November 19, 2002

CHILD LIVING ARRANGEMENTS BY RACE AND INCOME:
A SUPPLEMENTARY ANALYSIS
by Wendell E. Primus (1)

Executive Summary

PDF of this report

Related Report:
ArrowDeclining Share of Children Lived with Single Mothers in the Late 1990s

Additional Analyses

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

This supplementary analysis to the study "Declining Share of Children Lived with Single Mothers in the Late 1990s,"(2) which the Center issued on June 15, 2001, employs an alternative methodology to provide a clearer picture of changes in living arrangements within different income groups. It also briefly discusses possible explanations for those changes.

The original study concluded that overall, children were significantly less likely to live with single mothers in 2000 than in 1995 and were more likely to live with cohabiting mothers.(3) It also found that the trend away from living with married parents(4) had stopped. The new methodology used in this report does not alter these overall results, although it does alter the findings on the magnitude of the changes that occurred.

The original study also examined changes in living arrangements among lower-income children (defined as those in families with incomes below 200 percent of the poverty line) and higher-income children (those above 200 percent of the poverty line). Between 1995 and 2000, it found, the proportion of lower-income children living with single mothers declined 1.4 percentage points and the proportion of lower-income children living with cohabiting mothers increased 1.2 percentage points. Conversely, the proportion of higher-income children living with married parents decreased 1.4 percentage points during this period.

While these results for specific income groups are mathematically accurate, they understate the changes in living arrangements in the late 1990s because of the method the study used for comparing living arrangements across income. This analysis uses a different method, one that is more illuminating for the purposes of this study: it defines the lower-income group as the bottom 39 percent of children by income and the higher-income group as the top 61 percent of children. These percentages are derived from the fact that in 2000, about 39 percent of children lived in families with incomes below 200 percent of the poverty line, while about 61 percent of children lived in families with incomes above this level.

This methodology produces results somewhat different from the original study. Specifically, for the 1995 - 2000 period, it shows:

Table 1
Percent of Children by Race In Lower and Higher Income Families
Fixed Percentage 200% of Poverty Level
  1995 2000 1995 2000
Overall  
  Lower Income 39% 39% 44% 39%
  Higher Income 61% 61% 56% 61%
White Non-Hispanic  
  Lower Income 27% 27% 32% 27%
  Higher Income 73% 73% 68% 73%
Black Non-Hispanic  
  Lower Income 61% 61% 67% 61%
  Higher Income 39% 39% 33% 39%
Hispanic  
  Lower Income 64% 64% 72% 64%
  Higher Income 36% 36% 28% 36%
  • no change in the proportion of higher-income children living with married parents (compared to a 1.4 percentage-point decline in the original study).

In general, using the new methodology, the trend away from single-mother-only families seems to be associated primarily with lower-income families. The trend away from single mothers was evident among all racial groups, but primarily confined to each group's lower-income population. Among blacks and Hispanics, there also was a significant increase between 1995 and 2000 in the percentage of children living with married parents — either natural, adoptive, or step. Only in the black community were there any significant changes in the living arrangements of children among higher-income families.

The reason why the new methodology produces different results is the growth in income between 1995 and 2000. The proportion of children living below 200 percent of the poverty line fell from 44 percent to 39 percent during this period. In the original study, therefore, children in the 40th through 44th percentiles by income were included in the lower-income category in 1995 but were in the higher-income category in 2000. Since children living in higher-income families are less likely to live with single mothers than are lower-income children, the shift of relatively high-income children out of the lower-income category effectively increased the percentage of lower-income children living with single mothers, partially obscuring the trend away from single mothers in this income group. This supplementary analysis, by using a fixed percentage of children rather than a fixed level of income as the dividing line between income categories, avoids this flaw.

Table 2
Comparison of the Change in Child Living Arrangements from 1995 to 2000 by Income,
Using 200 Percent of the Poverty Level and Fixed Percentage Analysis
Fixed Percentage 200 % of Poverty Level
Overall 1995 2000 1995 - 2000 1995 2000 1995 - 2000
  Married Parents 69.5% 70.1% 0.6% 69.5% 70.1% 0.6%
  Single Mother 9.9% 18.4% -1.5% 19.9% 18.4% -1.5%
  Cohabiting Mother 2.6% 3.0% 0.4% 2.6% 3.0% 0.4%
Lower Income Bottom 39% (Fixed) Below 200% of Poverty
  Married Parents 48.3% 50.5% 2.2% 50.8% 50.4% 0.4%
  Single Mother 36.6% 32.7% -3.9% 34.2% 32.8% -1.4%
  Cohabiting Mother 4.8% 6.0% 1.2% 4.8% 6.0% 1.2%
Higher Income Top 61% (Fixed) Above 200% of Poverty
  Married Parents 83.8% 83.7% -0.1% 85.0% 83.6% -1.4%
  Single Mother 9.5% 9.5% 0.0% 9.0% 9.5% 0.5%
  Cohabiting Mother 1.2% 1.1% -0.1% 0.9% 1.1% 0.2%
Bold type indicates statistical significance at the 90 percent confidence level.
 

Possible explanations for the shift away from single-mother families include the decline in teen birth rates, changing attitudes toward marriage, a strong economy, and policy changes such as welfare reform and the strengthening of the child support enforcement program. However, it is difficult to weigh these factors or to judge the impact of changing living arrangements upon children's well-being.

Methodology(5)

There are at least two ways to compare living arrangements across income. (See Table 1). The first is to compare populations below and above a given income level over time. The original paper uses this method: "lower-income" children are defined as those in families with incomes below 200 percent of the poverty line and "higher-income" children as those above 200 percent of the poverty line. While the relative sizes of the two groups will change over time to reflect the changing number of families above and below the income threshold, the threshold itself remains constant. Using this approach, 44 percent of all children were below 200 percent of poverty in 1995 compared to 39 percent in 2000.

Table 3
Living Arrangements of All Children,
Based on a Fixed Percentage of Children Each Year
1985 1990 1995 2000 1985 - 1990 1995 - 2000
Overall
  Married Parents 74.3% 73.1% 69.5% 70.1% -1.2% 0.6%
  Single Mother 18.8% 18.9% 19.9% 18.4% 0.1% -1.5%
  Cohabiting Mother 1.5% 1.9% 2.6% 3.0% 0.4% 0.4%
 
Lower Income (Poorest 39 Percent)
  Married Parents 55.1% 52.9% 48.3% 50.5% -2.2% 2.2%
  Single Mother 35.0% 35.3% 36.6% 32.7% 0.3% -3.9%
  Cohabiting Mother 3.0% 3.7% 4.8% 6.0% 0.7% 1.2%
 
Higher Income (Richest 61 Percent)
  Married Parents 87.1% 86.4% 83.8% 83.7% -0.7% -0.1%
  Single Mother 8.6% 8.5% 9.5% 9.5% -0.1% 0.0%
  Cohabiting Mother 0.5% 0.7% 1.2% 1.1% 0.2% -0.1%
Bold type indicates statistical significance at the 90 percent confidence level.

Another way to compare living arrangements across income is to examine an identical percentage of children arranged by income across time. This is the method used here. For comparability to the earlier study, "lower income" is defined across all years as the percentage of children living below 200 percent of the poverty line in 2000, or the bottom 39 percent of children. Similarly, "higher income" is defined across all years as the percentage of children living above 200 percent of the poverty line in 2000, or the top 61 percent of children. With this method, the income threshold dividing the two groups changes over time, while the relative sizes of the groups remain constant.

Applying the fixed-percentage method to the data from the original study, the first two columns of Table 2 compare the bottom 39 percent and the top 61 percent of children overall in 2000 to identical percentages of children in 1995. Among children in the bottom 39 percent, those living with married parents rose by 2.2 percentage points, those with single mothers fell by 3.9 percentage points, and those with cohabiting mothers rose by 1.2 percentage points. The top 61 percent of children did not exhibit any statistically significant changes in living arrangements during this period.

These results differ from those reported in the original study, which found that between 1995 and 2000, the proportion of lower-income children living with single mothers declined 1.4 percentage points and the proportion of higher-income children living with married parents decreased 1.4 percentage points.

Despite the difference in results between the two studies, both are both mathematically accurate; the apparent discrepancy arises from the effects in the original study of the rise in incomes in the late 1990s. Between 1995 and 2000, the proportion of children living below 200 percent of the poverty line fell from 44 percent to 39 percent (See Table 1). As a result, in the original study, children in the 39th through 44th percentiles by income were included in the lower-income category in 1995 but were in the higher-income category for 2000.

Since children living in higher-income families are less likely to live with single mothers than are lower-income children, the shift of the relatively high-income children out of the lower-income category in 2000 effectively increased the percentage of lower-income children living with single mothers, partially obscuring the trend away from single mothers in this income group. This supplementary analysis, in contrast, uses a fixed percentage of children rather than a fixed level of income as the dividing line between income categories, so the rise in incomes during the late 1990s does not affect its results.

Changes in Living Arrangements by Race and Income

White, Non-Hispanic Children

Among white children overall, shifts in living arrangements between 1995 and 2000 were statistically insignificant. However, among lower-income white children, there were several statistically significant changes over this period. About 27 percent of white children lived below 200 percent of the poverty line in 2000. Comparing the poorest 27 percent of white children in 1995 and 2000, the proportion of white children living with single mothers decreased by 2.9 percentage points over this period, while the proportion living with cohabiting mothers increased by 1.7 percentage points. Both of these changes were significant. These trends contrast to a 2.9 percentage-point decrease in the proportion of lower-income white children living with married parents in the earlier 1985 - 1990 period. The living arrangements of white children in the top 73 percent of incomes did not change significantly between 1995 and 2000.

Black, Non-Hispanic Children

In contrast to white children, in recent years black children experienced statistically significant changes in living arrangements overall as well as by income subgroups. The proportion of black children overall who lived with married parents increased by 4.1 percentage points between 1995 and 2000; the proportion living with single mothers decreased by 4.0 percentage points; and the proportion living with cohabiting mothers increased by 1.3 percentage points.

Some 61 percent of black children lived below 200 percent of the poverty line in 2000. Between 1995 and 2000, the proportion of children in this low-income group living with married parents rose by 3.2 percentage points; the proportion living with single mothers fell by 4.7 percentage points; and the proportion living with cohabiting mothers rose by 2.2 percentage points.

Among the 39 percent of black children in the upper-income group, there was a sizable  although statistically insignificant decline in the proportion living with single mothers and a statistically significant increase of 5.2 percentage points in the proportion living with married parents. Black children were the only racial/ethnic group in which children in the higher-income subgroup were more likely to live with married parents in 2000 than in 1995. (It should be noted, however, that the proportion of higher-income black children living with married parents in 1995 was much lower than the proportion among other subgroups.)

WHY DOES THE NEW METHODOLOGY GET DIFFERENT
AND MORE INDICATIVE RESULTS?
AN ILLUSTRATIVE PROOF

To demonstrate how the two sets of results are both mathematically accurate and to illustrate why the fixed percentage analysis is more illuminating for the purposes of this study, consider a group of 100 children, distributed among various living arrangements as in Table A. Assume that 45 percent of the children lived below 200 percent of poverty in 1995, and, because of increasing incomes, 40 percent lived below 200 percent of the poverty line in 2000.

Table A:
Hypothetical Percentage of
Children Living with Married Parents,
Ratio of Income to Poverty Threshold
 

1995

2000

Poorest 40% of all children

20%

23%

40% to 45% (next 5%)

40%

41%

Richest 55% of all children

60%

61%

Overall

43.0%

45.8%

Analyzing the changes in the living arrangements of children in this example by poverty level and again by comparing a fixed percentage of low-income children yields different results. Table B, which is derived from Table A, shows that comparing living arrangements from 1995 to 2000 using 200 percent of poverty as the threshold between lower- and higher-income children yields a 0.8 percentage-point increase in the number of lower-income children living with married parents, and a 0.7 percentage-point decrease in the number of higher-income children living with married parents. On the other hand, fixing the threshold between lower- and higher-income children at the bottom 40 percent and top 60 percent of incomes results in a 3.0 percentage-point increase in the number of lower-income children living with married parents, and a 1.0 percentage-point increase in the number of higher-income children living with married parents. This example shows that the analysis by poverty level produces misleading results and does not reflect the actual situation depicted in Table A.

Table B:
Hypothetical Percentage of Children Living with Married Parents,
by Income Under Alternative Methodologies
 

1995

2000

Change

Overall

43.0%

45.8%

1.8%

Below 200% of Poverty

22.2%

23.0%

0.8%

Above 200% of Poverty

60.0%

59.3%

-0.7%

Poorest 40% of Children

20.0%

23.0%

3.0%

Richest 60% of Children

58.3%

59.3%

1.0%

The income increase built into this example drives the difference between the two sets of results. Using poverty level analysis rather than fixed percentage analysis, if 45 percent of children lived below 200 percent of poverty in 1995 and 40 percent did in 2000, then, all other things remaining the same, the higher-income group for 2000 would include an additional five percent of children who would have been included in the lower-income group using the 1995 threshold. In 1995, those five percent of children were better off than the rest of the lower-income group and also were the most likely of those children to live with married parents, effectively raising the percentage of lower-income children living with married parents in that year. In 2000, because of income gains, those five percent of children fall just above 200 percent of poverty and shift into the higher-income group. Within this group, they are the poorest children and the least likely to live with married parents, thus lowering the percentage of higher-income children living with married parents in 2000. In sum, these results reflect shifts in both income and living arrangements; the income shifts partially mask the changes in living arrangements.

This example, though artificially simple, shows how the results of analyzing child living arrangements can differ depending on the methodology used for sorting children into lower- and higher-income groups and that the fixed percentage method is more indicative of what actually occurred.

 

Table 4
Living Arrangements of White Non-Hispanic Children,
Based on a Fixed Percentage of Children Each Year

1985 1990 1995 2000 1985 - 1990 1995 - 2000
Overall  
  Married Parents 81.9% 81.1% 78.5% 78.2% -0.8% -0.3%
  Single Mother 12.4% 12.4% 12.8% 12.0% 0.0% -0.8%
  Cohabiting Mother 1.4% 1.7% 2.6% 2.8% 0.3% 0.2%
Lower Income (Bottom 27 Percent)
  Married Parents 64.7% 61.8% 57.8% 57.6% -2.9% -0.2%
  Single Mother 26.0% 26.9% 27.5% 24.6% 0.9% -2.9%
  Cohabiting Mother 3.7% 4.6% 6.2% 7.9% 0.9% 1.7%
Higher Income (Top 73 Percent)
  Married Parents 88.6% 88.6% 86.7% 86.6% 0.0% -0.1%
  Single Mother 7.3% 7.0% 7.4% 7.6% -0.3% 0.2%
  Cohabiting Mother 0.5% 0.7% 1.2% 1.0% 0.2% -0.2%
Bold type indicates statistical significance at the 90 percent confidence level.

Hispanic Children

Like black children, Hispanic children displayed significant changes between 1995 and 2000 both overall and by income group. Overall, the percentage of Hispanic children living with single mothers fell by a statistically significant 3.3 percentage points from 1995 to 2000. Some 64 percent of Hispanic children fell below 200 percent of the poverty line in 2000; within this group, the proportion living with married parents increased by 3.9 percentage points between 1995 and 2000, while the proportion living with single mothers decreased by 5.0 percentage points. There was no change in the percentage living with cohabiting mothers. Among the top 36 percent of Hispanic children, there were no significant changes in living arrangements.

Overall Trends

When fixed percentages are used to compare lower- and higher-income children, the trend away from children living in households with only single mothers seems to be associated primarily with lower-income families. Some 390,000 fewer lower-income children lived in single-mother-only families in 2000 than in 1995. There was a statistically significant decrease in the proportion of lower-income children in each of the three major racial subgroups living with single mothers during this period. Significantly larger percentages of lower-income minority children lived with married parents, including about 290,000 Hispanic children and 220,000 black children.

Table 5
Living Arrangements of Black Non-Hispanic Children,
Based on a Fixed Percentage of Children in Each Year
1985 1990 1995 2000 1985 - 1990 1995 - 2000
Overall  
  Married Parents 40.5% 38.6% 34.8% 38.9% -1.9% 4.1%
  Single Mother 47.1% 47.0% 47.1% 43.1% -0.1% -4.0%
  Cohabiting Mother 2.3% 2.9% 2.9% 4.2% 0.6% 1.3%
Lower Income (Bottom 61 Percent)
  Married Parents 24.8% 22.9% 20.2% 23.4% -1.9% 3.2%
  Single Mother 61.8% 60.5% 61.1% 56.4% -1.3% -4.7%
  Cohabiting Mother 3.0% 3.9% 3.6% 5.8% 0.9% 2.2%
Higher Income (Top 39 Percent)
  Married Parents 65.8% 63.9% 59.1% 64.3% -1.9% 5.2%
  Single Mother 24.9% 26.4% 26.9% 23.5% 1.5% -3.4%
  Cohabiting Mother 1.3% 1.4% 1.8% 1.8% 0.1% 0.0%
Bold type indicates statistical significance at the 90 percent confidence level.
 

The trend away from single-mother-only households was evident among all racial groups but was primarily confined to each group's lower-income population. The trend toward living with married parents was evident in black children overall and in all lower-income groups except for whites. Note that among whites, the previous trend of a drop in the proportion of low-income children living with married parents has halted. No significant changes were observed among the higher-income children in any racial/ethnic group except for the increase in the proportion of higher-income black children living with married parents, but here, too, a previous adverse trend appears to stop.

Possible Explanations

Since the publication of the Center's original child living arrangements study, the Urban Institute has released another study, "'Honey, I'm Home': Changes in Living Arrangements in the Late 1990s," by Gregory Acs and Sandi Nelson.(6) Richard Bavier also has analyzed CPS data and reached similar conclusions.(7) All in all, there seems to be considerable evidence that the living arrangements of children, particularly lower-income children, have changed.

What is not clear is why this change occurred. Our earlier study carefully avoided explanations of what caused the shifts in child living arrangements. Here we will briefly discuss some possible explanations. One should not conclude that these changes in the living arrangements of America's children is primarily due to welfare policy change, although that may well have been a contributing factor.

 
Table 6
Living Arrangements of Hispanic Children,
Based on a Fixed Percentage of Children in Each Year
1985 1990 1995 2000 1985 - 1990 1995 - 2000
Overall  
  Married Parents 68.6% 68.4% 64.2% 66.2% -0.2% 2.0%
  Single Mother 24.4% 23.5% 24.6% 21.3% -0.9% -3.3%
  Cohabiting Mother 1.1% 1.6% 2.4% 2.5% 0.5% 0.1%
Lower Income (Bottom 64 Percent)
  Married Parents 59.6% 61.7% 56.5% 60.4% 2.1% 3.9%
  Single Mother 33.1% 30.3% 31.9% 26.9% -2.8% -5.0%
  Cohabiting Mother 1.5% 2.1% 3.4% 3.4% 0.6% 0.0%
Higher Income (Top 36 Percent)
  Married Parents 85.5% 82.1% 79.2% 78.7% -4.3% -0.5%
  Single Mother 9.2% 11.6% 12.1% 12.0% 2.4% -0.1%
  Cohabiting Mother 0.3% 0.7% 0.7% 0.9% 0.4% 0.2%
Bold type indicates statistical significance at the 90 percent confidence level.
 

Teen Births

How much does the decline in teen birth rates account for the decline in children living in single-mother-only families? A large decline in the teen birth rate could affect the number of children in single-mother homes, since an average of about 75 percent of teenagers who gave birth in the 1990s were unmarried(8) and since teens who were married at the time of birth tend to have less-stable unions than couples who delay marriage and childbearing.(9)

The teen birth rate is defined as the number of women under age 20 for each 1,000 such women who give birth in a given year. This birth rate has dropped substantially over the past decade, particularly among blacks.(10) From 1990 to 1994, the average teen birth rate overall was 60.2. (In other words, for each 1,000 women under 20, some 60.2 of them gave birth that year.) The average teen birth rate dropped in the second half of the decade to 52.8, a decrease of 7.4 points or 12.3 percent. For black teenagers, the teen birth rate decrease was even steeper, from 110.8 in the first half of the 1990s to 88.4 in the second. This represented a 22.3 point or 20.2 percent decline, a substantial change for such a short period.

However, while teen birth rates decreased substantially from the early 1990s to the late 1990s, the number of teen females increased significantly during this time. As a result, the number of teen births overall did not decline nearly as much as suggested by the decline in the birth rate. This is the more relevant statistic for our study, because the number and living arrangements of children, not families, are the focus of this study.

According to National Vital Statistics Reports on births, women under age 20 gave birth to about 2.5 million babies between 1995 and 1999. This represents a decrease of only about 118,000 births (or a decline of 4.6 percent) from the 1990-1994 period. This trend probably had some effect on child living arrangements, but can account for only a small fraction of the 1.1 million fewer children living with single mothers in 2000 as compared to 1995.(11) (Nevertheless, if the teenage birth rate had remained unchanged, there would have been about 340,000 more births to teenage mothers, and the number of children living with single mothers would have been higher than it was).

Changes in the number of black teen births probably had a somewhat larger effect on child living arrangements than changes in overall teen births. The number of black teen births decreased steadily over the entire period of the 1990s. From 1995 to 1999, 665,000 babies were born to black women under age 20. This is about 100,000 fewer births to teenage mothers than occurred during the preceding five years, or a 13.1 percent decline. Between 1995 and 2000, the number of black children living with a single mother decreased by about 470,000. Relative to this number, the decrease in black teen births is probably a more significant factor.

Welfare Reform

Given the timing of the shift away from single mothers and the fact that it is associated primarily with lower-income families, many readers might assume its main cause is welfare reform. In addition, it may seem that because this shift results in more children living with two adults and fewer with a single adult, it is primarily "good." Both of these assumptions oversimplify to some extent what has taken place.

As with any demographic or economic trend, there are a number of possible explanations. In this case, they include growing income, which lessens stress in a family; more people in the labor force; changes in demographic trends, particularly the increase in children living in Hispanic homes (which can partially explain overall trends, but obviously does not explain trends among the other subgroups); changes in social views about marriage; fewer births to teenage mothers; the fatherhood movement; the overall decline in births; and changes in the age distribution of births.

Policy changes also may have played a role. These include the establishment of the Temporary Assistance to Needy Families (TANF) program and other changes in the safety net, some of which (such as food stamp cuts) reduced assistance to poor families, while others (such as expansion of the Earned Income Tax Credit and expanded health insurance coverage for low-income children) increased assistance, particularly for working families.

Another policy change, the strengthening of the child support enforcement system, also may have affected people's choices about family structure. Research has shown that strong child support enforcement reduces divorce and out-of-wedlock child bearing.(12) TANF programs generally require mothers to comply with the child support system to receive assistance, which may have had some effects on child living arrangements. In addition, the substantial recent increase in paternity establishment rates, coupled with greater child support enforcement efforts, has increased the probability that a low-income, unmarried father is involved to some extent with the child support system; this may be affecting male behavior as well.

However, it is difficult to measure and estimate how these policy changes — individually or collectively — have affected family structure decisions. It is likely that "welfare" changes can only explain a relatively modest proportion of the changes in living arrangements of children since 1995.

Table 7
Single Mother Family Living Arrangements of Children:
By Mothers' Earnings
1995 2000 Point Change
Lower-Earning Single Mothers
  Single Mother Alone 62.9% 59.4% -3.5%
  W/ Relatives 22.8% 23.0% 0.2%
  W/ Male 9.5% 12.8% 3.3%
  Other 4.8% 4.8% 0.0%
Middle-Earning Single Mothers
  Single Mother Alone 61.5% 64.1% 2.6%
  W/ Relatives 20.5% 17.3% -3.1%
  W/ Male 12.8% 14.9% 2.2%
  Other 5.2% 3.6% -1.6%
Percentage Point Differences in Changes between Lower and Middle Earning Single Mothers
  Single Mother Alone -6.1%
  W/ Relatives 3.4%
  W/ Male 1.1%
  Other 1.7%
Bold type indicates statistical significance at the 90 percent confidence level.
 

Strong Economy

Another potential factor explaining the trend away from single mother families is the strong economy between 1995 and 2000. Historically, when parents have jobs and real wages are increasing, the family is under less stress, and family breakups or separations are less frequent. Census data indicate that the rate of increase in the percentage of children living with just one parent was approximately four times greater during periods of increasing poverty relative to periods of decreasing poverty. Specifically, between 1968 and 2000, the increase in the percentage of children living with a single parent averaged 0.67 percentage points per year during years of increasing poverty but only 0.16 percentage points per year during years when poverty was falling.(13) The corollary is that the percentage of children living with married parents declines faster during periods of rising poverty relative to years when poverty is falling. Census data confirm this as well. More recent Census data illustrate that once again, the economy influences child living arrangements. As the country entered into a recession in early 2001, the number and percentage of children living with single mothers increased while the number and percentage living with two parents declined. This demonstrates that the performance of the economy is correlated with changes in the living arrangements of children.

While this analysis can not measure precisely the impact of a strong economy upon child living arrangements, it should be noted that the expansion of the 1990s differed substantially from the expansion of the 1980s. Unemployment in late 1999 reached 3.9 percent, substantially lower than the 5.0 percent unemployment rate reached in March 1989. Also, labor force participation rates were much higher in the late 1990s than in the late 1980s. Finally, real wage growth for low wage earners was substantially greater in the late 1990s than the late 1980s. Thus, the economy of the 1990s may have played a much larger role in the declining percentage of children living with single mothers than in average years.

Effect on Child Well-being

Proceeding one step further and assessing the impact of changing living arrangements upon child outcomes is extremely difficult. While the declining proportion of children living in single-mother-only families may have a positive impact upon many children, the data studied here do not directly measure child well-being or the quality or stability of family arrangements. Two other trends, however, should give us pause. Data from the National Survey of America's Families suggest that between 1997 and 1999, there was a small but statistically significant increase in the percentage of children living with neither of their biological parents.(14) While those children may be better off living with kin or within foster care homes, an increase in the size of that population cannot be labeled "good."

There also is evidence that between 1995 and 2000, children in single-mother families(15) with the lowest earnings became less likely to live only with a single mother. The table above examines, for the 1995 - 2000 period, the change in living arrangements of children who live with single mothers in two income groups: "lower-earning" (which is defined as the bottom 40 percent of children by income) and "middle-earning" (defined as the next 30 percent of children by income). Children in married parent families are excluded from this analysis. (Note: While the percentage of children in each group was held constant between 1995 and 2000, the earnings cutoff for the mothers in the lower-earning group increased significantly from $1,690 in 1995 to $8,000 in 2000 due to an overall rise in earnings and a greater proportion of mothers with earnings.)

In the lower-earning group, the proportion of children living with a single mother alone declined between 1995 and 2000, while the proportion living with a single mother and a cohabiting male increased. In the middle-earning group, the proportion of children living with a single mother alone did not change significantly during this period, while the proportion of children living with a single mother and other relatives declined.

In other words, between 1995 and 2000, the decline in the proportion of children living with a single mother alone was concentrated in the lower-earning group, while the decline in the percentage of children living with a single mother and other relatives was concentrated in the middle-earning group. These data when combined with other studies suggest that those mothers experiencing the most difficulty in adjusting to changes in the welfare system those with the lowest incomes may have coped with those income losses by living with unrelated men to a greater degree.

Census data for 1997 through 1999 suggest that about 750,000 of the lowest earning single mother families had lower disposable income (earnings plus cash assistance payments plus any food stamp or housing benefits plus the earned income tax credit less federal income and payroll taxes) than in 1995.(16) Unpublished 2001 Census data also show that the poorest 10 percent of single mother families on average had 28 percent less disposable income than in 1995, after adjusting for inflation. These families show little earnings gains and large losses of government benefits. That data in this study indicate that to cope with those income losses, a small percentage of these mothers no longer live with just their children while more live with unrelated men. Information from the National Survey of American families suggests that not all of the males who live with these mothers are the natural fathers of these children. Given child welfare studies, it is unclear whether this change is better for the children. The data also suggest that those mothers with greater earnings were somewhat less likely to live in extended families.

The data in this study thus suggest that overall trends in child living arrangements, while apparently positive, include some subtrends that may or may not be positive for children. This underscores the need to exercise caution in interpreting these trends.


End Notes:

1. The author wishes to thank Allen Dupree, Jessica Goldberg, and Kathleen Romig for their assistance in preparing this paper.

2. Allen Dupree and Wendell Primus, Declining Share of Children Lived with Single Mothers in the Late 1990s, <http://www.cbpp.org/archiveSite/6-15-01wel.htm> (Washington, DC: Center on Budget and Policy Priorities, June 2001).

3. In this analysis, we use the term "cohabiting mother" to refer to a living situation in which an unmarried mother lives with a single adult male who is not related to her and with no other adults. This definition is consistent with the Census Bureau's concept of the "Persons of the Opposite Sex Sharing Living Quarters," used in its analysis of cohabitation trends over time.

4. The term "married parents" does not mean married biological parents. Married parents also includes adoptive and step-parents. "Cohabiting mother" and "single mother" are mutually exclusive categories. The latter indicates single mothers living alone with their children or residing with relatives. The percentages in most of the tables do not total to 100 percent because they exclude children living with fathers and children not living with either of their biological parents.

5. In this analysis, cohabiting mother is defined as a mother living with her children and one or more unrelated adults, usually the mother's boyfriend. Also, the 700,000 foster children in this country are included in the breakouts by race, but not included in any of the income subdivisions in this analysis.

6. Gregory Acs and Sandi Nelson, "Honey, I'm Home": Changes in Living Arrangements in the Late 1990s (Washington, DC: Urban Institute, 2001).

7. Richard Bavier, "Recent Increases in the Share of Young Children Living with Married Mothers," Unpublished manuscript. March 2001.

8. Stephanie J. Ventura, Sally C. Curtin, and T. J. Matthews, "Variations in Teenage Birth Rates, 1991 - 98: National and State Trends," National Vital Statistics Reports, Vol. 48, No. 6 (Hyattsville, MD: Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, April 2000); Joyce A. Martin, Brady E. Hamilton, and Stephanie J. Ventura, "Births: Preliminary Data for 2000," National Vital Statistics Reports, Vol. 49, No. 5 (Hyattsville, MD: Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, July 2001).

9. Bramlett MD, Mosher WD. "First marriage dissolution, divorce, and remarriage: United States," Advance Data from Vital and Health Statistics; No. 323. Hyattsville, Maryland: National Center for Health Statistics. 2001.

10. Martin, Hamilton, and Ventura, "Births: Preliminary Data for 2000."

11. Ibid.; "Live Births by Age of Mother and Race: United States, 1933 - 98." [1985 - 98] Even though the overwhelming majority of teenage parents are single, not all are single. Furthermore, some of them marry after several years, either to the father of the child or to someone else. Thus, the decline of 118,000 births to teenage mothers does not translate into an equivalent decline in the number of children living in single mother families.

12. Nixon, Lucia A. 1997. "The Effects of Child Support Enforcement on Marital Dissolution," Journal of Human Resources, 32; Case, Anne. 1998. "The Effects of Stronger Child Support Enforcement on Non-Marital Fertility." In Irwin Garfinkel, Sara McLanahan, Daniel Meyer, and Judith A. Seltzer, (eds.), "Fathers Under Fire: The Revolution in Child Support Enforcement." New York: Russell Sage Foundation; Plotnic, Robert, Irwin Garfinkel, Daniel Gaylin, Sara McLanahan, and Inhoe Ku, 1998. "Better Child Support Enforcement: Can It Reduce Teenage Premarital Childbearing?". Paper presented at Population Association of American 1998 Annual Meeting; Blau, Francine, Lawrence Kahn, and Jane Waldfogel, 1999. "Does Welfare Influence Young Women's Marriage, Single Parenthood, and Female Headship Decisions?". Working paper, Columbia University.

13. Removing the 1993 to 2000 expansion increases the 0.16 figure only slightly.

14. Acs and Nelson, Ibid. Table 2.

15. In this section of the paper only, "single-mother families" means all children residing in with a single mother including mothers who are cohabiting.

16. Ron Haskins and Wendell Primus. "Welfare Reform and Poverty," Welfare Reform and Beyond, Policy Brief No. 4. Washington, DC: The Brookings Institution. July 2001.

Wendell Primus. Comments on "Effects of Welfare Reform on Family Income and Poverty" in Rebecca Blank and Ron Haskins, (eds.), The New World of Welfare Reform. Washington, DC: Brookings Institution Press, 2001.