby boomers could be faring worse (not just in wealth, but income) because there are so many of
them. They’ve had to compete more for jobs, housing, investment opportunities, etc., than did the
less-numerous generation before them. That so-called silent generation (born 1925-1944) was relatively small because birth rates dropped during the Great Depression; the “scarcity” of people then
worked to their advantage during the post-WW II economic boom.
• Although young families are often depicted by other observers as “poor,” such labeling might be a
stretch. The average young family has always been pinching pennies, given that it is just starting to
make money and has a lot of major expenses (marriage, children, house, etc.). Young families that
want to increase their chances of being wealthy should emulate the financial decision-making of
older families: maintain an emergency fund, pay down debt, put extra money in high-return investments, such as stocks, etc. Delaying the purchase of a house can also help in multiple ways.
4 Federal Reserve Bank of St. Louis
Essay No. 3
Age, Birth Year and Wealth
By William R. Emmons and Bryan J. Noeth
T
he first two essays in this series described large
and persistent differences in financial behaviors
and financial outcomes across racial and ethnic
groups and across education groups.1 In the first
essay, we showed that non-Hispanic whites and
Asians are much more likely to be thriving financially than blacks and Hispanics of any race, who
were more likely to be struggling. In the second
essay, we showed that families with higher levels of
education generally fared better than those with less
education along a number of important dimensions.
This essay documents significant differences in
financial choices and financial outcomes across the
life cycle (that is, at different stages of life) and across
birth-year cohorts (that is, comparing different
groups of people who were born at about the same
time). Like race-, ethnicity- and education-related
disparities, the age- and birth year-related differences described here have existed at least since
1989, when our data begin. We show that gaps in
several financial behaviors and financial outcomes
related to age and year of birth have grown larger in
recent years.
The existence of a life-cycle effect in economic
and financial matters is uncontroversial. Yet many
important discussions—such as changes in the
overall income or wealth distributions—sometimes
ignore the life cycle with misleading implications.
Young families typically start out with little or no
wealth—hence are “poor” on a simplistic ranking of
all families—but some will accumulate wealth rapidly
as they grow into middle age. Thus, it is misleading
in many cases to include young families among the
ranks of the poor. At the other end of the life cycle,
a typical family’s wealth does not begin to decrease
until quite an advanced age. Many older adults,
therefore, appear “rich.” Yet some older families with
above-median wealth may not be particularly welloff if their current income is low or their expenses,
such as health care, are high.
A less well-known fact is that your year of birth
influences your wealth, too. Following the work of
others, we found that a family headed by someone
born about 1970, for example, was likely to have
about 40 percent less wealth (inflation-adjusted) at
any given age than an otherwise identical family
headed by someone born about 1940 when he or
she was the same age (holding constant the level of
education, race or ethnicity, health status, income,
etc.). This positive birth-year effect for people born
about 1940 reinforces the positive life-cycle effect
on wealth accumulation, making people currently
in their 70s the beneficiaries of two purely demographic influences on their wealth—their age and
the era in which they were born.
This essay begins by describing the two alternative
frameworks we used to analyze the connections
between when someone was born and their income
and wealth—a life-cycle approach and a birth-year
cohort approach. The second section contains brief
qualitative snapshots of the current income, wealth
and key financial behaviors of each of three broad
age groups—young, middle-aged and old families—and an overview of some important patterns
across different birth-year cohorts. The third section
provides detailed characterizations of family balance
sheets and financial behaviors across a typical life
cycle during the past quarter-century, based on the
The Demographics of Wealth 5
Survey of Consumer Finances (SCF).2 The final section isolates birth-year cohort effects on income and
wealth, using results from a regression framework;
the details are in a technical appendix, which might
be of interest primarily to researchers.
I. Stages in the Adult Life Cycle and Birth-Year
Cohorts
We used a family head’s year of birth in two
different ways by applying both a life-cycle framework and a cohort, or generational, perspective. The
life-cycle framework highlights effects that operate
on all or most families in a certain age range when-
ever they reach it in historical time (such as the
middle or late 20th century or the early 21st century).
For example, all young people face decisions about
education, employment, marriage, child-bearing,
homeownership, saving, investment and a host
of other things. Old people face some of the same
decisions but in the opposite direction—whether
or when to exit employment, to stop saving, draw
down investments, etc. The underlying assumption of the life-cycle framework is that the stage of
life itself is important in its own right regardless of a
family’s experiences before reaching that age or the
time period in which it is lived.
Sidebar 1: Classifying Individuals and Families by Age and Birth-Year Cohort
We assigned each family to a life-cycle group based on
the age of the family head (or single individual) at the time
of the survey.3 We used the term young when referring to
families whose head was under 40 years of age at the time
of the survey. We used the term middle-aged in referring
to families whose head was between 40 and 61 at the time
of the survey. Old families were headed by someone 62
years or older at the time of the survey. Each family was
assigned a birth year based on the birth year of the family
head, even if other members of the family had birth years
that would place them in a different life-cycle group.
We chose cut-offs at 40 and 62 years of age to
create three groups that each represented approximately
one-quarter to one-half of the sample in each survey
year. (See Table 1.) The groups were designed to capture
broad features of typical economic and financial life cycles,
including homeownership, employment and receipt of
retirement benefits. For example, all five-year age groups
under 40 (for example, 35-39) have had homeownership
rates below the national average, while all five-year age
groups 40 or older have had homeownership rates above
the national average since 1990.4 Thus, young families in
our classification system were more likely to be renters
than either of the other age groups.
According to the Bureau of Labor Statistics, the
employment-to-population ratio for each age group
increases until ages 35-44 and then declines.5 Thus,
families we classified as young were more likely to include
members who were not yet in their working careers or
were early in their working careers than were families in
the older groups, which were more likely to include mem6 Federal Reserve Bank of St. Louis
bers who were later in their working careers or had left
work altogether.
At the upper break point of 62, the employment-topopulation ratio is about equal to the overall population’s
employment-to-population ratio. In fact, the employment-to-population ratio for the 55-59 age group was 9.3
percentage points above the population average in 2014,
while the ratio for the 60-64 age group was 5.7 percentage points below the population average.6 Thus, the group
of families 62 or older is significantly less likely to include
working members than families in the middle-aged group.
In addition, the youngest age at which Social Security
benefits can be drawn is 62. Thus, families in the old group
were far more likely to be receiving a fixed retirement
income than were families headed by someone under 62.
We did not assign families to life-cycle groups based
on their actual homeownership, employment or retirement status because the specific timing of each of these
decisions is endogenous, that is, each family or individual
chooses if or when to make each transition (or may be
forced to or prevented from making a transition by individual circumstances). There may be factors other than age or
birth year that contribute systematically to these decisions,
factors such as race, ethnicity or educational attainment. In
this essay, we wanted to isolate the effects of age and birth
year on economic and financial choices and outcomes;
therefore, we imposed a fixed life-cycle age structure in
much of our discussion. We used a more-granular—but
still fixed—life-cycle and birth-year cohort structure in the
regression models described in Section 4.
Table 1. Share of Families in the Survey of Consumer Finances by Age Group
Percent
1989
1992
1995
1998
2001
2004
2007
2010
2013
Old: 62 or older
25.9
26.3
25.9
24.2
24.5
24.7
25.5
26.9
28.9
Middle-aged: 40-61
35.9
36.1
37.7
41.0
42.6
43.4
43.8
43.7
42.1
Young: Under 40
38.3
37.6
36.3
34.7
32.8
31.9
30.7
29.4
29.0
Unless otherwise noted, the source for all tables and figures is the Federal Reserve’s Survey of Consumer Finances.
Table 2. Median Wealth of Families by Age of Family Head
Median wealth
in 1989
Percent of families in upper
half of nation’s wealth
distribution in 1989
$85,575
50.0%
$81,456
50.0%
–4.8%
Old: 62 or older
$149,728
62.3%
$209,590
70.3%
40.0%
Middle-aged: 40-61
$153,759
63.8%
$106,094
54.0%
–31.0%
$19,830
28.7%
$14,220
23.9%
–28.3%
All families
Young: Under 40
Median wealth
in 2013
Percent of families in upper
half of nation’s wealth
distribution in 2013
Percent change in
median wealth between
1989 and 2013
All dollar amounts are expressed in 2013 dollars, deflated by the CPI-U-RS (Consumer Price Index for Urban Consumers, Research Series).
To capture the three main stages of the adult life
cycle, we assigned each family in each survey year
to one of the following groups:
• young: families headed by someone who is
younger than 40 at the time of the survey (representing 29 percent of all families in the 2013
Survey of Consumer Finances);
• middle-aged: those between 40 and 61 years old
(42 percent);
• old: those 62 years or older (29 percent). (See
Table 1.)
Cohort analysis is an alternative and complementary analytical framework. Cohort analysis
illuminates the possibility that certain groups of
families born at one point in time (or a range of
years) may experience a life-cycle stage differently than other groups born in different years
(for example, people born in the 1940s vs. people born in the 1960s). By following through time
various cohorts of families defined by their (that is,
the family head’s) year of birth, we may be able to
identify unique aspects of their life courses that are
not strictly life-cycle regularities. Instead, there may
be experiences common to birth-year cohorts or
larger generations (such as the baby boomers) that
systematically influence their choices and outcomes.
II. Financial Snapshots of Three Age Groups and
Five Generations
A simple way to describe typical financial differences across the life cycle and how they have
changed is to compare the respective income or
wealth distributions of young, middle-aged and old
families at different points in time. (See Sidebar 2.)
For example, the median wealth of old, middleaged and young families in 1989 was $149,728,
$153,759 and $19,830, respectively, expressed in
dollars adjusted to the purchasing power of 2013.
(See Table 2). Within the overall distribution of family
wealth in 1989, 62.3 percent of old families ranked
above the median, while 63.8 percent of middleaged families had above-median wealth—but only
28.7 percent of young families did.
By 2013, the wealth medians were $209,590,
$106,094 and $14,220 for the old, middle-aged and
young, respectively. Within the overall distribution
of family wealth in that year, 70.3 percent of old
families ranked above the median. Only 54.0 percent of middle-aged families had above-median
wealth, while just 23.9 percent of young families
were in the top half. Combined with the increasing
share of old families in the population (recall Table
1), the large increase in typical wealth among old
families (an increase of 40.0 percent between 1989
and 2013) means that wealthy families in 2013 were
The Demographics of Wealth 7
Sidebar 2: Family Wealth and Income
Wealth is a family’s net worth, consisting of the excess
of its assets over its debts at a point in time. Total assets
include both financial assets, such as bank accounts,
mutual funds and securities, as well as tangible assets,
including real estate, vehicles and durable goods. Total
debt includes home-secured borrowing (mortgages),
other secured borrowing (such as vehicle loans) and
unsecured debts (such as credit cards and student loans).
Debt incurred in association with a privately owned
business or to finance investment real estate is subtracted
from the asset’s value, rather than being included in the
family’s debt. All wealth figures in the essay are adjusted
for inflation to be comparable to values recorded in 2013.
To measure income for the SCF, the interviewers
requested information on the family’s cash income, before
taxes, for the full calendar year preceding the survey. The
components of income in the SCF are wages, self-employment and business income, taxable and tax-exempt
interest, dividends, realized capital gains, food stamps and
other related support programs provided by government,
pensions and withdrawals from retirement accounts,
Social Security, alimony and other support payments, and
miscellaneous sources of income for all members of the
primary economic unit in the household.7
much more likely to be old than they were in 1989.
Meanwhile, the median middle-aged family in 2013
had 31.0 percent less wealth than the median middle-aged family in 1989, while the median young
family had 28.3 percent less wealth in 2013 than its
counterpart did in 1989.
Another way to illustrate how different the financial lives of families are at different stages in their
life cycles is to calculate the odds of having at least
$1 million in net worth. In 1989, the odds of a family headed by someone under 40 having $1 million
of wealth or more were about 1 in 51. By 2013, the
odds had lengthened slightly to 1 in 55. Among
middle-aged families, 1 in 12 had at least $1 million
in 1989 and 1 in 9 had that much wealth in 2013.
Among old families, 1 in 11 was a millionaire in 1989
but 1 in 7 old families had at least $1 million in 2013.
In other words, typical wealth differences between
young and old families were large in 1989 and had
increased notably by 2013.
A snapshot of families headed by someone
under 40. The head of a randomly chosen family headed by someone under 40 was somewhat
more likely to be black and much more likely to be
Hispanic than his or her respective race’s share in
the overall population in 2013.8 A randomly chosen
young family was much less likely to be white than
the white share of the overall population of families.
A young family’s wealth was likely to be far below
the levels of middle-aged and old families, but the
young family’s income was similar to that of an old
family and less than that of a middle-aged family. A
typical young family’s cash reserves were likely to be
similar to those of middle-aged and old families as a
share of total assets, but because young families typically have few assets, this means the actual amount
of liquid assets to meet emergencies was small. The
young family’s financial or business assets, which
are important for diversification and long-term
wealth accumulation, probably were small both in
absolute value and relative to the family’s total assets.
A young family was more likely to have debt than
a family over 40, and the young family’s borrowing
probably was much greater than that of middle-aged
and old families relative to assets or income. Low
asset diversification and high debt made the typical
young family’s financial life risky. Moreover, a typical
young family scored poorly on our index of sound
day-to-day financial decision-making, including
saving and cash-management choices.9
A snapshot of families headed by someone at
least 40 but less than 62 years old. The racial and
ethnic composition of the group of families headed
by someone between 40 and 61 was almost identical to the overall population averages throughout
1989-2013.10 A typical middle-aged family’s wealth
was likely to be between the wealth of old and young
families. Its income, however, was likely to be significantly higher than those of old or young families.
A typical middle-aged family’s cash reserves, share
of financial and business assets in total assets, and
total debt relative to total assets were likely to be
8 Federal Reserve Bank of St. Louis
Figure 1. Difference between a Racial or Ethnic
Group’s Share of Older Families and the Group’s
Share of All Families
15
Percentage points
10
5
0
–5
–10
1989 1992 1995 1998 2001 2004 2007 2010 2013
White
Black
Hispanic
Asian
Families headed by a non-Hispanic white person 62 years or
older constituted 79.9 percent of all families headed by someone 62 or older in 2013. The share of all families headed by a
non-Hispanic white person of any age in 2013 was 70.1 percent.
The difference between these two shares—9.8 percentage
points—is shown in the top line in the figure. White families
have been over-represented among old families compared
with their share in the total population by a substantial amount
throughout the period for which we have SCF data.
All other points in the figure are derived analogously for each
racial or ethnic group in each year. All groups other than whites
have been under-represented among old families compared
with their shares of the total population of families since 1989.
Unless otherwise noted, the source for all tables and figures is the
Federal Reserve’s Survey of Consumer Finances.
close to the population medians and between those
of old and young families. Perhaps surprisingly,
a typical middle-aged family scored only slightly
better than young families on our index of sound
day-to-day financial decision-making and far below
the typical old family.
A snapshot of families headed by someone
62 years or older. The most striking aspect of the
racial and ethnic composition of families headed by
someone at least 62 years old is the over-representation of whites. (See Figure 1.) Whites represented 79.9
percent of old families in the 2013 SCF, while they
constituted only 70.1 percent of all families.11 A typical old family’s wealth was likely to be much higher
than the wealth of middle-aged and young families.
Its income, however, was likely to be slightly lower
than that of a young family and significantly less than
that of a middle-aged family. A typical old family’s
share of financial and business assets to total assets
was likely to be much higher and its total debt relative
to total assets much lower than those of either of the
other groups. A typical old family scored far above
young and middle-aged families on our index of
sound day-to-day financial decision-making.
A snapshot of families headed by someone
born before 1925. The generation born during the
first quarter of the 20th century has been called the
“greatest generation” or the “G.I. generation” because
it contributed most of the men and women who
fought in World War II.12 This group of Americans
was overwhelmingly white and had much less
education than later generations. For example, in
the 1989 SCF, more than 80 percent of families 62
or older were white (representing family heads born
in 1927 or earlier). Only 17 percent of these families
had two- or four-year college degrees. A further 36
percent of family heads had completed high school,
but 47 percent of family heads had not completed
high school. We cannot use the SCF to describe
this generation’s young or middle-aged income or
wealth because these families were in retirement
already when our data begin. Figure 2 shows the
median incomes of the 1915-17 cohort at three-year
intervals when these family heads were in their early
70s through their mid-90s. Figure 3 shows that, in
contrast to income, the median wealth levels of the
1915-17 cohort did not decline in retirement.13 (Once
again, all figures are adjusted for inflation.)
A snapshot of families headed by someone born
between 1925 and 1944. The family heads born
during the 20 years following the greatest generation sometimes are referred to as the “silent generation” because members reached adulthood after
World War II and were humbled by the accomplishments of the generation that preceded them. This
generation played an understated role in American
life also because its numbers were small; birth rates
plunged during the Great Depression. This generation was overwhelmingly white (about 76 percent
of middle-aged families observed in the 1989 SCF)
The Demographics of Wealth 9
Figure 2. Median Income of Three-Year Birth
Cohorts
80,000
Figure 3. Median Net Worth of Three-Year Birth
Cohorts
400,000
70,000
50,000
40,000
30,000
1915-17
1933-35
20,000
1954-56
1975-77
10,000
0
1915-17
1933-35
300,000
2013 dollars
2013 dollars
60,000
1954-56
1975-77
1993-95
200,000
100,000
1993-95
19
25 31
37 43 49 55 61 67 73 79 85 91
0
Average age at SCF survey date
19
25 31
37 43 49 55 61 67 73 79 85 91
Average age at SCF survey date
All dollar amounts are expressed in 2013 dollars, deflated by the CPI-U-RS (Consumer Price Index for Urban Consumers, Research Series).
To illustrate the life cycle of family incomes (left) and wealth (right) and differences across birth-year cohorts, we assigned each
family in each survey to a three-year birth cohort. The oldest selected cohort shown here includes all families whose head was born
during 1915-17. They were observed in the SCF at ages 72-74 in 1989, which is represented on the horizontal axis as average age 73;
ages 75-77 in 1992 (average age 76); and so on through ages 95-97 in 2013. Because there were so few observations for the cohort in
2013, we show this group only through 2010, when they were 93-95 (average age 94).
The 1954-56 cohort was observed nine times—at ages 33-35 in 1989 (average age 34); ages 36-38 in 1992 (average age 37); and so
on through ages 57-59 in 2013 (average age 58).
The 1975-77 cohort was observed seven times—at ages 18-20 in 1995 (average age 19); ages 21-23 in 1998 (average age 22); and so
on through ages 36-38 in 2013 (average age 37).
The 1993-95 cohort was observed once at ages 18-20 in 2013 (average age 19).
Unless otherwise noted, the source for all tables and figures is the Federal Reserve’s Survey of Consumer Finances.
but less so than the greatest generation.14 The silent
generation was much better-educated than the
greatest generation, as 32 percent of family heads
among the former achieved a two- or four-year
college degree, 49 percent received high school
diplomas and only 19 percent did not finish high
school. The silent generation probably benefited
economically and financially from the post-WW II
economic boom and from this generation’s own
small numbers. Figure 2 shows that the 193335 cohort enjoyed substantially higher median
incomes than the 1915-17 cohort at the same ages.
In fact, the 1933-35 cohort’s median incomes were
22.5, 12.6 and 11.3 percent higher than the 1915-17
cohort’s median incomes at ages 72-74, 75-77 and
10 Federal Reserve Bank of St. Louis
78-80, respectively. Figure 3 shows that the 193335 cohort’s median wealth levels surpassed those
of the 1915-17 cohort at the same ages by an even
larger margin: by 47.5, 9.5 and 50.9 percent,
respectively.
A snapshot of families headed by someone
born between 1945 and 1964. The members of the
generation born during the two decades after World
War II commonly are called baby boomers because
the birth rate skyrocketed during this period. The
baby boom generation was the most racially and
ethnically diverse generation to that time in American history, with about 73 percent of families
headed by someone who was a non-Hispanic
white.15 It also was the best-educated generation to
that time, with 42 percent of family heads receiving at least a two- or four-year college degree. A
further 48 percent of family heads had high school
diplomas, while only 11 percent had not finished
high school. Baby boomers also benefited from the
strong post-WW II economy. Figure 2 shows that the
1954-56 cohort enjoyed median incomes 4.7 and
9.3 percent higher than the 1933-35 cohort at ages
54-56 and 57-59, respectively. However, Figure 3
reveals that the median wealth of the 1954-56 cohort
fell short of the median wealth of the 1933-35 cohort
at ages 54-56 and 57-59, by 13.5 and 29.8 percent,
respectively. These two observations correspond to
the 2010 and 2013 SCF dates, after asset prices had
been reduced by the Great Recession.
A snapshot of families headed by someone
born between 1965 and 1984. People born during
the two decades after the baby boom are sometimes called members of Generation X because
their place in history was uncertain. Generation X
was even more racially and ethnically diverse than
the baby boomers, with only about 70 percent of
families headed by someone who was a non-Hispanic white.16 Educational attainment in Generation X probably will surpass that of baby boomers,
but the extent of improvement is still uncertain
since some people do not complete their education
until well into their 30s.17 Members of Generation X
appear to have benefited much less from rising living standards than did previous generations. Figure
2 shows that the 1975-77 cohort received median
incomes 3.7 percent below the median income of
the 1954-56 cohort at ages 33-35 and 15.5 percent
higher than the 1954-56 cohort at ages 36-38. Figure 3 shows that the median wealth of the 1975-77
cohort was unambiguously lower than the median
wealth of the 1954-56 cohort at ages 33-35 and
36-38, by 35.4 and 32.7 percent, respectively. These
observations are from 2010 and 2013, after asset
prices had declined.
A snapshot of families headed by someone
born after 1984. People born in the early to mid1980s and beyond have been termed “millennials”
or members of Generation Y. It is too early in their
adult lives to say very much about how millennials will fare economically or financially compared
with their elders. It is safe to say, however, that the
millennial generation will be the most diverse and
perhaps the best-educated of any generation.
III. Income, Wealth, Balance Sheets and Financial
Behaviors
Striking and persistent differences in the typical
family’s income, wealth, balance-sheet structure
and a measure of financial decision-making that
we call financial health are evident across the life
cycle. Some differences are evident across birth-year
cohorts or generations, too. There is no a priori reason to expect life-cycle patterns to change over time
in any particular way—for example, for the young or
the old to become relatively better or worse off compared to other stages of the life cycle.
In some important ways, however, we found that
differences across the typical life cycle have grown
larger since 1989, when our data begin. For example,
old families long have had more wealth than young
families, but the gap has grown substantially in
recent years and it is not obvious why. Likewise, we
found substantial deviations from a simple upward
trend in income and wealth over time, which one
would expect as living standards rise.
Income. A family’s income generally follows a
hump-shaped pattern over its life cycle, starting at
a low level, peaking in middle age, then declining
in old age. Figure 2 illustrates this pattern using five
cohorts of family heads born throughout the 20th
century. Figure 4 combines the observed trajectories
of median family incomes for virtually all three-year
cohorts born during the 20th century. Individual
cohorts were observed up to nine times at threeyear intervals between 1989 and 2013.
The logarithmic vertical scale in Figure 4 makes it
easy to see the smoothness of changes in the growth
rate of the typical life cycle in family incomes. Constant rates of change are represented on a logarithmic scale by straight lines so that growth rates
are simply the slope of a line segment. Hence, the
figure reveals that the income of a typical family in
its 20s rises rapidly in percentage terms. The typical
rate of growth slows through the 30s and 40s, with
median family income peaking at about age 50.
After a plateau during the 50s, median incomes
decline more or less continuously from the early
60s onward.
The relative constancy of the life cycle of family
income over time is shown in Figure 5. Middleaged families had the highest median incomes by
far throughout 1989-2013, and there was little net
change on balance. Old families had the lowest
The Demographics of Wealth 11
Figure 4. Median Family Income for Three-Year Birth Cohorts Observed during 1989-2013
100,000
2013 dollars: logarithmic scale
1901
1904
1907
1910
1913
1916
1919
1922
1925
1928
1931
1934
1937
1940
1943
1946
1949
1952
1955
1958
1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
10,000
19
25
31
37
43
49
55
61
67
73
79
85
Age of family head at center of 3-year birth cohort when observed
All dollar amounts are expressed in 2013 dollars, deflated by the CPI-U-RS (Consumer Price Index for Urban Consumers, Research
Series). The vertical scale is logarithmic.
To illustrate the life cycle of family incomes, we assigned each family in each survey to a three-year birth cohort, beginning with all
families whose head was born between 1900 and 1902. We call this the 1901 cohort; these families were headed by someone who was
between 87 and 89 years old in 1989, the first survey year. The 1904 cohort consists of all families whose head was born between 1903
and 1905, and so on. The last group we include is the 1994 cohort, consisting of families headed by someone born between 1993 and
1995. These family heads were between 18 and 20 years old in 2013, the last survey year.
Each three-year birth cohort is observed a maximum of nine times at three-year intervals. The 1970 cohort, for example, was aged
18-20 in 1989; 21-23 in 1992; and so on through ages 42-44 in 2013.
Unless otherwise noted, the source for all tables and figures is the Federal Reserve’s Survey of Consumer Finances.
median incomes throughout the period, although
this group was the only one to experience a net
increase over the 24 years. Figure 6 shows more
clearly that the median income of old families
increased significantly relative to the overall median
income, rising from 61.3 percent in 1989 to 83.0
percent of the median income of all families in 2013.
Both middle-aged and young families’ median
12 Federal Reserve Bank of St. Louis
incomes declined slightly, relative to the overall
median income.
Wealth. A simple measure of a family’s financial
strength is its net worth, or wealth. Figure 7 shows
that median wealth generally is a non-decreasing
function of age. This contrasts sharply with median
income, which tends to decline significantly in later
life. Another contrast is that a typical family’s wealth
Figure 6. Median Family Income by Age of Family
Head Relative to Overall Median Family Income
80,000
160
70,000
140
60,000
120
50,000
100
Percent
2013 dollars
Figure 5. Median Income of Families by Age of
Family Head
40,000
30,000
20,000
10,000
0
80
60
Old (62+)
Middle-aged (40-61)
Young (<40)
All families
1989 1992 1995 1998 2001 2004 2007 2010 2013
All dollar amounts are expressed in 2013 dollars, deflated by
the CPI-U-RS (Consumer Price Index for Urban Consumers,
Research Series).
Median family income is the value of cash income, before taxes,
for the full calendar year preceding the survey for the family
that ranks exactly in the middle of a ranking by income. The
median income among all families decreased from $46,762 in
1989 to $46,668 in 2013, or 0.2 percent. See Sidebar 2 for more
information.
increases more gradually over a longer period of
time than is the case for income. Median wealth
rises at a slowly decreasing rate from the early
20s until about age 60. Thereafter, median wealth
appears to remain flat or decline slightly.
Figure 8 portrays the pronounced life cycle of
rising wealth over time. The typical young family
has very little wealth. The typical middle-aged family has accumulated wealth in the low six figures,
and the typical old family had over $200,000 of net
worth in recent survey years. The vastly different
experiences over time of families at different stages
in the life cycle are illustrated in Figure 9. Relative
to the overall median wealth—which declined
between 1989 and 2013 by $4,119 in 2013 dollars, or
4.8 percent—the typical old family fared extremely
well. The median wealth of an old family was 257.3
percent of the overall median in 2013, compared
with just 175.0 percent of the median in 1989. The
sharp increase in the ratio in 2010 and 2013 was
40
20
Old (62+)
Middle-aged (40-61)
Young (<40)
0
1989 1992 1995 1998 2001 2004 2007 2010 2013
Median family income is the value of cash income, before taxes,
for the full calendar year preceding the survey for the family that
ranks exactly in the middle of a ranking by income. See Sidebar 2
for more information.
The figure shows that the median income among families headed
by someone 62 or older in 2013 was 83.0 percent of the median income among all families. The median family headed by
someone aged 40 to 61 was 130.4 percent of the overall median
income. The median family headed by someone under 40 had
87.0 percent as much income as the overall median family.
due to the huge declines in overall median wealth
in those years, as Figure 8 shows. Although the
median wealth of old families decreased by 16.7
percent between 2007 and 2013, the median wealth
of all families decreased by 39.8 percent. Hardest hit
were middle-aged families, whose median wealth
decreased by 47.3 percent. Young families’ median
wealth declined by 35.5 percent.
An important conclusion we draw from longterm trends in median income and median wealth
is that old families generally have fared better than
either middle-aged or young families. This was true
in the data through 2007, even before the financial crisis and the Great Recession. Developments
reflected in the 2010 and 2013 SCF data accentuated the longer-term trends. Most strikingly, the
median-wealth gap between old and young families has increased from $129,899 in 1989 to $195,370
in 2013. Expressed as ratios, the median wealth of
old families increased from 7.6 times the median
The Demographics of Wealth 13
Figure 7. Median Family Net Worth for Three-Year Birth Cohorts Observed during 1989-2013
2013 dollars: logarithmic scale
1,000,000
1901
1904
1907
1910
1913
1916
1919
1922
1925
1928
1931
1934
1937
1940
1943
1946
1949
1952
1955
1958
1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
100,000
10,000
1,000
19
25
31
37
43
49
55
61
67
73
79
85
Age of family head at center of 3-year birth cohort when observed
All dollar amounts are expressed in 2013 dollars, deflated by the CPI-U-RS (Consumer Price Index for Urban Consumers, Research
Series). The vertical scale is logarithmic.
To illustrate the life cycle of family wealth, we assigned each family in each survey to a three-year birth cohort, beginning with all
families whose head was born between 1900 and 1902. We call this the 1901 cohort; these families were headed by someone who
was between 87 and 89 years old in 1989, the first survey year. The 1904 cohort consists of all families whose head was born between
1903 and 1905, and so on. The last group we included is the 1994 cohort, consisting of families headed by someone born between
1993 and 1995. These family heads were between 18 and 20 years old in 2013, the last survey year.
Each three-year birth cohort is observed a maximum of nine times at three-year intervals. The 1970 cohort, for example, was aged
18-20 in 1989; 21-23 in 1992; and so on through ages 42-44 in 2013.
wealth of a young family in 1989 to 14.7 times
in 2013.
Overall balance-sheet health. A household’s
balance sheet lists its assets and liabilities. Although
there is no such thing as a perfect balance-sheet
configuration or a one-size-fits-all set of prescriptions on how best to choose assets and liabilities,
several principles of wealth accumulation and
retention are reasonably clear. All else equal, each
of the following balance-sheet choices is likely to
support greater wealth accumulation:
14 Federal Reserve Bank of St. Louis
• Greater balance-sheet liquidity can support greater
wealth accumulation over time by buffering a
family against financial shocks, which can lead
to high-cost borrowing, distressed-asset sales, or
costly default on debts and other obligations;
• Greater asset diversification—including highreturn assets like stocks or a small business—can
lead to greater wealth on average over time due
to lower volatility for any given level of expected
return on assets (or equivalently, a higher
expected return for a given level of volatility),
Figure 8. Median Family Net Worth by Age of
Family Head
Figure 9. Median Family Net Worth Relative to
Overall Median Net Worth
300,000
250,000
300
Old (62+)
Middle-aged (40-61)
250
Young (<40)
All families
Young (<40)
200
150,000
Percent
2013 dollars
200,000
150
100,000
100
50,000
50
0
Old (62+)
Middle-aged (40-61)
1989 1992 1995 1998 2001 2004 2007 2010 2013
0
1989 1992 1995 1998 2001 2004 2007 2010 2013
All dollar amounts are expressed in 2013 dollars, deflated by
the CPI-U-RS (Consumer Price Index for Urban Consumers,
Research Series).
All dollar amounts are expressed in 2013 dollars, deflated by
the CPI-U-RS (Consumer Price Index for Urban Consumers,
Research Series).
Median family net worth is the value of total assets minus total
debts for the family that ranks exactly in the middle of a ranking
by net worth. The median wealth among all families declined
from $85,575 in 1989 to $81,456 in 2013, or 4.8 percent. See Sidebar 2 for more information.
Median family net worth is the value of total assets minus total
debts for the family that ranks exactly in the middle of a ranking
by net worth. See Sidebar 2 for more information.
Unless otherwise noted, the source for all tables and figures is the
Federal Reserve’s Survey of Consumer Finances.
reducing the likelihood of encountering costly
financial distress; and
• Lower leverage (debt-to-assets ratio) can lead
to greater wealth on average over time both
because borrowing itself is expensive and
because balance-sheet leverage amplifies any
shock to a family’s asset values into larger percent changes in net worth, raising the risk of
insolvency and of costly default on debt or other
obligations.
These balance-sheet practices can be described
as elements of prudent or conservative financial
decision-making. Figure 10 shows that old families
typically have much higher ratios of liquid assets
to total assets than middle-aged or young families. Median young and middle-aged families have
increased their holdings of liquid assets some-
The figure shows that the median net worth among families
headed by someone 62 or older was 257.3 percent of the median
net worth among all families. The median family headed by
someone aged 40 to 61 was 130.2 percent of the overall median
wealth. The median family headed by someone under 40 had 17.5
percent as much wealth as the overall median family.
what in recent years while the liquid-assets ratio
declined for the median old family, making the differences much smaller across age groups. Given the
larger asset holdings of middle-aged and old families, their actual holdings of safe and liquid assets still
are much greater than those of young families. In
2013, the median holdings of safe and liquid assets
were $82,766 among old families, $46,324 among
middle-aged families and only $14,021 among
young families.
Figure 11 shows that old families typically have
a much greater share of their assets invested in
financial and business assets, which provide both
asset diversification and higher average returns in
the long run than a portfolio consisting mostly of
tangible assets like a house, vehicles or other durable
goods.18 In 2013, the median share of assets held in
The Demographics of Wealth 15
Figure 10. Median Ratio of Safe and Liquid Assets to
Total Assets
Figure 11. Median Ratio of Financial and Business
Assets to Total Assets
12
45
Old (62+)
Middle-aged (40-61)
10
40
Young (<40)
All families
35
30
6
Percent
Percent
8
4
25
20
15
10
2
5
0
1989 1992 1995 1998 2001 2004 2007 2010 2013
Safe and liquid assets are defined as checking and saving
accounts, certificates of deposits, bonds and savings bonds.
These are assets that can be drawn upon quickly at low or no
cost in terms of fees or potential loss of value when selling on
short notice.
The figure shows the median among all families in each age
group of the percent of total assets held in the form of safe and
liquid assets.
0
Old (62+)
Young (<40)
Middle-aged (40-61)
All families
1989 1992 1995 1998 2001 2004 2007 2010 2013
Financial assets include all securities and accounts that can be
turned into cash. Business assets include the value of all privately
owned businesses minus its debts, shares in private businesses
minus the debts of the business for which the person is responsible, and investment real estate minus associated debt. Financial
and business assets include all of a family’s assets except tangible
assets, which include real estate, vehicles and other real property.
Financial and tangible assets are counted independently of any
debts owed by the person; business assets are expressed net of
the associated debt.
The figure shows the median among all families in each group
of the percent of total assets held in the form of financial and
business assets.
the form of financial and business assets was 34.8
percent among old families, 28.3 percent among
middle-aged families and only 18.5 percent among
young families. These ratios changed little on balance between 1989 and 2013.
Figure 12 shows the median ratios of debt to assets
for each age group. The median old family had
little or no debt throughout the period. The median
debt-to-assets ratios among middle-aged families
increased by 10 percentage points between 1989
and 2013, rising from 14.2 percent to 25.3 percent.
Among young families, the median debt-to-assets
ratio increased from 34.4 percent to 44.9 percent,
with a peak of 53.0 percent in 2010.
Putting together the patterns and trends in family
income and balance-sheet choices described above,
we can provide two answers to the question of how
16 Federal Reserve Bank of St. Louis
the typical old family has accumulated so much
wealth. Saving from peak earnings in middle age
represents a crucial first step; young families simply do not have enough income, nor have they had
enough time, to accumulate wealth from saving.
Another ingredient appears to be prudent balance-sheet management. Substantial cash reserves
remain part of the median old family’s balance sheet
even as income risk and balance-sheet leverage
decline. Asset diversification beyond housing and
durable goods into financial and business assets
becomes increasingly evident as families age. Debt
is paid down over time, freeing up cash flow, eliminating the high cost of borrowing and reducing the
risk of costly default on any obligations.
Our data document a correlation between prudent
balance-sheet management and wealth across the
Figure 12. Median Ratio of Total Debt to Total
Assets
60
50
Percent
40
Old (62+)
Middle-aged (40-61)
Young (<40)
All families
30
20
10
0
1989 1992 1995 1998 2001 2004 2007 2010 2013
The chart shows the median of the ratio of total debt to total
assets among all families in an age group.
life cycle, but we cannot identify causation. Do the
diversification and debt choices of middle-age and
old families cause them to be wealthier, or does
greater wealth cause or allow them to be better
diversified and less leveraged? We suspect there are
elements of two-way causation, that is, stronger
balance sheets probably do contribute to greater
wealth accumulation, but it may also be the case that
the passage of time—as described by the life cycle—
makes it easier to diversify and pay off debt when
greater wealth is available.
Nonetheless, an important question for future
research is the extent to which young families
might accumulate wealth more rapidly if their balance-sheet choices resembled those of old families
more closely. An example would be to delay purchase of a home with its attendant debt burden until
it was possible to buy a house that did not make the
family’s balance sheet dangerously undiversified
and highly leveraged.
Financial behaviors and financial health. A third
factor that may contribute to greater wealth accumulation among old families is the improving quality of
routine financial decision-making over the life cycle,
which we quantify with a financial health scorecard.
(See Sidebar 3.) The logic behind our scorecard is that
a family’s ability to make good everyday financial
decisions—its financial health—and its ability to accumulate wealth over time are likely to be correlated.
Each financial choice we examined includes two feasible alternatives, one of which is more likely to lead
to financial success. For example, saving is clearly
preferred to not saving, even if only a small amount
is saved. Paying one’s bills on time is clearly preferred
to missing a payment, and so on. For the question
about credit cards, we applied a series of screens if
a family did not have any credit cards. Having been
denied a card or choosing not to apply because the
family member expected to be rejected were scored
as negative signals, earning a score of zero. Owning
no credit cards by choice was a positive sign, earning
a score of one.
Table 3 shows that old families demonstrated the
highest average financial health scores in the period
1992-2013 by a wide margin. Because the standard
errors of estimation for each group covering the
entire sample ranged from 0.008 (for middle-aged
families) to 0.011 (for old families), we are highly
confident in a statistical sense that the average
scores are different from one another. Based on our
measure of financial health, old families on average
make far better every-day financial choices than
middle-aged families, who, in turn, make somewhat
better choices than young families.
IV. Cohort Effects in Family Income and Wealth
The technical appendix contains regression
results that support the hypothesis that levels of
income and wealth rose during the first several
decades of the 20th century, but then stopped
rising for most families around mid-century. The
analysis that produced these results holds constant
a number of important demographic characteristics, such as age, educational attainment, race or
ethnicity, family structure and health status. One
explanation for stagnating income and wealth for a
given set of demographic characteristics is that the
arrival of the baby boomers somehow disrupted the
rise of living standards. Another possibility is that
the ends of the Great Depression and World War II
were associated with social, political and economic
changes that favored generations born before the
baby boomers.
An obvious place to start is with relative cohort
sizes. The idea is that relatively small cohorts may
The Demographics of Wealth 17
Sidebar 3: A Financial Health Scorecard That Predicts Wealth Accumulation
To characterize the quality of basic financial decisionmaking by a typical family, we calculated a financial health
scorecard for each family in each wave of the SCF.19 The
scorecard consists of five questions that were asked of
each of the 38,385 families that participated in the survey
between 1992 and 2013:20
• Did you save any money last year?
• Did you miss any payments on any obligations in the
past year?
• Did you have a balance on your credit card after the
last payment was due?
• Including all of your assets, was more than 10 percent
of the value in liquid assets?
• Is your total debt service (principal and interest) less
than 40 percent of your income?
How we scored the responses to these questions and
the average number of points all respondents received on
each question are in Table 4.21 To investigate the predictive power of the scorecard for financial success, we split
the SCF sample in each survey year into 48 unique group
combinations, based on:
Table 3. Average Group Scores for Families on the
Financial Health Scorecard
Average financial health score
in all years, 1992- 2013
All families
3.01
Old: 62 or older
3.46
Middle-aged: 40-61
2.88
Young: Under 40
2.82
A family’s score on the financial health scorecard is the sum of the
individual scores to questions listed in Table 4, with a range of zero
to five. A score of five indicates the highest financial health; a score
of zero indicates the lowest financial health.
The standard errors of estimation covering the entire sample were
0.009 for young families, 0.008 for middle-aged families and 0.011
for older families. Thus, we are highly confident in a statistical
sense that the average score of each group is different from each
other group.
18 Federal Reserve Bank of St. Louis
• Three age groups: younger than 40, 40-61, and 62 and
older;
• Four education groups: less than high school diploma;
high school diploma, General Educational Development
(GED) certificate or vocational/technical certificate;
two- or four-year college degree only; and graduate or
professional degree;
• Four racial and ethnic groups: black, Hispanic, Asian
and white.
The individual item and overall index scores in 2013
were remarkably similar to the averages computed over all
eight waves of the SCF for which all the data were available (1992-2013). In other words, the elements of financial
health we estimated appear to be stable over time.
The average group scores are financially meaningful,
too—the simple correlation co-efficient between the
average financial health score of a group and the 1992-2013
average of median inflation-adjusted net worth (expressed
as a logarithm) for each of the 48 groups was 0.67. In other
words, our financial health scorecard was very good at
predicting how much wealth a group was likely to have.
have attracted scarcity premiums in labor, housing and financial markets, whereas relatively large
cohorts paid crowding penalties in those markets.22
Figure 13 displays the number of babies (under 1
year old) in the United States between 1896 and
2015. The striking 20-percent decline in the number of infants between 1925 and 1937 likely reflects
the massive economic and social disruption of the
Great Depression, as well as the slowdown in the
population’s natural growth rate after earlier high
immigration rates declined. The infant population
began to rise again in the late 1930s, but it was not
until 1945 that the size of this population reached
the level of 20 years earlier. Thus, these very small
birth-year cohorts may have been favored later in
life in a variety of ways, including relatively higher
earnings, lower house prices and stronger growth
in financial-asset prices. Sociologist Elwood Carlson called the generation born between 1929 and
1945 “the lucky few” precisely because it was the
Table 4. Questions in the Financial Health Scorecard
Scoring
Mean score in eight SCF
waves, 1992-2013
Mean score in 2013
SCF only
Less = 1;
Same or more = 0
0.56
0.53
0.84
0.85
0.44
0.47
Questions
1. After adjusting for any purchases of durable goods or
investments you made, did you spend more, the same or
less than your income in the past year?
2. Does either of these statements apply to you?
“We sometimes got behind or missed payments;” or
“Considering all the various loan or mortgage payments
we made during the last year, not all of the payments were
made the way they were scheduled; sometimes, they were
made later or missed.”
No, neither one
applies = 1;
Yes, one or both
apply = 0
3. Do any of these statements apply to you?
“We carried over a credit-card balance after we made our
last payment;” or
“We have been turned down in the past five years by a
particular lender or creditor when I (or my {husband/wife/
partner}) made a request for credit, or we were not given
as much credit as we applied for;” or
“There was a time in the past five years that we thought of
applying for credit at a particular place, but changed our
minds because we thought we might be turned down.”
No, none applies
or no credit cards by
choice = 1;
Yes, one or more
apply = 0
4. Including all of your assets, was more than 10 percent
of the value in safe and liquid assets, defined as liquid
accounts (checking, saving or money-market accounts),
certificates of deposits, bonds or savings bonds?
Yes = 1, No = 0
0.27
0.26
5. Is your total debt service, including both scheduled repayment of principal and interest owed, less than 40 percent
of your income?
Yes = 1, No = 0
0.91
0.92
Total score
0 to 5 possible
3.01
3.03
A family’s score on the financial health scorecard is the sum of the individual scores, with a range of zero to five. A score of five indicates the highest financial health; a score of zero indicates the lowest financial health.
Splitting the sample in each SCF wave into 48 unique groups, based on three age groups (less than 40, 40-61, and 62 and over),
four education groups (less than high school, high school or GED or vocational/technical certificate, two- or four-year college only, and
graduate or professional degree), and four racial and ethnic groups (African-American, Hispanic of any race, Asian and non-Hispanic
white), the simple correlation co-efficient between a group’s average financial health scorecard score for 1992-2013 and the group’s
inflation-adjusted median logarithm of net worth averaged across the eight waves is 0.67.
first American generation to be smaller in number
than those that came before. Carlson argued that
African-Americans and women born in those years
also enjoyed historically unprecedented opportunities throughout their adult lives.
The infant population recovered during the
baby boom that began in the 1940s. Peaking in
the early 1960s, the infant population doubled in
little more than two decades. Given trends in births
both before and after, the baby boom now appears
to have been a historical aberration. Hence, it is
plausible that baby boomers suffered from crowd-
ing in labor, housing and financial markets. This
may have resulted in unfavorable developments in
income and wealth accumulation.
Another set of explanations of the apparent end
of rising levels of income and wealth for a given set
of demographic factors relates to changes in economic growth and social policies. Post-World War
II economic growth was very rapid, and the value
of housing and financial assets increased strongly.
Rather than attracting any special advantages
related to their absolute numbers, people born in
the first half of the 20th century simply may have
The Demographics of Wealth 19
Figure 13. Estimated Number of Children under
1 Year Old
5.0
4.5
4.0
3.5
Millions
3.0
2.5
2.0
1.5
1.0
1895
1900
1905
1910
1915
1920
1925
1930
1935
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
2015
0.0
Estimated number of children under 1 year old
SOURCES: Census Bureau and authors’ own estimates.
been in the right place at the right time as they
were lifted by a rising tide. A related channel of
causation is the postwar expansion of the safety net,
especially for retired people. The steadily increasing
generosity of Social Security, the creation of Medicare in the 1960s and, 40 years later, the significant
expansion of Medicare in the prescription drug benefit greatly increased the resources directed to adults
reaching retirement age in the 1990s and 2000s.
Will the favorable income and wealth trends
observed among today’s old adults resume at some
point? We cannot know for sure, but it appears
unlikely that baby boomers—who are just now
entering retirement in large numbers—will accrue
the same incomes and wealth that pre-baby boomers
received for given demographic characteristics. First,
the baby boomers already have significantly lower
demographically adjusted incomes and wealth, as
we documented above. There is little time to make
up these shortfalls and little reason to believe that
social policies will change to assist these adults.
Second, reductions in social programs that help old
adults appear to be more likely than increases.
Among cohorts born after the baby boomers,
the members of Generation X stand out for having
low incomes and wealth for a given set of demo20 Federal Reserve Bank of St. Louis
graphic characteristics. Someone born in the 1970
cohort, for example, appears to have an income 25
percent lower and wealth 40 percent lower than an
otherwise identical person born in 1940. Given their
youth, it is impossible to say whether or to what
extent people born in the 1980s and 1990s will fare
better in the years ahead. The most we can say is
that there is, as of 2013, no convincing evidence
that the millennials will do appreciably better than
the members of Generation X.
V. Conclusion
We documented a very strong association between
a family head’s age and the family’s level of income
and wealth. There also is a strong association
between age and several indicators of balance-sheet
strength and financial health. If anything, the associations have become stronger over time, and the gaps
between age groups have widened.
We also found evidence of significant birth-year
cohort effects on income and wealth. Family heads
born about 1940 earned higher incomes and accumulated more wealth than family heads born before
or later, holding constant a host of demographic,
idiosyncratic and period-specific factors.
Technical Appendix
Disentangling the Effects of Birth Year, Life-Cycle
Stage and Historical Time Period on Income and
Wealth: A Regression Analysis
Multiple-regression analysis helps us sort out
and quantify life-cycle, birth-year cohort and other
factors’ influences on income and wealth in the SCF.
To be sure, we would be much more confident about
our estimates if we had data for hundreds of thousands of families observed over many decades or,
even better, thousands of the same families observed
over their entire life cycles. Instead, we have only
about 40,000 family observations collected over
a 24-year period in a series of cross sections. Key
assumptions, such as the constancy of life-cycle
effects over time and across cohorts and the lack of
important interaction effects among the independent
variables, must be kept, therefore, in mind. Nonetheless, the exploratory models we describe here appear
to provide important insights into the demographics
of income and wealth. We highlight cohort effects in
what follows.
Cohort effects in family income. Table 5 contains
estimation results from a regression of the logarithm
of family income on demographic, idiosyncratic,
birth-year cohort and time variables. We have 41,485
observations across the nine survey waves, and the
fit of the pooled regression model is good. The R2 is
43.3 percent, and estimated levels of statistical significance are high for many coefficients.
We regressed the logarithm of family income in a
given year on a cubic function of age to control for
life-cycle effects; on standardized (that is, demeaned
by the family’s demographic profile) measures of
marital status, family size, saving behavior and health
status to isolate idiosyncratic factors potentially
correlated with wealth accumulation; on education
and race or ethnicity indicators to capture the effects
of human capital, social class and potential legacies of discrimination; on year dummies to capture
time effects during the year of observation; and, of
primary interest, on a large set of birth-year cohort
indicator variables.
We constructed five-year birth-year cohorts,
beginning with families born between 1898 and
1902. We refer to this as the 1900 cohort. The last
cohort includes family heads born between 1988 and
1992, or the 1990 cohort. We chose the 1940 cohort as
the omitted category because it is near the middle of
the sample of birth years and because it turns out to
be a good example of families that appeared to experience an unusually favorable cohort effect.
Estimates of the coefficients on demographic and
idiosyncratic variables are reported in the upper
portion of Table 5. In general, the estimates are highly
significant with the expected signs. Family income
rises with the age of the family head, but it grows at a
decreasing rate. There is an important degree of curvature in the age-income profile, too. Idiosyncratic
factors reliably associated with increases in family
income include marital status, having more children
than average, saving money regularly and enjoying
above-average health.23 Higher levels of educational
attainment are very strongly predictive of higher
income. The regression results imply that, if all other
factors are held constant, being African-American or
Hispanic predicts significantly lower income than an
otherwise similar Asian family (the excluded category). Asian families, in turn, earn significantly lower
incomes than whites.
Time dummies for the 1989–2013 sample dates
are shown near the bottom of Table 5. The significant values on the estimated coefficients for the
2001, 2004 and 2007 dummy variables imply that
families’ incomes generally were higher than would
have been expected based on the overall sample.
Equivalently, incomes in 1992, 1995, 1998, 2010 and
2013 appear more “normal” in that they were not
significantly different from incomes in 1989, the
omitted year.
The variables of greatest interest in Table 5 are
the cohort dummy variables. Figure 14 displays the
information contained in the coefficient estimates,
which we interpret as the marginal effect of a family’s birth cohort on its inflation-adjusted income. If
The Demographics of Wealth 21
Table 5. Pooled Regression of Logarithm of Family Income on Demographic, Idiosyncratic,
Birth-Year-Cohort and Time Variables
Variable
Beta
T–Statistic
Intercept
Variable
Beta
T–Statistic
–4.11
–1.02
868.36
31.53
Birth year 1933–37 indicator
Age in years
14.41
12.18
Birth year 1938–42 (omitted)
Age squared
–0.186
–8.62
Birth year 1943–47 indicator
–4.08
–1.11
5.40
Age cubed
Standardized marital status
0.00
95.05
Birth year 1948–52 indicator
–15.26
–2.94
68.61
Birth year 1953–57 indicator
–13.17
–1.88
–17.35
–1.92
Standardized number of children
4.01
6.71
Birth year 1958–62 indicator
Standardized saving indicator
18.60
14.35
Birth year 1963–67 indicator
–23.10
–2.08
Standardized health status
49.30
29.68
Birth year 1968–72 indicator
–25.39
–1.93
High school dropout indicator
–175.55
–82.21
Birth year 1973–77 indicator
–17.49
–1.14
High school graduate or GED indicator
–124.99
–81.97
Birth year 1978–82 indicator
–18.42
–1.06
Some college indicator
–89.52
–49.81
Birth year 1983–87 indicator
–16.30
–0.83
Birth year 1988–92 indicator
–12.19
–0.55
29.42
9.26
Year 1989 (omitted)
–35.29
–9.69
Year 1992 indicator
–2.86
–0.87
–17.85
–4.57
Year 1995 indicator
–4.02
–1.03
Year 1998 indicator
2.16
0.45
3.65
0.14
Year 2001 indicator
16.23
2.77
–3.88
–0.21
Year 2004 indicator
17.11
2.45
Birth year 1908–12 indicator
–27.39
–1.87
Year 2007 indicator
23.47
2.87
Birth year 1913–17 indicator
–23.34
–1.93
Year 2010 indicator
–7.78
–0.84
Birth year 1918–22 indicator
–21.40
–2.21
Year 2013 indicator
–6.28
–0.60
Birth year 1923–27 indicator
–18.43
–2.47
R–squared of first regression
Birth year 1928–32 indicator
–14.87
–2.64
Observations
College graduate (omitted)
White indicator
African–American or black indicator
Hispanic of any race indicator
Asian or other (omitted)
Birth year 1898–1902 indicator
Birth year 1903–07 indicator
.433
41,485
NOTE: GED = General Educational Development test. Dependent variable is logarithm of inflation-adjusted family income in year t, excluding all nonpositive observations. Sample years are 1989, 1992, 1995, 1998, 2001, 2004, 2007, 2010 and 2013. Coefficients are expressed in
percent; for example, the value –17.35 for birth year 1958–62 indicator means negative 17.35 percent.
Unless otherwise noted, the source for all tables and figures is the Federal Reserve’s Survey of Consumer Finances.
there were no birth-year cohort effects determining family income, all of these parameter estimates
would be statistically indistinguishable from zero.
Coefficient estimates that are statistically different from zero at the 10 percent level are shown as
dark-colored bars; those that are not statistically
sufficient are light-colored.
As it turns out, cohort effects on income appear
to be important. Families born in the five years centered on 1940 (the omitted group in the regression)
do not differ to a statistically significant degree
from those born during the 1900 or 1905 cohorts. It
may be that there really are cohort effects, but the
22 Federal Reserve Bank of St. Louis
small number of sample members born before 1908
leads to imprecise estimates—or there may be no
such effects. Survivorship bias also may be important because those family heads born before 1908
and still alive at the time of the surveys may not be
representative of the entire original cohort to which
they belong. In particular, they may be relatively
better off in terms of health, education, lifetime
earnings and wealth.
All cohorts between 1910 and 1930 (including
family heads born between 1908 and 1932), however,
have statistically significantly lower incomes than
those of families with heads born in the 1940 cohort
Figure 14. Marginal Effect of Family Head’s Birth
Year on Logarithm of Family Income Relative to
Being Born in the Period 1938-42
10
5
0
Percent
–5
–10
–15
–20
–30
1900
1905
1910
1915
1920
1925
1930
1935
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
–25
Bars represent the estimated percentage difference in income
between a family in a five-year birth-year cohort centered around
the given year and the cohort of families with heads born in the
five-year cohort centered around 1940. The bars are co-efficient
estimates from the regression reported in Table 5.
Solid green bars are statistically different from zero at the 10-percent confidence level. Outlined bars are not statistically significantly different from zero at the 10-percent confidence level.
(1938-42). The estimated magnitudes of difference—
increasing monotonically from a 27 percent lower
level among the 1910 cohort to a 15 percent lower
level among the 1930 cohort—are consistent with a
generally rising level of family income across successive birth-year cohorts: Typical levels of income
rose with overall standards of living.
In terms of family income, the regression results
suggest that families headed by someone born
in the 1935, 1940 or 1945 cohort (that is, between
1933 and 1947) are statistically indistinguishable
from one another. However, beginning with the
1950 cohort (i.e., families headed by someone
born between 1948 and 1952), successive cohorts
through 1970 (born between 1968 and 1972) had
statistically significantly lower incomes than those
of the 1940 cohort. Moreover, the estimated magnitudes are economically significant: between 13
and 25 percent lower than the 1940 cohort. Finally,
all five-year cohorts that began in 1975 or later had
estimated income shortfalls of 12 to 18 percent.
However, these effects were not measured precisely. The fading of a statistically significantly negative cohort effect after the 1970 cohort may be due
to a true diminution of the effect or it may be due
to relatively small sample sizes and high variability
among younger families’ incomes.
Cohort effects in family net worth. Given evidence of important cohort effects in family income,
it would not be surprising to find similar effects in
family wealth. After all, unusually high incomes for
the 1935, 1940 and 1945 cohorts might have supported higher saving rates than those observed
among earlier- and later-born cohorts.
A logarithmic specification of net worth is problematic because about 8 percent of all family-year
observations of net worth are zero or negative in
the SCF. Dropping these observations could alter
our estimates of important relationships in the data
because we know the dropped observations would
not be a random sample of the population. Instead,
they are much more likely to be young, less-educated and nonwhite families. They also would be
more likely to be baby boomers and members of
Generation X.
An alternative transformation of net worth—the
inverse hyperbolic sine (IHS) function—allows us to
include zero or negative wealth observations while
retaining an interpretation of results that is similar
to the interpretation of results in the log model.24
We applied the Halvorsen–Palmquist transformation to coefficient estimates of indicator variables
and report the results in Table 6, which presents
estimates for a set of independent variables that is
identical to the set used in the log of income specification with one exception. We include the standardized—that is, de-meaned–square root of family
income to capture idiosyncratic variations in family
income—that is, an income that is unusually high
or low compared with the family’s demographically
predicted income, holding constant all of the other
variables included in the model. It is important
to control for unusual circumstances, like lottery
winners or people who have suffered large business
losses, in order to separate the random effects of
income (windfalls and catastrophes) from wealth
accumulation. The R2 is 66.2 percent, and estimated
levels of statistical significance are high for many
coefficients.
The Demographics of Wealth 23
Figure 15. Marginal Effect of Family Head’s Birth
Year on Transformed Family Wealth Relative to
Being Born in the Period 1938-42
20
10
0
Percent
–10
–20
–30
–40
–50
–70
1900
1905
1910
1915
1920
1925
1930
1935
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
–60
Bars represent the estimated percentage difference in wealth
between a family in a five-year birth-year cohort centered around
the given year and the cohort of families with heads born in the
five-year cohort centered around 1940. The bars are co-efficient
estimates from the regression reported in Table 6.
Solid green bars are statistically different from zero at the 10-percent confidence level. Outlined bars are not statistically significantly different from zero at the 10-percent confidence level.
The co-efficient estimate for the 1990 cohort is 150.0, but the
number of observations is very small.
24 Federal Reserve Bank of St. Louis
Figure 15 illustrates our estimates of the birthyear cohort’s marginal effect on a family’s wealth.
As before, the wealth regression holds demographic, idiosyncratic and time effects constant. The
dark-colored bars in the figure represent effects that
are statistically significant at the 10-percent level,
while the light-colored bars represent effects that are
not significant.
The marginal effect on wealth of being born
during 1898-1902 rather than during 1938-42 is not
statistically significantly different than zero, even
though the point estimate is about negative 35 percent. The small number of family heads born around
1900 and still alive in 1989 makes the estimate
imprecise. Beginning with the 1905 cohort, however, all cohorts through 1935 appear to have statistically significantly less wealth than the 1940 cohort,
holding constant many other factors. The shortfalls
rise from negative 66.2 percent to 14.2 percent, and
they generally are estimated precisely.
We cannot distinguish between the wealth of the
1940 and 1945 cohorts, and the estimated co-efficient on the 1945 cohort is small. Cohorts between
1950 and 1965 have decreasing point estimates
between negative 17.2 and negative 31.5 percent,
but none of them is estimated precisely. The point
estimate for the 1970 cohort is negative 40.0 percent, and it is marginally significant. Estimates are
imprecise between 1975 and 1985. The 1990 estimate
is large, positive and statistically significant, but we
hesitate to attach any importance to it because it
represents a small group of very young families.
Table 6. Pooled Regression of Transformed Family Net Worth on Demographic, Idiosyncratic,
Birth-Year-Cohort and Time Variables
Variable
Beta
Intercept
–41.34
–0.79
Birth year 1933–37 indicator
0.30
125.96
Birth year 1938–42 (omitted)
Age in years
26.53
8.24
Birth year 1943–47 indicator
Age squared
–0.10
Standardized square root of income
Age cubed
Standardized marital status
Standardized number of children
Standardized saving indicator
T–Statistic
Variable
Beta
T–Statistic
–14.23
–1.60
–8.32
–0.98
–1.84
Birth year 1948–52 indicator
–17.31
–1.49
0.00
–1.25
Birth year 1953–57 indicator
–19.99
–1.29
401.42
47.84
Birth year 1958–62 indicator
–20.48
–1.03
6.44
4.32
Birth year 1963–67 indicator
–31.53
–1.39
79.67
18.97
Birth year 1968–72 indicator
–39.96
–1.57
Standardized health status
218.41
29.77
Birth year 1973–77 indicator
–24.04
–0.73
High school dropout indicator
–98.68
–82.84
Birth year 1978–82 indicator
–15.98
–0.41
High school graduate or GED indicator
–94.65
–77.35
Birth year 1983–87 indicator
13.03
0.25
Some college indicator
–88.27
–49.80
Birth year 1988–92 indicator
149.98
1.68
College graduate (omitted)
White indicator
Year 1989 (omitted)
5.40
Year 1992 indicator
–1.66
–0.21
African-American or black indicator
–83.46
50.12
–20.66
Year 1995 indicator
0.32
0.03
Hispanic of any race indicator
–67.00
–11.81
Year 1998 indicator
–0.50
–0.04
Year 2001 indicator
17.86
1.13
Birth year 1898–1902 indicator
–34.84
–0.63
Year 2004 indicator
–4.91
–0.29
Birth year 1903–07 indicator
–66.22
–2.37
Year 2007 indicator
2.01
Birth year 1908–12 indicator
–58.85
–2.48
Year 2010 indicator
–49.83
–3.00
Birth year 1913–17 indicator
–56.53
–2.86
Year 2013 indicator
–58.82
–3.41
Birth year 1918–22 indicator
–30.97
–1.58
R squared of first regression
Birth year 1923–27 indicator
–39.27
–2.74
Observations
Birth year 1928–32 indicator
–29.94
–2.62
Scaling parameter, theta
Asian or other (omitted)
0.10
.662
41,485
.0001
NOTE: GED = General Educational Development test. Dependent variable is inflation-adjusted net worth after application of the inverse
hyperbolic-sine transformation: ASINH(Net Worth * Theta)/Theta, where Theta = .0001. Estimates shown for coefficients for indicator
variables are expressed after application of the Halvorsen-Palmquist transformation {100 * [exp(theta * beta) − 1]}. Sample years are
1989, 1992, 1995, 1998, 2001, 2004, 2007, 2010 and 2013. Interpretation of coefficients for indicator variables is analogous to the log
specification; for example, the value –20.48 for birth year 1958–62 indicator means negative 20.48 percent.
The Demographics of Wealth 25
1 See Emmons and Noeth (2015a, 2015b).
2 See Bricker et al. for an extensive discussion of
the SCF design and methodology.
3 See Bricker et al. for information on how the head
of a family is determined in the SCF.
4 See Census Bureau.
5 See Bureau of Labor Statistics.
6 See Bureau of Labor Statistics.
7 See Bricker et al.
8 In this essay, we use the term “white” to mean
non-Hispanic white. We use the terms “black,”
“non-Hispanic black” and “African-American”
interchangeably. Hispanics may be of any race.
The category “Asian or other origin” includes not
only people with origins in Asia but also those who
identify as American Indian, Alaska native, native
Hawaiian or Pacific Islander. Because Asians represent about 80 percent of this group in population
estimates published by the Census Bureau, we refer
to the group as Asian in what follows.
9 We provide detailed information about each of
these financial indicators in the next section.
10 This is somewhat surprising given the significant
and increasing over-representation of nonwhites
among young families and the over-representation of whites among old families during this
period. It turns out that the shares of middleaged families represented by each of the four
racial and ethnic groups changed at about the
same rate as their respective population shares
between 1989 and 2013. For example, the white
share of middle-aged families declined from
75.9 percent in 1989 to 69.8 percent in 2013.
The white share of all families declined from
74.8 percent to 70.1 percent; so, whites moved
from being slightly over-represented among
middle-aged families in 1989 to being slightly
under-represented in 2013.
11 Whites made up 69.8 percent of middle-aged
families and 60.7 percent of young families
in 2013.
12 See Howe and Elliott.
13 Survivorship bias may be important in explaining the failure of median wealth to decline at
advanced ages. If families with higher average
wealth tend to live longer—for example, families
with more education—then the median wealth
26 Federal Reserve Bank of St. Louis
14
15
16
17
18
19
20
21
22
23
24
among the surviving population at any time
could be constant or increasing even if most
individual surviving families’ wealth levels are
declining because low-wealth families are disappearing more rapidly. This is one reason why
we use a regression framework below. It allows
us to hold constant a number of important
demographic, idiosyncratic and time-period
influences on income and wealth to isolate
cohort effects.
These statistics about race for the silent generation represent family heads born during 1928-49.
These statistics refer to middle-aged families
observed in the 2004 SCF, representing birth
years 1943-64.
These statistics refer to young families observed in
the 2004 SCF, representing birth years 1965-86.
See Emmons and Noeth for discussion of trends
in educational attainment (2015b).
See Emmons and Noeth (2013, Tables 1 and 2) for
evidence from the Survey of Consumer Finances
that financial assets have produced much higher
returns than housing over long time periods.
See Emmons and Noeth (2014) for more discussion of the scorecard and its correlation with
wealth accumulation.
We excluded 1989 because it did not contain a
satisfactory version of the first question in our
scorecard.
The questions in the text are paraphrases; the
precise wording of the questions is in Table 4.
Easterlin documents the influence of cohort size
on economic and social outcomes with particular emphasis on the baby boom generation.
The explanatory variables expressed in standardized form are simple deviations from a
family’s predicted value of that variable based
on its demographics. For example, a family with
three children when its age-, education- and
race/ethnicity-predicted value was two is coded
as one. A family’s health status is expressed as
the difference between its actual reported health
status and the average of its demographic
group, and so on.
See Pence (2006) and Gale and Pence (2006) for
extensive discussion and application of the IHS
transformation to balance sheet data.
CD15-32466 07-15
Endnotes
References
Bricker, Jesse; Dettling, Lisa J.; Henriques, Alice;
Hsu, Joanne W.; Moore, Kevin B.; Sabelhaus,
John; Thompson, Jeffrey; and Windle, Richard
A. “Changes in U.S. Family Finances from 2010
to 2013: Evidence from the Survey of Consumer
Finances.” Federal Reserve Bulletin, September
2014, Vol. 100, No. 4.
Bureau of Labor Statistics. Labor Force Statistics from
the Current Population Survey, 2015.
Carlson, Elwood. The Lucky Few: Between the
Greatest Generation and the Baby Boom.
New York: Springer, 2008.
Census Bureau. Household Estimates for the United
States, by Age of Householder, by Family Status:
1982 to Present, 2015.
Easterlin, Richard A. Birth and Fortune: The Impact
of Numbers on Personal Welfare, 2nd ed. Chicago:
University of Chicago Press, 1987.
Emmons, William R.; and Noeth, Bryan J. “Economic Vulnerability and Financial Fragility.” Federal
Reserve Bank of St. Louis Review, September/
October 2013, Vol. 95, No. 5, pp. 361-88.
Emmons, William R.; and Noeth, Bryan J. “Five
Simple Questions that Reveal Your Financial
Health and Wealth.” Federal Reserve Bank of
St. Louis In the Balance, December 2014, No. 10.
Federal Reserve Board, Survey of Consumer
Finances. See www.federalreserve.gov/
econresdata/scf/scfindex.htm.
Gale, William G.; and Karen M. Pence. “Are Successive Generations Getting Wealthier and, If So,
Why? Evidence from the 1990s.” Brookings Papers
on Economic Activity, 2006, No. 1, pp. 155–234.
Halvorsen, Robert, and Raymond Palmquist. “The
Interpretation of Dummy Variables in Semilogarithmic Equations.” American Economic Review,
1980, Vol. 70, No. 3, pp. 474–75. See www.jstor.org/
stable/1805237.
Howe, Neil; and Elliott, Diana. “A Generational
Perspective on Living Standards: Where We’ve
Been and Prospects for the Future.” Forthcoming
in Economic Mobility: Research & Ideas on
Strengthening Families, Communities & The Economy, proceedings of a Federal Reserve System
Community Development Research Conference,
Washington, D.C.: Federal Reserve Board, 2015.
Pence, Karen M. “The Role of Wealth Transformations: An Application to Estimating the Effect of
Tax Incentives on Saving.” Contributions to Economic Analysis and Policy, 2006, Vol. 5, No. 1.
Sabelhaus, John. “Comment on Gale and Pence.”
Brookings Papers on Economic Activity, 2006,
No. 1, pp. 220–25.
Emmons, William R.; and Noeth, Bryan J. “Race,
Ethnicity and Wealth.” The Demographics of
Wealth, Federal Reserve Bank of St. Louis, February
2015a, No. 1.
Emmons, William R.; and Noeth, Bryan J. “Education
and Wealth.” The Demographics of Wealth, Federal
Reserve Bank of St. Louis, May 2015b, No. 2.
Emmons, William R.; and Noeth, Bryan J. “The
Economic and Financial Status of Older Americans: Trends and Prospects.” Chapter 1 in Financial
Capability and Asset Holding in Later Life: A Life
Course Perspective, Nancy Morrow-Howell and
Margaret S. Sherraden, editors, Oxford University
Press, pp. 3-26, 2015c.
The Demographics of Wealth 27
All of the essays in this series, along with their videos, can be read on the website of the
Center for Household Financial Stability at www.stlouisfed.org/hfs.