In 2001, 1.35 million children were born outside marriage. This represents 33.5 percent of all children born in the United States in that year. Children raised by never-married mothers are seven times more likely to be poor when compared to children raised in intact married families. The obvious nexus between single-parent families and child poverty has led President George W. Bush to propose a new trial program aimed at increasing child well-being and reducing child poverty by promoting healthy marriage.
Critics have rejected President Bush's proposal as illogical. They argue that increasing marriage would not significantly reduce child poverty for two reasons: first, that there is a substantial shortage of suitable males for single mothers to marry, and second, that even if single mothers married the father of their children, the earnings of the fathers are so low that they would not lift the family out of poverty.
However, new light has been shed on the status of non-married parents though the recent Fragile Families and Child Well-Being Study. The Fragile Families survey is a nationwide effort to collect data on both married and non-married parents at the time of a child's birth. The survey reveals that most of the claims about marriage and non-married fathers made by the opponents of the Bush "healthy marriage" proposal are wildly inaccurate.
The Fragile Families Study shows the following:
- The median age of non-married mothers is 22 at the time of birth of the child.
- Nearly three-quarters of non-married mothers are in a relatively stable romantic relationship with the expectant father at around the time of birth of their child.
- The expectant non-married fathers who have a romantic involvement with the mother-to-be are quite "marriageable." Very few have drug, alcohol, or physical abuse problems.
- On average, the earnings of non-married expectant fathers are higher than the earnings of expectant mothers in the year before the child's birth.
- The median annual earnings of non-married fathers are approximately $17,500 per year.
In this CDA Report, the Fragile Families data are used to calculate how much marriage could reduce poverty among couples who are not married at the time of the child's birth. This analysis finds that marriage would dramatically reduce poverty among the non-married mothers who are romantically involved with the fathers at the time of the child's birth.
Specifically, if these mothers do not marry but remain single, about 55 percent will be poor. By contrast, if all the mothers married the child's father, the poverty rate would fall to less than 17 percent. Thus, on average, marriage would reduce the odds that a mother and a child will live in poverty by more than 70 percent.
The contention, made by critics of the President's marriage-strengthening policy, that increased marriage will not reduce child poverty because fathers do not earn enough to lift a family out of poverty is inaccurate. While marriage of mothers and fathers would not eliminate child poverty in every case, in the overwhelming majority of cases, marriage would lift families out of poverty. Overall, the insights culled from the Fragile Families dataset and described in this CDA Report strongly indicate that a policy aimed at promoting healthy marriage among young parents has enormous potential to reduce child poverty.
The data used in this analysis are taken from the Fragile Families and Child Well-Being Study, developed jointly by Princeton University's Center for Research on Child Well-Being and Columbia University's Social Indicators Survey Center. The Fragile Families Study provides the best data available on the characteristics of non-married parents around the time of a child's birth.
These couples are of particular public policy interest since they are likely to be a high-priority target group for President Bush's proposed program to promote healthy marriage. Thus, the non-married parents in the Fragile Families survey are an excellent population for assessing the potential economic consequences of increasing marriage.
Characteristics of Non-Married Parents
Some 38 percent of the mothers in the Fragile Families Study were not married at the time of their child's birth. Popular opinion sees out-of-wedlock childbearing as occurring mainly to young girls of high-school age who lack stable relationships with their child's father. This perception is erroneous. The median age for mothers who give birth outside marriage is 22.
Nor are non-married mothers alone and isolated at the time of birth. As Table 1 shows, nearly 50 percent of these mothers are cohabiting with the expectant father at around the time of the child's birth. Another 23 percent describe themselves as "romantically involved" with the father, although the couple is not cohabiting.
The characteristics of non-married fathers who are cohabiting or romantically involved with the mother are generally more favorable than the popular stereotype. (See Table 2.) Around 67 percent of the fathers have at least a high-school degree. Some 97 percent were employed during the prior year, and 82 percent were employed at the time of the child's birth. The median annual income of these romantically involved/cohabiting fathers was between $15,000 and $20,000.
Among romantically involved or cohabiting couples, physical abuse is rare: A full 98 percent of the women in this group report that the father has never slapped them when angry. While some fathers do have drug and alcohol problems, the level is less than might be expected: Around 12 percent of the mothers report arguing with their boyfriends about a drug or alcohol problem in the last month; 2.5 percent report that drugs or alcohol impede the boyfriend's ability to hold a job.
On average, the non-married expectant fathers have higher earnings than the expectant mothers in the year before the child's birth. The median wage rate of fathers is $8.55 per hour, compared to $7.00 per hour for the mothers.
Nearly all couples that are romantically involved or cohabiting are interested in developing a long-term, stable relationship. Some 95 percent believe that there is at least a 50/50 chance they will marry in the future.
The purpose of this CDA Report is to calculate the reduction in poverty that would occur if non-married women married the fathers of their new children around the time of the child's birth. As shown in Table 1, some non-married pregnant women do not have positive and stable relationships with their child's father. In these cases, marriage is not, for the most part, a reasonable option. Therefore, we have restricted our initial marriage simulation to the 73 percent of non-married couples who were cohabiting or romantically involved but living apart at the time of their child's birth. We shall henceforth refer to these couples as the "marriageable group."
To determine the impact of marriage on the poverty of children and mothers, we first estimate what the poverty rate of the mothers would be if they remained single. We then calculate what the poverty rate would be if the mother and father marry. The difference between the poverty rate of the mothers when single and the rate for mothers when married demonstrates the potential for marriage to reduce child poverty and maternal poverty.
Employment and Earnings
The Fragile Families survey contains data on the annual earnings of new fathers during the year in which the child was born. We employ these annual earnings figures in our analysis. The study also provides annual earnings for mothers in the year before birth. However, women's participation in the labor force may be altered significantly by the birth of a child. Because of this, the paper estimates mothers' post-birth earnings based on a range of assumptions concerning the hours of employment.
Specifically, we have calculated the effect of marriage on poverty according to three separate scenarios relating to the mothers' employment after the child's birth.
Scenario #1: The mother has zero annual employment after the birth.
- Scenario #2: The mother is employed part-time for a total of 1,000 hours per year after the birth.
- Scenario #3: The mother is employed full-time throughout the year after the birth for a total of 2,000 hours.
In each scenario, the annual earnings of the fathers are assumed to be the same as the earnings in the year before the child's birth. The annual earnings of the mother are derived by multiplying the mother's hourly wage rate by the specified hours worked. In each scenario, the employment and earnings of a mother are assumed to be unchanged by marriage; that is, the mother is assumed to earn the same amount when married as when single.
The simulation assumes that single mothers will be eligible for TANF (Temporary Assistance for Needy Families), EITC (Earned Income Tax Credit), and food stamps. The level of benefits that a single mother would receive from each program is determined by the number of children in the family and the mother's annual earnings. Simulations for married couples assume that they are eligible only for food stamps and the EITC. The couple's earnings and family size determine the value of benefits. It is assumed that no married couples will receive TANF benefits.
Results of the Marriage Simulation
Under each scenario, we calculate the percentage of mothers who would be poor if they lived as single parents and the percentage who would be poor if they were married to the child's father.
Scenario #1: The mother is unemployed
Chart 1 shows the impact of marriage on maternal and child poverty under Scenario #1. In this scenario, the mothers are not employed after the birth of the child. When single, the mothers are solely dependent on welfare (TANF and food stamps). When married, the mothers are solely dependent on the father's earnings plus EITC and food stamps.
As Chart 1 shows, if mothers remain single and unemployed, they will be poor 100 percent of the time. This is because welfare benefits alone rarely, if ever, provide enough income to raise a family above the poverty level. By contrast, if the mother marries the child's father, the poverty rate drops dramatically to 35 percent. In other words, nearly two-thirds of the non-married fathers within the marriageable group earn enough by themselves to support a family above poverty without any employment on the part of the mother.
Under the conditions of Scenario #1, marriage more than doubles the family income of mothers and children. If unmarried, the mothers would have a median income of around $8,800. Marriage would raise the mothers' median family income by over $11,000 to $20,226. (See Table 3.)
Scenario #2: The mother is employed part-time
Chart 2 shows the impact of marriage on child poverty under Scenario #2. In this scenario, mothers are assumed to be employed part-time for a total of 1,000 hours per year after the birth of their child. This scenario closely matches the employment rates of single mothers with young children as reported by the U.S. Bureau of the Census. Thus, it is the most realistic of the three scenarios.
Single mothers are assumed to receive income from earnings, EITC, food stamps, and, in some cases, TANF. Married couples are assumed to receive income from earnings, EITC, and food stamps. In this scenario, mothers are assumed to work 1,000 hours per year, whether single or married.
As Chart 2 shows, 55 percent of the mothers in the Fragile Families Study will live in poverty if they remain single and are employed part-time. By contrast, if the mothers marry, their poverty rate plummets to 17 percent. In other words, the father's normal earnings, combined with the part-time earnings of the mother, are sufficient to raise 83 percent of the families above the poverty line.
Under conditions of part-time maternal employment in Scenario #2, marriage increases family income by 75 percent. If unmarried, mothers would have a median income of around $13,500. Marriage would raise the mothers' median family income by around $10,000 to a level of $23,700.
Marriage combined with part-time maternal employment not only raises nearly all families above poverty, but in many cases also raises family income well above the poverty level. For example, under Scenario #2, less than 4 percent of single mothers would have family incomes above 150 percent of the poverty level. By contrast, about 46 percent of married couples would have an income above 150 percent of the poverty level.
Scenario #3: The mother is employed full-time
Full-time/full-year employment is very effective in reducing poverty among single mothers. Some 90 percent of single mothers could maintain their families above poverty if they worked full-time throughout the year. (Full-time/full-year employment is equivalent 2,000 annual hours of employment or 40 hours per week for 50 weeks.) Census Bureau data reveal that approximately 30 percent of single mothers with children under four are employed 2,000 hours or more per year.
Since very few single mothers who were employed full-time/full-year would remain poor, marriage has little effect in reducing poverty in this scenario. (Nearly 96 percent of married couples would have incomes above the poverty level, compared to 90 percent of single mothers.) However, marriage would raise the family incomes of many full-time working mothers well above poverty and into middle-class levels.
Full-time working mothers would have a median income of around $17,500 per year. If these mothers married their child's father, median family income would rise to $29,000 per year. As Chart 4 shows, nearly two-thirds of these married couples would have incomes above 150 percent of the poverty level. By contrast, only 20 percent of full-time working single mothers would have incomes above that level.
Summary of Results 
Tables 3 and 4 summarize the results of the three scenarios for marriageable couples. As Table 3 shows, marriage would increase median family income of mothers in the study by between $10,200 and $11,400 per year. (The increase in median family income is less than the median annual earnings of the fathers-$17,500-because marriage entails an offsetting loss of welfare benefits for the mother.)
Table 4 summarizes the impact of marriage on poverty. In each scenario, marriage reduces the probability that mothers will live in poverty by at least two-thirds. Marriage would lift the incomes of many mothers above 150 percent of the poverty level. In Scenario #1, some 30 percent of married families would have incomes above 150 percent of the poverty level. In Scenario #3, nearly two-thirds of married families would have incomes above that level.
Alternative Marriage Simulation for All Non-married Couples
As noted, the marriage simulation data presented in Charts 1, 2, and 3 and Tables 3 and 4 pertain to the "marriageable" couples within the Fragile Families survey-i.e., those who, at the time of the child's birth, are cohabiting or living separately but are still romantically involved. These couples represent 73 percent of all non-married couples at the time of a child's birth.
Expectant mothers and fathers in the marriageable group have somewhat higher earnings than other non-married couples in the year before their child's birth. Therefore, marriage may have a substantially greater effect in reducing poverty among the marriageable group than among non-married couples in general.
To investigate that possibility, the marriage simulation was rerun for all non-married couples in the Fragile Families survey; thus, the new simulation included both the "marriageable" and "non-marriageable" couples. The results of this expanded simulation were extremely similar to those for the marriageable subset. For example, under Scenario #2, 56.5 percent of mothers will be poor, if unmarried, compared to 18.4 percent of mothers, if married. (For the marriageable sub-group, the figures were 55.1 percent and 16.9 percent, respectively.) The complete results of the expanded simulation for all non-married couples are shown in Appendix Table A.
Each year, more than 1.3 million children in the United States are born outside marriage. This represents 33.5 percent of all births. The Fragile Family survey shows that in 73 percent of out-of-wedlock births, the mother and father are romantically involved and have a relatively stable relationship.
Nearly half of non-married expectant mothers are cohabiting with the father at around the time of their child's birth. Overall, some 95 percent of non-married mothers express positive attitudes about marrying their new baby's father in the future. Yet only 9 percent of couples will actually marry within a year after their child's birth. Within a few years, the relationships of most of the non-married parents will deteriorate and the mother and father will split up.
As a new strategy for reducing child poverty and improving child well-being, President George W. Bush has proposed a new pilot program to promote healthy marriage. A principal target population of the President's proposed program would be romantically involved non-married couples at or around the "magic moment" of a child's birth. This target group is the precise population analyzed in this CDA Report.
Participation in the President's marriage program would be voluntary. The program would seek to increase healthy marriage by providing target couples with:
Accurate information on the value of marriage in the lives of men, women, and children;
Marriage-skills education that will enable couples to reduce conflict and increase the happiness and longevity of their relationship; and
Experimental reductions in the current financial penalties against marriage that are contained in all federal welfare programs.
The programs would utilize existing marriage-skills education programs that have proven effective in decreasing conflict and increasing happiness and stability among couples. The pro-marriage initiative would not seek merely to increase marriage rates among target couples, but would provide ongoing support to help at-risk couples maintain healthy marriages over the long term.
The President proposes to spend $300 million per year on his pilot program to promote healthy marriage. This modest sum represents spending only one cent to promote healthy marriage for every five dollars the government currently spends subsidizing single-parent families.
Critics of President Bush's proposal have charged that increasing the number of healthy marriages would not reduce child and maternal poverty. These claims are false and misleading.
For example, in a widely publicized paper entitled "Let Them Eat Wedding Rings: The Role of Marriage Promotion in Welfare Reform," an organization called Alternatives to Marriage asserted that "Marriage is not an effective solution to poverty." In fact, the actual study cited in the paper shows the opposite: Marriage would eliminate poverty for the majority of poor single mothers surveyed. Nevertheless, Alternatives to Marriage argues that marriage is not "an effective solution to poverty" because marriage would not eliminate poverty in every instance. The error of such an argument needs no further elaboration.
The analysis presented in this paper shows that marriage has an enormous potential to reduce poverty among couples who are unmarried at the time of their child's birth. In general, a 10 percent increase in the marriage rate of poor single mothers would reduce poverty among that group by 7 percentage points.
Increasing the number of healthy marriages would also have substantial non-economic benefits for children. Children who are raised in marriage by their biological mother and father are dramatically less likely to have emotional and behavioral problems, to be physically abused, to become involved in crime, to fail in school, to abuse drugs, and to end up on welfare as adults.
The erosion of marriage and the increase in single-parent families are major causes of child poverty and welfare dependence in the United States. Nearly three-quarters of government means-tested welfare aid to children goes to single-parent families. Over 80 percent of long-term child poverty occurs in broken or never-married families.
There is a widespread misconception that single mothers have little contact with the fathers of their children. In reality, surveys show that most non-married expectant mothers are romantically involved with their child's father at around the time of the child's birth. Most of these couples express positive attitudes about marriage and hope to become married in the future. Yet relatively few will, in fact, marry. Most will split apart a few years after the child's birth.
President George W. Bush has proposed a pilot program aimed at promoting healthy marriage, especially in low-income communities. A key target group for this policy would be non-married mothers and fathers around the time of the "magic moment" of a child's birth. This CDA Report demonstrates that policies to increase marriage among these parents could have a very large impact in reducing child poverty. In general, a 10 percent increase in marriage among poor single mothers would reduce child poverty within that group by 7 percentage points.
Healthy marriage is critical to the well-being of children, women, and men. President Bush's marriage-strengthening initiative should therefore be an essential part of any future welfare policy.
Robert E. Rector is Senior Research Fellow, Kirk A. Johnson, Ph.D., is Harry and Jeanette Weinberg Fellow in Statistical Welfare Research in the Center for Data Analysis, Patrick F. Fagan is William H. G. Fitzgerald Research Fellow in Family and Cultural Issues, and Lauren R. Noyes is Director of Research Projects at The Heritage Foundation.
As noted in the text, this CDA Report is based on data from the Fragile Families database (sometimes called the Survey of New Parents) conducted by the Center for Research on Child Well-Being at Princeton University and the Social Indicators Survey Center at Columbia University. The survey poses questions to roughly 3,500 families in order to gauge a nationally representative sample of parents, especially with regard to the nature of the relationship between unwed mothers and fathers.
In that respect, this is an especially useful database with which to simulate the effects of marriage on child poverty, especially given that the period of time around a child's birth is seen as a "magic moment" where unwed parents may decide to get married.
The data employ a national population weight that is designed to estimate properly the number of births in major U.S. cities (those with populations of at least 200,000). On a weighted basis, these data represent some 1.2 million babies; however, since the majority of births in America are to married couples, only about 38 percent of the births in the Fragile Families survey are to unwed mothers.
Of the out-of-wedlock births, nearly 73 percent ofchildren are born to parents who are romantically involved with each other. (See Table 1, above.) Clearly, marriage is likely to fail if there is not a romantic involvement between the mother and the father. For that reason, only mothers who self-report a current romantic involvement with the father of the child are included in the analysis.
To conduct the analysis properly, data are needed both on the number of the mother's children who live with her (if this is not her first birth) and on earnings for the mother (for her last paying job before any maternity leave) and the father. In some cases, surveys were completed for the mother but not the father, rendering those observations unusable. Further, some surveys were completed, but questions relating to earnings or income were not reported either by the mother or by the father. Because of these data limitations, this analysis includes only cases for just over 225,000 children, on a weighted basis. On an unweighted basis, the simulation includes nearly 1,250 observations, or about two-thirds of the "romantically involved but not married" subset of the survey.
The analysis simulates the child poverty rates these mothers are likely to experience if they remain single versus their poverty rates if they married the child's father. Income for the father is assumed to remain unchanged from the last year reported in the survey. Income for the mother is based on three core scenarios:
The mother does not work.
- Scenario #2:
The mother works 1,000 hours per year (or an average of 20 hours per week over the course of 50 weeks per year).
- Scenario #3:
The mother works 2,000 hours per year (or an average of 40 hours per week over the course of 50 weeks per year).
The Fragile Families data allow an hourly earnings figure to be computed for the mother to facilitate this analysis. In a few rare cases, the computed hourly income figure is less than the statutory federal minimum wage of $5.15 per hour. In those cases, the hourly rate is set at $5.15 per hour and the analysis is continued. The poverty rates are then calculated based on two family outcomes:
- Outcome A:
The mother does not marry the father of her new baby (and does not marry anyone else).
- Outcome B:
The mother does marry the father of her new baby.
One of the valid criticisms of the simple use of income as the basis for poverty determination is that it ignores program benefits and tax effects. When low-income families receive food stamps, for instance, their ability to consume increases and their economic situation is thereby improved. Because of this, income is adjusted to take into consideration the following four factors:
Cash welfare benefits are added to any income of single women (in this simulation, no married couples may receive TANF and, operationally, few married couples qualify for cash TANF benefits). Since actual TANF benefits do vary from state to state, the reasonable median-benefit state of Kansas is used in the simulation. In Kansas, a single mother with two children who has no income would have received $429 per month (or $5,148 per year) in benefits in 1999. TANF benefits are assumed to fall by 50 cents for every dollar earned. (Put another way, the TANF "disregard rate" is set at 50 percent.) If a single mother earns more than $10,296, she will not receive any TANF benefits.
- Food Stamps
Food stamps are calculated on the basis of the 1999 formula benefit levels. In 1999, a family of three could receive a maximum of $329 in food stamps per month. Since the food stamp program counts TANF as income for purposes of benefit calculation, any TANF is counted against food stamp eligibility. This simulation assumes that if families qualify for food stamps and/or TANF, they will apply for and receive benefits under current law formulas.
- Earned Income Tax Credit
If a family (whether married or not) has earnings, the EITC is calculated and included as income for purposes of poverty determination. The EITC has no interaction with either TANF or food stamps.
- Payroll (FICA) Taxes
Payroll taxes on earnings (employee side only) are subtracted from income as a last step before determining poverty rates.
Since initial interviews for the Fragile Families survey took place over the course of three years in the late 1990s and 2000, the data from the midpoint year of 1999 were chosen for income, program participation, and poverty calculation. The poverty thresholds are the ones published for that calendar year by the Census Bureau.
In all cases, poverty rates drop substantially with an increase in marriage. Tables 3 and 4 show how marriage would lift many of these families out of poverty-in some cases to more than 150 percent of the poverty level. (For a family of four, 150 percent of the poverty level is $25,543 per year.)
 For a detailed description of the Fragile Families and Child Well-Being Study, see http://crcw.princeton.edu/fragilefamilies. The Fragile Families Study is a survey of roughly 4,700 new and in many cases unwed parents. Information about them and their new children will be tracked over the course of five years. The analysis here deals only with the first year or "baseline" survey.
 The survey of new births was collected to garner information on the changing relationships of unwed parents. As noted above, these births will be tracked over the course of five years in order to analyze changing dynamics. This analysis deals only with the first-year "baseline" data, as the initial one-year follow-up public use data were not available when this research was completed in early 2003.
 The wage rates for the mothers are inferred, based on the last job they held, given that most of these women would be on maternity leave or another work break at the time of the survey.
 In reality, some mothers might choose to work less when married. For these mothers, marriage would produce an increase in both income and "leisure." For simplicity of presentation, this option of working less after marriage has not been included in the simulation.
 According the U.S. Census Bureau, a family is deemed "poor" if its annual income falls below specified poverty income thresholds. These thresholds vary according to family size. In 1999, the government's poverty income thresholds were $13,423 for a three-person family, $16,954 for a four-person family, and $19,882 for a five-person family. If a father and mother marry, the father is added to the count of family members and the poverty income threshold is increased accordingly; the average increase in the poverty threshold is roughly $3,000 per additional family member. Thus, if a father adds more than $3,000 in net income to a family, marriage will reduce the probability that the mother and children will be poor. See U.S. Bureau of the Census, Poverty in the United States: 1999, Current Populations Reports Series P60-210 (Washington D.C.: U.S. Government Printing Office, September 2000), p. A4.
 According to Census Bureau data, the median annual number of hours of employment for single mothers with children under four is 1,040.
 The benefits a mother would receive from EITC, food stamps, and TANF are contingent on her annual earnings, which are calculated by multiplying her hourly wage rate (as reported in the Fragile Families survey) by 1,000 hours.
 The increase in family income due to marriage is less than the median earnings level of the father because the couple would suffer a substantial reduction in welfare benefits if they marry.
 In 1999, a family of four would have an income above 150 percent of the poverty level if it had an income above $25,342. A family of three would have an income above 150 percent of the poverty level if it had an income of $20,135.
 As noted, nearly 50 percent of the non-married couples in the "marriageable group" are cohabiting at the time of the child's birth. According to Census Bureau methodology, cohabiting fathers are not considered as part of the mother's family unit. Thus, neither the father nor his earnings would be counted when determining whether the mother is poor. If a couple marries, the father and his income are included as part of the family unit. In practice, non-married cohabiting fathers are likely to contribute some income to the mother and child. This means that the government is likely to overestimate the de facto poverty rate of single mothers who are cohabiting. Consequently, the short-term impact of marriage in reducing poverty of single mothers may be somewhat overstated. On the other hand, the relationships of cohabiting parents are unstable. Such couples are likely to separate within a short period, and the mother and child will fall into true poverty. Healthy marriage-promotion policies are intended to increase not merely marriage, but also a couple's commitment and stability. By increasing the stability and longevity of the parents' relationships, marriage-promotion programs would have a substantial effect in reducing long-term poverty among mothers and children.
 Dorion Solot and Marshall Miller, "Let Them Eat Wedding Rings: The Role of Marriage Promotion in Welfare Reform," Alternatives to Marriage Project, 2002, executive summary, p. 1.
 Wendy Sigle-Rushton, "For Richer or Poorer," Working Paper 301-17 FF, Center for Research on Child Well-Being, Princeton University, 2001, Table 2. This study also employs data from the Fragile Families survey and obtains results similar to those presented in this CDA Report. For example, the study estimates that 95 percent of single mothers would escape poverty if they were married and worked part-time.
 Patrick F. Fagan, Robert E. Rector, Kirk A. Johnson, Ph.D., and America Peterson, The Positive Effects of Marriage: A Book of Charts (Washington, D.C.: The Heritage Foundation, April 2002), at http://www.heritage.org/Research/Features/Marriage/index.cfm.
 Ibid., Chart 8.
 Ibid., Chart 5.
 For an elongated critique of these income issues, see Robert E. Rector and Rea S. Hederman, Jr., "Income Inequality: How Census Data Misrepresent Income Distribution," Heritage Foundation Center for Data Analysis Report No. 99-07, September 29, 1999, at http://www.heritage.org/Research/Labor/CDA99-07.cfm.
 For the official 1999 poverty thresholds, see http://www.census.gov/hhes/poverty/threshld/thresh99.html. This analysis used the "weighted average" thresholds.