Published in the Journal of Economic Inequality, Volume 2, Number 2, August 2004
The Concept and Measurement of Asset Poverty:
Levels, Trends and Composition for the U. S., 1983-2001
Robert Haveman and Edward N. Wolff
In spite of the recession that began the new millennium, the United States prosperity of the second
one-half of the 1980s together with the booming economy of the 1990s created the impression that
American households have done well, particularly in terms of wealth acquisition. As we shall show, this is
decidedly not the case for many households. In this paper, we develop the concept of “asset poverty” as a
measure of economic hardship, distinct from and complementary to the more commonly used concept of
Asset poverty measures the extent to which American households have a stock of assets which is
sufficient to sustain a basic needs level of consumption during temporary hard times. We would note that
this concept of poverty, based on only the extent of asset-holdings, does not take into account the income
level of the household; the question is: Do the assets held by the household enable it to live at a minimum
level of consumption for a temporary period, should other source of income—e.g., earnings—be unavailable
during this period. As such, this measure complements standard measures of income poverty. We note that
income poverty measures identify as poor households whose annual income fails to support some socially
determined minimum level of consumption, abstracting from the household’s assets; our asset poverty
measure analogously identifies the poor as those households whose wealth or assets are insufficient to
enable them to live at this same minimum level, abstracting from the income available to the family.
We begin the paper with a brief discussion of income poverty measures, focusing on the official U.
S. income poverty indicator that serves as the basis for assessing the status of the nation’s least well-off
citizens. We then present our asset poverty concept, and the measures of this hardship indicator that we use
in the paper; we also describe the data sources that we use in our analysis. We report two indicators of the
level of asset poverty in the U. S. from 1983 to 2001. They reveal that in the face of the large growth in
overall assets in the U. S. over this period, the level of asset poverty has been increasing. In addition to
showing the level and trends in overall asset poverty in the U. S., we describe both the patterns of asset
poverty rates for various socioeconomic (e.g., race, age, schooling, family structure) groups over the 1983-
2001 period. We also compare the trends in asset poverty with those of income poverty, and report
differences in the prevalence and composition of asset poverty and income poverty.
I. The Concept of Poverty: Resources and Needs
Although poverty reduction is a universal goal among both nations and international organizations,
there is no commonly accepted way of identifying who is poor. Some argue for a multidimensional poverty
concept that reflects the many aspects of well-being. In this context, people deprived of social contacts (with
friends and families) are described as being socially isolated, and hence poor in this dimension. Similarly,
people living in squalid housing are viewed as “housing poor,” and people with health deficits as “health
poor.” Economists tend to prefer a concept of hardship that reflects the resources available to families, or
their “economic position” or “economic well-being,” somehow measured. Income is typically taken as the
measure of available resources, which is then compared to the income needs of the family. This economic
concept underlies the official United States poverty measure, and the proposed revision of it based on the
National Research Council (NRC) Panel Report.
Virtually all measures of economic poverty identify those families whose economic position
(defined in terms of command over resources) falls below some minimally acceptable level. There are two
requirements for such a measure—a precise definition of “economic resources” and a measure of the
minimum acceptable level of well being (or “needs”) in terms that are commensurate with the concept of
“resources” that is used.
Acceptable poverty measures also allow for differentiation according to household
size and composition. Given their economic basis, such measures exclude many factors that may affect
“utility” but are not captured by the concept of “resources” that is used.
Within this economic perspective, there are substantial differences regarding the specific economic
well-being indicators believed to best identify those whose economic position lies below some minimally
acceptable level. For example, the official U.S. poverty measure relies on the annual cash income of a
family, and compares this to some minimum annual income standard or “poverty line.” An alternative--and
equally legitimate--position is that the level of annual consumption better reflects a family’s access to
resources, or that a measure of a family’s income generating capacity is a more comprehensive indicator.
II. Official U. S. Poverty and Median Incomes: 1983-2001
The official definition of poverty in the United States has played a very special role in the
development of social policy in that country. A case can be--indeed, has been--made that the most important
contribution of the War on Poverty in the 1960s was the establishment of an official, national poverty line.
This official measure (including the recently proposed revision in it) has several distinct characteristics.
First, it is a measure of “income” poverty; the purpose is to identify those families that do not have sufficient
annual cash income (in some cases, including close substitutes to cash income such as Food Stamps) to meet
what is judged to be their annual basic needs. As such, it compares two numbers for each living unit—the
level of their annual income and the level of income that a unit of its size and composition requires in order
to secure a minimum level of consumption. By relying solely on annual income as the indicator of resources,
this measure ignores many potential sources of utility or welfare (e.g., social inclusion, or “security”) that
may be weakly tied to annual income flows. Second, distinct from the measures of relative poverty so
common in Europe (e.g., poverty lines defined relative to median income), the U. S. indicator is an absolute
poverty measure. As a result, decreases in inequality are reflected in reductions in poverty only if those
families with incomes below the absolute income cutoff are raised above it; a growing gap between those
with the least money income and the rest of society need not affect the official poverty rate.
The economic resources concept on which the U. S. measure rests (annual cash income) has been
subject to many criticisms.
Similarly, the arbitrary nature of the denominator of the poverty ratio--the
minimum income needs indicator--has also been criticized (Ruggles, 1990). Given its conceptual basis and
the crude empirical evidence on which the dollar cutoffs rest, the U.S. official poverty lines are essentially
arbitrary constructs. Finally, adjustments in the poverty line to account for different family sizes and
structures also rest on weak conceptual and empirical foundations.
In spite of criticisms of it, the official U. S. poverty measure provides a baseline against which to
judge estimates of asset poverty. Table 1 presents estimates of the percent of families in the U. S. that were
poor in those years over the 1983-2001 period for which we are able to study asset poverty, together with
estimates of median family income for these years.
Both the poverty and median income indicators of well being closely followed macroeconomic
conditions since the beginning of the 1980s. The official income poverty rate stood at over 12 percent at the
end of the severe recession of the early-1980s. During the several years of economic growth following that
recession, poverty fell steadily reaching a level of 10.3 percent by 1989. By 1992, family poverty had again
risen as the recession early in that decade also took its toll. However, in the prolonged expansion of the
1990s, official poverty again fell, to 10.8 percent in 1995, 10.0 percent in 1998, and to its lowest level since
the 1970s—9.2 percent---in 2001.
This pattern parallels changes in median family income over this period. Median family income
grew from $41,400 in 1983 to $47,200 in 1989, before falling to $45,200 during the recession of the early-
1990s. Persistent growth during the 1990s led to growth in median family income to its highest level during
the period of $51,400 in 2001.
III. Asset Poverty: Concepts and Data
With this background of trends in official poverty and median family income over the 1983-2001
period, we now turn to the definition and measurement of “asset poverty,” a concept that was first advanced
by Oliver and Shapiro (1997). We view families without a ‘safety-net cushion’ composed of asset holdings
to be in a vulnerable economic position; if alternative sources of income support such as the labor market or
public transfers are not available, only assets are left to avoid destitution. We define a household with
insufficient assets to enable it to meet basic needs for a period of time (three months) to be asset poor. This
measure does not consider the annual income position of the person, and hence serves to complement
indicators of poverty based on income flows alone.
A more demanding measure than either an income or an asset poverty measure would consider both
income and assets in defining poverty. Such a joint income/asset measure might label as poor households
income nor assets sufficient to sustain a minimum level of consumption for some period of
We present results from such a joint income/asset resource perspective below. Using this measure,
households are poor if they have neither annual income in excess of the poverty line nor assets in excess of
.25 of the poverty line.
A. Definitions and Conventions
We define a household or a person as being ‘asset poor’ if their access to wealth-type resources
insufficient to enable them to meet their basic needs
for some limited period of time. Clearly, this definition
leaves open a number of issues on which judgments are required.
What are ‘Basic Needs’?
We begin with the assumption that household needs can be met by access to financial resources,
such as income or real assets (e.g, owned homes), that can be valued in monetary units. Clearly, there is no
commonly accepted standard for measuring basic needs, as the variety of poverty thresholds used across
countries and research analyses varies widely. As indicated above, some measure the level of minimum
adequacy by referring to norms existing within a nation at a point in time, such as median income. Others
use professionally established minimum consumption standards. Our definition of asset poverty requires us
to make a choice of a standard for minimally acceptable needs.
What Period of Time?
The poverty thresholds indicate the level of basic resource needs for households of various sizes
measured over the course of a year; it is an annual “need-for-resources” concept. When this standard is
compared to the income flow over the course of a year, an income poverty measure is obtained. For our
purpose, the question is: How can these annual thresholds be used to indicate the adequacy of a stock of
wealth-type resources? How much of an asset stock should a household have in order to meet this annual
level of basic needs, were other resources not available. Over how long a period should asset holdings be
expected to provide a safety net cushion?
What is ‘Wealth’?
The third issue concerns the concept of wealth that we will employ in measuring asset poverty. A
number of issues must be considered, of which the following are representative. First, should housing
equity be included in the definition of assets; should families be expected to sell their homes in order to
obtain resources that are sufficient to provide this a protective cushion for periods of inadequate income?
Second, how should assets in the form of expected pensions or other forms of retirement saving be handled;
should families be expected to sacrifice these provisions for future security in order to support current
needs? Finally, in measuring available asset holdings, how should indebtedness be treated; are net asset
stocks the appropriate measure?
Two Measures of Asset Poverty
Based on these considerations, we propose and apply two measures of asset poverty.
based on the following choices. First, although there is no commonly accepted standard for the minimum
amount of financial resources that are required to meet needs, we use the family-size conditioned poverty
thresholds recently proposed by a National Academy of Science panel.
The panel recommended that the
thresholds should represent a dollar amount for food, clothing, shelter (including utilities), and a small
additional amount to allow for other common, everyday needs (e.g., household supplies, personal care, and
nonwork-related transportation). We employ a threshold developed for a reference family consisting of two
adults and two children using data from the U. S. Consumer Expenditure Survey, and then adjust this
threshold to reflect the needs of different family sizes and geographic differences in the cost of living. These
thresholds are based on the three-parameter equivalence scale for reflecting the needs of families of various
sizes and structures.
The 2001 threshold for a reference family of two adults with two children is $17,653,
which compares with the 2001 official income threshold of $17, 960.
Second, we need to stipulate a period of time over which assets should be expected to cushion
income losses. We propose the following standard: a family should have an asset cushion that would allow
them to meet their basic needs—the threshold poverty line--for three months (25 percent of a year), should
all other sources of support fail. Consistent with this standard, we compare the stock of asset holdings at a
point in time to 25 percent of the annual family-size specific poverty threshold. Hence, a four person family
would have asset needs equal to $4,413 (.25 x $17,653). With this standard, a family of four that held net
assets of less than $4,413 in 2001 would be declared "asset poor." Similarly, a one-person family with
assets below $2,303 or a six person family with assets below $6,229 would likewise fall below the basic
needs threshold. Again, note that no other source of resource support, such as earnings from work or other
forms of income are considered in measuring asset poverty.
Finally, we need to stipulate the definitions of “wealth” that we will use in constructing our asset
poverty measure. Our primary measure of assets is net worth, defined as the current value of all marketable
or fungible assets less the current value of debts. Net worth is thus the difference in value between total
marketable assets and total liabilities (or debt).
We take this net worth concept as our primary measure of
wealth as it reflects wealth as a store of value that can be liquidated in a short period of time, and therefore a
source of potential consumption. We judge that this concept best reflects the level of well-being associated
with a family's holdings; thus only assets that can be readily monetized are included.
We view this asset poverty measure as an indicator of the long-run economic security of families. A
portfolio of assets as complete as net worth is a point-in-time stock that reflects prior saving and other asset
accumulation decisions taken over a long period of time. The issue is, have these prior decisions provided a
sufficient cushion to enable a family to support itself for some period of time, should alternative sources of
support, such as earnings, fail? Relative to standard measures of income poverty that compare a single
year’s flow of income to a basic needs standard, this measure of asset poverty reflects the long-term ability
of a family to meet a minimum consumption standard.
Our second measure of wealth is based on a more restrictive definition of assets, namely liquid
assets, defined as cash or financial assets that can be easily monetized, excluding IRAs and pension assets.
This measure excludes the equity position in housing and real estate, the cash surrender value of defined
contribution pension plans, net equity in unincorporated businesses, and equity in trust funds. It also ignores
all forms of debt, including mortgage and consumer debt. This measure is appropriately thought of as an
‘emergency fund availability’ indicator of the ability of a family to ‘get by.’
Given these assumptions, our two standards of asset poverty are as follows:
• A family is asset poor if its net worth is less than 25 percent of the poverty line for families
of their size and composition—net worth < .25 family-specific poverty line.
• A family is asset poor if its liquid asset holdings are less than 25 percent of the poverty line
for families of their size and composition—liquid assets < .25 family-specific poverty line.
A Measure of Income-Asset Poverty
As we discussed above, we present evidence on the level of poverty when households are both
income poor and asset poor, a measure of joint income/asset poverty. In this measure, we combine the
income poverty measure with the asset poverty measure based on net worth. By this definition, a family is
joint income/asset poor if they have neither the income necessary to meet the income poverty standard nor
the assets necessary to meet the net worth asset poverty standard.
B. Data Sources
The data that we use in this study are the 1983, 1989, 1992, 1995, 1998, and 2001 Surveys of
Consumer Finances (SCF) conducted by the Federal Reserve Board. Each survey consists of a core
representative sample combined with a high-income supplement. The supplement is drawn from the Internal
Revenue Service's Statistics of Income data file. For the 1983 SCF, for example, an income cut-off of
$100,000 of adjusted gross income is used as the criterion for inclusion in the supplemental sample. The
advantage of the high-income supplement is that it provides a much "richer" sample of high income, and
therefore potentially very wealthy, families. The SCF also has the advantage of providing exceptional detail
on both assets and debt (several hundred questions are asked). For example, it asks each household to
identify both first and second mortgages and home equity credit lines, as well as the institutions granting the
loans and the interest rates charged. Credit card balances are asked for each credit card held by the family, as
well as interest charges.
IV. Asset Poverty in the U. S.: 1983-2001
Our overall estimates of the level of asset poverty in the U.S. are shown in Table 2 for the years
1983-2001. As expected, the more inclusive measure of assets, that based on net worth, yields the lower
poverty rates; the values range from a low of 22.4 percent in 1983 to 25.5 percent in 1998. Subsequent to the
recession of the early-1980s, net worth poverty rose by about 2 percentage points by 1989, then fell slightly
during the recession of the early-1990s, and again rose during the prolonged period of growth during the
decade of the 1990s. By this standard, the level of asset poverty in 1998 is the highest level recorded over
the 1983-2001 period. By 2001, net worth poverty had fallen to 24.5 percent..
When the liquid assets concept is used as the definition of assets, the asset poverty rate increases
substantially. By this measure, asset poverty is lowest in 1983 at 33 percent, and reaches a peak of nearly 44
percent in 1995. From the low levels during the 1980s, liquid asset poverty increased substantially in the
1990s. Even at the end of the 1990s growth period, liquid asset poverty stood at nearly 40 percent; the rate
fell slightly to about 37 percent by 2001.
For both measures, asset poverty at the end of the period equaled or exceeded both its 1983 level,
and its level during the recession of the early-1990s. Interestingly, the time pattern of asset poverty rates
does not closely reflect macroeconomic conditions, and does not parallel that of income poverty or median
These findings on the trend in asset poverty over the period after 1983 are consistent with, and
complementary to, estimates of the trend in wealth inequality reported by Wolff (2001) and D’Ambrosio
and Wolff (2001). Our results indicate that the increase in wealth inequality over this period affected not
only the wealthiest families, pulling them further from families with average levels of wealth, but also led to
decreases in wealth among those with the least assets. By focusing the level and prevalence of asset poverty,
we are able to identify the characteristics of those among the asset poor that have gained and lost position
during this period of increased asset poverty and inequality.
V. The Prevalence of Asset Poverty in 2001
Table 3 presents descriptive statistics on asset poverty for different demographic and labor market
groups in 1983, 1992, and the final year for which data are available, 2001. The population groupings that
we discuss include divisions by (a) race/ethnicity, (b) age of family head, (c) education of family head (d)
housing tenure status, and (e) marital status and presence of children.
The racial disparities in poverty rates indicated in the table are enormous, with the asset poverty
rates for minorities (Blacks/Hispanics) more than twice those for whites.
Using the net worth measure of
assets together with the 3-month cushion criterion, the asset poverty rates for whites range from 17 percent
to 19 percent over the 1983-2001 period; the range for Blacks/Hispanics is from 43 to 47 percent. Using the
liquid asset measure of assets, about 30 percent of white households are in asset poverty, while about 62
percent of Black/Hispanic households have inadequate liquid financial reserves to tide them over a 3-month
period at a level of living equal to the poverty line.
On the basis of the life cycle model of saving behavior (Modigliani and Brumberg, 1954), young
people borrow to support consumption while investing in human capital while those in their years of high
earnings save for retirement years. Consistent with this framework, we would expect high asset poverty rates
for families headed by a young person and low asset poverty rates for those at or beyond their peak earnings
years. Table 3 also shows the 2001 asset poverty rates for households headed by various age groups. The
pattern seen there is consistent with the life cycle framework. Irrespective of the measure used, households
headed by people less than 25 years of age have remarkably high asset poverty rates—for example, more
than 72 percent do not have net worth or liquid assets sufficient to support poverty line consumption for a
three month period. Both of these asset poverty rates fall monotonically by age. For households headed by a
person aged 35 to 49, net worth poverty rates are about one-third of the rates for the young households;
liquid asset poverty rates for the prime age group are about one-half of those for the youngest group. Those
aged 62 or more have the lowest asset poverty rates using either criterion; an average of about 10 percent
over the entire period for the net worth measure and about 25 percent for the liquid asset measure.
As with age, the asset poverty rates fall monotonically by the education of the head. Asset poverty
rates for households headed by a person with four or more years of college are about one-fourth of those of
families with a head who has not completed a high school degree. For example, while 60 percent of families
headed by a person with less than a high school degree are in liquid asset poverty, about 15 percent of the
college graduates have insufficient liquid assets to enable them to meet the three months of poverty line
The pattern of 2001 asset poverty rates by housing tenure shown in Table 3 is revealing. For
homeowners, the net worth asset measure that includes the value of home equity indicates an asset poverty
rate of about 6 percent, compared to rates of over 60 percent for renters. While the rates between these
tenure categories become closer when the liquid asset measure that excludes home equity is used, the asset
poverty rates of renters remain more than double those of homeowners. Indeed, nearly two-thirds of renters
have insufficient liquid assets to provide them the three-month cushion of poverty line consumption. It
seems clear that homeownership implies more than home equity, and is associated with the ownership of a
wide range of financial assets.
Table 3 also indicates that asset poverty rates in 2001 also vary substantially by family type. The
lowest asset poverty rates are observed among married couple families aged 65 years or older. Using the
three-month cushion standard, asset poverty rates for elderly married couples range from 5 percent when
home equity is included in the asset definition to 16 percent using the liquid asset definition. The rates for
two-parent families with children range from about 22 percent to 42 percent across the two asset poverty
measures, while the rates for families with children and a female single-parent range from 56 percent to 71
percent. This family type has among the highest asset poverty rates shown in the table.
This cross tabulation of poverty rates by subgroups of families does not indicate the independent
relationship of the racial, age, education, home owner, and family type characteristics to the probability of
being in poverty by either of these measures. To estimate the independent effect of these socio-economic
characteristics on the probability of being poor by any measure we fit a probit model to the observations in
each year. We defined the dependent variable as being in poverty (using several poverty measures, including
income poverty, asset poverty, joint income/asset poverty) and the characteristics of the families serving as
Table 4 shows the probit model fit to net worth poverty status in 1983 and 2001 using the individual
characteristics in Table 3 as right hand side variables. The excluded characteristics are generally those with
the lowest asset poverty rate (for example, being a home owner in the case of housing tenure status). The
probit results indicate that being Black/Hispanic (relative to being White or other) has a statistically
significant positive independent effect on the probability of being net worth poor. Second, the coefficients
by age group are all positive and generally decline with age (relative to the oldest age group, which is
excluded). They are significant for all age groups in 2001 and for all except age group 50-61 in 1983. Third,
the coefficients are all positive but decline with level of education (relative to the excluded group, college
graduates) except for college 1-3 in 1983. All of the coefficients are significant at the one percent level.
Fourth, the coefficient for being a renter (relative to being a home owner) is very high in the two years and
significant at the one percent level.
Finally, in the case of family type, there are some notable changes between 1983 and 2001. It should
be noted that since asset poverty for the base case, single males under the age of 65, remains virtually
unchanged between 1983 and 2001 (see Table 3), the changes in coefficients reflect changes in the asset
poverty propensity for these groups rather than for the base case. In 1983, the coefficients for all family
types are positive and statistically significant relative to single males under the age of 65. In both of the
years, the largest coefficient is that for female headed families with children under age 65. While married
couple families have significantly higher asset poverty rates than single males in 1983, by 2001 this
difference has disappeared. Apparently married couple families have reduced their indebtedness or increased
their savings more than single males over the 1983-2001 period. Interestingly, while the coefficient on
married and aged 65 or over and on older single female were positive and significant in 1983, by 2001 these
coefficients became negative and in the case of older married couples negative and statistically significant.
VI. Trends in Asset Poverty: 1983-2001
Table 3 also shows the percentage point change in asset poverty rates between 1983 and 2001, a
eighteen-year period. Note that the first year, 1983 is a recession year, while 2001 is at the end of a
prolonged recovery, with the economy at full employment. Given these different macroeconomic conditions,
it is expected that the rates of asset poverty would have fallen over this period. For both the net worth and
liquid asset poverty measures, the time pattern of change fails to meet our expectation. Increases in asset
poverty of 2.1 percentage points (9 percent) and 4.3 percentage points (13 percent) are recorded for these
definitions. In spite of the enormous increase in financial and pension wealth holdings over this period, 25
percent of the nation remains in net worth poverty and 38 percent is in liquid asset poverty in 2001.
While the overall patterns of asset poverty also describe the levels and trends for the white
population, the situation is quite different for Blacks and Hispanics. For Blacks/Hispanics, decreases in
asset poverty rates are observed for both asset poverty measures. The decreases are small however, and
range from 1 percent to 3 percent across the two measures.
Irrespective of definition, households headed by people less than age 50 experienced the largest
increases in asset poverty over this period. Using our two measure of asset poverty, the increases ranged
from 29-30 percent for the youngest group, from 15-22 percent for the 25-34 year olds, and by 27 percent
for the 35-49 year olds.
Across education groups, all of the groups except those with some college education experienced an
increase in net worth poverty over this 18-year period, with substantial increases experienced by the two
lowest schooling groups—35 percent for those with less than a high school degree, and a 33 percent increase
for high school graduates. Using the measure based on the liquid asset measure of wealth, a very large
increase in asset poverty over the period is recorded for all of the schooling groups. However, the increase
was substantially smaller for the group with some college than for the remaining groups. Surprisingly, the
increase in liquid asset poverty is exceptionally large for families headed by a college graduate; asset
poverty by this measure grew by one-third over the period, from 12 percent to 16 percent. Perhaps such high
education families are increasingly willing to rely on their ability to obtain loans and credit to provide short-
Asset poverty for renters grew substantially over the period, using both measures. Net worth
poverty rose by 8.8 percentage points (16 percent), and liquid asset poverty increased from 52 percent in
1983 to 64 percent in 2001, or by 24 percent over the period. In contrast, asset poverty for homeowners rose
by 2.2 percentage points, or by about 60 percent albeit from a very low base of 4 percent in 1983. The
ostensible reason is the very high growth in mortgage debt as a percent of house value, which almost
doubled over the period from 1983 to 2001. When the net asset value of the own home is excluded from the
asset base (the liquid asset poverty measure), the rate of asset poverty for homeowners increased by 2.1
percentage points—less than 10 percent.
Among families headed by a person less than 65 years, the largest increases in asset poverty are
recorded for childless married couples—a near doubling using the net worth poverty measure and an
increase of one-third using the liquid asset poverty measure. Nonelderly female headed families with
children experienced the lowest percentage increases in asset poverty—ranging from 16 percent for the net
worth measure to 12 percent for the liquid asset measure. Among families headed by a person aged 65 years
or more, the change in asset poverty levels varies substantially by type. Female headed families in this
category—primarily widows—experienced modest increases in asset poverty. However, for both aged
married couples and older single male households, decreases in asset poverty are recorded for both
measures. For older single male households, the reductions in asset poverty range from 25 to 30 percent.
In sum, then, overall asset poverty grew modestly over this 15-year period from 1983 to 2001.
Among population subgroups, however, the patterns of changing poverty prevalence vary substantially—
large increases in the rate of asset poverty are recorded for:
• whites relative to racial minorities,
• families headed by a person aged less than 50 years relative to those headed by an older person,
• families headed by a person with little schooling, relative to those with some college,
• renters relative to homeowners, and
• families headed by a person less than 65 years (irrespective of marital status and the presence of
children), relative to families headed by a person 65 years or older.
VII. Sub-period Asset Poverty Trends—1983-1992 and 1992-2001
The trends discussed in the previous section and shown in Table 3 summarize asset poverty
developments over the entire period from 1983 to 2001—from a distant recession year to a recent full
employment year. These long-period trends can be decomposed into trends over two separate periods—
from the recession year 1983 to another recession year, 1992, and from that year to the beginning of a
recession after the unprecedented growth experienced during the 1990s. These patterns are also shown in
For the entire population, the bulk of the increase in asset poverty came in the earlier of the two
periods; the period from 1992 to 2001 saw virtually no increase in overall asset poverty, irrespective of the
measure. This pattern also holds for white families; Black/Hispanic poverty, however, increased in the latter
period using the net worth measure, but decreased using the liquid asset measure.
For all of the age groups, the large increases in asset poverty occurred during the first period. For
families headed by a person over the age of 50, asset poverty either decreased in the latter period, or
increased only slightly. The subperiod pattern seen for the entire population is also observed for most of the
schooling groups; asset poverty grew substantially during the early period, with especially large increases
recorded for families headed by a college graduate. This contrasts with the pattern during the 1990s, during
which time asset poverty declined for college graduates—by over 20 percent using the net worth measure
and by 15 percent using the liquid asset measure. The fabled run-up in financial asset holdings for those with
education and schooling did improve the economic status of the lowest wealth holders of this group, but the
reductions in asset poverty for those with college degrees seem small by comparison to the overall gains by
For renters, asset poverty levels increased substantially for both of the measures during both sub-
periods. However, for homeowners, asset poverty using the liquid asset definitions fell during the latter
period. Unexpectedly, asset poverty increased for homeowners during the latter period, using the net worth
measure, which includes the equity value in owned homes; liquid asset poverty for homeowners decreased
during the latter period.
The patterns for the various family types are complex. Consider, first, families headed by a person
less than age 65. During the early period from 1983 to 1992, asset poverty increased by both measures for
all of the family types. However, during the most recent decade, from 1992 to 2001, asset poverty rose for
both intact and female-headed families with children by both measures. However, for families without
children, asset poverty fell during the recent period by both measures. Some surprising twists are also seen
for the families headed by a person aged 65 years or more. For older single females, net worth poverty fell
during the early period, but increased over this period using the liquid asset measure. During the later
period, asset poverty increased for both of the measues. For the other families headed by an older person—
married couples and single men—asset poverty fell over both of the subperiods using both measures.
VIII. Trends in Asset Poverty vs. Income Poverty
An interesting question concerns the difference in the trends of asset poverty relative to the official
income poverty measure. Table 5 presents the pattern of income poverty in the U. S. for the same three
years—1983, 1992, and 2001—of asset poverty tracked in Table 3. While overall asset poverty rose by
more than 10 percent according to both measures, the rate of income poverty fell from 14.7 percent to 13.2
percent, or by 11 percent.
For nearly all of the groups shown in Table 5, income poverty rose between 1983 and 1992, in some
cases substantially; the primary exceptions are those living in families headed by a person aged 50-61, intact
nonelderly families with children, and elderly married couples and single males. Much the same pattern
holds for both of the asset poverty measures, with only a few subgroups recording decreases. The trends in
asset and income poverty during this early period are very similar.
It is during the latter period—1992-2001—that substantial differences between the income and asset
poverty measures appear. During this recent period, overall asset poverty appears to have increased slightly,
while income poverty fell substantially from 16 percent to 13.2 percent, or by 18 percent. Of the 19
subgroups shown on Tables 3 and 5, net worth poverty rose for 14 of them over this period; liquid asset
poverty rose for 12 of the 19 subgroups. However, over this same time period, income poverty fell for 16 of
the 19 subgroups. Apparently, the gains in income experienced by the income poor during the economic
growth period of the 1990s did not find its way into the holding of assets by the asset poor. This pattern is
consistent with evidence on the low rates of saving by the poor, even when income is increasing.
IX. Toward a Joint Income/Asset Poverty Indicator
Given the two resource criteria that we have used to analyze the prevalence of poverty—annual
income and assets—it is possible to join the two measures and estimate the share of the nation’s families
that is both income poor and asset poor, and their composition. In Table 6, we present this comparison for
both 1983 and 2001, using the revised poverty lines and the net worth poverty measure of assets.
In 1983, when 14.7 percent of U. S. families had income below the poverty line, and 22.4 percent
were asset poor; 7.6 of the nation’s families were both asset and income poor. These joint poverty families
include 52 percent of the families who were income poor, and 34 percent of the families who were asset
poor. Between 1983 and 2001, the joint poverty rate increased from 7.6 to 7.9, or by about 4 percent,
suggesting that the upward trend in the asset poverty rate over time dominated the downward trend in the
income poverty rate over this period. In 2001, 60 percent of the families that were income poor were in joint
poverty, and 32 percent of asset poor families were poor by the joint asset/income poverty measure. Over
the 18 years, then, an increasing share of the income poor families were also asset poor, while among the
asset poor, a smaller proportion was also income poor.
Certain groups of the population have especially high rates of joint asset/income poverty, including
Blacks/Hispanics, those living in a family headed by a person aged less than 25 years, those in a family
headed by a person with less than a high school degree, renters, and female-headed families with children.
All of these groups have a rate of joint poverty in excess of 15 percent in both 1983 and 2001.
exception of the single unmarried mothers, all of these groups experienced large increases in the rate of joint
poverty between 1983 and 2001.
These patterns of joint income/asset poverty prevalence are also reflected in the composition of the
poor population by the various measures, as shown in Table 7. In 1983, when Blacks/Hispanics families
were 16 percent of all U. S. families, they comprised about 35 percent of all income or asset poor families,
but 47 percent of the families in joint poverty. Indeed, by 2001, minorities made up more than half (54
percent) of families classified as both income and asset poor. In 1983, households under the age of 35
constituted 31 percent of the total population but 48 percent of those in joint poverty. In 2001, their share of
total households fell to 23 percent while their share of families in joint poverty remained high, at 42 percent.
In 1983, families headed by someone with less than a high school degree comprised 29 percent of all
families but 58 percent of those in joint poverty. Between 1983 and 2001, their share of total households
declined by 11 percentage points, to 18 percent, while their proportion of families in joint poverty felly by
only 7 percentage points, to 51 percent. Renters made up about a third of all families in both 1983 and 2001
but close to 95 percent of those who were both asset and income poor. Those living in a family headed by a
female (both with children and without children) comprised 18 percent of all households in 1983 and 16
percent in 2001 but 44 percent of those in joint poverty in 1983 and 39 percent in 2001. Clearly, the
composition of the poor as determined by this joint poverty criterion is more heavily weighted toward these
vulnerable groups than is either the income or asset poverty measures.
X. Summary and Conclusions
The patterns of asset poverty over the period from 1983-2001 are discouraging in that very high
rates of asset poverty for the U. S. population are revealed, irrespective of the measure used. In 2001, one
fourth of American families have insufficient net worth to enable them to get by for 3 months at a poverty
line level of living, and over one third have insufficient liquid assets to support poverty level living for a 3
These high levels of asset poverty for the entire population disguise even higher rates for various
groups. Using the net worth poverty standard, the following indicates asset poverty rates in 2001 for some
of the groups most disadvantaged in terms of wealth holdings:
• Blacks/Hispanics 62 percent
• Head aged less than 25 years 72 percent
• Head aged 25-34 years 52 percent
• Head with less than a high school degree 60 percent
• Renters 64 percent
• Nonaged Female heads with children 71 percent
The growth in asset poverty over time is also discouraging. For both of our measures, the
prevalence of asset poverty grew from 1983 to 2001; an increase of 9 percent in net worth poverty, and an
increase of 13 percent for liquid asset poverty.
The patterns of growth in asset poverty over the two sub-periods—1983-1992 and 1992-2001—are
also revealing. For the population as a whole, asset poverty increased substantially from 1983 to 1992, even
though both were recession years. However, during the years of rapid income growth from 1992-2001, when
prosperity seemed to affect all groups, asset poverty did not fall and if anything edged up slightly. This is in
contrast to the substantial decrease in income poverty over this period. Apparently those least well off in the
U. S. economy used their increased incomes during this period for consumption rather than asset
Caner, Asena and Edward N. Wolff. 2002. "Asset Poverty in the United States, 1984-1999: Evidence from
the Panel Study of Income Dynamics." Working Paper No. 356. Levy Economics Institute of Bard College:
Citro, Constance F., and Robert T. Michael (eds.). 1995. Measuring Poverty: A New Approach
Washington, DC: National Academy Press.
D’Ambrosio, Conchita, and Edward Wolff. 2001. “Is Wealth Becoming More Polarized in the United
States?” Working Paper No. 330. Levy Economics Institute of Bard College: Annandale-on-Hudson, N.Y.
Haveman, Robert, and Melissa Mullikin. 2001. “Alternative Measures of National Poverty: Perspectives and
Assessment.” In Ethics, Poverty and Inequality and Reform in Social Security
. ed. Erik Schokkaert. London:
Ashgate Publishing Ltd.
Lerman, Donald L. and James J. Mikesell. 1988. "Impacts of Adding Net Worth to the Poverty Definition."
Eastern Economic Journal
, XIV:4, pp. 357-70.
Modigliani, Franco, and R. Brumberg. 1954. “Utility Analysis and the Consumption Function: An
Interpretation of Cross-section Data.” In Post-Keynesian Economics
, ed. K. K. Kurihara, New Brunswick,
NJ: Rutgers University Press.
Moon, Marilyn. 1977. The Measurement of Economic Welfare: Applications to the Aged.
Oliver, Melvin L., and Thomas M. Shapiro. 1997. Black Wealth/White Wealth. New York: Routledge Press.
Ruggles, Patricia. 1990. Drawing the Line: Alternative Poverty Measures and Their Implications for Public
Policy. Washington, D.C.: Urban Institute Press.
Rendall, Michael S., and Alden Speare, Jr. 1993. “Comparing Economic Well-Being among Elderly
Americans.” Review of Income and Wealth.
39, 1 (March) 1-21.
Sen, Amartya. 1992. Inequality Reexamined
. Cambridge, MA: Russell Sage Foundation and Harvard
U.S. Census Bureau. 1999. "Experimental Poverty Measures 1990 to 1997," Current Population Reports,
Weisbrod, Burton, and W. Lee Hansen. 1968. “An Income-Net Worth Approach to Measuring Economic
Welfare,” American Economic Review
, 58, 5 (December) 1315-29.
Wolff, Edward N. 2001. "Recent Trends in Wealth Ownership, 1983-1998", in Thomas M. Shapiro and
Edward N. Wolff eds. Assets for the Poor: The Benefits of Spreading Asset Ownership
. New York: Russell
Ziliak, James P. 2001. “Income Transfers and Assets of the Poor.”Institute for Research on Poverty,
University of Wisconsin-Madison, Discussion Paper 1233-01.
Table 1. Official Income Poverty Rates for Families and Median Family
Official Poverty Rate Median Family Income
Year For Families (percent) ($ thousands 2001)
1983 12.3 41.4
1989 10.3 47.2
1992 11.9 45.2
1995 10.8 46.8
1998 10.0 50.7
2001 9.2 51.4
Source: U.S. Bureau, website:
a. Based on the CPI-U-RS deflator.
Table 2. Asset Poverty Rates by Definition and Year for Households, 1983-2001
(figures are in percent)
Net Worth < Liquid Assets <
Year .25 Poverty Line .25 Poverty Line
1983 22.4 33.2
1989 24.7 36.4
1992 24.0 37.5
1995 25.3 43.8
1998 25.5 39.7
2001 24.5 37.5
Source: Authors' calculations from the 1983, 1989, 1992, 1995, 1998, and 2001 SCF.
Table 3. Asset Poverty Rates for Households by Demographic Group, 1983-2001
(figures are in percent)
Net Worth < .25 Poverty Line Liquid Assets < .25 Poverty Line
Grouping Category 1983 1992 2001 Change, 83-01 1983 1992 2001 Change, 83-01
All Households 22.4 24.0 24.5 2.1 33.2 37.5 37.5 4.3
Race Whites 17.1 19.1 18.0 0.9
26.9 29.8 30.4 3.5
Blacks/Hispanics 47.4 43.2 46.7 -0.7
63.8 66.8 62.1 -1.7
Age Less than 25 55.6 66.9 72.1 16.5 56.1 70.3 72.3 16.2
25-34 36.3 41.8 44.3 8.0 44.8 49.4 51.5 6.7
35-49 17.7 21.7 22.5 4.8 30.9 39.2 39.3 8.4
50-61 13.8 13.9 13.7 -0.1 26.2 26.2 28.7 2.5
62 or older 9.9 10.6 10.8 0.9 22.5 26.7 24.0 1.5
Education Less than HS Grad. 29.8 37.6 40.1 10.3 50.0 62.8 60.1 10.1
HS Graduate 20.9 26.4 27.8 6.9
33.6 40.9 45.8 12.2
College 1-3 25.5 20.8 25.4 -0.1
31.1 33.7 36.8 5.7
College Graduate 11.3 14.0 11.0 -0.3
11.8 18.5 15.8 4.0
Tenure Home owner 3.6 4.7 5.8 2.2
22.6 25.4 24.7 2.1
Renter 54.8 58.4 63.6 8.8
51.7 58.9 64.2 12.5
Family LT 65 years, Married, 21.6 21.6 22.3 0.7
37.6 37.9 42.2 4.6
Type with Children
LT 65 years, Married, 12.9 20.4 18.9 6.0
19.9 27.6 26.7 6.8
LT 65 years, Female Head 48.1 49.7 55.8 7.7
63.4 66.5 71.2 7.8
LT 65 years, Male Head 37.8 33.5 35.4 -2.4
38.5 43.3 41.6 3.1
65 or older, Married 5.5 4.9 4.8 -0.7
17.4 16.0 15.6 -1.8
65 or older, Female Head 15.3 13.8 18.3 3.0
29.0 32.8 33.5 4.5
65 or older, Male Head 21.1 21.2 14.6 -6.5
40.2 36.1 30.2 -10.0
Memo: Percent of Asset Poor 69.1 75.0 71.7 2.6
46.6 48.0 46.8 0.2
with Zero or Negative Net Worth
Source: Authors' calculations from the 1983, 1992, and 2001 SCF.
Table 4. Probit Estimates for Net Worth Poverty, 1983 and 2001
(Standard errors are in parentheses)
Variable 1983 2001
Intercept -3.501 *** -2.690 ***
Black or Hispanic 0.658 *** 0.316 ***
Age LT 25 0.588 *** 1.067 ***
Ages 25-34 0.807 *** 0.879 ***
Ages 35-49 0.362 * 0.404 ***
Ages 50-61 0.230 0.147 *
Less Than HS Graduate 0.623 *** 1.208 ***
HS Graduate 0.409 *** 0.664 ***
College 1-3 0.421 *** 0.476 ***
Renter 1.794 *** 1.713 ***
Married with Children, Under 65 0.921 *** 0.042
Married and Childless, Under 65 0.660 *** 0.003
Female Head with Children, Under 65 1.084 *** 0.330 ***
Female Head, Childless, Under 65 0.742 *** 0.083 *
Married, 65 or older 0.517 ** -0.452 ***
Female Head, 65 or over 0.784 *** -0.068
Male Head, 65 or over 1.039 ** 0.026
Number of Observations 4262 22210
Wald 1715.1 *** 6452.9 ***
Chi Square 1090.5 *** 10783.4 ***
Source: Authors' calculations from the 1983 and 2001 SCF. Exluded groups: (1) Whites and
other races; (2) Age group 62 and over; (3) College graduates; (4) Home owners; and
(5) Male heads under age 65. Key:
*** Significant at 1% level. ** Significant at 5% level. * Significant at 10% level
Table 5. Income Poverty Rates for Households by Demographic Group, 1983-2001
(figures are in percent)
Grouping Category 1983 1992 2001 Change, 1983-2001
All Households 14.7 16.0 13.2 -1.6
Race Whites 10.9 11.0 8.6 -2.3
Blacks/Hispanics 32.8 34.6 27.5 -5.3
Age Less than 25 26.7 43.1 33.6 6.9
25-34 13.1 16.8 13.6 0.5
35-49 11.8 12.3 10.5 -1.3
50-61 12.0 9.8 10.9 -1.1
62 or older 17.8 18.1 13.5 -4.3
Education Less than HS Grad. 29.5 36.9 35.6 6.1
HS Graduate 11.8 15.3 12.1 0.3
College 1-3 10.0 12.4 9.6 -0.4
College Graduate 3.1 4.0 3.2 0.1
Tenure Home owner 9.1 9.3 6.7 -2.5
Renter 24.5 27.8 26.8 2.3
Family LT 65 years, Married, 9.7 9.1 10.0 0.3
Type with Children
LT 65 years, Married, 4.9 6.7 4.8 -0.1
LT 65 years, Female Head 39.8 42.8 38.2 -1.6
65 or older, Married 11.6 6.8 7.1 -4.5
65 or older, Female Head 28.4 29.5 24.4 -4.0
65 or older, Male Head 31.0 15.0 11.7 -19.2
Memo: Percent of Income Poor 37.9 43.2 42.5 4.6
with Zero or Negative Net Worth
Source: Authors' calculations from the 1983, 1992, and 2001 SCF. Income poverty is based on the
NAS 3-parameter scale (see the text for details).
Table 6. Joint Income/Asset Poverty Rates for Households by Demographic Group, 1983-2001
(figures are in percent)
Asset Poor & Asset Poor Income Poor Asset Poor & Asset Poor Income Poor
Grouping Category Income Poor Only Only Income Poor Only Only
All Households 7.6 14.8 7.2 7.9 16.6 5.3
Race Whites 4.5 12.6 6.5
4.0 14.0 4.6
Blacks/Hispanics 21.7 25.6 11.1
32.7 14.0 -5.2
Age Less than 25 18.7 36.9 7.9 26.9 45.2 6.6
25-34 9.5 26.7 3.6 10.5 33.8 3.1
35-49 5.8 11.8 6.0 7.0 15.5 3.5
50-61 6.0 7.8 6.0 4.9 8.8 5.9
62 or older 5.2 4.6 12.6 5.1 5.7 8.4
Education Less than HS Grad. 15.1 14.7 14.4 22.2 17.9 13.4
HS Graduate 6.0 15.0 5.9
7.2 20.6 4.9
College 1-3 5.3 20.3 4.7
5.1 20.3 4.5
College Graduate 1.8 9.6 1.4
1.9 9.1 1.3
Tenure Home owner 0.4 3.2 8.7
0.6 5.2 6.0
Renter 20.0 34.8 4.5
23.0 40.6 3.8
Family LT 65 years, Married, 5.1 16.5 4.6
6.5 15.8 3.5
Type with Children
LT 65 years, Married, 2.0 10.9 2.9
2.8 16.1 2.0
LT 65 years, Female
Head 28.7 19.4 11.0
27.9 27.9 10.2
65 or older, Married 2.0 3.6 9.6
2.6 2.2 4.5
65 or older, Female Head 9.8 5.5 18.6
9.6 8.7 14.8
65 or older, Male Head 11.8 9.2 19.2
6.5 -6.5 5.3
Source: Authors' calculations from the 1983 and 2001 SCF. Income poverty is based on the
NAS 3-parameter scale (see the text for details). Asset Poverty is based on: Net Worth < .25 Poverty Line
Table 7. Composition of Joint Income/Asset Poor Households by Demographic Group, 1983 and 2001
(figures are in percent)
Percent Asset & Asset Income All All Percent Asset & Asset Income All All
of all Income Poor Poor Asset Income of all Income Poor Poor Asset Income
Grouping Category Households Poor Only Only Poor Poor Households Poor Only Only Poor Poor
All Households 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Race Whites 80.9 47.7 69.2 73.1 61.9 60.0
76.2 38.9 64.1 66.4 56.0 49.9
Blacks/Hispanics 16.3 46.6 28.2 25.2 34.4 36.2
13.0 54.1 11.0 -12.8 24.8 27.2
Age Less than 25 8.0 19.8 20.0 8.9 19.9 14.5 5.6 19.3 15.3 7.1 16.6 14.4
25-34 22.6 28.4 40.8 11.3 36.6 20.1 17.2 22.9 35.1 10.2 31.2 17.8
35-49 27.6 21.1 22.1 23.2 21.8 22.1 33.6 29.9 31.3 22.3 30.8 26.8
50-61 18.4 14.5 9.7 15.3 11.3 14.9 19.0 11.9 10.0 21.3 10.6 15.7
62 or older 23.5 16.2 7.3 41.2 10.3 28.4 24.6 16.0 8.4 39.1 10.8 25.3
Education Less than HS Grad. 29.0 57.7 28.8 58.3 38.6 58.0 18.1 51.0 19.6 46.1 29.7 49.0
HS Graduate 30.2 23.7 30.6 24.7 28.2 24.2
29.6 27.1 36.7 27.5 33.6 27.2
College 1-3 19.6 13.7 26.9 13.0 22.4 13.3
22.6 14.6 27.6 19.2 23.4 16.5
College Graduate 21.2 4.9 13.7 4.1 10.7 4.5
29.6 7.3 16.2 7.2 13.3 7.2
Tenure Home owner 63.4 3.5 13.7 77.1 10.3 39.2
67.7 5.6 21.1 76.9 16.1 34.2
Renter 36.6 96.5 86.3 22.9 89.7 60.8
32.3 94.4 78.9 23.1 83.9 65.8
Family LT 65 years, Married, 31.0 20.9 34.6 20.0 30.0 20.5
26.9 22.1 25.6 18.0 24.4 20.4
Type with Children
LT 65 years, Married, 20.0 5.4 14.8 8.0 11.6 6.6
22.3 7.8 21.6 8.5 17.2 8.1
LT 65 years, Female 8.7 32.9 11.4 13.4 18.7 23.4
8.5 30.2 14.3 16.5 19.4 24.7
Head with Children
65 or older, Married 9.8 2.5 2.4 13.1 2.4 7.7
11.2 3.6 1.5 9.6 2.2 6.0
65 or older, Female Head 9.1 11.8 3.4 23.6 6.2 17.5
7.2 8.8 3.8 20.1 5.4 13.3
65 or older, Male Head 0.4 0.6 0.2 1.0 0.3 0.8
2.7 2.3 1.3 2.7 1.6 2.5
Source: Authors' calculations from the 1983 and 2001 SCF. Income poverty is based on the NAS 3-parameter scale
(see the text for details). The categories do not necessarily sum to 100.0 because of the exclusion of certain categories.
Definition of Asset Concepts
Net worth = the gross value of owner-occupied housing
+ other real estate owned by the household
+ cash and demand deposits
+ time and savings deposits
+ certificates of deposit and money market accounts
+ government, corporate, and foreign bonds, and other financial securities
+ the cash surrender value of life insurance plans
+ the cash surrender value of defined contribution pension plans, incl. IRAs, Keogh ,401(k)s
+ corporate stock and mutual funds
+ net equity in unincorporated businesses
+ equity in trust funds
- mortgage debt
- consumer debt, including auto loans and credit card balances
- other debt.
Liquid = cash and demand deposits
+ time and savings deposits
+ certificates of deposit, and money market accounts
+ government, corporate, and foreign bonds, and other financial securities
+ the cash surrender value of life insurance plans
+ corporate stock and mutual funds.
Robert Haveman is John Bascom Emeritus Professor of Economics and Public Affairs at the University of
Wisconsin-Madison, and Adjunct Professor at the Research School of Social Sciences of Australian National
University. Edward Wolff is Professor of Economics at New York University and a Senior Scholar at the Jerome Levy
Economics Institute of Bard College. The authors would like to gratefully acknowledge the financial support of the
Russell Sage Foundation and the Ford Foundation.
This proposed revision is described in the report of the Panel on Poverty and Family Assistance, which was
appointed by the Committee on National Statistics of the National Research Council of the National Academy of
Sciences (Citro and Michael, 1995).
Sen (1992) considered the needs standard (or poverty line) to have “some absolute justification of its own,” it
being a level below which “one cannot participate adequately in communal activities, or be free of pubic shame from
failure to satisfy conventions” (p. 167).
Haveman and Mullikin (2001) discuss the advantages and disadvantages of these alternatives.
The most fundamental criticisms of the official measure focus on the basic social objective on which it rests;
cash income may not be the most salient indicator of well-being or position. Similarly, in assessing poverty trends over
time, perhaps the general trend in the overall level of living should be taken into account, as is the case with relative
measures of poverty. Aside from taking exception to the social objective that underlies the official measure, most other
criticisms of it focus on the adequacy of the annual income measure of “economic resources.” While the current cash
income numerator of the poverty ratio may reflect the extent to which the family has cash income available to meet its
immediate needs, it indicates little about the level of consumption spending potentially available to the family. For
many families, annual income fluctuates substantially over time. Unemployment, layoffs, the decision to undertake
mid-career training or to change jobs, health considerations, and especially income flows from farming and self-
employment may all cause the money income of a household to change substantially from one year to the next. Even as
an indicator of a family’s ability to meet its immediate needs, the current cash income measure is flawed--it reflects
neither the recipient value of in-kind transfers (e.g., Food Stamps and Medicaid, both of which are major programs in
the United States supporting the economic well being of low income families), nor the taxes for which the family is
liable. Although the Earned Income Tax Credit (EITC), a component of the tax system, has expanded into a major
form of income support for the low income working population, the refundable payments from the credit are viewed as
negative taxes and hence not included in the definition of income used in the official poverty measure. Similarly,
whereas current cash income--and hence the official poverty measure--reflects financial flows in the form of interest
and dividends from the assets held by individuals, the assets themselves are not counted, nor is the value of leisure (or
voluntary nonwork) time reflected in the measure. (This is less the case for the NRC-proposed revision to the official
poverty measure, as it attempts to account for some in-kind benefits in assessing the relationship of resources to needs.)
The official poverty measure is also silent on the differences in the implicit value that families place on income from
various sources. Income from public transfers, market work, and returns on financial assets are treated as being
equivalent in contributing to the family’s well-being.
One reviewer suggested we use the expression "low assets" rather than that "asset poverty" to indicate asset
deprivation. Since our definition of asset poverty is identical in concept to definitions of income poverty, we do not
think it appropriate to use the term “low assets” rather than “asset poverty.”. Otherwise the reviewer should also object
to the term “income poverty” and prefer that of "low income".
Alternatively, one might define as poor households those whose income over a period of time plus their
assets were not sufficient to maintain the required level of consumption for the stipulated period. This would be a less
demanding measure than the joint income/asset poverty measure.
Two strands of economics literature have studied the relationship between the resource flow (income) and
resource stock (wealth) dimensions of economic well-being. In an early contribution, Weisbrod and Hansen (1968)
proposed an ‘income-net worth’ measure of economic well-being. In this framework, well-being was measured by
adding to annual income the annual value of asset holdings when annuitized over the expected remaining years of
lifetime. They presented estimates of the level and distribution of this value, which indicated substantially higher levels
of well-being for older families, with more assets and fewer years over which to annuitize them. More recently, Moon,
(1977), Lerman and Mikesell (1988), and Rendall and Speare (1993) have refined the income-net worth measure and
used it to measure the poverty of U. S. families. When measured over all families, the rate of income-net worth poverty
is lower than the rate of income poverty, with substantial decreases in poverty rates for older families.
An alternative approach to understanding the links between income and savings (wealth holdings) has been
stimulated by empirical observations that the ratio of wealth to permanent income increases monotonically with
lifetime income, contrary to the prediction by the life cycle hypothesis of a constant ratio across families with varying
lifetime incomes. Ziliak (2001) has empirically investigated these potential explanations, and concludes that eligibility
for asset-tested transfer income accounts in part for the low level of liquid wealth for those with low permanent
income, and that high labor market earnings partially explains why the wealth to permanent income ratio is higher than
expected for families with high permanent income. These approaches both complement the joint income/asset poverty
measures on which we present evidence, and suggest further research regarding the determinants of the probability of
being joint asset-income poor relative to being either income or asset poor.
Caner and Wolff (2001) have also analyzed the level and trend in asset poverty using data from the Panel
Study of Income Dynamics.
Citro and Michael (1995).
Three-parameter scale = (ratio of the scale for 2 adults to one adult is 1.41. For single parents (adults + .8 +
.5 * children - 1)
; all other families (adults + .5 * children)
Our poverty line calculation is drawn from "U.S. Census Bureau (1999); Table C1: CPI-U adjustment,
Table C2: Three-parameter scale; and from the U.S. Census Bureau website:
Note that net worth excludes social security and defined benefit pension wealth (that is, the present value of
future expected social security and defined benefit pension payments, respectively). Such future expected payments
cannot be drawn against to finance current consumption. Defined contribution pensions, however, can be liquidated to
support consumption, albeit with a penalty. The value of vehicles that may be owned is also excluded. The rationale for
excluding vehicles is that for most families, particularly poor families, autos tend to be necessary for work-related
transportation, and therefore not readily available for sale to meet consumption needs.
Both asset measures are defined more completely in the Appendix.
We have combined African-Americans and Hispanics into a single group for two reasons. The first is the
relatively small sample sizes for these two groups and the associated sampling variability. The second is some changes in
the wording of questions on race and ethnicity over the five SCF surveys. In particular, in the 1995 and 1998 surveys, the
race question does not explicitly indicate non-Hispanic whites and non-Hispanic blacks for the first two categories, so that
some Hispanics may have classified themselves as either whites or blacks. In the case of the former, there is no way to
correct the classification.
The results of the other estimated probit regressions for income poverty and joint income and asset poverty
are available from the authors upon request.
The bottom row of Table 3 indicates the extent to which households who are asset poor by the two
definitions have zero or negative net worth. Of the households who are asset poor in the three years shown, about 70
percent have no (or negative) net worth. For these households, no asset cushion exists to provide support should
income from the labor market or the public sector fail. Of the households who are asset poor by the liquid asset
measure, at least 45 percent have no net worth cushion on which to draw. Those households with no asset cushion at
all experience the most severe levels of asset poverty. For both the net worth and the liquid asset measure, the percent
of these asset poor populations with no net worth at all has increased from 1983 to 2001.
The bottom row of Table 6 indicates the proportion of the income poor with no (or negative) levels of net
worth in each of the years. Although the percent of the households who are income poor has fallen over the 1983-2001
period, among the income poor households the proportion with zero or negative net worth has increased. The increase
in this proportion from 37.9 to 42.5 indicates that among the shrinking share of the population that is income poor, the
absence of any short-term asset cushion has in fact increased.
An alternative indicator of the concentration of families in joint asset/income poverty rate for a subgroup is
the share of the families in the subgroup who are in joint poverty relative to the share of the families who are either
income or asset poor. For example, 23.6 percent of white families are either income or asset poor and 4.5 percent are
both income and asset poor; 19 percent of the families who are either income or asset poor are both income and asset
poor. For Black/Hispanic families, this percentage is 37 percent. For 1983, the subgroups with rates of conjoint
asset/income poverty (among the families who are poor by either standard) in excess of one-third are Blacks/Hispanics,
those living in families headed by a person with less than a high school degree, renters, and female-headed families
with children. These same groups, plus those living in a family headed by a person aged less than 25 years, have an
indicator of joint poverty in excess of one-third in 2003.