Number B-6 in Series, "New Federalism: National Survey of America's Families"
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
Nearly one out of every five American children
is poor, making children twice as likely to be poor as adults in this country. This brief 1 examines the disparity in resources available to children in different types of families and in different parts of the country using data from the 1997 National Survey of America's Families
(NSAF).2
Economic well-being among children varies substantially, according to the
NSAF, both by type of family and by state.3 For example, on average children living with both parents have more than twice as much income available to them
as children living with only one parent. Additionally, the average
income available to children in a high-income state, such as New Jersey, is twice as high as that in a low-income state, such as Mississippi. However, most (five-sixths) of the income inequality among children comes not from family type or state of residence but from differences in the income-producing ability of the adults with whom they live.
We also find that states with high child poverty rates also tend to have high levels of inequality. This suggests that even "low-income" states—which also tend to have high child poverty rates—may have a substantial number of higher-income families and thus have additional resources within their borders that can be brought to bear on the needs of their poorest children.
Family Income Available to Children
When measuring family income available to children, the first question is, Whose income should be counted? In calculating family income, the Census Bureau includes the incomes of all family members living together who are legally related to one another through blood, marriage, or adoption. But this may be too narrow a definition to capture the complexity of today's families. Many children live in families that are also home to adults who are not legal relatives. For this reason, except
where specifically noted, the discussion here focuses on a wider definition of family—the social family. Social families include all legal family members who live with the child, plus all parental partners who are not formally married and any other nonfamily housemates
whose resources are presumably available to the children with whom they live.4
Since the income available to a child depends not only on the dollars of income but also on the size of the family (more members mean fewer resources available to each), an additional adjustment needs to be made to obtain a more complete income picture for children. An easy
way to adjust for family size is to convert cash income into a measure that expresses a family's income as a proportion of the official federal poverty line for a family of that size—called the
income-to-needs ratio. A family with income exactly at the poverty line, for example, is at 100 percent of poverty and has an income-to-needs ratio of 1.00.
Table 1 shows how the median family incomes available to children differ by type
of family and living arrangement. The most conspicuous finding is the enormous income disparity between children in two-parent families and children in one-parent families. For the legal definition of family (as defined by the Census Bureau), children in two-parent families have almost three times the resources available to them as do children in one-parent families. For the social definition, children in two-parent families still have over two and one-half times the family resources available to them as do children in one-parent families.
Whether the focus is on legal or social families makes little difference to median family income for two-parent families ($49,401 versus $49,556). For single-parent families, however, the difference is almost $2,000 a year ($17,328 versus $19,200), highlighting the importance of nonparental income sources for the well-being of these children.
Adjusting for family size reduces family income disparities somewhat, but does not change the basic picture. For two-parent families, the definition of family makes little difference to the median family income-to-needs ratio (2.84 for legal families versus 2.87 for social families). For single-parent families, the family income available to children, as before, is greater for social than for legal families (1.27 versus 1.15). And children in two-parent families still have much more income available to them than do their counterparts in one-parent families when family size is taken into account.
Table 1
Median Family Income Available to Children, by Living Arrangement, 1996
|
| |
Legal Families
|
Social Families
|
| |
Income |
Income-to-Needs Ratio* |
Income |
Income-to-Needs Ratio* |
| All Children |
$39,000 |
2.36 |
$40,000 |
2.39 |
| Children in Two-Parent Families |
49,401 |
2.84 |
49,556 |
2.87 |
| Children in One-Parent Families |
17,328 |
1.15 |
19,200 |
1.27 |
Source: Urban Institute tabulations from the 1997 National Survey of AmericaŠs Families.
* See text for definition. |
The sources of support available to children also vary considerably by their living arrangements. Figure 1 shows that, on average, over three-quarters of the resources available to all children come from parental earnings. For children in two-parent families, income from parental earnings accounts for over 90 percent of the total. For children in one-parent families, the proportion is considerably lower, but still more than half. Public transfers and social insurance make up a greater proportion of income for children in one-parent families (23.6 percent versus 3.9 percent for children in two-parent families). Similarly, non-parental earnings and private transfers, such as cash from extended family or friends, represent larger shares of income for children in one-parent families than for children in two-parent families.
Inequality in Family Incomes Available to Children across and within States
Child economic well-being varies widely by state. Table 2 lists the 13 states that are a particular focus of the NSAF in descending order of median family income available to children. States in the South tend to have lower median incomes than states in the Northeast and Midwest. Mississippi has the lowest median income available to children and New Jersey generally has the highest, with almost twice as much income available to the median child (an income-to-needs ratio of 3.27 compared with 1.67 for Mississippi).
Within states, median income available to children varies by family type. When we examine children in two-parent and one-parent families separately, we find that children in New Jersey do not universally have the highest median incomes available to them: New Jersey ranks 1st in income for children in two-parent families but only 5th for children in one-parent families. New York ranks much higher for children in two-parent families (5th) than one-parent families (10th). Wisconsin, in contrast, ranks 1st for children in one-parent families, but only 6th for children in two-parent families. Mississippi consistently ranks 13th out of 13 states.
The median family income measure used so far tells us the family income available to the child in the middle of that state's income distribution. It says nothing about how much better off some children in that state are relative to others in the same state. To gain a picture of inequality in the family incomes available to children within a state, we measure overall inequality among children using two indexes of inequality, both developed to summarize six standard measures of income inequality.5
The simple inequality index measures family income inequality among all children taken together.6 The composite inequality index takes into account income differences by living arrangement—combining the separate inequality scores for all children, children in two-parent families, and children in one-parent families.7 The two summary measures tell essentially the same story.
States vary considerably in how unequally family income available to children is distributed. Table 3 arranges the 13 focal states into high-, medium-, and low-inequality groups. California, Mississippi, New York, and Texas are high-inequality states on both indexes. Minnesota, Washington, and Wisconsin are low-inequality states on both indexes.
Table 2
State Differences in Social Family Median Income Available to Children, by Living
Arrangement (1996 Income-to-Needs Ratios)*
|
| |
All Children
|
Children in
Two-Parent Families
|
Children in
One-Parent Families
|
| |
Income-to-
Needs Ratio |
Descending
Order |
Income-to-
Needs Ratio |
Descending
Order |
Income-to-
Needs Ratio |
Descending
Order |
| United States |
2.39 |
|
2.87 |
|
1.27 |
|
| |
| New Jersey |
3.27 |
(1) |
3.85 |
(1) |
1.44 |
(5) |
| Massachusetts |
3.00 |
(2) |
3.60 |
(2) |
1.42 |
(6) |
| Minnesota |
2.87 |
(3) |
3.20 |
(3) |
1.67 |
(2) |
| Wisconsin |
2.76 |
(4) |
3.05 |
(6) |
1.70 |
(1) |
| |
| Michigan |
2.72 |
(5) |
3.14 |
(4) |
1.31 |
(7) |
| Washington |
2.65 |
(6) |
2.93 |
(8) |
1.65 |
(3) |
| Colorado |
2.61 |
(7) |
2.97 |
(7) |
1.51 |
(4) |
| New York |
2.39 |
(8) |
3.17 |
(5) |
1.05 |
(10) |
| Florida |
2.14 |
(9) |
2.77 |
(10) |
1.22 |
(8) |
| |
| Alabama |
2.08 |
(10) |
2.80 |
(9) |
0.88 |
(12) |
| California |
2.03 |
(11) |
2.60 |
(11) |
1.04 |
(11) |
| Texas |
2.02 |
(12) |
2.46 |
(12) |
1.13 |
(9) |
| Mississippi |
1.67 |
(13) |
2.48 |
(13) |
0.81 |
(13) |
Source: Urban Institute calculations from the 1997 National Survey of AmericaŠs Families.
* See text for definition.both across and within states.
|
Factors Driving Income Inequality among Children
The overall picture, therefore, is one of substantial child inequality both across and within states. The results presented so far, however, do not identify the factors driving the inequality. If resource differences across states are important, for example, policies that redress economic imbalances between states may effectively reduce inequality and aid low-income children. If differences in incomes between two- and one-parent families contribute heavily to inequality among children, policies aimed at promoting two-parent families and/or providing help specifically to one-parent families may be important. If neither state nor living arrangements account for the bulk of the inequality, the variation in the ability of adult family members to generate income is likely the main factor, and policies that raise the earning capacity of low-income families directly may be the most effective way to reduce income inequality.
An effective way to identify the relative contributions of the various factors driving child inequality involves a regression analysis, which allows the analyst to disentangle the relative importance of each factor taken by itself.8 State differences, other things being equal, account for very little (1 percent) of the overall inequality in family incomes available to children. Living arrangements account for considerably more, but still only a modest share of the total. State of residence and living arrangement together account for only one-sixth (17 percent) of the overall inequality in family incomes available to children. These results suggest that differences in adult earnings and unearned income (primarily public transfers) are the main cause of inequality in the resources available to children.9
The Link between Inequality and Child Poverty
Not surprisingly, states with lower average incomes available to children tend to have higher child poverty rates, which also vary considerably from state to state. But even states with low median incomes have both poor and rich families. If the high child poverty states are also states with high child income inequality, they may have more resources that could potentially be devoted to the needs of poor children than their median incomes would suggest.
Comparing the state ordering of income (see table 2) with the state ordering of inequality (see table 3) already suggests that the poorer states tend to be the most unequal ones. California, Mississippi, and Texas, for example, are the three poorest states in terms of median resources available to children (table 2) and are all in the high-inequality category (table 3). Minnesota and Wisconsin rank in the top four in median income and are in the low-inequality group. A formal analysis confirms this link, yielding statistical correlations between median family income and the simple and composite indexes of child inequality of 0.87 and 0.77, respectively (out of a possible maximum of 1.00 if the two were perfectly correlated and 0 if there were no link at all).
These findings suggest that, even though the high-inequality states tend to have lower average incomes available to children, they may well have resources that can be brought to bear on the needs of their poorer children. (Whether they use any additional public resources for children, of course, depends on the extent to which they also have childless persons with significant unmet needs as well as on the priority they assign to helping children versus other needy groups.)
For states that wish to reduce the amount of income inequality among children, several actions are possible. The most obvious would be for states with high inequality to increase their cash transfers to poor families with children. Other research shows, for example, that these states tend to have the least-generous welfare benefits (Moffitt 1999). But even the welfare benefits offered by high-benefit states leave families with no other income well below the poverty line. In addition, raising cash benefit levels could have the disadvantage of drawing low-income working families out of jobs and onto the welfare rolls.
An alternative to raising the incomes of low-income families is for high-inequality states to focus their resources on helping working families. A particularly effective policy to supplement the incomes of working families is the earned income tax credit, which supplements the earnings of low-income family members who work but is not available to families that have no earnings. States with high child inequality could either introduce or increase an existing earned income tax credit, thus enhancing the incomes of their poorer children while also encouraging work.
Notes
1. This brief is drawn from Acs and Gallagher (1999).
2. The first wave of the NSAF collected economic, health, and social information on 44,000 households between February and November of 1997. The survey oversamples households with incomes under 200 percent of the federal poverty level and households in each of 13 targeted states. The NSAF provides information on a nationally representative sample of the civilian, noninstitutionalized population under age 65 and their families. A second wave of this survey is being fielded in 1999. For more information, including the survey methods and data reliability, see Dean Brick et al. (1999).
3. There is an important distinction between focusing on child income inequality (that is, differences in family incomes available to children) and family inequality (that is, differences in incomes among families). In a family-based approach, for example, a wealthy family with one child and a poor family with two children would count as one rich family and one poor family. A child-based approach would count the same information as one rich child and two poor children.
4. "Income" as used here includes all earnings of adults in the family and public transfers and social insurance, private transfers, and other cash income (such as interest received by any family member). It excludes the value of near-cash benefits like food stamps. Income figures are for 1996. "All children" entries in the exhibits include children who live with neither parent as well as children who live with one or two parents.
5. Three of these are measures of inequality over the entire income distribution (the Gini coefficient, the coefficient of variation, and the variance of the natural logarithm of income). The other three are ratios that measure the distance between the income of someone at a particular point in the distribution and the income of another person at some lower point (the 50/20 percentile ratio, the 80/20 percentile ratio, and the 80/50 percentile ratio). For further explanation and detail, see Acs and Gallagher (1999).
6. This is referred to as the "all children index" in Acs and Gallagher (1999).
7. Both summary indexes are computed by assigning each state a score based on whether it ranks high (2), medium (1), or low (0) on each of the 6 separate measures of income inequality. The 6 measures are the Gini coefficient, the coefficient of variation, the variance of the natural logarithm of income, the ratio of income of the 80th percentile to the 20th percentile, the ratio between the median and the 20th percentile, and the ratio between the 80th percentile and the median. There are 6 measures for the simple index, because all children are treated as the same group. But there are 18 (3 times 6) measures in total for the composite index, because the composite index treats all children, children in two-parent families, and children in one-parent families as three distinct groups. Thus, if a state ranks high on each inequality measure, it scores 12 (or 36) on the summary measures. If it scores 0 on all 6 (or 18) inequality measures, it scores 0 on the summary measures. The state score is then divided by the maximum number of points to yield indexes that go from 0 to 1. For the ranges shown in table 3, "high" is roughly from 1.0 to 0.7, "medium" roughly from 0.6 to 0.4, and "low" roughly from 0.3 to 0.1. For more detail, see Acs and Gallagher (1999).
8. For computational convenience, this part of the analysis uses the variance of the natural logarithm of income as the measure of inequality and dollar incomes rather than income-to-needs ratios. The results presented here are based on a set of three regressions on the natural logarithm of income: In the first regression, the explanatory variables are a series of dummies indicating the children's state of residence; in the second, the explanatory variables are a series of dummies indicating children's living arrangements; and in the third, the explanatory variables are a series of dummy variables capturing both the state and living arrangements of children (e.g., one-parent family in Alabama, two-parent family in Alabama, two-parent family in Washington). For details, see Acs and Gallagher (1999).
9. The combined influence of living arrangement and state comes to slightly more than the total of the two separate influences because they influence one another to some degree.
References
Acs, Gregory, and Megan Gallagher. 1999. Sources of Support and Income Inequality among America's Children. Washington, D.C.: The Urban Institute. Assessing the New Federalism Discussion Paper No. 99-15.
Dean Brick, Pat, Genevieve Kenney, Robin McCullough-Harlin, Shruti Rajan, Fritz Scheuren, Kevin Wang, J. Michael Brick, and Pat Cunningham. 1999. 1997 NSAF Survey Methods and Data Reliability.
Washington, D.C.: The Urban Institute. Assessing the New Federalism Methodology Report No. 1.
Moffitt, Robert. 1999. "Explaining Welfare Reform: Public Choice and the Labor Market." International Tax and Public Finance 6 (3): 289-316.
About the Authors
Gregory Acs is a senior research associate in the Urban Institute's Income Benefits Policy Center. His research focuses on issues of social insurance, social welfare, and the compensation of workers. In recent work, he has studied the employment patterns of young women and the impact of disabilities on the duration of welfare receipt and the ability of welfare recipients to work.
Megan Gallagher is a research associate in the Urban Institute's Income and Benefits Policy Center. For the Assessing the New Federalism project, she conducted a case study on social services in Washington and currently is analyzing income and employment findings from the National Survey of America's Families.
