Federal Reserve Bank of Minneapolis Quarterly Review Vol. 21, No. 2, Spring 1997, pp.

321

Dimensions of Inequality: Facts on the U.S. Distributions of Earnings, Income, and Wealth
Javier Daz-Gimnez Visiting Professor Department of Economics Universidad Carlos III Vincenzo Quadrini Assistant Professor Department of Economics Universitat Pompeu Fabra Jos-Vctor Ros-Rull Senior Economist Research Department Federal Reserve Bank of Minneapolis and Assistant Professor Department of Economics University of Pennsylvania

Abstract This article describes some facts about nancial inequality in the United States that a good theory of inequality must be able to explain. These include the facts that labor earnings, income, and wealth are all unequally distributed among U.S. households, but the distributions are signicantly different. Wealth is much more concentrated than the other two. Wealth is positively correlated with earnings and income, but not strongly. The movement of households up and down the economic scale is greater when measured by income than by earnings or wealth. Differences across the three variables remain when the data are disaggregated by age, employment status, educational level, and marital status of the heads of U.S. households. Each of these classications also has signicant differences across households. All the facts are based on data taken from the 1992 Survey of Consumer Finances and the 198485 and 198990 Panel Study of Income Dynamics.

The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Minneapolis or the Federal Reserve System.

The purpose of this article is to report facts on the distributions of labor earnings, income, and wealth in the United States. We provide a quantitative description of these three most often discussed dimensions of inequality.1 Providing such a description is no easy task, mainly because, in abstract terms, inequality means very little, and when we try to give inequality a concrete meaning, we discover its multidimensional nature. The basic question that any study of inequality has to address is, Inequality of what? When people talk about inequality, they talk about the unequal distributions of opportunities, talents, earnings, income, wealth, consumption, leisure, bequests, luck, and so on. Often people treat some of these variables, especially income and wealth, as if they are more or less the same. But are they? In our view, an accurate description of inequality should acknowledge its multidimensional nature, and it should consider as many of these dimensions as possible. Given this multidimensional nature of inequality, our specic objectives in this article are to use the available data to document some of the dimensions of inequality and to highlight the main features of the data in a coherent and summarized fashion. Creating a precise description of inequality based on available data is difficult. We cannot use established theory to provide us with guidance because there is no such thing as an established theory of inequality. Given this lack of an established theory, we have attempted to provide the data in a format that allows researchers to analyze the data with whatever theory they have in mind and to use the data to test the implications of any theory.2 We have found two reliable and systematic sources of data on inequality among U.S. households: the Survey of Consumer Finances (SCF) and the Panel Study of Income Dynamics (PSID). (The SCF is conducted by the National Opinion Research Center at the University of Chicago and is sponsored by the Federal Reserve with the cooperation of the Department of the Treasury. The PSID is conducted by the Survey Research Center of the University of Michigan and is funded primarily by the National Science Foundation.) Every fact that we report in this article has been constructed from the data obtained from one of those two sources. The sample years we use are 1992 for the SCF and 1984, 1985, 1989, and 1990 for the PSID. (Earnings and income reported in these sample years are for the preceding calendar year. All other data reported are for the sample year.) We discuss some of the technical details of the SCF and the PSID in the Appendix. The dimensions of inequality which we describe in this article are the following: Earnings, Income, and Wealth. The dimensions of inequality that are perhaps the most frequently studied and most easily confused are earnings, income, and wealth. This confusion arises in part from the twisted nature of the relationships among them, especially the relationship between income and wealth. First, labor earnings is one of the components of income, the one related to labor input. Next, income is dened as revenue from all sources before taxes but after transfers. Among other components, this variable includes both

labor earnings and income generated by wealth. Finally, wealth is dened as the net worth of the household, both the stock of unspent past income and one of the sources from which income (capital income) is obtained. Moreover, given that labor income and capital income are perfect substitutes as far as their purchasing power is concerned, wealth also plays a potentially important role in the labor supply decision and, hence, in the determination of labor earnings. (See the Appendix for details on more precise denitions of these three variables.) Additional evidence that earnings, income, and wealth are easily confused is provided by the ambiguous meanings of rich and poor. When people talk about the rich, it is not clear whether they are referring to the earnings-rich, the income-rich, or the wealth-rich. Also confusing are references to the poor, including the earnings-poor, the income-poor, and the wealth-poor. We document unambiguously that these concepts of rich and poor are not all the same. To document some of the earnings, income, and wealth inequality facts, we partition our sample into groups along each of these three dimensions. Since people do move up and down the economic scale, we also report some facts about earnings, income, and wealth mobility. Contrary to common belief, many of the characteristics of the earnings, income, and wealth distributions are signicantly different. We nd that wealth is by far the most concentrated of the three variables, earnings ranks second, and income is the most dispersed of the three. Furthermore, we nd that even though earnings and income are highly correlated, the correlations between earnings and wealth and between income and wealth are surprisingly low, 0.230 and 0.321, respectively. We contend that a good theory of inequality should be able to account for the differences among the distributions of earnings, income, and wealth that we document in this article. Given the interdependences among these variables, accounting for these differences is difficult. Age. The measures of inequality are different if we consider yearly earnings, income, or wealth or if we study those same variables throughout the life cycles of the people in the household. Inequality measurements also differ across age cohorts. We partition our sample into age groups to document some of these differences. We nd that people of retirement age play an important role in accounting for inequality. Employment Status. To document the relationship between income sources and inequality, we partition our sample into workers (people who are employed by others), the self-employed, retirees, and nonworkers (people who do not work but do not consider themselves retired) according to the employment status of the household head. We report the average earnings, income, and wealth; the shares of income accruing from different sources; and the average size of the households in this partition. We nd that the self-employed are rich along all three dimensions. Education. Education (or human capital accumulation) increases the market value of peoples time, and

therefore, it plays a potentially important role in the determination of labor earnings and, hence, in the distributions of earnings, income, and wealth. To characterize the implications of inequality in education, we partition our sample into college graduates, high school graduates, and nonhigh school graduates according to the education level of the head of the household. We report the average earnings, income, and wealth; the shares of income accruing from different sources; and the average size of the households in this partition. It turns out that according to the SCF data set, there is a close association between the education level and the economic performance of households. Marital Status. Finally, we partition the households in our sample according to the marital status of the household head. We report the inequality in earnings, income, and wealth of married households and of single households with and without dependents. The groups of singles with and without dependents are further partitioned by sex. We nd that as far as the economic performance of households is concerned, it seems to pay off to be married. We do not discuss the following potentially important dimensions of inequality, primarily because we have not found a reliable source of data for them: Inherited Ability and Tastes. Two dimensions of inequality, inherited ability and tastes, play important roles in labor/leisure choices, and they are, therefore, potentially important determinants of the earnings distribution and, indirectly, the distributions of income and wealth. Tastes pose additional problems for theory because they play a crucial role in most model economies and are hard to measure. Bequests. Bequests also condition labor/leisure choices, and hence, they help determine the distributions of earnings, income, and wealth. Bequests are an additional motive for altruistic households to save, and hence, they foster earnings, income, and wealth inequality. Finally, bequests are a vehicle for the intergenerational transmission of wealth inequality, and hence, they increase the persistence of that inequality. Luck. Luck probably plays an important role in the determination of inequality. But it is hard to separate luck from some other variables. Are talent, effort, judgment, or luck reasons that make some people better off than others? Why do champions tend to get lucky? The difficulties that arise when trying to answer these and other related questions justify in part our decision not to discuss luck in this article. We also do not describe other forms of inequality, such as differences in the levels of consumption or the number of hours worked.3 Looking at inequality from those points of view perhaps should be our ultimate concern, since to some extent those variables show how the households perceive their own present and future opportunities, and hence, they give us a better indication of inequality in welfare. We leave this approach for a future project. Finally, in light of the inequality facts we document, we describe in our conclusion what we consider to be

the essential ingredients of a successful theory of inequality. Earnings, Income, and Wealth The SCF data set unambiguously shows that earnings, income, and wealth are unequally distributed across the households in the sample. The values of the concentration statistics that we have computed are large, and the density functions of the earnings, income, and wealth distributions are skewed; they present a fat lower tail and a thin upper tail. Perhaps more surprisingly, the data show that while earnings and income are highly correlated, the correlations between income and wealth and between earnings and wealth are much smaller. We report a set of statistics that describe the earnings, income, and wealth partitions. Then we use those statistics to summarize some of the earnings, income, and wealth inequality facts.

Concentration Lets start by examining how concentrated earnings, income, and wealth are in the United States. Table 1 reports the Gini indexes, the coefficients of variation, and the ratios of the shares earned or owned by the top 1 percent and the bottom 40 percent of the earnings, income, and wealth distributions. We have chosen to report this last statistic because the bottom 40 percent is the smallest group that holds a positive share in all three distributions. The three sets of statistics unambiguously show that wealth is the variable that is by far the most concentrated. The households that belong to the top 1 percent of the wealth distribution own 29.55 percent of the wealth, and they are on average 875 times wealthier than those that belong to the bottom 40 percent of the wealth distribution. This difference between the top and bottom groups is about 10 times larger than the difference for the same groups in the income partition and about 4 times larger than that difference in the earnings partition. The concentration statistics that we have computed also show that labor earnings are signicantly more concentrated than income. We conjecture that one of the reasons for this fact is the equalizing effect of income transfers, which we include in our denition of income and which we do not include in our denition of earnings. Transfers make it possible for some people to receive welfare payments and not have to enter the labor force, and hence, these peoples income is signicantly greater than their labor earnings. In Chart 1, we plot the Lorenz curves for the earnings, income, and wealth distributions. In this chart, we observe that in the lower part of the distribution, the Lorenz curve for earnings lies below the Lorenz curve for income. This offers additional support for our conjecture about the equalizing effect of income transfers, since it shows that income is indeed less concentrated than earnings in the lower tail. The Lorenz curve for earnings crosses the Lorenz curve for income at approximately the 87th percentile, which is the point at which the concentration of income increases as the share of capital income starts to dominate. In the entire domain, the Lorenz curve for wealth lies signicantly below the Lorenz curves for earnings and income.

Skewness Table 2 reports the percentiles at which the earnings, income, and wealth means are located and the mean-tomedian ratio for each of the three distributions. In symmetric distributions, the mean is located in the 50th percentile; consequently, the ratio of the mean to the median is 1. As the concentration of a variable increases, so does the mean-to-median ratio, and the location of the mean moves to a higher percentile. We nd that the wealth distribution is the most skewed of the three and that income is somewhat more skewed than earnings. Charts 25 display the histograms of the three distributions and the histogram of nonzero earnings. Correlation To describe the joint distributions of earnings, income, and wealth, we compute the correlation coefficients among these three variables, and we report them in Table 3. As we could have expected, our data show that earnings, income, and wealth are positively correlated and that the correlation between earnings and income is high. This should, indeed, be the case, given that labor earnings account for approximately 72 percent of household income. A more interesting fact is that the correlation between income and wealth is only 0.321. This fact becomes more remarkable if we take into account the high correlation between capital income and wealth. The correlation between earnings and wealth, 0.230, is even lower than that between income and wealth. The low correlation between earnings and wealth could arise for a variety of reasons. For example, it could be a result of the fact that wealthy households assign a signicantly smaller amount of time to the labor market, perhaps because a large fraction of the households are comprised of retirees, or it could be because wealthy households command lower wages. The Poor and the Rich As we have already mentioned, the common usage of the concepts of rich and poor is fairly ambiguous. To avoid this ambiguity, we distinguish between rich and poor in terms of earnings, income, and wealth. In this section, we discuss some of the facts reported in Tables 4, 5, and 6. We organize these facts into two groups: those that pertain to the households that belong to the lower tail of the different partitions, which we refer to generically as the poor, and those that pertain to the households that belong to the upper tail of the different partitions, which we refer to generically as the rich. We have chosen this organization criterion because most of the existing theories of inequality have trouble justifying the two tails of the distributions. We hope that this characterization of the data will point to the possible reasons the existing theories seem to fail.4 The Earnings-Poor Lets start with the earnings-poor. (The data on the earnings partition are displayed in Table 4.) As many as 24 percent of the households in the SCF sample have zero earnings, and an additional 0.42 percent have negative earnings, because there are a large number of

households with members outside the labor force. This is also the case for households with a retired head. In fact, most of the earnings-poor are apparently retirees. (The average age of the heads of the households that belong to the lowest earnings quintile is 65.42 years.) Moreover, households in the lowest quintile earn a signicant share of income (7.93 percent), which consists mostly of transfers and capital income, and they own a sizable share of wealth (17.92 percent).5 The households with negative earnings are mostly headed by business owners in nancial distress. In spite of these business losses, the total income of these households is positive, since they receive signicant shares of transfers and capital income. Moreover, in the SCF sample, the households with negative earnings are wealthy. Specically, the households that are in the lowest 1 percent of the earnings distribution own almost three times the average wealth, which puts them above the 80th percentile of the wealth distribution.

The Earnings-Rich Next, we consider the earnings-rich. (See Table 4.) The households that belong to the top 1 percent of the earnings distribution make almost 15 times the samples average earnings, and those that belong to the top quintile make just over 3 times the average earnings. A large share of the income of the earnings-rich comes from business sources, which include income from professional practices, businesses, and farms. Moreover, this type of income is increasing with earnings. Many of the earnings-rich are married, and they tend to live in large households. (The average household size in the top quintile of the earnings distribution is 3.09 people, while that in the lowest quintile is only 1.73 people.) In fact, across the earnings distribution, except for the lower and upper tails, both the proportion of married households and the average household size are clearly increasing with earnings. The Income-Poor Now lets turn to the income-poor. (See Table 5.) Only 0.96 percent of the households in the SCF sample have zero income. The fraction of households with zero earnings, recall, is 24 percent. If we discount households with heads over age 65, which constitute 20.26 percent of the SCF sample, we still nd at least 3 percent of sample households with positive income and zero earnings. (Their income is either capital income or transfers.) Some of this income is operating as a safety net. An additional 0.25 percent of the households have negative income. (The fraction of households with negative earnings, again, is 0.42 percent.) Two other important facts must be kept in mind when interpreting these numbers. One is that 1991, the year of the SCF data for earnings and income, was a recession year. The other is that the share of income earned by the lowest quintile is procyclical. Hence, the long-term number of the income-poor might be somewhat smaller than these annual data suggest.6 As we could have expected, the negative-income households are, once again, headed by business owners in nancial distress. Given that 1991 was a recession year, the number of business failures that affected the households in our sample was probably above average.

A perhaps more surprising fact is that the incomepoor own above-average wealth. Specically, Table 5 shows that the households that are in the lowest 1 percent of the income distribution own 1.54 percent of total wealth, which puts them in the 85th percentile of the wealth distribution. Moreover, the households that are in the lowest 15 percent of the income distribution own 0.63 percent of total wealth, which puts them in the second quintile of the wealth distribution. Across the whole income distribution, the percentage of income obtained from transfers is decreasing with income. Transfers account for 75.19 percent of the income earned by the households that belong to the lowest income quintile and for only 3.23 percent of the income earned by the households that belong to the top income quintile. Perhaps more surprising is the fact that without transfers, 12.78 percent of the sample households would have zero income. As far as their marital status is concerned, a very large percentage of the income-poor are single, both with and without dependents. Specically, while singles without dependents account for about half of the households in each of the lowest two quintiles, they represent only 31.18 percent of the total sample. The share of singles with dependents is also signicantly larger in the lowest quintile (21.12 percent) than in the total sample (11.41 percent), and the share of singles with dependents decreases as income increases.

The Income-Rich Turning to the income-rich, we nd that the households that belong to the top 1 percent of the income distribution earn about 19 times the samples average income, but when we consider those households that belong to the top quintile, this number is reduced to 3 times. Here, as was the case in the earnings partition, the income-rich receive a signicant share of their income from business sources. Specically, business income accounts for 27.49 percent of the income of the households that are in the top 1 percent of the income distribution and for 16.16 percent of the income of the households that are in the top income quintile. The income-rich also tend to be earnings- and wealth-rich. In fact, the households that are in the top income quintile hold very similar shares of earnings, income, and wealth: 58.36 percent, 59.91 percent, and 62.73 percent, respectively. Finally, the income-rich are mostly middle-aged and married, and they tend to live in large households. Specically, in the top income quintile, 85.5 percent of the household heads are between 31 and 65 years old, 88.21 percent of the top quintile household heads are married, and their average household size is 2.95 people. Moreover, across the income distribution, both the share of married households and the household size are clearly increasing with income. The Wealth-Poor Next, we discuss the wealth-poor. (Table 6 shows the wealth partition.) Approximately 3.4 percent of the sample households have zero wealth, and another 3.5 percent have negative wealth. However, in spite of this reasonably small number of propertyless households, wealth is by far the most unequally distributed of the

three variables that we consider in this section. The households that are in the lowest 40 percent of the wealth distribution own only 1.35 percent of the total sample wealth, and those in the lowest 80 percent own only 20.51 percent of the total sample wealth. The SCF data also show that some of the wealthpoor are reasonably well-to-do, in terms of both earnings and income. Specically, the earnings of the lowest 1 percent of the wealth-poor households are only slightly lower than median earnings, and their income is slightly above median income. Furthermore, given that these households have a signicant ability to borrow (with average debts that amount to approximately 50 percent of average wealth), there must be some sense in which these households are not actually poor. The average net worth of the rest of the households in the lowest wealth quintile is approximately zero. These households, however, also make a signicant amount of income, which puts them in the second and third quintiles of the income distribution. The wealth-poor tend to be young and single. A total of 33.64 percent of the households in the lowest wealth quintile have a head under age 31. This percentage is more than twice the sample average (16.44 percent). The percentage of households in the lowest wealth quintile that are single is 64.48, and 24.93 percent of them are single with dependents, which, again, is more than twice the sample share of singles with dependents.

The Wealth-Rich Finally, lets look at the wealth-rich. Table 6 shows that the households in the top 1 percent of the wealth distribution own 29.55 percent of the total sample wealth, and those in the top quintile own an impressive 79.49 percent. Moreover, this last group of households is both earnings-rich and income-rich. (The households in the top quintile earn 41.21 percent of total earnings and make 46.15 percent of total income.) The top quintile wealth-rich obtain signicant shares of their income from capital (18.39 percent) and from business sources (17.95 percent). Wealth-rich households tend to be both older and married. The percentage of household heads in the top wealth quintile over age 65 is 29.70, while the sample share of that age is only 20.26 percent, and 77.07 percent of the household heads in the top wealth quintile are married, while the sample share of married households is 57.41 percent. Mobility People move up and down the economic scale; they do not stay in the same earnings, income, and wealth groups forever.7 Different reasons make households change earnings, income, and wealth groups. Perhaps aging is the main cause of mobility for most households, but it is certainly not the only one. Mobility is also affected by the results of business projects and other ventures that can bring about signicant changes in earnings to lucky or unlucky entrepreneurs. There can also be some other radical expressions of good luck (such as gambling) or bad luck (such as accidents). Yet some other changes are a consequence of the conscious effort of households to smooth consumption over time.

Whatever its cause, economic mobility makes inequality an essentially dynamic phenomenon, and in our opinion, a trustworthy theory of inequality should be able to account for at least some of the earnings, income, and wealth mobility facts that we report in this section. All the facts reported so far are based on data from the SCF. However, since the SCF is not a panel study, it does not track people over time.8 Therefore, to construct our mobility measures, we use data from the PSID.9 We use data on household net worth from the PSID for the years 1984 and 1989 (reported in the 1984 and 1989 PSIDs) and combine them with data on earnings and income for the same households for those two years (reported in the 1985 and 1990 PSIDs). We use the PSID data to construct Table 7, where we report the transition matrices for the 1984 earnings, income, and wealth quintiles. For example, the entry in the rst row and the rst column of Table 7 reports that 85.8 percent of the households in the lowest earnings quintile in 1984 were also in the lowest earnings quintile in 1989. To avoid the role of retirees in shaping the mobility of households with zero earnings, Table 8 reports the transition matrices in earnings for households with positive earnings in both sample periods. To partially control for the role played by age in shaping the properties of the mobility of earnings, Table 8 also reports the transition matrices of earnings for those households with heads between the ages of 35 and 45 in 1984. As far as earnings is concerned, the households in the lowest quintile are by far the least mobile. This lack of mobility is probably mostly attributable to agerelated issues, but it could also reect some form of earnings poverty trap. (Recall that the lowest quintile is made up of a large fraction of retirees, and retirees seldom move out of retirement.) In general, the lowest and highest quintiles should be the least mobile, since the households in those quintiles can only move either up or down the economic scale, while those in the middle quintiles can move both up and down. The households in the three middle quintiles are clearly the most mobile. When we exclude the households with zero earnings, the second-highest eigenvalue for earnings is only 0.687. In this case, earnings becomes the most mobile of the three variables and wealth the most persistent. As far as income is concerned, again, the households in the lowest quintile are the least mobile, but they are more mobile than those in the lowest quintile of the earnings partition. In contrast, in the wealth partition, the households in the top quintile are the least mobile. This suggests that wealth is the most persistent of the three variables under study. These transition matrices include a lot of information. We want to use a simple, one-dimensional summary statistic. One such statistic is the second-highest eigenvalue.10 The closer this eigenvalue is to 1, the more persistent is the variable under study. The secondhighest eigenvalues for earnings, income, and wealth are 0.807, 0.742, and 0.804, respectively. Therefore, according to these statistics, the mobility among income

quintiles is greater than the mobility among earnings or wealth quintiles. Other Dimensions of Inequality Some characteristics of households that are closely related to earnings, income, and wealth are age, employment status, education, and marital status. Here we discuss in detail how those dimensions shape inequality among households in terms of earnings, income, and wealth.

Age Some of the differences in earnings, income, and wealth across households can be safely attributed to the differences in peoples age.11 There are two main ways to quantify the earnings, income, and wealth differences. One way is to compare peoples lifetime statistics with their yearly statistics. This would require following a sample of households through their entire life cycle. Unfortunately, we do not have a long enough panel to do that. Thus, we choose instead to use the other way of quantifying differences: to partition the population of the household heads into age cohorts and compute the relevant statistics for each cohort. We report these statistics in Table 9. Specically, we report, for 10 age cohorts and for the entire sample, the U.S. sample averages and the Gini indexes for earnings, income, and wealth; the percentages of income from various income sources; the relative cohort size; and the number of people per primary economic unit. We nd that earnings are monotonically increasing with the age of household heads until age 50, when earnings start to decline. As we could have expected, the earnings of households with a head over age 65 drop signicantly to only about 15 percent of the samples average earnings. The income of the different age cohorts displays a similar behavior. Income is moderately increasing until age 55, and then it declines, albeit signicantly more gradually than earnings. The average income of households with a head over age 65 is close to 62 percent of the average income in the total sample. Wealth is also monotonically increasing in the early stages of the life cycle and peaks a little before age 60, ve years after income does. The group over age 65 owns more wealth than any of the groups age 45 and under. With some exceptions, the Gini index for earnings is moderately increasing with age, and it is highest for the group over age 65.12 The Gini index for income displays a similar behavior. A perhaps more surprising fact is that age seems to make little difference for wealth inequality. The maximum difference in this statistic between ages is only 0.089, and if we do not take into account the youngest cohort (whose wealth is the most concentrated), then this difference drops to only 0.036. As far as the income sources of the different age cohorts are concerned, it appears that they are almost monotonic in age for all types of income. With the exception of the youngest and the 3645 age groups, the share of labor income decreases as age increases. The share of capital income tends to increase with age. The share of business income is lowest at both ends of the age distribution and highest in the 3645 and 6165 age groups. The smaller shares in the 4660 age group

are hard to explain. The share of transfers is quite low for all ages except, of course, for the older cohorts. It increases somewhat in the 6165 age group, and it peaks in the over-65 age group; transfers account for almost half of the latter groups income. In the case of households with heads age 25 and under, transfers also account for a larger share of income than in the middleaged groups.

Employment Status Next, in order to document the relationship between income sources and inequality, we partition the SCF sample into workers, the self-employed, retirees, and nonworkers according to the employment status of the household head. In Table 9, we report the 1992 U.S. sample averages and Gini indexes for earnings, income, and wealth; the percentages of income from various sources; the relative cohort size; and the number of people per primary economic unit for these four employment status groups and for the entire sample. It turns out that the differences across these groups are substantial. Workers are the largest group; they account for 54.9 percent of the sample. The average earnings of workers are about 25 percent higher than the sample average, and their average income is nearly 6 percent higher than the sample average. Also, workers average wealth is signicantly lower than the sample average. (Workers own about 67 percent of the sample average.) Although self-employed households make up only 10.9 percent of the sample, they enjoy a remarkably good nancial situation. Their income is almost twice the sample average, and they own an even greater share of wealthmore than three times the sample average. Retirees account for 18.1 percent of the sample. Their income is about 78 percent of the average. As we could have expected, while both the earnings and the income of retirees are below the sample average, their wealth is above the average (almost 24 percent above it). Households with a head who does not work are both income-poor and wealth-poor. The earnings of these households are less than one-third of the average earnings, which account for half of their income. Another important source of income for this group is transfers. Education Next, in order to document the relationship between education and inequality, we partition the SCF sample into three groups based on the level of education attained: a group labeled college, which includes households with a head who has at least a college degree; a group labeled high school, which includes households with a head who has a high school degree but has not completed college; and a group labeled no high school, which includes households with a head who has not completed high school. In Table 9, we report, for these three education groups and for the entire sample, the averages and Gini indexes for earnings, income, and wealth; the percentages of income from various sources; the group size; and the number of people per primary economic unit. According to the SCF data set, there is a close association between the education level and the economic performance of households. High school households

make up 50.2 percent of the SCF sample; college households, 28.6 percent; and nohigh school households, the remaining 21.2 percent. On average, college and high school households have earnings that are, respectively, about six and three times higher than the earnings of nohigh school households. The differences in wealth holdings are also large, about ve and two times larger, respectively. Finally, the differences in income, although still largeabout four and two times, respectivelyare somewhat smaller due in part to the equalizing effect of transfers, which account for 28.2 percent of the income of nohigh school households. The Gini indexes show that the concentrations of income and wealth are very similar across education levels, while earnings are most concentrated in households with no high school. As far as the income sources are concerned, college households obtain more income from business and capital sources than do other groups; households that have completed high school are mostly laborers; and among all these groups, households with no high school receive the largest share of income from transfers and the lowest share from labor, capital, and business sources. Finally, it is also the case that the average size of the SCF primary economic unit is slightly increasing with the amount of education of the head of the household.

Marital Status In this section, we document the relationship between marital status and inequality. For this purpose, we partition the SCF sample into married and single households according to the marital status of the head of the household. We also partition singles according to whether or not they have dependents, and we subdivide these two partitions according to the sex of the head of the household. We refer to these groups as the marital status partition.13 In Table 9, we report the averages and Gini indexes for earnings, income, and wealth; the percentages of income from various income sources; the relative cohort sizes; and the number of people per primary economic unit for these marital status groups and for the entire sample. The main properties of the marital status partition are the following: compared to single households with or without dependents, married households make substantially higher earnings and income and own a substantially higher amount of wealth. This is still the case if we divide the earnings, income, and wealth of married households by two to account for double-income households. We nd that singles without dependents are signicantly better off nancially than singles with dependents. Not only are the earnings of singles without dependents about 12 percent higher, but their income is about 30 percent higher, and their wealth is close to an impressive 120 percent higher than singles with dependents. However, the average household size of singles without dependents is only about one-third of the average household size of singles with dependents. The percentage of income from transfers is about three times larger for single households than for married households. As we could have expected, the percentage of income from transfers is the largest for singles with dependents.

As far as the Gini indexes are concerned, both the earnings and the income of single households without dependents are the most unequally distributed, while the greatest concentration of wealth corresponds to single households with dependents. Finally, single females signicantly outnumber single males in the SCF sample, with sample shares of 27.5 percent and 15.1 percent, respectively. This difference is consistent with the fact that females live longer than males. Single females both with and without dependents earn less labor earnings (47 percent less), make less income (38 percent less), and own less wealth (31 percent less) than their male counterparts. Also, single females with dependents account for a large part of the sample (9.3 percent), and they are in a particularly bad nancial position: their earnings, income, and wealth are only about 36 percent, 41 percent, and 24 percent, respectively, of the sample averages. Conclusion So far, economists have no satisfactory theory of inequality. Such a theory must simultaneously account for all of the properties of the U.S. distributions of earnings, income, and wealth that we have discussed here: concentration, skewness, and correlation. Moreover, such a theory of inequality must account for the dynamic features of such distributions, that is, the mobility of individual households up and down the economic scale over time, which we have also discussed here. In light of the inequality facts in this article, we suggest that the following elements are important ingredients for a reliable theory of inequality:

cessful theory should account for how the patterns of household formation and dissolution shape inequality. All this probably amounts to a somewhat tall order for theorists, but work in this direction has begun. In a companion article in this issue, Quadrini and Ros-Rull assess the performance of existing theories of inequality to account for the U.S. wealth distribution and discuss new directions of research that take into account the dimensions of inequality we discussed in this article.

Transfers. Income transfers distort the labor/ leisure decision, and they allow households to survive without work. They are an important source of income for earnings- and wealth-poor households; hence, they should play an important role in any attempt to account for the lower tails of the distributions. Businesses. Businesses in nancial distress account for the sizable amount of negative income earned by many U.S. households. Moreover, business income is an important source of income for the households in the upper tails of the distributions. These facts suggest that both business successes and business failures should be important elements for any theory of inequality. Retirees. Retirees hold a large share of total wealth. Moreover, their labor earnings are zero. These facts spell trouble for any theory of inequality that abstracts from elements of the life cycle. Education. Households whose head has a college education have more than twice the earnings, income, and wealth of those households whose head has a high school education. Understanding the determinants of the acquisition of education becomes a crucial part of understanding inequality. Marital Status. The better nancial performance of married households over single households cannot be accounted for only by family size. A suc-

Body Notes *For contributions to this work, the authors thank research technical support staff at the Minneapolis Fed and the editors and referees of this journal. 1 This article is by no means unique in its attempts to account for U.S. inequality in earnings, income, and wealth. For example, Weicher (1995) describes the changes in the U.S. wealth distribution between 1983 and 1989. Using preliminary data, Kennickell, Starr-McCluer, and Sundn (1997) detail recent changes in the income, net worth, assets, and liabilities of U.S. families. Wolff (1987) produces estimates of wealth inequality for the 196283 period. In contrast to these studies, we attempt to provide a global view of inequality that relates earnings, income, and wealth rather than concentrate on how the distribution of one or more of these variables changes over time. 2 Quadrini and Ros-Rull (in an article in this issue of the Quarterly Review) review some recent theories of inequality, and they evaluate these theories according to how well the theories account for some of the data we report here. 3 See Slesnick 1993, 1994 for a discussion of inequality in consumption. 4 In our discussion of the rich, we highlight the characteristics of the top 1 percent because the households that belong to this small group make 14.76 percent of total earnings and 18.57 percent of total income, and they own 29.55 percent of total wealth. 5 These wealth holdings would put the households that belong to the lowest quintile of the earnings distribution well into the third quintile of the wealth distribution. 6 For details on the cyclical behavior of the income distribution, see Castaeda, Daz-Gimnez, and Ros-Rull 1995. 7 Note that this could be the case, and we could still have invariant distributions of earnings, income, and wealth. 8 Actually, in the 1983 and 1986 SCFs, there was a limited effort to follow households over time. See Kennickell and Starr-McCluer 1994. 9 An important shortcoming of the PSID is that, unlike the SCF, it has not been designed to address issues related to wealth holdings, and therefore, the data for these variables are of lower quality, especially the data that pertain to the wealthand income-rich. For a discussion of the PSID, see the Appendix. 10 Note that in probability transition matrices, the highest eigenvalue is always 1. 11 In fact, there is a large quantitative literature that uses models in which differences in peoples age are the main source of differences in earnings, income, and wealth across households. See, for example, Auerbach and Kotlikoff 1987, Fullerton and Rogers 1993, and Ros-Rull 1996. 12 In fact, for this group, the Gini index shows a rarely seen value higher than 1, because there are a nontrivial number of households with negative earnings. 13 Note that singles without dependents do not necessarily live alone; they may also live with other nancially independent adults.

Appendix Data Sources and Variable Denitions

Here we describe where we got the data and how we dene the variables discussed in the preceding article. Data Sources Our primary data source is the 1992 Survey of Consumer Finances (SCF) conducted by the National Opinion Research Center at the University of Chicago and sponsored by the Federal Reserve with the cooperation of the Department of the Treasury. The SCF is probably the most comprehensive source of data on the earnings, income, and wealth of U.S. households.

The SCF uses a two-part sampling strategy designed to obtain a sufficiently large and unbiased sample of wealthier households. The 1992 sample includes 3,906 households, out of which 2,456 were selected using standard multistage areaprobability sampling methods. The remaining 1,450 households were selected using tax report data. This second group of households was specically selected to oversample wealthier households. To enhance the reliability of the data, the SCF also makes weighting adjustments for survey nonrespondents. (See Kennickell and Starr-McCluer 1994 and the references contained therein for details on the properties of this data set. Also see Kennickell, McManus, and Woodburn 1996 for the statistical apparatus for understanding the signicance of the results.) Our secondary data source is the Panel Study of Income Dynamics (PSID) conducted by the Survey Research Center of the University of Michigan and funded primarily by the National Science Foundation. The PSID follows households over time, and we have used its data to construct our measures of household mobility. The only two years for which PSID data on household wealth are available are 1984 and 1989.1 We combine these data with data on earnings and income from 1985 and 1990 that refer to 1984 and 1989. Unlike the SCF sample, the PSID sample includes a very small number of income-rich and wealth-rich families; therefore, the statistics computed for the right tail of the distribution based on the PSID data set are less reliable. Variable Denitions Households The households in this article are the primary economic units of the SCF. A primary economic unit includes a person or a couple of persons who live together and all the other persons who live in the same household who are nancially dependent on them. For example, underage children and, in some circumstances, older relatives are considered dependents. A nancially independent person who lives in the same house, such as a roommate or a brother-in-law, is not considered to be a member of the unit. We also follow the SCF convention as far as the determination of the head of the household is concerned. The SCF considers the male of a couple to be the head of the household.2

Children, food stamps, and other forms of welfare and assistance; income from Social Security and other pensions, annuities, compensation for disabilities, and retirement programs; income from all other sources including settlements, prizes, scholarships and grants, inheritances, gifts, and so on. In other words, the notion of income that we use attempts to include all before-tax income received during the year. It approximately corresponds to the payments to the factors of production owned by the household plus transfers. However, it does not include imputed income from the services of some assets such as owner-occupied housing. (See Slesnick 1992, 1993 for details.)

Wealth We dene wealth as the net worth of households. This includes the value of nancial and real assets of all kinds net of various kinds of debts. Specically, the assets that we consider are the following: residences and other real estate; farms and all other businesses; checking accounts, certicates of deposit, and other banking accounts; IRA/Keogh accounts, money market accounts, mutual funds, bonds and stocks, cash and call money at the stock brokerage, and all annuities, trusts, and managed investment accounts; vehicles; the cash value of term life insurance policies and other policies; money owed by friends, relatives, businesses, and others; pension plans accumulated in accounts; and other assets.3 Our denition of wealth differs slightly from those used in other studies. Wolff (1995), for instance, provides several denitions of household wealth. The denition of his that is closest to ours is what he calls marketable wealth. The main difference between this denition and ours is that he does not include vehicles and pension plans accumulated in accounts, while we do. Kennickell and Starr-McCluers (1994) denition differs from ours in that they include the current face value of term life insurance policies that build up a cash value (that is, the cash amount paid in case the particular event occurs), while we include only the cash value of these policies.
The SCF and the U.S. NIPA Other data available on income and wealth are consistent with the SCF sample data. For example, in the 1992 SCF sample, average household income for the calendar year of 1991 was $45,924, and average household income excluding transfers for that year was $41,610. In comparison, 1991 personal income minus government transfers, as measured by the U.S. national income and product accounts (NIPA), was slightly over $40,000.4 Also, in the 1992 SCF sample, average household wealth in 1992 was $184,308, and the resulting ratio of wealth to income minus transfers was 4.43. The ratio between the Federal Reserve ow of funds accounts measurement of household net worth and the NIPA denition of national income was 4.31 in 1988. Notwithstanding the differences in the denitions of income and wealth, these two ratios are fairly similar.5

Earnings, Income, and Wealth The key variables that we consider in this article are labor earnings, income, and wealth. The denitions of these variables are as follows. Earnings We dene labor earnings as wages and salaries of all kinds plus a fraction of business income. Business income includes income from professional practices, businesses, and farm sources. The value for the fraction of business and farm income that we impute to labor earnings is the samplewide ratio of unambiguous labor income (wages plus salaries) to the sum of unambiguous labor income and unambiguous capital income. In the sample that we consider, this ratio is 0.864. Income We dene income as all kinds of revenue before taxes. Hence, our denition of income includes both government and private transfers. Specically, the sources of income that we consider are the following: wages and salaries; income (whether positive or negative) from professional practices, businesses, and farm sources; interest income, dividends, gains or losses from the sale of stocks, bonds, and real estate; rent, trust income, and royalties from any other investment or business; unemployment and worker compensation; child support and alimony; Aid to Dependent Children, Aid to Families with Dependent

Appendix Notes 1 At the time this article was written, 1994 PSID data on household wealth were not available. 2 In single households, the nancially independent person of either sex is considered to be the household head. 3 Note that in our denition of wealth, we have not included the present value of pension plans not accumulated in accounts. 4 These calculations are based on a population size of 250 million and an average household size of 2.4 people. 5 To rene our comparison, we should subtract from the NIPA denition of national income the following components: corporate prots minus personal dividends, employer contributions to Social Security, and the rent imputed to owneroccupied houses. We should also subtract from the Federal Reserve ow of funds accounts measurement of household net worth the value of all consumer durables

other than vehicles. These corrections would reduce both the numerator and the denominator of the wealth-to-income ratio, and we conjecture that the corrected value for that ratio would not differ by much from the one that we have quoted.

References

Auerbach, Alan J., and Kotlikoff, Laurence J. 1987. Dynamic scal policy. Cambridge, Mass.: Cambridge University Press. Castaeda, Ana; Daz-Gimnez, Javier; and Ros-Rull, Jos-Vctor. 1995. Unemployment spells and income distribution dynamics. Working Paper 95-9. Federal Reserve Bank of Philadelphia. Fullerton, Don, and Rogers, Diane Lim. 1993. Who bears the lifetime tax burden? Washington, D.C.: Brookings Institution. Kennickell, Arthur B.; McManus, Douglas A.; and Woodburn, Louise R. 1996. Weighting design for the 1992 Survey of Consumer Finances. Manuscript. Board of Governors of the Federal Reserve System.

Kennickell, Arthur B., and Starr-McCluer, Martha. 1994. Changes in family nances from 1989 to 1992: Evidence from the Survey of Consumer Finances. Federal Reserve Bulletin 80 (October): 86182. Kennickell, Arthur B.; Starr-McCluer, Martha; and Sundn, Annika E. 1997. Family nances in the U.S.: Recent evidence from the Survey of Consumer Finances. Federal Reserve Bulletin 77 (January): 124. Ros-Rull, Jos-Vctor. 1996. Life-cycle economies and aggregate uctuations. Review of Economic Studies 63 (July): 46589. Slesnick, Daniel T. 1992. Aggregate consumption and saving in the postwar United States. Review of Economics and Statistics 74 (November): 58597. ___________. 1993. Gaining ground: Poverty in the postwar United States. Journal of Political Economy 101 (February): 138. ___________. 1994. Consumption, needs and inequality. International Economic Review 35 (August): 677703. Weicher, John C. 1995. Changes in the distribution of wealth: Increasing inequality? Federal Reserve Bank of St. Louis Review 77 (January-February): 523. Wolff, Edward N. 1987. Estimates of household wealth inequality in the U.S., 19621983. Review of Income and Wealth 33 (September): 23156. ___________. 1995. Top heavy: A study of the increasing inequality of wealth in America. New York: Twentieth Century Fund Press.

Tables 13

Measures of U.S. Earnings, Income, and Wealth

Table 1 Concentration
Ratio of Top 1% to Bottom 40%

Variable

Gini Index

Coefficient of Variation

Earnings Income Wealth

.63 .57 .78

4.19 3.86 6.09

211 84 875

Table 2 Skewness
Percentile Location of Mean Ratio of Mean to Median

Variable

Earnings Income Wealth

65 71 80

1.65 1.72 3.61

Table 3 Correlation
Correlation Coefficient

Variables

Earnings and Income Earnings and Wealth Income and Wealth

Source: 1992 Survey of Consumer Finances

.928 .230 .321

Chart 1

The Lorenz Curves for the U.S. Distributions of Earnings, Income, and Wealth
What % of All Households Have What % of All Earnings, Income, and Wealth % 100

80

Earnings
60

40

20

Income
0 10 0

Wealth

20 40 60 80 % of Households (Ranked by Amount)

100

Source: 1992 Survey of Consumer Finances

Charts 25

U.S. Distributions of Earnings, Income, and Wealth

With Levels Normalized So That the Mean is 1*

% 24

Chart 2 All Earnings

% 24

Chart 4 Income

12 10 8 6 4 2 0 2

12 10 8 6 4 2 0 2

4 6 Normalized Level

10

4 6 Normalized Level

10

Chart 3 Nonzero Earnings

% 24

Chart 5 Wealth
% 24

12 10 8 6 4 2 0 2

12 10 8 6 4 2 0 2

4 6 Normalized Level

10

4 6 Normalized Level

10

*On all of these charts, the last plotted bar represents the frequency of households with more than 9.91 times the average level. Source: 1992 Survey of Consumer Finances

Table 4

U.S. Households Ranked by Earnings . . .

Characteristics of Sample Households in Each Earnings Group*

The Earnings-Poor
Household Characteristics

Bottom 1%
.40 .41 .29 2.92

Households in Earnings Quintiles

1st .40 .02 7.93 17.92 2nd 3.19 .16 7.87 12.24 3rd 12.49 .62 11.24 8.30 4th 23.33 1.17 19.75 13.71 5th 61.39 3.07 53.21 47.83

The Earnings-Rich
Top 105%
12.38 2.47 10.60 8.84 51% 16.37 4.09 14.80 14.59 1% 14.76 14.76 13.09 14.68 Total Sample 100.00 1.00 100.00 100.00

15% .00 .00 1.54 3.32

510% .00 .00 1.78 4.09

Share of Total Sample

(% of $)

Earnings Ratio of Earnings Group Average to Sample Average Income Wealth

Share of Each Groups Income

(% of $)

Source of Income
Labor Capital Business Transfers Other 33.28 101.58 156.07 91.42 29.79 .00 28.14 .00 59.46 12.40 .00 31.58 .00 63.81 4.61 1.20 31.58 5.63 55.13 17.72 27.75 18.63 1.70 37.00 14.92 76.97 5.50 3.49 8.67 5.37 81.48 5.74 4.17 3.26 5.36 67.72 7.92 17.79 .93 5.65 75.79 6.45 9.68 1.45 6.62 67.76 9.30 13.75 .51 8.68 43.28 11.67 43.85 .43 .77 63.06 9.93 10.37 9.39 7.24

Share of Each Groups Sample

(% of People)

Age of Household Head

Share of Each Group 30 and Under 3145 4665 Over 65 Average Age (Years) .55 19.41 20.49 59.55 65.22 5.90 10.48 19.60 64.02 65.82 7.34 8.06 18.95 65.65 66.03 6.42 10.67 19.59 63.31 65.42 22.17 25.67 24.11 28.05 49.63 26.24 39.39 29.06 5.32 41.18 19.55 45.65 31.87 2.93 41.61 7.79 49.19 41.31 1.72 44.32 7.50 55.06 35.69 1.75 43.59 5.02 42.52 50.77 1.69 46.26 2.02 34.01 59.27 4.70 49.22 16.44 34.11 29.19 20.26 48.43

Marital Status of Household Head

Married Single Without Dependents With Dependents 42.95 44.07 12.98 32.68 53.71 13.61 29.55 55.59 14.86 32.13 53.22 14.65 38.29 43.33 18.38 53.65 30.75 15.60 72.13 20.21 7.65 90.86 8.39 .75 94.50 4.60 .90 93.91 5.99 .10 87.69 11.13 1.18 57.41 31.18 11.41

Average Household Size (Number of People)

2.05

1.66

1.72

1.73

2.06

2.41

2.79

3.09

3.09

3.23

3.01

2.41

*Percentages may not sum to 100% due to rounding. Source: 1992 Survey of Consumer Finances

Table 5

. . . Ranked by Income . . .
Characteristics of Sample Households in Each Income Group*

The Income-Poor
Household Characteristics

Bottom 1%
.30 .31 .37 1.54

Households in Income Quintiles

510% .54 .11 .19 .86 1st 2.18 .11 .71 5.29 2nd 6.63 .33 5.37 7.05 3rd 11.80 .59 12.60 9.95 4th 19.47 .97 22.95 14.98 5th 59.91 3.00 58.36 62.73

The Income-Rich
Top 105%
10.72 2.14 12.32 11.97 51% 15.87 3.97 16.76 22.25 1% 18.57 18.57 11.73 16.32 Total Sample 100.00 1.00 100.00 100.00

15% .16 .04 .07 .63

Share of Total Sample

(% of $)

Income Ratio of Income Group Average to Sample Average Earnings Wealth

Share of Each Groups Income

(% of $)

Source of Income
Labor Capital Business Transfers Other 28.76 1.56 135.55 6.31 1.08 30.11 7.30 .97 60.56 1.05 24.06 .68 1.45 73.70 .12 39.35 3.11 18.28 75.19 .62 57.46 5.23 1.03 34.79 1.49 74.82 4.94 2.44 16.62 1.18 81.61 5.88 3.76 7.94 .80 56.20 13.00 16.16 3.23 11.41 72.98 10.85 11.34 3.79 1.03 62.05 16.30 16.19 2.56 2.90 21.75 17.08 27.49 .69 32.98 63.06 9.93 10.37 9.39 7.24

Share of Each Groups Sample

(% of People)

Age of Household Head

Share of Each Group 30 and Under 3145 4665 Over 65 Average Age (Years) 10.46 28.73 39.89 20.92 52.46 28.13 20.10 29.71 22.06 47.96 20.39 19.20 25.18 35.23 53.80 19.49 22.44 22.55 35.52 53.18 19.64 25.03 24.82 30.51 51.21 22.64 37.18 23.49 16.69 45.11 14.05 42.28 33.21 10.46 45.28 6.36 43.64 41.86 8.13 47.37 6.62 44.28 42.13 6.96 46.94 3.21 39.50 47.36 9.92 49.50 3.72 29.40 52.47 14.40 52.54 16.44 34.11 29.19 20.26 48.43

Marital Status of Household Head

Married Single Without Dependents With Dependents 32.95 52.82 14.23 21.93 56.07 22.00 16.39 58.79 24.82 22.30 56.58 21.12 43.85 40.35 15.80 56.87 29.78 13.35 75.83 19.01 5.16 88.21 10.19 1.60 89.13 10.59 .28 90.99 8.29 .72 80.86 16.72 2.41 57.41 31.18 11.41

Average Household Size (Number of People)

1.76

1.87

1.87

1.87

2.08

2.40

2.76

2.95

2.90

3.12

2.69

2.41

*Percentages may not sum to 100% due to rounding. Source: 1992 Survey of Consumer Finances

Table 6

. . . And Ranked by Wealth

Characteristics of Sample Households in Each Wealth Group*

The Wealth-Poor
Household Characteristics

Bottom 1%
.52 .52 .83 .75

Households in Wealth Quintiles

510% .01 .00 .88 1.26 1st .39 .02 7.05 6.90 2nd 1.74 .09 14.50 12.55 3rd 5.72 .29 16.48 14.87 4th 13.43 .67 20.76 19.54 5th 79.49 3.97 41.21 46.15

The Wealth-Rich
Top 105%
12.62 2.52 8.43 9.05 51% 23.95 5.99 12.34 13.80 1% 29.55 29.55 7.65 9.59 Total Sample 100.00 1.00 100.00 100.00

15% .02 .00 1.18 1.11

Share of Total Sample

(% of $)

Wealth Ratio of Wealth Group Average to Sample Average Earnings Income

Share of Each Groups Income

(% of $)

Source of Income
Labor Capital Business Transfers Other 78.87 11.48 .64 8.79 .22 75.78 .09 .36 22.26 1.50 50.30 .21 .36 34.76 14.36 72.15 1.68 1.65 20.06 4.46 81.30 .53 2.22 7.17 8.77 76.53 2.40 3.83 11.30 5.94 71.56 4.64 5.75 10.93 7.11 48.80 18.39 17.95 7.14 7.72 57.86 15.04 10.62 9.76 6.71 43.40 20.27 24.31 4.94 7.07 30.50 33.54 31.22 2.39 2.34 63.06 9.93 10.37 9.39 7.24

Share of Each Groups Sample

(% of People)

Age of Household Head

Share of Each Group 30 and Under 3145 4665 Over 65 Average Age (Years) 11.64 56.77 24.67 6.92 42.16 31.19 45.45 16.68 6.69 38.62 33.17 30.57 18.45 17.80 43.14 33.64 36.48 17.06 12.81 40.90 27.49 39.13 21.39 11.99 42.32 13.64 37.48 26.81 22.07 49.43 5.27 35.79 34.21 24.72 52.61 2.14 21.69 46.46 29.70 56.89 1.06 18.56 51.32 29.06 57.71 1.34 21.11 50.53 27.01 56.33 .49 14.00 54.82 30.70 59.28 16.44 34.11 29.19 20.26 48.43

Marital Status of Household Head

Married Single Without Dependents With Dependents 70.79 23.56 5.64 29.82 38.47 31.71 21.84 45.18 32.98 35.52 39.55 24.93 49.82 36.28 13.90 59.42 30.96 9.62 65.23 28.42 6.35 77.07 20.70 2.24 79.30 18.45 2.24 83.83 14.94 1.23 85.27 12.06 2.68 57.41 31.18 11.41

Average Household Size (Number of People)

2.51

2.44

2.28

2.37

2.36

2.44

2.43

2.47

2.38

2.74

2.58

2.41

*Percentages may not sum to 100% due to rounding. Source: 1992 Survey of Consumer Finances

Table 7

Three Measures of the Economic Mobility of U.S. Households

Percentage of Households in Each Quintile in 1984 That Were in Each Quintile in 1989

1989 Quintile Measure 1984 Quintile

1st

2nd

3rd

4th

5th

Earnings

1st 2nd 3rd 4th 5th

85.8 18.6 7.1 7.5 5.8

11.6 40.9 12.0 6.8 4.1

1.4 30.0 47.0 17.5 5.5

.6 7.1 26.2 46.5 18.3

.5 3.4 7.6 21.7 66.3

Income

1st 2nd 3rd 4th 5th

71.0 19.5 5.1 2.5 1.9

17.9 43.8 25.5 10.7 2.1

7.0 22.9 37.2 23.4 9.5

2.9 10.1 24.9 42.5 20.3

1.3 3.7 7.3 20.8 66.3

Wealth

1st 2nd 3rd 4th 5th

66.7 25.4 5.8 1.8 .7

23.4 46.6 24.4 4.6 .8

6.6 20.4 44.9 22.4 5.7

2.9 5.4 20.5 49.6 21.6

.4 2.3 4.6 21.6 71.2

Source: 1984, 1985, 1989, and 1990 Panel Study of Income Dynamics

Table 8

A Closer Look at the Economic Mobility of U.S. Households

Percentags of Households in Each Earnings Quintile in 1984 That Were in Each Earnings Quintile in 1989

1989 Quintile Type of Household 1984 Quintile

1st

2nd

3rd

4th

5th

With Positive Earnings in Both 1984 and 1989

1st 2nd 3rd 4th 5th

58.8 20.2 9.7 7.7 3.6

25.1 45.6 20.2 6.1 2.9

9.0 21.6 40.4 20.0 9.0

5.1 8.6 21.9 45.9 18.4

2.0 4.0 7.8 20.4 66.1

With Heads 3545 Years Old in 1984

1st 2nd 3rd 4th 5th

63.3 23.6 4.7 6.9 1.1

27.2 44.3 16.7 8.1 4.0

4.0 22.3 47.0 20.2 6.4

3.3 7.3 25.1 44.6 19.1

2.3 2.4 6.6 20.1 69.3

Source: 1984, 1985, 1989, and 1990 Panel Study of Income Dynamics

Table 9

Other Dimensions of U.S. Inequality

Breakdown of U.S. Household 1992 Sample by Characteristics of Household Head*

Average Level (1992 $)

Characteristic Age 25 and Under 2630 3135 3640 4145 4650 5155 5660 6165 Over 65 Earnings Income Wealth

Concentration (Gini Index)

Earnings Income Wealth Labor

Source of Income (%)

Capital Business Transfers Other

% of Sample

Average Household Size

(Number of People)

16,210 29,937 39,164 47,123 48,367 52,301 49,207 43,352 29,722 4,927

18,908 34,009 47,701 54,618 58,616 62,914 63,884 57,411 53,119 28,442

26,207 35,732 76,060 102,234 187,820 254,922 299,256 357,254 300,240 251,850

.528 .410 .466 .542 .506 .473 .509 .613 .793 1.032

.471 .418 .494 .555 .513 .499 .550 .609 .679 .611

.808 .734 .755 .719 .753 .753 .755 .751 .744 .725

84.0 86.4 75.0 66.4 71.4 74.9 71.3 67.0 45.4 12.5

1.7 1.7 3.2 3.3 8.3 9.1 10.0 14.3 14.8 26.8

2.0 1.9 8.2 23.0 12.8 9.5 6.6 9.9 12.2 5.5

6.4 2.6 3.1 2.4 4.0 3.0 2.7 4.7 15.8 43.4

5.8 7.3 10.5 4.9 3.4 3.5 9.3 4.1 11.8 11.7

6.8 9.7 12.1 11.4 10.6 8.6 7.0 6.3 7.3 20.3

2.23 2.44 3.12 3.02 3.12 2.94 2.13 2.08 1.86 1.51

Employment Status Worker 41,247 Self-Employed 64,429 Retired 10,438 Nonworker 9,491 Education College High School No High School

48,532 90,483 35,714 18,386

123,958 580,934 228,269 72,363

.439 .606 .955 .786

.467 .618 .653 .563

.740 .758 .689 .818

83.0 45.5 13.5 50.3

5.4 15.9 19.1 10.7

2.3 29.8 18.2 1.5

3.0 3.2 35.1 30.9

6.3 5.6 14.0 6.6

54.9 10.9 18.1 16.1

2.67 2.71 1.62 2.23

60,231 27,225 10,236

81,188 36,694 20,146

353,270 136,923 68,275

.564 .554 .733

.556 .485 .551

.764 .734 .752

61.0 69.5 46.6

11.8 8.2 7.2

15.3 5.4 4.9

4.7 10.9 28.2

7.2 5.9 13.1

28.6 50.2 21.2

2.50 2.41 2.31

Marital Status Married 46,580 Single Without Dependents 15,308 With Dependents 13,653 Single Without Dependents Male Female Single With Dependents Male Female Total Sample

61,692 26,306 20,186

249,398 113,063 51,426

.545 .729 .583

.522 .589 .474

.759 .760 .803

65.3 53.0 64.8

9.6 12.8 5.5

11.8 6.0 3.3

6.7 17.3 22.0

6.6 10.9 4.4

57.4 31.2 11.4

3.09 1.00 2.90

21,365 10,984

33,696 21,030

125,897 103,899

.690 .745

.625 .534

.805 .717

55.7 50.0

11.2 14.6

8.9 2.6

11.8 23.6

12.3 9.2

13.0 18.2

1.00 1.00

21,125 11,991 33,074

25,491 19,006 45,924

85,757 43,790 184,308

.451 .609 .628

.396 .487 .573

.754 .811 .781

75.4 61.6 63.1

6.2 5.3 9.9

8.6 1.8 10.4

9.8 25.7 9.4

.0 5.7 7.2

2.1 9.3 100.0

2.66 2.96 2.41

*Percentages may not sum to 100% due to rounding. Source: 1992 Survey of Consumer Finances