Journal Issue: Children and Poverty Volume 7 Number 2 Summer/Fall 1997
Effects of Income on Child Outcomes
As illustrated in Table 1, poor children suffer higher incidences of adverse health, developmental, and other outcomes than nonpoor children. The specific dimensions of the well-being of children and youths considered in some detail in this article include (1) physical health (low birth weight, growth stunting, and lead poisoning), (2) cognitive ability (intelligence, verbal ability, and achievement test scores), (3) school achievement (years of schooling, high school completion), (4) emotional and behavioral outcomes, and (5) teenage out-of-wedlock childbearing. Other outcomes are not addressed owing to a scarcity of available research, a lack of space, and because they overlap with included outcomes.
While this review is organized around specific outcomes, it could also have been organized around the various ages of childhood.9-11 Five age groups are often distinguished—prenatal to 2 years, early childhood (ages 3 to 6), late childhood (ages 7 to 10), early adolescence (ages 11 to 15), and late adolescence (ages 16 to 19). Each age group covers one or two major transitions in a child's life, such as school entrances or exits, biological maturation, possible cognitive changes, role changes, or some combination of these. These periods are characterized by relatively universal developmental challenges that require new modes of adaptation to biological, psychological, or social changes.10
Somewhat different indicators of child and youth well-being are associated with each period. For example, grade retention is more salient in the late childhood years than in adolescence (since most schools do not hold students back once they reach eighth grade12). Furthermore, low income might influence each indicator differently. As an illustration, income has stronger effects on cognitive and verbal ability test scores than it has on indices of emotional health in the childhood years.Physical Health
Compared with nonpoor children, poor children in the United States experience diminished physical health as measured by a number of indicators of health status and outcomes (see Table 1). In the 1988 National Health Interview Survey, parents reported that poor children were only two-thirds as likely to be in excellent health and almost twice as likely to be in fair or poor health as nonpoor children. These large differences in health status between poor and nonpoor children do not reflect adjustment for potentially confounding factors (factors, other than income, that may be associated with living in poverty) nor do they distinguish between long- or short-term poverty or the timing of poverty. This section reviews research on the relationship of poverty to several key measures of child health, low birth weight and infant mortality, growth stunting, and lead poisoning. For the most part, the focus is on research that attempts to adjust for important confounding factors and/or to address the effect of the duration of poverty on child health outcomes.
Low birth weight (2,500 grams or less) and infant mortality are important indicators of child health. Low birth weight is associated with an increased likelihood of subsequent physical health and cognitive and emotional problems that can persist through childhood and adolescence. Serious physical disabilities, grade repetition, and learning disabilities are more prevalent among children who were low birth weight as infants, as are lower levels of intelligence and of math and reading achievement. Low birth weight is also the key risk factor for infant mortality (especially death within the first 28 days of life), which is a widely accepted indicator of the health and well-being of children.13
Estimating the effects of poverty alone on birth outcomes is complicated by the fact that adverse birth outcomes are more prevalent for unmarried women, those with low levels of education, and black mothers—all groups with high poverty rates. One study that used data from the NLSY to examine the relationship between family income and low birth weight did find, however, that among whites, women with family income below the federal poverty level in the year of birth were 80% more likely to have a low birth weight baby as compared with women whose family incomes were above the poverty level (this study statistically controlled for mothers' age, education, marital status, and smoking status). Further analysis also showed that the duration of poverty had an important effect; if a white woman was poor both at the time when she entered the longitudinal NLSY sample and at the time of her pregnancy (5 to 10 years later), she was more than three times more likely to deliver a low birth weight infant than a white woman who was not poor at both times. For black women in this sample, although the odds of having a low birth weight baby were twice the odds for white mothers, the probability of having a low birth weight baby was not related to family poverty status.14
Other studies that used county level data to examine the effects of income or poverty status and a number of pregnancy-related health services on birth outcomes for white and black women also found that income or poverty status had a statistically significant effect on both low birth weight and the neonatal mortality rate for whites but not for blacks.15,16
Although overt malnutrition and starvation are rare among poor children in the United States, deficits in children's nutritional status are associated with poverty. As described more fully in the Child Indicators article in this journal issue, stunting (low height for age), a measure of nutritional status, is more prevalent among poor than nonpoor children. Studies using data from the NLSY show that differentials in height for age between poor and nonpoor children are greater when long-term rather than single-year measures of poverty are used in models to predict stunting. These differentials by poverty status are large even in models that statistically control for many other family and child characteristics associated with poverty.17
Harmful effects of lead have been documented even at low levels of exposure. Health problems vary with length of exposure, intensity of lead in the environment, and the developmental stage of the child—with risks beginning prior to birth. At very young ages, lead exposure is linked to stunted growth,18 hearing loss,19 vitamin D metabolism damage, impaired blood production, and toxic effects on the kidneys.20 Additionally, even a small increase in blood lead above the Centers for Disease Control and Prevention (CDC) current intervention threshold (10 µg/dL) is associated with a decrease in intelligence quotient (IQ).21
Today, deteriorating lead-based house paint remains the primary source of lead for young children. Infants and toddlers in old housing eat the sweet-tasting paint chips and breathe the lead dust from deteriorating paint. Four to five million children reside in homes with lead levels exceeding the accepted threshold for safety,22 and more than 1.5 million children under six years of age have elevated blood lead levels.23
Using data from NHANES III (1988-1991), one study found that children's blood lead levels declined as family income increased.23 All other things being equal, mean blood lead levels were 9% lower for one- to five-year-olds in families with incomes twice the poverty level than for those who were poor. Overall blood levels were highest among one- to five-year-olds who were non-Hispanic blacks from low-income families in large central cities. The mean blood lead level for this group, 9.7 µg/dL, was just under the CDC's threshold for intervention and almost three times the mean for all one- to five-year-olds.Cognitive Abilities
As reported in Table 1, children living below the poverty threshold are 1.3 times as likely as nonpoor children to experience learning disabilities and developmental delays. Reliable measures of cognitive ability and school achievement for young children in the Children of the NLSY and IHDP data sets have been used in a number of studies to examine the relationship between cognitive ability and poverty in detail.6,24-26 This article reports on several studies that control for a number of potentially important family characteristics and attempts to distinguish between the effects of long- and short-term poverty.
A recent study using data from the Children of the NLSY and the IHDP compared children in families with incomes less than half of the poverty threshold to children in families with incomes between 1.5 and twice the poverty threshold. The poorer children scored between 6 and 13 points lower on various standardized tests of IQ, verbal ability, and achievement.25 These differences are very large from an educational perspective and were present even after controlling for maternal age, marital status, education, and ethnicity. A 6- to 13-point difference might mean, for example, the difference between being placed in a special education class or not. Children in families with incomes closer to, but still below, the poverty line also did worse than children in higher-income families, but the differences were smaller. The smallest differences appeared for the earliest (age two) measure of cognitive ability; however, the sizes of the effects were similar for children from three to eight. These findings suggest that the effects of poverty on children's cognitive development occur early.
The study also found that duration of poverty was an important factor in the lower scores of poor children on measures of cognitive ability. Children who lived in persistently poor families (defined in this study as poor over a four-year span) had scores on the various assessments six to nine points lower than children who were never poor.25 Another analysis of the NLSY that controlled for a number of important maternal and child health characteristics showed that the effects of long-term poverty (based on family income averaged over 13 years prior to testing of the child) on measures of children's cognitive ability were significantly greater than the effects of short-term poverty (measured by income in the year of observation).26
A few studies link long-term family income to cognitive ability and achievement measured during the school years. Research on children's test scores at ages seven and eight found that the effects of income on these scores were similar in size to those reported for three-year-olds.25 But research relating family income measured during adolescence on cognitive ability finds relatively smaller effects.27 As summarized in the next section, these modest effects of income on cognitive ability are consistent with literature showing modest effects of income on schooling attainment, but both sets of studies may be biased by the fact that their measurement of parental income is restricted to the child's adolescent years. It is not yet possible to make conclusive statements regarding the size of the effects of poverty on children's long-term cognitive development.School Achievement Outcomes
Educational attainment is well recognized as a powerful predictor of experiences in later life. A comprehensive review of the relationship between parental income and school attainment, published in 1994, concluded that poverty limited school achievement but that the effect of income on the number of school years completed was small.28 In general, the studies suggested that a 10% increase in family income is associated with a 0.2% to 2% increase in the number of school years completed.28
Several more recent studies using different longitudinal data sets (the PSID, the NLSY, and Children of the NLSY) also find that poverty status has a small negative impact on high school graduation and years of schooling obtained. Much of the observed relationship between income and schooling appears to be related to a number of confounding factors such as parental education, family structure, and neighborhood characteristics.28-30 Some of these studies suggest that the components of income (for example, AFDC) and the way income is measured (number of years in poverty versus annual family income or the ratio of income to the poverty threshold) may lead to somewhat different conclusions. But all the studies suggest that, after controlling for many appropriate confounding variables, the effects of poverty per se on school achievement are likely to be statistically significant, yet small. Based on the results of one study, the authors estimated that, if poverty were eliminated for all children, mean years of schooling for all children would increase by only 0.3% (less than half a month).30
Why do not the apparently strong effects of parental income on cognitive abilities and school achievement in the early childhood years translate into larger effects on completed schooling? One possible reason is that extra-familial environments (for example, schools and neighborhoods) begin to matter as much or more for children than family conditions once children reach school age. A second possible reason is that school-related achievement depends on both ability and behavior. As is discussed in the Emotional and Behavioral Outcomes section, children's behavioral problems, measured either before or after the transition into school, are not very sensitive to parental income differences.
A third, and potentially crucial, reason concerns the timing of economic deprivation. Few studies measure income from early childhood to adolescence, so there is no way to know whether poverty early in childhood has noteworthy effects on later outcomes such as school completion. Because family income varies over time,31 income measured during adolescence, or even middle childhood, may not reflect income in early childhood. A recent study that attempted to evaluate how the timing of income might affect completed schooling found that family income averaged from birth to age 5 had a much more powerful effect on the number of school years a child completes than does family income measured either between ages 5 and 10 or between ages 11 and 15.7 For low-income children, a $10,000 increase in mean family income between birth and age 5 was associated with nearly a full-year increase in completed schooling. Similar increments to family income later in childhood had no significant impact, suggesting that income may indeed be an important determinant of completed schooling but that only income during the early childhood years matters.Emotional and Behavioral Outcomes
Poor children suffer from emotional and behavioral problems more frequently than do nonpoor children (see Table 1). Emotional outcomes are often grouped along two dimensions: externalizing behaviors including aggression, fighting, and acting out, and internalizing behaviors such as anxiety, social withdrawal, and depression. Data regarding emotional outcomes are based on parental and teacher reports. This section reviews studies that distinguish between the effects of long- and short-term poverty on emotional outcomes of children at different ages.
One study of low birth weight five-year-olds using the IHDP data set found that children in persistently poor families had more internalizing and externalizing behavior problems than children who had never been poor. The analysis controlled for maternal education and family structure and defined long-term poverty as income below the poverty threshold for each of four consecutive years. Short-term poverty (defined as poor in at least one of four years) was also associated with more behavioral problems, though the effects were not as large as those for persistent poverty.6
Two different studies using the NLSY report findings consistent with those of the IHDP study. Both found persistent poverty to be a significant predictor of some behavioral problems.26,32 One study used data from the 1986 NLSY and found that for four- to eight-year-olds persistent poverty (defined as a specific percentage of years of life during which the child lived below the poverty level) was positively related to the presence of internalizing symptoms (such as dependence, anxiety, and unhappiness) even after controlling for current poverty status, mother's age, education, and marital status. In contrast, current poverty (defined by current family income below the poverty line) but not persistent poverty was associated with more externalizing problems (such as hyperactivity, peer conflict, and headstrong behavior).32
The second study used NLSY data from 1978–1991 and analyzed children ages 3 to 11. On average children living in long-term poverty (defined by the ratio of family income to the poverty level averaged over 13 years) ranked three to seven percentile points higher (indicating more problems) on a behavior problem index than children with incomes above the poverty line. After controlling for a range of factors including mother's characteristics, nutrition, and infant health behaviors, the difference remained though it dropped in magnitude. This study also found that children who experienced one year of poverty had more behavioral problems than children who had lived in long-term poverty.26
The above studies demonstrate that problematic emotional outcomes are associated with family poverty. However, it is important to note that the effects of poverty on emotional outcomes are not as large as those found in cognitive outcomes. Also these studies do not show that children in long-term poverty experience emotional problems with greater frequency or of the same type as children who experience only short-term poverty. These studies analyzed data for young children. Few studies have examined the link between emotional outcomes and poverty for adolescents. One small study of 7th- to 10th-graders in the rural Midwest did not find a statistically significant relationship between poverty and emotional problems, either internalizing or externalizing.33 Self-reporting by the adolescents rather than maternal reporting, as used in the data sets on younger children, may account for the differences found in the effects of income on emotional outcomes in this study as compared with the previously reviewed research. It may also be that younger children are more affected by poverty than older children.
These findings point to the need for further research to improve understanding of the link between income and children's emotional outcomes.Teenage Out-of-Wedlock Childbearing
The negative consequences for both mothers and children associated with births to unwed teen mothers make it a source of policy concern.34 Although the rate of out-of-wedlock births among poor teens is almost three times as high as the rate among those from nonpoor families (see Table 1), the literature on linkages between family income and out-of-wedlock childbearing is not conclusive. A recent review of the evidence put it this way: "[P]arental income is negative and usually, but not always, significant. . . . The few reports of the quantitative effects of simulated changes in variables suggest that decreases in parental income . . . will lead to small increases in the probability that teen girls will experience a nonmarital birth."28
A recent study, which used data from the PSID to investigate factors in teen out-of-wedlock births, found that variations in income around the poverty threshold were not predictive of a teenage birth but that the probability of a teenager's having an out-of-wedlock birth declined significantly at family income levels above twice the poverty threshold.35 The duration and timing of poverty had no effect on the probability of a teen out-of-wedlock birth. These findings are somewhat different from those reported for cognitive outcomes and school achievement. In the case of cognitive outcomes for young children, the variation in income mattered most to children at very low levels of income; for school achievement, the timing and duration of poverty seemed to have important differential effects on outcomes.
Why should poverty status matter more for schooling than for childbearing? This difference is consistent with the more general result that parental income appears more strongly linked with ability and achievement than with behavior. The factors influencing teenage out-of-wedlock childbearing are less well understood than the factors influencing schooling completion: interventions have generally been much less successful in altering teen birthrates than in keeping teens in school.36,37